Documentos de Académico
Documentos de Profesional
Documentos de Cultura
Villanueva
Nuevas Metodologías en la Evaluación de la Peligrosidad Nuevas Metodologías en la Evaluación de la Peligrosidad
y el Riesgo por Avenidas en Cuencas de Montaña y el Riesgo por Avenidas en Cuencas de Montaña
New Methods in the Flash Flood Hazard and Risk Analysis in Mountain Basins New Methods in the Flash Flood Hazard and Risk Analysis in Mountain Basins
New Methods in the Flash Flood Hazard and Risk Analysis in Mountain Basins
El objetivo de este trabajo de investigación ha sido desarrollar nuevas metodologías que permitan mejorar
la evaluación del riesgo por avenidas en zonas montañosas. En primer lugar, se han integrado y combinado
diferentes fuentes de datos (instrumentales, históricas o documentales y dendrogeomorfológicas) para
recopilar y caracterizar las avenidas pasadas, e incorporar esta información al análisis de frecuencia.
A continuación, se ha incluido la carga leñosa en los análisis de peligrosidad y riesgo por avenidas.
A pesar de la importancia que tiene su presencia en la dinámica fluvial, este tipo de carga raramente ha
sido tenida en cuenta en los estudios de inundaciones. Para poder incluirla, se ha desarrollado un modelo
The main objective of this research paper is therefore to develop new methods to improve flood risk evaluation
in mountain areas. First, the data from different instrumental, historical, documentary and dendrogeomorphologilcal
sources was integrated to compile and characterize past flash-floods and include this combined information in frequency
analysis. The woody load was then included in the flash-flood hazard and risk analyses. In spite of the significant impact
of the presence of this type of load on fluvial dynamics, it is seldom taken into account in flood studies. To allow this
factor to be included, a numerical model was developed to simulate the woody material transport in conjunction with
the hydrodynamics. Finally, the uncertainty at each stage of the risk analysis (flow rate estimate, frequency analysis,
hydrodynamic simulation and vulnerability assessment) was evaluated from a probabilistic viewpoint.This methodology
was implemented in various basins or river reaches in the Sierra de Gredos (Spanish Central System).
Directores:
Dr. José María Bodoque del Pozo Dr. Andrés Díez Herrero
Universidad de Castilla-La Mancha Instituto Geológico y Minero de España
UNIVER
RSIDAD CO
OMPLUTEN NSE DE MAD
DRID
FACULTAD DEE CIENCIASS GEOLÓG
GICAS
DEPARTAMEENTO DE GGEODINÁM MICA
MEMORIA
A PARA OP
PTAR A LA OBTENCIÓ
ÓN DEL TÍTULO DE DOCTOR
PH
HD DISSERTA
ATION
“NUE
EVAS MET
TODOLOG GÍAS EN LA
L EVALU
UACIÓN DE LA
P
PELIGROSIDAD Y EL RIESG
GO POR AVENIDAS
A S EN CUE
ENCAS DE
E
M
MONTAÑA A”
“NEW MET
THODS FOR
R THE ANALLYSIS OF FLA
ASH FLOOD HAZARD AAND RISK IN
N
MOUUNTAIN BASSINS”
Virgin
nia Ruiz Villanueva
Dire
ectores:
Dr. JJosé M. Bo
odoque de
el Pozo Dr. And
drés Díez H
Herrero
I. AGRADECIMIENTOS .............................................................................................. I
II. RESUMEN .................................................................................................................... V
III. ABSTRACT ............................................................................................................... IX
IV. ESTRUCTURA DE LA MEMORIA.................................................................... XI
A mi abuelo
AGRADECIMIENTOS
I. AGRADECIMIENTOS
Desde un punto de vista institucional este trabajo se enmarca dentro de los proyectos
de investigación del Plan Nacional de I+D+i Dendro-Avenidas (CGL2007-62063-HID) y
MAS Dendro-Avenidas (CGL2010-19274) y ha estado financiado por una beca pre-
doctoral FPI (BES-2008-003725) del Ministerio de Educación y Ciencia, y por el Instituto
Geológico y Minero de España (IGME). También se ha contado con la colaboración de la
Agencia Estatal de Meteorología (AEMET), la Confederación Hidrográfica del Tajo, la
Junta de Castilla y León y los Ayuntamientos de Navaluenga, Arenas de San Pedro y
Guisando.
En primer lugar quiero agradecer a los directores de esta tesis, Andrés Díez Herrero y
José M. Bodoque del Pozo, el que un día, hace ahora 5 años, decidieran darme la
oportunidad de emprender este proyecto con ellos, depositando en mí toda su confianza.
En estos años me han acompañado a lo largo de todo el camino, ofreciéndome siempre su
conocimiento, su apoyo, su visión crítica y sus consejos. A ellos les debo el afán de
superación, entusiasmo y dedicación, y también la constancia y tozudez, con la que he
tratado de realizar siempre mi trabajo.
A continuación quiero recordar a todos mis compañeros del IGME. En especial, Juan
A. Ballesteros, compañero de despacho, de campo y de proyectos, del que he aprendido
mucho y con quien he compartido múltiples experiencias y espero seguir haciéndolo en el
futuro. Eulogio Pardo, Carolina Guardiola y Mario Hernández que con su colaboración
han contribuido notablemente al desarrollo de este trabajo. Además, en estos años, he
tenido la gran suerte de coincidir con Marta Díez, Teresa Herrero, Rut Sánchez, Laura
Becerril, Jorge Buzzi, Manuela Chamizo, Alberto Jiménez, Héctor Aguilera y el resto de
PIFs (personal investigador en formación), y compartir con ellos jornadas, congresos y
otras alegrías. Ángel Prieto, Carlos Lorenzo, Margarita Sanabria y Gerardo Herrera, a
quienes he acudido en innumerables ocasiones solicitando ayuda y consejos y siempre me
han resuelto las dudas con una sonrisa y una mejor predisposición; además de ser excelentes
compañeros de tupper y de buena conversación. Miguel A. Rodríguez Pascua y Raúl Pérez,
compañeros del Área de Riesgos Geológicos, y de cafés, quienes me descubrieron una nueva
manera de mirar los restos arqueológicos. Luis Carcavilla, me introdujo en el mundo de la
divulgación científica y el patrimonio geológico, gracias por hacerme partícipe de los
Geolodías, y por todo lo aprendido en Sobrarbe.
También quiero trasladar mi gratitud a todos los miembros de los proyectos Dendro-
Avenidas y MAS Dendro-Avenidas, que han seguido muy de cerca, y han participado de
forma activa, en el desarrollo de este trabajo. En particular, Miguel Ángel Éguibar (UPV),
que además me recibió en la Universidad Politécnica de Valencia para poner a mi
disposición sus conocimientos de hidráulica; también Mar Génova (UPM), por su ayuda
con la botánica y los análisis de muestras; José A. Fernández Yuste (UPM) y Carolina
Martínez (UPM) por sus gratificantes consejos y sugerencias; y Juan M. Rubiales (UPM).
También quiero agradecer a Ignacio Gutiérrez, Luis Fernández y Luis Barca su
colaboración en los trabajos de topografía, su predisposición a ayudar y facilitar lo más
posible esta tarea.
Sin duda, piezas indispensables en el desarrollo de este trabajo han sido las estancias
breves de investigación realizadas en estos años y financiadas por el Ministerio de Ciencia e
Innovación. Sin estas estancias, y sin la colaboración de las personas que se ofrecieron
generosamente a incorporarme a sus equipos de investigación como un miembro más, esta
tesis no hubiera sido igual. Además, en el plano personal también han supuesto unas
experiencias muy enriquecedoras y por ello guardaré siempre un recuerdo muy especial.
evento. Quiero resaltar la acogida, el apoyo y el compañerismo de todos los miembros del
departamento, en especial a mis compañeros de despacho, Davide Zocatelli y Francesco
Marra, pero también a Daniele Penna, Paolo Tarolli, Nicola Mantesse, Alberto de Luca,
Alberto Gobbi, Enrico Pozza, etc. Además esta experiencia también me permitió colaborar
con Lorenzo Marchi y Eric Gaume que han actuado como evaluadores de esta tesis.
A mis amig@s (la cuchipandi, no os nombro a tod@s porque como la familia no para
de crecer, no terminaría…). Con ellos lo he compartido todo, y aguantan mi insistente
esfuerzo por enseñarles a mirar el paisaje con ojos de geólogo; gracias por la paciencia, el
II. RESUMEN
Los resultados han demostrado también que el transporte de material leñoso influye
significativamente en las consecuencias de una avenida, y por tanto, este proceso no puede
ser obviado. Con la obtención de las curvas `calado-obstrucción´ es posible interpretar el
comportamiento hidráulico y la sobrelevación que producirá una determinada reducción
de la sección, debida al transporte y depósito de la carga leñosa. Asimismo, la amplificación
de las consecuencias de la avenida por causa de esta obstrucción puede definirse mediante el
nuevo concepto de `periodo de retorno equivalente´.
Para entender los mecanismos de incorporación de este material leñoso a los cauces y
poder realizar estimaciones de volúmenes, es necesario definir las áreas contribuyentes y
realizar un análisis de la vegetación. Con el fin de llevar a cabo estos análisis, tanto a escala
regional como de cuenca, los sistemas de información geográfica junto con la construcción
de matrices de lógica difusa, constituyen una herramienta eficaz cuando se dispone de los
datos necesarios. De esta forma, es posible cuantificar el volumen de material que puede
llegar a un tramo de río y que puede ser un riesgo potencial por su transporte durante una
avenida. Para analizar este transporte y su influencia en la hidrodinámica se ha desarrollado
un modelo numérico integrado en un modelo hidráulico bidimensional. El desarrollo de
este modelo supone una aportación novedosa ya que no existía hasta el momento una
herramienta similar que permitiera simular el movimiento de material leñoso junto con la
hidrodinámica. La validación mediante ensayos con modelos físicos de laboratorio, y su
aplicación en dos zonas de estudio ha puesto de manifiesto su utilidad en los análisis de
peligrosidad y riesgo por avenidas. El modelo permite reconstruir avenidas con abundante
carga leñosa, simulando los diferentes regímenes de transporte y reproduce los patrones de
transporte y depósito del material. Por todo ello, este modelo puede ser también empleado
en otros estudios de la dinámica fluvial, de restauración o ecológicos.
III. ABSTRACT
The main objective of this research paper is therefore to develop new methods to
improve flood risk evaluation in mountain areas. First, the data from different
instrumental, historical, documentary and dendrogeomorphological sources was integrated
to compile and characterize past flash-floods and include this combined information in
frequency analysis. The woody load was then included in the flash-flood hazard and risk
analyses. In spite of the significant impact of the presence of this type of load on fluvial
dynamics, it is seldom taken into account in flood studies. To allow this factor to be
included, a numerical model was developed to simulate the woody material transport in
conjunction with the hydrodynamics. Finally, the uncertainty at each stage of the risk
analysis (flow rate estimate, frequency analysis, hydrodynamic simulation and vulnerability
assessment) was evaluated from a probabilistic viewpoint. This methodology was
implemented in various basins or river reaches in the Sierra de Gredos (Spanish Central
System).
reconstructed events into the frequency analysis reduced uncertainty in the quantile
estimates. However, depending on how these data are included in the analysis (as censored
data, value intervals or data with negligible uncertainty), the resulting quantiles may vary.
This variation must therefore be taken into account when using these data in risk analysis.
The results have also shown that the transport of woody material has a significant
effect on the outcome of a flash flood, and therefore this process cannot be ignored.
Obtaining the clogging curves allows the interpretation of the hydraulic behaviour and
water surface elevation produced by a specific cross-sectional reduction due to the
transport and deposition of the wood load. The increased effect of the flash flood because
of this obstruction can also be defined using the new concept of ‘equivalent return period’.
In conclusion, a risk analysis that combines different data, the use of indirect
methods to reconstruct events, the inclusion of woody loads, a probabilistic approach and
scenario generation leads to improvements in the study of flash floods in mountain areas.
Although the methodology proposed in this paper has been applied to a specific sector of
the Sierra de Gredos, it can be easily extrapolated to other mountain areas.
1. INTRODUCCIÓN, ANTECEDENTES
Y OBJETIVOS
1.1 Introducción
1.1.1. Características de las cuencas y ríos de montaña
1.1.2. Avenidas e inundaciones en zonas montañosas
1.1.3. Fuentes de datos en el estudio de avenidas
1.2. Antecedentes
1.2.1. Antecedentes en los estudios de avenidas en zonas de
montaña
1.2.2. Antecedentes en España
1.1. Introducción
Los ríos sensu lato están presentes en la mayoría de ambientes y paisajes de la Tierra y
han sido estudiados por científicos de forma sistemática a lo largo de la historia de la
humanidad (Plinio, Leonardo Da Vinci, Humboldt, etc.). De las diferentes tipologías de
ríos descritas en la bibliografía, este trabajo de investigación aborda el estudio de ríos o
arroyos de montaña, cuyas características particulares los diferencian de los ríos que
discurren por las llanuras o zonas bajas (Surrel, 1841; Dana, 1850; Shaler, 1891). A pesar
de que una región montañosa pueda asociarse de forma intuitiva con una determinada
altitud y/o pendiente, no existe un acuerdo generalizado en su definición. El programa de
medio ambiente de las Naciones Unidas (UNEP-WCMC) da ciertos criterios para
delimitar regiones montañosas basados en estos dos parámetros (altitud y pendiente); al
igual que la Comisión Europea de análisis de regiones montañosas en Europa (European
Commission contract No 2002.CE.16.0.AT.136 en 2004) que además añade como
criterio la combinación de ambos (Directiva 75/268/CEE y Reglamento CE-1257/1999).
Figure 1.1: Global Topography and Bathymetry. Mountain areas colored in brown.
(Wavemetrics, 2012).
Figura 1.1: Topografía y batimetría mundial. En colores marrones las zona de montaña.
Longitud y latitud en grados (Wavemetrics, 2012).
Además, son las zonas montañosas las que acumulan más de la mitad de los recursos
de agua dulce de la Tierra (Viviroli et al, 2003; Weingartner et al., 2003), pudiendo
representar a escala regional el caudal generado en estas zonas, el 95% del caudal total de
una cuenca (Liniger et al., 1998; Viviroli y Weingartner, 2004). Sin embargo,
tradicionalmente han recibido menos atención que las cuencas de los grandes ríos de las
regiones más llanas (Montgomery and Buffington, 1997), debido principalmente a la
dificultad de acceso y explotación de recursos como la pesca (Benda et al., 2005). Aunque
esta tendencia parece haber cambiado en las últimas décadas, al menos desde el punto de
vista medioambiental o en los estudios de cambio climático. En 1992, en la Cumbre sobre
la Tierra (Conferencia de las Naciones Unidas sobre el Medio Ambiente y el Desarrollo)
de Río de Janeiro, se incluyó un Capítulo 13 titulado «Ordenación de los ecosistemas
frágiles: desarrollo sostenible de las zonas de montaña» en el Programa 21, principal
documento de la reunión. La Comisión de las Naciones Unidas sobre Desarrollo
Sostenible designó después a la FAO como coordinador de tareas para este capítulo de las
montañas. Una colaboración de amplitud sin precedentes entre organismos de las Naciones
Unidas, gobiernos nacionales, organizaciones internacionales, ONG e instituciones de
investigación, hizo posible presentar en el período extraordinario de sesiones de la
Asamblea General de las Naciones Unidas, Cumbre sobre la Tierra + 5, en 1997, un
extenso informe titulado Mountains of the world: a global priority (Montañas del mundo:
una prioridad mundial; Messerli e Ives, 1997), y un documento complementario sobre
políticas, Mountains of the world: challenges for the twenty-first century (Montañas del
mundo: desafíos para el siglo XXI; Bisaz et al., 1997).
Figure 1.2: Stylized and simplified sketch of the three primary zones of a mountain
drainage basin based on hydrogeomorphic processes.
Figura 1.2: Esquema idealizado y simplificado de las diferentes zonas en que puede dividirse
una cuenca hidrológica de montaña en función de los procesos hidrogeomorfológicos.
De esta forma, otros autores proponen considerar el río y el ecosistema que lleva
asociado como un geosistema lineal, en el que los procesos que tienen lugar aguas abajo
están condicionados por los que suceden aguas arriba (Garzón-Heydt, 1999).
Figure 1.3: Differences in disturbance processes among colluvial. confined alluvial, and
unconfined alluvial channels in mountain drainage basins (based on Montgomery and
Buffington, 1997).
Figura 1.3: Procesos y morfologías asociadas a los diferentes canales de una cuenca de
montaña: canales coluviales, confinados aluviales y aluviales no confinados (Tomado y
traducido de Montgomery y Buffington, 1997).
Teóricamente, los ríos de alta-media pendiente que presentan una elevada capacidad
de transporte no deberían acumular grandes depósitos de sedimento (Montgomery et al.,
1996). Sin embargo, la propia presencia de bloques y bolos actúa como obstrucción,
produciendo una reducción local de la pendiente y de la velocidad del flujo, incrementando
la retención de sedimento (Benda et al., 2005). También los deslizamientos y flujos que
tienen lugar en las partes altas aportan cantidades elevadas de material a los cauces.
Los bosques modernos cubren casi un tercio de la superficie de la Tierra (Atjay et al.
1979), pero la madera ha estado presente en los arroyos y ríos durante más de 400 millones
de años, aunque su peso relativo en los ecosistemas de agua dulce ha variado
sustancialmente a través del tiempo geológico (Schumm 1968; Cotter, 1978; Montgomery
et al., 2003; Fig. 1.5). Tradicionalmente la morfología de los cauces se ha relacionado
fundamentalmente con la carga de sedimentos; sin embargo, la presencia de vegetación y
material leñoso también ha sido determinante. La evolución de las plantas coincide con
una mayor abundancia de depósitos de ríos meandriformes en el registro geológico (hasta
ese momento los depósitos más abundantes eran de tipo braided; Schumm 1968; Cotter,
1978). El desarrollo de un patrón de meandros requiere la cohesión de las barras y bancos
de orilla (Schumm 1963); esta cohesión pudo ser proporcionada por la vegetación de
ribera. Esta teoría se confirmaría con la datación de abundantes restos vegetales en
depósitos meandriformes del Carbonífero, y pudo ser así al menos hasta el Pérmico-
Triásico, momento en el que vuelven a ser más abundantes los depósitos de tipo trenzado,
coincidiendo con una reducción drástica de la vegetación terrestre (Ward et al., 2000). El
desarrollo de las coníferas y en general de la cubierta forestal evolucionó durante períodos
interglaciares cálidos en el Pleistoceno (Porter, 1983; Velichko 1984; Huntley y Webb
1988), llegando a su óptimo en Europa en el Holoceno (Wright, 1983; Velichko 1984).
Figure 1.5: Hypothesized relative influence of wood on the world’s rivers throughout
geologic time (from Montgomery et al., 2003).
Figura 1.5: Influencia relativa (hipotética) de la presencia de material leñoso en los ríos a lo
largo del tiempo geológico (Tomado y traducido de Montgomery et al., 2003).
Figure 1.6. (A) Upstream view of Santa María stream crossing Canturral area in La Adrada
(Ávila, Gredos mountain range, Spain, January 2012). The boulder size is decimetre.
Figura 1.6. (A) Vista hacia aguas arriba de la Garganta Santa María en La Adrada a su paso
por la urbanización Canturral (Ávila, Sierra de Gredos, enero 2012). El tamaño de los
bloques es decimétrico.
Figure 1.6: (B) Woody jam with tilted trees in the flow direction of the Teverga River in
Tuñón (Asturias, Picos de Europa mountain range, October 2011).
Figura 1.6: (B) Acumulación de material leñoso entre árboles inclinados en la dirección de
la corriente en el río Teverga en las cercanías de Tuñón (Asturias, Picos de Europa, octubre
2011).
Figure 1.6: (C) Pull up trees in a plane bed reach of the College Burn River in the North of
NewCastle (England, November 2009).
Figure 1.6: (D) Wood and blocks and boulders in the Kaligandaki River in the Annapurna
region (Nepal, Hymalaya, February 2012).
Figura 1.6: Restos de árboles entre los cantos y bloques en un tramo del río Kaligandaki en
la región de los Annapurna (Nepal, Himalaya, febrero 2012)
(Georgakakos, 1992; Collier, 2007). Son fenómenos muy localizados espacialmente que
tienen lugar en cuencas de pocos cientos de kilómetros cuadrados o menos y con tiempos
de respuesta muy cortos (pocas horas o menos; O’Connor y Costa, 2004; Borga et al., 2007
y 2008).
Figure 1.8: Headwaters of the Tajo River after an important flood, large quantities of wood
were mobilized and deposited in jams (1920-30´). Picture taken from the collection of
Complutense University of Madrid.
Figura 1.8: Río Tajo en cabecera, tras una crecida importante, que movilizó grandes
cantidades de madera y la colocó formando una represa (años 1920-30).Fotografía de la
colección fotográfica histórica del Departamento de Geodinámica de la Universidad
Complutense de Madrid.
Los datos que se han empleado para completar y/o extender los registros
instrumentales de avenidas son de diversa naturaleza, y su uso es la base de la
paleohidrología (Baker, 2002; Benito y Thorndycraft, 2005); son los llamados indicadores
de paleonivel (palaestage indicators, PSI).
circulantes mediante modelos hidráulicos (Lang et al., 2004) y asignarles una determinada
probabilidad de ocurrencia, introduciéndolos como datos complementarios en el análisis
estadístico de caudales procedentes del registro instrumental (Francés, 2004; Payrastre et
al., 2005; Marquínez et al., 2006; Brázdil et al., 2006).
paredes de cañones y terrazas permiten acotar las fechas de fases de incisión o de abrasión
vinculadas a las avenidas, y por lo tanto datar éstas y sus límites, a partir de los que se
infieren las magnitudes (Gregory, 1976).
El uso de todas estas fuentes de datos se basa en la premisa de que la geometría del
canal es conocida, y por tanto pueden ser combinados con métodos hidráulicos como
modelos numéricos unidimensionales o bidimensionales (Web y Jarret, 2002; Rico et al.,
2001; Benito et al., 2004; Ballesteros et al., 2011) para obtener parámetros como caudal,
altura de lámina de agua, velocidad o energía de la corriente.
Todas estas fuentes de datos pueden combinarse y emplearse para tratar de entender
las variaciones en magnitud y frecuencia de avenidas en los últimos cientos o miles de años
y así tratar de inferir las causas o factores climáticos responsables de esas variaciones (Ely et
al., 1993; Thorndycraft et al., 2002).
1.2. Antecedentes
Tras la avenida súbita ocurrida en 1976 en Colorado (Estados Unidos) en el río Gran
Thompson (Big Thompson River), donde murieron 144 personas (USGS, 1979), algunos
científicos estadounidenses pusieron de manifiesto que los métodos tradicionales, usados
hasta ese momento en el estudio de grandes inundaciones en tramos de ríos medios o bajos,
no eran igual de eficaces para el estudio de avenidas en ríos de montaña (Jarret y Costa,
1983; Jarret, 1987). Como resultado, estos autores propusieron metodologías diferentes
para mejorar los estudios de avenidas. Por un lado, propusieron separar la población de
datos de caudales máximos del registro de estaciones de aforo en función de su origen
meteorológico. El objetivo era diferenciar aquellos eventos originados por lluvias intensas
de los originados por deshielo, y de esta forma entender mejor la relación causa-efecto y
frecuencia de ambos procesos (Jarrett y Waythomas, 1995).
Además analizaron los valores propuestos por otros científicos (Chow, 1959; Barnas,
1967) para cuantificar la rugosidad del lecho, y observaron que estos valores (que habían
sido propuestos principalmente en tramos medios y bajos de ríos) no eran adecuados para
ríos de montaña. Así, propusieron utilizar la siguiente ecuación para predecir los valores
más adecuados para estos ríos:
Estos mismos autores (Costa, 1978; Baker, 1983; Costa, 1986; Jarret, 1987)
emplearon diferentes métodos paleohidrológicos para extender y completar el registro de
aforos en cuencas de montaña. Y propusieron utilizarlos junto con los datos instrumentales
para estimar la frecuencia de avenidas en cuencas no aforadas a través de análisis regional.
De las diferentes técnicas empleadas habitualmente en paleohidrología analizaron los
resultados con el objetivo de cuantificar los posibles errores cometidos. Jarret (1987)
concluye que el uso del método área-pendiente (slope-area) lleva asociado un error de hasta
el 75%, y propone como mejor método el conocido como calado-crítico (critical-depth).
Aunque otros autores afirman que es posible obtener medidas precisas utilizando ambos
métodos de forma adecuada, llegando solo a un 10% de error en la estimación (Benson y
Dalrymple, 1967); también en esta línea, Costa (1983) demostró que las reconstrucciones
de caudales llevadas a cabo a partir del tamaño de las partículas transportadas por el flujo
llevaba asociado un 28% de error; y Webb y Jarret (2002) en una revisión posterior de
diversos métodos paleohidrológicos estimaron que el error podía llegar al 30%. Con el
objetivo de reducir las incertidumbres en las estimaciones, Williams y Costa (1988)
describen con detalle las observaciones geomorfológicas que deben tenerse en cuenta para
estudiar una inundación pasada.
Sin embargo, las incertidumbres y los errores no solo están relacionados con las
estimaciones indirectas, sino también con los datos instrumentales; Pelletier (1988) y Jarret
(1990) resaltaron las dificultades de diseñar las redes de medida en zonas de montaña y
observaron que los datos hidrometeorologicos registrados también podían llevar asociados
importantes errores. Por ejemplo, Spreafico et al. (1985, 2002) analizaron los errores
relacionados con medidas directas de caudal y velocidad de la corriente durante avenidas en
ríos de montaña, y concluyeron que éstos pueden llegar al 17%, cuando en tramos medios o
bajos de ríos estos errores pueden ser del 2-3%.
Estos errores están asociados a las características particulares de los flujos en estos ríos,
tal y como se ha explicado anteriormente. Grant et al. (1997) investigaron de forma
analítica estas características, estudiando principalmente los cambios de régimen y la
resistencia al flujo, y observaron que el régimen crítico no se conserva más que en tramos de
muy corta longitud (Trieste, 1992; Tinkler, 1997; Magirl et al., 2009). Aunque, por el
contrario, estudios más recientes, como los llevados a cabo por Comiti et al. (2009)
concluyeron que, al menos para tramos de escalón-poza es posible asumir condiciones de
flujo crítico en los estudios post-evento o paloehidrológicos.
Estudiando además los perfiles verticales de velocidad del flujo y junto con los
trabajos llevados a cabo por Marchand et al. (1984), Jarret (1990) definió estos perfiles
como no logarítmicos, con formas sinuosas, con velocidades mucho menores cerca del
lecho, debido a la presencia de bloques y cantos en el lecho, y velocidades mayores cerca de
la superficie. Años después, en 2007, Wilcox y Wohl observaron esta misma estructura
Los efectos que las avenidas producen en la morfología de los ríos y en el transporte
de sedimentos ha sido objeto de investigación desde trabajos pioneros como los llevados a
cabo por Wolman y Miller (1960) o Wolman y Gerson (1978); o los que se fueron
desarrollando después (Gupta, 1975, 1988; Baker, 1984; Baker and Pickup, 1987; Miller,
1990; Wohl, 1992; Grimm et al., 1995). Actualmente estos estudios se pueden agrupar en
dos grandes líneas: aquéllos que estudian la producción de sedimentos y su transporte
durante avenidas y los que analizan su distribución morfológica y su depósito. Estos
trabajos han puesto de manifiesto que los efectos geomorfológicos de una avenida no
siempre se relacionan con la magnitud de la misma. Así, inundaciones de elevada magnitud
pueden causar impactos leves (geomorfológicamente hablando), o por el contrario avenidas
de menor magnitud pueden llevar asociados importantes impactos geomorfológicos. E
incluso, para una misma avenida, los efectos pueden variar espacial y temporalmente
(Baker, 1977; Gupta, 1983; Nanson, 1986; Baker y Pickup, 1987; Osterkamp y Costa,
1987; Lewin, 1989; Pickup, 1991; Costa y O’Connor, 1995; Dury, 1973; Gomez et al.,
1995; Kochel, 1988; Magilligan et al., 1998; Wolman y Eiler, 1958; Butler y Malanson,
1993; Costa, 1974; Magilligan, 1992; Miller, 1990; Nanson, 1986). Actualmente, estos
estudios se han visto potenciados por el uso cada vez mas frecuente de la tecnología LIDAR
(Cavalli et al., 2008; Trevisani et al., 2009; Vianello et al., 2009; David et al., 2010).
papel de la turbulencia en la disipación de la energía del flujo (Carollo et al., 2005), y por el
hecho de que la movilización se produce a partir de unos umbrales de movimiento
vinculados a eventos de baja frecuencia (como se ha explicado en el epígrafe 1.1.1). Por
tanto, la movilidad de material se incrementa considerablemente durante las avenidas
(Lenzi et al., 2006; Mao y Lenzi, 2007; Mao et al., 2008; Turowski et al., 2009; Yager,
2006) y puede ser determinante en las consecuencias de éstas, junto con el otro
componente habitual en los ríos de montaña, el material leñoso.
Por último, en relación a los análisis de riesgo por avenidas, el parámetro que suele
usarse más frecuentemente como indicador de los daños potenciales es el calado o altura de
lámina de agua (Kaplan y Garrick, 1981, Mileti, 1999). Como observaron Grigg y Helweg
(1975), en edificios o estructuras de similares características, una inundación con un calado
determinado producirá daños o pérdidas económicas similares en ellos. Por tanto, las
pérdidas directas causadas por una avenida dependerán, entre otros factores, del calado y
del uso que tenga un determinado edificio o estructura (Smith, 1994; Penning-Rowsell et
al., 1994; USACE, 1996, Merz et al., 2004). Además las características propias de los
edificios o estructuras también serán importantes en cuanto a los daños que se puedan
producir (Kreibich et al., 2009).
fallecidos, la rotura del embalse de Tous en 1982 con 38 victimas, las inundaciones del País
Vasco en 1983 con 40 muertos, los episodios ocurridos en la vertiente mediterránea de
1989 con 42 fallecidos, o la catástrofe del camping de Biescas en 1996 con 86 victimas
mortales (Francés et al., 2000; Ayala-Carcedo, 2002; Fig. 1.9). Recientemente, habría que
mencionar las inundaciones ocurridas en Asturias en febrero de 2012, o las avenidas que
afectaron a Murcia, Almería, Málaga, Valencia y Aragón en octubre de ese mismo año, con
al menos una docena de fallecidos y más de 120 millones de euros en pérdidas económicas.
Figure 1.9: Sediment and wood deposits after the 1996 flash flood in the Arás stream in
Biescas (Spain), Las Nieves Camp Site. 86 people died (picture taken by Carlos Sancho and
courtesy of Virginia Rubio).
Desde una perspectiva de las causas meteorológicas, o los cambios meteorológicos y/o
climáticos asociados a inundaciones se pueden citar los trabajos realizados por Llasat y
Barnolas (UB-SMC), o Rodrigo (UAL), Morales y Ortega (UVA), entre otros. Y con el
mismo objetivo, es decir, analizar las causas meteorológicas, pero utilizando para ello datos
de Radar meteorológico y modelización hidrológica, cabe destacar el grupo de investigación
de Sempere-Torres, Berenguer y Velasco (CRAHI-UPC).
Además de los trabajos desarrollados por estos y otros grupos de investigación, existen
en la legislación española numerosos instrumentos de ordenación en el ámbito de las
inundaciones (normativa estatal, regional o local). Sin embargo, no hay una normativa
Sin embargo, en ningún caso en los desarrollos normativos se especifica cómo deben
abordarse estos estudios de peligrosidad y/o riesgo en zonas de montaña. A este respecto, se
publicó en 2008 el libro Mapas de peligrosidad por avenidas e inundaciones. Guía para su
elaboración de la serie Riesgos Geológicos del IGME (Díez-Herrero et al., 2008) donde se
recogen los métodos y criterios para la elaboración de diferentes mapas de peligrosidad por
avenidas e inundaciones. En el año 2007, el actual Ministerio de Agricultura, Alimentación
y Medio Ambiente (MAGRAMA, antes Ministerio de Medio Ambiente, y Medio Rural y
Marino) puso en funcionamiento el Sistema Nacional de Cartografía de Zonas Inundables
(SNCZI) como un instrumento de apoyo a la gestión del espacio fluvial, la prevención de
riesgos, la planificación territorial y la transparencia administrativa. Y en 2011 se publicó la
Guía Metodológica para el desarrollo del Sistema Nacional de Cartografía de Zonas
Inundables (MAGRAMA, 2011) que recoge recomendaciones técnicas para la
determinación del Dominio Público Hidráulico y las zonas inundables de acuerdo con el
RD 9/2008 que modificó el Reglamento del Dominio Público Hidráulico y el RD
903/2010 en lo relativo a los mapas de peligrosidad de inundación. En ambos libros se hace
referencia a las zonas montañosas (como sistemas torrenciales) y sus particularidades, que
ponen de manifiesto la necesidad de aplicar métodos diferentes a los habitualmente
empleados en el medio fluvial. Particularmente en lo que se refiere a la estimación de la
magnitud y frecuencia de eventos, se señala como mayor inconveniente la carencia de datos
de aforos y precipitaciones en la propia cuenca, así como la influencia de otros procesos,
principalmente flujos tractivos y mecanismos de transporte en masa. De esta forma, y según
la guía del MAGRAMA, en la zonificación de la peligrosidad en zonas de montaña, a
diferencia de las zonas fluviales, no se plantea la asignación de recurrencias concretas sino
una clasificación cualitativa, definiendo niveles de peligrosidad torrencial alta, media, baja e
indiferenciada; esta clasificación, no exenta de subjetividad, puede resultar en diferentes
fuentes de incertidumbre que afectan a la previsibilidad de los patrones de riesgo (Hoffman
y Hammonds, 1994; Paté-Cornell, 1996).
Hay que tener en cuenta además, que el cambio socioeconómico acontecido en las
últimas décadas ha llevado a un incremento en el desarrollo de los entornos montañosos y
regiones ribereñas (Comiti et al., 2008). Este cambio se refleja en un uso cada vez mayor de
las zonas de montaña con respecto al turismo y a procesos de urbanización. Como
consecuencia, se pueden desencadenar alteraciones en los procesos geomorfológicos
(Hooke, 1999; Cendrero et al., 2006). Pero además, se produce un aumento de los bienes
tangibles e intangibles (debido a la presión antrópica) expuestos a los desastres naturales.
Esto, junto con posibles cambios en los regímenes fluviales, incluyendo los impactos
humanos sobre estos sistemas (Briggs et al., 2006), o los efectos del cambio climático sobre
las tasas de precipitación y los regímenes de escorrentía (Schmocker-Fackel y Naef, 2010),
puede suponer un aumento significativo del riesgo (Mitchell, 2003; Benito, 2006: Fush et
al., 2005).
Al mismo tiempo, la percepción social del riesgo en zonas donde los procesos
catastróficos tienen asociada una baja frecuencia también puede suponer un problema
añadido. Siendo la situación ideal la no ocupación de estas áreas, una vez han sido
ocupadas, se hace necesario estudiar los procesos en profundidad, para mejorar el
conocimiento de los mecanismos de generación y sus consecuencias, y con ello poder
prevenir o mitigar los daños.
Las graves consecuencias de las avenidas ocurridas en los últimos años en diferentes
ambientes montañosos en Europa (Eslovenia, Austria, Italia, Suiza, Polonia, España),
ponen de manifiesto la importancia del estudio de estos procesos. En concreto, durante
estos episodios de avenidas, los efectos en forma de cambios de morfología de los cauces y la
reducción del área de la sección transversal debido a la obstrucción por material
transportado tuvieron como consecuencia una amplificación significativa en la magnitud
del proceso (Diehl, 1997; Lyn et al., 2007; Comiti et al., 2008; Mao y Comiti, 2010;
Mazzorana et al., 2011). Además, quedó demostrado que en algunos casos, los mapas de
peligrosidad existentes resultaron no ser tan fiables como se esperaba (Bezzola y Hegg,
2007). Con el objetivo de mejorar los análisis de riesgos y para apoyar la toma de decisiones
de los gestores, los análisis de inundación en estas zonas de montaña necesitan ser
reformulados para examinar estas cuestiones (Girod y Mieg, 2008) y además deben ser
capaces de hacer frente y tomar en consideración las diferentes fuentes de incertidumbre
(Paté-Cornell, 1996; Merz et al., 2008).
Y por tanto, el quinto objetivo será integrar todo el proceso en un análisis de riesgo
desde una perspectiva probabilista que incorpore las incertidumbres asociadas a todas
las etapas del análisis (análisis de frecuencia, modelación hidrodinámica, curvas calado-
daño, etc.), e integre la carga leñosa mediante la generación de escenarios.
El último objetivo (sexto) de esta investigación será sintetizar todos los análisis
anteriores y, a partir de una discusión de las ventajas y limitaciones de cada fuente de
datos y metodologías proponer una recomendaciones para la mejora de las evaluaciones
de peligrosidad y riesgo por avenidas en cuencas de montaña.
2. MATERIALES Y MÉTODOS
2.1. Análisis territorial
2.1.1 Caracterización geomorfológica
2.1.2. Estudio de la vegetación
2.2. Análisis dendrogeomorfológico
2.6. Modelización hidrodinámica y estimación
de caudales
2.6.1. Modelización unidimensional
2.6.2. Modelización bidimensional
2.7. Estudio de la incorporación de carga
leñosa a los ríos
2. Materiales y Métodos
Hay que destacar que los límites entre estos grupos de métodos son en ocasiones
difusos, puesto que enlazan o se integran unos con otros conformando a su vez una
metodología más compleja. En cada uno de los artículos (apartados) que forman el capítulo
de resultados (Capítulo 4 de esta Memoria) se detalla la metodología específica, las
integraciones o combinaciones de métodos, y los desarrollos metodológicos llevados a cabo
para la consecución de los objetivos. En este capítulo, se describen de forma genérica, en
diferentes secciones (nueve apartados), las distintas disciplinas metodológicas seguidas y
desarrolladas a lo largo de todo el trabajo de investigación (figura 2.1).
Figure 2.1:
2 1: Methodological flowchart
flowchart. Numbers correspond with sections in this chapter.
chapter
CAPÍTULO /CHAPTER 2
Además se realizaron recorridos dde campo para p compro obar la carttografía, intterpretar
y caraacterizar zon
nas problem máticas, evaaluar el grad
do de activiidad de los procesos activos,
a y
determminar la pootencia y oriigen de las fformacionees superficiaales, con esppecial incid
dencia en
aquélllas de origen
n fluvial-colluvial
A escala dee cuenca se definieron los princip pales parámetros o índdices morfom
métricos
relacioonados conn la hidrologgía de aveniddas (resumiidos en la figura 2.3).
VIRGINIA
A RUIZ-VILLANUEEVA 47 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER
C 2
VIRGINIA RUIZ-VILLANUEVA
A 48 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 2
Figure 2.4: (a) Unit bars features found in gravel-bed channels; (b) Bar types and different
parts. cb: head bar; lb: tail bar; cp: main channel; cs: secondary channel; cl: lateral channel.
(From Church and Jones, 1982 and Ramos and Friend, 1982).
Figura 2.4: (a) Tipos de barras en sistemas aluviales de gravas; (b) Esquema de tipos de
barras y sus zonas. cb: cabeza, lb: cola de barra; cp: canal principal o exterior; cs: canal
secundario o interior; cl: canales laterales. (Extraído de Church y Jones, 1982 y Ramos y
Friend, 1982).
El análisis detallado de tamaños de grano, facies u otras formas menores no fue objeto
de este estudio. No obstante, sí se pudo relacionar cualitativamente las meso-y macro-
formas con la energía de la corriente, análogamente a la propuesta de Allen (1983).
Todo ello, tanto geomorfología, como vegetación y otros mapas, como usos y tipos de
suelo fueron integrados en un sistema de información geográfica (ArcGis 9.3; ESRI, 2011),
lo que permitió su análisis espacial conjunto.
Figure 2.5: Dendrogeomorphological evidence useful in the study of torrential floods. 1.)
“candelabrum” growths; 2) Tilted with feet tipped over; 3) Decapitated; 4) Impacted by
sediment load; 5) Branches torn off; 6) Grazing by other falling boles; 7) Erosion on the
trunk; 8) Trunk bends and angles; 9) Sudden narrowing of trunk; 10) Bifurcations; 11)
Bends and angles in branches; 12) Regrowth from buried trunks; 13) Exposed roots with
stripped bark and eroded surface; 14) Roots aloft without contact with substratum; 15)
New roots from buried trunks. (Díez-Herrero , 2009 and 2012).
VIRGINIA RUIZ-VILLANUEVA
A 56 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 2
Para realizar las mediciones de anchura de los anillos se empleó una mesa de medida
semiautomática LINTAB asociada al programa TSAP Win (Time Series Analysis and
Presentation; Rinntech, 2010) y acoplada a un ordenador, que permite una precisión en la
medida de hasta 0,01 mm. Otros equipos que se utilizaron fueron: lupa binocular y fuente
de luz fría (Fig. 2.7).
TESIS DOCTO
Figura 2.7: Trabajo de laboratorio: (a) preparación de muestras; (b) análisis y medida de anchura de anillos de crecimiento; (c)
curvas de crecimiento obtenidas en el análisis; (d) examen visual de secciones; (e) sección con diversas heridas externas e
C
De esta manera es posible inferir las causas de las avenidas, caracterizar los eventos
por su severidad, en función de los daños recogidos en la documentación, así como analizar
su estacionalidad.
Figure 2.8: (a) Example of a news item featuring a historical flood in the study site (taken
from ABC newspaper); (b) 1982 flood plaque placed on a memorial statue in Luis Manuel
Lopez camp site in memorian of a fatality.
Figura 2.8: (a) Ejemplo de noticia en la que se describe un evento de avenida en la zona
de estudio (extraído del periódico ABC). (b) placa ubicada sobre un busto
conmemorativo en el camping Luis Manuel López en recuerdo a una víctima de una
avenida en 1982.
Duración: 1= un día o pocos, 2= varios días a una semana; 3= una semana o más, 0=
duración desconocida.
Con objeto de estudiar la posible relación entre los datos y su variación se utilizaron
diversos análisis: El análisis de componentes principales consiste en describir la variación
producida por la observación de p variables aleatorias, en términos de un conjunto de
nuevas variables correlacionadas entre sí, cada una de las cuales es combinación lineal de las
variables originales. El análisis de correspondencias se utilizó para representar posibles
asociaciones existentes entre variables categóricas, con objeto de establecer patrones o
estructuras en los datos obtenidos con su observación. También con el propósito de
clasificar las observaciones de manera que los datos de estos grupos sean lo más similares
posible y los grupos lo más diferentes posible, se llevó a cabo el análisis de conglomerados
(cluster en la nomenclatura anglosajona) obteniéndose así una clasificación de los datos
multivariante.
Una vez que se dispone de las fechas de los eventos de avenida y de la estimación de
sus caudales, además de las series anuales de caudales máximos diarios para cada estación de
aforos, se procedió a su análisis estadístico con objeto de estimar la distribución de
frecuencias (por ejemplo en los apartados 4.2 y 4.3 de Resultados). Se considera que los
datos de estas series son independientes e idénticamente distribuidos y, en consecuencia,
los caudales máximos se comportan como una variable aleatoria.
De entre los métodos de estimación de cuantiles tres son los más utilizados:
La función GEV es la más recomendada (NERC, 1975). Por tanto es la que más se ha
empleado en este trabajo de investigación.
al., 2009). Además se utilizó la topografía como variable secundaria para las simulaciones
secuenciales gaussianas siguiendo la propuesta de Goovaerts (2000).
Figure 2.9: Variogram and variogram model (modified from Guardiola, 2009).
El método de Monte Carlo permite determinar esa función de distribución y/o algún
aspecto de la misma (media, varianza) generando n muestras y evaluando repetidamente el
estadístico sobre ellas hasta aproximar su distribución mediante la distribución empírica de
valores obtenidos. Para generar estas n muestras se puede utilizar la técnica denominada
Cadenas de Markov. En la teoría de la probabilidad, se conoce como cadena de Markov a
un tipo especial de proceso estocástico discreto en el que la probabilidad de que ocurra un
evento depende del evento inmediatamente anterior.
Uno de los métodos más sencillos que se utilizaron en este trabajo para transformar la
lluvia en escorrentía fue el método racional, en su versión adaptada a la península Ibérica
(Témez, 1987 y 1991; Ferrer, 1993), pero dándole un enfoque estocástico (ver Capítulo
4.2 de Resultados). Este método empírico se basa en la transformación de precipitación con
intensidad I (que empieza en forma instantánea y continúa de forma indefinida) a una
escorrentía que continuará hasta que se alcance el tiempo de concentración (Tc), momento
en el cual toda la cuenca está contribuyendo al flujo. En ese momento de equilibrio se
alcanzará el caudal punta (Qp); el volumen entrante al sistema será el producto de la
intensidad de precipitación por el área de la cuenca (I·A), y se ve reducido por un
coeficiente de escorrentía (C, entre 0 y 1) que representa la proporción de agua retenida en
las abstracciones iniciales. Las unidades de las variables se expresan en el sistema
norteamericano, por lo que su conversión al Sistema Internacional (Q, m3 s-1; I, mm h-1; A,
km2) precisa dividir el producto entre 3,6, resultando como fórmula general:
=( ∙ ∙ ∙ )/3,6 [1]
= 0,3 ∙ ( ,
)^0,76 [3]
Para reforzar conocimientos en esta rama de la ciencia, así como en análisis post
evento y manejo de datos de radar meteorológico en modelización hidrológica distribuida
se realizó una estancia breve de investigación de 4 meses en la Facultad de Ciencias de la
Universidad de Padova (Italia); además se realizó el curso "Model building, inference and
hypothesis testing in conceptual hydrological modelling" organizado e impartido en el
Gabriel Lippmann Institute de Luxemburgo, con una duración de 40 horas.
= [5]
La parte de precipitación que no infiltra y pasa por tanto a ser escorrentía directa
genera el hidrograma de flujo superficial. Para calcular el hidrograma se utilizó el método
del hidrograma unitario del Soil Conservation Service, basado en el parámetro Tlag:
= /1,67 [6]
=∆ / [7]
= ∙ (1 − ) [8]
∙ ∙ ∙∆
Figure 2.11: Evidence and marks found and analysed in one of the study sites. (picture
taken by Andrés Díez)
Figura 2.11: Marcas y evidencias encontradas y analizadas en una de las zonas de estudio
(fotografía de Andrés Díez).
Figura 2.12: Esquema ilustrativo de los depósitos, marcas y evidencias de utilidad paleohidrológica en zonas de montaña
(adaptado de Jarret , 2002 y Benito y Thorndycraft, 2004).HWM significa marca de aguas altas.
CAPÍTULO /CHAPTER 2
Se puede simplificar el flujo del agua como un flujo unidimensional asumiendo que la
profundidad y velocidad sólo varían en la dirección longitudinal del canal, cuyo eje se
supone aproximadamente una línea recta, y la velocidad es constante en cualquier punto de
una sección transversal. En este supuesto, el movimiento del agua está gobernado por las
ecuaciones de Saint-Venant unidimensionales: continuidad y conservación de la cantidad
de movimiento:
Ecuación de la continuidad,
El flujo uniforme tiene lugar cuando la pendiente del lecho (S0) es igual a la pendiente
de fricción (Sf) y todos los demás términos de la ecuación de la conservación de la cantidad
de movimiento se consideran despreciables (slope-convenyance method). En general, se
sobreentiende que además el flujo es permanente, con lo cual la línea de energía, la
superficie de la lámina de agua y el lecho son líneas paralelas. Partiendo de esta
simplificación metodológica, derivada de la ecuación de Manning-Strickler, se relaciona la
velocidad con parámetros morfométricos y con un índice de rugosidad para los distintos
materiales del lecho para estimar los caudales:
/
= ∙ / [11]
Durante un evento de avenida es habitual que el flujo del agua no quede restringido al
canal principal, sino que el caudal desborde ocupando las márgenes y llanura de
inundación. En estas condiciones el flujo debe considerarse bidimensional y el modelo a
utilizar debe resolver las ecuaciones de aguas someras promediadas en profundidad,
también conocidas como 2D Shallow Water Equations (2D-SWE) o ecuaciones de Saint
Venant bidimensionales, éstas son:
h (uh) (vh)
Continuidad: 0
t x y [13]
u u u h
Dinámica: u v g Rx
t x y x [14]
v v v h
u v g Ry
t x y y [15]
En este trabajo se ha utilizado el software libre Iber (Corestein et al., 2010; Bladé et
al., 2012) para la simulación hidrodinámica en dos de las zonas de estudio (ver apartados
4.5, 4.6 y 4.7 de Resultados).
Figure 2.13: Topographical survey in Arenal River: (a) Total station: (b) GPS.
Figura 2.13: Levantamiento topográfico en el Río Arenal: (a) estación total (b) GPS.
Una vez obtenida la geometría se genera la malla de cálculo. Para ello se divide el
dominio de estudio en celdas de tamaño relativamente pequeño (Fig. 2.14).
Los modelos trataron de ser calibrados allí donde se disponía de marcas y medidas,
especialmente en los parámetros de rugosidad, y se sometieron a diferentes análisis de
sensibilidad, tanto de las condiciones de contorno como de la rugosidad y variaciones
geométricas.
Figure 2.14: (a) Digital elevation model of the Arenal River crossing Arenas de San Pedro
village; (b) Calculation mesh created for Cabrera stream.
Figura 2.14: (a) Modelo digital d elevaciones del río Arenal a su paso por Arenas de San
Pedro. (b) Malla de cálculo generada para un tramo del arroyo Cabrera.
Además de analizar las zonas susceptibles a sufrir estos procesos, para delimitar las
áreas contribuyentes, estas zonas deben estar conectadas con el cauce (McDade et al., 1990;
Robison y Beschta, 1990; Van Sickle y Gregory, 1990; Bragg y Kershner, 2004); de manera
que en función de parámetros como la altura del árbol, la pendiente del terreno y la
distancia al cauce se determina la probabilidad de que una pieza de madera llegue al río
(Fig. 2.15).
Figure 2.15: (a) Illustration of variables used for mass balance of wood; (b) Potential tree
fall area showing total tree height, Ht; effective tree height, He; and total arc distance, 2π
He; (c) Schematic illustrating the probability of woody debris falling into a stream from a
tree located (A) at the edge of the stream, (B) at a distance less than the effective tree
height, and (C) at a distance greater than the effective tree height. (from Robison and
Beschta 1990; Martin and Benda, 2001)
Figura 2.15: Esquema de las variables que se tienen en cuenta en los procesos de
incorporación de madera a los cauces; (b) área potencial de caída de un individuo; (c)
ilustración esquemática de la probabilidad de que una pieza de madera llegue al cauce en
función de su localización: (A) en el banco de orilla; (B) a una distancia menor que su
altura efectiva; (C) a una distancia mayor. (tomado y traducido de Robison y Beschta
1990; Martin y Benda, 2001).
El inicio del movimiento de una pieza de madera (suponiendo ésta como un cilindro)
situada en un cauce, se determinó mediante un análisis de las fuerzas que actúan sobre ella,
siguiendo los trabajos de Braudrick y Grant (2000), Bocchiola et al. (2002), Merten et al.
(2010) y Mazzorana et al. (2011), y aportando algunas modificaciones. Estas fuerzas son,
por un lado, la fuerza gravitacional y la fuerza de arrastre que facilitan el movimiento, y por
otro lado, la fuerza de rozamiento o fricción que se opone al movimiento. Dependen de
parámetros como la densidad de la madera, su longitud y diámetro, el ángulo que forma
con respecto al flujo, las condiciones hidrodinámicas del flujo (calado y campo de
velocidades) y unos coeficientes de fricción con el lecho y de arrastre (Fig. 2.16).
Figure 2.16: Schematic and body-force diagrams of some of the components of the force
balance acting on a log (from Braudrick and Grant, 2000).
Figura 2.16: Diagramas en los que se muestra de forma esquemática algunas de las fuerzas
involucradas en el inicio de movimiento de una pieza de madera (tomado y traducido de
Braudrick y Grant, 2000).
Figure 2.17: Transport regimes based on wood density (ρw), flow density (ρ), water depth
(h) and log diameter (Dw). A, B and D, rolling or sliding; C, floating.
Todos los ensayos se grabaron con una cámara digital de gran angular instalada unos
metros por encima del canal. Las videograbaciones fueron procesadas mediante un código
desarrollado en Matlab para determinar la trayectoria de cada pieza en el canal. Así se
pudieron comparar los resultados del modelo matemático con los ensayos en el modelo
físico.
Figura 2.18: (a) Canal de experimentación del Instituto FLUMEN (UPC); (b) medida de calados con limnímetro; (c)
medida del campo 3D de velocidad del flujo con un velocímetro microacústico Doppler; (d) piezas de madera utilizadas.
CAPÍTULO /CHAPTER 2
Figure 2.19: Uncertainty bounds in the frequency analysis (1) and hydraulic modeling (2).
Los resultados se integraron de forma que los mapas obtenidos están representados
con una aproximación probabilista, es decir, no marcados por un único limite, sino
representando una zonación gradualmente probable.
Figure 2.20: Illustration of the terms flood hazard, vulnerability and risk (from Merz and
Thieken, 2004)
Figura 2.20: Esquema ilustrativo de los términos peligrosidad, vulnerabilidad y riesgo por
inundación (tomado y traducido de Merz y Thieken, 2004).
Figure 2.21: Illustration of the risk analysis process and the configuration of scenarios with
large wood.
Figura 2.21: Esquema ilustrativo del proceso seguido en el análisis de riesgo y la generación
de escenarios con y sin carga leñosa.
Geográficamente la Sierra de Gredos forma parte del Sistema Central español (Fig.
3.1). Se trata de una cadena montañosa con cerca de 500 km de desarrollo que atraviesa una
parte de la Península en dirección SO-NE. La Sierra de Gredos se sitúa en su tramo medio,
entre las sierras de Guadarrama y Béjar. En ella se localiza la mayor elevación de toda la
cadena, el pico Almanzor, con 2592 m de altitud. Gredos posee un relieve acusado, y
asimétrico presentando la vertiente sur un mayor desnivel y laderas más escarpadas. Se
divide en tres sectores: oriental hasta el Puerto del Pico, central hasta el Puerto de
Tornavacas y occidental o Sierra de Béjar. Se distinguen además dentro de Gredos varias
sierras y macizos montañosos, entre los que destacan de Este a Oeste: la sierra del Valle, la
sierra de la Paramera, sierra de Villafranca, La Serrota, Sierra Llana, sierra del Barco, sierra
de Tormantos, el Macizo Central, sierra de Béjar, sierra de Candelario, cuerda del
Calvitero, sierra de la Alberca, sierra de Tonda, montes de Tras La Sierra y sierra de
Francia.
D
Desde el pun nto de vistaa hidrológi co, la Sierrra de Gredoos forma laa divisoria natural
n
entre laas cuencas del
d Duero (aal Norte) y eel Tajo (al Sur),
S quedaando ademáás circunscrita por
algunoss de sus afluuentes prin
ncipales: el Tiétar y el Alberche por p el Sur y el Norestte, y el
Tormess y el Jerte por
p el Nortee y el Suroesste.
A
Administratiivamente, la l Sierra dee Gredos ocupa
o funddamentalmeente el Surr de la
provinccia de Ávilaa (Castilla y León), auunque algun
nas sierras penetran een la provin
ncia de
Salamannca, Cácerees (Extremadura) y Tolledo (Castillla-La Manccha).
D
Desde un puunto de vista geológicco, forma parte de uno de los fragmentoss de la
Cadenaa Varisca, ell Macizo Ib bérico. Denttro de la claasificación del Macizoo Ibérico, laa Sierra
de Greddos perteneece a la Zo ona Centroo-Ibérica (L Lotze, 1945 5; Julivert eet al., 1972
2; Díez
Balda et al., 1990; figura
f 3.2).
Figur
ure 3.2: Geoological andd structural ssketch of thhe Sierra de Gredos andd surroundi
dings
(from
m Tectonic and geologi gic map of Spain
Sp 1:2000 0000 by IGMME).
VIRGINIA RUIZ-VILLANUEV
VA 92 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 3
Este zócalo granítico se encuentra limitado por las cuencas cenozoicas del Tajo y
Duero, constituidas por depósitos de origen continental. Además calizas y dolomías (de
poco espesor) flanquean al norte y sur en las zonas adyacentes más alejadas de la Sierra.
Estas rocas de edad cretácica son testimonio de la transgresión marina del mar de Thetis
ocurrida a final de este periodo, y pueden considerarse la prolongación más occidental de la
rama castellana de la Cuenca Ibérica mesozoica.
Durante la orogenia Alpina tuvo lugar una reactivación de las fracturas preexistentes
con movimientos en la vertical (Bea y Moreno, 1985). Además se han considerado como
accidentes con posible actividad neotectónica, aquellas fallas alpinas principales que
delimitan los bloques elevados del Sistema Central (Magnas 554-579, IGME). Diversos
autores posteriores (Pedraza, 1973, 1978; Gutiérrez Elorza y Rodríguez Vidal, 1978;
Garzón Heydt, 1980; Garzón, Pedraza y Ubanell, 1982; Pedraza, 1989; y Centeno
Carrillo, 1989) entre otros, han realizado estudios detallados sobre la fracturación y
evolución geomorfológica del Sistema Central en su conjunto.
Figura 3.3: (a) Superficie de cumbres formada por galayares en la zona de Los Galayos,
Guisando, Ávila.
Figure 3.3 (b): High and medium steep slopes;view from road N-502 in the Mombeltrán
area.
Figura 3.3: (b) Laderas escarpadas y débilmente escarpadas; vista de la zona de Mombeltrán
desde la carretera N-502.
Figure 3.3 (c): Granitic forms in the pathway from San Estebán del Valle to Cuevas del
Valle.
Figura 3.3 (c): Morfologías graníticas en el camino de San Estebán del Valle a Cuevas del
Valle.
Figure 3.3 (d): Fluvial system of the Arenal River crossing Arenas de San Pedro village.
Figura 3.3 (d): Sistema fluvial del río Arenal a su paso por Arenas de San Pedro.
3.2 Context
C to climáático
Figu
gura 3.4: Claasificación climática
c dee Köppen-G
Geiger en la península IIbérica (AE
EMET,
2011).
VIRGINIA
A RUIZ-VILLANU
UEVA 99 TESIS DOCTTORAL/ PhD THESIS
H (2012)
CAPÍTULO /CHAPTER
C 3
R
Respecto a laas precipitaciones, la siituación es algo más co
ompleja. Enn principio cabría
esperar una mayorr cantidad de d precipitaación al noorte, donde encontram mos mayor altitud
media y posición laatitudinal más
m septentrrional; sin embrago,
e ucede lo conntrario (Fig.. 3.5).
su
FFigura 3.5: P
Precipitacióón media annual de la peenínsula Ibéérica (AEM
MET, 2011).
).
VIRGINIA RUIZ-VILLANUEV
VA 100 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 3
El hecho de que las precipitaciones sean muy superiores al sur, debe relacionarse con
la apertura de la cuenca del Tajo (en concreto de su afluente el Tiétar) hacia el Oeste, lo
que permite que la influencia atlántica afecte a buena parte de la vertiente meridional de la
Sierra. Por el contrario, al norte, esa posible llegada de aire húmedo occidental se ve
obstaculizada por la particular disposición de los relieves existentes en la zona (como la
Sierra de Candelario con cotas de hasta 2400 m). Así también las precipitaciones
registradas en el puerto del Pico son inferiores a las registradas en Guisando, el Hornillo o
el Arenal, localidades situadas a cotas altitudinales inferiores pero ubicadas entre
contrafuertes de dirección N-S.
VIRGINIA RUIZ-VILLANUEV
VA 102 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 3
Los tipos de suelo, y según el Atlas del Medio Natural y de los Recursos Hídricos de la
provincia de Ávila (IGME, 2010), están formados por tres órdenes principales, de acuerdo
a la clasificación americana (USDA, 2006), a saber: Entisoles, Inceptisoles y Alfisoles. De
forma general, los Entisoles ocupan las laderas de la Sierra, son suelos jóvenes formados
sobre materiales difíciles de alterar, sin apenas diferencia de horizontes en el perfil. Los
Inceptisoles son algo más evolucionados, aunque el mayor grado de evolución le
corresponde a los Alfisoles con un horizonte subsuperficial argílico. En las áreas de cumbre
pueden aparecer algunos suelos turbosos ligados a las formas glaciares.
Piso mesomediterráneo: se corresponde con la zona de más baja cota (entre 400 y
800 m), y de forma general es el piso de la encina y el pino piñonero. Este piso
ocupa las zonas de fondo de valle o parte baja de las laderas, que es donde se
asientan los principales núcleos de población y en sus inmediaciones las zonas de
cultivo, las zonas de pastos y zonas industriales. Los encinares (Quercus ilex) se
encuentran mejor conservados en la vertiente sur, mientras que en la vertiente
norte, más continental, lo hacen formando un monte bajo en las laderas de solana.
Los pinares de pino piñonero (Pinus pinea) también se concentran en los
piedemontes y lomas bajas, aunque están en general poco representados. También
se pueden encontrar rodales de alcornoque (Quercus suber) en el tramo medio
alto del valle del Tiétar. Además, en este piso basal a la encina le acompañan
enebros (Juniperos oxycedrus), jaras (Cistus ladanifer), retamas (Retama
sphaerocarpa) y algunos brezos (Erica australis).
Piso supramediterráneo: zona de media montaña (de 900 a 1500 m) o piso del
melojo y del pino resinero. Cuando la humedad aumenta y la temperatura
disminuye, la encina empieza a mezclarse con el roble melojo (Quercus pirenaica)
o el pino resinero o negral (Pinus pinaster; Fig. 3.7). A éstos les acompañan como
arbustos el piorno (Cytisus scoparius), la estepa (Cytisus laurifolius) y retama
(Genista falcata). Algún tejo (Taxus baccata), acebo (Ilex aquifolium) y serbal
(Sorbus aucuparia) aparecen junto al roble. Hay que destacar que, debido la
intensa actividad dedicada a la extracción de resina, muchos de los pinares son
repoblaciones artificiales. Otros cultivos como el de castaños (Castanea sativa) ha
Figure 3.8: Cytisus scrubs in the high altitude range (Sierra del Valle, Venero Claro,
Navaluenga).
Figura 3.8: Matorrales de Cytisus en la zona de alta montaña (Sierra del Valle, Venero
Claro, Navaluenga).
Figure 3.9: Riparian forest in the Arenal River crossing Arenas de San Pedro.
Figura 3.9: Bosque de ribera en el río Arenal a su paso por Arenas de San Pedro.
Figure 3.10: Schematic vegetation catena of two cross sections of the Sierra de Gredos (up)
western section; (down) eastern section. (Adapted from Génova , 2009).
Hay que resaltar una práctica habitual de manejo que afecta a la vegetación de ribera
particularmente. Este manejo comprende labores combinadas de desbroce, limpieza y
podas selectivas del bosque de ribera. A ello se añade el destoconado previa tala. El desbroce
selectivo por medios manuales y mecánicos (desbrozadora) se suele llevar a cabo en tanto en
la ribera como en la llanura de inundación de ríos y gargantas.
Figure 3.11: Main predominant flood types: 1, flash floods; 2, torrential floods; 3, flooding
of big rivers; and 4, flooding of endorheic areas (from Díez-Herrero, 2009).
El deshielo del manto nival en primavera tras un periodo de grandes nevadas, bien
por ascenso térmico, situaciones de inestabilidad más templadas (paso de frentes
cálidos y húmedos), o precipitaciones líquidas sobre la nieve, pueden generar
importantes subidas del caudal.
Los ríos de la cuenca del Tajo han generado avenidas que muchas veces se han
traducido en grandes inundaciones llegando a producir muertos y cuantiosas pérdidas
materiales. Entre las avenidas históricas conocidas de este siglo destacan las siguientes: las
del 16 al 22 de febrero de 1936, la del 22 al 26 de enero de 1941, la del 3 al 6 de marzo de
Figure 3.12: Studied basins and rain gauge and stream gauge networks in the Gredos
mountain region.
4. RESULTADOS/RESULTS
4.1. Dendrogeomorphic analysis of flash floods ....115
REFERENCE:
Ruiz-Villanueva, V.; Díez-Herrero, A.; Stoffel, M.; Bollschweiler, M.; Bodoque, J.M.;
Ballesteros, J.A. (2010). Dendrogeomorphic analysis of flash floods in a small
ungauged mountain catchment (Central Spain). Geomorphology, 118, 383-392.
DOI:10.1016/j.geomorph.2010.02.006
ABSTRACT
Flash floods represent one of the most significant natural hazards with serious death
tolls and economic damage at a worldwide level in general and in Mediterranean mountain
catchments in particular. In these environments, systematic data is often lacking and
analyses have to be based on alternative approaches such as dendrogeomorphology. In this
study, we focus on the identification of flash floods based on growth disturbances (GD)
observed in 98 heavily affected Mediterranean pine trees (Pinus pinaster Ait.) located in or
next to the torrential channel of the Pelayo River in the Spanish Central System. Flash
floods are quite common in this catchment and are triggered by heavy storms, with high
discharge and debris transport rates favoured by high stream gradients. Comparison of the
anomalies in tree morphology and the position of the trees in the channel showed that the
intensity of the disturbance clearly depends on geomorphology. The dating of past flash
flood events was based on the number and intensity of GD observed in the tree-ring series
and on the spatial distribution of affected trees along the torrent, thus allowing seven flash
flood events during the last 50 years to be dated, namely in 1963, 1966, 1973, 1976, 1996,
2000, and 2005.
1. Introduction
Therefore, the aim of this paper is to date past flash flood events in order to improve
frequency analysis and hazard estimation. We focus on the identification of flash floods
based on growth disturbances in heavily affected trees located in or next to the torrential
channel of the Pelayo River in the Spanish Central System and present (i) a detailed
geomorphic map (1:1 000) of the torrent; (ii) an analysis of the relation between
geomorphology and external tree disturbances; and (iii) a tree-ring-based dating of past
flash flood events.
2. Study Site
The study presented in this paper was conducted within the channel and on the
lateral banks of the Pelayo River, a torrent with a length of 10 km and a catchment area of
20.6 km2 that is located on the southern slopes of the Gredos Mountain Range (Spanish
Central System; Fig. 1).
Fig. 1. (A) Location of the study area in the Gredos Mountain Range (Spanish Central
System); (B) overview of the Pelayo River catchment and the study site of this analysis; (C)
the Pelayo River at Guisando village during normal conditions and (D) during the flash
flood event in December 2008 (picture of 2008 flash flood event, courtesy of Gloria Suarez
García, used with permission).
The torrent originates at 2300 m asl and crosses the village of Guisando (765
inhabitants, 40º13'49" N., 5º9'38" W.; 764 m asl) and two recreational areas (Luis Manuel
López Camp and Los Galayos Camping). At 500 m asl, the torrent merges with the Arenal
River, a tributary of the Tagus River. The main morphometric characteristics of the study
site are summarized in Table 1.
The climate at the study site is continental Mediterranean and vegetation consists of
the Mediterranean forest. It is characterized by Mediterranean pine (Pinus pinaster Ait.),
oak (Quercus), chestnut (Castanea), walnut (Juglans), and hazel (Corylus). The altitude in
this zone ranges from 800 to 1600 m asl; the annual average temperature varies from 8 to
12ºC; and the zone is characterized by abundant rainfall. In fact, the average annual rainfall
amounts to 2000 mm. Geology consists of Upper Palaeozoic plutonic and metamorphic
bedrock and surficial Quaternary formations (colluvium and weathering mantles) made up
of conglomerates, gravels, and sands.
Flash floods are common in the Pelayo River catchment and are favoured by heavy
storms and the high stream gradient, which make extremely high discharge and debris transport
rates possible. Meteorological data are available from the “El Risquillo” station located in
the lower part of the catchment (05º09´ W., 40º13´ N.; 766 m asl.). In contrast, for this
torrential system, flow data is not available.
The methods applied in this study are outlined in Fig. 2 and can be divided into three
main sections, namely: (i) field procedures; (ii) tree-ring and field data analyses; and (iii)
flash flood reconstruction.
Fig. 2. Flow diagram illustrating the different analysis methods applied, from the data
sources to the results.
The geomorphic analysis of the torrential system was carried out applying
physiographic and morphometric mapping methods, interpretation of aerial photographs
taken in 2007 (Plan Nacional de Ortofotografía Aérea, PNOA), and field verification.
Physiographic units (i.e., high steep slopes, low steep slopes, and piedmont) of the
Pelayo River catchment have been defined based on several thematic factors, such as
altitude, geology, vegetation, and land use.
Morphometric elements have been classified as slope or torrential elements, but the
focus was clearly on torrential landforms. Based on the classification suggested by
Cenderelli and Cluer (1998), we distinguish five types of gravel bars: (i) medial-
longitudinal gravel bars with equal size channels on both sides of the bar; (ii) longitudinal-
lateral gravel bars with a main or external channel on one side and an internal or secondary
channel on the other side; (iii) lateral-point gravel bars, where the bar is close to the river
banks and a secondary channel; (iv) transverse-diagonal gravel bars with one main channel
and a secondary channel active during flash floods; and (v) levees (i.e., natural deposits on
the river banks). The type and number of bars were noted in the field, their area calculated,
and the spatial distribution of bars along the channel analyzed.
Flash floods may affect the morphology of a tree in different ways, leading to
different growth responses in the tree-ring series (Stoffel and Bollschweiler, 2009). The
most frequent disturbances and associated reactions in trees growing inside or adjacent to
flash flood torrents are shown in Figs. 3 and 4: flash floods may (i) tilt the stems of trees
growing in the river, on gravel bars, or river banks through the unilateral pressure of the
flow or through the impact of individual boulders (Braam et al., 1987; Fantucci and
Sorriso-Valvo, 1999; Stoffel et al., 2005); (ii) expose roots as a result of river bank or gravel
bar erosion (Lamarche, 1968; Carrara and Carroll, 1979; McAuliffe et al., 2006); (iii)
remove the bark from the stem and injure the cambium from the impact or the abrasion of
boulders and wood transported in the flow (Sachs, 1991; Larson, 1994; Bollschweiler et al.,
2008); or (iv) sheared off trees and cause the formation of candelabra growth following
severe impact by boulders (Butler et al., 1987; Shroder and Butler, 1987).
The sampling strategy was based on these external evidence of flash flood activity in
trees. We selected a total of 98 P. pinaster that were influenced by flash floods and were
located within the channel, on the gravel bars, or on the banks of the Pelayo River.
In addition, 16 undisturbed trees were selected in the medium and upper reaches of
the valley slopes in order to obtain a reference chronology representing local growth
conditions.
For each tree sampled, additional information was noted, such as (i) determination
of its coordinates using a GPS; (ii) its geomorphic position; (iii) its position with respect to
neighbouring trees; (iv) description of the growth disturbances (GD); (v) tree diameter;
(vi) tree height; and (vii) the position of cores sampled. We extracted increment cores
using increment borers (Grissino-Mayer, 2003) with a length of 40 and 60 cm (internal
diameter in both cases: 5.5 mm).
Fig. 4. External evidence of flash flood activity in trees growing along the Pelayo River: (A)
abrasion scars; (B) bark erosion; (C) stem tilting; (D) stem burial; (E) floating roots; (F)
exposed roots. The blue arrows indicate the flow direction.
Two cores were usually sampled per tree, one in the flow direction of flash floods and
one from the opposite side. In trees with exposed roots or buried stem bases, samples were
taken as close to the ground as possible in order to obtain the largest number of tree rings.
In the case of tilted trees, two cores were extracted at the height of the maximum bend,
where compression wood was most likely to be present. Finally, at least three cores were
selected in trees with visible injuries: two samples from the lateral edges of the injury where
signs of the impact were visible in the tree-ring series but where no rings were missing
because of abrasion. In addition, one more sample was taken from the opposite side of the
stem. In the case of the undisturbed trees, two cores were extracted perpendicular to the
slope direction and as close to ground level as possible.
A total of 269 increment cores were collected in the field: 241 samples from the 98
disturbed P. pinaster trees growing within the active channel or in its vicinity as well as 28
cores from the 16 undisturbed reference trees.
Samples collected from the disturbed and undisturbed trees were analyzed in the
laboratory using the standard methods described by Stoffel and Bollschweiler (2008):
After air-drying the samples and preparing the core surfaces (polishing), we counted tree
rings and measured ring widths using a digital LINTAB positioning table coupled to a
Leica stereomicroscope and TSAPWin 4.6 software (Time Series Analysis and
Presentation; Rinntech, 2010). This program also allows the representation of measured
tree-ring series, as well as cross-dating and quality checks of the growth curves. Ring widths
were measured with an accuracy of 1/100 mm. Increment curves of the disturbed trees
were then cross-dated with the reference chronology (Rinntech, 2008) in order to correct
faulty tree-ring series derived from disturbed samples (e.g., false or missing rings) and to
determine initiation of abrupt growth suppression or release. Furthermore, samples were
analyzed visually, and tree rings showing compression wood or callus tissue were noted in a
dedicated file.
The dating of past flash flood events was based on (i) the GD observed in the tree-
ring series (i.e., abrupt growth suppression or release, compression wood, eccentric growth,
callus tissue, and injuries); (ii) the intensity of the GD signal in the tree-ring record; (iii)
the overall number of trees affected by an event; as well as (iv) on the spatial distribution of
affected trees along the torrent.
The different parameters were then quantified and specific weights (wGD from 0.1
to 1) were assigned to the different GD (Table 2). The number of trees affected in a
particular year and the trees being present for analysis in the same year were then used to
derive the percentage of damaged trees (%DT). The spatial distribution of affected trees in
the torrent (SD) was attributed a value of 1 in the case of heterogeneous distribution (i.e.,
several trees that were concentrated in the same area), and a value of 2 was chosen for
homogeneous distribution (i.e., trees evenly spaced along the torrent).
The product of these parameters (wGD × %DT × SD) was then used to date past
flash flood events; and event years were, in addition, tested for significance using the non
parametric Wilcoxon-Mann-Whitney test.
Table 2: Weights used for the quantification of different growth disturbances with
different intensitiesa
a
Intense means that the GD is well recognizable directly in the increment core, and weak means that
the GD is hardly recognizable directly, but well recognizable with the help of a magnifying glass and the
disturbed growth curves.
4. Results
From the 67 gravel bars identified in the study reach, we note a predominance of
medial-longitudinal, lateral-longitudinal, or lateral-point gravel bars and a scarcity of
transverse-diagonal gravel bars or levees (Table 3).
Fig. 5. Geomorphological mapping of the upper part of the study reach. The legend shows
the classification of units and elements.
Table 3: Types of gravel bars and area (absolute and relative) in the study reach
The spatial distribution of gravel bars changes along the torrent (Table 4). In the
upper part, medial bars are more abundant (60% of medial bars are located in this part of
the reach); whereas lateral point gravel bars are most significant in the central part of the
study site (52% of observed point bars are located in this part).
Table 4: Spatial distribution of bars along the studied reach (upper, central, and lower
part) (values are given per gravel bar type)
Thereafter, external disturbances in trees were related with tree positions in the
channel. A total of 208 macro disturbances in tree morphology were identified in the 98 P.
pinaster trees (Table 5). Several signs were normally observed per tree sampled, with tilted
trunks (96%), exposed roots (62%), and bark erosion (23%) being the most common
features observed. In contrast, apex decapitation (5%) or stem burial (7%) were only
occasionally found.
Table 5: Trees showing external disturbances and located in different geomorphic settingsa
a
Gravel bars are grouped in medial (medial-longitudinal) and lateral (including lateral-longitudinal.
lateral-point and transverse-diagonal bars). Abbreviations: HB: head of bar; TB: tail of bar; M: margin;
EM: external margin; IM: internal margin; SR: straight reach; EC: external curve of a meander (i.e.
undercut slope); IC: internal curve of a meander (i.e., slip-off slope); C-D. disturbance in the flow
direction. and A-B. disturbance perpendicular to the flow direction.
Table 5 also shows that a majority of trees with the most energetic disturbances (i.e.,
stem tilting or exposed roots) are located in the zones with greatest stream power and most
significant geomorphic work (i.e., banks and external margin of lateral bars). For instance,
trees with tilted stems are located in different geomorphic settings, but their relative
number clearly increases with the increasing process energy involved. The relation between
the energy of the external evidence and the geomorphic sides is shown in Figure 6.
Fig. 6. Relation between the energy of the external evidence and the geomorphic sides. Ball
size represents the percentage of trees showing a specific disturbance located in a specific
geomorphic side related to the total number of trees showing that disturbance. The legend
may be checked in Table 5.
Figure 6 shows a good relation between the most significant disturbances in tree
morphology and the most energetic geomorphic tree positions within the channel. As a
general rule, the relative number of tilted trees (floating or exposed roots) increases with
increasing process energy. In the case of apex decapitation (bark erosion or stem burial)
different peaks can be observed; and the relative number of disturbances in tree
morphology does not seem to increase linearly with increasing process energy.
Tree-ring counting of the flash flood affected trees enabled an assessment of tree age
and their distribution along the torrent (Fig. 7). The trees used in this study are on average
38 years old (STDEV: 22 years). The oldest tree sampled attained sampling height in A.D.
1869 and the youngest tree sampled in 1998. The upper sector of the torrent contains the
youngest trees.
Fig. 7. Age structure of the sampled trees in the torrent. Left: upper sector of the study
reach. Right: lower sector.
A total of 280 GD could be detected in the tree-ring series of the disturbed trees
(Table 6), most frequently in the form of growth decreases (present in more than 70% of
the samples). Injuries and callus tissue were, in contrast, scarce and only observed in 7% of
the samples.
The dating of past flash flood events was based on specific weights and the product of
GD. The percentage of trees affected and the distribution of affected trees is shown in
Table 7. In total, GD are observed in 22 different years since A.D. 1943. Based on the
Wilcoxon test (95% confidence interval) and p 0.04, we consider seven of the years with
GD as flash flood years namely: 1963, 1966, 1973, 1976, 1996, 2000, and 2005.
Based on the tree morphology and the GD in the tree-ring series, the evidence is good
for the existence of events in the other years as well, but the small number of trees available
for analysis did not allow them to be considered events with equal confidence. One
example is the year 1989, where the confidence interval is lower but significant. The spatial
distribution of trees affected in particular years is shown in Fig. 8.
Table 7: Calendar year, damaged trees, living trees, percentage of damaged trees, weight
values for the detected growth disturbances, weight for the spatial distribution along the
torrent, and the calculated coefficient of confidence.
Fig. 8. Spatial distribution of living and damaged trees of four dated events.
5. Discussion
In the present study, 98 P. pinaster trees affected by flash floods have been analyzed
with 241 increment cores to reconstruct flash flood events at Pelayo River (Spanish
Central System) covering the past 50 years. While the time covered by this reconstruction
is much shorter compared to other, non-Mediterranean environments (e.g., Stoffel et al.,
2008; Mayer et al., in review.), our study clearly shows the potential of flash flood analyses
based on information contained in growth-ring series of affected trees. Tree-ring analysis
was complemented with detailed geomorphic mapping focusing on gravel bar formation
according to the classification of Church and Jones (1982) and Cenderelli and Cluer
(1998). The spatial distribution of the gravel bars changes along the torrent. The proximal-
distal ocurrence of channel bars follows the schematic diagram proposed by Church and
Jones (1982). In very high flows, like in the present case, sediment is mobilized from the
channel and flow resistance is accordingly modified when the greatest water and sediment
loads are transported through the channel.
In addition, we also identify years with a more limited number of primarily weak
signatures, such as in 2003, 1999, 1997, 1990, 1970, and 1956. Based on the reactions
observed in the tree-ring series, we assume that these GD would be the result of flash floods
as well. In some cases, the tree-ring signatures appear to be weak because of strong tree
responses in the growth-ring series resulting from preceding events.
For other years, as e.g., 1989, neighbouring catchments registered intense flash flood
events (Ballesteros et al., 2010) that make the presence of a simultaneous event in the
Pelayo River more probable. Another limitation of dendrogeomorphic work based on P.
pinaster becomes obvious for the possible event of 1999, when late and weak GD were
preferably added to an event in 2000 for which several injuries could be used for accurate
dating of a high intensity event.
Based on the above considerations, we believe that there are two types of flash flood
events at Pelayo River: on one hand, we identify flash floods characterized by high
magnitude but low frequency. These events are capable of causing major damage to the
vegetation and could be reconstructed in the growth-ring series of a large number of trees.
On the other hand, we identify a second type of event of smaller magnitude and higher
frequency, but still large enough to occasionally cause less severe damage to the vegetation.
Such a separation of higher frequency–lower magnitude from lower frequency–higher
magnitude events may improve magnitude-frequency analysis of flash floods and hazard
estimation in the future.
6. Conclusions
Acknowledgements
The authors want to express their gratitude to the Ministry of Science and Innovation, for financial
support and for the grant for a 3 months research stay in Bern (Switzerland). This work was founded by
CICYT, Dendro-Avenidas project (CGL2007-62063) and for the Geological Survey of Spain. We are
grateful to the Tagus Water Authority, Enviroment Department of Castilla y León in Ávila, and Guisando
Council for their collaboration. We want to adreess our special thanks to forester Gloria Suárez-García for
her assitance in the field. Last but not least, we offer our gratitude to the anonymous reviewers for their
helpful comments.
REFERENCES
Alestalo, J., 1971. Dendrochronological interpretation of geomorphic processes. Fennia 105.
Alía, M., Martín, S., De Miguel, J., Galera, R., Agúndez, D., Gordo, J., Catalán, G., Gil, L., 1996. Regiones
de procedencia Pinus pinaster Aiton., Dirección general de Conservación de la Naturaleza, Madrid,
Spain.
Ballesteros, J.A.; Stoffel, M.; Bodoque, J.M.; Bollschweiler, M.; Hitz, O.; Díez-Herrero, A. 2010. Changes in wood
anatomy in tree rings of Pinus pinaster Ait. following wounding by flash floods. Tree Ring Bulletin, accepted
for publication.
Begin, Y., 2001. Tree-ring dating of extreme lake levels at the subarctic–boreal interface. Quaternary
Research 55, 133–139.
Blanco, E., Casado, M., Costa, M., Escribano, R., Gracía Antón, M., Génova, M., Gómez Manzaneque, G.,
Gómez Manzanaque, F., Moreno, J., Morla, C., Regato, P., Sainz Ollero, H., 1997. Los bosques
ibéricos. Editorial Planeta, Barcelona, España, 572 pp.
Bogino, S.M., Bravo, F., 2008. Growth response of Pinus pinaster Ait. to climatic variables in central
Spanish forests. Ann. For. Sci. 65, 506.
Bollschweiler, M., Stoffel, M., Schneuwly, D.M., 2008. Dynamics in debris-flow activity on a forested cone,
a case study using different dendroecological approaches. Catena 72, 67–78.
Borga, M., Boscolo, P., Zanon, F., Sangati, M., 2007. Hydrometeorological analysis of the August 29, 2003
flash flood in the eastern Italian Alps. Journal of Hydrometeorology 8 (5), 1049–1067.
Braam, R.R.,Weiss, E.E.J., Burrough, A., 1987. Spatial and temporal analysis of mass movement using
dendrochronology. Catena 14, 573–584.
Butler, D.R., 1987. Teaching general principles and applications of dendrogeomorphology. J. Geol. Educ.
35, 64–70.
Carrara, P.E., Carroll, T.R., 1979. The determination of erosion rates from exposed tree roots in the
Piceance Basin, Colorado. Earth Surf. Process. 4, 307–317.
Cenderelli, D., Cluer, B., 1998. Depositional processes and sediment supply in resistant-boundary
channels: examples from two cases studies. In: Whol, E., Tinkler, K., (Eds.), Rivers over rock: Fluvial
processes in bedrock channels. Geophysical monograph 107. American Geophysical Union.
Church, M., Jones, D., 1982. Channel bars in gravel-bed rivers. In:. Hey, R. D., Bathurst, J. C. and Thorne,
C.R. (eds), Gravel-bed Rivers. John Wiley and sons Ltd. Chichester.
Fantucci, R., 2007. Dendrogeomorphological analysis of shore erosion along Bolsena lake (Central Italy).
Dendrochronologia 24, 69–78.
Fantucci, R., Sorriso-Valvo, M., 1999. Dendrogeomorphological analysis of a slope near Lago, Calabria
(Italy). Geomorphology 30, 165–174.
Gaume, E., Bain, V., Bernardara, P., Newinger, O., Barbuc, M., Bateman, A., Blasˇkovicˇova´, L., Blöschl,
G., Borga, M., Dumitrescu, A., Daliakopoulos, I., Garcia, J., Irimescu, A., Kohnova, S., Koutroulis,
A., Marchi, L., Matreata, S., Medina, V., Preciso, E., Sempere-Torres, D., Stancalie, G., Szolgay, J.,
Tsanis, I., Velasco, D., Viglione, A., 2008. A collation of data on European flash floods. Journal of
Hydrology, accepted for publication.
Grissino-Mayer, H.D., 2003. A manual and tutorial for the proper use of an increment borer. Tree-Ring
Res. 59, 63–79.
LaMarche, V.C., 1968. Rates of slope degradation as determined from botanical evidence, White
Mountains, California. U.S. Geol. Surv. Prof. Pap., 352-I.
Larson, P.R., 1994. The vascular cambium. Development and structure, Springer, Berlin, 725 pp.
Mayer, B., Stoffel, M., Bollschweiler, M., Hübl, J., Rudolf-Miklau, F., (in review). Frequency and spread of
hyperconcentrated flows on fans: a dendrogeomorphic case-study from a dolomite catchment in the
Austrian Alps. Geomorphology.
McAuliffe, J.R., Scuderi, L.A., McFadden, L.D., 2006. Tree-ring record of hillslope erosion and valley floor
dynamics: Landscape responses to climate variation during the last 400yr in the Colorado Plateau,
northeastern Arizona. Glob. Planet. Change 50, 184–201.
Rinntech, 2010. LINTAB: Precision ring by ring. http://www.rinntech.com/Products/ Lintab.htm.
Sachs, T., 1991. Pattern formation in plant tissue, Cambridge University Press, Cambridge, 246 pp.
Scheuren, J.M., O. le Polain de Waroux, R. Below, Guha-Sapir, D., 2008 Annual Disaster Statistical
Review: The Numbers and Trends 2007. Centre for Research on the Epidemiology of Disasters
(CRED), Catholic University of Louvain (UCL), Brussels.
Shroder, J.F., 1978. Dendrogeomorphological analysis of mass movement on Table Cliffs Plateau, Utah,
Quat. Res. 9, 168–185.
Shroder, J.F., Butler, D.R., 1987. Tree-ring analysis in the earth sciences. In: Jacoby, G. C. and Hornbeck, J.
W (Eds.), Proc. Int.Symp. Ecological Aspects of Tree-Ring Analysis. Lamont Doherty Geological
Observatory, 186–212.
St. George, S., Nielsen, E., 2003. Palaeoflood records for the Red River, Manitoba, Canada, derived from
anatomical tree-ring signatures. Holocene 13 (4), 547–555.
Stoffel, M., Bollschweiler, M., 2009. What Tree Rings Can Tell About Earth-Surface Processes: Teaching
the Principles of Dendrogeomorphology. Geography Compass 3, 1013–1037.
Stoffel, M., Bollschweiler, M., 2008. Tree-ring analysis in natural hazards research: an overview. Natural
Hazards and Earth System Sciences 8, 187–202.
Stoffel, M., Lievre, I., Conus, D., Grichting, M., Raetzo, H., Gärtner, H. W., Monbaron, M., 2005. 400
years of debris flow activity and triggering weather conditions: Ritigraben, Valais, Switzerland, Arct.
Antarc. Alp. Res. 37(3), 387–395.
Yanosky, T.M., 1999. Tree ring analysis in hydrometry. In: Herschy, R.W. (Ed.), Hydrometry: Principles
and Practises. Wiley, New York, 265–289.
Yanosky, T.M., Jarrett, R.D., 2002. Dendrochronologic evidence for the frequency and magnitude of
paleofloods, in House, P.K., Webb, R.H., Baker, V.R., and Levish, D.R.,(eds) Ancient Floods,
Modern Hazards: Principles and Applications of Paleoflood Hydrology. American Geophysical
Union, Water Science and Application Series 5, 77-89.
Zielonka, T., Holeksa, J., Ciapala S., 2008. A reconstruction of flood events using scarred trees in the Tatra
Mountains, Poland. Dendrochronologia 26, 173–183.
(4.2) 5. Discussion
5.1. Interpretation and value of data from
documentary sources
5.2. Interpretation and value of tree-ring data
5.3. Seasonality, meteorological causes and flood
envelopes
5.4. Flash flood frequency
(4.2) 6. Conclusions
REFERENCE:
Ruiz-Villanueva, V., Díez-Herrero, A., Bodoque, J.M., Ballesteros, J.A., Stoffel, M.
(2012). Characterization of flash floods in small ungauged mountain basins of
Central Spain using an integrated approach. Catena, Submitted 28 August. Under
review 14 September.
ABSTRACT
One of the main problems of flood frequency analysis (FFA) is the lack of
hydrological input data and related uncertainties. In an attempt to overcome these
problems, we use archival records, dendrogeomorphic time series and instrumental data
(daily rainfall and discharge) from four ungauged and poorly gauged mountain basins in
Central Spain to reconstruct and compile information on 41 flash flood events since the
end of the 19th century. Estimation of historical discharge and the incorporation of
uncertainty for the at-site and regional FFA were performed with an empirical rainfall-
runoff assessment as well as stochastic and Bayesian Markov Chain Monte Carlo
(MCMC) approaches. Results for each of the ungauged basins include fluctuations in
flood frequency, flood severity, their seasonality and triggers (synoptic meteorological
situations). The reconstructed data series clearly demonstrate how uncertainty can be
reduced by including historical information, but also points to the considerable influence
of different approaches on quantile estimation. This uncertainty should be taken into
account when these data are used for flood risk management.
1. Introduction
precision (Stoffel et al., 2008), and may, moreover, provide information regarding river
dynamics (Arbellay et al., 2012).
Based on the above considerations, this study takes into account non-instrumental
data sources (tree-ring records and documents) in order to improve at-site and regional
FFA in ungauged and poorly gauged mountain basins of central Spain (Sierra de Gredos)
where the lack of instrumental data (flow data) prevents the use of traditional methods for
the characterization of flash floods. In particular, the paper aims at (i) reconstructing the
most complete catalogue of past flash floods in the study area, (ii) analyzing the fluctuation
in frequency and severity, their seasonality, synoptic meteorological causes as well as the
human impacts in terms of damage to infrastructures and fatalities. Moreover, (iii) we
address the estimation of historical peak discharge taking into account uncertainties
regarding antecedent conditions and land-use changes so that the results of this study can
be incorporated in flood risk management.
2. Study Region
The basins analyzed in this study are located in the eastern Sierra de Gredos massif
(Province of Avila) of the Spanish Central System (Fig.1) belong to the Tiétar River basin,
a tributary of the Tagus River. Geology in the region is mainly composed of granites
(Upper Palaeozoic granitoids) and covered by a sandy weathering mantle.
Vegetation is abundant in the area and dominated by Pinus pinaster Ait. in the
headwaters and deciduous forests (Quercus pyrenaica Wahl. and Quercus ilex L.) in the
lower parts of the study region. European alder (Alnus glutinosa (L.)) and narrow-leaved
ash (Fraxinus angustifolia Vhal.) predominate in the river corridors. Tree clearance has
been practiced in the study areas in the past and is still being practiced today.
This study focuses on four ungauged (Arenal and Pelayo Rivers) or poorly gauged
(Arenal River at Arenas de San Pedro and Santa Maria River at Candeleda) basins. The
Pelayo is a tributary of the Arenal river; it crosses the village of Guisando (765 inhabitants)
and two camp sites. The Arenal is a tributary of the Tiétar River and crosses the village of
Arenas de San Pedro (~6900 inhabitants), where a stream gauge (automatic hydrological
information system (AHIS) station) has been installed in 2001 (119 km2 drainage area).
The neighboring basin of the Santa Maria River crosses the village of Candeleda (~5145
inhabitants) where a stream gauge exists since 1973. The main characteristics of these
basins are summarized in Table 1.
Despite the fact that at least 18 people have been killed by floods in the study area
and vicinity over the past 70 years, flood frequency studies or flood risk management plans
are still missing so far in the area.
Table 1: Main morphometric characteristics of the four basins analyzed in this study. Note
the existence of an automatic hydrological information system (AHIS) station for the
Arenal River and of a stream gauge at Candeleda for the Santa Maria River (located in
Figure 1).
For the purpose of this study, different approaches were combined and data
integrated in order to obtain a detailed characterization of flash floods in the study area
(Fig. 2).
3.1. Histo
orical Do
ocumenttation
Written doocumentaryy sources (ee.g., municip pal archivess, newspapeers, civil protection
documments and scientific papers) weere screeneed systemattically to oobtain (in--) direct
mation on the impact that past pprecipitatioon and floo
inform od events haave had on n people,
their ggoods and services. Qualitative
Q ddata gathered from thhese sourcess were tran nsformed
into nnumerical indices
i folllowing Barrriendos an nd Coeur (2004) andd Bullón (2 2011) to
derivee data aboutt the type, in
ntensity, duuration and
d date of passt events.
VIRGINIA
A RUIZ-VILLANUEEVA 149 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
In the Pelayo River, tree-ring records of 114 P. pinaster trees (269 samples) were
analyzed along a 2.5-km stretch upstream the village of Guisando so as to detect growth
disturbances (GD) induced by flash floods. Following the protocol suggested by Ruiz-
Villanueva et al. (2010), dating of flash flood events in the Pelayo River was based on (i)
the nature of GD observed in the tree-ring series (i.e., abrupt growth suppression or release,
compression wood, eccentric growth, callus tissue, and injuries); (ii) the intensity of the
GD signal in the tree-ring record; (iii) the overall number of trees affected by an event; as
well as (iv) on the spatial distribution of affected trees along the stream.
In the reach of the Arenal River crossing Arenas de San Pedro village, riparian trees
(A. glutinosa and F. angustifolia) were analyzed over a distance of 3 km. As a result of a
clean-up (logging) of the channel by the Water Authority in 2008, 84 cross-sections could
be analyzed on stumps and thus allowed convenient dating of visible and internal wounds.
At the same time, the absence of entire stems prevented an assessment of scar heights and
presumably prevented detection of scars inflicted at higher positions on the stem. As usual
in dendrogeomorphic studies (Stoffel and Bollschweiler, 2008, 2009), additional
information was obtained for each tree sampled, namely its geomorphic position, tree
coordinates, or description of external disturbances. In the case of the Arenal River, the
dating of past flood events was accepted if at least two scars were identified in different
trees at the same year.
The rain gauge network of central Spain is sparse and unevenly distributed, covering
mainly valley and lowland areas, thus leaving mountainous regions behind with very
limited data (Buytaert et al., 2006). In addition, hydrometeorological networks in the area
mainly provide daily data starting at best in the early twentieth century. Usually, however,
time series start in the 1950 only or even later, and many stations have been removed,
relocated or completely stopped in the early 21st century.
This study builds on three rain gauges located in the area (Fig. 1) providing daily data
for the time window covered with documentary and tree-ring sources. The AHIS are
continuously recording meteorological parameters; and they have been operational only
since 2001.
The network of flow gauges is very much centered in large river basins, and the few
existing records from smaller mountain catchments are usually short and were not installed
before 1973. Figure 1 illustrates that eight stream gauges exists in the mountain ranges of
the wider study area and in basins similar to those studied here.
Documentary sources and tree-ring data were then used to infer flood frequency.
Unlike Ballesteros et al. (2011), where a large number of scars in injured trees could be
used to reconstruct peak discharge as well, we had to derive flood magnitude from daily
rainfall records in the present case.
Daily precipitation data was first transformed into mean areal precipitation for each
of the four basins using traditional techniques (Chow et al., 1988). Precipitation analysis
was different for the event of 1936, since only one station was available in this case and the
correlation (linear regression) between this station and the others had to be tested in order
to estimate the mean areal precipitation.
Based on the precipitation estimates for each flood event we performed an empirical
rainfall-runoff assessment (based on rational method following Témez 1991, Ferrer 1993
and Rico et al. 2005) and used a stochastic approach to estimate peak discharge. The runoff
coefficient C (0≤ C ≥1) is an important diagnostic variable for catchment response
(Marchi et al., 2010) and depends mainly on antecedent conditions, land use and soil
cover. Since these conditions may not be well known for past events, the runoff coefficient
C has been treated as a stochastic variable here with a known uniform distribution over an
interval (a, b), where a represents the lower bound equal to 0.2, and b is the upper bound
was equal to 0.6. These boundaries were selected based on the values proposed by Marchi
et al. (2010) who found the range 0.23 to 0.56 for extreme flash floods in the
Mediterranean region. Based on the normalized and cumulative histograms showing the
stochastic tendencies of the data set, a goodness-of-fit table is displayed with various
probability distributions and the best estimates for their parameters. The Monte Carlo
simulation then provided the mean and standard deviation for each simulated discharge
event.
The procedure described above was then validated using the records from the stream
gauge located in Candeleda and in Arenal AHIS (for the time period between 2001 and
2008) so as to analyze the accuracy of the approach and the assumed error.
The flood frequency analysis is applied to three cases with different quantities and
types of data: (i) at-site frequency analysis for Candeleda combining historical information
from 1936 and instrumental data series since 1973 (ii) at-site frequency analysis for the
Arenal AHIS station using the short instrumental data series (2001–2011) and the longer
historical estimates; and (iii) a regional flood frequency analysis (RFFA) using all available
flow data from the stream gauges located in the mountains and based on the flood-index
method.
For the FFA, a Bayesian Markov Chain Monte Carlo (MCMC) procedure (Reis and
Stedinger, 2005) was used, as it can handle information from historical and instrumental
observations through adequately defined likelihood functions in a straightforward way,
and more importantly, as it can account for uncertainties in hydrologic extremes and as it
provides estimates of confidence bounds for the estimated quantiles (Gaal et al., 2010).
In a final step, results of this study were comapred with an existing regional model of
maximum discharge for the western part of Tagus Basin (Mediero and Jiménez, 2001;
CEDEX, 2011).
4. Results
Results from the Pelayo study showed that tree ages reach up to ~100 years. Scars or
injuries were only occasionally found (<10%) whereas other GD were much more
frequently observed (280 cases) in the 114 trees investigated. GD could be found in 22
years, of which eight showed a large number of visible GD in a representative number of
trees; these eight years were considered as the most reliable flash flood events.
In the Arenal River, the age of the riparian trees reached up to ~50 years. A total of
147 injuries was detected in the 84 trees analyzed and in 24 different years. The oldest scar
was inflicted in 1953. Several samples contained more than 2 injuries (and up to 7 in the
same cross section). The largest number of scars (16 which is the 11% of all the entire
population) was found in 1997. Based on the number of trees affected at the same time
(and by the same event), reconstruction of past flash floods was considered very reliable for
12 events.
Comparison of tree-ring records from the two catchments points to 34 years with
dendrogeomorphic evidence of flash flood activity, of which 14 coincide between the two
rivers. The documentary sources provided information on 49 events since 19th century.
The 14 event years identified in the tree-ring records of the Pelayo and Arenal Rivers were
then contrasted with the events identified in the documentary sources in order to assign
occurrence dates to each event. The final database of archival used for further analysis
therefore contains 41 events (Fig. 3) and is restricted to flash floods registered in
documentary sources, events registered both documents and tree-ring records as well as
events simultaneously recorded in a large number of trees but without analogues in written
sources.
A
As shown inn the figure 4, 40% of them have occurred in the sprinng-summer season
while 60% in autumn-winter.. This seasoonality coulld be related
d to differennt meteoro
ological
forcing expressed in the Fig.4..
VIRGINIA RUIZ-VILLANUEVA
A 154 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
4.2. Estim
mates of flash floo
od dischaarges and
d uncertaainties
VIRGINIA
A RUIZ-VILLANUEEVA 155 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
Table 2: Discharge estimates (mean and standard deviation) for the four study sites (units
are in m3 s–1). No data was available for cells given in grey.
4.3. Floo
od frequeency analysis
VIRGINIA
A RUIZ-VILLANUEEVA 157 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
4.3.2. At-site FFA using a short data series and longer historical
information
The at-site FFA at Arenal River AHIS station is based on a short data series of
annual maximum discharge since 2001 and estimated discharge from longer historical
records (Fig. 6), whereby different mechanisms of frequency fitting are applied.
Figur
ure 6: Fittedd GEV distriributions baased on instr trumental data
da and (a)) iincluding the t four
large
gest historica
cal events (i.ee. the higheest magnituude from alll estimates)) w
with uncerrtainty;
insttrumental records
r andd (b) all estim
mates with uncertainty
u y based on hhistorical reecords.
VIRGINIA
A RUIZ-VILLANUEEVA 159 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
Table 4: Estimation of the discharge quantiles Q(ML) for Arenal River at AHIS station
corresponding to the return periods T = 10, 50, 100, 500. CI0.05 (CI0.95) is the 5% (95%)
confidence limit of the estimates Q(ML), ΔCI = CI0.95- CI0.05. Grey highlights the lowest
ΔCI and ΔCI/Qt values.
Fig
ig. 7: Fittedd GEV distriibutions forr the Pelayoo River at Guisando
G (A
A and B) andd the
Arennal River at Arenas
A de San
S Pedro vvillage (C an nd D), withh discharge eestimates trreated as
instr
trumental reecords (A and
an C) and aas instrumeental record ds with the eexception of
o the 4
(5)) largest even
ents which were
w conside dered historirical events with
w uncerttainty (B an nd D).
VIRGINIA
A RUIZ-VILLANUEEVA 161 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER
C 4
T
The inclusioon of some events as hhistorical afffects the co onfidence iintervals, an
nd this
effect reemains visibble indepenndently of tthe numberr of large evvents (only 1936, the two
t or
the fourr largest eveents with un
ncertainty) included in
n the analysis (data nott shown).
4.3.4. Regional
R flood frequuency analysis (RFFA
A)
R
Regional quaantiles weree estimated based on th he index-floood methodd and the MCMC M
approacch using eigght concord dant stationns from a homogeneou
h us region (H H=1.39, D<2.14,
N=8). A total of 11 probability disttribution functions f were
w testedd (K=0.37)) with
estimateed initial parameters
p of
o 0.6067, 0.5548, and -0.1183. The MCM MC approacch was
first useed for the gaauged dataset and incluuding the esstimates forr the three bbasins (Fig. 8).
VIRGINIA RUIZ-VILLANUEVA
A 162 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
5. Discussion
In this study 41 past floods have been documented based on historical sources and
dendrogeomorphic evidence. The veracity of the compiled data is supported by its high
level of temporal and spatial coherence and by the correspondence between the events
described in the documents and those recognized in tree-ring records.
The only available common data between the datasets was daily rainfall, which was
consequently used to estimate peak discharge for most of the documented events. We
combined an empirical rainfall-runoff assessment with a stochastic approach so as to
transform daily data into peak discharge for each event. This simple empirical assessment
has been extensively used in Spain (Díez-Herrero, 2001; Potenciano, 2004; Rico et al.,
2005; Ortega, 2007) for the assessment of peak flows in natural watersheds with surface
areas up to 3000 km2 and concentration times ranging from 1 to 24 hours. Validation of
the method revealed that estimates were associated with 17% error, which is relatively
small for past flood reconstructions (peak discharge estimates of historical floods by
hydraulic calculations have errors up to 25% according to Reis and Stedinger, 2005).
The most important assumption of this study was the treatment of the runoff
coefficient as a stochastic variable since antecedent conditions, land use and other
parameters are not known for past events. Boundaries (0.2-0.6) were established based on
the findings of Marchi et al. (2010) who obtained values in the range of 0.23 to 0.56 for the
Mediterranean region. Merz et al. (2006), in contrast, related runoff coefficients with flood
type, and found that smallest runoff coefficients (median 0.15) are associated with flash
floods, short-rain floods (0.36), whereas slightly larger values are produced by long-term
rainfalls (0.38) and the largest runoff coefficients (0.63) are associated with snowmelt
floods.
20th century (i.e. four out of seven high intensity events occurred in 1902, 1922, 1927, and
1928). As floods were classified based on damage to infrastructures (bridge and roads
damages) and not on casualties (Barriendos and Coeur, 2004), it seems possible that the
great differences in quality of structures may have influenced the classification. On the
other hand, however, events that caused fatalities in the region (i.e. 1936, 1959, 1982,
1990, and 1999), were not necessarily high-intensity (high-magnitude) events, but floods
took place during late summer or early autumn during these years, when tourism causes the
population in this region to increase substantially.
Dendrogeomrphic evidence in riparian trees has been used here to date the
occurrence and frequency of past floods, but not for a direct estimation of flood
magnitude. Although there is no perfect correlation between the number of scarred trees
(or growth disturbances) and discharge estimates, we still observe a tendency for more
evidence in trees during larger events, especially at Pelayo River where a confidence
coefficient (Ruiz-Villanueva et al., 2010) was used rather than the absolute number of
injuries (as in the Arenal River).
Tree age was the most important limitation of dendrogeomorphology at this study
site (Stoffel et al., 2010), as vegetation growing within the area of fluvial activity does not
normally reach high ages (<70 yr in the present study).
Woody material with different stages of decay and various dimensions was present
along the study reach of the Arenal River, as were granite boulders in both the Arenal and
Pelayo Rivers (boulders are much larger and more abundant in the latter). The large
presence of woody debris may explain the predominance of scarred trees at Arenal River
and the formation of impacts by floating wood, whereas the scarcity of woody debris and
the predominance of bedload transport may explain the rare occurrence of injuries in trees
in the Pelayo River,. In addition, differences in tree species composition (P. pinaster trees
with thick and protective bark in Pelayo River and more vulnerable, thin-barked riparian
trees in Arenal River), tree density and river slope may have contributed to these
differences as well.
Most of the disturbance events recorded in the tree-ring records could be successfully
related to documentary information of flood events, and those without analogues in the
written archives occurred in the year after the largest events. Noteworthy, the largest
estimated magnitude floods (1936, 1963, 1997, 1999, 2000, and 2005) are in concert with
those years best represented in the dendrogeomorphic record in both the Arenal and
Pelayo Rivers, with the exception of 1936 for which tree age was not sufficient to be
recorded in the trees.
For the years where dendrogeomorphic evidence was clearly visible but documentary
information missing (1966, 1973, 1976, and 1984), we hypothesize that large quantities of
recently deposited solids (sediment and wood) could have been remobilized even by
moderate flows. We therefore speculate that dendrogeomorphic evidence is not always and
necessarily related to extreme events, but that damage can also be inflicted to trees by
ordinary high flows. In that sense, and provided that tree-ring evidence is used in rivers in
absence of other types of documentary data, dendrogeomorphic time series should be
analyzed carefully and more in terms of a contribution to river dynamics rather than as an
exclusive indicator of high flows.
Benito, 2006 and 2008) in which the study sites are located. At the local (basin) scale, a
negative trend exists between discharge and the NAO index, with the largest flood events
documented here being related to highly negative values of the NAO index in 1936, 1963,
1996, or 1997. As such they may be associated with wet conditions in the western
Mediterranean and North Africa (Wanner et al., 1994) and cold air in northern Europe.
For the older period for which discharges could not be estimated (1849-1936), local
data was compared with records at the basin scale. According to Benito et al. (2003), the
period 1870–1900 shows a substantial increase in the frequency of extraordinary floods in
the Tagus Basin, whereas Barriendos (1997) identifies a phase between 1840–1859 for
which a sharp increase in floods and a decrease in droughts. Similar data also exists for
catchments of Eastern and Southeast Spain (Barriendos and Martín-Vide, 1998)
particularly for the period 1830–1870, with a maximum period from 1848 to 1868 and
peaks between 1851–1857 and 1859–1860, and another period between 1916 and 1951,
with peaks between 1917 and 1928. The most exceptional floods affecting the entire Tagus
basin were recorded in 1936, 1941, and 1947, as well as in 1917, 1924, and 1928, in its
headwaters (Benito et al., 2003). Our findings corroborate the body of evidence on past
floods in the Tagus basin and we observe the documented events of 1849, 1853, 1856,
1865, 1919, 1922, 1927, 1928, and 1932 in study region as well. Their occurrence over
large areas also leads us to hypothesize that they were at least of medium or high intensity
and that they were related to intense rainfalls affecting a large part of the Tagus basin.
This seasonality of floods observed in this study also agrees with the distribution
documented by Gaume et al. (2009) or Marchi et al. (2010) for European flood events in
general and for the Mediterranean region in particular.
Our study also shows that the largest magnitude flood of the 20th century in the
study area occurred in 1936 when flow records and gauging stations did not exist in the
wider study region. We therefore compared the estimates of this flood event with data
from the catalogue of maximum observed floods in Spain (IAHS, 2004) and the regional
envelope curve (REC). Our analysis shows quite clearly that the 1936 flood is located in
the upper bound of the REC (Fig. 9).
nalytical steep, this studdy aimed at incorporaating data oon historicaal floods
In a last an
gatherred from documentaryy sources aand tree-rin ng records into at-site and regionnal flood
frequeency analyssis (FFA an nd RFFA, respectivelyy). For thiss purpose a Bayesian Markov
chain Monte Caarlo (MCMC) framew work (Kuczeera, 1999; Reis
R and Steedinger, 200 05; Gaal
VIRGINIA
A RUIZ-VILLANUEEVA 167 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
et al., 2010; Gaume et al., 2010) was used and a likelihood function was built so as to
properly handle the information on historical floods. Following Gaume et al. (2010), we
considered that accurate data on extremes (in terms of discharge) are not absolutely
necessary in such analysis as the dominant information needed is the number of non-
exceedances of a perception threshold (X0) and the historical period (h).
For the period covered by instrumental records, floods are characterized by their
peak flow values, whose determination are neither easy nor accurate, and may be
accompanied by uncertainty or measurement errors. Therefore, it is essential to verify the
value of the maximum peak stage, because approaching the river during a flood may be
impossible (Brazdil et al., 2006). This uncertainty has been avoided in this work due to the
lack of data; just the uncertainty in the discharge estimates has been taken into account.
The challenge of our study is caused by the flood of 1936, as this event was larger
than any discharge recorded during periods covered by continuous hydrological records
and therefore influences the flood frequency analyses. In the Santa Maria basin at
Candeleda village, for instance, the inclusion of historical information evidently reduces
uncertainty and affects quantile estimates. The relatively small reduction in quantiles may
be explained by the fact that the historical estimate of the 1936 event is fraud by relatively
large errors (Reis and Stedinger, 2005), but results including this historical extreme still
seem more reliable since the confidence bounds are reduced significantly.
The particular case of Arenal river at AHIS station is the most difficult to address
with the available data. A short data series exists, and a longer series of historical estimates
has been incorporated in different ways. In all cases, quantiles provided by the regional
approach (as proposed by CAUMAX) seem to be overestimated (underestimated), but
they generally remain within the same order of magnitude and the confidence intervals.
The most important differences between conventional series and those including
tree-ring and documentary sources are found in the regional approach, for which larger
quantiles are observed. This is similar to findings by Gaume et al. (2010) for the Gard
region in France, where the quantile estimation including ungauged events was multiplied
by a factor >2, presumably because of a limited set of ungauged events and a short period
covered by the gauged series. Figure 10 illustrates differences in quantile estimation based
on different approaches, and the related uncertainties expressed by μ1, σ1 μ2 and σ2 (i.e.
mean and standard deviation of quantiles).
Note how thet estimation of the llargest quan ntiles (500 years)
y may bbe strongly affected
depennding on the t treatm ment of thee data, beecoming up p twice vaalue (Fig.10 0). The
underrestimation may resultt in an undderestimatioon of the risk, r but at the same timet the
overesstimation could result in an econoomic overesstimation as well in thee risk manaagement.
Becauuse of this uncertainty in
i the quanntile estimattion, flood hazard
h mapps may be co onceived
since a probabiility appro oach (Rom manowicz and a Beven,, 2003; B ates et al.,. 2004;
Pappeenberger et al., 2006) as
a determiniistic approaaches normaally do not ttake any acccount of
ncertaintiess (Bates et al., 2004; M erz et al., 20005, 2007; Di
the un D Baldassaarre et al., 2010).
VIRGINIA
A RUIZ-VILLANUEEVA 169 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
6. Conclusions
Acknowledgement:
The authors would like to express their gratitude to the Spanish Ministry of Science and Innovation,
for financial support. This work was founded by the MAS Dendro-Avenidas project (CGL2010-19274)
and the Geological Survey of Spain (IGME). We are grateful to the Tagus Water Authority and
Meteorological Agency (AEMET) for having provided meteorological data; the Environment Department
of Castilla y León in Ávila, Arenas de San Pedro and Guisando Councils (particularly to Nuria Blázquez,
Gloria Suárez and Sixto Díaz) for their collaboration. Our special thanks to Fernando Palacios for
providing historical data and his help in the search of archival data. We would also like to mention the
collaboration of TRAGSA in the sampling of cross-sections of trees. The authors want to express her
gratitude to J.A. Fernández-Yuste for his helpful comments on frequency analysis.
REFERENCES
Alestalo, J., 1971. Dendrochronological Interpretation of Geomorphic Processes, 105. Fennia.
Ballesteros Cánovas, J.A., Bodoque, J.M., Díez-Herrero, A., Sanchez-Silva, M., Stoffel, M., 2011a.
Calibration of floodplain roughness and estimation of palaeoflood discharge based on tree-ring
evidence and hydraulic modelling. Journal of Hydrology 403,103-115.
Ballesteros Cánovas, J.A., Eguibar, M., Bodoque, J.M., Díez-Herrero, A., Stoffel, M., Gutiérrez-Pérez, I.,
2011b. Estimating flash flood discharge in an ungauged mountain catchment with 2D hydraulic
models and dendrogeomorphic paleostage indicators. Hydrological Processes 25, 970–979.
Ballesteros, J.A., Stoffel, M., Bodoque, J.M., Bollschweiler, M., Hitz, O., Díez-Herrero, A., 2010a. Wood
anatomy of Pinus pinaster Ait. following wounding by flash floods. Tree-Ring Research 66, 93-103.
Ballesteros, J.A., Stoffel, M., Bollschweiler, M., Bodoque, J.M., Díez-Herrero, A., 2010b. Flash-flood
impacts cause changes in wood anatomy of Alnus glutinosa, Fraxinus angustifolia and Quercus
pyrenaica. Tree Physiology 30, 773-781.
Barriendos, M., Coeur, D., 2004. Flood data reconstruction in historical times from non-instrumental
sources in Spain and France. In: Systematic, Palaeoflood and Historical Data for the Improvement
of Flood Risk Estimation. Methodological Guidelines (ed. by G. Benito & V. R. Thorndycraft), 29–
42. Centro de Ciencias Medioambientales, Madrid, Spain.
Barriendos, M., 1997. Climatic variations in the Iberian Peninsula during the late Maunder Minimum (AD
1675-1715): an analysis of data from rogation ceremonies. The Holocene, 7, 1, E. Arnold, pp. 105-
111.
Barriendos, M. y Martín Vide, J., 1998. Secular Climatic Oscillations as Indicated by Catastrophic Floods
in the Spanish Mediterranean Coastal Area (14th-19th Centuries). Climatic Change, 38, Kluwer
Academic Publishers, pp. 473-491.
Bates, P.D., Horritt, M.S., Aronica, G., Beven, K., 2004. Bayesian updating of flood inundation likelihoods
conditioned on flood extent data. Hydrological Processes 18, 3347–3370.
Benito and Machado, 2012. Floods in the Iberian Peninsula. Changes in Flood Risk in Europe IAHS
Special Publ. 10.
Benito, G., Díez, A., Fernández de Villalta, M., 2003a. Magnitude and frequency of flooding in the Tagus
basin (Central Spain) over the last millennium. Climatic change, 58. 171-192 pp.
Benito, G., Rico, M., Thorndycraft, V. R., Sánchez-Moya, Y., Sopeña, A., Díez Herrero, A., Jiménez, A.,
2006. Palaeoflood records applied to assess dam safety in SE Spain. In: River Flow 2006. Balkema
(Taylor & Francis Group), Oxford, UK.
Benito, G., Sanchez-Moya, Y., Sopeña, A., Machado, M.J., Pérez Gonzalez, A., 2003b. Palaeoflood record of
the Tagus river (Central Spain) during the late Pleistocene and Holocene. Quaternary Science
Reviews 22, 1737-1756.
Benito, G., Thorndycraft, V.R., 2005. Palaeoflood hydrology and its role in applied hydrological sciences.
Journal of Hydrology 313, 3–15.
Benito, G., Thorndycraft, V.R., Enzel, Y., Sheffer, N.A. Rico, M., Sopeña, A., Moya, Y., 2004. Palaeoflood
data collection and analysis. En: G. Benito y Thorndycraft, V.R. (eds). Systematic, Palaeoflood and
Historical Data for the Improvement of Flood Risk estimation. Methodological Guidelines. CSIC,
Madrid. 15-28pp.
Benito, G., Thorndycraft, V.R., Rico, M., Sánchez-Moya, Y., Sopeña, A., 2008. Palaeoflood and floodplain
records from Spain: Evidence for long-term climate variability and environmental changes.
Geomorphology 101, 68–77.
Brázdil, R., Kundzewicz, Z. W., Benito, G., 2006. Historical hydrology for studying flood risk in Europe,
Hydrolog. Sci. J. 51, 739–764.
Brázdil, R., Pfister, C., Wanner, H., Storch, H., Luterbacher, J., 2005. Historical Climatology in Europe –
The State of the Art, Climatic Change 70, 363–430.
Brázdil, R., Pfister, C.,Wanner, H., von Storch, H., and Luterbacher, J., 2010. Historical climatology in
Europe – the state of the art, Climate Change 70, 7–40.
Bullón, 2011. Relationships between precipitation and floods in the fluvial basins of Central Spain based
on documentary sources from the end of the 16th century. Nat. Hazards Earth Syst. Sci. 11, 2215–
2225.
Buytaert, W., Celleri, R., Willems, P., Bièvre, B. D., Wyseure, G., 2006. Spatial and temporal rainfall
variability in mountainous areas: A case study from the south Ecuadorian Andes. Journal of
Hydrology 329, 413-421. 770 DOI:10.1016/j.jhydrol.2006.02.031.
Castellarin, A., 2007. Probabilistic envelope curves for design flood estimation at ungauged sites, Water
Resour. Res. 43.
CEDEX., 2011. Mapa de caudales máximos. Memoria tecnica. http://
www.marm.es/es/agua/temas/gestion-de-los-riesgos-de-inundacion/memoria_tecnica_v2.0_tcm7-
162773.pdf
Chow, V.T., Maidment, D., Mays, L.W., 1988. Applied Hydrology, McGraw‐Hill, New York.
Dalrymple, T., 1960. Flood frequency analysis. US Geological Survey Water Supply paper, 1543A.
Di Baldassarre, G., Schumann, G., Bates, P.D., Freer, J., Beven, K., 2010. Floodplain mapping: a critical
discussion on deterministic and probabilistic approaches, Hydrological Sciences Journal 55, 364-
376,
Díez Herrero, A., 2001. Geomorfología e Hidrología fluvial del río Alberche. Modelos y SIG para la gestión
de riberas. Tesis doctoral. Universidad Complutense de Madrid (inédita).
Díez-Herrero, A., Ballesteros-Cánovas, J.A., Bodoque, J.M., Ruiz-Villanueva, V., 2012. A new
methodological protocol for the use of dendrogeomorphological data in flood risk analysis.
Hydrology Research.
Ferrer, J., 1993. Recomendaciones para el cálculo hidrometeorológico de Avenidas. CEDEX Monografía
M-37. 76 pp.
Ferrer, J., 1996. El modelo de función de distribución SQRT-ET MÁX en el análisis de máximos
hidrológicos: aplicación a lluvias diarias. Tesis doctoral. Universidad Politécnica de Madrid.
Gaál, L., J. Szolgay, S. Kohnová, K. Hlavcová, A. Viglione, 2010. Inclusion of historical information in flood
frequency analysis using a Bayesian MCMC technique: a case study for the power dam Orlík, Czech
Republic, Contributions to Geophysics and Geodesy 40, 121-147.
Gaume, E., Gaál, L., Viglione, A., Szolgay, J., Kohnov´a, S., and Blöschl, G., 2010. Bayesian MCMC
approach to regional flood frequency analyses involving extraordinary flood events at ungauged sites,
J. Hydrol. 394, 101–117.
Gaume, E., Valerie, B., Pietro, B., Newinger, O., Barbuc, M., Bateman, A., Blaˇskovicov´a, L., Bl¨oschl, G.,
Borga, M., Dumitrescu, A., Daliakopoulos, J., Garcia, J., Irimescu, A., Kohnova, S., Koutroulis, A.,
Marchi, L., Matreata, S., Medina, V., Preciso, E., Sempere-Torres, D., Stancalie, G., Szolgay, J.,
Tsanis, I., Velasco, D., and Viglione, A., 2009. A compilation of data on European flash floods, J.
Hydrol. 367, 70–78, doi:10.1016/j.jhydrol.2008.12.028.
Gottesfeld, S., 1996. British Columbia flood scars: maximum flood-stage indicators. Geomorphology 14,
319-325.
Hosking JRM, Wallis JR., 1997. Regional Frequency Analysis. Cambridge University Press: 224 pp.
Johnson RM, Warburton J., 2002. Flooding and geomorphic impacts in a mountain torrent: Raise Beck,
central Lake District, England. Earth Surface Processes and Landforms 27, 945-969.
Kuczera, 1999. Comprehensive at-site flood frequency analysis using Monte Carlo Bayesian inference.
Water Resources Research, VOL. 35, NO. 5, PAGES 1551-1557.
Marchi, L., Borga, M., Preciso, E., and Gaume, E., 2010.Characterisation of selected extreme flash floods in
Europe and implications for flood risk management, J. Hydrol., 394, 118–133,
doi:10.1016/j.jhydrol.2010.07.017.
Martín-Vide J., Barriendos M., 1995. The use of rogation ceremony records in climatic reconstruction. A
case study from Catalonia Spain. Climatic Change 30, 201-221.
Mediero, L., Jiménez, A., 2007. Regional analysis for frequency estimation of annual flood peaks in
ungauged basins of Spain. EGU2007-A-03251. Abstracts of the Contributions of the EGU General
Assembly 2007, Vienna (Austria), 15-20 April 2007. European Geosciences Union. ISSN: 1029-
7006.
Merz, B., Thieken, A.H., Gocht, M., 2007. Flood risk mapping at the local scale: Concepts and challenges.
In: Flood Risk Management in Europe: Innovation in Policy and Practice (Begum, S., Stive, M.J.F.,
Hall, J.W.; Eds.). Series: Advances in Natural and Technological Hazards Research, 25, Springer,
Dordrecht, Chapter 13, 231-251.
Merz, R. and Bloschl, G., 2005. Flood frequency regionalization spatial proximity vs. catchment attributes,
Journal of Hydrology 302, 283–306,
Ortega, J.A., 2007. Paleocrecidas, avenidas recientes e hidroclimatología en la cuenca media y baja del rio
Guadiana. Tesis doctoral. Universidad Complutense de Madrid (inédita).
Pappenberger, F., Matgen P, Beven, K.J., Henry, J.B., Pfister, L., de Fraipont, P., 2006. Influence of
uncertain boundary conditions and model structure on flood inundation predictions, Advances in
Water Resources 29, 1430–1449.
Potenciano, A., 2004. Las inundaciones históricas en el centro-sur de la Península Ibérica. Condicionantes
geomorfológicos y climáticos. Tesis doctoral. Universidad Complutense de Madrid (inédita).
Reis, D. S., Stedinger, J. R., 2005. Bayesian MCMC flood frequency analysis with historical information.
Journal of Hydrology, 313(1-2), 97-116. doi:10.1016/j.jhydrol.2005.02.028.
Rico, M., 2004. Las paleocrecidas en la cuenca media del Río Segre durante el Pleistoceno superior-
Holoceno: registros morfosedimentarios y análisis hidroló- gico. PhD thesis, University of Zaragoza,
Spain.
Rico, M., Benito, G., Barnolas, A., 2001. Combined palaeoflood and rainfall–runoff assessment of
mountain floods (Spanish Pyrenees). Journal of Hydrology, 245(1-4), 59-72. doi:10.1016/S0022-
1694(01)00339-0.
Romanowicz, R., Beven, K., 2003. Estimation of flood inundation probabilities as conditioned on event
inundation maps, Water Resources Research 39 (3), 1073–1085.
Ruiz-Villanueva, V., Díez-Herrero, A., Stoffel, M., Bollschweiler, M., Bodoque, J.M., Ballesteros, J.A., 2010.
Dendrogeomorphic analysis of flash floods in a small ungauged mountain catchment (Central
Spain). Geomorphology 118, 383-392.
Shroder, J.F., 1980. Dendrogeomorphology: review and new techniques of treering dating. Progress in
Physical Geography 4 (1): 161-188.
St. George, S., Nielsen, E., 2003. Palaeoflood records for the Red River, Manitoba, Canada, derived from
anatomical tree-ring signatures. Holocene 13 (4), 547-555.
Stoffel, M., Bollschweiler, M., 2008. Tree-ring analysis in natural hazards research – an overview. Natural
Hazards and Earth System Sciences 8: 187–202.
Tarolli, P., Borga, M., Morin, E., & Delrieu, G., 2012. Analysis of flash flood regimes in the North-Western
and South-Eastern Mediterranean regions. Natural Hazards and Earth System Science, 12(5), 1255-
1265. doi:10.5194/nhess-12-1255-2012.
Temez, J. R., 1991. Extended and improved Racional Method. Version of the Highways Administration of
Spain. Proc. XXIV Congress. España. Vol. A. 33-40pp.
Thorndycraft, V.R., Barriendos, M., Benito, G., Rico, M., Casas, A., 2006. The catastrophic floods of AD
1617 in Catalonia (northeast Spain) and their climatic control. Hydrological Sciences Journal 51
(5), 899–912.
Trigo R.M., Pozo-Vazquez D., Osborn T.J, Castro-Diez Y., Gámis-Fortis S., Esteban-Parra M.J., 2004.
North Atlantic Oscillation influence on precipitation, river flow and water resources in the Iberian
Peninsula. International Journal of Climatology 24, 925-944.
Viglione, A., 2009. nsRFA: Non-supervised Regional Frequency Analysis, R package.
Wanner H., Brazdil R., Frich P., Frydendahl K., Jonsson T., Kington J.A., Pfister C., Rosenorn and
Wishman E., 1994. Synoptic interpretation of monthly weather maps for the late Maunder
Minimum 1675-1704. In: Frenzel B., Pfister C. y Glaser B. (Eds.). Climatic Trends and Anomalies
in Europe.
Yanosky, T.M., Jarrett, R.D., 2002. Dendrochronologic Evidence for the Frequency and Magnitud of
Palofloods. Ancient Floods, Modern Hazards: Principles and Application of Paleoflood Hydrology.
Water Science and Application 5, 77-89.
Zielonka, T., Holeksa, J., Ciapala, S., 2008. A reconstruction of flood events using scarred trees in the Tatra
Mountains, Poland. Dendrochronologia 26, 173–183.
REFERENCE:
Ruiz-Villanueva, V., Bodoque, J.M., Díez-Herrero, A., Eguibar, M.A. Pardo-Igúzquiza, E.
(2012). Reconstruction of a flash flood with large wood transport and its influence
on hazard patterns in an ungauged mountain basin. Hydrological Processes.
DOI: 10.1002/hyp.9433
ABSTRACT
The reconstruction of past flash floods in ungauged basins leads to a high level of
uncertainty, which increases if other processes are involved such as the transport of large
wood material. An important flash flood occurred in 1997 in Venero Claro (Central
Spain) causing significant economic losses. The wood material clogged bridge sections,
raising the water level upstream. The aim of this study was to reconstruct this event,
analysing the influence of woody debris transport on the flood hazard pattern. Since the
reach in question was affected by backwater effects due to bridge clogging, using only
HWM or PSI may overestimate discharges, and so other methods are required to estimate
peak flows. Therefore, the peak discharge was estimated (123 ± 18 m3 s-1) using indirect
methods; but 1D-hydraulic simulation was also used to validate these indirect estimates
through an iterative process (127 ± 33 m3 s-1), and reconstruct the bridge obstruction to
obtain the blockage ratio during the 1997 event (~48%) and the bridge clogging curves.
Rainfall-runoff modelling with stochastic simulation of different rainfall field
configurations also helped to confirm that a peak discharge greater than 150 m3 s-1 is very
unlikely to occur, and that the estimated discharge range is consistent with the estimated
rainfall amount (233 ± 27 mm).
It was observed that the backwater effect due to the obstruction (water level ~7 m)
made the 1997 flood (~35-year return period) equivalent to the 50-year flood. This
allowed the equivalent return period to be defined as the recurrence interval of an event of
specified magnitude which, where LWD is present, is equivalent in water depth and extent
of flooded area to a more extreme event of greater magnitude. These results highlight the
need to include obstruction phenomena in flood hazard analysis.
Keywords: flood hazard analysis, large wood material, ungauged basin, clogging
curves, equivalent return period.
1. Introduction
The high potential risk associated with flash floods in mountain areas is due to a
rapid and complex catchment response (Weingartner et al., 2003). Besides high water
levels in the drainage network, important geomorphic changes and large transport of
material (i.e., wood) must be considered as additional factors.
Spatial and temporal scales of flash floods, combined with the space-time scales of
conventional measurements of rainfall and discharge networks, make it particularly
difficult to characterize these events properly (Borga et al., 2008; Marchi et al., 2009;
Smith et al., 2011). This type of flood phenomenon may therefore be poorly documented
(in terms of data quantity and quality), particularly in mountainous areas (Sivapalan et al.,
2003).
The mobilization of woody material in rivers has been considered in the past, but
very few studies have included this phenomenon in flood hazard and risk analyses (see
Mazzorana et al. (2011) and references cited therein).
In Spain, hydrometeorological networks mainly provide daily data, but the most
dangerous floods occur in the space of a few hours. This may mean that the hydrologic
response is measured with an inappropriate time interval (Camarasa and Tilford, 2002).
As well as this limitation, the rain gauge network in the Spanish Central System is sparse
and unevenly distributed, covering mainly valley and lowland areas, a common problem in
mountainous regions (Buytaert et al., 2006). Flow gauges are mainly installed in large river
basins, but even where they exist in smaller mountain catchments, the flow time series may
be wrongly recorded during extreme events, due to the damage often caused to gauges by
the event itself or because discharge exceeds recordable levels. This means that a lack of
input data often implies a high level of uncertainty in the assessment of flood hazard and
risk in ungauged mountain basins.
An important flash flood occurred in 1997 in the Arroyo Cabrera catchment, a small
ungauged forested mountain catchment in Central Spain. The documentation of this flash
flood reveals high peak discharges and a complex flood response. The main aim of this
study, therefore, was to reconstruct this event in terms of magnitude and also to
understand the process interaction, particularly how the obstructive effect of wood at
bridges affected the hazard level. Three stages were defined to achieve this objective: (i)
estimation of the flash flood water level and peak discharge; (ii) reconstruction of the
clogging process. (iii) reconstruction of the hydrological catchment response and
estimation of the rainfall required to generate the peak discharge, evaluating different
rainfall field configurations.
This paper also attempts to answer the important question of how large wood affects
the event magnitude and hazard patterns in general. To answer this question the clogging
curves were designed and the concept of an equivalent return period was defined.
The study site is in the Arroyo Cabrera catchment, a torrential tributary of the
Alberche River in the Tagus River Basin, on the northern slopes of the Sierra de Gredos,
Spanish Central System (see Figure 1 A and B). The catchment area is over 15.5 km2. The
maximum difference in height within the watershed is 1188 m, and the main channel is
5500 m long, with an average slope of 21.6%.
The local forest stand is formed predominantly by Pinus pinaster Ait., Pinus
sylvestris L. and Quercus pyrenaica Willd. Riparian broadleaved species Alnus glutinosa
(L.), and Fraxinus angustifolia (Vhal.) can be found on both river banks.
Figure 1: (A) Location of the study site within the Spanish Central System. (B) Drainage
basin to the bridge section. The Arroyo Cabrera watershed has been monitored since 2004,
with 6 rain gauges and 1 stream gauge installed (shown on map). (C) Detail of the study
area, with main reaches described in the text. The study site (Reach 5) is marked in red.
Three small lobes were deposited upstream of the avulsion point. The secondary
channel developed and extended 1.3 km, breaking away from the main channel after 160
m, with a maximum width of 35 m (between Reaches 3 and 4). While most of the
sediment material was deposited in the lobes or carried away by the main channel, most of
the large woody debris (LWD) delivery was recruited in the secondary channel.
The main channel extended 1.4 km, with maximum width 40 m, until it met the
secondary channel again and they were reunited into a single channel 1 km upstream of the
confluence with the Alberche River.
Reach 3 includes the Trampalones bridge, which was demolished by a large wood jam
deposit transversal to the flow direction. The volume of the wood jam was approx. 36 m3
but most of the air pockets in the jam were filled with fine sediment or coarse woody debris
(inferred from photos). In Reach 3, 400 m of a rural path was destroyed.
In Reaches 4 and 5 (Figure 2A) evidence of the water level can be clearly seen (high
water mark (HWM) on the small white building), with toppled and other damaged trees
(scars as palaeostage indicators (PSI); Ballesteros et al., 2011a and 2011b). The most
important action in this reach was the destruction of a small bridge across the river from
the small building to the other side. Only the bridge abutments are still standing, and a
gravel bar has developed upstream on the right bank. Some wood pieces, average length 6
m and average diameter 0.4 m, were deposited parallel to the flow direction and with the
root system upstream.
Finally, Reach 5 (Figure 2B), with the largest obstacle in the stream, is the most
critical reach for flood hazard. Here important deposits of wood and boulders were
observed, with some imbricate boulders lying on top of logs on the right bank. Some coarse
woody debris was also found on top of the larger pieces. Some of the logs are parallel to the
flow direction but others are transversal. The logs are of average length 7 m and average
diameter 0.4 m (inferred from photos). The volume on the right side of the bridge was
estimated at 85 m3, and most of this non-wood volume was filled with sediment. The
volume on the left side was 60 m3 (volumes inferred from photos). The evidence from
observations, images and other indicators (e.g. fine sand and coarse wood deposits found
on the bridge deck), show that the bridge was flooded due to the clogged wood.
Figure 2: Images taken the day after the 1997 event. Photos correspond to Reach 5 in Fig.1.
The reach studied is therefore in Reach 5 (length 250 m, altitude 805-755 m a.s.l.).
The stream bed morphology is a constrained reach without wide floodplains, typical of a
bedrock river. This reach was not severely modified, guaranteeing that changes in channel
geometry were minimal after this flood event.
Table 1: Main physical characteristics of the Arroyo Cabrera catchment and reach studied
3. Methodology
As described above, there was no data available for this event from stream gauges,
radar rainfall estimates or sub-daily rain gauges and the event was characterized by a
backwater effect due to large wood transport. This resulted in some difficulties and
uncertainties in the event reconstruction. To reduce these uncertainties different
approaches were used to estimate the main flood parameters, although the method was
focused on analyzing the influence of wood on the hazard pattern.
estimate the peak discharge and evaluate different scenarios to quantify the influence of the
large woody debris on the flood magnitude, due to bridge clogging processes. (3)
Reconstruction of the hydrological response of the catchment combined with the
stochastic simulation of the precipitation required for the estimated peak discharge.
The first step was to identify the HWM and PSI and survey them to estimate the
water level in different reaches (between main Reaches 5 and 6 marked in Fig.1). These
included the HWM visible on the small building located on the streamside (Fig. 2A) and
on the bridge deck (Fig. 2B). However, the reach studied was affected by backwater effects
due to bridge clogging. Therefore, using only HWM or PSI may obtain overestimated
discharges, and other methods are required to estimate discharge when this occurs.
First, the peak discharge was assessed using the slope-conveyance method (Gaume
and Borga, 2008; Marchi et al., 2009) in the section furthest upstream of the bridge, less
influenced by obstruction.
Indirect peak discharge estimates of this type are usually potentially affected by
different sources of uncertainty, which may be induced by errors in HWM assessment and
in the choice of roughness coefficients, in the use of the post-flood geometry and its survey,
in the assumptions concerning the energy line slope and, in this particular case, by the
backwater effect. Taking these uncertainties into account, a discharge range was established
and compared with other approaches (rainfall-runoff and hydraulic modelling).
An iterative hydraulic simulation was carried out in the studied reach using the step-
backwater method (HEC-RAS 4.1; USACE, 2010b) in steady conditions.
The reach geometry was produced using a combination of differential GPS and Total
Station (cross section 10 m maximum spacing). The bridge structure was mapped more
precisely (measuring the road on the deck, the arch and the entire structure). Topographic
data surveyed (accuracy(z) =10 cm) with the available DEM (pixel size 5 m, accuracy(z) =
25 cm) were used to obtain a more detailed DEM of the study reach with a spatial
resolution of pixel size 0.65 m (accuracy(z) = 10 cm in the riverbed, 25 cm in the rest).
This new DEM was used for further analysis with HEC-GeoRAS 4.3 (USACE, 2011) to
compute the water surface elevation maps.
The Manning roughness coefficient was established from field observations and
aerial orthophotos from 2000 - 2010, taking into account the evolution of the vegetation
within the stream to assess the correct values for the vegetation stand in 1997. This
roughness coefficient was established on both banks and riverbed for each cross section
according to the criteria used by Ballesteros et al. (2011a) in the same area.
Several discharges (ranging from 10 to 200 m3 s-1) were then input into the model
varying the hydraulic radius in the bridge section. This was carried out reducing the cross
sectional area by 10% each time in an iterative procedure using the block obstruction tool.
This obtained the ratio between the percentage of occlusion and the water level for a given
discharge, called here clogging curves. HWM and PSI were then used again to assess a
range of peak discharges, now taking into account the bridge occlusion in the simulation.
The next step was to analyze the flash flood frequency for the estimated discharge.
To attempt to assign a return period value to the 1997 peak flow, data obtained from an
available regional study of maximum discharges (CEDEX, 2011) was examined. In this
study, frequency analyses of catchments smaller than 50 km2 are carried out using the
rational method.
However, the authors’ own short recorded data series allowed to be computed with
the estimated discharge for the 1997 event, adjusted using different frequency distributions
(GEV, Gumbel and LogPearson III)
Where the precipitation data network does not have rainfall gauges located at high
elevations, the main risk is underestimating precipitation values at higher points (Vélez et
al., 2007). Therefore, simulations are needed to model stochastic behaviour of the climatic
system. It is also advisable to use secondary information from an easy-to-measure variable,
with complete coverage of the study area and correlated with rainfall, such as altitude
obtained through a DEM. The advantage has been shown of using the DEM together with
rain gauge data to estimate daily rainfall (Pardo-Igúzquiza, 1998; Goovaerts, 2000).
Stochastic simulation of these daily data from the 10 stations provided a description
and a measure of uncertainty of the spatial variability of this phenomenon. This was done
by generating multiple realizations (500) of the stochastic process modelling the spatial
distribution under study. All the simulation tools implemented belong to the sequential
simulation category using the Stanford Geostatistical Modeling Software (SGeMS, Remy
et al., 2009), and the COSGSIM algorithm which allows a co-kriging solving system to be
chosen.
To limit the simulation results some data available for the storm was used. According
to eye witnesses and news reports (extracted from Díez-Herrero, 2003), on that day it
started to rain in the Sierra de Gredos around 18.00 (local time), the most intense rainfall
was at 23.30, and by around 06.00 the frontal rain was over the Mediterranean sea.
Therefore, a storm duration of around 12 h can be assumed with an approximately
symmetrical hyetograph.
In addition, previous work in the study area (Ruiz-Villanueva et al., 2011) established
an intensity-duration precipitation threshold for triggering the 1997 shallow landslide.
According to this study the critical rainfall amounts ranged from 215-324 mm depending
on the storm duration (6-24h). These amounts are the minimum values needed to trigger
the landslide. The total duration of the storm was about 12 hours (~264 mm based on this
threshold), but the range was extended here to 6-24 h (215-324 mm) to establish
minimum and maximum boundaries. The stochastic simulations that met this imposed
condition (minimum precipitation 215 mm and maximum precipitation 325 mm) and
were representative of the known orographic pattern of rainfall usual in this area, were
assumed to be the most reliable and were selected for further evaluation.
The intensive monitoring of this small basin since 2004, with 6 rain gauges and 1
flow gauge, together with an in-depth knowledge of the landscape and geomorphic
structure of the area, allowed the calibration and validation of the hydrological simulation.
The excess precipitation was transformed into runoff using the SCS dimensionless
unit hydrograph method. Channel routing was simulated implementing the Muskinghum-
Cunge method. For the baseflow, the recession method was implemented.
Calibration and validation of the hydrological model was limited by the short time
series, with records for only 7 years. It was performed for 9 selected recorded events
exceeding a threshold of 12 m3 s-1 (water level 1 meter) in the bridge section with a stream
gauge installed since 2004. (These events were the highest recorded in the basin until
2010). An automatic calibration routine was used for this where the sum squared residuals
and univariate method were used as objective function and search algorithm, respectively.
Once the model had been calibrated, parameters were kept constant, with the
exception of those describing the antecedent soil moisture conditions. For the 1997 event
the antecedent conditions were wet, based on the accumulated precipitation over the 5
previous days (see Ruiz-Villanueva et al., 2011 for details). Several scenarios were simulated
and the basin response was analysed to different design hyetographs resulting from the
disaggregation of the reliable stochastic realizations. These 15-minute time interval
hyetographs were obtained by disaggregating daily data using the available intensity-
duration curves for Spain (Salas and Fernández, 2007). Three hyetotypes and three storm
durations (6, 12 and 24 h) were used. These hyetotypes (Montescarchio et al., 2009) were:
(A) triangular increasing rate hyetograph; (B) triangular decreasing hyetograph; (C)
isosceles triangular hyetograph.
4. RESULTS
According to the field work, including surveying of high water marks (HWM), water
surface slope and cross-sectional geometry at multiple sites along the stream reach, the
water depth in the bridge section was 7.1 meters. The same procedure was carried out
using the HWM observed on the small building and PSI scars on two trees. The water level
upstream of this section was~3 m (HWMS6= 7.1 m; HWMS5= 2.9 m; PSI1= 2.8 m; PSI2=
3.2 m).
0.08 m1/2 s-1 to 0.15 m1/2 s-1 on vegetated banks; and from 0.08 m1/2 s-1 to 0.1 m1/2 s-1 in the
main channel (higher values where large boulders were present).
The results obtained from the slope-conveyance method provided values ranging
from 126 - 183 m3 s-1 (central value 155 m3 s-1; 126 m3 s-1 in the section furthest from the
bridge, and therefore less affected by the backwater). However, because of the clogging
process, these values are likely to be overestimated. Flow velocity ranges from 4.68 to 1.84
m s-1 depending on the section, but the mean in the entire reach is 3.61 m s-1. The average
reach Froude number is 0.7 (ranging from 0.34 to 0.83).
The average diameter of the larger particles ranged from 400 to 600 mm. According
to our calculations the flow velocity required to transport this type of material ranges from
3 to 4 m s-1 (central value 3.5 m s-1); using the cross sectional area (section 6, 35 m2) the
discharge range is estimated in the bridge section at 105-140 m3 s-1 (central value 123 m3 s-
1
). The average Froude number according to this velocity is 0.6.
The estimated discharge in the less affected section using the slope-conveyance
method (126 m3 s-1) is similar to the central value obtained by the competence method
(123 m3 s-1), therefore the range 105-140 m3 s-1 is used here as an indirect discharge
estimation.
T
The results ofo the iteraative proceddure describbed in section 3.2 aree expressed in the
cloggingg curves shoown in Fig. 3.
VIRGINIA RUIZ-VILLANUEVA
A 192 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
When a peak discharge of 123 m3 s-1 was simulated with an obstruction of 48% of the
bridge section, the result satisfies the water level estimates based on field indicators (HWM
and PSI). Flow velocity computed by the hydraulic model ranged up to 4 m s-1 (mean 2.85
m s-1), confirming the estimates provided earlier.
The simulation was also run without any bridge obstruction for the same discharge
to analyze the influence of the obstruction (Fig.4).
This simulation without bridge obstruction was calibrated using the rating curve
obtained for the bridge cross-section by hydraulic modelling and with the recorded data
and field measurements (Fig.5).
Figuure 5: Ratinng curve forr the bridgee section witth stream gaauge installe
led since 200
004.
VIRGINIA RUIZ-VILLANUEVA
A 194 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
Figure 6: Water depth for the estimated discharge of the 1997 flash flood, (A) without any
obstruction at the bridge, (B) taking into account the LWD and obstruction at the bridge,
(C) water surface with LWD (in red) and without LWD (blue). Gas deposits and camping
site are shown on the maps.
Fig. 6 shows clearly that the water surface elevations are different in the absence of
wood. This suggests that specific areas will flood more frequently than expected if large
wood is transported.
Frequency curves are plotted in Fig. 7. The return period of the 1997 estimated
discharge applying the CEDEX method is 80-200 years (continuous black line ; depending
on the range 123 ± 18 m3 s-1). However, if the authors’ own short recorded data series is
computed together with the estimated discharge for the 1997 event, and adjusted using
different frequency distributions (GEV, Gumbel and LogPearson III; Fig. 10, dotted grey
lines) different results are obtained. For the 1997 event the recurrence time would range
from 25-45 years (depending on the range 123 ± 18 m3 s-1).
For higher discharges the uncertainty increases due to the short data series. But the
most important observation is that these curves are all above the other curve shown, so that
the frequency using the rational method may be underestimated.
n conclusion
In n, the 1997 7 estimatedd discharge would
w havee a return pperiod of ~335±10
years (bbased on thhe mean of our approxximations),, however the t wood cllogging maakes its
consequuences equiivalent in water
w level too a higher magnitude
m event
e (~50 year event). This
hypotheesis enabless us to intro oduce the nnew concep pt of the eq
quivalent ret
eturn periodd, used
here as tthe recurren
nce intervall for an evennt of a givenn magnitudde that in thhe presence of any
obstrucction (in th his case wood)
w is eqquivalent inn water levvel to an event of greater
g
magnituude. An iterative process was carrried out agaain to analyyse this effeect. Simulattion of
several discharges and differrent obstruuctions alloowed the curves c showwing the defined
d
equivaleent return period
p to be
b computedd (the curvve estimated d with the aauthors’ reccorded
data and the one obtained
o frrom CEDE EX data) an nd the perceentage of brridge obstrruction
(Fig. 8)..
VIRGINIA RUIZ-VILLANUEVA
A 196 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
Figgure 8: Equi
uivalent retuurn period. C
Curve repreesenting the effect of w
wood cloggiing in
criticaal sections in
i flood hazzard analysi s. The brokken red line is the estim
mated return n period
for thhe 1997 eveent based on he continuo
n the recordded short data series; th uous blue lin
ne is the
returnn period proovided by CEDEX
C . Th
The dotted black
bl line is the
t mean an and the greyy shading
shows tthe uncertaainty area.
VIRGINIA
A RUIZ-VILLANUEEVA 197 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER
C 4
T
The Gaussian n variogram
m model choosen for thee three variograms (raiinfall, topoggraphy
and croossing varioogram) was exponentiaal as is usu
ual in rainfaall simulatiions (Grimes and
Pardo-IIgúzquiza, 2010)
2 s equal to 1. The varriogram ran
with no nugget effect and sill nge was
46,200 m (Fig. 9).
Fi
Figure 9: Expperimental variogramss for DEM anda precipit
itation. Blacck line is the
he
exponeential modeel applied. Numbers
N ovver squares are
a data paiirs for precip
ipitation dat
ata (for
DEM varriogram moore than 10000 pairs werre used).
O
Of the 500 computed
c stochastic
s siimulations,, 70 fulfilled
d the condiitions impo
osed in
the stuudy, and 588 of these agreed wi th the reggional orogrraphic patttern. The spatial
represenntation of one
o of thesee reliable sim
mulations iss shown in Fig.10.
F
VIRGINIA RUIZ-VILLANUEVA
A 198 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
Figure 10: Example of stochastic realization which fulfilled the conditions and so is reliable
for use in hydrological simulation. Daily registered data is placed over the station (blue
squares).
The calibration results show a good fit between simulated and observed peak
discharges, with a Nash–Sutcliffe efficiency equal to 0.73 and a correlation coefficient of
0.79.
Therefore, 432 independent stochastic simulations were performed for all the
combination of hyetotypes and rainfall volume, to obtain the spatially distributed daily
rainfall including the orographic effect.
T
The total prrecipitation
n in the cattchment froom the selected scenaarios was used
u to
extrapoolate from recent
r rainffall amountts (recorded
d within thee catchmennt) the amo
ount of
rainfall needed to produce a discharge of 104 - 150 m s (central valuue of 123 m3 s-1)
3 -1
VIRGINIA RUIZ-VILLANUEVA
A 200 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
5. Discussion
The discharge estimation was carried out first by means of empirical equations.
Assuming the backwater effect due to bridge obstruction, the slope-conveyance method
was applied and result obtained in the furthest section from the bridge was 126 m3 s-1.
Costa (1983) reviewed four theoretical and empirical methods for velocity
estimation for boulders transported by flash floods, finding differences in velocity between
the four methods ranging from 17 to 30%, and obtaining the equation V = 0.18 d0.487.
Costa (1983) demonstrated that the uncertainty of palaeodischarge estimates derived from
this approach has an average error of 28% (USACE, 2003). This method has been applied
by different researchers (Costa, 1983; Gregory, 1983; Williams, 1983; Baker et al., 1984;
Johnson and Warburton, 2002; Kehew et al., 2010; among others) to reconstruct past
floods.
Since the estimated discharge using the slope-conveyance method (126 m3 s-1) for the
section least affected (by the backwater effect) was similar to the central value obtained by
the competence method (123 m3 s-1), the range 105-140 m3 s-1 was used here as indirect
discharge estimation. This range was also corroborated by the iterative hydraulic process
(with the range between 95 and 160 m3 s-1; 127± 33 m3 s-1).
Although the objectives, the studied reach, data and methods used differ from those
applied in this study, our estimations were also compared with a previous study of the 1997
event (Ballesteros et al., 2011a). These authors investigated the use of scarred trees for the
peak discharge estimation in a different reach downstream of the bridge. They estimated
145 m3 s-1 as the minimum peak discharge for which all scars are submerged. This value
therefore represents a PSI reference level for estimating the hypothetical flow depth. Then,
by using several scenarios they established a peak discharge of 79 ± 14 m3 s-1 according to
the minimized deviation between tree scar height and simulated water elevation. As they
point out, several bridges upstream of the study site could reduce the peak flow in their
study reach. It should be pointed out that the bridge obstruction triggered a critical flow
downstream and therefore the flow conditions above and below this section were also
different. On the one hand, their estimates are affected by the bridge lamination, and on
the other hand, the flow may be affected by damming resulting in several pulses or peaks
(interferences in the flow) in the discharge. The HWM and PSI analysed are assumed to be
the maximum envelope of these peaks.
Therefore, considering these effects and the uncertainties of both methods, the order
of magnitude agrees with our results (according to the upper and lower limits of both
estimations the discharge would be of the same order of magnitude: 103 ± 38 m3 s-1).
Flow velocity was also estimated as approximately 3.5 m s-1. This velocity is,
according to Marchi et al. (2010), characteristic of flash floods for catchments of less than
350 km2 (Lumbroso and Gaume, 2012).
The Froude number was approximately 0.65, which is also consistent with several
studies which argue that supercritical flow cannot be sustained over reaches longer than 20
m (Jarret, 1984; Grant, 1997; Lumbroso and Gaume, 2012), particularly when bed
material is transported (Bathurst et al., 1979). Nevertheless, a supercritical flow could be
expected for the reach geometry based on channel slope. In the studied reach, the
subcritical flow is also imposed by the bridge obstruction and the flow condition changes
to supercritical downstream.
The hydraulic simulation was run using the HEC-RAS model. This was used for two
different purposes; first, to reproduce the clogging process (computing the clogging
curves); and also to reconstruct the 1997 event validating the indirect estimates
(obstructions between 30 to 60% need discharges between 95 and 160 m3 s-1, central value
of 127± 33 m3 s-1, to fit the observed water level). This model has very often been used
successfully to reconstruct past floods (some examples are Agget and Wilson, 2009;
Balasch et al., 2010; Benito et al., 2003a and b; Gül et al., 2009; Horrit and Bates, 2002;
Jacoby et al., 2008; Koutroulis and Tsanis., 2010; Ortega and Garzón, 2009; Thorndycraft
et al., 2005).
Although roughness coefficients could not be calibrated along the main channel due
to lack of data, the uniform bedrock present in the study reach greatly helped the
assignment of Manning’s values. In addition, we applied the same criterion used by
Ballesteros et al. (2011a) and field observations and photographs. Nevertheless, the
Manning coefficient is considered to be of low sensitivity when calculating high magnitude
flood flow discharges (O'Connor and Webb, 1988; Enzel et al., 1994) showing that a 20 %
change in this parameter produced a change of less than 5 % in the corresponding
discharge. This is particularly the case here, since the energy losses due to roughness are
lower if we take into consideration that the main process is the obstruction and the
subsequent change in flow conditions and backwater effect.
In addition, in this study the bridge structure and the cross-sectional reduction were
simulated with HEC-RAS (using the block obstruction tool), and also the water level. The
choice of the block tool was based on the bridge geometry, with a single arch without piers.
As can be observed in the photos (Fig. 2), this bridge design determined the wood
accumulation. In this case, the logs were deposited on the bridge laterals, and no wood was
found on the bridge deck. The tool allows areas of the cross-section to be defined that will
be permanently blocked on both sides of the bridge. This method differs from that used by
Mazzorana et al. (2011a), since the bridge geometry is also different.
concept of the equivalent return period has been established. These important
contributions have enabled improved risk analysis taking this process into account. Three
types of behaviour were observed in the clogging curves: the lower threshold where the
slight increase in water level is linear when the clogging percentage increases; the critical
and most important threshold, where small increments in clogging percentage make the
water level rise rapidly, until the upper threshold is reached, where the maximum water
level remains almost constant.
This behaviour can be explained by flow regime changes. First, in the LT region the
flow regime is supercritical, and the water level apparently increases linearly (sub- parallel
curves) with increasing obstruction ; then, the regime changes to subcritical, and the water
level increases abruptly (CT). Last, in the UT region (higher discharges and/or total
obstruction) the wetted perimeter increases very slowly , then the curve slope is less steep
than in the first region. These curves are only valid for this bridge, but the methodology to obtain
them is can be easily reproduced, and the three main sectors found in the envelopes will be similar for
other similar bridges, since these are related to the hydraulic conditions.
The equivalent return period is defined as the recurrence interval for an event of a
given magnitude that in the presence of obstructions corresponds to a more extreme event.
On the study site the peak discharge with no bridge obstruction increases in magnitude up
to eight times depending on the increase in obstruction.
Radar rainfall estimates are of great importance in flash flood studies (Marchi et al.,
2009). However, radar data have not been used here because there was no weather radar
data available for the 1997 event and there were only 10 daily rain gauges irregularly
distributed in the study area in 1997. Therefore, secondary information was used with the
altitude obtained from a DEM as an easy-to-measure variable, with complete coverage of
the study area and correlated with rainfall. The advantages of using the DEM together
with rain gauge data in estimating daily rainfall have been shown (Pardo-Igúzquiza, 1998;
Goovaerts, 2000). The same is true when generating geostatistical simulations of daily
rainfall.
The HEC-HMS rainfall-runoff model has also been applied (Scharffenberg and
Fleming, 2005; Fleming and Neary, 2004; Maskey et al., 2004; Cunderlik and Simonovic,
2007) to estimate the possible rainfall field configuration and duration.
Here, moderate events (but the largest recorded, most of them in autumn-winter,
therefore with similar initial conditions to the 1997 event) have been used to calibrate the
rainfall-runoff model to simulate a more extreme event. This procedure could have
resulted in increased model uncertainty. The main limitation could be related to the initial
abstractions and therefore the peak discharge obtained could be underestimated, due to the
model calibration using medium-high frequency events. As a result, probably the losses
occurring in an extreme event such as the one studied here would be lower than those
simulated by us. The wet initial conditions (low initial abstractions) established for the
1997 event could help to reduce this uncertainty.
On the other hand, the inverse case, the identification of parameters in an extreme
flood for use in moderate events, has previously been used successfully (Ruiz-Villanueva et
al., accepted for publication). The work presented here shows that calibrating parameters
of moderate events for use in more extreme flood cases may be also useful when initial
conditions are properly accounted for.
The simulations showed that a 12 hour storm (the approx. total duration of the
storm), triangular increasing rate and isosceles triangular rate hyetographs, with simulated
maximum rainfall intensities ranging up to 25 mm h-1 and total precipitation over the
catchment ranging from 205 to 260 mm reproduce the catchment response during the
1997 event.
This event may be considered as a prototype of the likely storms responsible for flash
flood peaks in Central Spain, and this assumption agrees with the seasonality found by
Marchi et al. (2010), that in Mediterranean regions (Spain, France and Italy) the highest
occurrence of flash floods is in autumn, whereas events in inland Continental regions
(Austria, Rumania and Slovakia) commonly occur in summer; this, as well as that events in
the Mediterranean region are usually larger and more intense than in Continental regions,
reveal different climatic forcing.
6. Conclusions
This study presents the reconstruction of the 1997 flash flood with large wood
transport occurring in Sierra de Gredos (Central Spain). There was no available recorded
data from stream gauges, radar rainfall estimates or sub-daily rain gauges for this event. The
event was characterized by a backwater effect due to large wood transport, and this leads to
difficulties and uncertainties in the indirect estimations. Therefore, indirect methods
together with hydraulic simulation and rainfall-runoff modelling have been applied to
obtain the main flash flood parameters.
This procedure also allowed the clogging curves to be obtained for this bridge. The
obstruction due to the wood load influenced hazard patterns, resulting in estimated water
discharge levels different from those expected in the absence of wood (with 2 m difference
in this case). This allowed the new concept of the equivalent return period to be defined as
the recurrence interval of an event where an obstruction is present, the consequences of
which are equivalent in water depth and therefore in flood extent to a more extreme event.
The reconstruction of these past events will provide a better understanding of the
catchment response and they will also improve risk analysis. The order of magnitude
estimation of precipitation and discharge, and the post-event evaluation showed that the
magnitude of this event was seriously influenced by the large quantities of wood
transported in the flow and how this affected critical sections like bridges. These
observations highlight the need to improve the study of woody debris transport during
floods. Finally, the concepts of the clogging curves and the equivalent return period may
help to understand the influence of this (and other processes) in flood hazard and risk
analysis.
Acknowledgements
The authors would like to express their gratitude to the Spanish Ministry of Science and Innovation,
for financial support. This work was founded by the MAS Dendro-Avenidas project (CGL2010-19274)
and the Geological Survey of Spain (IGME). We are grateful to the Tagus Water Authority, Environment
Department of Castilla y León in Ávila, Caja Ávila, Asocio de Ávila and Navaluenga Council for their
collaboration. Our special thanks the forester José Luis Galán for his assistance in the field. We would also
like to mention Juan Ballesteros, Ignacio Gutierrez, Tasio Fernández, Leticia Salas, Carolina Guardiola and
Ángel Prieto for contributing suggestions and interesting discussions which significantly improved this
paper. Thanks to the contribution of 2 anonymous reviewers the quality of the early version of this
manuscript was improved.
REFERENCES
Aggett GR, Wilson JP. 2009. Creating and coupling a high-resolution DTM with a 1-D hydraulic model in
a GIS for scenario-based assessment of avulsion hazard in a gravel-bed river. Geomorphology 113:
21-34. DOI: 10.1016/j.geomorph.2009.06.034.
Andreoli A, Comiti F, Lenzi MA. 2007. Characteristics, distribution and geomorphic role of large woody
debris in a mountain stream of the Chilean Andes. Earth Surf. Process. Landforms 32, 1675–1692.
DOI:10.1002/esp.
Baker VR. 1984. Flood Sedimentation in Bedrock Fluvial Systems. In: Sedimentology of gravels and
conglomerates, Koster, E.H., Steel, R.J. (eds.). Canadian Society of Petroleum Geologists Memoir
10: 87–98.
Balasch JC, Ruiz-Bellet JL, Tuset J, Martín de Oliva J. 2010. Reconstruction of the 1874 Santa Tecla’s
rainstorm in Western Catalonia (NE Spain) from flood marks and historical accounts. Natural
Hazards and Earth System Science, 10: 2317-2325. DOI:10.5194/nhess-10-2317-2010.
Ballesteros JA, Eguíbar M, Bodoque JM, Díez-Herrero A, Stoffel M, Gutiérrez-Pérez I. 2011a. Estimating
flash flood discharge in an ungauged mountain catchment with 2D hydraulic models and
dendrogeomorphic paleostage indicators. Hydrol. Process. 25: 970–979.
Ballesteros, J. a., Bodoque, J. M., Díez-Herrero, a., Sanchez-Silva, M., Stoffel, M. 2011b. Calibration of
floodplain roughness and estimation of flood discharge based on tree-ring evidence and hydraulic
modelling. Journal of Hydrology 403: 103-115. doi:10.1016/j.jhydrol.2011.03.045.
Bathurst, J. C., Li, R. M., Simons, D. B. 1979. Hydraulics of mountain rivers, Exp. Stn. Rep. CER 78-79
JCB-RLM-DBS5S, 229 pp., Colo. State Univ., Ft. Collins.
Benito G, Díez-Herrero A, Fernandez de Villalata M. 2003. Magnitude and frequency of flooding in the
Tagus basin (Central Spain) over the last millennium. Climatic Change 58: 171–192.
Bocchiola D, Catalano F, Menduni G, Passoni G. 2002. An analytical–numerical approach to the
hydraulics of floating debris in river channels. Journal of Hydrology 269: 65-78.
DOI:10.1016/S0022-1694(02)00195-6.
Bodoque JM, Eguibar MA, Díez-Herrero A, Gutiérrez-Pérez I, Ruiz-Villanueva V. 2011. Can the discharge
of a past hyperconcentrated flow be estimated based on palaeoflood evidence?. Water Resources
Research. Accepted for publication.
Borga M, Gaume E, Creutin JD, Marchi L. 2008. Surveying flash floods: gauging the ungauged extremes.
Hydrological Processes 22: 3883-3885. DOI:10.1002/hyp.7111.
Buytaert, W., Celleri, R., Willems, P., Bièvre, B. D., Wyseure, G. 2006. Spatial and temporal rainfall
variability in mountainous areas: A case study from the south Ecuadorian Andes. Journal of
Hydrology 329: 413-421. DOI:10.1016/j.jhydrol.2006.02.031.
Camarasa AM, Tilford KA. 2002. Rainfall-runoff modelling of ephemeral streams in the Valencia region
(eastern Spain). Hydrological Processes 16: 3329-3344. DOI:10.1002/hyp.1103.
CEDEX 2011. Mapa de caudales máximos. Memoria tecnica.
http://www.marm.es/es/agua/temas/gestion-de-los-riesgos-de-
inundacion/memoria_tecnica_v2.0_tcm7-162773.pdf.
Comiti F, Andreoli A, Mao L, Lenzi MA. 2007. Wood storage in three mountain streams of the Southern
Andes and its hydro-morphological effects. Earth Surf. Process. Landforms 32: 1675–1692.
DOI:10.1002/esp.
Comiti F, Mao L, Preciso E, Picco L, Marchi L, Borga M. 2008. Large wood and flash floods: evidence from
the 2007 event in the Davča basin (Slovenia). Monitoring, Simulation, Prevention and Remediation
of Dense Debris Flows II 60: 173-182. DOI:10.2495/DEB080181.
Costa JE. 1983. Paleohydraulic reconstruction of flash-flood peaks from boulder deposits in the Colorado
Front Range. Geological Society of America Bulletin 94: 986–1004.
Cunderlik JM, Simonovic SP. 2007. Inverse flood risk modelling under changing climatic conditions.
Hydrological Processes 21: 563-577.
Diehl TH. 1997. Potential Drift Accumulation at Bridges. Publication No. FHWA-RD-97-
028.U.S.Department of Transportation, Federal Highway Administration Research and
Development, Turner-Fair- bank Highway Research Center, Virginia. On the web at: http://
www.tn.water.usgs.gov/pubs/FHWA-RD-97-028/drfront1.htm.
Díez-Herrero A. 2003. Geomorfología e Hidrología fluvial del río Alberche. Modelos y SIG para la gestión
de riberas. Serie Tesis Doctorales 2. Publicaciones del Instituto Geológico y Minero de España
(Ministerio de Ciencia y Tecnología, Madrid. 587 pp.
Enzel Y, Ely LL, House PK, Baker VR, Webb RH. 1993. Palaeoflood evidence for a natural upper bound to
flood magnitudes in the Colorado River basin. Water Resour. Res. 29: 2287–2297.
Fleming M, Neary V. 2004. Continuous hydrologic modeling study with the Hydrologic Modeling System.
Journal of Hydrologic Engineering 93: 175-183.
Gaume, E., Bain, V., Bernardara, P., Newinger, O., Barbuc, M., Bateman, A., Blaškovicová, L., Blöschl, G.,
Borga, M., Dumitrescu, A., Daliakopoulos, I., Garcia, J., Irimescu, A., Kohnova, S., Koutroulis, A.,
Marchi, L., Matreata, S., Medina, V., Preciso, E., Sempere-Torres, D., Stancalie, G., Szolgay, J.,
Tsanis, I., Velasco, D., Viglione, A., 2009. A compilation of data on European flash floods. J. Hydrol.
367: 70–78.
Gaume, E., Borga, M. 2008. Post-flood field investigations in upland catchments after major flash floods:
proposal of a methodology and illustrations. Journal of Flood Risk Management 1: 175-189.
DOI:10.1111/j.1753-318X.2008.00023.x.
Goovaerts P. 2000. Geostatistical approaches for incorporating elevation into the spatial interpolation of
rainfall. Journal of Hydrology 228: 113-129.
Grant, G. E. 1997. Critical flow constrains flow hydraulics in mobile-bed streams : A new hypothesis.
Water Resources 33: 349-358.
Gregory, K.J. 1983. Background to Paleohydrology- a perspective: New York, John Woley, 486 p.
Grimes D, Pardo-Igúzquiza E. 2010. Geostatistical analysis of rainfall. Geographical Analysis 42: 136-160.
Gül GO, Harmancıoğlu N, Gül A. 2009. A combined hydrologic and hydraulic modeling approach for
testing efficiency of structural flood control measures. Natural Hazards 54: 245-260.
DOI:10.1007/s11069-009-9464-2.
Herschy, R.W., 2002. The world’s maximum observed floods. Flow Meas. Instrum. 13: 231–235.
Horritt M, Bates P. 2002. Evaluation of 1D and 2D numerical models for predicting river flood
inundation. Journal of Hydrology 268: 87-99. DOI:10.1016/S0022-1694(02)00121-X.
International Association of Hydrological Sciences (IAHS), 2003. World Catalogue of Maximum
Observed Floods. IAHS Press.
Jacoby Y, Grodek T, Enzel Y, Porat N, McDonald EV, Dahan O. 2008. Late Holocene upper bounds of
flood magnitudes and twentieth century large floods in the ungauged, hyperarid alluvial Nahal
Arava, Israel. Geomorphology 95: 274-294. DOI:10.1016/j.geomorph.2007.06.008.
Jarrett, R. D. 1984.Hydraulics of high-gradient streams, J. Hydraul. Eng. 110: 1519–1539.
Johnson RM, Warburton J. 2002. Flooding and geomorphic impacts in a mountain torrent: Raise Beck,
central Lake District, England. Earth Surface Processes and Landforms 27: 945-969.
DOI:10.1002/esp.386.
Johnson SL, Swanson FJ, Grant GE, Wondzell SM. 2000. Riparian forest disturbances by a mountain
flood ? the influence of floated wood. Hydrological Processes 14: 3031-3050. DOI:10.1002/1099-
1085(200011/12)14:16/17<3031:AID-HYP133>3.0.CO;2-6.
Kehew AE, Milewski A, Soliman F. 2010. Reconstructing an extreme flood from boulder transport and
rainfall–runoff modelling: Wadi Isla, South Sinai, Egypt. Global and Planetary Change 70: 64–75.
Koutroulis AG, Tsanis IK. 2010. A method for estimating flash flood peak discharge in a poorly gauged
basin: Case study for the 13–14 January 1994 flood, Giofiros basin, Crete, Greece. Journal of
Hydrology 385: 150-164.
Lumbroso, D., Gaume, E. 2012. Reducing the uncertainty in indirect estimates of extreme flash flood
discharges. Journal of Hydrology: 414-415. DOI:10.1016/j.jhydrol.2011.08.048.
Lynn, B.H., R. Healy, L.M. Druyan, 2007: An analysis of the potential for extreme temperature change
based on observations and model simulations. J. Climate 20, 1539-1554, doi:10.1175/JCLI4219.1.
Maizels, J.K. 1983. Palaeovelocity and palaeodischarge determination for coarse gravel deposits. In Gregory,
K.J., Ed. Background to paleohydrology. Chichester, Wiley, 101–139.
Mao L, Comiti F, Lenzi MA. 2010. Bedload Dynamics in Steep Mountain Rivers : Insights from the Rio
Cordon Experimental Station ( Italian Alps ). Main, 253-265.
Mao, L., Comiti, F., 2010. The effects of large wood elements during an extreme flood in a small tropical
basin of Costa Rica. In: De Wrachien, D., Brebbia, C.A. (Eds.), Debris Flow III. WIT Press, UK, pp.
225–236.
Marchi, L., Borga, M., Preciso, E., Gaume, E. 2010. Characterization of selected extreme flash floods in
Europe and implications for flood risk management. Journal of Hydrology 394: 118-133.
DOI:10.1016/j.jhydrol.2010.07.017.
Marchi, L., Borga, M., Preciso, E., Sangati, M., Gaume, E., Bain, V., Delrieu, G. Bonnifait, L., Pogacnik, N.
2009. Comprehensive post-event survey of a flash flood in Western Slovenia : observation strategy
and lessons learned. Hydrological Processes 23: 3761-3770. DOI:10.1002/hyp.
Maskey S, Guinot V, Price RK. 2004. Treatment of precipitation uncertainty in rainfall-runoff modelling: a
fuzzy set approach. Adv. Water Res. 27: 889–898.
Mazzorana B, Comiti F, Volcan C, Scherer C. 2011. Determining flood hazard patterns through a
combined stochastic–deterministic approach. Natural Hazards. DOI:10.1007/s11069-011-9755-2.
Mazzorana B, Hübl J, Zischg A, Largiader A. 2010. Modelling woody material transport and deposition in
alpine rivers. Natural Hazards 425-449. DOI:10.1007/s11069-009-9492-y.
Mazzorana, B., & Fuchs, S. (2010a). Fuzzy Formative Scenario Analysis for woody material transport
related risks in mountain torrents. Environmental Modelling & Software 25: 1208-1224.
DOI:10.1016/j.envsoft.2010.03.030.
Mazzorana, B., Comiti, F., Fuchs, S. 2011a. A structured approach to enhance flood hazard assessment in
mountain streams. Natural Hazards. DOI:10.1007/s11069-011-9811-y.
Mazzorana, B., Comiti, F., Volcan, C.,Scherer, C. 2011b. Determining flood hazard patterns through a
combined stochastic–deterministic approach. Natural Hazards 59: 301-316. DOI:10.1007/s11069-
011-9755-2.
Mazzorana, B., Hübl, J., Zischg, A., Largiader, A. 2010b. Modelling woody material transport and
deposition in alpine rivers. Natural Hazards 425-449. DOI:10.1007/s11069-009-9492-y.
Mazzorana, B., Zischg, A., Largiader, A., Hübl, J. 2009. Hazard index maps for woody material recruitment
and transport in alpine catchments. Natural Hazards and Earth System Science 9: 197-209.
DOI:10.5194/nhess-9-197-2009.
Montesarchio V, Lombardo F, Napolitano F. 2009. Rainfall thresholds and flood warning: an operative
case study. Natural Hazards and Earth System Science 9: 135-144. DOI:10.5194/nhess-9-135-2009.
O’Connor JE, Webb RH. 1988. Hydraulic modelling for paleoflood analysis. In Flood Geomorphology,
Baker VC, Kochel RC, Patton PC (eds). John Wiley & Sons, Inc.: New York; 393–402.
Ortega JA, Garzón-Heydt G. 2009. Geomorphological and sedimentological analysis of flash-flood
deposits. The case of the 1997 Rivillas flood (Spain). Geomorphology 112: 1-14.
doi:10.1016/j.geomorph.2009.05.004.
Palacios D, Garcia R, Rubio V, Vigil R. 2003. Debris flows in a weathered granitic massif: Sierra de Gredos,
Spain. Catena 51: 115–140.
Palacios D, Marcos J, Vázquez-Selem L. 2011. Last glacial maximum and deglaciation of Sierra de Gredos,
Central Iberian Peninsula. Quaternary Internacional 233: 16–26.
Pardo-Igúzquiza E. 1998. Comparison of geostatistical methods for estimating the areal average
climatological rainfall mean using information of precipitation and topography. International
Journal of Climatology 18:1031-1047.
Remy N, Boucher A, Wu J. 2009. Applied Geostatistics with SGeMS: A User's Guide. Cambridge
University Press. 264pp.
Ruiz-Villanueva V, Bodoque JM, Díez-Herrero A, Calvo C. 2011. Triggering threshold precipitation and
soil hydrological characteristics of shallow landslides in granitic landscapes. Geomorphology 133:
178-189.
Ruiz-Villanueva, V., Bladé-Castellet, E., Sánchez-Juny, M., Martí, B., Bodoque , J.M., Díez Herrero, A. 2D
modelling of wood transport in rivers. Under review. Journal of Hydrology.
Ruiz-Villanueva, V., Borga, M., Zocatteli, D., Marchi, L., Gaume, E., Ehret, U., Zehe , E. Extreme flood
response to short-duration convective rainfall in South-West Germany. Accepted for publication in
HESSD.
(4.4). 4. Results
4.1. Wood contributing areas due to
landslides, bank erosion and fluvial
transport during floods.
4.2. Potentially recruitable wood volume:
theoretical predictions
(4.4). 5. Discussion
5.1. Large wood recruitment: sources and
volumes
5.2. Potential errors and limitations
5.3. Potential applications
(4.4). 6. Conclusions
REFERENCE:
Ruiz-Villanueva, V., Díez-Herrero, A., Ballesteros, J.A., Bodoque, J.M. (in press). Large
Woody Debris potentially recruitable due to landslides, bank erosion and floods in
mountain basins: a quantitative estimation approach. River Research and
Applications . DOI: 10.1002/rra.2614.
ABSTRACT
An in-depth knowledge of the fluvial corridor and surrounding slopes and forest
vegetation is needed for a better understanding of wood recruitment or inputs to rivers.
The information available in Central Spain on hydrogeomorphic processes and forest
distribution enabled the evaluation of potential wood recruitment from three sources:
landslides, bank erosion and fluvial transport during floods on a regional scale. The
method presented here is based on a geographical information system (GIS) and on multi-
criteria and multi-objective assessment using fuzzy logic principles. First, the areas
potentially affected by landslides, bank erosion and floods were delineated, and a
vegetation analysis was carried out to obtain the vegetation resistance and forest density.
Several scenarios were proposed based on the process frequency and severity. Using this
method the volume of potentially available wood can be estimated for each scenario.
14 river basins in populated areas were selected for further analyses and field survey.
Observations of in-stream storage of woody debris and tree disturbances were used to
interpret the woody debris dynamics throughout the watershed and validate the obtained
results.
1. Introduction
Vegetation affected but not killed by hydrogeomorphic processes will react to these
disturbances (Stoffel and Wilford, 2011), and may also provide information on the river
dynamics (i.e. dendrogeomorphology) and LWD.
Within the stream bed and floodplain, erosion processes occur along the banks
altering the static equilibrium of trees, so these may topple or slide into the channel. This
process is related to the stream flow power and its geomorphological configuration.
Woody material is also available and mobile within the channel bed and on forested bars
and may be transported and deposited by fluvial mechanisms.
The presence of wood in streams has been shown to be linked to positive ecological
effects, in contrast to past perceptions that streams need to be freed from obstructions (see
references in Kasprak et al., 2011). A better understanding of LWD entrainment, or the
process by which woody material is transported to the river, is therefore needed when
considering the effects of LWD in rivers from an ecological perspective, for analysing
geomorphological processes and for flood hazard assessment. Therefore, the rate of LWD
delivered to streams has been the subject of several studies in recent years. Martin and
Benda (2001) constructed a LWD budget using a proposed quantitative framework to
evaluate spatial and temporal controls on LWD recruitment rate transport. Some years
later, Benda and Sias (2003) evaluated the mass balance of in-stream organic debris making
quantitative estimates of wood flux. May and Gresswell (2003) identified wood
Therefore, the aims of this paper are: (i) to define areas that may contribute woody
material to streams, observing the importance of different recruitment processes and
creating reliable scenarios based on the process severity; (ii) to provide estimates of the
order of magnitude of recruitable wood volumes for each scenario; and (iii) to study the
LWD dynamics in those catchments identified as potentially risk-prone.
Because the focus here was on the role of hydrogeomorphic processes in supplying
woody material to the channel, a spatially inclusive approach is adopted regarding the
volume of potentially recruitable wood from landslides, bank erosion and fluvial transport
during floods (Fig.1). Other factors such as wind, snow, fires or decay which may also
deliver woody material are outside the scope of this study.
Figure 1: The three main wood recruitment processes on a catchment scale studied in this
analysis: (Ls) landslides; (Ft) fluvial transport during floods; (Be) stream bank erosion.
2. M
Method
dology
The proposed method d is dividedd into three main steps (Fig. 2): (i)) terrain an
nalysis to
establish wood sources
s suchh as areas affected byy landslides, floods annd bank ero osion, in
additiion to takin
ng into accoount the diffferent scennarios (based on the prrocess severrity); (ii)
vegetaation analyysis to obtaain the treee resistancee and foresst density (iii) wood volume
estimaates for each
h scenario. Finally
F the results are evaluated
e with
w a field ssurvey
Figgure 2: Gene
neral methoddology flow
wchart
VIRGINIA
A RUIZ-VILLANUEEVA 219 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
The importance of wood recruitment due to mass wasting depends on the type and
size or area of the landslide, the age or size of trees recruited and the connectivity to the
channel (source areas intersecting a channel segment of a given length; Benda and Sias,
2003). Existing and published natural hazard maps for the province of Ávila (Díez-
Herrero and Ballesteros, 2009; Ballesteros et al., 2012) were used to establish potential
landslide areas, as they include landslide susceptibility information. A buffer area of
influence was also established around these areas, to include toppled trees that may be
recruited indirectly by the action of landslides.
A number of studies have shown the importance of tree proximity to the channel on
overall recruitment (McDade et al., 1990; Robison and Beschta, 1990; Van Sickle and
Gregory, 1990; Bragg and Kershner, 2004). Therefore, the connectivity between those
areas and the stream was analysed as this is decisive in the final wood entrainment. Van
Sickle and Gregory (1990) quantified the probability of a tree entering the stream as a
function of the distance from the nearest channel boundary and the tree height (Ht);
Mazzorana et al. (2009) also took the slope gradient into account. Connectivity was
established as a function of both the distance to the channel and the slope. If a tree is
located in a landslide-prone area or in the toppling influence area then it will reach the
channel if it is at a close enough distance (D; D= k·Ht; k is the toppling coefficient and the
value assigned here was 2 (tree height x 2), or if it is further away but on a steep slope
(>40%).
VIRGINIA
A RUIZ-VILLANUEEVA 221 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
Bank erosion often occurs during floods, and it recruits trees at rates depending on
the erodibility of banks, flood frequency and stand density (Benda and Sias, 2003).
Following Nakamura and Swanson (1994) and Kasprak et al. (2011) the hypothesis
put forward here was that the measurement of channel sinuosity quantifies how prone the
stream is to lateral migration. Transport capacity is also used here as an indicator of the
potential stream capacity for bank erosion.
Empirical equations have been proposed for steep mountain streams to estimate
orders of magnitude for sediment concentration (Smart and Jaeggi (1983); Mizuyama and
Shimohigashi (1985); Bathurst et al. (1987); Meunier (1989); Rickenmann (1990 and
1991)). This concentration is the ratio between solid and liquid discharges and it can be
used as an indicator of stream transport capacity. In general these are regression equations
based on field and laboratory data, and related to stream gradient.
According to our hypothesis, stream reaches with a high sinuosity (high lateral
migration of the channel) and high transport capacity (high sediment concentration) will
be prone to bank erosion. Therefore, the reasoning is as follows: trees located on the
riverbed or flood plains (defined in the previous subsection) will be easily entrained during
floods, and bank erosion will significantly increase the entrainment probability (Fig. 4).
Figur
ure 4: Probab
ability in perrcentage off w
wood recruuited withinn the fluviall corridor based
ba on
floood frequencycy, flood sevverity and sttream bankk erosion cappacity.
VIRGINIA
A RUIZ-VILLANUEEVA 223 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
The Forest Map of Spain and National Forest Inventory produced by the Ministry of
the Environment, Spain (MFE, 2011) include fifteen different descriptor fields for the
ecology and structure of the forest mass. Up to three different species of forest trees are
identified, each with their development stage, percentage occupation (percentage of the
total forest occupied by this species), canopy cover for total trees (percentage of ground
covered by the horizontal projection of the canopy), and DBH (diameter at breast height),
among others.
In general terms, the vegetation resistance increases when the area has been reforested
or increasingly managed (Hutte, 1968; Stumbles, 1968). Given that planted riparian
vegetation can take decades to mature fully, it is assumed that the natural rate of LWD
recruitment would be significantly reduced in rehabilitated riparian corridors in
comparison to remnant corridors (Webb and Erskine, 2003). The type of tree (conifer,
deciduous or riparian) is also relevant to this resistance, since an abundant below-ground
root biomass anchors the tree to the substrata. In riparian vegetation the root system depth
depends on the water table, whereas in conifers it may be deeper, and therefore the
resistance to toppling may be greater (Naka, 1982; Abernethy and Rutherfurd, 2001). This
means that the recruitment potential will increase with trees in their natural state, and
even more so if they are riparian species.
Once the source areas are defined and classified into different categories, the number
of trees from each area is estimated and a statistical analysis is performed.
The National Forest Map and Inventory provide the required data, however some
simplification has to be assumed. Data is provided for the three main species in any given
area and the total canopy cover (Ci) as a percentage of the total area covered by forest. The
inventory for the province of Ávila contains estimates of tree density (expressed as number
of trees per area) for each species for the whole forested territory. This latter density is
called here relative density per species (RDSPi) and is used together with the species
occupation and canopy cover to estimate final volumes in a given area. Ai is the
contributing area defined for a specific recruitment process and with an established
occurrence probability or severity:
PRWSPi=Ai·Ci·RDSPi [1]
To take into account the defined vegetation resistance and the severity of the
potential recruitment mechanism, a volume correction factor (Fc) was defined which is
equivalent to a recruitment probability. As shown in Fig. 3 and 4, this can be 1, 0.5 or 0.1.
This factor therefore reduces the total volume of potential recruitable wood in those areas
where susceptibility to the process is lowest and/or vegetation resistance is highest.
PRWt=PRWls+PRWbe+PRWft [3]
where PRWsl is the potentially recruitable wood volume due to landslides; PRWbe the
potentially recruitable wood volume due to bank erosion; PRWft the potentially
recruitable wood volume due to fluvial transport during floods.
Different scenarios were created based on process frequency and severity. Then, the
potential source areas were defined and classified using fuzzy associative matrices, and the
potentially available recruitable wood volumes were also estimated based on the forest data
for each scenario:
The classification of potential wood source areas was based on fuzzy logic principles
(Zadeh, 1965 and 1968; Zimmermann, 2001) which usually use IF-THEN rules such as
fuzzy associative matrices (ANNEX 1). In fuzzy logic applications non-numeric linguistic
variables are often used to facilitate the expression of rules and facts. A linguistic variable
may have a value or its antonym, or something in between, and they are useful because they
can be modified using linguistic hedges applied to primary terms.
This procedure allows the areas to be classified by the likelihood that they will recruit
wood material based on the severity of the potential recruitment process and the
vegetation resistance and abundance. Fuzzy associative matrices take into account three
categories or impact levels (high, medium and low) for all factors (process severity or
susceptibility, vegetation resistance, correction factor, etc.).
In those basins where a river flows through a populated area scenarios were
established, and the potentially recruitable wood volume was calculated. Statistical analyses
(using the R programming language, www.r-project.com) were carried out to interpret the
results, and this procedure allowed us to compare the data of the different selected basins.
2.4. Field
d observvations
Solid mateerial (sedim ment and w wood) transsported in the stream flow may damage
trees, and this is i the basiss of dendrrogeomorph hology, whiich uses trree-ring anaalysis to
interppret geomorphic proceesses (Stofffel and Bolllschweiler, 2008; Ruiiz-Villanuevva et al.,
2010; Ballesteros et al., 2011). Evidennce from surviving
s trrees may bee used to interpret
LWD D dynamics,, which can n be particuularly usefuul when wo oody debris has been removed
r
from the river or o transporrted longer distances. Díez-Herrrero et al. ((2012) pro oposed a
unifieed classificattion for treee disturbannces caused by
b floods, and
a here theese disturbaances are
relatedd to wood recruitment
r t, delivery aand transport. For instaance, candeelabra growtth (FDE
4), decapitated trrees (FDE 6), 6 scars (FD DE 7 and 9),
9 branchess torn off (FFDE 8), naarrowing
in truunks (FDE 12) and bifurcations
b s (FDE 13)) may be related
r to L
LWD delivvery and
transpport, whereas vegetatio on patterns (FDE 1, 2,, 3), tilted trees (FDE 55) or expossed roots
(FDE E 17 and 18)) are eviden nce of LWD D recruitmen nt and poteential deliveery.
VIRGINIA
A RUIZ-VILLANUEEVA 227 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
and external disturbances in living trees. These observations were compared with the
results obtained in the wood volume estimations.
3. Test site
The study site is in the central and eastern massif of the Sierra de Gredos, the highest
section (Almanzor Peak, 2592 m a.s.l.) of the Spanish Central System, which crosses the
Iberian Peninsula from SW to NE. The sector of Sierra de Gredos studied here is in the
south of the province of Avila, where the information required for the analysis was
available; the total area of the study site is 4658 km2.
The Sierra de Gredos is mainly composed of granite tectonic blocks sloping towards
the north and separated by a network of fault lines, running N-S and E-W. The bedrock
consists of Upper Palaeozoic granitoids. Regolith covers most of the slopes in Sierra de
Gredos, reaching a depth of 3-4 m in some places. The weathering mantle is not firmly
anchored to the slopes, so heavy rains weaken it and initiate shallow landslides and debris
flows (Palacios et al., 2003; Ruiz-Villanueva et al., 2011).
The area studied here is surrounded by the main tributaries of these two important
rivers: the Tiétar to the S, the Adaja to the NE, the Tormes to the NW, the Alberche to
the E and the Jerte to the W (Fig. 6).
Several factors also determine the occurrence of frequent flash floods. These include
the high drainage density, the morphology of the headwaters and the generally small basin
areas with very steep average slope, resulting in a very short response times.
There is abundant forested mass in the area (2358 km2 or 50.6% of the study site),
and different species are found depending on altitude (Fig.7). At higher altitudes, where
the extreme climatic characteristics impede the development of tree or shrubs the main
grass species are Festuca sp. and Carex sp. At lower levels shrubs (Cytisus sp.) and stretches
of high mountain conifers such as Pinus sylvestris are present. At lower altitudes this latter
species is replaced by Pinus pinaster which usually appears with Genista florida, Ilex
aquolifolium, Erica arborea and Sorbus aucuparia. Lower down, the forest is composed
mainly of deciduous trees such as Quercus pyrenaica and Quercus ilex. The vegetation
found on the river banks is predominantly Alnus glutinosa and Fraxinus angustifolia.
Figure 7: Land use and forested canopy cover of the study region.
Basins with a river flowing through a populated area (assuming potential flood risks)
were selected for further analysis and field surveys (Table 1) and are highlighted in Fig 7.
4. Results
The obtained results are grouped in three sections: (i) wood contributing areas; (ii)
potentially recruitable wood volumes; and (iii) field observations.
The landslide-prone areas according to the stability model results make up around 10
% (461 km2) of the study site. Close to 55% of these areas (254 km2) fall into the medium
susceptibility category, and just 30% (138 km2) and 15% (71 km2) are considered to have
high and low landslide susceptibility respectively.
A 50 m buffer zone was set around these landslide-prone areas as the toppling
influence area. This increases the potential area that could be affected by landslides by
233.11 km2 (695 km2 in total). 75% of this area is very or fairly well connected to the
channel which means that trees in these areas may reach the channel due to landslide
recruitment. However, only 158 km2 (32%) of this area is forested according to the
National Forest database.
From the flood hazard assessment of the whole area, the flood severity for around
10% of the watersheds is classified as very high, 32% as high, 43% as medium and 15% as
low.
The Quaternary units delineated in relation to flood processes occupy a total area of
610 km2. Of these, 50.5% (308 km2) are stream bed or channels, 10.5% (65 km2)
floodplains and bars, 3% (17 km2) terraces, and 36% (221 km2) other formations (fans,
etc.).
The likelihood of potential recruitment of wood material in these areas due to floods
was classified as: high, 47 % (288 km2), medium, 15% (88 km2), and low, 38% (232 km2).
Only 20% (123 km2) of the defined areas are forested. The analysis of these parameters and
variables of each selected basin is shown in Table 2.
Table 2: Defined forested areas, areas prone to landslides, floods and bank erosion for
selected basins. Percentages indicate the ratios between the total forested area and the
delineated area defined in the previous column.
The forested area was classified based on the vegetation resistance following the
methodology described above (see Fig. 8). The vegetation resistance to recruitment of the
total forested area (2358 km2) is classified as: low, 40%, medium, 47% , high, 13%.
After analysing the contributing areas defined for the different recruitment
processes, the vegetation resistance and the forest density (3 main species, canopy cover
and relative density) theoretical predictions were obtained of the number of trees
potentially recruitable in each scenario (Table 3 and Fig. 9). Since there may be some
uncertainties in the sources used for the analysis, these numbers are taken here as orders of
magnitude, and used to identify the main recruitment processes and compare basins.
VIRGINIA RUIZ-VILLANUEVA
234
CAPÍTULO /CHAPTER 4
Figure 9: Number of potentially recruitable trees for the selected 14 basins for (A) Scenario
1, (B) Scenario 2, (C) Scenario 3.
Some observations can be made from the results. In Scenarios 1 and 2 the major
recruitment process is slope movements in terms of the number of recruitable trees, except
for the basin called Navaluenga 1. There are no major differences between Scenarios 1 and
2 regarding the importance of landslide and fluvial transport during floods, whereas bank
erosion seems to be less important in Scenario 2 (except for Navaluenga 1 where it is the
main recruitment process together with floods in both scenarios).
Based on the interpretation of Fig. 9, some basins seem to show different behaviour
in all scenarios. To analyse this observation in detail a cluster analysis (hierarchical
clustering using Ward's method) was carried out based on the number of recruitable trees,
basin area, forested area, landslide-prone area, flood-prone area and bank erosion-prone
area for each scenario (Fig. 10).
For Scenarios 1 and 2 three different groups are defined, whereas for Scenario 3, two
groups were distinguished. This reveals different behaviour between several groups, but
highlights similarities between some basins: Arenas, Candeleda, Guisando, La Adrada,
Navaluenga 1, Navaluenga 2 and Venero.
(C)
Figure 10: Dendograms of hierarchical clustering using Ward's method. Dendograms for
(A) Scenario 1. (B) Scenario 2. (C) Scenario 3.
Figure 11: (A, B, C) Graphs of a correspondence analysis for each scenario. Basin: basin
area; Veg_Area: forested area; Lnd: landslide-prone area: H_Lnd: high severity landslide-
prone area; Ft: flood- and bank erosion-prone area; H_Ft: high severity flood-prone area;
H_Be: high severity bank erosion-prone area.
The field work allowed us to interpret the results, and certain indicators were used to
corroborate the hypotheses proposed (Table 4), although in some cases they may not be
perfectly correlated.
Table 4: Summary of field observations in the selected basins: YES (with shading)=
observed: NO = not observed. In bold : Group 1 and 2 for Scenarios 1 and 2, and group 2
for Scenario 3. *See comments.
Table 4 shows that two or more field indicators were found in some basins. These
basins (shown in bold) also showed the highest wood volume estimates and were grouped
together in the cluster analysis: Arenas, Candeleda, Guisando, La Adrada, Navaluenga 1,
Navaluenga 2 and Venero.
5. Discussion
The following section discusses the proposed methodology and the results obtained,
analyses some errors and limitations and highlights some potential applications.
The aim of this paper was to propose a method to define recruitment areas due to
landslides, floods and bank erosion, and to estimate the volume of potentially recruitable
wood depending on process severity.
Other recruitment processes such as wind, snow, fire or chronic wood decay are
outside the scope of this study. Wood decay could be included as a percentage of the total
wood volume (if data are available), since tree mortality and fall rates can generate some
variation in wood recruitment (Benda et al., 2003).
The whole analysis was based on existing hazard maps, available morphometric
information derived from DEMs, geological and geomorphological spatial information and
forestry inventory and maps. A GIS was used to obtain a spatially distributed analysis of
potential LWD source areas and estimated wood volumes. Multi-criteria and multi-
objective evaluation and fuzzy logic principles were used to define reliable scenarios,
classifying areas by the likelihood of wood material recruitment based on potential
recruitment processes, vegetation resistance and abundance. Fuzzy associative matrices
allowed all the available information to be used reliably based on the three categories or
impact levels defined. The fuzzy algebraic operations assign consistency to all
combinations, and the end result of the procedure is a series of scenarios (similar to the
formative scenarios defined by Scholz and Tietje, 2002), which include all the levels
mentioned above.
The potentially recruitable number of trees was successfully estimated for each
scenario using this method. ‘Potentially recruitable’ has been defined as the number of
trees that may contribute wood to the channel throughout the basin. This concept has
been used before (Kasprak et al., 2011). However, these authors evaluated the number of
vegetation pixels (based on LIDAR DEM) that were tall enough to span the channel,
rather than individual trees.
The amount of wood in streams (wood budget) was estimated by Martin and Benda
(2001), Benda and Sias (2003) and May and Gresswell (2003). Based on in-channel wood
volume, they developed simplified mathematical expressions to estimate LWD flux; here
the number of standing trees located in the recruitment source areas that could reach the
streams was computed. Mazzorana et al. (2009) defined an expression to estimate the
relative availability of recruitable woody material. However, this is a synthetic indicator for
the relative propensity of a basin for entrainment and delivery of woody material under
given transport conditions. In a later work, Mazzorana et al. (2011) established using pixels
the maximum amount (m3) of transported woody material during a flood event; here
different scenarios based on severity levels of recruitment processes (not a given event)
have been taken into account.
The three main scenarios proposed are based on previous knowledge of the area and
the processes involved. Thus, Scenario 1 was defined as unlikely to occur because the
probability of a landslide occurring in all the delineated areas is very low (defined by
different susceptibility levels). However, this is a high risk flood-prone region (Díez-
Herrero, 2003), and therefore Scenario 3 is the most realistic. Analysing the three cases also
allowed us to compare the potential behaviour of all the basins (recruitment capacity) and
highlight those areas where recruitment processes may occur.
The results of the analyses show clearly the different behaviour of some basins in
terms of the number of potentially recruitable trees, basin area, forested area and
recruitment source areas. These basins (highlighted in bold in Table 4) may have a higher
wood recruitment capacity based on the results of the analyses and field observations.
Although the sources and mechanism of wood recruitment may vary across regions, the
method presented here clarifies and explains these differences on a regional scale, enabling
areas for detailed studies or field surveys to be identified. The field work highlighted the
need to establish indicators to interpret the LWD dynamics at reach scale, especially when
no direct evidence is found and woody debris has been removed from the river or
We acknowledge some limitations of the conceptual model, and probably one of the
most striking findings is the validation of volume estimations. Although exact numbers are
obtained, this analysis may be accompanied by some uncertainties and the results must
therefore be considered as orders of magnitude. One source of uncertainty is the input data
used (the level of detail of the procedure is restricted to the detail defined (scale) of hazard,
geological, topographical and forestry maps). Possible inaccuracies in earlier estimated
landslide and flood susceptibility are outside the scope of this work. Besides the previously
published landslide and flood maps, we defined stream bank erosion-prone areas. The
factors controlling stream bank erosion are complex and interrelated and may require
detailed field data. However, the aim here was to analyse this process as a potential wood
recruiting mechanism; the stream bank erosion susceptibility was inferred from two other
indicators, the stream sinuosity and transport capacity, and the erosion-prone areas were
defined following the method used for flooding.
The use of different scales and various information sources may also alter the results
to some extent. It is not possible to test the theoretical predictions made in this paper fully,
but to estimate the uncertainty related to the delineation of source areas a sensitivity test of
this parameter over the final recruitable tree results was carried out. A 10% interval
reduction in the recruitment area was computed and recruitable trees were recalculated.
Fig. 12 shows the case for the landslide-prone area. The total wood volume showed a linear
decrease ranging from 10% to 35%. Therefore, up to 25% error is associated with the
source area delineation.
Figure 12: Model sensitivity analysis. Graph shows landslide-prone area reduction (%) and
recruitable trees reduction (%).
In the delineated source areas the probability of a tree entering the stream was
quantified based on the paper by Robison and Beschta (1990). These probabilities were
incorporated into the method using the volume correction factor. This factor therefore
reduces the total volume of potential recruitable wood in those areas where the process
susceptibility is lowest and/or vegetation resistance is highest. The vegetation resistance
was established based on the tree species and forest stage, and on previous work by Hutte
(1968), Stumbles (1968), Naka (1982), Abernethy and Rutherford (2001) and Webb and
Erskine (2003). This concept may be equivalent to the structural classification of forested
areas made by Blaschke et al. (2004) and the woody debris availability indicator used by
Mazzorana et al. (2011). The volume correction factor can also be subjective, and the
modification of this coefficient may change the final results. However, in the authors’
opinion, 100% recruitment is not reliable; therefore, the maximum probability value (%)
was used to obtain maximum potential recruitable volumes.
Analysis of the results focused on the basins crossed by rivers provided an overview of
those areas potentially at risk from large woody debris transport during floods. According
to our findings, 7 of the 14 basins studied showed enough evidence in the field and in the
theoretical calculations to affirm that LWD transport takes place and may be a potential
hazard during flood events. Many studies show that woody material transported in the
flow is responsible for increasing flood hazard impact (Comiti et al, 2008; Mao and
Comiti, 2010; Rickenmann and Koschni, 2010). Hence, woody material transport cannot
be ignored in reliable hazard assessment. The authors of this paper also assume that the
WD contribution is due to an episodic process, a landslide or flood (Wohl et al., 2011), so
that transport will be congested (Braudrick et al., 1997). This type of transport may result
in clogging or obstructions at critical stream configurations such as bridges. Some studies
have included this phenomenon in flood hazard analysis (Merten et al., 2010; Mazzorana
et al., 2010, 2011; Ruiz-Villanueva et al., 2012). In all these studies estimates of wood
volumes are required as input.
The method presented here may be useful as a first step in the identification of those
basins where there may be significant amounts of LWD and in the preliminary definition
of wood loads for physically-based LWD modelling (Ruiz-Villanueva et al., under review).
The proposed methodology can also be used for river and forest restoration and
management (Hilderbrand et al., 1998). Knowing the spatial patterns of LWD
recruitment can provide a watershed context for understanding geomorphic and ecological
processes associated with LWD (Martin and Benda, 2001). This may help land managers
to identify the relative importance of different recruitment processes, knowing where and
how much LWD is recruited. Forecasts of future conditions could also be simulated using
different forestry cover. This allows estimates of changing conditions of source areas and
wood volumes in a changing perspective of land use or stand dynamics (Swanson et al.,
1998). Scenarios for climate change could also be incorporated. The same is true for
predictions based on changing recruitment processes (i.e. types, frequency and severity), so
that the recruitment capacity can be analysed at a basin scale.
6. Conclusions
This paper presents a method to define areas that may contribute to the delivery of
woody material to streams on a regional scale. The method takes into account the
importance of different recruitment processes (landslides, fluvial transport and bank
erosion during floods), creating reliable scenarios based on process severity and also
provides estimates of recruitable wood volumes for each scenario.
The method provides a suitable approach to set realistic basin-wide targets defining
the LWD recruitment capacity and estimate the order of magnitude of the maximum
wood volume contributed. In the study area, 7 of the 14 basins showed enough evidence in
the field and in theoretical calculations to conclude that LWD transport takes place and it
could be a potential risk with the highest potential recruitable wood volume (expressed as
thousands of trees). The results can be used to characterize the dynamics of woody debris
from the perspective of potential risk from its transport during floods. To improve this
hazard analysis it is crucial to predict the relative propensity of streams for the entrainment
and delivery of woody material.
Acknowledgements
This work was funded by MAS Dendro-Avenidas project and the Geological Survey of Spain
(IGME). The first author is most grateful to Ángel Prieto and Margarita Sanabria (IGME) for their
suggestions in the GIS analysis, and the Spanish Ministry of the Environment for providing the forest map
and inventory data.
REFERENCES
Abernethy, B., Rutherfurd, I.D. 2001. The distribution and strength of riparian tree roots in relation to
riverbank reinforcement. Hydrological Processes. 15: 63-79.
Ballesteros-Cánovas, J.A., Díez-Herrero, A. & Bodoque, J.M. 2012. Searching for useful non-systematic
tree-ring data sources for flood hazard analysis using GIS tools. Catena 92: 130–138.
Ballesteros Cánovas, J. A., Eguibar, M., Bodoque, J. M., Díez-Herrero, A., Stoffel, M., Gutiérrez-Pérez, I.
2010. Estimating flash flood discharge in an ungauged mountain catchment with 2D hydraulic
models and dendrogeomorphic palaeostage indicators. Hydrological Processes.
doi:10.1002/hyp.7888
Ballesteros, J. A., Bodoque, J. M., Díez-Herrero, A., Sanchez-Silva, M., Stoffel, M. 2011. Calibration of
floodplain roughness and estimation of flood discharge based on tree-ring evidence and hydraulic
modelling. Journal of Hydrology. Elsevier B.V. doi:10.1016/j.jhydrol.2011.03.045Ballesteros-
Cánovas, J. A., Díez-Herrero, A., Bodoque, J. M. Searching for useful non-systematic tree-ring data
sources for flood hazard analysis using GIS tools. Catena 92: 130-138.
Bathurst, J.C., Graf, W.H., Cao, H.H. 1987. Bed load discharge equations for steep mountain rivers”, en
Thorne, C.R.; Bathurst, J.C. y Hey, R.D. (Ed.): Sediment transport in gravel-bed rivers. Wiley, pp.
453-491.
BCN25, 2011. Base Cartográfica Numérica digital 1:25000. Instituto Geográfico Nacional,
http://www.ign.es/ign/layoutIn/actividadesCartografia.do
Benda, L. E., Sias, J. C. 2003. A quantitative framework for evaluating the mass balance of in-stream organic
debris. Forest ecology and management 172: 1-16.
Bodoque, J. M., Eguibar, M. a., Díez-Herrero, A., Gutiérrez-Pérez, I., Ruíz-Villanueva, V. 2011. Can the
discharge of a hyperconcentrated flow be estimated from paleoflood evidence? Water Resources
Research 47: 1-14. 10.1029/2011WR010380
Bragg, D. C., Service, U. F., Uam, P. O. B., Kershner, J. L., Unit, F. E. 2004. Sensitivity of a Riparian Large
Woody Debris Recruitment Model to the Number of Contributing Banks and Tree Fall Pattern.
USDA Forest Service Gen. Tech. Rep. PSW-GTR-181 19: 117-122.
Braudrick, C. A., Grant, G. E., Ishikawa, Y., Ikeda, H. 1997. Dynamics of Wood Transport in Streams: A
Flume Experiment. Earth Surface Processes and Landforms, 22 (7), 669-683. 10.1002/(SICI)1096-
9837(199707)22:7<669::AID-ESP740>3.3.CO;2-C
Comiti F, Mao L, Preciso E, Picco L, Marchi L, Borga M. 2008. Large wood and flash floods: evidences
from the 2007 event in the Davca basin (Slovenia). In: De Wrachien D, Brebbia CA, Lenzi MA
(eds) Monitoring, simulation, prevention and remediation of dense and debris flow II. WIT Press,
Southampton, pp 173–182.
Díez Herrero, A., Ballesteros, J.A. 2009. Procesos activos. Los riesgos geológicos. El riesgo por avenidas e
inundaciones. En: C. Marchán (Coord. Gral.), Atlas del medio natural y de los recursos hídricos de
la provincia de Ávila. Capítulo 21, págs. 201-210, Instituto Geológico y Minero de España y Excma.
Diputación de Ávila, 240 pp.
Díez-Herrero, A., 2003. Geomorfología e Hidrología fluvial del río Alberche. Modelos y SIG para la gestión
de riberas. Serie Tesis Doctorales 2. Publicaciones del Instituto Geológico y Minero de España,
Madrid. 587 pp.
Díez-Herrero, A., Ballesteros-Cánovas, J.A., Bodoque, J.M., Ruiz-Villanueva, V. 2012. A new
methodological protocol for the use of dendrogeomorphological data in flood risk analysis.
Hydrology Research. In press.
Directive 2007/60/EC of the European Parliament and of the Council of 23 October 2007 on the
assessment and management of flood risks.
GEODE, 2011. Plan de Cartografía Geológica Continua Digital 1:50000. Instituto Geológico y Minero de
España (IGME; Geological Survey of Spain), http://cuarzo.igme.es/sigeco/default.htm.
Hilderbrand, R.H., Lemly, A.D., Dolloff, C.A., Harpster, K.L., 1998. Design Considerations for Large
Woody Debris Placement in Stream Enhancement Projects. North American Journal of Fisheries
Management 18:161–167
Hutte, P., 1968. Experiments on windflow and wind damage in Germany: Site and susceptibility of spruce
forests to storm damage. Forestry 41: 20-26.
Johnson, S. L., Swanson, F. J., Grant, G. E., Wondzell, S. M. 2000. Riparian forest disturbances by a
mountain flood? the influence of floated wood. Hydrological Processes 14: 3031-3050.
Kasprak, A., Magilligan, F. J., Nislow, K. H., Snyder, N. P. 2011. A LIDAR derived evaluation of
watershed-scale large woody debris sources and recruitment mechanisms : coastal Maine, USA. River
Research and Applications. doi:10.1002/rra.
Kraft CE, Warren DR. 2003. Development of spatial pattern in large woody debris and debris dams in
streams. Geomorphology 51: 127–39.
Lienkaemper, G. W., F. J. Swanson. 1987. Dynamics of large woody debris in streams in old-growth
Douglas-fir forests. Canadian Journal of Forest Research 17:150–156.
MAGRAMA, 2012. Sistema Nacional de Cartografía de Zonas Inundables (SNCZI). Guía Metodológica
para el desarrollo del Sistema Nacional de Cartografía de Zonas Inundables
http://www.magrama.gob.es/es/agua/temas/gestion-de-los-riesgos-de-
inundacion/snczi/default.aspx
Mao, L., Comiti, F., Lenzi, M. A. 2010. Bedload Dynamics in Steep Mountain Rivers : Insights from the
Rio Cordon Experimental Station (Italian Alps). Geological Survey Scientific Investigations Report
2010-5091.
Martin, D.J., Benda, L.E. 2001. Patterns of instream wood recruitment and transport at the watershed
scale. Transactions of the American Fisheries Society 130: 940-958.
May, C.L., Gresswell, R.E. 2003. Large wood recruitment and redistribution in headwater streams in the
southern Oregon Coast Range, U.S.A. Canadian Journal of Forest Research 33: 1352.
Mazzorana, B., Comiti, F., Fuchs, S. 2011. A structured approach to enhance flood hazard assessment in
mountain streams. Natural Hazards. doi:10.1007/s11069-011-9811-y.
Mazzorana, B., Hübl, J., Zischg, A., Largiader, A. 2010. Modelling woody material transport and deposition
in alpine rivers. Natural Hazards, 425-449. 10.1007/s11069-009-9492-y
Mazzorana, B., Zischg, A., Largiader, A., Hübl, J. 2009. Hazard index maps for woody material recruitment
and transport in alpine catchments. Natural Hazards and Earth System Science 9: 197-209.
DOI10.5194/nhess-9-197-2009
Mcdade, M.H., F.J. Swanson, W.A. Mckee, J.F. Franklin, J. Van Sickle. 1990. Source distances for coarse
woody debris entering small streams in western Oregon and Washington. Canadian Journal of
Forest Research 20:326–330.
Meisina, C., Scarabelli, S. 2007. A comparative analysis of terrain stability models for predicting shallow
landslides in colluvial soils. Geomorphology 87: 207-223.
Merten, E., Finlay, J., Johnson, L., Newman, R., Stefan, H., Vondracek, B. 2010. Factors influencing wood
mobilization in streams. Water Resources Research, 46(10), 1-13. 10.1029/2009WR008772
Meunier, M. 1989. Essai de synthèse des connaissances en érosion et hydraulique torrentielle, en La Houille
Blanche. 5, pp. 361-375.
MFE ,2011. El Mapa Forestal de España 1:25000 (MFE), Ministerio de Agricultura, Alimentación y Medio
Ambiente (MARM) http://www.magrama.es/es/biodiversidad/temas/montes-y-politica-
forestal/mapa-forestal/mfe_25.aspx
Mizuyama, T., Shimohigashi, H. 1985. Influence of fine sediment concentration on sediment transport
rates, en Jap. Civil Eng. Journal. 27-1.
Montgomery, D.R., T.M. Massong, C.S. Hawley. 2003. Influence of debris flows and log jams on the
location of pools and alluvial channel reaches, Oregon Coast Range. Geological Society of America
Bulletin 115: 78-88.
Naka, K. 1982. Community dynamics of evergreen broadleaf forests in southwestern Japan. I. Wind
damaged trees and canopy gaps in an evergreen oak forest. Journal of Plant Research 95: 385-399.
Nakamura, F., Swanson, F. J., Wondzell, S. M. 2000. Disturbance regimes of stream and riparian systems ? a
disturbance-cascade perspective. Hydrological Processes 14: 2849-2860. doi:10.1002/1099-
1085(200011/12)14:16/17<2849::AID-HYP123>3.0.CO;2-X.
Nakamura, F., Swanson, F.J., 1994. Distribution of coarse woody debris in a mountain stream, western
Cascade Range, Oregon. Canadian Journal of Forest Research 24: 2395–2403.
Palacios, D., Garcia, R., Rubio, V., Vigil, R., 2003. Debris flows in a weathered granitic massif: Sierra de
Gredos, Spain. Catena 51: 115–140.
Palacios, D., Marcos, J., Vázquez-Selem, L. 2010. Last glacial maximum and deglaciation of Sierra de
Gredos, Central Iberian Peninsula. Quaternary Internacional 233: 16–26.
Rickenmann, D., 1990. Bedload transport capacity of slurry flows at steep slopes. Mitteilung VAW 103.
Zürich.
Rickenmann, D., 1991. Hyperconcentrated flow and sediment at steep slopes, en J. of Hydr. Eng. 117:
1419-1439.
Rickenmann, D., Koschni, A. 2010. Sediment loads due to fluvial transport and debris flows during the
2005 flood events in Switzerland. Hydrological Processes 24: 993-1007.
Robison, Beschta,. 1990. Identifying trees in riparian areas that can provide coarse woody debris to streams.
For. Sci. 36: 790-801
Rosso R, Rulli MC, Bocchiola D. 2007. Transient catchments hydrology after wildfires in a Mediterranean
watershed: runoff, sediment and woody debris. Hydrol Earth Syst Sci. 11: 125–40.
Ruiz-Villanueva, V., Bodoque, J. M., Díez-Herrero, A., Calvo, C. 2011. Triggering threshold precipitation
and soil hydrological characteristics of shallow landslides in granitic landscapes. Geomorphology
133: 178-189. 10.1016/j.geomorph.2011.05.018
Ruiz-Villanueva, V., Bodoque, J.M., Díez-Herrero, A., Eguibar, M. A., Pardo-Igúzquiza, E. 2012.
Reconstruction of an ungauged flash flood event with large wood transport and its influence on
hazard patterns. Hydrological Processes. First on line: DOI: 10.1002/hyp.9433.
Ruiz-Villanueva, V., Díez-Herrero, A., Stoffel, M., Bollschweiler, M., Bodoque, J. M., Ballesteros, J. A.
2010. Dendrogeomorphic analysis of flash floods in a small ungauged mountain catchment (Central
Spain). Geomorphology 118: 383-392. 10.1016/j.geomorph.2010.02.006
Ruiz-Villanueva, V.,. Bladé-Castellet, E., Sánchez-Juny, M., Martí, B., Bodoque , J.M., Díez Herrero, A. 2D
modelling of woody debris transport in rivers. Under review. Journal of Hydrology.
Scholz, R.W., Tietje, O. 2002, Embedded case study methods: integrating quantitative and qualitative
knowledge. Thousand Oaks, Sage Publ.
Seo, J.I., Nakamura F., Akasaka, T., Ichiyanagui, H., Chun, K.W. 2012. Large wood export regulated by the
pattern and intensity of precipitation along a latitudinal gradient in the Japanese archipelago. Water
Resources Research 48. doi:10.1029/2011WR010880.
Smart G.M., Jaeggi M. 1983. Sediment transport on steep slopes. Mitteilung VAW. 64. Zürich.
Sobota D, Gregory S, Sickle JV. 2006. Riparian tree fall directionality and modeling large wood recruitment
to streams. Canadian Journal of Forest Research 36: 1243–1254.
Stoffel, M., Bollschweiler, M. 2008. Tree-ring analysis in natural hazards research – an overview. Natural
Hazards and Earth System Science 8: 187-202. 10.5194/nhess-8-187-2008
Stoffel, M., Wilford, D. J. 2011. Hydrogeomorphic processes and vegetation: disturbance, process histories,
dependencies and interactions. Earth Surface Processes and Landforms. 10.1002/esp.2163.
Stumbles, R.E. 1968. How wind influences silviculture and management as a district officer sees it. Forestry
Supplement, 45-50.
Swanson, F.J. 2003. Wood in rivers: a landscape perspective. American Fisheries Society Symposium 37:
299-313.
Swanson, F.J., S.L. Johnson, S.V. Gregory, and S.A. Acker. 1998. Flood disturbance in a forested mountain
landscape – interactions of land use and floods. Bioscience 48: 681-689.
Van Sickle, J., Gregory, S. V. 1990. Modeling inputs of large woody debris from falling trees. Canadian
Journal of Forest Research 20: 1593–1601.
Webb, A.A., Erskine, W.D. 2003. A practical scientific approach to riparian vegetation rehabilitation in
Australia. Journal of Environmental Management 68: 329–341.
Wohl, E., Bolton, S., Cadol, D., Comiti, F., Goode, J. R., Mao, L. 2011. A two end-member model of wood
dynamics in headwater neotropical rivers. Journal of Hydrology, 1-10.
Zadeh, L.A. 1965. Fuzzy sets. Information and Control 8: 338–353.
Zadeh, L.A. 1968. Fuzzy algorithms. Information and Control 12: 94–102.
Zimmermann, H. 2001. Fuzzy set theory and its applications. Boston: Kluwer Academic Publishers.
(4.5). 3. Results
3.1. General settings
3.2. Results of log motion
REFERENCE:
Ruiz-Villanueva, V., E. Bladé Castellet, M. Sánchez Juny, B. Martí, J.M. Bodoque, A. Díez
Herrero. Floating wood transport: 2D Model developing and flume experiments.
ABSTRACT
The transport of wood material in rivers has been the subject of various studies in
recent years. Most research has focused on the geomorphologic role of woody debris, its
recruitment processes and spatial distribution in streams. In this paper, a numerical
hydrodynamic model has been developed to simulate wood transport in rivers, together
with other fluvial processes such as sediment transport and turbulence. This tool has been
implemented as a new computational module into the Iber two-dimensional hydraulic
simulation software. The new module calculates the position and velocity of differently
shaped logs using a kinematic and a dynamic approach, both based on the balance of forces
involved in the wood movement. These approaches consider the interaction between the
logs themselves and between the logs and the channel configuration. Prototype-scale flume
experiments validated the capacity of the model to reproduce accurately the movement of
floating logs in a straight channel with constraints. This paper presents a first attempt to
include wood transport in two-dimensional hydrodynamic river simulation. This new
modelling capacity will potentially improve hazard and risk analyses, optimize
infrastructure design and facilitate river management and restoration tasks.
1. Introduction
The flood events which occurred in Switzerland and western Austria in 2005 were
characterized by increased impact and associated losses due to the transport of woody
material, as described by Mazzorana et al., 2010a and 2010b. The damage and extreme
water stages observed in a mountain basin in the Slovenian Alps in 2007 were ascribed, at
least partly, to the collapse of temporary dams formed by sediments and large woody
debris, according to Comiti et al., 2008a and 2008b. The effects of changing channel
morphology during flood events and the reduction of cross-sectional area due to wood
clogging were found to amplify process intensities significantly (e.g., Lyn et al., 2007; Mao
and Comiti, 2010). In central Spain, the influence that wood transport during floods has
on hazard patterns may also be considerable (Ruiz-Villanueva et al., under review).
The Japanese government makes enormous efforts to protect cities from so-called
‘floating log disasters’, yet little is known about the physical dynamics of large woody debris
(LWD) transport (Braudrick et al., 1997).
On the other hand, the influence of LWD on flow resistance has only been examined
by a few researchers (Manga and Kirchner, 2000; MacFarlane and Wohl 2003; Wilcox et
al., 2008). Some studies have attempted to characterize the transport and hydrodynamics
of wood in rivers in quantitative terms. In this research, WD mobilization has been
investigated in laboratory flumes either by combining detailed hydraulic predictions of
flume studies with real data from field studies, and/or through analytical approaches
(Braudrick and Grant, 2000; Braudrick et al., 2001; Haga, 2002; Bocchiola et al., 2002,
Buxton, 2010; Wohl et al., 2011 and Wohl, 2011). These studies successfully predicted
woody material entrainment based on the balance of hydrodynamic and resistance forces
and some of them dealt with transport regimes.
In recent years, a few studies have been carried out to incorporate this phenomenon
into flood hazard and risk analysis, e.g. as a stochastic process (Merten et al., 2010;
Mazzorana et al., 2010a; Mazzorana et al., 2010b), but there is no tool currently available
to simulate hydrodynamic and other fluvial processes (sediment transport, turbulence etc.)
together with wood transport.
The main aim of this paper is to develop a numerical method for simulating the
transport of wood in rivers based on the water levels and flows obtained with a two
dimensional hydrodynamic model.
The model presented here was developed in Fortran programming language and
implemented as a module (Woody Iber) into the Iber (Corestein et al., 2010) two-
dimensional hydrodynamic software, developed by CIMNE, GEAMA group (UDC) and
Flumen Research Institute (UPC). Iber (www.iberaula.es) is a numerical tool for the
two-dimensional simulation of turbulent free surface unsteady flow and sediment
transport in watercourses. At present, Iber has three main computational modules: a
hydrodynamic module, a turbulence module and a sediment transport module. The
hydrodynamic module solves the depth-averaged shallow water equations (2D-SWE)
The wood transport module presented here allows the inclusion of wood pieces in
the simulation, combining an Eulerian scheme for water flow with a Lagrangian approach
for wood transport. The method uses the flow variables calculated with the hydrodynamic
module to update the position and velocity of the wood logs at each time step. The wood
transport method can thus be combined with any 2D hydraulic computer modelling
package using an explicit scheme, such as the majority of newly developed schemes based
on the finite volume and the finite element methods.
Flume experiments were carried out to validate the model. The behaviour of wood in
rivers was simulated by the model and reproduced on a prototype-scale for different flume
configurations and flow conditions. Several types of dowels served as LWD. The
experiments were recorded and the log trajectories obtained from the videos were
compared with the results of the numerical simulation.
The structure of this paper is as follows: Section 2.1 shows the model
implementation which contains: (i) analyses of the incipient motion and wood transport
regimes implemented; (ii) calculation of the wood velocity and orientation ; (iii)
description of the influence of turbulence; (iv) interactions between logs and obstructions;
(v) effect of logs on the hydrodynamics; (vi) the initial and boundary conditions of the
wood. Section 2.2 describes the model validation procedure by flume experiments. Section
3 presents and compares experimental and numerical results. Finally, Section 4 discusses
the results obtained, highlighting the implications and limitations of the proposed
numerical model at present. A discussion of potential applications and use in real rivers is
also provided.
2. Model set-up
First of all, the incipient motion and transport regimes have to be established. Then,
if the entrainment is considered, the wood will react according to the forces acting on each
log. Therefore, flow exerts an influence on the logs but the opposite is also true, as the
presence of logs will affect the flow hydrodynamics. The Lagrangian scheme developed for
wood pieces computes all the log parameters (position, velocity, rotation angle, shear stress
etc.) in the centre of mass and at the ends of each log. At the same time the hydrodynamic
parameters (affected by the influence of the log presence) are computed using the Eulerian
scheme.
The gravitational force ( Fg ) acting on the log is equal to the effective weight of the
log ( Fw ) in the downstream direction and is equal to
Fig.1:: Schematicc and body-fforce diagraam of the foorce balancee componennts acting on
n a piece
of wood.d.
where Lw is the piecce length, w and are the densitiesd off wood annd water
respecctively, is
i the bed angle
a of thee bed slopee, g is gravvity, Aw is the area off the log
perpenndicular to piece lengtth (Fig.1)
Aw Dw / 4
2
[2]
and Asub iss the submeerged area oof the log perpendicul
p lar to piecee length exp
posed to
drag.
Asub is a fu
unction of the
t log drafft ( y ) and piece diam
meter ( Dw ),, which forr a right-
circulaar cylinder resolves to
Asub 2 ccos 1 1 2 y / Dw sinn 2 cos 1 1 2 y / Dw Dw / 8
2
[3]
VIRGINIA
A RUIZ-VILLANUEEVA 259 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
The friction force in the opposite flow direction ( F f ) is equal to the normal force (
Fn Fw cos ) acting on the log times the coefficient of friction between the wood and
the bed:
where bed is the coefficient of friction between the wood and the bed.
The drag force ( Fd ), also in the flow direction, is the downstream drag exerted on
the log by the water in motion and is equal to
Where Uflow is the water velocity, U log is the log velocity and Cd is the drag
coefficient of the wood in water. This last parameter has been analysed extensively before.
Brooks et al., (2006) proposed 1.2 for wood in real streams, and Bocchiola et al., (2006)
found a value of 1.41 for dowels in flume experiments. In this study it is assumed constant
but the modeller can change its value in each simulation.
The combination of these three equations yields the force balance at incipient
motion for a circular cylinder lying on the river bed: F f Fg Fd
g w Lw Aw g Lw Asub bed cos sin
[7]
U flow / 2 C d Lw h sin Asub cos
2
They used the average flow velocity ( U flow ) as reference velocity for the moving
woody material for a wide range of flow conditions, and then the velocity along the
transport trajectory for each moving woody material ( U log ) model log is estimated as
The transport inhibition parameter ( C * ) was recalculated using the relative velocity
of the water in relation to the log ( U r U flow U log )
Given these results and based on the log density, log diameter and water depth, four
main scenarios can be distinguished and related to the transport regimes (Fig. 2).
In Fig. 2, scenario A, the log density is assumed to be greater than the water density,
and the water depth is lower than the log diameter and equal to the submerged log
diameter. In this situation if U lim U flow the resting condition is imposed on the log, and if
U lim U flow the log will move, sliding with velocity U log U flow U lim .
When the log density is greater than the water density and the water depth is higher
than the log diameter and higher than the submerged log diameter, i.e. the log is
completely submerged (Fig. 2, scenario B), then again if U lim U flow the log will move and
slide, otherwise it will repose. The log velocity is U log U flow U lim
Fig.2: Cross section of a piece of wood in a stream in different conditions. (A): Log density
greater than water density, and water depth lower than log diameter. (B): Log density
greater than water density, and water depth higher than log diameter (C): Log density
lower than or equal to water density, and water depth higher than or equal to log diameter.
(D): Log density lower than or equal to water density, and water depth lower than log
diameter. Log diameter (Dw), water depth (h), submerged log area (Asub) and submerged
log diameter (y).
When the log density is equal to or lower than the flow density and the water depth
is higher than the log diameter and higher than the submerged log diameter (Fig. 2,
scenario C), then the log will float, U lim 0 , and will be transported with a velocity equal
to the flow velocity, except in the case of interaction with other logs or channel walls.
In Fig. 2, scenario D, the log density is equal to or lower than the flow density and the
water depth is lower than the log diameter and equal to the submerged log diameter, then
U lim is calculated using equation 8. If U lim U flow the resting condition is imposed, and if
U lim U flow the log will slide with velocity: U log U flow U lim . The log will only float
when w Aw Asub , going back to scenario C.
As well as sliding, a piece may also move by rolling (Bocchiola et al., 2006) in
scenarios A, B and D, but the moment forces involved are beyond the scope of this study.
Here the focus is on a floating regime since this is the main wood transport regime during
flood events (Ballesteros et al., 2011).
The movement of wood logs includes both translation and rotation. Two methods
have been implemented to estimate wood velocity. First, the kinematic method is based on
the work of Mazzorana et al., (2011) with some new approaches described above, thus the
velocity of a wood log is obtained from equation [11]. The second or dynamic method,
applies Newton’s second law to every log exposed to gravity, friction and drag forces, now
in vectorial form: Fg , Ff , Fd . For each time step the log velocity, now considered as a two
dimensional vector in the horizontal plane, is updated with:
n 1
Ulog Ulog
n
t a [13]
where Ulog
n
is the velocity vector at time n, a is the log acceleration
Fg Ff Fd
a [14]
w Lw Aw
and t is the time step. The log position at time n X log
n
( xlog
n
, ylog
n
) is then updated
to time n+1 with
n 1 1
X log X log
n
t U log
n
a t 2 [15]
2
Apart from the translatory movement described, logs turn driven by the velocity
distribution across the flow section. Because one end of the piece of wood is moving faster
than the other end, the piece rotates towards a more flow-parallel orientation. Since the
velocity field still varies across the piece, under uniform flow conditions the piece
continues to rotate towards a flow-parallel orientation until it reaches a stable orientation.
Naturally, if the flow is not uniform and the river is non-prismatic, the process produces
continuous turning of the logs which change their orientation over time. To simulate these
orientation changes the velocities at each end of every log were obtained from the flow
model. Depending on the mesh size with respect to the log size, one log can be contained in
a single mesh element (finite volume) or in more than one, the flow velocity U1flow ,2
at each
end (1 or 2) of the log is calculated from the flow velocity at the log centre U flow , the
cm
velocity gradients and the relative position of the log ends with respect to its centre:
U flow
U1flow
,2
Ucm
flow (X1,2 Xcm ) [16]
X
The angular velocity of the log can be obtained from the velocity values at the log
ends and thus the change in log orientation at every time step.
2.1.3. Turbulence
Although the flow in river channels is turbulent, for river flow where the geometry is
smooth enough, no re-circulation zones appear and roughness acts as the principal factor
of vortex stabilization, the inclusion of turbulence models has little or no effect on the
velocity field (Cea, 2005). Nevertheless, even in these cases it is important to consider
turbulence while modelling the transport of suspended substances or sediments as
dispersion can be affected by the turbulent viscosity. Similarly, in mountain rivers, small
swirls may appear and disappear with an almost chaotic movement, showing the
complexity of the turbulent motion and this turbulence may affect wood transport. Iber
includes several turbulence models, but for the influence of turbulence on wood transport
the Rastogi-Rodi k-ε model (Rastogi and Rodi, 1978) is used, as it can give information on
the magnitude of the turbulent kinetic energy. The k-ε model considers turbulence caused
by bed friction, velocity gradients, dissipation and convective transport.
For wood transport, therefore, turbulence is included in the simulation with the k-ε
model, and the wood velocity is re-calculated based on the turbulent velocity (Ut). This
turbulent velocity is an approximation to the instantaneous velocity reconstructed from
the mean velocity and the ratio between turbulent kinetic energy k , obtained from the k-ε
model, and mean kinetic energy K 1 / 2 U flow 2 where Uflow is the mean flow velocity.
Ut U flow 1 R Ti [17]
where R is a random number between 1 and -1, and Ti is the coefficient of turbulent
restitution
k
Ti [18]
K
Ti is the ratio between turbulent fluctuations, resulting from the internal chaotic
fluid motions that characterize turbulent flow, and mean velocity. Including this
coefficient of turbulent restitution in the wood transport simulation basically means
introducing a random component into the movement of logs placed in a turbulent flow.
Thus, identical logs dropped into the same spot may end up in different places.
The interactions between the logs and the channel configuration and among the logs
themselves have been taken into account in the model. The focus was on the changes in log
velocity due to contact with the banks or with other pieces, but the effect of branches or
roots has been ignored .
If one piece of wood is floating in the flow and meets another piece (floating or
resting) then the two may collide and continue moving with a different velocity (Fig. 3,
(A)). This new velocity, referred to as final velocity Vf is calculated from the initial velocity
Vf for both pieces as
Vf (1 e) Vcm e Vi [19]
where
m1 V1 m2 V2
Vcm [20]
m1 m2
is the centre mass velocity during the collision, e is the restitution coefficient (equal
to 1 assuming elastic interaction), m1 and m2 are the log masses, and V1 and V2 are the
initial velocities.
When a piece of wood reaches the bank, it can be entrapped if the slope of the
riverbank is not steep and the water level is falling or if it reaches a dry area, or it can
bounce and change its trajectory and velocity.
In the first case (Fig. 3, (B)), the driving forces decrease due to the reduction of the
submerged area, but the resisting forces are still active around the log, therefore the initial
motion condition is re-calculated.
Fig. 3: Schematic illustrations of log interactions with each other and with the river
banks. (A) In time interval 1, one piece of wood is moving in the flow direction and meets
another piece; they collide in time interval 2 and after the collision they continue moving
with different velocities. (B) Part of the log is out of the river (dry area). (C) Log hits the
bank and slides parallel to it before being reincorporated into the flow. (D) Log hits the
bank and bounces off. Different time steps are represented as t1, t2 and t3,; α is the
incidence angle.
If the log hits a bank or boundary, the log trajectory and velocity may change with a
different pattern depending on the incidence angle, as was observed in the flume
experiments. If the incidence angle is lower than a given value (a threshold of 45º is
assumed here, but this value can be modified by the modeller) the movement of the log is
treated as sliding (Fig. 3, (C)). On the other hand, if the incidence angle is higher the log
bounces and the velocity on the side of the bank decreases, and the log trajectory changes in
terms of the flow field velocity (Fig. 3, (D)).
In natural streams, pieces of wood are often braced against obstructions in the
channel, stream banks, live vegetation, or other pieces of wood (Fig. 4; Merten et al., 2010).
F
log s
d
τ wood ,i [21]
Vi
where τwood,i is the shear stress at every finite volume, or mesh element, i, Fd the drag
forces as in [6] and Vi the volume of the 2D finite volume, or area of mesh element, i.
Fig. 4: Sketch illustrating the influence of woody debris (deposits or individual logs) on
hydrodynamics.
The initial conditions must be assigned to the whole domain. To do this, the initial
position of each log (x, y coordinates of mass centre and angle with respect to the flow), its
length, diameter and density are specified.
Inlet boundary conditions can also be assigned, specifying a number of pieces per
minute (or volume) as a condition. Based on the knowledge of the fluvial corridor, the
riparian vegetation, and the wood availability, ranges of maximum and minimum lengths,
diameters and density of wood are established. Then stochastic variations of these
parameters together with position and angle need to be used to characterize each piece of
wood entering the simulation.
The numerical model was validated by means of flume experiments. The objective
here was not to study the patterns or behaviour of logs in rivers, but to test the accuracy
and performance of the numerical model described above. The flume was simulated at
scale 1:1.
Flume experiments were conducted at the Flumen Research Institute (UPC) using a
re-circulating flume with known bed-slope and roughness. The experiments were set up
based on the previous work by Braudrick et al., 1997; Braudrick and Grant, 2001;
Bocchiola et al., 2006a and 2006b; and Wilcox and Wohl, 2006, but some specific
characteristics were adapted due to the different aim of this study.
The flume is 0.6 m wide and 10 m long, with a rectangular cross section and smooth
sidewalls (see Fig. 5), and known roughness. The channel is fed with water from a re-
circulatory system. Upstream of the flume there is a cobble-filled baffle to dissipate energy
and ensure a uniform flow. All flume runs were completed with a non-erodible boundary
and no sediment transport.
These experiments were designed to validate the log transport numerical model in a
simple geometry and limit uncertainties as far as possible. Further complications with
changing bed geometry or sediment interactions were left for future research.
Round white beech wooden dowels were used to mimic the movement of a piece of
LWD under the influence of water flow. Five sizes of dowels were introduced into the
flume at three different density levels, with three different orientations in five different
geometric configurations. This procedure allowed us to obtain different scenarios (several
log shapes, log interactions, different log velocity etc.) to test the model using several
calculations, and this was the main aim of this work.
The five types of LWD pieces were: (i) type 1: short dowels (0.1 m, less than half the
flume width) with medium diameter (10 mm); (ii) type 2: medium dowels (0.2 m, close to
half the flume width) with small diameter (8 mm); (iii) type 3: medium dowels (0.2 m,
close to half the flume width) with large diameter (12 mm); (iv) type 4: long dowels (0.5
m, more than half the flume width) with medium diameter (10 mm); (v) type 5: medium
dowels (0.2 m ) with very large diameter (18 mm). The average density of this type of wood
is 720 kg m-3. The logs were painted in bright colours to facilitate monitoring.
Foor individuual pieces on nly, the orieentation waas changed manually iin the initiaal time
step. Thhree preferrred orientattions were ttested: paraallel, oblique and perpeendicular to
o flow,
to forcee different trajectories and
a log rotaations.
T
The straightt channel was w modifiied with seeveral lateraal constricttions and central
c
obstaclees (Fig. 5) too form diffe
ferent two-ddimensional velocity fields, differeent dischargges (12
-1
and 18 l s ) and booundary con c and 9.7 cm weir heeight) were run to
nditions (crritical, 5.8 cm
work w
with low (higgh flow velo ocity) and hhigh levels (llow flow veelocity).
Fig. 5: Flum
me geometrric configura
rations. Geoometry num
mbers refer tto Table 1.
T
The 3D (u,, v, w) vellocity fieldd was meassured with a MicroAAcoustic Doppler
Velocim
meter (Vecttrino, www
w.nortekusa..com). Thee sampling volume off the ADV was 7
VIRGINIA RUIZ-VILLANUEVA
A 270 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
mm length cylinder with 1.8 mm transmit length. The control volume was located 5 cm
away from the probe to reduce flow interference. The maximum sampling rate of the
Vectrino used in the experiments was 200Hz.
The calibration of the velocimeter and data recording was managed through
Vectrino Plus software (User Guide, 2009) and the raw data obtained was processed and
filtered with WinADV software (Wahl, 2000). The filtering parameters were 80%.
All flume runs were recorded using a wide-angle digital camera installed above the
flume to obtain a perpendicular overhead view. The video frames encompassed the entire
flume and were recorded at a rate of 30 frames per second.
A code was developed in Matlab to determine the log trajectory throughout the
flume. This code computes the log centre and orientation on each frame. The input data
for each run include : length, width and colour of the tested logs, and approximate angular
orientation and position of the log on the initial frame. Log dimensions and location
coordinates are specified in image pixels.
For each frame, the code computes a combination of the red, green and blue image
channels, yielding a new channel hereafter referred to as N. This combination depends on
the log colour and enhances the contrast between the log pixels and those corresponding to
the surrounding water surface and flume walls. N channels of consecutive frames (i and
i+1) are then subtracted on a pixel basis. In the resultant Ni+1 –Ni channel, maximum pixel
values indicate the log position on frame i+1, while the minima correspond to the ith
position. Then, a template with the log length and width is shifted on the Ni+1 –Ni channel
around the coordinates of the known log centre at frame i, using a range of different
orientations around the known ith orientation. For each template position and orientation,
the sum of Ni+1-Ni pixels within the template is computed. The template position giving
the maximum sum indicates the log location on frame i+1. These centre coordinates and
orientation angle are stored as those corresponding to frame i+1 and used as initial
conditions in the next iteration to compute the log location on frame i+2.
3. Results
To solve a differential equation using the finite volume method, first the studied
domain must be spatially discretized by dividing it into relatively small cells. To do this, the
different flume geometry configurations were input into the model creating different 1 cm
resolution structured meshes. For all the discharges studied the flow was subcritical at the
boundaries. At the inlet boundary the total water discharge was distributed over a uniform
unit discharge profile perpendicular to the boundary. The bed friction is defined as a
Manning roughness coefficient, which is assigned to each element of the mesh; the value of
the roughness is determined based on numerical and experimental comparison of steady
backwater profiles (0.01 s m-1/3). The simulations were run with and without turbulence.
When turbulence was included, the k-ε model was selected.
The hydrodynamic simulation represents the flow conditions in the flume with good
accuracy (numerical and experimental comparison for geometry 1, in Fig. 6); the x-
direction velocity has a correlation coefficient of 0.91 (R2=0.84). The deviation was due to
some 3D effect in the flow which cannot be simulated by the 2D model.
VIRGINIA
A RUIZ-VILLANUEEVA 273 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
As described above, four different geometries were used to obtain different flow
conditions (Table 1). Geometry 1 and 3 present similar water depth conditions although
discharge, velocity field and constrictions were different; geometries 2 and 4 present higher
water depth ratio (difference between inlet or maximum and outlet or minimum; see
Table 1). The highest ratio was observed in geometry 2. Geometry 3 presents the lowest
values for flow velocity and geometry 4 the highest.
The first verification carried out was the proper simulation of the log movement,
including both translation and rotation. Various logs were placed in the flume with
different orientations as described above, and these runs were recorded and simulated. As
expected, logs placed parallel to the flow direction continued to move with this orientation
following the centre line; however, logs placed initially oblique (45º) or perpendicular
(90º) to the main flow direction turned until they reached the same orientation as in the
previous case (Fig. 7).
After these initial trials, the model log trajectories and log velocity values were
validated, testing the kinematic and dynamic methods implemented, described in the
methodology section. The simulation results were compared with the post-processed
records, analysing the model accuracy with the correlation coefficient, absolute and
squared errors and statistical significance difference (p-value); one of these simulations is
shown in Fig. 8 and statistics are provided in Table 2.
Figure 7: Numerical
N model resuults (geomeetry 3): thrree logs plaaced with different
d
otated to a more stabl
orienttations (on the left) ro ble hydrodyn namic positition, paralleel to the
flow ddirection. Log trajectorries are show
wn as blackk lines. Flow
w direction lleft to right.
VIRGINIA
A RUIZ-VILLANUEEVA 275 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER
C 4
VIRGINIA RUIZ-VILLANUEVA
A 276 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
The main difference between tthe two approaches iss that the kinematic method
calculates log veloocity based on the flow w velocity and
a the tran nsport inhibbition param
meter, as
both wwood and flow
f velocities are veryy similar (coorrelation co
oefficient > 0.8) but not
n equal
due too turbulencce. (When there
t is no tturbulence,, log velocity and flow velocity aree exactly
equal)). The dynaamic metho od, on the oother hand, takes log accelerationn into consid deration
(basedd on actingg forces) wh hen calculaating log veelocity (Fig. 9). This w will be discussed in
detail in the Disccussion section.
VIRGINIA
A RUIZ-VILLANUEEVA 277 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER
C 4
Foor the previous test caase, simulatiion results using the kinematic
k m
method are shown
in Fig. 110 together with the main
m hydroddynamic parrameters.
T
Turbulence is an impo ortant issuee to consideer when modelling wa water system ms, and
thereforre a key asppect of the model validdation was the analysiis of the efffect of turb
bulence
when mmodelling loog transportt.
VIRGINIA RUIZ-VILLANUEVA
A 278 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
When turbbulence is simulated, tthe hydrodyynamic con nditions chaange, and represent
more precisely thhe flume flo
ow conditionns. In the previous
p exaample, if theere is no turrbulence
presennt, flow velocity increaases and thherefore thee logs move faster, incrreasing the error in
log poosition (Fig.. 11).
VIRGINIA
A RUIZ-VILLANUEEVA 279 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER
C 4
Fig. 12:
2: (A and B):
): Log (masss centre) traj
ajectory in geometry
g 2 observed
o annd simulateed with
and wwithout turrbulence. Correlation
C ccoefficient. > 0.8. (C) Log
L positioon over timee in:
observeed and simuulated. (D) Absolute
A errror in log position
p cooordinates (m
meters) over
er time.
(E) SSquared errror.
VIRGINIA RUIZ-VILLANUEVA
A 280 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
As already seen above, log trajectories do not change significantly, but log velocity
changes due to flow conditions. Table 3 shows the statistics for these simulations.
Table 3: Statistics for the test case shown in Fig.12. X_obs and Y_obs are X, Y coordinates
observed in the flume and extracted from the records. Sim and sim_k-e are X, Y
coordinates resulted from the simulation applying kinematic method without and with
turbulence respectively.
A general observation is that error seems to increases in the flume outlet in all cases.
This has led us to consider including another source of error in the validation process. This
will be discussed in Section 4.
Fi
Figure 13: Numerical
N model
m results
ts (kinematiic method) of Geometr try 2: water depth,
flow veelocity (moodule), Frouude numberr and turbullence viscossity are show wn togetherr with
l trajector
log ory (black linnes). Flow direction
d lefft to right .
VIRGINIA RUIZ-VILLANUEVA
A 282 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
One of thee most impoortant effects of wood in hydrodyynamics is thhe backwatter effect
due too clogging. Flume
F conffiguration 4 allowed uss to test thiss effect (Fig.
g. 15).
VIRGINIA
A RUIZ-VILLANUEEVA 283 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER
C 4
4. D
Discussion
A numerical model wass developedd to simulatte the hydrodynamics and other fluvial
processees (turbulen
nce etc.) with wood traansport. Th his model reepresents thhe first attem
mpt to
introduuce wood trransport intto the 2D hhydrodynam mic simulation of riverrs. The mod del has
been teested and vaalidated wiith flume eexperimentss. An in-depth discusssion of these and
other isssues (compparison witth other appproaches, potential
p uses
u and appplications etc.) e is
provideed, togetherr with a careeful analysiss of all the parameters
p described
d abbove.
VIRGINIA RUIZ-VILLANUEVA
A 284 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
Figure
re 16: Geommetry 4 simuulation resul
ults. (a) Logs
gs blocked (ccoloured linnes) and traj
ajectories
(bla
lack lines); (b)
( Water depth
d with aand withouut LWD ; (cc) Velocity m module witth and
with
thout LWD D ; (d) Froudde number w with and without
w LWWD ; (e) Turrbulence visscosity
with aand withoutut LWD .
VIRGINIA
A RUIZ-VILLANUEEVA 285 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
The main simplification assumed by the model was the shape of the logs as cylinders,
avoiding the effect of branches or roots. This geometry is not representative of woody
debris with complex shapes, but it provides a good approximation which has been used by
several researchers (see Bocchiola et al., 2008; Braudrick et al., 1997; Buxton, 2010;
Mazzorana et al., 2011), and may provide a picture of non-rooted and defoliated logs often
occurring in rivers as a result of fluvial transport, wood harvesting and forest fires
(Bocchiola, 2011). Braudrick and Grant (2000) proposed a theoretical method to attach
the rootwad to a log using a disk on the cylinder ends. This will be taken into account for
future developments.
The general approach of initial motion is based on previous works (see Braudrick and
Grant, 2000; Haga et al., 2002; Mazzorana et al., 2011) but some modifications have been
made. Since the numerical model calculates the flow and log velocities at every time step
when the incipient motion is computed we modified the transport inhibition parameter (
C * ) proposed by Mazzorana et al., 2011, using the relative velocity of the water in relation
to the log velocity, instead of the flow average velocity.
Logs may move by rolling, sliding or floating. These transport regimes are simulated
by the model using the relations between log and flow densities, water depth and log
diameter. When these parameters are calculated the model assigned one of these regimes to
each log (each time step), and with this method, floating logs can be simulated and also
rolling or sliding logs (although the moment forces involved are outside the scope of this
study at this stage), but the model does not take the Z coordinate into account.
Then the log velocity is calculated as described above, using kinematic and dynamic
methods; velocity is broken down into two dimensions (X and Y axes). Most research on
log motion uses continuity or Manning equations to estimate flow velocity based on float
and stop watches (Braudrick and Grant, 2000 and 2001; Bocchiola et al., 2008; Bocchiola,
2011) or measures reach-average velocity using salt dilution methods (Wilcox and Wohl,
2006); some others used 1D or 2D models (Mazzorana et al., 2011; Merten et al., 2010),
first computing the hydraulics (1D or 2D) and then using the results to calculate the wood
mobilization. Here the two horizontal components of log velocity are calculated for each
time-step at each end of every log and at the log centre based on the flow model at the same
time as the flow velocity. According to our findings, at present the kinematic method
reproduces the log movement more precisely (in time and space), and therefore the
dynamic method will have to be improved.
Here the log translation is calculated and also the log rotation and orientation for
every time step. This has been achieved by computing the calculations in the log centre
mass and at both log ends.
The drag coefficient of the wood in water is an important parameter in the log
velocity calculation that must be discussed here. Natural pieces of wood may also have a
greater drag due to skin friction (Merten et al., 2010) but this has not been taken into
account in this paper. In this approach, the drag coefficient is used as a constant value,
which can be modified according to each case simulated. This coefficient varies with the
log shape, its position in the stream and Reynolds number. However, the drag coefficient
on many different shapes of bodies has been studied extensively. Brooks et al., (2006)
reviewed the literature and selected a value of 1.2 for wood in streams, whereas Bocchiola
et al., (2006) used 1.41 for dowels in flume. We take the value proposed by this latter
author. According to our findings, this parameter is less important for the kinematic
method, but for the dynamic method it is more sensitive (Fig.17). Therefore, we do not
discard modifying this approach and calculating it as a variable in the future.
The displacement of wood may be affected by turbulence, and the model is able to
incorporate this effect into the simulation. When the turbulence is computed the model
recalculates the log velocity using a coefficient of turbulence restitution as explained in
Section 2.3 above. Furthermore, the results also showed how turbulence may change due to
the presence of logs; therefore in the authors’ opinion this is an important contribution to
the hydrodynamic simulation.
The interaction between the logs themselves was solved assuming the simplest case,
elastic interaction, but a real-life case could be much more complex. The interaction
between logs and banks was simplified as well, assuming an incident angle as threshold
(based on flume observations). These interactions will be analysed and improved shortly in
a future study. As far as the authors are aware, there are no similar approaches previously
published in the relevant literature.
T
The influencce of woodyy debris onn hydrodynamics is weell establishhed, e.g. as Gippel
G
nts, reducingg average veelocity and locally
(1995) states “debrris acts as laarge roughnness elemen
ng the wateer surface profile”.
elevatin p Soome method ds can be found in th the literaturre, e.g.
Manga and Kirchn ner, 2000; Bocchiola,
B 22011; Hygeelund and Manga,
M 200 3. This effeect was
VIRGINIA RUIZ-VILLANUEVA
A 288 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
solved here by including drag force as an additional term similar to roughness following the
methodology proposed by Cea and Vázquez-Cedón (2009). They considered an effective
porosity parameter and additional drag to reduce available storage volume due to small-
scale obstructions which are not resolved by the numerical mesh. Of these two terms,
which are not present in the classical depth-averaged shallow water equations, drag is more
relevant, and even more so in the case of logs where only part of the depth is affected.
Probably one of the most relevant steps is the establishment of initial and boundary
conditions because, as is usual in environmental models, the results are conditioned by
them. An in-depth understanding of the fluvial corridor and the riparian vegetation is
needed to establish the availability of wood and the ranges of its maximum and minimum
lengths, diameters and density. However, methodologies have been developed before on
the characterization of wood, its distribution along the river and potential availability, and
its recruitment processes (see : Abbe and Montgomery, 2003; Andreoli et al., 2007;
Mazzorana et al., 2009; Wohl and Jaeger, 2009; Wohl et al., 2011). But here the number of
input trees (logs) has to be estimated, and logs have to be characterized individually. This
has not been achieved before, and is the aim of a forthcoming paper (Ruiz-Villanueva et al.,
submitted).
Flume experiments have been used here to test and validate the correct performance
of the numerical model. Analyses of behaviour, log patterns, distance travelled or jam
formation are not included .
The first challenge was to reproduce the hydrodynamics of the flume accurately,
because the geometries used produced an important two-dimensional flow with re-
circulation zones and a slight 3D component. Although the model was well-calibrated
deviations were observed. These deviations between the simulated hydrodynamics of the
channel and the numerical model validation results condition the reproduction of log
transport. There are several reasons for these differences between the observed and the
simulated. One is that the 3D effect cannot be reproduced by a two-dimensional model.
Another reason is that collecting data with Vectrino, attempted near the surface (since logs
mainly move with hydrodynamic surface conditions) meant the equipment had to be
submerged a few centimetres in the flow. The highest errors were observed in the flume
outlet where there is a sharp crested weir. The lower accuracy could be because of the
limitations of the 2D model when simulating a 3D flow such as the one that occurs near
the weir. However, good correlation was found for each geometry (correlation coefficient
>0.8).
Iber is a well known two-dimensional model for simulating turbulent free surface
unsteady flow and environmental processes in river hydraulics. Since this model has been
extensively used to simulate both rivers or physical models (Corestein et al., 2010; Bladé et
al., 2012) the application of the new module (Woody Iber) in real rivers or streams seems
to be straightforward.
The flume experiments were not designed as a scale model of any real reach or
prototype. Nevertheless, if the experiments were to be extrapolated to a real stream using
Froude similarity, if λ is the geometric scale, the scales to be used for time, velocity and
discharge would be: λt=λ1/2, λv=λ1/2, λQ=λ1/2.
For example, where λ=20, the flume would be a physical model of a reach 12 m wide
and 100 m long , discharge would be around 32 m3 s-1, with maximum velocities around 3
m s-1, and logs ranging from 2 to 10 m length and from 0.2 to 0.36 m in diameter, which is
reliable.
Some trials have been run for different cases (Fig. 18 shows one example), the model
works, and the results are reasonably credible; however, it is very difficult to validate this
type of model in real cases.
In the simulation shown in Fig.18, log interactions, logs entrapped in the banks and
jam forming could be observed, as well as log transport in the flow. At present, there is no
detailed field data regarding LWD available to the authors to validate the method with
field data, but work towards this is underway in a study area where a well-known flood
event with LWD loading took place (Ruiz-Villanueva et al., 2012).
Fig. 18: LWD model transport simulation in Fluviá River (Catalonia). Water depth is
shown by colours and logs by brown lines.
Nevertheless, in real rivers, the important process of sediment transport has also to be
taken into account. The important consequences of woody debris in fluvial corridors on
sediment transport and geomorphic effects has been studied before (Comiti et al., 2006
and 2007; Abbe and Montgomery, 2003; Wohl, 2011; Andreoli et al., 2007; Gomi et al.,
2004). One of the main effects of LWD on sediment dynamics is that wood may force the
formation of bars upstream of LW dams, and an impoundment may be formed where
deposition takes place and the bed slope is lower than the channel average. Wood also
influences channel width, narrowing some sections or causing localized erosion. In
addition, wood can force spatial variations in shear stress and stream power that alter bed
and bank topography (Buffington and Montgomery, 1999).
As described in the first part of this paper, Iber has a computational module to
simulate sediment transport. The sediment transport module solves the non-cohesive
sediment non-stationary transport equations including the bedload transport and the
suspended sediment transport. However, the influence of LWD presence on sediment
storage and transport has not been tested yet with the present model, and this important
issue will be investigated in the future.
In terms of this process in relation to flood hazard analysis, the main problem is the
clogging of critical sections such as bridges (Mazzorana et al., 2011a). The simulation of
LWD transport at critical stream geometry configurations may therefore be an important
contribution. In terms of drift entrainment and transport, the effects of blockage due to
LWD accumulations at river bridges may be studied from a perspective of either flood
hazard (backwater effects; see Mazzorana et al., 2011c) or infrastructure evaluation
(failure, pier scour, or sizing; see Schmocker and Hager, 2010).
This model may also be useful in other analytical approaches. The ecological
importance of LWD is reflected in zones with decreased flow velocity and increased flow
depth, where habitat suitability for fish species colonization is enhanced (Bocchiola, 2011),
so log deposition areas could be an ecological indicator. LWD has also become a central
theme in river management and restoration (Benda and Sias, 2003).
commercial modelling tools available based on finite volume or finite element methods
that satisfy this requirement.
5. Conclusions
A numerical model has been developed to simulate the hydrodynamics and other
fluvial processes (turbulence) with wood transport. This tool has been implemented as a
new computational module into the Iber two-dimensional hydraulic simulation software.
The new module calculates the position and velocity of logs with different shapes using
kinematic and dynamic approaches, both based on the balance of forces involved in wood
movement. These approaches account for the interactions between the logs themselves and
between the logs and the channel configuration. Prototype-scale flume experiments
validated the model’s capacity to reproduce accurately floating log movement in a straight
channel with constraints.
This model as it is at present is a first attempt to incorporate wood transport into the
two-dimensional hydrodynamic simulation of rivers.
Acknowledgements
This work was funded by CICYT MAS Dendro-Avenidas project and the Geological Survey of
Spain (IGME). The lead author acknowledges the grant obtained from the Spanish Ministry of Science and
Innovation, for a 3-month research period at the Flumen Research Institute at University of Catalunya
(UPC). The authors are grateful to Georgina Corestein and Hans Sánchez for their help with
programming, to Soledad Estrella for her help with the laboratory experiments, to Ceferino Robledo who
was responsible for recording the flume experiments and to Joaquim Rabada for his help with IT.
REFERENCES
Abbe, T., Montgomery, R., 2003. Patterns and processes of wood debris accumulation in the Queets river
basin, Washington. Geomorphology, 51(1-3), 81-107.
Andreoli, A., Comiti, F., Lenzi, M. A., 2007. Characteristics, distribution and geomorphic role of large
woody debris in a mountain stream of the Chilean Andes. Earth 1692, 1675-1692.
Ballesteros Cánovas, J.A., Bodoque, J.M., Díez-Herrero, A., Sanchez-Silva, M. & Stoffel, M. 2011b.
Calibration of floodplain roughness and estimation of palaeoflood discharge based on tree-ring
evidence and hydraulic modelling. Journal of Hydrology 403,103-115.
Benda, L. E., Sias, J. C., 2003. A quantitative framework for evaluating the mass balance of in-stream
organic debris. Forest Ecology and Managment, 172, 1-16.
Bladé, E., Cea, L., Corestein, G., Escolano, E., Puertas, J., Vázquez-Cendón, M. E., Dolz, J., and Coll, A.
2012. “Iber - Herramienta de simulación numérica del flujo en ríos.” Revista Internacional de
Métodos Numéricos para Cálculo y Diseño en Ingeniería (in press).
Bocchiola, D., Rulli, M. C., Rosso, R., 2008. A flume experiment on the formation of wood jams in rivers.
Water Resources, 44, 1-17.
Bocchiola, D., 2011. Hydraulic characteristics and habitat suitability in presence of woody debris: A flume
experiment. Advances in Water Resources, 1304-1319.
Bocchiola, D., Catalano, F., Menduni, G., Passoni, G., 2002. An analytical–numerical approach to the
hydraulics of floating debris in river channels. Journal of Hydrology 269, 65-78.
Bocchiola, D., Rulli, M. C., Rosso, R., 2006b. Flume experiments on wood entrainment in rivers. Advances
in Water Resources, 1182-1195.
Bocchiola, D., Rulli, M., Rosso, R., 2006a. Transport of large woody debris in the presence of obstacles.
Geomorphology, 166-178.
Braudrick, C. A., Grant, G. E., Ishikawa, Y., Ikeda, H., 1997. Dynamics of Wood Transport in Streams: A
Flume Experiment. Earth Surface Processes and Landforms, 22(7).
Braudrick, C. A., Grant, G. E., Northwest, P., Forest, U. S., 2001. Transport and deposition of large woody
debris in streams : a flume experiment. Geomorphology, 263-283.
Braudrick, C., Grant, G. E., 2000. When do logs move in rivers? Water Resources Research 36, 571.
Brooks, A. P., Abbe, T., Cohen, T., Marsh, N., Mika, S., Boulton, A., Broderick, T., Borg, D., Rutherfurd, I.,
2006. Design guidelines for the reintroduction of wood into Australian streams, Land & Water
Australia, Canberra.
Buffington, J. M., Montgomery, D. R., 1999. Effects of hydraulic roughness on surface textures of gravel-
bed rivers. Water Resources Research, 35(11), 3507.
Buxton, T. H., 2010. Modeling entrainment of waterlogged large wood in stream channels. Water
Resources Research 46.
Cea, L., 2005. An unstructured finite volume model for unsteady turbulent shallow water flow with wet-
dry fronts: Numerical solver and experimental validation. PhD Dissertation. Universidad de A
Coruña, Departamento de Métodos Matemáticos y de Representación. Programa de Doctorado en
Ingeniería Civil.
Manga, M., Kirchner, J. W., 2000. Stress partitioning in streams by large woody debris. Water Resources
36, 2373-2379.
Mazzorana, B., Comiti, F., Volcan, C., Scherer, C., 2011. Determining flood hazard patterns through a
combined stochastic–deterministic approach. Natural Hazards, 301-316.
Mazzorana, B., Fuchs, S., 2010a. Fuzzy Formative Scenario Analysis for woody material transport related
risks in mountain torrents. Environmental Modelling & Software, 25(10), 1208-1224.
Mazzorana, B., Hübl, J., Zischg, A.M., Largiader, A., 2010b. Modelling woody material transport and
deposition in alpine rivers. Natural Hazards, 425-449.
Mazzorana, B., Zischg, A., Largiader, A., Hübl, J., 2009. Hazard index maps for woody material recruitment
and transport in alpine catchments. Natural Hazards and Earth System Science, 197-209.
Merten, E., Finlay, J., Johnson, L., Newman, R., Stefan, H., Vondracek, B., 2010. Factors influencing wood
mobilization in streams. Water Resources Research 46, 1-13.
Montgomery DR, Piegay H., 2003. Wood in rivers: interactions with channel morphology and processes.
Geomorphology 51, 1-5.
Montgomery, D., 2003. Wood in rivers: interactions with channel morphology and processes.
Geomorphology 51, 1-5.
Moulin, B., Piegay, H., 2004. Characteristics and temporal variability of large woody debris trapped in a
reservoir on the River Rhone (Rhone): implications for river basin management. River Research and
Applications 20, 79-97.
Rastogi, A., Rodi, W., 1978. Predictions of heat and mass transfer in open channels. Journal of Hydraulics
Division 104, 397-420.
Ruiz-Villanueva, V., Bodoque, J.M., Díez-Herrero, A., Eguibar, M.A., Pardo-Igúzquiza, E., 2011.
Reconstruction of an ungauged flash flood event with large wood transport and its influence on
hazard patterns. Hydrological Processes (under review).
Schmocker, L., Hager, W. H., 2010. Drift accumulation at river bridges. River Flow 2010 - Dittrich, Koll,
Aberle & Geisenhainer (eds.), 2010. Bundesanstalt für Wasserbau, ISBN 978-3-939230-00-7.
Swanson, F.J., 2003. Wood in Rivers: a landscape perspective. American Fisheries Society Symposium 37,
299-313.
Thevenet, A., Citterio, A., Piegay, H., 1998. A new methodology for the assessment of large woody debris
accumulations on highly modified rivers (example of two French Piedmont rivers). Regulated
Rivers: Research & Management 14, 467-483.
Vázquez-Cendón, M. E., 1999. Improved Treatment of Source Terms in Upwind Schemes for the Shallow
Water Equations in Channels with Irregular Geometry. Journal of Computational Physics 148, 497-
526.
Vectrino User Guide, 2009. www.nortekusa.com.
Wahl, T.L., 2000. Analyzing ADV Data Using WinADV. Joint Conference on Water Resources
Engineering and Water Resources Planning & Management. July 30, August 2, 2000. Minneapolis,
Minnesota.
Wilcox, A. C., Peckarsky, B. L., Taylor, B. W., Encalada, A. C., 2008. Hydraulic and geomorphic effects on
mayfly drift in high-gradient streams at moderate discharges. Ecohydrology 186, 176- 186.
Wilcox, A. C., Wohl, E. E., 2006. Flow resistance dynamics in step-pool stream channels: 1. Large woody
debris and controls on total resistance. Water Resources Research 42, 1-16.
Wohl, E., 2011. Threshold-induced complex behavior of wood in mountain streams. Geology 39, 587–590.
Wohl, E., Bolton, S., Cadol, D., Comiti, F., Goode, J. R., Mao, L., 2011. A two end-member model of wood
dynamics in headwater neotropical rivers. Journal of Hydrology 1-10.
Wohl, E., Jaeger, K., 2009. A conceptual model for the longitudinal distribution of wood in mountain
streams. Earth, 329-344.
(4.6). 1. Introduction
(4.6). 2. Study area
(4.6). 3. Methodology
3.1. Modelling large wood transport in rivers
3.2. Wood and critical stream configurations:
internal conditions
3.3. Inlet boundary conditions, scenarios and
validation
(4.6). 4. Simulation results and validation
(4.6). 5. Discussion
(4.6). 6. Conclusions
REFERENCE:
Ruiz-Villanueva, V.; Bladé, E.; Díez-Herrero, A.; Bodoque, J.M.; Sánchez-Juni, M. 2D-
hydrodynamic modelling of large wood transport at critical stream geometry
configurations.
Submitted to Earth Surface Processes and Landforms.
ABSTRACT
Large woody material (LW) transported in the flow may be entrapped at critical
stream geometry configurations (e.g. at bridges), and therefore may dramatically increase
the destructive power of floods. This was the case in a mountain catchment where a flood
event with LW loading took place in 1997, and relevant data was available. The aim of this
study was to simulate a bridge clogging process, modelling individual pieces of wood
moving with the water flow and interacting with the bridge. First, the boundary conditions
(inlet discharge and wood) were established, and then a 2D model was developed to
simulate the transport of LW together with the hydrodynamics. A combination of weir
and gate internal conditions was used to represent the bridge geometry, and the model
simulated the way in which the wood’s interactions with it affect the hydrodynamics.
Different scenarios for the wood budget allowed us to study the influence of inlet
boundary conditions in bridge clogging. For the studied event, the scenario which best
reproduced the bridge clogging effect and the flood characteristics was the one in which
60% of total wood entered before the peak discharge. This dropped to 30% at the peak
itself, and finally fell to 10% during the recession curve. In addition, the accumulation
patterns of LW along the reach were compared with photographs, showing that the model
succeeded in predicting the deposition patterns of wood.
1. Introduction
The role of large wood (LW) in rivers has been widely studied. An extensive
literature now exists describing the influence of wood on stream ecology (Gurnell et al.,
2002; Martin ad Benda, 2001; Gippel and White, 2000), since wood provides a habitat for
fishes and riverine species (Jackson and Sturm, 2002) and regulates water flows and
nutrient fluxes (Welty et al., 2002). In addition, it has been demonstrated that LW plays a
key role in the geomorphology of rivers with many implications in erosion and
sedimentation processes, channel morphology, hydraulics, etc. (Montgomery et al., 2003).
Recent research has focused on the mobilization of woody material during floods,
since transported woody material that is entrapped at critical stream geometry
configurations (e.g. at bridges) results in a substantial increase in the destructive power of
floods (Shields and Gippel, 1995; Darby and Thorne, 1995; Dudley et al., 1998; Shields et
al., 2001; Diehl, 1997; Braudrick et al., 1997).
In forested mountain catchments, most trees fall into the stream as a result of a
variety of mechanisms such as mass wasting, channel migration and bank undercutting
(May and Gresswell, 2003; Swanson, 2003), stochastic mechanisms such as windthrow and
fire (Benda and Sias, 2003; Rosso et al., 2007) or simple tree mortality (Benda et al., 2003).
Since these geomorphological process sequences often occur during major floods
(Nakamura et al., 2000), when the number of wood pieces likely to be transported
increases, the recruitment and dynamics of woody material have been studied from the
perspective of its potential responsibility in increasing flood hazard impacts (see examples
in Comiti et al., 2008; Mao and Comiti 2010; Rickenmann and Koschni, 2010;
Mazzorana et al., 2010; Ruiz-Villanueva et al., 2012). But this also makes field
measurements difficult. As a result, there are few direct observations and/or measurements
of conditions of wood entrainment and transport (see one pioneer study in MacVicar and
Piegay, 2012). Physical models and flume experiments have been used to overcome these
constraints, and this has contributed greatly to our present knowledge of LW transport
(Braudrick and Grant, 2000; Braudrick et al., 2001; Haga, 2002; Bocchiola et al., 2002,
Buxton, 2010; Wohl et al, 2011).
Since the seventies, several published studies have described instances in which LW
contributed to bridge failure or damage in the U.S., New Zealand and Canada (Brice et al.,
1978a and 1978b; Harik et al., 1990; Dongol, 1989; Lyn et al., 2007). Flood damage
caused by LW, although not specifically bridge damage, is also described in Europe
(Comiti et al., 2006; Comiti et al. 2008a and 2008b; Mao and Comiti 2010). However,
despite the effects that LW may have on floods, it has very often been overlooked and just a
few studies propose methods to incorporate this process in flood hazard or risk analysis
(Merten et al., 2010; Mazzorana et al., 2010a and 2010b, Mazzorana et al., 2011; Ruiz-
Villanueva et al., 2012). Moreover, in those cases when a bridge is included in the research,
the approach has been the modification of the bridge geometry in order to simulate the
effect of the clogged wood (see Mazzorana et al., 2011). However, there has been no
general model for wood movement in streams so far. A numerical model was developed by
Ruiz-Villanueva et al. (under review) to simulate the hydrodynamics of rivers, including
turbulence, together with woody debris transport. It was added as a new module (Woody
Iber) to Iber, a two-dimensional hydrodynamic software application (www.iberaula.es)
based on the finite volume method (Corestein et al., 2010; Bladé et al., 2012). This model
succeeded in introducing woody debris transport into the 2D hydrodynamic simulation of
rivers.
The main aim of this study was to test the proper operation of the application and its
ability to provide a suitable model for a critical stream configuration, such as a bridge.
Therefore, the purpose is to simulate a bridge clogging process caused by LW transport by
modelling individual wood pieces in the simulation. The interaction between logs and
bridge is then simulated and the effect on the hydrodynamics is analysed.
The area selected for the study is a well-known mountain catchment where a flood
event with LW loading took place in 1997 (see details in Bodoque et al., 2011 and Ruiz-
Villanueva et al., 2012) and where the necessary data is available. Since the event evolved
from landslide to flash flood, the LW delivery was episodic (Wohl et al., 2011) and the
transport was congested (Braudrick et al., 1997).
2. Study area
The study site is a reach of the Arroyo Cabrera stream, a tributary of the Alberche
River in the Tagus River Basin, on the northern slopes of the Sierra de Gredos, in the
Spanish Central System (Figure 1). This small ungauged forested mountain catchment
covers an area of over 15.5 km2, the maximum difference in height within the watershed is
1188 m, and the main channel is 5500 m long, with an average slope of 21.6%. The geology
consists mainly of Upper Palaeozoic granitoid covered by superficial Quaternary
formations of conglomerates, gravels, sands and silts.
The local forest stand is formed predominantly by Pinus pinaster Ait., Pinus
sylvestris L. and Q. pyrenaica Willd. In addition, riparian broadleaved species A. glutinosa
(L.) and F. angustifolia (Vhal.) can be found on both river banks.
Torrential rainfall events usually occur in autumn and winter, resulting in abundant
surface runoff, sediment mobilization and related flash-flood events. The most important
event identified to date was a 1997 flash flood. The chronological description of the 1997
process can be summed up as follows: a complex sediment-laden flow was triggered by a
shallow landslide (Ruiz-Villanueva et al., 2011); this evolved into a hyper-concentrated
flow (Bodoque et al., 2011). 3.2 km downslope most of the sediment transport was
partially deposited and an avulsion took place (Díez-Herrero, 2003). This process led to a
major change in the drainage network pattern, and remobilized large quantities of
sediment, damaging the vegetation located within the stream and on the banks and
resulting in a substantial wood recruitment process.
Upstream from the point where it flows into the Alberche River there is a critical
section in the outlet of the catchment; here there is a bridge which is the biggest obstacle in
the stream and where large deposits of wood and boulders were observed after the 1997
event (Figure 1). The evidence from observations, images and other indicators (e.g. fine
sand and coarse wood deposits found on the bridge deck) show that the bridge was flooded
due to the clogged wood. Indirect methods together with hydraulic simulation and
rainfall-runoff modelling have been applied to obtain the main flash flood parameters.
Thus, peak discharge (123 ± 18 m3 s-1), simulated hydrograph, water level (~7 m at the
bridge section) and blockage ratio (48 ± 8% of the bridge section) were estimated for this
event in a previous study (for details see Ruiz-Villanueva et al., 2012).
Figure 1: Location of the study area: Arroyo Cabrera stream basin with vegetation density
and species, and location and picture of the bridge affected by wood clogging.
3. Methodology
The methodology applied in this study is divided into 3 main stages (Figure 2): (1)
estimation of the potentially recruitable large wood at basin scale (analysing the
recruitment areas caused by an avulsion phenomenon, fluvial transport and bank erosion)
that enables us to set inlet boundary conditions for large wood (Qw); (2) setting initial and
boundary hydrodynamic conditions (Qi); (3) the flood event with large wood transport
modelling and analysis.
Figure 2: Proposed methodology scheme for the study and simulation of a flood event with
large wood transport. Qi and Qw are the inlet flow discharge and inlet wood budget
respectively.
The sections below describe these stages as follows: (i) first, a brief description of the
model developed for simulating the wood transport in rivers is provided; (ii) a detailed
description of the methods for simulating the interaction between wood and critical
sections such as bridges (weirs and gates) is given; (iii) the inlet boundary conditions
(discharge and wood budget) for the studied flood event are estimated.
The transport of wood in the stream has been simulated using a two-dimensional
hydrodynamic model recently developed for this purpose. The model is described in detail
in Ruiz-Villanueva et al. (under review) and therefore only a brief description is provided
here.
To deal with hydrodynamics and turbulence, the finite volume method with a
second order Roe Scheme is used. This method is especially suitable for flows in mountain
(torrential) rivers, where shocks and discontinuities can occur and flow hydrographs are
very sharp. The method is conservative, even when wetting and drying processes take place.
which is equal to the normal force acting on the log times the coefficient of friction
between the wood and the bed; and the drag force, also acting in the flow direction, which
is the downstream drag exerted on the log by the water in motion.
The movement of wood logs includes three possible transport regimes (floating,
rolling or sliding) and both translation and rotation, due to the fact that one end of the
piece of wood is moving faster than the other end (based on flow velocity field) and causes
the piece to rotate towards a more flow-parallel orientation.
Interactions between logs and the channel configuration and among logs themselves
are also taken into account in the model. Therefore, log velocity and trajectory may change
due to contact with the banks or with other logs. If one piece of wood is floating in the
flow and meets another piece (floating or resting) then the two may collide and continue
moving with a different velocity. Moreover, when a piece of wood reaches the bank, it can
be entrapped and the driving forces decrease due to the reduction of the submerged area,
but the resisting forces are still active around the log and therefore the initial motion
condition is re-calculated.
Hydraulic structures like gates, weirs or bridges, which change the conditions of the
system and, thus, cannot be represented by the Saint Venant equations are usually treated
as internal conditions. The internal conditions available in the Iber model (similar to most
2D hydraulic models) are: (i) flow below a gate; (ii) flow over a weir; (iii) the weir-gate
combination; and (iv) local loss determined by a loss coefficient. In the case of gates, the
standard equation of the flow under a gate, which can be free or submerged, is used. The
data to enter are the discharge coefficient, the gate’s bottom level, the gate opening height
and its width. To simulate weirs, the rectangular weir discharge equation is used, which can
also represent free flow and submerged flow. The data to input are the weir’s crest
elevation, the discharge coefficient and the weir’s length. In this study the condition used
to represent the bridge is a combination of the two previous ones, those for gates and weirs,
so all the above-mentioned parameters must be used. The total discharged flow is taken as
the sum below the gate and the flow above the weir.
If a piece of wood transported in the flow interacts with a gate or weir then the log
may go below the gate or over the weir, depending on the gate aperture and width, or the
weir length, the water level and the wood diameter (Figure 4). If the water level (h1) and
wood diameter are higher than the height of the gate (Zg) or lower than the height of the
weir (Zw), then the wood log will collide with the obstacle, in our case with the deck of the
bridge. In this case, from the point of view of wood transport the obstacle is treated as a
material wall, and logs interacting with it will bounce back and get stuck. The drag force on
trapped logs represents an opposite action to water flow, producing a rise in water levels
and a decrease in velocity.
Figure 4: Scheme of the gate and weir internal contour conditions and wood.
The inlet discharge (a peak discharge of approximately 120 m3 s-1) required for the
hydraulic model was a reconstructed event hydrograph using uncertainty intervals
obtained by hydrological simulation in a previous study (Ruiz-Villanueva et al., 2012).
Moreover, the wood boundary conditions were assigned to the whole domain. Based
on the knowledge of the fluvial corridor, the riparian vegetation, and the wood availability,
ranges were established for the diameters, density and lengths of wood. The stochastic
variation of these parameters, together with position and angle with respect to the flow,
were then used to characterize each piece of wood entering the simulation.
Post event surveys and previous studies allowed us to reconstruct the area affected by
the flood and the avulsion where most of the trees were recruited. Then, using the available
information on the forest stand (The Forest Map of Spain and National Forest Inventory
produced by the Ministry of the Environment, Spain; MFE, 2011), it was possible to
estimate recruited wood volumes (number of trees; WVr) following the methodology
proposed by Ruiz-Villanueva et al. (in press). This method allows us to establish the
boundaries for the recruitment source areas and estimate the wood volume potentially
delivered in this event. The model uses up to three different species of forest trees, each
with their development stage, percentage occupation (percentage of the total forest
occupied by this species), canopy cover for total trees (percentage of ground covered by the
horizontal projection of the canopy), and DBH (diameter at breast height), among others.
The National Forest Map and Inventory provide the required data but some
simplification has to be assumed. Data is provided for the three main species in any given
area and the total canopy cover (Ci) as a percentage of the total area covered by forest. The
inventory for the province of Ávila contains estimates of tree density (expressed as number
of trees per area) for each species for the whole forested territory. This latter density is
called here relative density per species (RDSPi) and is used together with species
occupation and canopy cover to estimate final volumes in a given area. Ai is the
contributing area defined for a specific recruitment process and with an established
probability of occurrence or severity:
PRWSPi=Ai•Ci•RDSPi [1]
To take into account the defined vegetation resistance and the severity of the
potential recruitment mechanism, a volume correction factor (Fc) was defined. This is
equivalent to a recruitment probability and can be 1, 0.5 or 0.1. This factor therefore
reduces the total volume of potential recruitable wood in those areas where susceptibility
to the process is lowest and/or vegetation resistance is highest.
Once we have the number of trees we can estimate the number of logs (pieces of
woody debris) as:
where k is a coefficient.
For example according to Benda and Sias (2003), a 40 m-tall tree is equivalent to 5
pieces of 8 m-long woody debris, so in this case k=5.
According to field observations and photographs, most of the trees were transported
short distances, resulting in few broken trees. However, due to the uncertainty of this rule
we estimated the number of trees (k=1).
A series of scenarios were then simulated assuming that the peak in wood transport
occurred prior to the peak discharge and varying the amount and temporal distribution of
the wood budget. The scenarios (Figure 3) were established based on previous knowledge
of the event studied and the observations made by MacVicar and Piegay (2012). These
authors recorded the transport of wood during floods in a river and they observed that the
peak in wood occurs prior to the peak discharge and that the transport rates are
approximately four times higher on the rising limb of the hydrograph than on the falling
limb. Therefore, the proposed scenarios are:
Scenario 1: 100% of total recruited wood enters just before and during the
peak discharge and wood transport occurs from the rising limb and during
the peak (~ 1 h).
Scenario 2: 60% enters before the peak discharge (~ 0.5 h), during the peak
this amount is reduced to 30% (~ 0.5 h), and finally during the recession
curve 10% of the total recruited wood is transported (~ 0.5 h).
Scenario 3: the total of recruited trees enter continuously during the whole
event until the middle of the recession curve (~ 3 h).
VIRGINIA RUIZ-VILLANUEVA
A 312 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
After analysing the contributing areas defined for this event, the vegetation resistance
and the forest density (3 main species, canopy cover and relative density) theoretical
predictions were obtained of the number of trees potentially recruited (WVr) for the reach
studied.
Length (m) 7 15 2
The simulated wood budget for the three scenarios was 213, 180 and 137 logs
respectively. According to the pictures taken some days after the event, the model’s results
satisfactorily reproduce the wood deposit location and the bridge clogging for the three
scenarios (Figure 4).
The main deposit areas or the areas prone to form wood jams are the two bridge
abutments (top right picture in fig.4), and along the left bank of the stream,
particularly the areas shown in the pictures at top and bottom left and bottom right in
fig.4. There is also the left bank downstream of the bridge, although this reach is
outside the scope of this study.
Figure 5: Maximum water depth for the simulated hydrograph without wood, and for the
three scenarios (SC1, SC2, and SC3). The black line represents the bridge, and the
grey polygon the small white building with the high water mark (see fig.4).
The maximum water depth for the simulated hydrograph without wood is around
4.7 metres in the bridge section; this value is in the same range as that obtained in the
previous study, which was around 4.5 m. Whereas the backwater effect due to bridge
clogging increases this level to 7.1 m according to previous work, 7.3 m was the result
obtained from scenario 2 of this reconstruction. In scenario 1 this level could increase to
more than 8 metres (1 metre above the bridge deck). Scenario 3 is an intermediate case,
where the bridge is less blocked and the water level does not reach the bridge deck, rising to
6.8 metres.
The other HWM used for the reconstruction was located in a small building placed
on the right bank. This mark was 2.9 metres; once again, scenario 2 gives the best
approximation with a value of 2.8 metres.
Figure 6: Maximum flow velocity for the simulated hydrograph without wood, and for the
three scenarios (SC1, SC2, and SC3). The black line represents the bridge, and the
grey polygon the small white building with the high water mark (see fig.4).
The main velocity estimated for this event was around 3.5 m s-1 to 4 m s-1. Although
these values are exceeded locally, the main flow velocity agrees with these previous
estimates. In addition, as the bridge is increasingly obstructed, it can be observed how the
flow velocity decreases upstream.
The main effect of bridge clogging due to wood transport on flow transport capacity
is the increase of this parameter on the right bank, whereas it decreases on the left bank.
This agrees with the deposition patterns observed in the post event pictures. The left side is
where most of the wood is deposited (see left hand pictures in Fig.4), but on the right side
bigger boulders are observed as well as erosive marks (see pictures on right in Fig.4).
Figure 7: Maximum transport capacity for the simulated hydrograph without wood, and
for the three scenarios (SC1, SC2, and SC3). The black line represents the bridge, and the
grey polygon the small white building with the high water mark (see fig.4).
5. Discussion
A bridge clogging process due to LW transport during a flash flood event was
reproduced by modelling individual wood pieces moving in the flow, using a two-
dimensional hydrodynamic model. This study represents the first attempt to simulate the
interaction between individual wooden logs and obstacles together with the
hydrodynamics in rivers.
The main simplification assumed by the model was the shape of the logs as cylinders,
disregarding the effect of branches or roots. This geometry provides a good approximation
for broadleaved trees, which has been used by several researchers (see Bocchiola et al., 2008;
Braudrick et al., 1997; Buxton, 2010; Mazzorana et al., 2011), and may provide a good
picture of non-rooted and defoliated logs often occurring in rivers as a result of fluvial
transport, wood harvesting and forest fires (Bocchiola, 2011). Although, according to the
findings of Schmocker and Hager (2010), the probability of blocking may increase with
the presence of roots, in the case of conifers, for example, in their experiments the bridge
geometry used was a truss bridge deck, which differs from the one simulated in this study.
The same authors have pointed out that, once a piece of wood is entrapped in the obstacle,
the likelihood of blocking increases; this is taken into account in our model, since log
interactions are included.
In the estimation of the wood volumes recruited (WVr), we first delineated the area
affected by the flood and avulsion process, based on previous work and field observations.
This is an important issue, and could be a source of uncertainty if this area is unknown, or
we are simulating future or design scenarios and not past events. The forest vegetation
present in the affected area was then analysed. The main recruitment processes were fluvial
transport and bank erosion. The probability of a tree being recruited was quantified based
on the work done by Ruiz-Villanueva et al. (2012) and using the volume correction factor.
This factor reduces the total volume of potential recruitable wood in those areas where
vegetation resistance is highest. Vegetation resistance was established based on the tree
species and stage of the forest, and on previous studies by Hutte (1968), Stumbles (1968),
Naka (1982) and Abernethy and Rutherford (2001). This concept may be equivalent to
the structural classification of forested areas made by Blaschke et al. (2004) and the woody
debris availability indicator used by Mazzorana et al. (2011). The volume correction factor
can be a source of subjectivity, and the modification of this coefficient could change the
final results. To limit this effect a 25% error margin was assumed and a range of maximum
and minimum WVr was provided.
The establishment of different scenarios for the wood budget allowed us to study the
influence of the inlet boundary conditions in the wood clogging of bridges. These scenarios
were established based on previous knowledge of the area and the event studied, and on the
findings provided by MacVicar and Piegay (2012). As expected, steady transport is
associated with a lower probability of blocking, whereas congested and episodic transport
resulted in the highest blocking rate. Our results agree with the observations made by these
authors: the wood budget of scenario 2, which best reproduces the 1997 event, was
distributed along the hydrograph (60, 30 and 10%) starting just before the peak of the
hydrograph.
On the other hand, sediment transport took place, as can be observed in the pictures,
and this should be taken into account, because it may also influence the bridge clogging
process and demonstrates the variation in transport capacity. However, this study focused
on the reconstruction and simulation of wood transport, so further analysis should be
carried out to better incorporate both wood and sediment transport processes.
The simulation results showed that the model successfully reproduces the areas
where the wood will be deposited and will form jams. This could be of interest not only for
flood hazard evaluation, but also for fluvial dynamics, fluvial restoration or ecological
analysis. The proposed methodology can thus also be used for river and forest restoration
and management (Hilderbrand et al., 1998). Knowing the spatial patterns of LWD
recruitment can provide a watershed context for understanding the geomorphic and
ecological processes associated with LW (Martin and Benda, 2001). Knowledge of the
wood deposit areas may help technicians to design more effective countermeasures, such as
filter dams and/or nets, to prevent the transport of wood downstream to critical channel
cross-sections, while allowing the growth of ecologically important riparian buffers
(Comiti et al., 2006). Since pools and other geomorphic elements are closely associated
with LW dams (Gurnell and Sweet, 1998; Montgomery and Buffington, 2003), the
reconstruction of the wood deposition areas could be used to predict potential changes in
geomorphic patterns. Furthermore, the ecological importance of wood has been
demonstrated (Gurnell et al., 2002), and this model could also be applied in ecological
studies, since LW deposits provide a nursery habitat, protection, and are a source of
nutrients for various organisms (Triska and Cromack, 1980; Harmon et al., 1986; Sedell et
al., 1988).
6. Conclusions
This work presents the reconstruction of a bridge clogging process due to wood
transport during a flash flood. This obstruction increased the water level and the flooded
area upstream of the bridge and, therefore, magnified the flood impacts. The
reconstruction was carried out by modelling the interaction between individual woody logs
and obstacles, together with the hydrodynamics. Previous available data were used to
determine the flood parameters and validate the model’s results. The influence of the wood
budget and the wood transport regimes were also investigated by means of scenarios.
Results showed how this model succeeded in reproducing the flash flood event and the
bridge obstruction due to wood transport, as well as the areas prone to wood deposition.
Therefore, this model represents a powerful tool for flood hazard analysis but could
also be useful in forest restoration and the management of geomorphic and ecological
processes associated with LW.
Acknowledgments
This work was funded by CICYT MAS Dendro-Avenidas project (CGL 2010-19274) and the
Geological Survey of Spain (IGME). Special mention to Ignacio Gutiérrez for his assistant with the
topographical survey.
REFERENCES
Abernethy, B., Rutherfurd, I.D. 2001. The distribution and strength of riparian tree roots in relation to
riverbank reinforcement. Hydrological Processes. 15: 63-79.
Benda, L. E., Sias, J. C. 2003. A quantitative framework for evaluating the mass balance of in-stream organic
debris. Forest ecology and management 172: 1-16.
Benda, L.; Miller, D.; Sias, J.; Martin, D.; Bilby, R.; Veldhuisen, C.; Dunne T. 2003. Wood recruitment
processes and wood budgeting. Ecology and Management of Wood in World Rivers. Volume: 37,
Pages: 49-73.
Bladé, E., Cea, L., Corestein, G., Escolano, E., Puertas, J., Vázquez-Cendón, M. E., Dolz, J., and Coll, A.
2012. “Iber - Herramienta de simulación numérica del flujo en ríos.” Revista Internacional de
Métodos Numéricos para Cálculo y Diseño en Ingeniería (in press).
Bocchiola, D., Catalano, F., Menduni, G., & Passoni, G. (2002). An analytical–numerical approach to the
hydraulics of floating debris in river channels. Journal of Hydrology, 269(1-2), 65–78.
doi:10.1016/S0022-1694(02)00195-6
Bocchiola, D., Rulli, M. C., Rosso, R., 2008. A flume experiment on the formation of wood jams in rivers.
Water Resources, 44, 1-17.
Bodoque JM, Eguibar MA, Díez-Herrero A, Gutiérrez-Pérez I, Ruiz-Villanueva V. 2011. Can the discharge
of a past hyperconcentrated flow be estimated based on palaeoflood evidence? Water Resources
Research. Washington.
Braudrick, C. A., Grant, G. E., Ishikawa, Y., Ikeda, H., 1997. Dynamics of Wood Transport in Streams: A
Flume Experiment. Earth Surface Processes and Landforms, 22(7).
Braudrick, C. A., Grant, G. E., Northwest, P., Forest, U. S., 2001. Transport and deposition of large woody
debris in streams: a flume experiment. Geomorphology, 263-283.
Braudrick, C., Grant, G. E., 2000. When do logs move in rivers? Water Resources Research 36, 571.
Brice, J.C., Blodgett, J.C., Carpenter, P.J., Cook, M.F., Craig, G.S., Jr., Eckhardt, D.A., Hines, M.S.,
Lindskov, K.L., Moore, D.O., Parker, R.S., Scott, A.G., and Wilson, K.V., 1978a, Countermeasures
for hydraulic problems at bridges, Volume 1 - analysis and assessment: Washington, D.C., Federal
Highway Administration Report No. FHWA-RD-78-162, 169 p. [Available from the National
Technical Information Service, Springfield, VA 22161 as PB-297 132.]
Brice, J.C., Blodgett, J.C., Carpenter, P.J., Cook, M.F., Craig, G.S., Jr., Eckhardt, D.A., Hines, M.S.,
Lindskov, K.L., Moore, D.O., Parker, R.S., Scott, A.G., and Wilson, K.V., 1978b, Countermeasures
for hydraulic problems at bridges, Volume 2 - case histories for sites 1-283: Washington D.C.,
Federal Highway Administration Report No. FHWA-RD-78-163, 542 p. [Available from the
National Technical Information Service, Springfield, VA 22161 as PB-297 685.]
Buxton, T. H., 2010. Modeling entrainment of waterlogged large wood in stream channels. Water
Resources Research 46.
Chang, F.F.M., and Shen, H.W., 1979, Debris problems in the river environment: Federal Highway
Administration Report No. FHWA-RD-79-62, 67 p
Chang, F.F.M., 1973, A statistical summary of the cause and cost of bridge failures: Federal Highway
Administration, Office of Research and Development, Report No. FHWA-RD-75-87, 42 p.
Comiti, F., Andreoli, A., Mao, L., Lenzi, M. A., 2008a. Wood storage in three mountain streams of the
Southern Andes and its hydro-morphological effects. Earth Surface Processes and Landforms 33,
242-262.
Comiti, F., Mao, L., Preciso, E., Picco, L., Marchi, L., Borga, M., 2008b. Large wood and flash floods:
evidence from the 2007 event in the Davča basin (Slovenia). Monitoring, Simulation, Prevention
and Remediation of Dense Debris Flows II, 60, 173-182. Southampton, UK: WIT Press.
Comiti, F., Andreoli, A., Lenzi, M., Mao, L., 2006. Spatial density and characteristics of woody debris in
five mountain rivers of the Dolomites (Italian Alps). Geomorphology 78, 44-63.
Corestein, G., Bladé, E., Cea, L., Lara, Á., Escolano, E., 2010. Iber, a river dynamics simulation tool. GiD, 1-
4.
Diehl TH. 1997. Potential Drift Accumulation at Bridges. Publication No. FHWA-RD-97-
028.U.S.Department of Transportation, Federal Highway Administration Research and
Development, Turner-Fair- bank Highway Research Center, Virginia. On the web at: http://
www.tn.water.usgs.gov/ pubs/FHWA-RD-97-028/drfront1.htm
Díez-Herrero A. 2003. Geomorfología e Hidrología fluvial del río Alberche. Modelos y SIG para la gestión
de riberas. Serie Tesis Doctorales. Publicaciones del Instituto Geológico y Minero de España
(Ministerio de Ciencia y Tecnología, Madrid. 587 pp.
Dongol, D.M., 1989, Effect of debris rafting on local scour at bridge piers: Auckland, University of
Auckland Department of Civil Engineering, School of Engineering Report No. 473, 100 p.
Dudley, S. J., Fischenich, J. C., & Abt, S. R. (1998). Effect of Woody Debris entrapment on Flow
Resistance. Journal of the American Water Resources Association, 34(5), 1189–1197.
doi:10.1111/j.1752-1688.1998.tb04164.x
Gippel, C. J., 1995. Environmental Hydraulics of Large Woody Debris in Streams and Rivers. Journal of
Environmental Engineering, 121(5), 388.
Gurnell, A. M, Piegay, H., Swanson, F. J, Gregory, S. V 2002; Large wood and fluvial processes. Freshwater
Biology 47, 601–619.
Haga, H., 2002. Transport and retention of coarse woody debris in mountain streams: An in situ field
experiment of log transport and a field survey of coarse woody debris distribution. Water Resources
Research 38.
Harik, I.E., Shaaban, A.M., Gesund, H., Valli, G.Y.S., and Wang, S.T., 1990, Bridge failures, 1951-1988:
Journal of Performance of Constructed Facilities, v. 4, no. 4, p. 272-277.
Hutte, P., 1968. Experiments on windflow and wind damage in Germany: Site and susceptibility of spruce
forests to storm damage. Forestry 41: 20-26.
Jackson, C.R. and C.A. Sturm, 2002. Woody Debris and Channel Morphology in First- and Second-Order
Forested Channels in Washington’s Coast Ranges. Water Resources Research 38:1177-1190.
Lyn D., Cooper T., Condon D., Gan L., 2007. Factors in debris accumulation at bridge piers, Washington,
US Department of Transportation, Federal Highway Administration Research and Development,
Turner-Fairbank Highway Research Center.
MacVicar, B., Piégay, H. 2012. Implementation and validation of video monitoring for wood budgeting in
a wandering piedmont river, the Ain River (France). Earth Surface Processes and Landforms, (519),
n/a–n/a. doi:10.1002/esp.3240
Mao, L., Comiti, F., Lenzi, M. A. 2010. Bedload Dynamics in Steep Mountain Rivers : Insights from the
Rio Cordon Experimental Station (Italian Alps). Geological Survey Scientific Investigations Report
2010-5091.
Martin, D.J., Benda, L.E. 2001. Patterns of instream wood recruitment and transport at the watershed
scale. Transactions of the American Fisheries Society 130: 940-958.
May, C.L., Gresswell, R.E. 2003. Large wood recruitment and redistribution in headwater streams in the
southern Oregon Coast Range, U.S.A. Canadian Journal of Forest Research 33: 1352.
Mazzorana, B., Comiti, F., Volcan, C., Scherer, C., 2011. Determining flood hazard patterns through a
combined stochastic–deterministic approach. Natural Hazards, 301-316.
Mazzorana, B., Fuchs, S., 2010a. Fuzzy Formative Scenario Analysis for woody material transport related
risks in mountain torrents. Environmental Modelling & Software, 25(10), 1208-1224.
Mazzorana, B., Hübl, J., Zischg, A.M., Largiader, A., 2010b. Modelling woody material transport and
deposition in alpine rivers. Natural Hazards, 425-449.
Mazzorana, B., Zischg, A., Largiader, A., Hübl, J., 2009. Hazard index maps for woody material recruitment
and transport in alpine catchments. Natural Hazards and Earth System Science, 197-209.
Merten, E., Finlay, J., Johnson, L., Newman, R., Stefan, H., Vondracek, B., 2010. Factors influencing wood
mobilization in streams. Water Resources Research 46, 1-13.
MFE ,2011. El Mapa Forestal de España 1:25000 (MFE), Ministerio de Agricultura, Alimentación y Medio
Ambiente (MARM) http://www.magrama.es/es/biodiversidad/temas/montes-y-politica-
forestal/mapa-forestal/mfe_25.aspx
Montgomery, D., 2003. Wood in rivers: interactions with channel morphology and processes.
Geomorphology 51, 1-5.
Naka, K. 1982. Community dynamics of evergreen broadleaf forests in southwestern Japan. I. Wind
damaged trees and canopy gaps in an evergreen oak forest. Journal of Plant Research 95: 385-399.
Nakamura, F., Swanson, F. J., Wondzell, S. M. 2000. Disturbance regimes of stream and riparian systems ? a
disturbance-cascade perspective. Hydrological Processes 14: 2849-2860. doi:10.1002/1099-
1085(200011/12)14:16/17<2849::AID-HYP123>3.0.CO;2-X.
Palacios, D., Marcos, J., Vázquez-Selem, L. 2010. Last glacial maximum and deglaciation of Sierra de
Gredos, Central Iberian Peninsula. Quaternary Internacional 233: 16–26.
Rickenmann, D., Koschni, A. 2010. Sediment loads due to fluvial transport and debris flows during the
2005 flood events in Switzerland. Hydrological Processes 24: 993-1007.
Rosso R, Rulli MC, Bocchiola D. 2007. Transient catchments hydrology after wildfires in a Mediterranean
watershed: runoff, sediment and woody debris. Hydrol Earth Syst Sci. 11: 125–40.
Ruiz-Villanueva, V.,. Bladé-Castellet, E., Sánchez-Juny, M., Martí, B., Bodoque , J.M., Díez Herrero, A. 2D
modelling of woody debris transport in rivers. Under review. Journal of Hydrology.
Ruiz-Villanueva, V., Bodoque, J. M., Díez-Herrero, A., Calvo, C. 2011. Triggering threshold precipitation
and soil hydrological characteristics of shallow landslides in granitic landscapes. Geomorphology
133: 178-189. 10.1016/j.geomorph.2011.05.018.
Ruiz-Villanueva, V., Bodoque, J.M., Díez-Herrero, A., Eguibar, M.A., Pardo-Igúzquiza, E., 2011.
Reconstruction of an ungauged flash flood event with large wood transport and its influence on
hazard patterns. Hydrological Processes. First on line: DOI: 10.1002/hyp.9433.
Shields, F.D; Gippel, C.J. 1995. Prediction of the effects of woody debris removal on flow resistant. Journal
of hydraulic Engineering, 121: 341-354.
Shields, F. D. Cooper, C. M. 2001. Woody vegetation and debris for in-channel sediment control, 15(1),
83–92.
Stumbles, R.E. 1968. How wind influences silviculture and management as a district officer sees it. Forestry
Supplement, 45-50.
Swanson, F.J. 2003. Wood in rivers: a landscape perspective. American Fisheries Society Symposium 37:
299-313.
Wohl, E., Bolton, S., Cadol, D., Comiti, F., Goode, J. R., Mao, L., 2011. A two end-member model of wood
dynamics in headwater neotropical rivers. Journal of Hydrology 1-10.
4.7. INFLUENCE OF LARGE WOOD
TRANSPORT IN FLOOD RISK ASSESSMENT
(4.7). 1. Introduction
(4.7). 2. Study area
(4.7). 3. Methodology
3.1. Extreme value statistics: flood frequency analysis
3.2. 2D- hydrodynamic modelling
3.3. Wood transport recruitment and 2D
hydrodynamic simulation
3.4. Potential flood damage and risk to population
(4.7). 4. Results
4.1. Expected 500-year flood event
4.2. Potential damages with and without wood
transport scenarios
(4.7). 5. Discussion
(4.7). 6. Conclusions
REFERENCE:
Ruiz-Villanueva, V., Bodoque, J.M., Díez-Herrero, A., Bladé, E. 2012. Influence of large
wood transport in flood risk assessment in a mountain village (central Spain).
Submitted to Natural Hazards.
ABSTRACT
An important issue that is not considered in most flood risk assessments is the
transport of solids associated with the flood flow, in this case, wood transport. The
transport and deposition of this wood may be a worse hazard than the flood flow itself. To
analyze this effect, scenario-based 2D hydrodynamic flood modelling was carried out. Since
flood risk assessment has considerable intrinsic uncertainty, probabilistic thinking was
complemented by possibilistic thinking, considering worst-case scenarios. The Bayesian
Markov Chain Monte Carlo (MCMC) procedure was applied to account for uncertainties
in the estimated discharge, and to provide confidence bounds for the quantiles obtained.
The water depth in the test reach was analyzed under uncertainty of the topographical
accuracy and the input Manning roughness coefficient values. This procedure obtained a
probabilistic flood map for a 500-year return period. Then, a series of scenarios was built
based on the wood budget to simulate wood transport and deposition. The potential
damage was estimated as well as a preliminary social vulnerability for all scenarios (with
and without wood transport). The results show that the transport and deposition of wood
during flooding increases potential damage by up to 50% and increases the number of
potentially exposed people by up to 35%. In this paper this phenomenon has been
successfully modelled and included in flood risk assessment.
1. Introduction
The mobilization of woody material in rivers has been considered in the past, but
very few studies have included this phenomenon in flood hazard and risk analysis (see
Mazzorana et al. 2011a, b, c and their references cited). The transport and deposit of wood
can cause a quick succession of backwater effects due to the reduction of cross-sectional
area, accompanied by bed aggradation, channel avulsion and local scouring processes,
which can ultimately lead to embankment/bridge collapse and floodplain inundation
(Diehl, 1997; Comiti et al., 2007; Lynn et al., 2007; Mao and Comiti, 2010). As a result,
flooded areas are likely to be different from those predicted from models where the
presence of wood is not considered (see an example in Ruiz-Villanueva et al., 2012a) and
therefore this may result in the incorrect/uncertain estimation of flood risk.
A wide range of factors can lead to uncertainty in flood risk estimation, but all of
them can be divided into two main types (Tung and Yen, 1993; Paté-Cornell, 1996): (A)
Natural variability (random uncertainty): refers to uncertainties associated with the
inherent randomness of natural processes; (B) Knowledge uncertainty (epistemic
uncertainty): results from incomplete knowledge of the system under consideration and is
related to the ability to understand, measure and describe the system (model, parameter
and data uncertainties).
terms of worst- and best-case assumptions. In this case, although the total variability of the
parameters can be captured, no detailed information on uncertainty is provided. When
input parameters are treated as random variables with known probability distributions
(Apel et al., 2004), the uncertainty is described in the most informative, but also the
strictest way. But if no prior information regarding the parameter probability distribution
is available, an alternative approach could be used instead, based on the imprecise
probability (i.e. fuzzy set approach), in which some probability theory axioms are relaxed
(Fellin et al., 2005).
The aim of this article, therefore, is to analyze how the transport and deposition of
wood influences flood risk estimation in a mountain village. But to take the uncertainties
mentioned above and to obtain the optimal (most likely) flood risk estimate, this
uncertainty was considered at all points in the flood risk analysis.
2. Study area
The study area is located in the eastern massif of the Sierra de Gredos (in the south of
Avila province), the highest section (Almanzor Peak, 2592 m a.s.l.) of the Spanish Central
System, which crosses the Iberian Peninsula SW - NE (Fig.1). These mountains are the
natural division between the river basins of the Tajo (Tagus) to the S. and Duero to the N.
The sector of Sierra de Gredos studied here is on the southern slopes, in the basin of the
Tiétar, a tributary of the Tajo. We studied a reach of the Arenal, a tributary of the Tiétar.
The Arenal flows through the town of Arenas de San Pedro (~6900 inhabitants).
The drainage area at Arenas de San Pedro is 67 km2 and the total length is 12.5 km. The
study reach is 1.2 km long (see Table 1), and the river flows in nearly natural conditions
(since there are no dams upstream), although part of the reach is channelled.
Table 1: Main morphometric characteristics of the Arenal river basin and the study reach.
There is abundant forested mass in the area, and different species are found
depending on altitude. At higher altitudes, where the extreme climatic characteristics
impede the development of tree or shrub species, grassland cover is found (mainly Festuca
sp., Carex sp.). At lower levels, there is a characteristic presence of shrubs (Cytisus sp.) and
stretches of high mountain conifers such as Pinus sylvestris. This species is replaced at
lower altitudes by Pinus pinaster which usually appears with Genista florida, Ilex
aquolifolium, Erica arborea and Sorbus aucuparia. Further down, the forest is composed
mainly of deciduous trees such as Quercus pyrenaica and Quercus ilex. The riparian
vegetation is predominantly Alnus glutinosa and Fraxinus angustifolia. Tree clearance has
been carried out in the past in the study area and is still ongoing.
The climate of the study area is determined by the frequent arrival of Atlantic
depressions from the SW during autumn, winter and spring and by the predominant
Azores anticyclone causing very dry summers (only 10% of annual precipitation). The
yearly precipitation recorded in the area is around 2000 mm (1913 mm at 800 m a.s.l.).
Figure 1: Location of the study site. There are four bridges along the studied reach, and the
Cuevas tributary stream joins the Arenal in mid reach.
Flash floods are common in the Arenal catchment area (Fig. 2) boosted by heavy
rains and the high stream gradient (>0.02 m·m-1), which lead to extremely high discharge
and wood transport rates.
Figure 2: (left) Flash flood in 1999; (right) Post-flood in 2009. Note deposited wood.
3. Meethodo
ology
T
The general methodolo ogical approoach was designed
d to analyze thhe flood rissk in a
mountaain village incorporati
i ing wood ttransport in nto the anaalysis. As m
mentioned in the
introduuction, the risk analyysis is accoompanied by b several sources off uncertaintty. To
consideer these unccertainties in
i the analylysis the meethodology was divideed into four steps
(Fig.3):
Step 1: For the flood freqquency anallysis (FFA), a Bayesiann Markov Chain
Montte Carlo (MMCMC) proocedure (R Reis and Steddinger, 20005) was used d. This
approoach can account for uuncertaintiees in hydrologic extrem mes as it prrovides
estim Gaal et al., 2010).
mates of conffidence bouunds for thee estimated quantiles (G 2
Step 2: The water
w depthh was analyyzed underr uncertainnty of the input
Mannning roughn ness coefficiient values and the acccuracy of thhe digital eleevation
modeel used for 2D
2 hydrodyynamic simu ulation.
Step 3: The rolee of large wwood transp port duringg floods waas included in the
analysis by modeeling differeent scenarioos.
Step 4: Damagge functio ns with predictive
p bounds annd flood hazard
h
thresh
holds were used to est imate the 500-year
5 flo
ood risk in tterms of expected
or potential direct damage ((Messner an
nd Meyer, 2005)
2 and inn terms of danger
d
or hazzards for peeople (HR W
Wallingford
d, 2005).
VIRGINIA RUIZ-VILLANUEVA
A 334 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
P[S´/(Z0-Z
Z´≤0)] : unccertainty inn the stage-ddamage relationship
R=P S´ whhere:
S´ is the ressult.
As several studies haave indicateed, the maajor source of uncertaainty in flo ood risk
assessmment has too be expectted in extreeme values statistics
s Merz et al., 2002). Thee Arenal
(M
river bbasin at Arrenas de San n Pedro is uungauged, so no flow measuremeents were available.
a
We ussed a compiilation of hiistorical disscharges estimated usinng rainfall-rrunoff meth hods and
accounting for uncertaintiess related to antecedentt condition ns and land--use changees. These
historrical floods were reco onstructed from docu umentary sources andd tree-ring analysis
(descrribed in deetail in Ruiiz-Villanuevva et al., su ubmitted). Then, thee Bayesian Markov
Chain n Monte CarloC (MC CMC) proocedure (R Reis and Sttedinger, 22005) was applied
accounting for uncertaintie
u es in discharrge estimattes, and pro
oviding connfidence bou unds for
the obbtained quaantiles (Gaaal et al., 20110; Gaume et al., 20100). Throughh the incorp poration
of (sevveral subsetts of) data on
o historicaal floods (w
with associatted uncertaiainty), the in
nfluence
of varyying approaaches on the estimationn of quantiiles in at-sitee analyses pprovided thee related
variabbility.
VIRGINIA
A RUIZ-VILLANUEEVA 335 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
The hydrodynamic simulation was carried out applying the Iber two-dimensional
hydrodynamic software (Corestein et al., 2010; Bladé, 2012; www.iberaula.es), developed
by CIMNE, GEAMA group (UDC) and Flumen Research Institute (UPC). Iber is a
numerical tool for 2D simulation of turbulent free surface unsteady flow and sediment
transport in watercourses. Iber uses the finite volume method, which is widely used in
computational fluid dynamics (Leveque, 2002; Versteeg and Malalasekera, 2007). To solve
the hydrodynamics and turbulence, the finite volume method with a second order Roe
Scheme (time explicit scheme) is used on non-structured meshes. This method is especially
suitable for flows in mountain rivers, where shocks and discontinuities can occur giving
very sharp hydrographs. The method is conservative even when wetting and drying
processes take place.
The calculation mesh was built with 47451 elements; initial, internal (bridges) and
boundary (inlet and outlet) conditions were defined; up to 23 minutes of computation
time was required to achieve model stability.
The water depth in the studied reach was analyzed under uncertainty of the
topographical accuracy and input Manning roughness coefficient values. The range of
possible roughness unit values was defined following the criteria established by Chow
(1959) and adapted for mountain streams by Jarret (1990). Assuming that there is a
statistical distribution of possible values of the Manning coefficient defined by normal
distribution, the distribution parameters are N(μ,σ). To obtain all (or almost all) possible
values from the distribution in a fuzzy set universe, 4σ was selected as the base interval of a
fuzzy set. In this case, 95% of the value population is in the fuzzy set universe.
In forested mountain catchments the supply of large woody debris (LWD) may be
caused by a variety of mechanisms including landslides, floods and bank erosion (May and
Gresswell, 2003; Swanson, 2003). Potentially recruitable wood volumes were estimated
depending on the recruitment process severity (Ruiz-Villanueva et al., 2012b), but focused
on fluvial transport during floods as the main recruitment process. The transport of wood
in the river was then simulated building reliable scenarios. A numerical model was
developed and integrated in the 2D Iber model to simulate the hydrodynamics and wood
transport (a detailed description can be found in Ruiz-Villanueva et al., (under review);
only a brief description is provided here). This Lagrangian model calculates the position
and velocity of logs with different shapes using kinematic and dynamic approaches, both
based on the balance of forces involved in wood motion. The movement of wood logs
includes both translation and rotation. The approach accounts for the interactions
between the logs themselves and between the logs and the channel configuration. If a piece
of wood transported in the flow interacts with a gate or weir then the log may pass below
the gate or above the weir, depending on the gate opening and width, weir length, water
depth and wood diameter. The flow turbulence may affect this movement, therefore wood
velocity is re-calculated based on the turbulent kinetic energy of the flow. A significant
contribution of this model is that hydrodynamics and wood transport are computed in
two ways; therefore, the hydrodynamics influence the wood transport, but the presence of
wood also influences the hydrodynamics. This latter effect is solved with additional drag,
which is not considered in the 2D-Saint Venant equations for clean water. The drag forces
exerted on the logs located in a finite volume become an additional shear stress source
term, similar to how roughness is usually treated in this kind of schemes.
Initial wood conditions must be assigned to the whole domain. To do this, the
initial position of each log (x, y coordinates of mass centre and angle with respect to the
flow), length, diameter and density are specified.
Inlet boundary conditions for logs are assigned, specifying number of pieces per
minute (or volume) and establishing ranges of maximum and minimum lengths, diameters
and wood density. Stochastic variations of these parameters together with position and
angle must be used to characterize each piece of wood entering the simulation.
We used the available information on the forest stand from the Forest Map of Spain
and National Forest Inventory produced by the Ministry of the Environment (MFE,
2011). This uses up to three different species of forest trees, each with their development
stage, percentage occupation (percentage of the total forest occupied by these species),
canopy cover for total trees (percentage of ground covered by the horizontal projection of
the canopy), and DBH (diameter at breast height).
A series of scenarios was then simulated assuming that peak wood transport
occurred just before and during peak discharge and varying the distribution of the wood
budget (according to the three transport regime definitions proposed by Braudrick et al.,
1997):
Since there was no information on the initial wood conditions no logs were placed in
the reach at the initial time step.
of flood risk management (Messner, FloodSite, 2006). Therefore, the estimation of direct
flood damage was assessed as follows: land use (from the cadastral data) together with
water depth maps were the inputs for the damage function. Since flooding events and their
resulting damage are usually highly spatially diverse, results were presented both as a
monetary figure and as a statistical analysis and on maps to show the spatial distribution of
damages.
All data used for our analysis is publicly available and updated. Most of the data was
obtained from municipal statistics (census and social data, Arenas de San Pedro). Seven
single land-use types were used for elements at risk: residential buildings (single storey,
multi-storey, with and without basement), industrial areas, agricultural areas (mainly fruit
and olive trees), and sports or recreation areas. Since there is a large riverside parking area,
potential damage to vehicles was also taken into account.
4. RESULTS
The compiled discharge data series used in previous work were used as explained for
the frequency analysis. As a first stage, the data was processed as instrumental records in the
MCMC frequency analysis, and then trials were run with some of the events as
instrumental and others as historical, including the estimated uncertainty through an
iterative process. Some results of this iterative process are summarized in Table 2.
Table 2: Arenal river at Arenas de San Pedro : estimated discharge quantiles Q(ML)
corresponding to return period T =500. CI0.05 (CI0.95) is the 5% (95%) confidence limit of
estimates Q(ML).
Note that the existing regional model of maximum discharges for the western section
of the Tajo river basin proposed by Mediero and Jimenez (2007) and CEDEX (2011) gives
a value of 641 m3∙s-1 for this quantile (without any confidence interval). The historical
event estimates for this river revealed that the highest magnitude flood in the 20th century
in this area was in 1936 with a discharge of 509 ± 145 m3∙s-1(Ruiz-Villanueva et al.,
submitted).
For the Cuevas tributary stream, since there is a dam upstream of its confluence with
the Arenal, we assumed the maximum discharge without overflooding to be 205 m3∙s-1.
Manning coefficient values were assigned to each cell of the mesh, ranging from
0.011 - 0.019 m1/2∙ s–1 in the built-up area, from 0.066 - 0.189 m1/2 ∙s–1 for forested areas,
0.06 - 0.14 m1/2 ∙s–1 for vegetated banks, and 0.023 - 0.084 m1/2∙ s–1 in the main channel
where large boulders were present.
The combination of the flood frequency results with the Manning values
distribution allowed us to obtain a probabilistic map for the expected flood-prone area for
the 500-year flood (Fig.4).
Wood transport scenarios were built based on the parameters (discharge and
Manning roughness values) used for the 90% probability inundation map. When wood
enters the simulation, the main effect is the backwater effect upstream of the bridges where
the wood is entrapped and deposited, reducing the cross-sectional area, and as a result the
water depth increases significantly in these areas (Fig.5).
Figure 4: Probabilistic flood map (without wood transport) for 500-year flood.
Figure 5: Water depth (in meters) for the 90% probabilistic flooding map (A) without
wood; (B) scenario 1 of wood transport and deposition; (C) scenario 2; (D) scenario 3.
The financial replacement value for damage estimation was based on average market
prices for this town: residential buildings - 1 200 €∙m-2; industrial buildings- 1 400 €∙m-2;
vehicles -12 000 €; sport and recreation areas - 60 €∙m-2. For residential and industrial uses,
the financial value of structural damage was taken as 15% of the average building price and
the value of contents damage as 20%, based on insurance data and previous studies
(PATRICOVA, 2002; Ballesteros et al., under review; Fig.6).
Figure 6: Examples of the damage functions used in this analysis. (A) Structural
vulnerability (B) contents vulnerability (depth-damage functions) for one-storey
residential buildings with basement.
Seven land-use types were analyzed, with most buildings in type 3: single-storey
without basement; Type 4: multi-storey without basement; and type 9: industrial. These
three types were therefore used for further analysis (Fig.7).
The total estimated expected damage in Arenas de San Pedro for the 500-year flood
(probability equal to 0.002) without wood transport is up to 14 000 € m-2 affected area
(with 85% probability ).
When wood transport is simulated these values change (Fig. 8) to up to 21 000 € m-2.
Figur
ure 8: Total potential
p damage
da in thhe three mai
ain types off buildings
b inn the area based
ba on
simuulations with
thout woodd (light bluee bars) and for
f the threee scenarios oof wood traansport
(wooody brown n bars).
VIRGINIA
A RUIZ-VILLANUEEVA 345 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
(22.53%) are 50 - 70; and 1209 (17.58%) are over 70. 29.15% of the population is
considered as dependant (over 65 or under 18). There are at least 453 illiterate inhabitants
(6.6%); 961 (13.4%) with no formal education; 1613 (23.46%) with basic qualifications;
187 (2.7%) university graduates; approx. 605 (8.8%) are unemployed (2008 census data).
There are approx. 2358 households (usual residence) with an average two occupants over
35 and one 16 - 34. The second most common type of household has 2 occupants with at
least one over 65.
These statistics suggest that approx. 30% of the population is highly vulnerable, and
based on the flood simulations without wood transport, a minimum of 105 residential
buildings would be affected, with 50% of these multi-storey; assuming that each storey is
one household, with an average 3 residents per house, the minimum number of people
potentially affected by flooding is 400 (with 120 of these highly vulnerable) . If we take
into account the worst-case wood transport scenario these figures increase to 120 buildings
with 540 (162 highly vulnerable) inhabitants.
In addition we analysed the human danger zone (based on water depth and flow
velocity) and the potential transport capacity of the flow (in terms of potentially mobile
sediment diameter), which could both be important factors in an evacuation process (Fig.
9).
For the worst-case wood transport scenario, the transport capacity does not change
significantly; the ellipse in Fig. 9C shows the area where this parameter increases (upstream
of the wood depositional reach). Note the backwater effect in the flooded areas
downstream; in this area the flow has a very low transport capacity. However, the human
danger zone increases significantly, particularly in the area marked in Fig. 9D with an
ellipse, where the effect of bridge clogging causes inundation.
The car park area has an approx. capacity for 75 vehicles, although there are also
other parking areas. But just considering this area, and the average market price of 12 000 €
per vehicle, the potential damages may be as high as 400 000-500 000 € (with 85%
probability) but if wood transport is included in the simulation this value increases to 700
000 € (Fig. 10).
Figure 9: (A) Flow transport capacity (diameter of sediment in meters) without wood and
(C) scenario 3 of wood transport; (B) Danger zone for people without wood and (D)
scenario 3 of wood transport. Maps are related to the 90% probability 500-year flood.
5. Discussio
on
T
This paper uses
u a probaabilistic miicro-scale ap pproach to compute aand allocatee flood
risk andd related unncertainty. Except
E for tthe sport an
nd recreatioon areas, annd the vehiccles, all
other eeconomic criteria
c werre quantifieed at the level of individual hoouseholds. Social
vulneraability was also
a partiallly evaluatedd. Howeverr since the aim of thiis research was w to
evaluatee the influence of wood transportt in flood risk analysis, the estimatted econom mic loss
was nott as detailedd as that forr individual houses (seee an examplle of a studyy at house level
l in
De Mora and Díez-Herrero, 2008), butt mean valu ues were useed instead. The same is true
for sociial vulnerabbility which h was only partially evaluated,
e based
b on geeneral popu ulation
statisticcs, and the potential average nuumber of people p affeccted was evvaluated without
w
consideering any oth her vulneraability factoors.
VIRGINIA RUIZ-VILLANUEVA
A 348 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 4
situations where there is considerable uncertainty is the additional dimension that this
process adds to the risk assessment and risk management strategies (Merz et al., 2010).
Data on historical floods previously gathered from documentary sources and tree-
ring records were incorporated into the at-site flood frequency analysis (FFA). For this
purpose a Bayesian Markov chain Monte Carlo (MCMC) framework (Kuczera, 1999; Reis
and Stedinger, 2005; Gaal et al., 2010; Gaume et al., 2010) was used and a likelihood
function was built in to handle the information on historical floods properly. Following
Gaume et al. (2010), we considered that accurate data on extremes (in terms of discharge)
are not absolutely necessary in such analysis as the main information needed is the
frequency of non-exceedance of a perception threshold (X0) and the historical period (h).
This analysis revealed large differences in quantile estimation based on different
approaches; e.g., the estimation of the 500-year quantile was strongly affected depending
on how the data was processed (as instrumental without uncertainty, or as historical with
uncertainty), increasing to more than twice the value. Underestimating this value may
result in risk underestimation, but at the same time overestimating it could result in
financial and risk management overestimation. Because of this uncertainty in the quantile
estimation, flood hazard maps have been produced using a probability approach
(Romanowicz and Beven, 2003; Bates et al., 2004; Pappenberger et al., 2006) as
deterministic approaches normally do not take uncertainties into account (Bates et al.,
2004; Merz et al., 2005, 2007; Di Baldassarre et al., 2010).
There was no rating curve or any other data available to calibrate the hydraulic
model. Therefore, the water depth was analyzed under uncertainty of the input Manning
roughness coefficient values of the main channel and inundation areas. The value range
was treated as normally distributed but using a fuzzy set approach. Komatina (2005)
compares this method with a Monte Carlo simulation and found them highly comparable,
with both methods giving almost the same results.
In order to estimate the damaged share of the asset values, depending on water depth,
relative depth-damage functions have to be applied. Such damage functions show the
average susceptibility of each sector against inundation depth. Since there was no organized
systematic collection of flood damage data in Arenas de San Pedro, we used functions
based on those used in other similar regions in Spain (PATRICOVA, 2002; Ballesteros et
al., under review), and computed predictive bounds to take the uncertainty into account.
Some studies apply aggregated values from insurance companies. For example, Briene et al.
(2002) in the Netherlands use an approximate value of 70 000 € per flat (year 2000).
German studies (e.g. Reese et al. 2003; Meyer 2005) often apply an average value of 700 €
per m 2 of living area (full replacement value) or 350 € per m 2 (depreciated value) which
also come from an insurance company. To estimate the damage we used the average market
replacement price for this town for residential and industrial buildings, cars, and sport or
recreation areas. For residential and industrial uses, the financial value of structural damage
was taken as 15% of the average building price; the value of contents damage was assessed
as 20%. The values are reliable approximations, but the calculations are in € m-2 because the
aim of this study was to make a probabilistic evaluation of uncertainty and to analyze the
influence of wood transport. The goal was to compare values, not to obtain exact monetary
losses.
The flow velocity can also have a significant influence on damages, because danger to
humans increases with velocity (Merz et al., 2010): People may be swept away when flow
velocities are above 0.5 m/s (Marco, 1994). However, a better indicator for human
instability in flood situations is the product of flow velocity v and water depth h. Abt et al.
(1989) or Russo et al. (2011) experimented in a test flume on human subjects standing on
different surfaces at various depths and velocities until the point of instability was reached .
Although to date velocity has rarely been taken into account in damage evaluation, we
attempted to include it in the analysis by means of hazardousness or dangerousness for
people.
The results of this paper show how wood transport and deposition may significantly
increase flood risk. When wood enters the simulation the main effect is the backwater
effect upstream in critical sections such as bridges, where wood is entrapped and deposited.
This reduces the cross-sectional area, and the water depth increases significantly producing
the inundation of nearby areas. As a result, the potential damage increases by up to 50%, as
well as the number of potentially exposed persons (35%). This increment may be
equivalent to the values obtained with the discharge of clear water (without wood)
corresponding to a higher return period (see the concept of equivalent return period in
Ruiz-Villanueva et al., 2012).
6. Conclusions
The presence of wood during a flood may influence the consequences and potential
damages of the flooding. The main effect is the obstruction of cross-sectional areas
particularly in critical sections such as bridges, which can trigger the inundation of nearby
areas. As a result, the flooded areas and flow conditions are different from those defined in
the absence of wood.
Since it may not be easy to obtain empirical data of the process and there may be
considerable uncertainty, the scenario approach represents a powerful alternative.
Including the worst-case scenario in these situations is another dimension that this process
adds to risk assessment and risk management strategies.
Acknowledgments
This work was funded by CICYT MAS Dendro-Avenidas project (CGL 2010-19274) and the
Geological Survey of Spain (IGME). We are grateful to the Confederación Hidrográfica del Tajo and
Meteorological Agency (AEMET) for having provided meteorological data; the Junta de Castilla y León in
Ávila, Ayuntamiento de Arenas de San Pedro (particularly to Nuria Blázquez, Gloria Suárez and Sixto
Díaz) for their collaboration. Special mention to Martí Sánchez-Juni (UPC) for his collaboration; and to
Ignacio Gutiérrez and Luis Fernandez for their assistance with the topographical survey.
REFERENCES
Abt SR, Wittler RJ, Taylor A, Love DJ (1989) Human stability in a high hazard flood zone. Water
Resources Bulletin 25(4):881–890
Apel, H., Thieken, A. H., Merz, B., & Bl, G. (2004). Flood risk assessment and associated uncertainty, 295-
308.
Bates, P.D., Horritt, M.S., Aronica, G., Beven, K., Bayesian updating of flood inundation likelihoods
conditioned on flood extent data. Hydrological Processes 18, 3347–3370, 2004.
Benda, L. E., Sias, J. C., 2003. A quantitative framework for evaluating the mass balance of in-stream
organic debris. Forest Ecology and Managment, 172, 1-16.
Bladé, E., Cea, L., Corestein, G., Escolano, E., Puertas, J., Vázquez-Cendón, M. E., Dolz, J., and Coll, A.
2012. “Iber - Herramienta de simulación numérica del flujo en ríos.” Revista Internacional de
Métodos Numéricos para Cálculo y Diseño en Ingeniería (in press).
Briene M, Koppert S, Koopman A, Verkennis A 2002, Financiele onderbouwing kengetallen
hoogwaterschade, NEI B.V.
CEDEX., 2011. Mapa de caudales máximos. Memoria tecnica. http://
www.marm.es/es/agua/temas/gestion-de-los-riesgos-de-inundacion/memoria_tecnica_v2.0_tcm7-
162773.pdf
Chow, V.T., 1959. Open-channel Hydraulics. McGraw-Hill, New York.
Comiti, F., Mao, L., Wilcox, A., Wohl, E., Lenzi, M., 2007. Field-derived relationships for flow velocity and
resistance in high-gradient streams. Journal of Hydrology 340, 48-62.
Corestein, G., Bladé, E., Cea, L., Lara, Á., Escolano, E., 2010. Iber, a river dynamics simulation tool. GiD, 1-
4.
De Mora Jiménez, E. Y Díez Herrero, A. (2008). Análisis del riesgo de inundación en localizaciones
puntuales: el edificio Sabatini (Toledo). En: I. Galindo Jiménez, L. Laín Huerta Y M. Llorente
Isidro (Eds.), El estudio y la gestión de los riesgos geológicos. Publicaciones del Instituto Geológico
y Minero de España, Serie: Medio Ambiente. Riesgos Geológicos, nº 12, Capítulo 3, 39-52,
Madrid, IGME y Consorcio de Compensación de Seguros (MEH). ISBN: 978-84-7840-755-2;
NIPO: 657-08-022-0; D.L. M-15342-2008.
Di Baldassarre, G., Schumann, G., Bates, P., Freer, J., Beven, K., 2010. Floodplain mapping: a critical
discussion on deterministic and probabilistic approaches. Hydrol. Sci. J. 55 (3), 364–376.
Diehl, T.H., 1997, Potential Drift Accumulation at Bridges, FHWA-RD-97-28, U.S. Department of
Transportation, Federal Highway Administration, Washington, D.C.
European Commission: A new EU Floods Directive 2007/60/EC, available at:
http://ec.europa.eu/environment/water/floodrisk/ index.htm
Faulkner H., Parker D., Green C. & Beven K. Developing a translational discourse to communicate
uncertainty in flood risk between science and the practitioner. Ambio 2007, 36, 692–703.
Fellin, W., Lessmann, H., Oberguggenberger, M., Vieider, R. 2005. Analysing Uncertainty in Civil
Engineering. ISBN 3-540-22246-4 Springer Berlin Heidelberg New York.
Flood Risk Estimation, in: International Conference on Flood Estimation, edited by Spreafico, M. and
Weingartner, R., CHR Report II-17, 577–585, 2002.
FLOODsite, 2009. Flood risk assessment and flood risk management. An introduction and guidance based
on experiences and findings of FLOODsite (an EU‐funded integrated project). Deltares‐Delft
Hydraulics, Delft.
Gaál, L., J. Szolgay, S. Kohnová, K. Hlavcová, A. Viglione, 2010. Inclusion of historical information in flood
frequency analysis using a Bayesian MCMC technique: a case study for the power dam Orlík,
Czech Republic, Contributions to Geophysics and Geodesy 40, 121-147.
Gaume, E., Gaál, L., Viglione, A., Szolgay, J., Kohnov´a, S., and Blöschl, G., 2010. Bayesian MCMC
approach to regional flood frequency analyses involving extraordinary flood events at ungauged
sites, J. Hydrol. 394, 101–117.
Gómez M., Macchione F. and Russo B. (2010). Hazard criteria related to urban flooding produced by
heavy storm events. Proceedings of 1st European Congress of the IAHR. Edinburgh, May 4-6,
2010
Haque, C., Etkin, D. 2007. People and community as constituent parts of hazards: The significance of
societal dimensions in hazards analysis. atural Hazards, 41(2), 271-282.
HR Wallingford, 2005. The flood risk to people methodology. Flood Risks to People, Phase 2 R&D
Output FD2321/TR1. Defra/ Environment Agency Flood and Coastal Defence R&D
Programme.
INE, 2008. Datos del Instituto Nacional de Estadística.
Komatina, D. (2005). Uncertainty analysis as a complement to flood risk assessment, (September).
Kreibich, H., Seifert, I., Merz, B., and Thieken, A. H.: Development of FLEMOcs – A new model for the
estimation of flood losses in the commercial sector, Hydrolog. Sci. J., submitted, 2010.
Kuczera, 1999. Comprehensive at-site flood frequency analysis using Monte Carlo Bayesian inference.
Water Resources Research, VOL. 35, NO. 5, PAGES 1551-1557.
LeVeque R. J. 2002 Finite Volume Methods for Hyperbolic Problems, Cambridge University Press.
Lyn D., Cooper T., Condon D., Gan L., 2007. Factors in debris accumulation at bridge piers, Washington,
US Department of Transportation, Federal Highway Administration Research and Development,
Turner-Fairbank Highway Research Center.
Mao L., A. Andreoli, F. Comiti, M.A. Lenzi. Geomorphic effects of large wood jams on a subantarctic
mountain stream. River Research and Applications, 24, 249-266, 2008.
Marco JB (1994) Flood risk mapping. In: Rossi G, Harmancioglu N, Yevjevich V (eds) Coping with floods,
Kluwer Academic Publishers, Dordrecht, 353–373
May, C.L., Gresswell, R.E. 2003. Large wood recruitment and redistribution in headwater streams in the
southern Oregon Coast Range, U.S.A. Canadian Journal of Forest Research 33: 1352.
Mazzorana, B., Comiti, F., Volcan, C., Scherer, C., 2011. Determining flood hazard patterns through a
combined stochastic–deterministic approach. Natural Hazards, 301-316.
Mazzorana, B., Fuchs, S., 2010a. Fuzzy Formative Scenario Analysis for woody material transport related
risks in mountain torrents. Environmental Modelling & Software, 25(10), 1208-1224.
Mazzorana, B., Hübl, J., Zischg, A.M., Largiader, A., 2010b. Modelling woody material transport and
deposition in alpine rivers. Natural Hazards, 425-449.
Mazzorana, B., Zischg, A., Largiader, A., Hübl, J., 2009. Hazard index maps for woody material recruitment
and transport in alpine catchments. Natural Hazards and Earth System Science, 197-209.
Mediero, L., Jiménez, A., 2007. Regional analysis for frequency estimation of annual flood peaks in
ungauged basins of Spain. EGU2007-A-03251. Abstracts of the Contributions of the EGU
General Assembly 2007, Vienna (Austria), 15-20 April 2007. European Geosciences Union. ISSN:
1029-7006.
Merz B, Kreibich H, Thieken A, Schmidtke R (2004) Estimation uncertainty of direct monetary flood
damage to buildings. Natural Hazards and Earth System Sciences 4:153–163
Merz, B., Hall, J., Disse, M., and Schumann, A.: Fluvial flood risk management in a changing world, Nat.
Hazards Earth Syst. Sci., 10, 509-527, doi:10.5194/nhess-10-509-2010, 2010.
Merz, B., Thieken, A. H., and Bl¨ oschl, G.: Uncertainty analysis for
Merz, B., Thieken, A.H., Gocht, M., Flood risk mapping at the local scale: Concepts and challenges. In:
Flood Risk Management in Europe: Innovation in Policy and Practice (Begum, S., Stive, M.J.F.,
Hall, J.W.; Eds.). Series: Advances in Natural and Technological Hazards Research, 25, Springer,
Dordrecht, Chapter 13, 231‐251, 2007.
Merz, R. and Bloschl, G., Flood frequency regionalisation‐spatial proximity vs. catchment attributes,
Journal of Hydrology, 302, 283–306, 2005.
Messner F., Meyer V. (2005) Flood damage, vulnerability and risk perception. Challenges for flood damage
research. UFZ, Leipzig, 26 p
Messner F., Penning-Rowsell E, Green C, Meyer V, Tunstall S, van der Veen A, 2007: Guide- lines for
socio-economic Flood Damage Evaluation. FLOODsite-Report T09-06-01, 176 pp
MFE ,2011. El Mapa Forestal de España 1:25000 (MFE), Ministerio de Agricultura, Alimentación y Medio
Ambiente (MARM) http://www.magrama.es/es/biodiversidad/temas/montes-y-politica-
forestal/mapa-forestal/mfe_25.aspx
NRC (National Research Council) (2000) Risk analysis and uncertainty in flood damage reduction studies,
National Academy Press, Washington, DC
Pappenberger, F., H. Harvey, K. Beven, J. Hall, R. Romanowicz and P. Smith (2006). Implementation Plan
for library of tools for uncertainty evaluation, Flood Risk Management Research Consortium.
Paté-Cornell, M. E. (1996b); “Uncertainties in risk analysis: six levels of treatment,” Reliability Engineering
and System Safety, Vol. 54, pp. 95-111.
Reese S., Markau H.-J., and Sterr H. (2003) MERK – Mikroskalige Evaluation der Risiken in
überflutungsgefährdeten Küstenniederungen. Abschlussbericht, Kiel.
Reis, D. S., Stedinger, J. R., 2005. Bayesian MCMC flood frequency analysis with historical information.
Journal of Hydrology, 313(1-2), 97-116. doi:10.1016/j.jhydrol.2005.02.028.
Romanowicz, R.J. and Beven, K. J., 2003, “Bayesian estimation of flood inundation probabilities as
conditioned on event inundation maps”, Water Resources Research, vol. 39, pp. 10731089.
Ruiz-Villanueva, V., Bladé-Castellet, E., Sánchez-Juny, M., Martí, B., Bodoque , J.M., Díez Herrero, A. 2D
modelling of woody debris transport in rivers. Under review.
Ruiz-Villanueva, V., Bodoque, J.M., Díez-Herrero, A., Eguibar, M. A., Pardo-Igúzquiza, E. 2012.
Reconstruction of an ungauged flash flood event with large wood transport and its influence on
hazard patterns. Hydrological Processes. First on line: DOI: 10.1002/hyp.9433.
Ruiz-Villanueva, V., Díez-Herrero, A., Ballesteros, J.A., Bodoque, J.M 2012b. Potential Large Woody
Debris recruitment due to landslides, bank erosion and floods in mountain basins: a quantitative
estimation approach. Accepted for publication. River Research and Applications.
Russo, B. (2009). Design of surface drainage systems according to hazard criteria related to flooding of
urban areas. PhD Dissertation,Technical University of Catalonia, Spain.
Russo, B.; Gómez, M.; Macchione, F. 2011. Experimental approach to determine flood hazard criteria in
urban areas. 12th International Conference on Urban Drainage, Porto Alegre/Brazil, 10-15
September 2011.
Swanson, F.J. 2003. Wood in rivers: a landscape perspective. American Fisheries Society Symposium 37:
299-313.
Yen, B.C., Tung, Y.K., 1993. Some recent progress in reliability analysis for hydraulic design, in: Yen, B.C.,
Tung, Y.K. (Eds.), Reliability and Uncertainty Analysis in Hydraulic Design: A Report Prepared
by the Subcommittee on Uncertainty and Reliability Analysis in Design of Hydraulic Structures of
the Technical Committee on Probabilistic Approaches to Hydraulics. ASCE, New York, pp. 35–
79.
Versteeg and. Malalasekera, An introduction to computational fluid dynamics. Harlow : Pearson Education
Limited, 2007. ISBN: 978-0-13-127498-3.
Wind HG, Nierop TM, de Blois CJ, de Kok JL (1999) Analysis of flood damages from the 1993 and 1995
Meuse flood. Water Resources Research 35(11):3459–3465
Yeo, S. W.: Controls on flood damages, Ba River Valley, Fiji, Unpublished Ph.D. thesis, Natural Hazards
Research Centre, School of Earth Sciences, Macquarie University, 1998.
5. SÍNTESIS, DISCUSIÓN Y
CONCLUSIONES
de 98 árboles). Las principales evidencias que se observaron en los árboles afectados por las
avenidas fueron: inclinación del fuste (93 %), raíces expuestas (64 %) y heridas con
descortezados (7%). También se muestrearon otros 16 árboles no afectados (con 28
testigos extraídos) que sirvieron para elaborar una serie de referencia o patrón y con ella
eliminar anomalías debidas a variaciones climáticas, epidemias o plagas.
Los únicos datos instrumentales disponibles comunes a las 4 cuencas y para este
intervalo temporal fueron las series de precipitación en 24 horas en diversas estaciones
pluviométricas de la zona. Estos datos se utilizaron para calcular la lluvia areal en las
cuencas de estudio, y mediante un sencillo método empírico de trasformación lluvia-
escorrentía se estimaron caudales para un gran número de eventos. Ya que no se contaba
con información referente a los cambios en los usos del suelo y/o las condiciones
antecedentes de humedad para estos eventos, el parámetro de coeficiente de escorrentía se
trató de forma estocástica en las estimaciones. Así, mediante simulaciones de Monte Carlo,
se obtuvieron los valores de caudales más probables con una estimación de la incertidumbre
asociada. La validación del método se llevó a cabo comparando los valores estimados con
los observados para la serie disponible y se obtuvieron unas tasas de error del 17%.
estacionalidad de los eventos también reveló importante información, y sirvió como base
en la interpretación de las causas meteorológicas. Un 60% de los eventos tuvieron lugar en
otoño-invierno, mientras que el 40% restante en primavera-verano. Por tanto, es posible
afirmar que la mayor parte de las avenidas en estas zonas se relacionan con perturbaciones
frontales de invierno. Así lo corroboró también la comparación con otros estudios previos
de la cuenca del Tajo, y la correlación encontrada entre los caudales de los eventos más
intensos con un elevado índice NAO negativo.
Este trabajo no sólo se basó en documentación histórica, sino que combinó esta
información con los resultados de estudios dendrogeomorfológicos, lo que permitió
obtener un mayor conocimiento en la interpretación de la dinámica de estos ríos.
Los mecanismos estudiados de incorporación de material leñoso a los ríos fueron los
movimientos del terreno (deslizamientos), la erosión de los bancos de orilla, y las propias
avenidas e inundaciones. Para conocer la dinámica de incorporación de carga leñosa es
necesario establecer las áreas contribuyentes que pueden aportar el material, entendiendo
estas áreas como aquellas susceptibles de sufrir un determinado proceso. Estas fueron
trazadas basadas en mapas existentes de la zona de estudio y para los cuales se analizaron los
factores de conectividad con el río. Así, creando unas matrices de lógica difusa se
establecieron criterios en función de la intensidad de los eventos y su frecuencia para
establecer la probabilidad de que un árbol sea incorporado como detrito leñoso desde las
laderas o desde el corredor fluvial. Estos criterios dependen de la distancia al cauce, la altura
del árbol, la pendiente de la ladera, la inundabilidad y la capacidad de erosión del río.
Además, se requiere un conocimiento detallado de la tipología y distribución de la
vegetación que definen una resistencia (o facilidad) a ser incorporado como detrito leñoso
en el río.
En este modelo, el inicio del movimiento de una pieza de madera (suponiendo ésta
como un cilindro) situada en un cauce, se determina mediante un análisis de las fuerzas que
actúan sobre ella. Por un lado, la fuerza gravitacional y la fuerza de arrastre que facilitan el
movimiento, y por otro lado la fuerza de rozamiento o fricción que se opone al
movimiento. Estas fuerzas dependen de parámetros como la densidad de la madera, su
longitud y diámetro, el ángulo que forma con respecto al flujo, las condiciones
hidrodinámicas del flujo (calado y campo de velocidades) y unos coeficientes de fricción
con el lecho y de arrastre. Así, una pieza de madera iniciará su movimiento y se desplazará,
bien por flotación (a una velocidad similar a la del agua) o por rodadura o arrastre de fondo
(a una velocidad diferente a la del agua).
En primer lugar se estimó el volumen de material que pudo llegar al tramo de estudio,
evaluando el tipo de vegetación situada en la zona afectada por la avenida y el proceso de
avulsión que tuvo lugar aguas arriba de este tramo. Siguiendo la metodología propuesta en
el apartado 4.4, en función de la especie y del estado de la vegetación se estimó la resistencia
de la misma a ser incorporada al cauce como detrito leñoso, y junto con la densidad
(número de individuos por área) se calculó el volumen. Este volumen se utilizó como
Los resultados permitieron observar los patrones de depósito del material leñoso y
éstos se compararon con la abundante información gráfica (fotografías) disponible
recopilada unos días después del evento. Además se analizó el calado en aquellas secciones
donde se disponía de información complementaria, y de esta forma se pudo comprobar que
el escenario de transporte escalonado es el que más se ajustaba a las observaciones hechas en
campo y a la reconstrucción del fenómeno de obstrucción del puente recogida en el
apartado 4.3. Mientras que los escenarios de transporte repentino y transporte constante
suponían un mayor y menor porcentaje de obstrucción del puente respectivamente.
VIRGINIA
A RUIZ-VILLANUEEVA 369 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER
C 5
Figura 5.1:
5 Esquem
ma conceptuual de la met
etodología propuesta
p paara estimar la
frecueencia de aven
enidas
Laa metodoloogía propuessta para inccorporar la cargac leñosaa en los anáálisis de riessgo por
avenidaas comienzaa con el anállisis de las ááreas contribbuyentes, ell estudio dee la vegetación y la
estimacción de loss volúmenes disponibbles. A coontinuación n, se simulla, junto con c la
hidrodiinámica, el transporte
t del
d materiaal en los ríoss, permitien ndo de este modo anallizar su
influenccia en la pelligrosidad y en el riesgoo (Fig. 5.2).
VIRGINIA RUIZ-VILLANUEVA
A 370 TESIS DOCTO
ORAL/ PhD THESIS (2012)
VIRGINIA
371 A RUIZ-VILLANUEEVA
Figure 5.2: Proposed methodology for incorporating woody material transport in flood risk analysis
CAPÍTULO /CHAPTER 5
mayoría de los ejemplares arbóreos estén inclinados y presenten raíces expuestas, y apenas
un porcentaje muy pequeño muestre heridas. Por otro lado, el tramo estudiado del río
Arenal se caracteriza por presentar un lecho con material más fino (cantos y gravas), y
podría definirse como una alternancia de tramos salto-poza y lecho plano (con zonas de
incisión sobre el lecho rocoso). En este caso aumenta la capacidad de transporte y el umbral
de movimiento de material esta cerca del caudal de bankfull (Dietrich, 1989; Buffington,
1995); además se observan abundantes restos de vegetación en el cauce y las márgenes. Este
aumento en la capacidad de transporte y el abundante material leñoso observado, que
generalmente será transportado por flotación, son los causantes de las numerosas heridas
encontradas en la vegetación a lo largo del río. Sin embargo, como ya se ha discutido en el
apartado 4.2, las heridas producidas en este caso no siempre están relacionadas con eventos
extraordinarios de avenidas, sino que están marcadas por el umbral de movimiento
mencionado y estarían condicionadas por la cantidad y el tipo de material disponible para
ser movilizado.
Por consiguiente, los datos extraídos de estos análisis deben ser examinados
cuidadosamente y si es posible, complementados con otras fuentes de información, como
las fuentes documentales.
analizadas; sería pues necesario recopilar más información para analizarlo con mayor
profundidad y poder interpretarlo en un contexto regional y reconstruirlo con más detalle.
Esta reconstrucción de eventos pasados, en el caso del evento de 1997 en la cuenca del
arroyo Cabrera (apartado 4.3), se realizó combinando diversos métodos: empíricos, como
el de la competencia del flujo, el llamado slope-conveyance, la simulación hidráulica y la
simulación hidrológica. La integración de estas metodologías, permitió estimar el caudal,
reproducir el proceso de obstrucción del puente principal y reconstruir el patrón de
precipitación para un evento que puede considerarse como un evento tipo de avenida
súbita para la zona de la Sierra de Gredos. La simulación hidrológica realizada empleó
como datos de entrada los resultantes de la simulación estocástica de la precipitación. De
nuevo, y al igual que ocurría con la aplicación del método racional modificado, fue
condicionante la falta de datos con suficiente resolución espacial y temporal. Los métodos
tradicionales de interpolación espacial pueden resultar limitados cuando se aplican a datos
de lluvia en regiones montañosas (Basist, et al., 1994; Daly et al., 1994; Prudhomme y
Reed, 1999; Marquínez et al., 2003); por esta razón se utilizó la simulación geoestadistica.
De esta forma se tuvo en cuenta también la topografía, al incorporar como variable
secundaria el modelo digital de elevaciones. Esta limitación en la resolución de los datos
puede ser superada en la actualidad con el desarrollo de la tecnología del radar
meteorológico; aunque no puede utilizarse para la reconstrucción de eventos históricos
(como se ha tratado en este trabajo de investigación), puesto que no se dispone de datos
históricos. Como se ha demostrado en diversos estudios, el uso de estos datos proporciona
una potente herramienta y abre nuevas perspectivas en el análisis de avenidas actuales
(Gaume y Borga, 2008; Marchi et al., 2009). No obstante, es importante también ser
consciente de los errores que pueden acompañar a los datos derivados del radar
meteorológico, especialmente en las zonas de montaña (Hossain et al., 2004).
Por otro lado, la elección de los métodos hidráulicos empleados en los apartados 4.3,
4.6 y 4.7 estuvo, una vez más, limitada por los datos de partida, pero en este caso también
estuvo condicionada por el tipo de tramo de río estudiado. Así, en el caso del tramo del
arroyo Cabrera, se utilizaron métodos empíricos, combinados con simulación hidráulica
unidimensional. Las características del tramo y del flujo permitían asumir las
simplificaciones de estos métodos (ver apartado 2.6). Las limitaciones más destacables del
uso de modelos 1D y de Hec-Ras en particular son: (i) La ecuación de la energía supone
siempre distribuciones hidrostáticas de presiones y la ecuación de fricción permanente de
Manning, además no tiene en cuenta la turbulencia. Por tanto, la solución no se ajustará a
la realidad en casos donde las presiones y las tensiones turbulentas se alejan del modelo
lineal; (ii) En principio sólo se deberían utilizar para modelizar ríos con pendientes
menores de 10º ya que no se tiene en cuenta la componente vertical del peso de la columna
de agua en las ecuaciones; (iii) Los saltos u obstáculos al flujo no son reproducidos
correctamente (ya que Hec-Ras no realiza balance de fuerzas).
Mientras que en los trabajos realizados en el río Arenal se optó por utilizar un modelo
bidimensional, puesto que el flujo desborda e inunda la zona urbana. El uso de modelos
2D, y de Iber en particular, permite resolver algunas de las limitaciones descritas para los
modelos 1D. Iber posibilita incluir la turbulencia en los cálculos, y utiliza el método de los
volúmenes finitos con un esquema de segundo orden de Roe. Este método es adecuado para
la simulación de flujos en ríos de montaña, ya que es capaz de modelizar cambios bruscos de
sección o discontinuidades a lo largo del tramo, así como singularidades (frentes de onda,
resaltos hidráulicos…), o tránsito de hidrogramas (régimen variable). Además se trata de un
método conservativo, incluso con variaciones en los frentes seco-mojado (permite eliminar
del cálculo los volúmenes finitos secos, e incorporarlos si se mojan, y se conserva
exactamente el volumen de agua incluso en presencia de fuertes irregularidades
geométricas). Sin embargo, también se asumen en estos modelos bidimensionales algunas
simplificaciones: (i) La profundidad de la lámina de agua debe ser pequeña con relación a
las otras dimensiones del problema (ecuaciones de aguas someras); (ii).También suponen
una distribución hidrostática de presiones en la vertical; (iii) De nuevo, deben ser utilizados
para modelizar tramos con pendiente reducida. El cumplimiento de estas hipótesis implica
que las componentes de la velocidad y aceleración en el eje z son despreciables frente a las
componentes en los otros ejes, y también que éstas últimas tienen una marcada
uniformidad vertical.
utilizado en este trabajo (ver apartados 4.6 y 4.7) ha sido la capacidad de transporte o
diámetro crítico que el flujo es capaz de transportar, basado en los criterios de Shields
(1936). Sería pues conveniente obtener datos para incorporar este tipo de material en los
análisis.
que puedan representar diversos estados incluyendo el denominado peor caso posible
(worstcase scenario). En el caso del trabajo presentado en el apartado 4.4, los escenarios
generados pueden clasificarse como de tipo formativo combinado con lógica difusa (fuzzy-
formative), que resulta en una composición cuantitativa/cualitativa basada en matrices de
impacto (Tietje, 2005). Mientras que, por otro lado, los escenarios creados en los apartados
4.6 y 4.7 pueden definirse como modelos (model scenario; Ruth y Hannon, 1997), basados
en caracterizar alguna variable desconocida dentro del sistema, en este caso el volumen y el
régimen de transporte del material leñoso. El objetivo final es obtener diferentes
posibilidades para evaluar situaciones probables o posibles (Meadows et al., 1974); en este
trabajo, relativas a la influencia del material leñoso en la peligrosidad y el riesgo por
avenidas.
La estimación del riesgo llevada a cabo en Arenas de San Pedro (apartado 4.7)
estuvo enfocada en la estimación probabilista de pérdidas potenciales asociadas a la avenida
de periodo de retorno 500 años, para analizar, por un lado las incertidumbres asociadas, y
por otro, cuantificar la influencia suponiendo además transporte de material leñoso.
Generalmente las evaluaciones de riesgo se realizan para diferentes caudales asociados a
distintos periodos de retorno de manera que al final se pueda hacer una estimación de las
denominadas pérdidas medias anuales (anual average damages; Meyer, 2007). También es
frecuente realizar estudios coste-beneficio cuando se quiere evaluar la eficacia de una
determinada medida de corrección o prevención frente a las avenidas (Tung, 2002; Zhu y
Lund, 2009). En cualquier caso, los análisis de riesgo siempre están sujetos a numerosas
fuentes de incertidumbre (Apel et al., 2004; Merz et al., 2010). Para tener en cuenta estas
incertidumbres Mazzorana y Fusch (2010) proponen una red de escenarios (derivada de los
trabajos de Scholz y Tietje, 2002) compuesta por diversas evaluaciones de los procesos
naturales (avenidas y flujos de derrubios), diferentes situaciones para los elementos
expuestos, y la vulnerabilidad de estos elementos, con el fin de obtener como resultado una
serie de evaluaciones de riesgo. En el caso de la generación de escenarios de procesos de
avenidas, estos autores contemplan la carga leñosa y generan las distintas situaciones
basadas en los trabajos de Mazzorana et al. (2009 y 2011) en los que resaltan las diversas
incertidumbres asociadas a este proceso.
Los resultados obtenidos en este trabajo permiten reducir y/o evaluar estas
incertidumbres relativas a la estimación del material que puede llegar a los ríos, al estudio
de su transporte en los ríos, su interferencia con estructuras como puentes y analizar su
influencia en la peligrosidad y riesgo por avenidas.
Por ejemplo, para completar el registro de avenidas puede ser posible emplear las
fuentes de datos geológicas, tal y como se explicaba en la introducción de esta Memoria,
aunque en este tipo de ambientes montañosos puede resultar difícil encontrar sedimentos
útiles en los estudios de paleoinundaciones, debido a la elevada capacidad de transporte, al
tamaño del sedimento, etc. Sin embargo, estudios realizados en ambientes de montaña
(Rico et al., 2001; Thordyncraft y Ruiz-Villanueva, 2010) demuestran que sí es posible
encontrarlos y emplearlos en la reconstrucción de avenidas. Si se diera el caso, entonces se
podrían combinar estas fuentes de datos con las utilizadas en este trabajo de investigación
(fuentes históricas y dendrogeomorfológicas).
En n relación al
a modelo desarrolladoo para simullar el transp porte de mmaterial leño oso en
ríos, see abren divversas líneass de investtigación. Prrincipalmen nte estaríann enfocadass en la
mejora de las capaccidades del modelo. Laa aplicación n en más ríos, de diversaa tipología (como
ríos braaided de grravas u otro os), y emplleando datoos observad dos de cam mpo, haría posible
p
corregirr y mejorar su funcion namiento. A Además es posible
p impplementar nnuevos algoritmos
que permitan sim mular piezaas de madeera con raaíces o ram mas. Siguieendo los trrabajos
desarrolllados por Braudrick
B y Grant (20000) presenttan una aprroximación consideran ndo los
detritoss con raíces como piezas cilíndric as unidas a un disco. ElE efecto dee las raíces supone
s
e extremo sobre el leccho del río, el área sum
que la ppieza está eleevada por ese mergida del tronco
t
disminuuye, lo que haría
h variar el balance dde fuerzas tal
t y como se representaa en la figurra 5.4.
VIRGINIA RUIZ-VILLANUEVA
A 382 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 5
Figu
gure 5.4: Schhematic andd body-force
ce diagrams of some off the
t componnents of the
he force
balance acting
a on a log
l with a rroot-wad (fr
from Braudr
drick and Grrant, 2000).
).
Figura 5.4:: Diagramass esquemátiicos del balaance de fuerrzas sobre uuna pieza dee madera
supooniendo éstaa como un cilindro
c y uun disco (ext
xtraído y trad
aducido de B Braudrick y Grant,
2000).
VIRGINIA
A RUIZ-VILLANUEEVA 383 TESIS DOCCTORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER
C 5
O
Otra capaciddad del mod delo que noo ha podidoo ser explorada en proffundidad durante
este trabbajo de inveestigación, ha
h sido la s imulación conjunta de
d hidrodinnámica, maaterial
leñoso y sedimentos. Las lim mitaciones ttemporales no permitiieron recabaar la inform
mación
indispen nsable paraa obtener los datos necesarios para la siimulación y validació ón del
transpoorte de seddimentos. Sin embarrgo, esta posibilidad
p abre una línea nueeva de
investiggación, no sólo
s en el esstudio de laas avenidas,, sino comppletando asíí el rompeccabezas
que form man los prinncipales commponentes modelizados de la dinámica fluviial (figura 5.5).
Figura 5.5:
5 Compoonentes prin incipales de la dinámicaa fluvial quee podrían seer
moodelizados tras
t los resul
ultados de esste trabajo de
d investigac
ación en fututuros trabajoos.
A
Además, la metodología
m a utilizada ppodría replicarse para adaptar
a el m
modelo a otrro tipo
mo el caso dee hielo, por ejemplo (K
de mateeriales, com Kolerski, 20 011). En la eescasa biblio
ografía
que exisste sobre este tema se afirma
a que eel transportte de hielo durante inuundaciones puede
produciir el mismoo efecto qu ue la carga leñosa, obbstruyendo puentes y amplifican ndo los
efectos de la inunddación.
VIRGINIA RUIZ-VILLANUEVA
A 384 TESIS DOCTO
ORAL/ PhD THESIS (2012)
CAPÍTULO /CHAPTER 5
5.4. Conclusiones
La relación encontrada entre las evidencias externas que muestran los árboles y la
posición geomorfológica que ocupan éstos en el cauce puede ser de utilidad para
optimizar los muestreos, focalizándolos en aquellos árboles que aporten mayor
información.
Debido a los efectos que produce la carga leñosa, se observó que las consecuencias de
una avenida pueden modificarse. Esto nos ha llevado a definir el nuevo concepto de
periodo de retorno equivalente, en este caso aplicado a los procesos de obstrucción
producidos por el transporte y deposito de material leñoso, pero que podría ser de
aplicación para otros procesos que modifiquen la intensidad de la avenida.
El desarrollo del modelo para simular el transporte de material leñoso en ríos es una
aportación importante ya que no existía hasta el momento ninguna herramienta
similar. La aplicación de este modelo a los casos de estudio ha puesto de manifiesto su
viabilidad y su potencial, aunque se hace necesaria la obtención de datos de campo
para su adecuada validación. Sin disponer de estos datos empíricos, se presenta como
Conclusions
The general conclusion drawn from this research is that the limited availability of
data in forested mountain river basins means that different types of source data need to be
integrated, and various methods must be combined to reconstruct past events. The
influence of the woody load on the hydrodynamic behaviour of rivers cannot be ignored
and must also be included in flash flood hazard and risk analysis.
Various specific conclusions can also be made from the research carried out and the
results obtained:
The relationship found between the external evidence obtained from the trees and
their geomorphological position within the stream bed can be used to optimize
sampling procedures, by focusing on the trees which provide the most information.
These botanical data sources also provide information on river dynamics and can be
used to interpret transport mechanisms for solid material (wood or sediment).
The information compiled from the documentation consulted shows the seasonal
nature of the flash floods occurring in the study area. This seasonal distribution is
conditioned by the meteorological mechanisms which generate the flash floods. We
are thus able to confirm that most events occur in the study area in autumn/winter
and are related to Atlantic frontal disturbances. However, the events occurring in
summer (associated with convective storms) are related to the highest death tolls.
Reconstruction of the 1997 event shows that indirect estimates of this type of events
should be approached from different viewpoints (empirical equations, hydraulic and
The methodology applied for obtaining the clogging curves for the bridge in the
Venero Claro area can easily be extrapolated to any other infrastructure. The
hydraulic behaviour interpreted in the three regions defined by these curves will also
be similar in other bridges with the same characteristics.
Due to the effects of the woody load, it was observed that the impact of a flash flood
may be modified. This has led us to define a new concept of the equivalent return
period, in this case applied to the obstruction processes resulting from the transport
and deposition of woody material, but which could also be applied to other processes
which alter the flash flood intensity.
The GIS based methodology applied for the study of the wood entrainment
processes would allow other processes to be included in the analysis. This
methodology is easy to apply in other areas and, as is shown, can obtain the orders of
magnitude of the potentially available woody material. The field indicators used to
study the woody load dynamics also allow valuable information to be extracted on its
entrainment and in-river transport.
The development of the model simulating the transport of woody material in rivers
is an important contribution, as no similar tool is available to date. Applying this
model to the case studies highlights its viability and potential, although field data
must be obtained to validate it appropriately. Even if this empirical data is not
available it is still an alternative tool for generating scenarios, which has proved to be
very useful in all the analyses carried out.
Risk analysis from a probabilistic viewpoint allowed the inclusion of different sources
of uncertainty in this research, and obtained reliable results even though calibration
data was not available.
The proposed methodologies, both for combining data sources and for including the
woody load in flash flood risk assessment, improve the analysis of this type of process,
particularly in forested mountain basins.
6. REFERENCIAS BIBLIOGRÁFICAS,
LISTA DE FIGURAS Y TABLAS
Abbe, T. B., Montgomery D. R. 2003. Patterns and processes of wood debris accumulation in the Queets
River basin, Washington, Geomorphology, 51, 81–107.
Abbe, T.B., Montgomery, D.R. 1996. Large woody debris jams, channel hydraulics, and habitat formation
in large rivers. Regulated Rivers: Research and Management 12, 201–221.
Abernethy, B., Rutherfurd, I.D. 2001. The distribution and strength of riparian tree roots in relation to
riverbank reinforcement. Hydrological Processes. 15: 63-79.
Abrahams, A. D., Atkinson, G. Li, J. F. 1995. Step-pool streams: Adjustment to maximum flow resistance,
Water Resour. Res., 31, 2593–2602.
Adenlof, K. A.,. Wohl, E. E. 1994Controls on bedload movement in a subalpine stream of the Colorado
Rocky Mountains, USA, Arct. Alp. Res., 26, 77– 85.
Ajami, N. K., Gupta, H., Wagener, T., Sorooshian, S. 2004: Calibration of a semi-distributed hydrologic
model for streamflow estimation along a river system. J. Hydrol., 298, 112–135.
Alestalo, J. 1971. Dendrochronological interpretation of geomorphic processes. Fennia 105.
Alonso Otero, F.; Arenillas Parra, M., Sáenz Ridruejo, C.: La morfología glaciar en las montañas de Castilla
la Vieja y León. Consejo General de Castilla y León. Burgos, 1982. ISBN 84-300-6572-5.
Anagnostou, E.N., Grecu, M., Anagnostou, M.N. 2006. X-band polarimetric radar rainfall measurements
in keys area microphysics project. J. Atmos. Sci. 63, 187– 203.
Anders, A. M., Roe, G. H., Hallet, B., Montgomery, D. R., Finnegan, N. J., Putkonen,. J. 2006.Spatial
patterns of precipitation and topography in the Himalaya, Tectonics, Climate, and Landscape
Evolution, edited by S. D.Willett, , Spec. Pap. Geol. Soc. Am.398, 39–53,
Andreoli, A., Comiti, F. Lenzi, M. A. 2007. Characteristics, distribution and geomorphic role of large
woody debris in a mountain stream of the Chilean Andes, Earth Surf. Processes Landforms, 32,
1675–1692.
Andrieu, H., Creutin, J. D. Delrieu, G., Faure, D. 1997. Use of a weather radar for the hydrology of a
mountainous area. Part 1: Radar measurement interpretation. J. Hydrol., 193, 1–25.
Andrieu, H., Delrieu, G., Creutin, J. D. 1995: Identification of vertical profiles of radar reflectivity using an
inverse method: 2—Sensitivity analysis and case study. J. Appl. Meteor., 34, 240–259.
Arcement, Schneider 1989. Guide for selecting Manning’s roughness coefficients for natural channels and
floodplains. US Geological survey, Water supply paper; 2339.
Archer, E., Martien, K.K. And Taylor, B.L. 2007. Use of an individual-based simulation of BCB bowhead
whale population dynamics to examine empirical genetic data. Paper SC/59/BRG17 submitted to
the Annual Meeting of the Scientific Committee of the International Whaling Commission, May,
2007. 28pp.
Atjay, G.L.., Ketner, P Duringneaud, P. 1979. Terrestrial primary production and phytomass, SCOPE 13:
The Global Carbon Cycle (B. Bolin, E. Degens, S. Kempe, and P. Ketner, editors), Wiley,
Chichester, United Kingdom, pp. 129–181.
Ballesteros, J.A., Stoffel, M., Bollschweiler, M., Bodoque, J.M., Díez-Herrero, A. 2010b. Flash-flood impacts
cause changes in wood anatomy of Alnus glutinosa, Fraxinus angustifolia and Quercus pyrenaica.
Tree Physiology 30, 773-781.
Ballesteros,J.A. 2011. Aplicación de métodos dendrogeomrofológicos para el análisis de peligrosidad y
riesgo por avenidas e inundaciones. Tesis Doctoral. Universidad Politécnica de Madrid. 266pp.
Barnes, B.1939. The structure of discharge recession curve. Am. Geophys.Union. Trans., IV 721-725.
Barnes, H.H..1967. Roughness characteristics of natural channels. U.S. Geol. Surv. Water-Supply Pap.,
1849: 213 pp.
Barnolas, M., Llasat, M.C. 2007. Metodología para el estudio de inundaciones históricas en España e
implementación de un SIG en las cuencas del Ter, Segre y Llobregat. Monografías CEDEX, 90. 264
p. + CD compacto, Madrid.
Barredo, J.I. 2009. Normalised flood losses in Europe: 1970–2006. Nat. Hazards Earth Syst. Sci., 9, 97–
104, 2009
Barriendos, M., Coeur, D.,2004. Flood data reconstruction in historical times from non-instrumental
sources in Spain and France. In: Systematic, Palaeoflood and Historical Data for the Improvement
of Flood Risk Estimation. Methodological Guidelines (ed. by G. Benito & V. R. Thorndycraft), 29–
42. Centro de Ciencias Medioambientales, Madrid, Spain.
Barry, P J., Anderson, R. L., Swain. K. M. Undated. How to Evaluate and Manage Storm-Damaged Forest
Areas. Southeastern Area, U.S. Forest Service. 15 p.
Basist A, Bell GD, Meentemeyer V 1994. Statistical relationships between topography and precipitation
patterns. J Clim 7:1305–1315.
Bates, P.D., Horritt, M.S., Aronica, G., Beven, K. 2004. Bayesian updating of flood inundation likelihoods
conditioned on flood extent data. Hydrological Processes 18, 3347–3370.
Bates, P.D., Horritt, M.S., Aronica, G., Beven, K. 2004. Bayesian updating of flood inundation likelihoods
conditioned on flood extent data. Hydrological Processes 18, 3347–3370.
Bathurst, J. C. 1990.Tests of three discharge gauging techniques in mountain rivers, Hydrology of
Mountainous Areas, edited by L.Molnar, IAHS Publ.190, 93–100.
Bathurst, J.C., Crosta, G.B., Garcia‐Ruiz, J.M., Guzzetti, F., Lenzi, M.A., Rios Aragués, S. 2003.
DAMOCLES: Debris‐fall assessment in mountain catchments for local end‐users. In: Debris‐Flow
Mitigation: Mechanics, Prediction and Assessment, Rickenmann & Chen (eds) 2003 Millpress,
Rotterdam.
Bathurst, J.C., Graf, W.H., Cao, H.H. 1987. Bed load discharge equations for steep mountain rivers”, en
Thorne, C.R.; Bathurst, J.C. y Hey, R.D. (Ed.): Sediment transport in gravel-bed rivers. Wiley, pp.
453-491.
Battan Louis J. 1973. Radar obsrevation of the Atmosphere
Bea, F. 1985. Los granitoides hercínicos de la mitad occidental del batolito de Ávila (sector de Gredos).
Aproximación mediante el concepto de superfacies. R. Acd. Ci. Ex. Fis. Nat., Madrid, LXXIX, 549-
572.
Bea, F. y Moreno-Ventas, I. 1985ª. Estudio petrológico de los granitoides del área centro-norte de la Sierra
de Gredos (Batolito de Ávila; Sistema Central Español). Stvd. Geol. Salmant., XX, 137-174.
Bea, F. y Moreno-Ventas, I. 1985b. Diferentes series de rocas con afinidades granodioríticas en la Sierra de
Gredos (Batolito de Ávila, Sistema Central Español). Criterios para su identificación. Stvd. Geol.
Salmant., XX, 175-203.
Bea, F. y Pereira, M.D. 1990. Estudio petrológico del Complejo Anatéctic o de Peña Negra (Batolito de
Ávila, España Central). Rev. Soc. Geol. España, 3 (1-2), 87-103.
Bea, F.; Carnicero, A.; Gonzalo, J.C.; López Plaza, M.; y Rodríguez Alonso, M.D. (eds.). 1985. Geología de
los granitoides y rocas asociadas del macizo Hespérico. Libro Homenaje a L.C. García de Figuerola.
Editorial Rueda, Madrid, 542 págs.
Bell and Moore. 1998. A grid-based distributed flood porecasting model for use with weather radar data.
Part 2. Case studies. Hydrol. Erath Syst. Sci., 2, 283-289.
Bellido, F.; Capote, R.; Casquet, C.; Fúster, J.M.; Navidad, M.; Peinado, M. y Villaseca, C. 1981. Caracteres
generales del Cinturón Hercínico en el Sector Oriental del Sistema Central Español. Cuad. Geol.
Ibérica, 7, 15-51.
Bellido, F.; Casquet, C.; Huertas, M.J.; Ibarrola, E. y Ruiz, M.T. 1990. Petrología. Mapa geológico de
España. Escala 1:
Belmonte A.C., Beltran F.S. 2001. Flood events in Mediterranean ephemeral streams (ramblas) in Valencia
region, Spain. Catena, 45, 229–249.
Benda, L. E. 1990.The influence of debris flows on channels and valley floors in the Oregon Coast Range,
USA, Earth Surf. Processes Landforms, 15, 457–466.
Benda, L. E., Hassan, M. A., Church, M., May, C. L 2005.Geomorphology of steepland headwaters: The
transition from hillslopes to channels, J. Am. Water Resour. Assoc., 41, 835–851,
Benda, L. E., Miller, D. Sais, J. Martin, D,. Bilby, R., Veldhuisen, C. ,Dunne, T. 2003b Wood recruitment
processes and wood budgeting, The Ecology and Management of Wood in World Rivers, edited by
S. V.Gregory, K. L.Boyer, A. M.Gurnell, 49–73, Am. Fish. Soc., Bethesda, Md.
Benda, L. E., Veldhuisen, C., Black, J. 2003a. Debris flows as agents of morphological heterogeneity at low-
order confluences, Olympic Mountains, Washington, Geol. Soc. Am. Bull., 115, 1110–1121.
Benda, L.E., Poff, N.L., Tague, C., Palmer, M.A., Pizzuto, J., Cooper, S., Stanley, E. and Moglen, G. 2002:
How to avoid train wrecks when using science in environmental problem solving. Bioscience 52(12),
1127–36.
Benito, G. 2002. La paleohidrología en el análisis de inundaciones. En: F.J. Ayala- Carcedo y J. Olcina
Cantos (Coords.), Riesgos Naturales. Editorial Ariel, Ariel Ciencia, Barcelona, Cap. 50, pp. 953-967.
Benito, G. 2006. Riesgos de inundaciones: tendencias históricas y perspectivas de acuerdo con el cambio
climático. Revista Cuaternario y Geomorfología, 20(3-4): 29-44.
Benito, G. Lang, M. Barriendos, M. Llasat, M.C. Francés, F. Ouarda, T. Thorndycraft, V. Enzel, Y.
Bardossy, A. Coeur, D. Bobée, B. 2004. Use of systematic, palaeoflood and historical data for the
improvement of flood risk estimation. Review of scientific methods. Natural Hazards 31: 623–643.
Benito, G. Rico, M. Thorndycraft, V.R. Sánchez-Mova, Y. Sopeña, A. Díez-Herrero, A., Jiménez, A. 2006.
Palaeoflood records applied to assess dam safety in SE Spain. En Ferreria Alves, Leal y Cardoso. River
Flow. Taylor and Francis Group, London. 2113- 2120.
Benito, G. Rohde R.F. Seely, M. Kulls, C. Dahan, O. Enzel, Y. Todd S. Botero, B. Grodek, T. 2010.
Management of alluvial aquifers in two southern African ephemeral rivers: Implications for IWRM.
Water Resources Management 24: 641-667.
Benito, G., Díez-Herrero, A., y Fernández De Villalta, M. 2003. Magnitude and Frequency of Flooding in
the Tagus Basin (Central Spain) over the Last Millenium. Climatic Change, 58 (1-2): 171-192.
Benito, G., Lang, M., Barriendos, M., Llasat, M.C., Frances, F., Ouarda, T., Thorndycraft, V., Enzel, Y.,
Bardossy, A., Coeur, D., Bobée, B. 2004. Use of Systematic Paleoflood and Historical Data for the
Improvement of Flood Risk Estimation. Review of Scientific Methods. Natural Hazards 31, 623-
643.
Benito, G., Thorndycraft, V.R. (Eds.). 2004. Systematic, palaeoflood and historical data for the
improvement of flood risk estimation, methodological guidelines. CSIC, Madrid, 115 pp.
Benito, G., Thorndycraft, V.R. 2005. Palaeoflood hydrology and its role in applied hydrological sciences.
Journal of Hydrology, 313: 3–15.
Benson, M.A., Dahymple, T. 1967. General field and office procedures for indirect discharge
measurements. U.S. Geological Survey, Techniques of Water-Resources Investigations, Book 3, ch.
Al, 30 pp.
Berg, N., Carlson, A., Azuma, D. 1998.Function and dynamics of woody debris in stream reaches in the
central Sierra Nevada, California, Can. J. Fish. Aquat. Sci., 55, 1807–1820.
Bezzola G, Sigg H, Lange D. 2004. Schwemmholzru¨ckhalt in der Schweiz. In: Internationales Symposion
Interpraevent 2004‘‘, Riva, Trient; Tagungspublikation, Band 3, pp 29–40.
Bilby, R. E., Likens, G. E. 1980. Importance of organic debris dams in the structure and function of stream
ecosystems, Ecology, 61, 1107–1113.
Bilby, R. E.. 1984, Removal of wood debris may affect stream channel stability, J. For., 609–613.
Bilby, R.E., Ward, J.W. 1989. Changes in characteristics and function of woody debris with increasing size
of streams in western Washington. Transactions of the American Fisheries Society 118, 368– 378.
Bisaz, A., Escher, F., Grosjean, M., Ives, J.D., Messerli, B., Price, M.F. 1997. Mountains of the world:
challenges for the twenty-first century. A contribution to Chapter 13, Agenda 21. Mountain
Agenda. Bern, Paul Haupt.
Bladé, E., Cea, L., Corestein, G., Escolano, E., Puertas, J., Vázquez-Cendón, M. E., Dolz, J., and Coll, A.
2012. Iber Herramienta de simulación numérica del flujo en ríos.. Revista Internacional de Métodos
Numéricos para Cálculo y Diseño en Ingeniería (in press).
Blahut, J., Westen, C. J. van., Sterlacchini, S. 2010.Analysis of landslide inventories for accurate prediction
of debris-flow source areas, Geomorphology, 119, 36–51,
Blong, R.J., Gillespie, R. 1978. Fluvially transported charcoal gives erroneous C-14 ages for recent deposits.
Nature, 271: 739- 741.
Blöschl, G., Reszler, C., Komma, J. 2008. A spatially distributed flash flood forecasting model.
Environmental Modelling & Software 23 (4), 464–478.
Bluck, B.J. 1979. Structure of coarse grained braided stream alluvium. Trans. R. Soc. Edinb., 70: 181-221.
Bocchiola, D., Catalano, F., Menduni, G., Passoni, G. 2002. An analytical–numerical approach to the
hydraulics of floating debris in river channels. Journal of Hydrology 269, 65-78.
Bocchiola, D., M.C. Rulli, y Rosso. R. 2006. Transport of large woody debris in the presence of obstacles.
Geomorphology 76: 166-178.
Bodoque, J.M.; Eguibar, M.A.; Díez-Herrero, A.; Gutiérrez-Pérez, I.; Ruiz-Villanueva, I. 2011. Can the
discharge of a hyperconcentrated flow be estimated from paleoflood evidence? Water Resources
Research, 47, W12535, 1-14.
Bohorquez, P., Darby, S.E. 2008. The use of one- and two-dimensional hydraulic modelling to reconstruct
a glacial outburst flood in a steep Alpine valley. Journal of Hydrology.
Bonnifait, Delrieu ,G., Lay, M., Boudevillain, B., Masson, A., Belleudy, P., Gaume, E., Saulnier., G. 2009.
Distributed hydrologic and hydraulic modelling with radar rainfall input: Reconstruction of the 8–9
September 2002 catastrophic flood event in the Gard region, France. Advances in Water Resources
32, 1077–1089.
Borga M, Anagnostou EN, Frank E. 2000. On the use of real-time radar rainfall estimates for flood
prediction in mountainousbasins. Journal of Geophysical Research 105(D2): 2269–2280.
Borga M. 2007. Hydrometeorological data resources and technologies for effective flash flood forecasting.
Research plan. University of Padova, 130.
Borga, M., Gaume, E., Creutin, J.D., Marchi, L. 2008. Surveying flash flood response: gauging the ungauged
extremes. Hydrol. Process. 22 (18), 3883–3885. doi:10.1002/hyp.7111.
Borga, M., Tonelli, F., Moore, R.J., Andrieu, H. 2002. Long-term assessment of bias adjustment in radar
rainfall estimation. Water Resour. Res. 38 (11), 1226.
Bowman, M. J. 1977. Nutrients Distribution And Transport In Long Island Sound. Estuarine Coastal
Mar. Sci. 5: 53 L-548. East Africa Meteorological. Deft. 1980. Climatological statistics for East
Africa. Nairobi.
Bracken, L. J., Croke, J. 2007. The concept of hydrological connectivity and its contribution to
understanding runoff-dominated geomorphic systems, Hydrol. Processes, 21, 1749–1763.
Bradley, C., Mosugu, M, Gerrard AJ. 2005. Simulation modelling of water movement in a cracking clay soil.
Soil Use and Management. 21, 386-395
Bradley, W.C., Fahnestock, R.K., Rowekamp,E.T. 1972. Coarse sediment transport by flood flows on the
Knik River, Alaska. Geol. Sot. Am. Bull., 83: 1261-1284.
Bragg, D. C., Service, U. F., Uam, P. O. B., Kershner, J. L., Unit, F. E. 2004. Sensitivity of a Riparian Large
Woody Debris Recruitment Model to the Number of Contributing Banks and Tree Fall Pattern.
USDA Forest Service Gen. Tech. Rep. PSW-GTR-181 19: 117-122.
Brath, A., Montanari, A. Toth, E. 2004. Analysis of the effects of different scenarios of historical data
availability on the calibration of a spatially-distributed hydrological model, J. Hydrol., 291, 232–253.
Braud, I.; Chancibault, K.; Debionne, S.; Lipeme Kouyi, G.; Sarrazin, B.; Jacqueminet, C.; Andrieu, H.;
Béal, D.; Bocher, E.; Boutaghane, H.; Branger, F.; Breil, P.; Chocat, B.; Comby, J.; Dehotin, J.;
Dramais, G.; Furusho, C.; Gagnage, M.; Gonzalez-Sosa, E.; Grosprêtre, L.; Honegger, A.;
Jankowfsky, S.; Joliveau, T.; Kermadi, S.; Lagouy, M.; Leblois, E.; Martin, J.-Y.; Mazagol, P.-O.;
Michel, K.; Molines, N.; Mosini, M.-L.; Puech, C.; Renard, F.; Rodriguez, F.; Schmitt, L.; Thollet, F.
and Viallet, P. 2010.The AVuPUR project (Assessing the Vulnerability of Peri-Urban Rivers):
experimental setup, modelling strategy and first results, Proceedings of the 7th Novatech 2010
Buffington, J. M., W. E. Dietrich, J. W. 1992. Kirchner, Friction angle measurements on a naturally formed
gravel streambed: Implications for critical boundary shear stress, Water Resour. Res., 28, 411–425.
Bullón, T. 2011. Relationships between precipitation and floods in the fluvial basins of Central Spain based
on documentary sources from the end of the 16th century. Nat. Hazards Earth Syst. Sci. 11, 2215–
2225.
Butler, D.R., Malanson, G.P. 1985. A history of high-magnitude snow avalanches, southern Glacier
National Park, Montana, USA. Mountain Research and Development 5: 175-182.
Buytaert, W., Célleri, R. De Bièvre, B. Cisneros, F. Wyseure, G. Deckers, J. Hofstede. R. 2006. Human
impact on the hydrology of the Andean páramos. Earth-Science Reviews 79:53-72.
Camarasa, A.M. 2002. Crecidas e inundaciones. En: F.J. Ayala-Carcedo y J. Olcina Cantos (Coords.),
Riesgos Naturales. Ariel Ciencia, Barcelona, 46: 859-877.
Camarasa, A.M. 2006. Inundaciones en España. Tipología. La importancia de las avenidas súbitas. Riesgos
Naturales y Desarrollo Sostenible. Impacto, Predicción y Mitigación, I.G.M.E., Serie Medio
Ambiente. Riesgos Geológicos, 10, 167-178.
Cantero, A. 2003. El paisaje forestal está ineludiblemente unido a las actividades socio-económicas de la
población rural. Actas de la II Reunión sobre Historia Forestal. Cuad. Soc. Esp. Cien. For. 16: 309-
313.
Carling, P. A., Cao, Z., Holland, M. J., Ervine, D. A., Babaeyan-Koopaei, K. 2002 Turbulent flow across a
natural compound channel, Water Resour. Res., 38(12), 1270, 10.1029/2001WR000902.
Carollo, F., Ferro, V., Termini, D. 2005. Analyzing Turbulence Intensity in Gravel Bed Channels J.
Hydraul. Eng., 131(12), 1050–1061.
Carpenter, T.M., Georgakakos, K.P. 2004a. Intercomparison of lumped versus distributed hydrologic
model ensemble simulations. Journal of Hydrology.
Carpenter, T.M., Georgakakos, K.P. 2004b. Impacts of parametric and radar rainfall uncertainty on the
ensemble streamflow simulation of a distributed hydrologic model. Journal of Hydrology, 298, No.
1-4, 202-221.
Carrasco, R.M., Pedraza, J. 1995. Morfología glaciar de las cimas en la sierra de Gredos: monteras de hielo.
En: T. Aleixandre y A. Pérez-González (eds.), Reconstrucción de paleoambientes y cambios
climáticos durante el Cuaternario, Monografías 3, 227-238, Centro de Ciencias Medioambientales
(CSIC).
Castellarin, A., Di Baldassarre, G., Brath, A. 2010. Floodplain management strategies for flood attenuation
in the Po river. River Research and applications,.
Cavalli M., Tarolli, P., Marchi, L., FontanaG.. 2008. The effectiveness of airborne LiDAR data in the
recognition of channel-bed morphology. Catena 73, 249–260.
Cea, L. 2005. An unstructured finite volume model for unsteady turbulent shallow water flow with wet-dry
fronts: Numerical solver and experimental validation. PhD Dissertation. Universidad de A Coruña,
Departamento de Métodos Matemáticos y de Representación. Programa de Doctorado en Ingeniería
Civil.
Cea, L..Vázquez-Cendón, M. E. 2010. Unstructured finite volume discretization of two-dimensional
depth-averaged shallow water equations with porosity. Int. J. Numer. Meth. Fluids. 63, 903–930.
Cendrero, A., J. Remondo, J. Bonachea, V. Rivas , J. Soto. 2006. Acción Humanay Procesos Geológicos
Superficiales ¿Qué Le Estamos Haciendo A La Faz De La Tierra?. Rev.R.Acad.Cienc.Exact.Fís.Nat.
(Esp) Vol. 100, Nº. 1, pp 187-209, 2006. VI Programa de Promoción de la Cultura Científica y
Tecnológica.
Chiari M, Friedl K, Rickenmann D. 2010. A one-dimensional bedload transport model for steep slopes.
Journal of Hydraulic Research 48(2): 152-160.
Chow, V.T. 1959. Open Channel Hydraulics. McGraw-Hill, New York, 680 pp.
Chow, V.T., Maidment, D. R. Mays L.W. 1988.Applied Hvdrologv, McGraw Hill, New York.
Chow, V.T., Maidment, D., Mays, L.W. 1988. Applied Hydrology, McGraw‐Hill, New York.
Church, M., Jones, D. 1982. Channel bars in gravel-bed rivers. In:. Hey, R. D., Bathurst, J. C. and Thorne,
C.R. (eds), Gravel-bed Rivers. John Wiley and sons Ltd. Chichester.
Cohen, S.,Willgoose,G.,Hancock,G. 2010. ThemARMspatiallydistributedsoil evolution model:three-
dimensionalmodelframework,andanalysisofhill- slope
andlandformresponses.JournalofGeophysicalResearch115
Collier, C.G. 2007. Flash flood forecasting: What are the limits of predictability?. Quarterly Journal of the
Royal Meteorological Society Volume 133, Issue 622, pages 3–23, January 2007 Part A.
Comiti, F., Andreoli, A., Mao, L., Lenzi, M. A. 2008a Wood storage in three mountain streams of the
southern Andes and its hydro-morphological effects, Earth Surf. Processes Landforms, 33, 244–262.
Comiti, F., Cadol, D., Wohl, E.E. 2009. Flow regimes, bed morphology, and flow resistance in self-formed
step-pool channels. Water Resources Research. 45: W04424.
Comiti, F., Mao, L., Preciso, E., Picco, L., Marchi, L., Borga, M. 2008b.Large wood and flash floods:
Evidence from the 2007 event in the Davča basin (Slovenia), Monitoring, Simulation, Prevention
and Retention of Dense and Debris Flow II, edited by D.de Wrachien, C. A.Brebbia, M. A.Lenzi,
WIT Trans. Eng. Sci.60, 173–182.
Cooke, R. 1991.Experts in Uncertainty. Oxford University Press.
Corestein, G., Bladé, E., Cea, L., Lara, Á., Escolano, E. 2010. IBER, a river dynamics simulation tool. GiD,
1-4.
Coriell, F. 2002. Reconstruction of a paleoflood chronology for the Middlebury River Gorge using tree
scars as flood stage indicators. Unpublish senior thesis, Middlebury College, Department of
Geology, Middlebury, VT 05753, 45 pp.
Costa JE. 1983. Paleohydraulic reconstruction of flash-flood peaks from boulder deposits in the Colorado
Front Range. Geological Society of America Bulletin 94: 986–1004.
Costa, J. E. 1974. Response and recovery of a Piedmont watershed from tropical storm Agnes. Wat. Resour.
Res. 10, 106-112.
Costa, J. E. 1983. Paleohydraulic reconstruction of flash-flood peaks from boulder deposits in the Colorado
Front Range, Geol. Soc. Am. Bull., 94, 986–1004.
Costa, J. E. 1986. Rhéologie, geomorphic, and sedimentologic differentiation of water floods,
hyperconcentrated flows and debris flows. In: Flood Geomorphology (ed. by V. R. Baker, R. C.
Kochel & P. C. Patton), 113-122. Wiley-Interscience, New York.
Costa, J. E., Jarrett, R. D. 1981. Debris flows in small mountain stream channels of Colorado and their
hydrologic implications, Bull. Assoc. Eng. Geol., 18, 309–322.
Costa, J. E., O'Connor, J. E. 1995. Geomorphically effective floods, Natural and Anthropogenic Influences
in Fluvial Geomorphology, edited by J. E.Costa, , 45–56, AGU, Washington, D. C.
Costa, J.E. 1978. Holocene stratigraphy in flood-frequency analysis. Water Resour. Res., 14.: 62&632.
Costa, J.E. 1983. Paleohydraulic reconstruction of flash-flood peaks from boulder deposits in the Colorado
Front Range. Geol. Sot. Am. Bull., 94: 986-10)4.
Costa, J.E. 1984. Physical geomorphology of debris flows. In: J.E Costa and P.J. Fisher (IEditors),
Developments and Applications of Geomorphology. Springer, New York, pp. 268-3 17.
Costa, J.E. 1988. Rheologic, geomorphic and sedimentologic differentiation of water floods,
hyperconcentrated flows and debris flows. In: V.R. Baker, R.C. Kochel and PC. Patton (Editors),
Flood Geomorphology. Wiley, New York, pp. 113-122.
Costa, J.E., Jarrett, R.D. 2008. An Evaluation of Selected Extraordinary Floods in the United States
Reported by the US Geological Survey and Implications for Future Advancement of Flood Science.
US Geological Survey, Scientific Investigations Report 2008-5164, 52 pp.
Cotter, E. 1978. The evolution of fluvial style, with special reference to the central Appalachian Paleozoic.
In: Miall, A.D. (Ed.), Fluvial Sedimentology: Canadian Society of Petroleum Geologists Memoir,
vol. 5, pp. 361–383.
Creutin JD, Borga M. 2003. Radar hydrology modifies the monitoring of flash flood hazard. 10·10
002/hyp.5122. Hydrological Processes 17(7): 1453–1456.
Cunge, J.A. 1967. On the Subject of a Flood Propagation Method. Journal of Hydrological Research, 7(2),
205-230.
Cunnane, C. 1978. Unbiassed plotting position, a review. Journal of Hydrology, 327 (3/4), 205-222.
Cunnane, C. 1987. Review of statistical models for flood frequency estimation. En: V.P. Singh (Ed.).
Regional Flood Frequency Analysis, 49-95, Reidel Publising Company.
Dalrymple, T. Benson, M.A. 1968. Measurement of peak discharge by the slope-area method. U.S.
Geological Survey Techniques of Water-Resources Investigations, Book 3, Ch. A2, 12 PP.
Daly, C., R.P. Neilson, Phillips. D.L. 1994. A statistical-topographic model for mapping climatological
precipitation over mountainous terrain. Journal of Applied Meteorology 33: 140-158.
Dana. 1850. Euphausia superba.
David, G. C. L., E. Wohl, S. E. Yochum, B. P. Bledsoe, Controls on spatial variations in flow resistance
along steep mountain streams, Water Resour. Res., 46, W03513, 10.1029/2009WR008134, 2010
Delrieu G, Ducrocq V, Gaume E, Nicol J, Payrastre O, Yates E, Kirstetter P-E, Andrieu H, Ayral PA,
Bouvier C, Creutin JD, Livet M, Anquetin A, Lang M, Neppel L, Obled C, Parent-du- Chatelet J,
Saulnier GM, Walpersdorf A, Wobrock W. 2005. The catastrophic flash-flood event of 8–9
September 2002 in the Gard region, France: a first case study for the C´evennes-Vivarais
Mediterranean Hydro-meteorological Observatory. Journal of Hydrometeorology 6: 34–52.
Delrieu, G., and J. D. Creutin, 1995: Simulation of radar mountain returns using a digitized terrain model.
J. Atmos. Oceanic Technol., 12, 1038–1049.
Delrieu, G., S. Caoudal, and J. D. Creutin, 1997: Feasibility of using mountain return for the correction of
ground-based X-band weather radar. J. Atmos. Oceanic Technol., 14, 368–385.
Di Baldassarre, G., Schumann, G., Bates, P.D., Freer, J., and Beven, K., Floodplain mapping: a critical
discussion on deterministic and probabilistic approaches, Hydrological Sciences Journal, 55(3), 364‐
376, 2010.
Di Baldassarre, G., Schumann, G., Bates, P.D., Freer, J., Beven, K. 2010. Floodplain mapping: a critical
discussion on deterministic and probabilistic approaches, Hydrological Sciences Journal 55, 364-
376,
Diehl, T.H. 1997, “Potential Drift Accumulation at Bridges”, FHWA-RD-97-28, U.S. Department of
Transportation, Federal Highway Administration, Washington, D.C.
Dietrich, W. E., Dunne, T. 1978. Sediment budget for a small catchment in mountainous terrain, Z.
Geomorphol., 29, 191–206,
Dietrich, W. E., Dunne, T. 1993.The Channel Head. Chapter 7 in Channel Network Hydrology, Edited
by K. Beven and M. J. Kirkby, John Wiley, p.175-219.
Dietrich, W. E., P. J. 1989.Whiting, Boundary shear stress and sediment transport in river meanders of
sand and gravel, River Meandering, edited by S.Ikeda, G.Parker, 1–50, AGU, Washington, D. C.,
Díez Herrero, A. 2001. Geomorfología e Hidrología fluvial del río Alberche. Modelos y SIG para la gestión
de riberas. Tesis doctoral. Universidad Complutense de Madrid (inédita).
Díez Herrero, A., Laín Huerta, L., Llorente Isidro, M. (eds.). 2006. Mapas de peligrosidad de avenidas e
inundaciones. Métodos, experiencias y aplicación. IGME. Madrid.
Díez-Herrero, A, Laín Huerta, L, Llorente, M. 2008a. Mapa de peligrosidad por avenidas e inundaciones.
Guía para su elaboración. IGME, Madrid, 190 pp.
Díez-Herrero, A., Ballesteros-Cánovas, J.A., Bodoque, J.M., Ruiz-Villanueva, V. 2012. A new
methodological protocol for the use of dendrogeomorphological data in flood risk analysis.
Hydrology Research. In press.
Díez-Herrero, A., Ballesteros-Cánovas, J.A., Bodoque, J.M., Ruiz-Villanueva, V. 2012. A new
methodological protocol for the use of dendrogeomorphological data in flood risk analysis.
Hydrology Research.
Dingman, S. L., Fluvial Hydrology, 383, W. H. Freeman, New York, 1991.
Diplas, P. 1987. Bed load in gravel-bed streams. /. Hydmul. Div. ASCE 113 (3), 277-292
Doviak, R. J., and D. S. Zrnic, 1993: Doppler Radar and Weather Observations. Dover, 562 pp.
Durán, J.J. 1985. Ccoordinación, F.J. Ayala Carcedo. Geología y prevención de daños por inundaciones.
Madrid : Instituto Geológico y Minero de España, D.L.. 421 pp. ISBN 84-7474-324-9.
Dury, G.H. 1976. Change prediction, present and former, from channel dimensions. Journal of Hydrology,
30: 219–245.
Ely, L. L., Y. Enzel, V. R. Baker, D. R. Cayan, 1993. A 5000-year record of extreme floods and climate
change in the southwestern United States, Science, 262, 410–412,
ESYRCE. 2007. Encuesta sobre superficies y rendimientos de cultivos. Ministerio de medio ambiente,
medio rural y medio marino. Madrid.
F. Gob, F. Petit, J.-P. Bravard, A. Ozer, A. Gob. 2003. Lichenometric application to historical and
subrecent dynamics and sediment transport ofa Corsican stream (Figarella River—France).
Quaternary Science Reviews 22, 2111–2124.
Ferrer, J. 1993. Recomendaciones para el cálculo hidrometeorológico de Avenidas. CEDEX Monografía M-
37. 76 pp.
Ferrer, J. 1996. El modelo de función de distribución SQRT-ET MÁX en el análisis de máximos
hidrológicos: aplicación a lluvias diarias. Tesis doctoral. Universidad Politécnica de Madrid.
Ferrer, M.; González de Vallejo, L.I., García López-Davalillo, J.C., Rodríguez, J.A., Estévez, H. y Trimboli,
M. 2004. Pérdidas por terremotos e inundaciones en España durante el periodo 1987-2001 y su
estimación para los próximos 30 años (2004-2033). Instituto Geológico y Minero de España y
Consorcio de Compensación de Seguros, Madrid, 126 pp.
Ferrer, M.; Rodríguez, J. y Estrela, T. 1995. Generación automática del número de curva con sistemas de
información geográfica. Ingeniería del Agua, vol. 2 (4), 43-58.
Fontaine, T. A., T. S. Cruickshank, J. G. Arnold, Hotchkiss, R. H. 2002. Development of a snowfall-
snowmelt routine for mountainous terrain for the Soil Water Assessment Tool (SWAT), J. Hydrol.,
262, 209–223.
Francés, F. 2004. Flood frequency analysis using systematic and non-systematic information. En:
Systematic, palaeoflood and historical data for the improvement of flood risk estimation:
Methodological guidelines, G. Benito y V.R. Thorndycraft (Eds.), CSIC y Comisión Europea,
Madrid, pp. 54-70.
Francés, F., J.D. Salas y D.C. Boes, 1991. Flood Frequency Analysis by Using Historical Information. Procc.
of XXIV IAHR International Congress, Madrid. A 11-21.
Francis, R. A., G. E. Petts, A. M. Gurnell, 2008Wood as a driver of past landscape change along river
corridors, Earth Surf. Processes Landforms, 33, 1622–1626,
Fryirs, K., Brierley, G. J., Preston, N. J. and Kasai, M. 2007a. Buffers, barriers and blankets: The
(dis)connectivity of catchment-scale sediment cascades. Catena, 70, 49-67.
Fuchs S., Keiler M., Zischg A., Bründl M. 2005. The long-term development of avalanche risk in
settlements considering the temporal variability of damage potential. Natural Hazards and Earth
System Sciences 5, 893-901.
Gaál, L., J. Szolgay, S. Kohnová, K. Hlavcová, A. Viglione, 2010. Inclusion of historical information in flood
frequency analysis using a Bayesian MCMC technique: a case study for the power dam Orlík, Czech
Republic, Contributions to Geophysics and Geodesy 40, 121-147.
García Pérez, A. 2008. Métodos avanzados de Estadística Aplicada. Técnicas avanzadas. Editorial UNED.
Colección Educación Permanente.
García Pérez, A. 2010. Métodos avanzados de Estadística Aplicada. Métodos robustos y de remuestreo.
Editorial UNED. Colección Educación Permanente.
García Pérez, A. 2011a. Estadística Aplicada: conceptos básicos. Editorial UNED. Colección Educación
Permanente.
García Pérez, A. 2011b. Estadística Aplicada con R. Editorial UNED. Colección Educación Permanente.
Garzón Heydt, M.G.; González Lodeiro, F. y Vegas, R. 1976. Estudio tectónico y microtectónico de la
fracturación en un segmento del Sistema Central español. Stvd. Geol. Salmant., X, 7-16.
Garzón, G.; Martínez, J. y Alonso, A. 1992. Los cambios antrópicos en la dinámica del río Jarama (Madrid)
como modificadores del riesgo
Garzón, M.G. 1972. Estudio morfoestructural de la Sierra de Gredos en la transversal Ávila-Sierra de San
Vicente. Tesis de Licenciatura, Universidad Complutense de Madrid, 84 págs, Madrid.
Garzón, M.G. 1977. Estudio preliminar de la morfología de la zona comprendida entre Talavera de la Reina
y Arenas de San Pedro (Borde meridional de la Sierra de Gredos). Actas II Reunión Nac. Grupo de
Trabajo del Cuaternario. Trab. Neog. Cuat., 6, 85-91.
Garzón, M.G. 1980. Estudio geomorfológico de una transversal en la Sierra de Gredos Oriental (Sistema
Central Español). Ensayo de una cartografía geomorfológica. Tesis Doctoral, Facultad de CC.
Geológicas, U. Complutense de Madrid, 425 p. (inédito).
Garzón, M.G. 1985. Las avenidas como fenómeno geológico. En: Ayala, F.J. (Coord.), Geología y
Prevención de daños por inundaciones, 5-53, IGME, Madrid.
Garzón, M.G. y Alonso, A. 1995. Variabilidad holocena e histórica de un río meandriforme de gravas (río
Jarama, Madrid). En: T. Aleixandre y A. Pérez-González (eds.), Reconstrucción de paleoambientes y
cambios climáticos durante el Cuaternario, CSIC, Centro de Ciencias Medioambientales,
Monografías 3, 79-92.
Garzón, M.G. y Alonso, A. 1996. El río Guadarrama, morfología y sedimentación actual en un cauce
arenoso de tipo braided. Cuad. Geol. Ibérica, 21, 369-393.
Garzón, M.G.; Fernández, P. y Centeno, J.D. 1991. La morfogénesis en el Sistema Central Ibérico. En:
M.G. Garzón, J.D. Centeno y E. Acaso (eds.), Problemas geomorfológicos del Centro y Noroeste de
la Península Ibérica, Ed. Universidad Complutense, Madrid, 61-72.
Garzón-Heydt, 1999; Daños y beneficios de las inundaciones de los ríos. Rev. R. Acad. Cienc. Exact. Fisc. y
Nat. 93,1; 115-125.
Gaume E., Livet M., Desbordes M., Villeneuve J.-P. 2004. Hydrologic analysis of the Aude, France, flash-
flood 12 and 13 November 1999. J Hydrol , 286, 135–154.
Gaume, E. 2006. On the asymptotic behavior of flood peak distributions. Hydrology and Earth System
Sciences 10 (2), 233–243.
Gaume, E., Bain, V., Bernardara, P., Newinger, O., Barbuc, M., Bateman, A., Blaskovicova, L., Bloschl, G.,
Borga, M., Dumitrescu, A., Daliakopoulos, I., Garcia, J.,Irimescu, A., Kohnova, S., Koutroulis, A.,
Marchi, L., Matreata, S., Medina, V.,Preciso, E., Sempere-Torres, D., Stancalie, G., Szolgay, J., Tsanis,
I., Velasco, D., and Viglione, A. 2009. A compilation of data on European flash floods. Journal of
hydrology 367 (1–2): 70–78.
Gaume, E., Borga, M. 2008. Post flood field investigations after major flash floods: proposal of a
methodology and illustrations. Journal of Flood Risk Management 1 (4), 175–189.
Gaume, E., Gaál, L., Viglione, A., Szolgay, J., Kohnov´a, S., and Blöschl, G. 2010. Bayesian MCMC
approach to regional flood frequency analyses involving extraordinary flood events at ungauged sites,
J. Hydrol. 394, 101–117.
Georgakakos, K. P. 1986. A generalized stochastic hydrometeorological model for flood and flash flood
forecasting. Part 1: Formulation. Water Resour. Res., 22, 2083–2095.
Geza, M., E. P. Poeter, J. E. McCray, 2009.Quantifying predictive uncertainty for a mountain-watershed
model, J. Hydrol., 376, 170–181,
Gippel, C.J. 1995. Potential of Turbidity Monitoring for Measuring the Transport of Suspended Solids in
Streams. Hydrological Processes. 9: 83-97.
Girod, B. and Mieg, H. 2008. Wissenschaftliche und politische Gründe für den Wandel der IPCC-
Szenarien, Gaia, 17 (3), 302-311.
Glancy, P.A., and Williams, R.P. 1994. Problems with indirect determinations of peak streamflows in steep,
desert stream channels: 1994 National Conference on Hydraulic Engineering, Proceedings, August
1-5, 1994, Buffalo, NY
Glaser, R., Stangl, H. 2003. Historical floods in the Dutch Rhine Delta, Nat. Hazards Earth Syst. Sci., 3,
605–613.
Glaser, R.: Klimageschichte Mitteleuropas, 2001. 1000 Jahre Wetter, Klima, Katastrophen, Wiss.
Buchgesellschaft, Darmstadt,
Gob, F., N. Jacob, J.-P. Bravard, F. Petit, 2005.Determining the competence of mountainous
Mediterranean streams using lichenometric techniques, Geomorphological Processes and Human
Impacts in River Basins, edited by R. J.Batalla, C.Garcia, IAHS Publ.299, 161–170.
Godet, M. 1986. Introduction to ‘la prospective’: Seven key ideas and one scenario method. Futures, vol. 2
no. 2, pp 134-57.
Gomi T, Sidle RC. 2003. Bedload transport in managed steep-gradient headwater streams of southeastern
Alaska. Water Resources Research 39: 1336.
Goovaerts P. 2000. Geostatistical approaches for incorporating elevation into the spatial interpolation of
rainfall. Journal of Hydrology 228: 113-129.
Grant, G. 1990. Hydrologic, geomorphic and aquatic habitat implications of old and new forestry. Pub No:
1144.
Grant, G. E., F. J. Swanson, M. G. Wolman, 1990.Pattern and origin of stepped-bed morphology in high-
gradient streams, western Cascades, Oregon, Geol. Soc. Am. Bull., 102, 340–352,
Grant, G. E., T. Mizuyama1991., Origin of step-pool sequences in high gradient streams: A flume
experiment, Japan-US Workshop on Snow Avalanche, Landslide, Debris Flow Prediction and
Control, edited by M.Tominaga, 523–532, Japan Sci. and Technol. Agency, Natl. Res. Inst. for
Earth Sci. and Disaster Prev., Tsukuba, Japan,
Gregory, K.J. 1976. Drainage basin adjustments and man. Geographia Polonica, 34, 155-173.
Gregory, K.J.(Ed.) 1977, River Channel Changes, John Willey & Sons, 448 pp., Chichester, N. Ireland.
Grigg, N. S., llelweg, O. J. 1975 Slate o f t h e art of estimating Hood damage in urban areas. War Resour.
Hull. 11(2), 379-390.
Grimm, M. M., E. E. Wohl, R. D. Jarrett, 1995.Coarse-sediment distribution as evidence of an elevation
limit for flash flooding, Bear Creek, Colorado, Geomorphology, 14, 199–210,
Grimm, M.M. 1993. Paleoflood history and geomorphology of Bear Creek basin, Colorado. M.S. thesis,
Colorado State University, Ft. Collins, Colorado, 126 pp.
Grissino-Mayer, H.D. 2003. A manual and tutorial for the proper use of an increment borer. Tree-Ring
Res. 59, 63–79.
Gupta, A. 1975. Stream characteristics in eastern Jamaica, an environment of seasonal flow and large floods.
Am. J. Sci., 275: 825- 847.
Gupta, A. 1988. Large floods as geomorphic events in humid tropics. In: V.R. Baker, R.C. Kochel and P.C.
Patton (Editors), Flood Geomorphology. Wiley, New York, pp. 301-315.
Gupta, V.K.; Rodríguez-Iturbe, I. y Wood, E.F. 1986. Eds.: Scale problems in Hydrology, D. Reidel,
Dordrecht, Holland.
Gupta, V.K.; Waymire, E. y Wang, C.T. 1980. A representation of an instantaneous unit hydrograph from
geomorphology. Water Resour. Res., 16(5), 855-862.
Gurnell A.M. y Sweet R. 1998. The distribution of large woody debris accumulations and pools in relation
to woodland stream management in a small low-gradient stream. Earth Surface Processes and
Landforms, 23: 1101–1121.
Gurnell, A. M., H. Piégay, F. J. Swanson, S. V. Gregory, 2002.Large wood and fluvial processes, Freshwater
Biol., 47, 601–619.
Gurnell, A.M., Petts, G.E., Hannah, D.M. 2001. Riparian vegetation and island formation along the gravel-
bed Riume Tagliamento, Italy. Earth Surface Processes and Landforms 26, 31– 62.
Gutiérrez-Elorza, M. 2008. Geomorfología. Ed. Pearson-Prentice Hall. ISBN: 9788483223895.
Hack, J. T. 1965. Geomorphology of the Shenandoah Valley, Virginia and West Virginia, and origin of the
residual ore deposits. Professional Paper.
Hack, J. T., J. C. Goodlett, 1960. Geomorphology and forest ecology of a mountain region in the central
Appalachians, U.S. Geol. Surv. Prof. Pap.347, 66,
Haga, H., T. Kumagai, K. Otsuki, S. Ogawa, 2002Transport and retention of coarse woody debris in
mountain streams: An in situ field experiment of log transport and a field survey of coarse woody
debris distribution, Water Resour. Res., 38(8).
Hazen, A. 1930. Water supply. American Civil Engineers Handbook. John Wiley & Sons, New York, NY.,
pp. 1444–1518.
Hitz, O.M., Gärtner, H., Heinrich, I., Monbaron, M. 2008. Application of ash (Fraxinus excelsior L.) roots
to determine erosion rates in mountain torrents. Catena 72: 248-258.
Hoffman, F. O. and Hammonds, J. S. 1994. Propagation of Uncertainty in Risk Assessments: The Need to
Distinguish between Uncertainty due to Lack of Knowledge and Uncertainty due to Variability.
Conference Proceedings: "When and How Can You Specify a Probability Distribution Function
when You Don't Know Much?" Risk Analysis 14(5):707-712.
Hooke, R.LeB. 1999, Spatial distribution of human geomorphic activity in the United States: Earth Surface
Processes and Landforms, v. 24, p. 687-692.
Horrit, M.S., Bates, P.D. 2002. Evaluation of 1D and 2D Numerical Models for Predicting River Flood
Inundation. Journal of Hydrology, 268, 87-99.
Horrit, M.S., Bates, P.D., Mattinson, M.J. 2006. Effects of mesh resolution and topographic representation
in 2D finite volume models of shallow water fluvial flow. Journal of Hydrology 329, 306–314.
Horton, R. E. 1945. Erosional development of streams and their drainage basins; hydrophysical approach
to quantitative morphology, Geol. Soc. Am. Bull., 56, 275–370,
Horton, R., y Wierenga, P. J. 1983. Estimating the Soil Heat Flux from Observations of Soil Temperature
Near the Surface. Soil Science Society of America Journal, 47: 14-20.
Hossain, S.C.; Mukherjee, S.; Pati, K.C. & Saha,R.N. Groundwater Arsenic Contamination And Its
Health Effects In The Ganga-Meghna- Brahmaputra Plain. Journal Of Environmental Monitoring,
2004. 6(6): p. 74N-83N.
http://ec.europa.eu/regional_policy/sources/docgener/studies/pdf/montagne/mount1.pdf
http://ec.europa.eu/regional_policy/sources/docgener/studies/pdf/montagne/mount9.pdf
http://www.unep-wcmc.org/datasets-tools-reports_15.html?types=JournalPaper,Report,Book&ctops=
http://www.unepwcmc.org/medialibrary/2010/09/10/71bb554a/Mountain_Watch.pdf
Hupp, C.R. 1988. Plant ecological aspects of flood geomorphology and paleoflood history. In: V.R. Baker,
R.C. Kochel and P.C. Patton (Editors), Flood Geomorphology. Wiley, New York, pp. 335-356.
Hupp, C.R., Osterkamp, W.R. 1985. Bottomland vegetation distribution along Passage Creek, Virginia, in
relation to fluvial landforms. Ecology 66: 670-681.
Hutte, P. 1968. Experiments on windflow and wind damage in Germany: Site and susceptibility of spruce
forests to storm damage. Forestry 41: 20-26.
Jacob N. 2003. The torrential crisis: temporal and spatial thresholds of change; the example of the
Cévennes (Ardèche, France). In Fouache E. (Ed.): The Mediterranean World Environment and
History. Elsevier, Paris, 123-136.
Jacobi, W., Zeglen, S., Jeffrey, Beale, D. 2008. Black stain root disease progression in coastal Douglas-fir in
British Columbia. Can. J. Plant Pathol. 30: 339–344.
Jacobson, R. B.,. Gran, K. B 1999. Gravel sediment routing from widespread, low-intensity landscape
disturbance, Current River basin, Missouri, Earth Surf. Processes Landforms, 24, 897–917,
Jakob, M., Jordan, P. 2001. Design flood estimates in mountain streams; the need for a geomorphic
approach, Can. J. Civ. Eng., 28, 425–439,
Jarret, R. Silverton. Paleoflood Research to Improve Flood Science. NOAA River Forecast Center
Anchorage. Project USGS National Research Program Denver.
http://www.uccs.edu/~bvogt/ges431/readings/week_8/jarrett-paleoflood-talk.pdf.
Jarrett RD, y England Jr JF. 2002. Reliability of paleostage indicators for paleoflood studies. En: House,
P.K., Webb, R.H., Baker, V.R., Levish,D.R. (Eds.), Ancient Floods, Modern Hazards: Principles and
Applications of Paleoflood Hydrology. Water Science and Application, American Geophysical
Union: Washington, D.C.; Vol. 5: 91–109.
Jarrett, R. D. 1984. Hydraulics of high-gradient streams, J. Hydraul. Eng., 110, 1519– 1539.
Jarrett, R. D. 1990. Paleohydrologic techniques used to define the spatial occurrence of floods,
Geomorphology, 3, 181–195.
Jarrett, R. D. 1990a. Hydrologic and hydraulic research in mountain rivers, Water Resour. Bull., 26, 419–
429,.
Jarrett, R. D., J. P. Capesius, D. Jarrett, J. F. England Jr. 1996. Where the past (paleoflood hydrology) meets
the present, understanding maximum flooding, Geological Society of America Annual Meeting,
Abstracts with Programs, A-110.
Jarrett, R.D. 1984. Hydraulics of high gradient streams. J. Hydraul. Div. Am. Sot. Civ. Eng., 110: 1519-
1539.
Jarrett, R.D. 1985. Determination of roughness coefficients for streams in Colorado. U.S. Geol. Surv.
Water-Resources Investigations Report 85-4004.54 pp.
Jarrett, R.D. 1986. Evaluation of the slope-area method of computing peak discharge. In: S. Subitzky
(Editor), Selected papers in the hydrological sciences 1986. U.S. Geol. Surv. Water-Supply Pap.,
2310: 13-26.
Jarrett, R.D. 1990. Paleohydrologic techniques used to define the spatial occurrence of floods.
Geomorphology, 3: 181-195.
Jarrett, R.D. 1993. Flood elevation limits in the Rocky Mountains. In: C.Y. Kuo (Editor), Engineering
Hydrology. American Society of Civil Engineering, pp. 180-185.
Jarrett, R.D. and Costa, J.E. 1986. Hydrology, geomorphology, and dam-break modeling of the July 15,
1982, Lawn Lake Dam and Cascade Lake Dam failures, Larimer County, Colorado. U.S. Geol. Surv.
Prof. Pap. 1369: 78 pp.
Jarrett, R.D. and Costa, J.E. 1988. Evaluation of the flood hydrology in the Colorado Front Range using
precipitation, streamflow, and paleoflood data. U.S. Geological Survey Water-Resources
Investigations Report 87-4117,37 pp.
Jarrett, R.D. and Malde, H.E. 1987. Paleodischarge of the late Pleistocene Bonneville Flood, Snake River,
Idaho, computed from new evidence. Geol. Sot. Am. Bull., 99: 127-134. Williams, G.P. and Costa,
J.E., 1988. Geomorphic measurements after a flood. In: V.R. Baker, R.C. Kochel and P.C. Patton
(Editors), Flood Geomorphology. Wiley, New York, pp. 65-77.
Jarrett, R.D., Costa, J.E. 1983. Multidisciplinary approach to the flood hydrology of foothill streams in
Colorado, In Johnson, A.I., and Clark, R.A. (eds.) International Syposium on Hydrometeorology,
Denver, Colo., 1982, Proceedings: Bethesda, Maryland, American Water Resources Association, p.
565-569.
Jarrett, R.D.,1987. Flood hydrology of foothill and mountain streams in Colorado. PhD dissertation,
Colorado State University, Ft. Collins, CO, 239 pp., unpubl.
Jeffries, R., S. E. Darby, D. Sear, A. 2003. The influence of vegetation and organic debris on flood-plain
sediment dynamics: Case study of a low-order stream in the New Forest, England, Geomorphology,
51, 61–80.
Julivert, M.; Fonboté, J.M.; Ribeiro, A. y Conde, L. 1972. Mapa tectónico de la Península Ibérica y Baleares.
E 1:1.000.000, Memoria explicativa, IGME, Madrid, 113 pp.
Kaplan, Garrick B. J. 1981. On the Quantitative Definition of Risk Risk Analysis, Volume 1, pp. 11-27.
Keller, E. A., Swanson, F. J. 1979. Effects of large organic material on channel form and fluvial processes,
Earth Surf. Processes, 4, 361–380.
Kennedy, J.F. 1975. Ice-jams mechanics. Proc. IAHR Symposium if Ice problems. Hanover.
Kite, G.W. 1988. Frequency and Risk Analyses in Hydrology, Water Resources Publications, Littleton,
Colorado, E.U.A. 187 pp.
Kochel, R.C., Baker, V.R. 1982. Paleoflood hydrology. Science 215: 353–361.
Kolerski, T. 2011. Numerical modeling of ice jam formation in the włocławek reservoir. Task quarterly 15
no 3–4, 283–295.
Krajewski, W.F., Smith, J.A. 2002. Radar hydrology–rainfall estimation. Advances in Water Resources 25,
1387–1394.
Kreibich, H. Piroth K., Seifert I., Maiwald, H.,. Kunert U, Schwarz, J., Merz, B., Thieken, A. H.. 2009. Is
flow velocity a significant parameter in flood damage modelling?. Nat. Hazards Earth Syst. Sci., 9,
1679–1692.
Krige, Danie G. 1951. A statistical approach to some basic mine valuation problems on the Witwatersrand.
J. of the Chem., Metal. and Mining Soc. of South Africa 52 (6): 119–139.
L.B. y Maddock, T. 1953. The hydraulic geometry of stream channels and some physiographic implications,
U.S. Geological Survey Professional Paper 252: 57 pp.
Lakshmi, V. 2005.Standing back looking forward: Role of satellite remote sensing in the prediction of
ungauged basins, Predictions in Ungauged Basins: International Perspectives on the State of the Art
and Pathways Forward, edited by S.Franks, , IAHS Publ.301, 118–124.
Lamb, M. P., W. E. Dietrich, S. M. Aciego, D. J. DePaolo, Manga, M. 2008a. Formation of Box Canyon,
Idaho, by megaflood: Implications for seepage erosion on Earth and Mars, Science, 320, 1067–1070.
Lamb, M. P., W. E. Dietrich, Sklar, L. S. 2008b. A model for fluvial bedrock incision by impacting
suspended and bed load sediment, J. Geophys. Res., 113,
Lamb, M. P., W. E. Dietrich, Venditti, J. G. 2008c.Is the critical Shields stress for incipient motion
dependent on channel-bed slope?, J. Geophys. Res., 113.
Lancaster, S. T., S. K. Hayes,. Grant, G. E. 2003. Effects of wood on debris flow runout in small mountain
watersheds, Water Resour. Res., 39(6).
Lane, S. N.1998. Hydraulic modelling in hydrology and geomorphology: A review of high resolution
approaches, Hydrol. Processes, 12, 1131–1150.
Langbein, W.B., Leopold, L.B. 1968. River channel bars and dunes theory of kinematic waves: USGS Prof.
Pap. 422-L..
Lange, D., Bezzola, G. R. 2006. Schwemmholz: Probleme und Lösungsansätze [Driftwood: Problems and
solutions]. VAWMitteilung 188, H.-E. Minor, ed., ETH Zürich, Zurich, 188.
Lenzi, M. A., L. Mao, Comiti, F. 2006. When does bedload transport begin in steep boulder bed streams?,
Hydrol. Processes, 20, 3517–3533.
Lenzi, M. A.,2004. Displacement and transport of marked pebbles, cobbles and boulders during floods in a
steep mountain stream, Hydrol. Processes, 18, 1899–1914.
Leopold, L. B., Miller, J. P. 1956. Ephemeral streams - hydraulic factors and their relation to the drainage
net, U.S. Geol. Surv. Prof. Pap.282-A, 1–37.
Leopold, L.B., Langbein, W.B. 1966. River Meanders. Scientific American, 214 (6), 60-70.
Leopold, L.B.; Wolman, M.G. y Miller, J.P. 1964. Fluvial Processes in Geomorphology. V.H. Freeman, San
Francisco, 522 pp.
Lewin, J. 1989. Floods in Fluvial Geomorphology. En: K. Beven y P. Carling (Eds.), Floods: Hydrological,
Sedimentological and Geomorphological Implications, 265-284, John Wiley & Sons, Chichester.
Lienkaemper, G. W., F. J. Swanson, 1987. Dynamics of large woody debris in streams in old-growth
Douglas-fir forests, Can. J. For. Res., 17, 150–156.
Liniger H, Gichuki Fn, Kironchi G, Njeru L. 1998. Pressure on the Land: The Search for Sustainable Use
in a Highly Diverse Environment [online]. Eastern and Southern Africa Geographical Journal
(ESAG), Vol. No. 8, Special Number. Available from:
http://www.cde.unibe.ch/CDE/pdf/ESAG_Journal.pdf.
Lisle, T. E. 1986. Stabilization of a gravel channel by large streamside obstructions and bedrock bends,
Jacoby Creek, northwestern California, Geol. Soc. Am. Bull., 97, 999–1011.
Lisle, T. E. 1987. Channel morphology and sediment transport in steepland streams, in Erosion and
Sedimentation in the Pacific Rim, Proceedings of the Corvallis Symp, Aug. 1987, edited by R.
L.Beschta, , IAHS Publ.165, 287–297.
Llasat, M.C., M. Barnolas, M. Ceperuelo, M. Llasat, Prat, M.A. 2004. Algunos aspectos del impacto social
de las inundaciones en Cataluña. Revista del Aficionado a la Meteorología, RAM 20.
www.meteored.com/ram/
Lumbroso, D., Gaume, E. 2012. Reducing the uncertainty in indirect estimates of extreme flash flood
discharges. Journal of Hydrology: 414-415. DOI:10.1016/j.jhydrol.2011.08.048.
Lyell, C. 1830. Principles of Geology. Being an Attempt to Explain the Former Changes of the Earth’s
Surface, by Reference to Causes now in Operation, vol. 1. John Murray, London. 512 pp.
Lyn D, Cooper T, Condon D, Gan L 2007. Factors in debris accumulation at bridge piers, Washington, US
Department of Transportation, Federal Highway Administration Research and Development,
Turner- Fairbank Highway Research Center.
Magilligan, F. J. 1992. Thresholds and the spatial variability of flood power during extreme floods,
Geomorphology, 5, 373–390.
Magilligan, F.J., Phillips, J.D., James, L.A. y Gómez, B. 1998. Geomorphic and Sedimentological Controls
on the Effectiveness o fan Extreme Flood. The Jorunal of Geology, 106: 87-95.
Magirl, C. S., J. W. Gartner, G. M. Smart, Webb, R. H. 2009.Water velocity and the nature of critical flow
in large rapids on the Colorado River, Utah, Water Resour. Res., 45.
MAGRAMA, 2011. http://www.magrama.gob.es/es/biodiversidad/temas/conectividad-ecologica-en-el-
territorio/carta_espaniola_montanias_julio05_tcm7-19550.pdf
Mao L. Surian N. 2010. Observations on sediment mobility in a large gravel-bed river. Geomorphology,
114, 326–337.
Mao, L., A. Andreoli, F. Comiti, Lenzi, M. A. 2008a. Geomorphic effects of large wood jams on a sub-
Antarctic mountain stream, River Res. Appl., 24, 249–266.
Mao, L., F. Comiti, A. Andreoli, L. Picco, M. A. Lenzi, A. Urciulo, R. Iturraspe, A. Iroumè, 2008b. Role
and management of in-channel wood in relation to flood events in Southern Andes basins,
Monitoring, Simulation, Prevention and Remediation of Dense and Debris Flow II, edited by D.de
Wrachien, C. A.Brebbia, M. A.Lenzi, WIT Trans. Eng. Sci.60, 207–216.
Mao, L., G. P. Uyttendaele, A. Iroumé,. Lenzi, M. A. 200 c. Field based analysis of sediment entrainment in
two high gradient streams located in Alpine and Andine environments, Geomorphology, 93, 368–
383.
Mao, L., Lenzi, M. A. 2007. Sediment mobility and bedload transport conditions in an alpine stream,
Hydrol. Processes, 21, 1882–1891.
Marchand, J.P., Jarrett, R.D., Jones, L.L. 1984. Velocity profile, water-surface slope, and bed-material size
for selected streams in Colorado.
Marchi, L., Borga, M., Preciso, E., and Gaume, E. 2010.Characterisation of selected extreme flash floods in
Europe and implications for flood risk management, J. Hydrol., 394, 118–133.
Marchi, L., Borga, M., Preciso, E., Sangati, M., Gaume, E., Bain, V., Delrieu, G., Bonnifait, L., Pogacˇnik, N.
2009. Comprehenive post-event survey of a flash flood in Western Slovenia: observation strategy
and lessons learned. Hydrological Processes 23 (26), 3761–3770.
Marks K, Bates PD. 2000. Integration of high resolution topographic data with floodplain flow models.
Hydrological Processes 14: 2109–2122
Marquínez, J., J. Lastra y P. García. 2003. Estimation models for precipitation in mountainous regions: the
use of GIS and multivariate analysis. Journal of Hydrology 270: 1-11.
Marquínez, J., Lastra. J., García, P. 2003: Estimation models for precipitation in mountainous regions: the
use of GIS and multivariate analysis, Journal of Hydrology 270, 1 – 11.
Marquínez, J.; Lastra, J. Y Fernández, E. 2006. Metodología utilizada para cartografiar la 25. Peligro,
vulnerabilidad y riesgo de inundación en ramblas mediterráneas Cuad. de Geogr. 83, 2008
peligrosidad de inundaciones en las cuencas del Norte. En: Díez Herrero, A.; Lain Huerta, L. y
Llorente Isidro, M. (Eds.). Mapas de peligrosidad de avenidas e inundaciones. Métodos, experiencias
y aplicación. Publicaciones del Instituto Geológico y Minero de España. Serie: Medio. Ambiente.
Riesgos Geológicos 7, 125-141. Madrid.
Martín Duque, J.F. 2000. La información geomorfológica en el contexto de los inventarios ambientales.
Mapas fisiográficos para la gestión territorial. Bol. R.Soc. Esp. Hist. Nat. (Sec. Geol.) 96 1-2: 33-46.
Martin, D.J., Benda, L.E. 2001. Patterns of instream wood recruitment and transport at the watershed
scale. Transactions of the American Fisheries Society 130: 940-958.
Martínez de Pisón, E. Muñoz Jiménez, J.1972. Observaciones sobre la morfología del Alto Gredos. Estudios
Geográficos. Madrid.
Martínez de Pisón, E. y Muñoz, J. 1972. Observaciones sobre la morfología del Alto Gredos. Estudios
Geogr., XXXIII, 129, 597-690.
Martín-Vide J., Barriendos M. 1995. The use of rogation ceremony records in climatic reconstruction. A
case study from Catalonia Spain. Climatic Change 30, 201-221.
Matheron, G. 1962. Traité de géostatistique appliquée. Editions Technip.
May, C.L., Gresswell, R.E. 2003. Large wood recruitment and redistribution in headwater streams in the
southern Oregon Coast Range, U.S.A. Canadian Journal of Forest Research 33: 1352.
Mazzorana, B., Comiti, F., Fuchs, S. 2011a. A structured approach to enhance flood hazard assessment in
mountain streams. Natural Hazards. DOI:10.1007/s11069-011-9811-y.
Mazzorana, B., Comiti, F., Volcan, C.,Scherer, C. 2011b. Determining flood hazard patterns through a
combined stochastic–deterministic approach. Natural Hazards 59: 301-316. DOI:10.1007/s11069-
011-9755-2.
Mazzorana, B., Hübl, J., Fuchs, S. 2009. Improving risk assessment by defining consistent and reliable
system scenarios. Natural Hazards and Earth System Sciences 9 (1), 145e159.
Mazzorana, B., Hübl, J., Zischg, A., Largiader, A. 2010. Modelling woody material transport and deposition
in alpine rivers. Natural Hazards..
McAlpine, Robert G. 1961. Yellow-poplar seedlings intolerant to flooding, J. Forest,59: 566-568.
McCuen, R.H. 1982. A guide to Hydrological Analysis Using SCS methods. Soil Conservation Service.
Prentice-Hall, Inc. Englewood Cliffs, New Yersey, USA.
Mcdade, M.H., F.J. Swanson, W.A. Mckee, J.F. Franklin, And J. Van Sickle. 1990. Source distances for
coarse woody debris entering small streams in western Oregon and Washington. Can. J. For.
Res.20:326–330.
McDonald, B. C., Banerjee, I. 1971. Sediments and bed formson a braided outwash plain: Canadian
Journal of Earth Sciences, v. 8, p. 1282-1301.
Meadows, D. L., Behrens, W. W., Meadows, D. H., Naill, R. F., Randers, J., and Zahn, E., 1974. Dynamics
of growth in a finite world. Wright-Allen, Inc. 238 Main Street, Cambridge. Massachusetts
Merten, E., Finlay, J., Johnson, L., Newman, R., Stefan, H., Vondracek, B. 2010. Factors influencing wood
mobilization in streams. Water Resources Research 46, 1-13.
Merz B., Hall J., Disse M., Schumann A. 2010. Fluvial flood risk management in a changing world. Natural
Hazards Earth System Science, 10 : 509–527,
Merz, B., Apel, H., Gocht, M. 2004. Entwicklung eines probabilistischen Ansatzes zur Bestimmung des
Deichversagens durch Überströmen. in Merz, B. andApel, H. (eds.), Risiken durch Naturgefahren,
Scientific Technical Report STR04/01. GeoForschungsZentrum Potsdam, Potsdam, pp. 77–84.
Merz, B., Kreibich, H., Apel, H. 2008. Flood risk analysis: uncertainties and validation, Östereichische
Wasserwirtschaft (ÖWAWI), 05-06, 1-6.
Merz, B., Thieken, A. H., y Blöschl, G. 2002. Uncertainty analysis for Flood Risk Estimation. En:
International Conference on Flood Estimation, Spreafico, M. and Weingartner, R., (Eds.,) II-17,
577–585.
Merz, B., Thieken, A.H. 2004. Flood Risk Analysis: Concepts and Challenges, OsterreichischeWasser und
Abfallwirtschaft, 56(3–4), 27–34.
Merz, B., Thieken, A.H., Gocht, M., Flood risk mapping at the local scale: Concepts and challenges. In:
Flood Risk Management in Europe: Innovation in Policy and Practice (Begum, S., Stive, M.J.F.,
Hall, J.W.; Eds.). Series: Advances in Natural and Technological Hazards Research, 25, Springer,
Dordrecht, Chapter 13, 231‐251, 2007.
Merz, B., y Thieken, A.H. 2005. Separating natural and epistemic uncertainty in flood frequency analysis.
Journal of hydrology, 309: 114-132.
Merz, R. and Bloschl, G., Flood frequency regionalisation‐spatial proximity vs. catchment attributes,
Journal of Hydrology, 302, 283–306, 2005.
Messerli, B., J. D. Ives, 1984.Mountain Ecosystems: Stability and Instability, Int. Mtn. Soc., Boulder, Colo.,
Meyer, G. A. 2001. Recent large-magnitude floods and their impact on valley-floor environments of
northeastern Yellowstone, Geomorphology, 40, 271–290,
Meyer, V. 2007 GIS-based Multicriteria Analysis as Decision Support in Flood Risk Management.
UFZ Discussion Papers 6/2007. Leipzig.
Mileti, D.S. 1999. Disasters by Design: A Reassessment of Natural Hazards in the United States. Joseph
Henry Press. Washington D.C.
Miller, A. J. 1990a. Flood hydrology and geomorphic effectiveness in the central Appalachians, Earth Surf.
Processes Landforms, 15, 119–134.
Miller, A. J. 1990b. Fluvial response to debris associated with mass wasting during extreme floods, Geology,
18, 599–602.
Mitchell J. 2003. European River Floods in a Changing World. Risk Analysis 23.
Montgomery, D. R., B. D. Collins, J. M. Buffington, T. B. Abbe, 2003a. Geomorphic effects of wood in
rivers, The Ecology and Management of Wood in World Rivers, edited by S. V.Gregory, K. L.Boyer,
A. M.Gurnell, 21–47, Am. Fish. Soc., Bethesda, Md.
Montgomery, D. R., Buffington, J. M. 1997. Channel-reach morphology in mountain drainage basins,
Geol. Soc. Am. Bull., 109, 596–611.
Montgomery, D. R., G. E. Grant, Sullivan, K. 1995b. Watershed analysis as a framework for implementing
ecosystem management, Water Resour. Bull., 31, 369–386.
Montgomery, D. R., J. M. Buffington, N. P. Peterson, D. Schuett-Hames, Quinn, T. P. 1996b. Stream-bed
scour, egg burial depths, and the influence of salmonid spawning on bed surface mobility and
embryo survival, Can. J. Fish. Aquat. Sci., 53, 1061–1070.
Montgomery, D. R., J. M. Buffington, R. D. Smith, K. M. Schmidt, G. Pess, 1995a. Pool spacing in forest
channels, Water Resour. Res., 31, 1097–1105.
Montgomery, D. R., T. B. Abbe, 2006. Influence of logjam-formed hard points on the formation of valley-
bottom landforms in an old-growth forest valley, Queets River, Washington, USA, Quat. Res., 65,
147–155.
Montgomery, D. R., T. B. Abbe, J. M. Buffington, N. P. Peterson, K. M. Schmidt, J. D. Stock,
1996a.Distribution of bedrock and alluvial channels in forested mountain drainage basins, Nature,
381, 587–589.
Montgomery, D. R., T. M. Massong,. Hawley, S. C. S. 2003b. Influence of debris flows and log jams on the
location of pools and alluvial channel reaches, Oregon Coast Range, Geol. Soc. Am. Bull., 115, 78–
88.
Montgomery, D.R. and Piégay, H. 2003: Editorial: wood in rivers: interactions with channel morphology
and processes. Geomorphology 51, 1–5.
Montgomery, D.R., and J.M. Buffington, 1998, Channel processes, classification, and response, in River
Ecology and Management, edited by R. Naiman and R. Bilby, Springer-Verlag, New York, NY, pp.
13-42.
Montgomery, D.R., Buffington, J.M., Smith, R.D., Schmidt, K.M., Pess, G.R. 1995. Pool frequency in
forest channels. Water Resources Research 31, 1097– 1105.
Morisawa, M. E. 1962. Quantitative geomorphology of some watersheds in the Appalachian Plateau, Geol.
Soc. Am. Bull., 73, 1025–1046.
Mueller, E. R., J. Pitlick,. Nelson, J. M 2005. Variation in the reference Shields stress for bed load transport
in gravel- bed streams and rivers, Water Resour. Res., 41.
Muñoz Jiménez, J., Sanz Herráiz, C.: Las Montañas. Guía Física de España. Alianza Editorial. Madrid,
1995. ISBN 84-206-0750-9.
Naka, K. 1982. Community dynamics of evergreen broadleaf forests in southwestern Japan. I. Wind
damaged trees and canopy gaps in an evergreen oak forest. Journal of Plant Research 95: 385-399.
Nakamura, F., Swanson, F. J., Wondzell, S. M. 2000. Disturbance regimes of stream and riparian systems ? a
disturbance-cascade perspective. Hydrological Processes 14: 2849-2860.
Nakamura, F., Swanson, F.J. 1993. Effects of coarse woody debris on morphology and sediment storage of a
mountain stream system in western Oregon. Earth Surface Processes and Landforms 18, 43– 61
Nanson, G.C. 1986. Episodes of vertical accretion and catastrophic stripping: a mode of disequilibrium
floodplain development. Bull. Geol. Soc. Am., 97: 1467-1475.
NERC 1975. Flood studies report, v.1-5, 1100 págs., Hydrologic Studies, Whitefriars LTD, London.
Nicol, S.J., Lieschke, J.A., Lyon, J.P. and Koehn, J.D. 2004. Observations on the distribution and
abundance of carp and native fi sh, and their responses to a habitat restoration trial in the Murray
River, Australia. New Zealand Journal of Marine and Freshwater Research 38, 541–551.
Norbiato, D., Borga, M., Degli Esposti, S., Gaume, E., Anquetin, S. 2008. Flash flood warning based on
rainfall thresholds and soil moisture conditions: An assessment for gauged and ungauged basins.
Journal of Hydrology 362 (3-4), 274–290.
O. Shearman, 1976. U.S. Department of the Interior, Geological Survey, Water Resources Division, Open-
File Report 76-499, Computer Applications for Step-Backwater and Floodway Analysis, Program E-
431, J
O’Connor, D., H. Renting, M. Gorman and J. Kinsella 2006 The Evolution of Rural Development in
Europe and the Role of EU Policy. In O’Connor, D., H. Renting, M. Gorman and J. Kinsella (Eds.)
Driving Rural Development: Policy and Practice in Seven EU Countries. Assen: Van Gorcum
Oberguggenberger, M. 2005. The mathematics of uncertainty: models, methods and interpretations. In
Analyzing Uncertainty in Civil Engineering. Springer Berlin Heidelberg, 51-72
O'Connor, J. E., J. E. Costa, 2004. Spatial distribution of the largest rainfall-runoff floods from basins
between 2.6 and 26,000 km2 in the United States and Puerto Rico, Water Resour. Res., 40.
O'Connor, J. E., R. H. Webb, Baker, V. R. 1986. Paleohydrology of pool-and-riffle pattern development:
Boulder Creek, Utah, Geol. Soc. Am. Bull., 97, 410–420.
O'Connor, J. E., Webb, R. H. 1988. Hydraulic modeling for paleoflood analysis, Flood Geomorphology,
edited by V. R.Baker, R. C.Kochel, P. C.Patton, 393–402, John Wiley, New York.
Ogden, F.L.; Sharif, H.O.; Senarath, S.U.S.; Smith, J.A.; Baeck, M.L. Y Richardson, J.R. 2000. Hydrologic
analysis of the Fort Collins, Colorado, flash flood of 1997. Journal of Hydrology, nº 228, 82-100.
Olcina Cantos, J. 2004. Riesgo de inundaciones y ordenación del territorio en la escala local. El papel del
planeamiento urbano municipal. Boletín de la Asociación de Geógrafos Españoles.
Olcina Cantos, J. 2006. ¿Riesgos Naturales? I. Sequías e inundaciones. Editorial DaVinci Continental.
Colección Geoambiente XXI. Barcelona, 220 p.
Olcina Cantos, J. 2006. ¿Riesgos Naturales? II. Huracanes, sismicidad y temporales. Editorial DaVinci
Continental. Colección Geoambiente XXI. Barcelona, 205 p.
Olcina Cantos, J. 2009. Hacia una ordenación sostenible de los territorios de riesgo en Europa y toma de
decisiones (Farinós, J. Romero, J. y Salom. J. (eds.)) Publicaciones de la Universitat de Valencia,
Valencia, pp. 153-182.
ONU 2004. Living with Risk: A GlobalOrtega JA, Garzón-Heydt G. 2009. Geomorphological and
sedimentological analysis of flash-flood deposits. The case of the 1997 Rivillas flood (Spain).
Geomorphology 112: 1-14. doi:10.1016/j.geomorph.2009.05.004.
Ortega, 2007. Paleocrecidas, avenidas recientes e hidroclimatología en la cuenca media y baja del río
Guadiana. Tesis Doctoral. Universidad Complutense de Madrid.
Ortega, J.A. y Garzón, G. 2006. Interpretación de los depósitos de avenida como clave para establecer la
dinámica de la llanura de inundación. En: A. Pérez Alberti y J. López Bedoya (Eds.), Geomorfología
y territorio. Actas de la IX Reunión Nacional de Geomorfología, Universidade de Santiago de
Compostela, 629-644
Osterkamp, W.R. y Hupp, C.R. 1987. Dating and interpretation of debris flows by geological and botanical
methods at Whitney Creek gorge, Mount Shasta, California. Geol. Soc. Am. Rev. Eng. Geol. 7, 157-
163.
Pappenberger, F., H. Harvey, K. Beven, J. Hall, R. Romanowicz, Smith, P. 2006. Implementation Plan for
library of tools for uncertainty evaluation, Flood Risk Management Research Consortium.
Parker, G., Klingeman P.C., McLean D. 1982. Bedload and Size Distribution in paved gravel-bed streams .
Journal of the Hydraulics Division. Vol. 108, No. HY4 : 544-571 pp
Paté-Cornell, M. E. 1996b. Uncertainties in risk analysis: six levels of treatment. Reliability Engineering
and System Safety, Vol. 54, pp. 95-111.
Patton, P. C. 1988a. Drainage basin morphometry and floods, Flood Geomorphology, edited by V.
R.Baker, R. C.Kochel, P. C.Patton, 51–64, John Wiley, New York.
Patton, P. C. 1988b. Geomorphic response of streams to floods in the glaciated terrain of southern New
England, Flood Geomorphology, edited by V. R.Baker, R. C.Kochel, P. C.Patton, 261–277, John
Wiley, New York.
Patton, P. C., Baker, V. R. 1976. Morphometry and floods in small drainage basins subject to diverse
hydrogeomorphic controls, Water Resour. Res., 12, 941–952.
Payrastre, O., E. Gaume, Andrie, H. 2005. Use of historical data to assess the occurrence of floods in small
watersheds in the mediterranean area, Adv. Geosci., 2, 313–320.
Payrastre, O., Gaume, E., and Andrieu, H. 2005. Use of historical data to assess the occurrence of floods in
small watersheds in the French mediterranean area, Adv. Geosci.
Pedraza, J. 1994. Periglaciarismo del Sistema Central español. En: A. Gómez Ortiz, M. Simón y F. Salvador
(eds.), Periglaciarismo en la Península Ibérica, Canarias y Baleares. Estudios significativos,
Monografías de la S.E.G., 7, 147-164, Universidad de Granada, Granada.
Pedraza, J. 1996. Geomorfología: Principios, Métodos y Aplicaciones. Ed. Rueda, Madrid, 414 pp.
Rickenmann, D. 1991. Hyperconcentrated flow and sediment at steep slopes, en J. of Hydr. Eng. 117:
1419-1439.
Rickenmann, D., Koschni, A. 2010. Sediment loads due to fluvial transport and debris flows during the
2005 flood events in Switzerland. Hydrological Processes 24: 993-1007.
Rico, M. 2004. Las paleocrecidas en la cuenca media del Río Segre durante el Pleistoceno superior-
Holoceno: registros morfosedimentarios y análisis hidroló- gico. PhD thesis, University of Zaragoza,
Spain.
Rico, M., Benito, G., Barnolas, A. 2001. Combined palaeoflood and rainfall–runoff assessment of
mountain floods (Spanish Pyrenees). Journal of Hydrology, 245(1-4), 59-72.
Rinntech, 2010. TSAP-WinTM. User Reference, Rinntech ®, 91 pp.
Ritter, D. F., R. C. Kochel, Miller, J. R. 1995. Process Geomorphology, 3, 546, Wm. C. Brown, Dubuque,
Iowa,
Roberson, J:A., Crowe, C.T. 1993. Fluid mechanics. Houghton Mifflin (Boston). 746 pp. ISBN
0395637899.
Robison, E. G., Beschta, R. L. 1990. Coarse woody debris and channel morphology interactions for
undisturbed streams in southeastern Alaska, USA, Earth Surf. Processes Landforms, 15, 149–156,
Robson A, Reed D. 1999. Flood Estimation Handbook. Volume 3. Statistical Procedures for Flood
Frequency Estimation. Institute of Hydrology: Wallingford
Rodriguez-Iturbe, I. 1993.The geomorphological unit hydrograph, Channel Network Hydrology, edited by
K.Beven, M. J.Kirkby, 43–68, John Wiley, Chichester, U. K.,
Romanowicz R. J., Beven, K. J. 2003. Estimation of flood inundation probabilities as conditioned on event
inundation maps, Water Resources Research, VOL. 39, NO. 3,
Ruiz-Villanueva, V., M. Borga, D. Zoccatelli, L. Marchi, E. Gaume, And U. Ehret. 2012c. Extreme flood
response to short-duration convective rainfall in South-West Germany. Hydrology and Earth
System Science, 16, 1543–1559.
Ruiz-Villanueva,V.; Thorndycraft, V.; Díez-Herrero, A. 2010. Análisis dendrogeomorfológico y registro de
paleoinundaciones en la cuenca del río Till (NE Inglaterra): resultados preliminares. En: X.
ÚBEDA, D. VERICAT, R.J. BATALLA (Eds), Avances de la Geomorfología en España, 2008-
2010. pp. 521-524. XI Reunión Nacional de Geomorfología, 20-24 septiembre de 2010. UB, CTFC,
ULl y SEG, Solsona (Lérida). D.L.: B-61.451-2010; ISBN: 978-84-693-4551-1.
Russell, I.C. 1909. Rivers of North America. G.P. Putnam’s Sons, New York, NY. 522 pp.
Ruth, Matthias and Hannon, Bruce. 1997. Modeling Dynamic Economic Systems. New York: Springer-
Verlag.
Salas L, Fernandez T. 2007. In-site regionalization to estimate an intensity– duration–frequency law: a
solution to scarce spatial data in Spain. Hydrological Processes 21:3507–3513.
Sangati, M., Borga, M., Rabuffetti, D., Bechini, R. 2009. Influence of rainfall and soil properties spatial
aggregation on extreme extreme flash flood response modelling: an evaluation based on the Sesia
river basin, North Western Italy. Advances Water Resources 32 (7), 1090–1106.
Sant George, 2003. Paleoflood records for the Red River, Manitoba, Canada derived
Sauvageot H. Radar Meteorology. Artech House, Inc; 1991. p. 315.
Schmocker, L., Hager, W.H. 2010. Drift accumulation at river bridges. River Flow Dittrich, Koll, Aberle &
Geisenhainer (eds). Bundesanstalt für Wasserbau ISBN 978-3-939230-00-7.
Scholz, R.W. and Tietje, O. 2002. Embedded Case Study Methods: Integrating Quantitative and
Qualitative Knowledge. Sage Publications.
Schulze, E.-D., F. S. Chapin III, and G. Gebauer. 1994. Nitrogen nutrition and isotope differences among
life forms at the northern treeline of Alaska. Oecologia 100:406–412.
Schumm, S. A. 1956. The role of creep and rainwash on the retreat of badland slopes, Am. J. Sci., 254, 693–
706.
Schumm, S. A. 1966The development and evolution of hillslopes, J. Geol. Educ., 14, 98–104,.
Schumm, S. A.1963a. A tentative Classification of Alluvial River Channels, U.S. Geol. Surv. Circ.477, 10.
Schumm, S. A.1963b The Disparity Between Present Rates of Denudation and Orogeny, U.S. Geol. Surv.
Prof. Pap.454-H, 13.
Schumm, S.A. 1977. The Fluvial System, John Willey & Sons, N. York, 338 págs.
Sear DA. 1994. River restoration and geomorphology. Aquatic Conservation 4: 169–177.
Sedell, J. R., J. E. Richey, F. J. Swanson, 1989.The river continuum concept: A basis for expected ecosystem
behavior of very large rivers?, Can. Spec. Publ. Fish. Aquat. Sci., 106, 110–127.
Serrar S., Delrieu G., Creutin J.D., Uijlenhoet R. 2000, Mountain Reference Technique: the use of
mountain returns to calibrate weather radar systems operating at attenuating wavelengths. J. of
Geophys. Res. Vol 105, No D2, 2281-2290.
Shaban, A., G. Faour, M. Khawlie, Abdallah, C. 2004.Remote sensing application to estimate the volume of
water in the form of snow on Mount Lebanon, Hydrol. Sci. J., 49, 643–653,
Shaler, N. S., 1891, The origin and nature of soils, in Powell, J. W., ed., USGS 12th Annual report 1890-
1891: Washington, D.C., Government Printing Office
Shaw, J., and R. Kellerhals (1977), Paleohydraulic interpretation of antidune bedforms with applications to
antidunes in gravel, Journal of Sedimentary Petrology, 47, 257-266.
Shields A 1936 Application of similarity principles and turbulence research to bed-load movement.
Mitteilunger der Preussischen Versuchsanstalt f¨ur Wasserbau und Schiffbau 26: 5–24
Shoecraft, R.P., 1875. Map of Township No. 31 North, Range No. 5 East, Willamette Meridian,
Washington Territory. U.S. Bureau of Land Management, Plates of Washington Territory,
Olympia, WA.
Shreve, R. L. (1969) Stream lengths and basin areas in topological!)' random channel networks../. Geol. 77,
397-414.
Shroder, J.F.1980. Dendrogeomorphology: review and new techniques of treering dating. Progress in
Physical Geography 4 (1): 161-188.
Shvidchenko, A., and Pender, G. 2000. “Flume study of the effect of relative depth on the incipient motion
of coarse uniform sediments.” Water Resour. Res., 362, 619–628.
Sidle, R. C. 1988. Bed load transport regime of a small forest stream, Water Resour. Res., 24, 201–218.
Sidle, R. C., S. Noguchi, Y. Tsuboyama, Laursen, K. 2001.A conceptual model of preferential flow systems
in forested hillslopes: Evidence of self-organization, Hydrol. Processes, 15, 1675–1692.
Sidle, R.C.; Pearce, A.J.; O’loughlin, C.L., 1985: Hillslope stability and land use. Water Resources
Monograph 11: 140 pp.
Sigafoos R.S. 1961. Vegetation in relation to flood frequency near Washington, D. C.: U. S. Geological
Survey Professional Paper 424-C, p. 248-249.
Sigafoos R.S. 1964. Botanical Evidence of Floods and Flood-Plain Deposition. U. S. Geological Survey
Professional Paper, 485A. 35 pp.
Sivapalan, M., K. Takeuchi, S.W. Franks, V.K. Gupta, H. Karambiri, V. Lakshmi, X. Liang, J.J. McDonnell,
E.M. Mendiondo, P.E. O’Connell, T. Oki, J.W. Pomeroy, D. Schertzer, S. Uhlenbrook and E. Zehe,
2003. AISH Decade on Predictions in Ungauged Basins (PUB), 2003-2012: Shaping an exciting
future for the hydrological sciences. Hydrological Sciences — Journal — des Sciences Hydrologiques
48: 857-880.
Smith, D. I. 1994. Flood Damage Estimation: A Review of Urban Stage-Damage Curves and Loss
Functions, Water SA, vol. 20, pp. 231-238.
Smith, J. D. 2004. The role of riparian shrubs in preventing floodplain unraveling along the Clark Fork of
the Columbia River in the Deer Lodge Valley, Montana, Riparian Vegetation and Fluvial
Geomorphology, edited by S. J.Bennett, A.Simon, 71–85, AGU, Washington, D. C.
Smith, J.B., N. Bhatti, G. Menzhulin, R. Benioff, M.I. Budyko, M. Campos, B. Jallow, and F. Rijsberman
(eds.), 1996: Adapting to Climate Change: An International Perspective. Springer- Verlag, New
York, NY, USA, 475 pp.
Smith, S. M.,. Prestegaard, K. L 2005. Hydraulic performance of a morphology-based stream channel
design, Water Resour. Res., 41.
Soil Conservation Service. 1972. Hydrology. In: National Engineering Handbook, Section 4. USDA Soil
Conservation Service, Washington DC, USA.
Spreafico, M , Schadler, B. & Sigrist, B. 1985. Gesehwindigkeitsverleilung in schweizerischen
Fliessgewiissern. Wasser. Energie, Luft, vol. 1/2, 77. Jahrgang J. Schweiz. Wasserwirtschal'tsverband,
CH-5401 Baden.
Spreafico, M, Hodel, H.P, Kaspar, H. 2001. Rauheiten in ausgesuchten schweizerischen Fliessgewässern.
Berichte des BWG, Serie Wasser Nr.1, Bern.
Spreafico, M., Weingartner, R., Barben, M., Ryser, A. 2003. Evaluation des crues dans les bassins versants de
Suisse - Guide pratique. Rapports de l'OFEG, Série Eaux No 4, Berne.
Stedinger, J.R., Baker, V.R. 1987. Surface water hydrology: historical and paleoflood information. Reviews
of Geophysics. Vol. 25, 119-124
Stedinger, J.R., Cohn, T.A. 1986. The value of historical and paleoflood information in flood frequency
analysis. Water Resour. Res. 22(5), 785-793.
Stedinger, J.R.. Vogel, R.M, Foufoula-Georgiou, E. 1993. Frequency analysis of extreme events. In: D.
Maidment (Editor), Handbook of Applied Hydrology. Mc-Graw Hill Book Co., New York,
Chapter 18.
Stoffel, M. Bollschweiler,M.,. Butler,D.,. Luckman B.2010. Tree Rings and Natural Hazards: A State-of-
Art. Springer.
Stoffel, M., Bollschweiler, M. 2008. Tree-ring analysis in natural hazards research: an overview. Natural
Hazards and Earth System Sciences 8, 187–202.
Strahler, A. N. 1964. Quantitative geomorphology of drainage basins and channel networks, Handbook of
Applied Hydrology, edited by V. T.Chow, 4-40–4-74, McGraw Hill, New York.
Stumbles, R.E. 1968. How wind influences silviculture and management as a district officer sees it. Forestry
Supplement, 45-50.
Sullivan, K., T. E. Lisle, C. A. Dolloff, G. E. Grant, Reid, L. M. 1987.Stream channels: The link between
forests and fishes, Streamside Management: Forestry and Fishery Implications, edited by E. O.Salo,
T. W.Cundy, 39–97, Univ. of Washington, Institute of Forest Resources, Seattle.
Surell, A. 1841. Etude sur les torrents des Hautes-Alpes, Tome 1. Dunod, Paris
Swanson, F.J. 2003. Wood in rivers: a landscape perspective. American Fisheries Society Symposium 37:
299-313.
Swanson, F.J., Lienkaemper, G.W. 1978. Physical Consequences of Large Organic Debris in Pacific
Northwest Streams. General Technical Report PNW-69, U.S.D.A. Forest Service, Pacific
Northwest Forest and Range Experiment Station, Portland, OR. 12 pp.
Swanson, F.J., Lienkaemper, G.W. 1984. Interactions among fluvial processes, forest vegetation, and
aquatic ecosystems, South Fork Hoh River, Olympic National Park. In: Starkey, E.E., Franklin, J.F.,
Matthews, J.W. (Eds.), Proceedings of the Second Conference on Scientific Research in the National
Parks. Oregon State Univ. Forest Research Lab. Publ., Corvallis, OR, pp. 30–34.
Swanston, D.N. 1991. Natural processes. In: Meehan, W.R. (Ed.), Influences of Forest and Rangeland
Management on Salmonid Fishes and Their Habitats. American Fisheries Society Special
Publication, vol. 19, pp. 139– 179. Bethesda, MD.
Takahashi, T., H. Nakagawa, Kuang, S. 1987.Estimation of debris flow hydrograph on varied slope bed,
Erosion and Sedimentation in the Pacific Rim, edited by R. L.Beschta, , IAHS Publ.165, 167–177.
Témez, J.R. 1987. Cálculo hidrometeorológico de caudales máximos en pequeñas y medianas cuencas
naturales. Textos de la Dirección General de Carreteras, Tecnología carreteras, MOPT, Madrid, pag.
var.
Témez, J.R. 1991. Extended and Improved Rational Method. Version of the Highways Administration o
Spain. Proc. XXIV I.A.H.R. Congress, Madrid (España), vol. A, 33-40.
Thomas, B. T. 1987. Spring shorebird migration through central Venezuela. Wilson Bull. 99:571- 578.
Thompson, D. M. 2006a. Changes in pool size in response to a reduction in discharge: A flume experiment,
River Res. Appl., 22, 343–351.
Thompson, D. M. 2006b. Did the pre-1980 use of in-stream structures improve streams? A reanalysis of
historical data, Ecol. Appl., 16, 784–796,
Thompson, D. M. 2006c. The role of vortex shedding in the scour of pools, Adv. Water Resour., 29, 121–
129.
Thorndycraft ,V., Ruiz-Villanueva, V., Palaeoflood hydrology and flood response to past climatic
variability: conceptual framework and preliminary results of the River Till case study (NE England).
British Hydrology Society Third International Symposium, Role of Hydrology in Managing
Consequences of a Changing Global Environment. Poster 19-23 July 2010. Newcastle University,
Newcastle upon Tyne, Inglaterra.
Thorndycraft V.R, Benito G., Rico M., Sánchez-Moya Y., Sopeña A. y Casas A. 2004. A Late Holocene
palaeoflood record from slackwater flood deposits of the Llobregat River, NE Spain. Journal
Geological Society of India 64: 549-559.
Thorndycraft, V. R., Benito, G., Llasat, M. C., and Barriendos, M. 2002. Palaeofloods, historical data and
climatic variability: applications in flood risk assessment, in: Palaeofloods, Historical Data and
Climatic Variability: Applications in Flood Risk Assessment, edited by: Thorndycraft, V. R., Benito,
G., Barriendos, M., and Llasat, M. C., CSIC, Madrid, Spain, 3–9,
Thorndycraft, V., Benito, G., Rico, M., Sopeña, A., Sánchez, Y & Casas, M. 2005. Palaeoflood hydrology of
the Llobregat River, NE Spain: a 3000 year record of extreme floods. Journal of Hydrology, 313 (1-
2), 16-31.
Thorndycraft, V.R., Benito, G., Barriendos, M., Llasat, M.C. 2003. Palaeofloods, Historical Data and
Climatic variability: Applications in Flood Risk Assessment. CSIC, Madrid. 378 pp.
Tietje, O. 2005. Identification of a small reliable and efficient set of consistent scenarios. Journal of Cleaner
Production, 162, 418-432.
Tinkler, K. J. 1997a.Critical flow in rockbed streams with estimated values for Manning's n,
Geomorphology, 20, 147–164.
Tinkler, K. J. 1997b. Indirect velocity measurement from standing waves in rockbed rivers, J. Hydraul.
Eng., 123, 918–921.
Trevisani, S., Cavalli, M., and Marchi, L. 2009. Variogram maps from LiDAR data as fingerprints of surface
morphology on scree slopes, Nat. Hazards Earth Syst. Sci., 9, 129–133.
Trieste, D. J. 1994. Supercritical flows versus subcritical flows in natural channels, Proceedings, ASCE
Hydraulic Engineering '94 Conference, August 1–5, Buffalo, New York, edited by G. V.Cotroneo,
R. R.Rumer, 732–736, Am. Soc. of Civ. Eng., New York.
Trieste, D. J., Jarrett, R. D. 1987.Roughness coefficients of large floods, Proceed., Conf. On Irrig. Systems
for the 21st Century, edited by L. G.James, M. J.English, 32–40, Am. Soc. of Civ. Eng., Portland,
Oreg..
Trieste, D.J. 1992.Evaluation of supercritical/subcritical flows in high- gradient channels, Journal of
Hydraulic Engineering, 118(8), 1107-1118.
Tung, Y.K., 1985. River flood routing by nonlinear Muskingum method. J. Hydraul. Engng. 111 (12),
1447–1460.
Turowski, J. M., D. Lague, Hovius, N. 2009a.nResponse of bedrock channel width to tectonic forcing:
Insights from a numerical model, theoretical considerations, and comparison with field data, J.
Geophys. Res., 114.
Turowski, J. M., E. M. Yager, A. Badoux, D. Rickenmann, Molnar, P. 2009. The impact of exceptional
events on erosion, bedload transport and channel stability in a step-pool channel, Earth Surf.
Processes Landforms, 34, 1661–1673.
US Army Corps of Engineers (USACE). 2009. HEC-HMS 3.5 User Manual, US Army Corps of
Engineers Hydrologic Engineering Center, Davis, CA.
US Army Corps of Engineers (USACE). 2010a. HEC-GeoHMS 4.2 User Manual, US Army Corps of
Engineers Hydrologic Engineering Center, Davis, CA.
US Army Corps of Engineers (USACE). 2010b. HEC-RAS River Analysis System: Hydraulic Reference
Manual, US Army Corps of Engineers Hydrologic Engineering Center, Davis, CA.
Van Sickle, J., Gregory, S. V. 1990. Modeling inputs of large woody debris from falling trees. Canadian
Journal of Forest Research 20: 1593–1601.
Versini, P.-A., Gaume, E., and Andrieu, H. 2010, Assessment of the susceptibility of roads to flooding based
on geographical information – test in a flash flood prone area (the Gard region, France), Nat.
Hazards Earth Syst. Sci., 10, 793–803.
Vianello, A., Cavalli, M., Tarolli, P. 2009. LiDAR-derived slopes for headwater channel network analysis.
Catena 76, 97–106.
Viglione, A. 2009. nsRFA: Non-supervised Regional Frequency Analysis, R package.
Vignal, B., H. Andrieu, and J. D. Creutin, 1999: Identification of vertical profiles of reflectivity from
volume scan radar data. J. Appl. Meteor., 38, 1214–1228.
Viviroli, D., R. Weingartner, Messerli, B. 2003. Assessing the hydrological significance of the world's
mountains, Mt. Res. Dev., 23, 32–40.
Viviroli, D., Weingartner, R. 2004. Hydrologische Bedeutung des europäischen Alpenraums.
Hydrologischer Atlas der Schweiz, Tafel 6.4, Bern.
Ward, J. V., K. Tockner, U. Uehlinger, Malard, F. 2001. Understanding natural patterns and processes in
river corridors as the basis for effective river restoration, Reg. Rivers Res. Manage., 17, 311–323.
Warren, D.R., and Kraft, C.E. 2008. Dynamics of large wood in eastern U.S. mountain stream. Forest
Ecology and Management. 256(808-814).
Wavemetrics,2012. http://www.wavemetrics.com/products/igorpro/dataaccess/gissupport.htm
Waythomas, C.F., Jarrett, R.D. 1994. Flood geomorphology of Arthurs Rock Gulch, Colorado, paleoflood
history, Geomorphology, 11, 15-40.
Webb RH, Jarrett RD. 2002. One-dimensional estimation techniques for discharges of paleofloods and
historical floods. In Ancient Floods, Modern Hazards: Principles and Applications of Paleoflood
Hydrology. Water Science and Application, House PK, Webb RH, Baker VR, Levish DR (eds).
American Geophysical Union: Washington, D.C.; Vol. 5: 111–126.
Webb, A.A., Erskine, W.D. 2003. A practical scientific approach to riparian vegetation rehabilitation in
Australia. Journal of Environmental Management 68: 329–341.
Weingartner, R., M. Barben, Spreafico, M. 2003.Floods in mountain areas: An overview based on examples
from Switzerland, J. Hydrol., 282, 10–24.
Whittaker, J. G., M. Jaeggi, N. R. 1982.Origins of step-pool systems in mountain streams, J. Hydraul. Div.
Am. Soc. Civ. Eng., 108, 758–773.
Williams, G.P., Costa, J.E. 1988. Geomorphic measurements after a flood», en BAKER, V.R.; KOCHEL,
R.C y PATTON, P.C. (eds.): Flood geomorphology, Wiley, New York, pp. 65-77.
Wohl, E. 2000. River Mountain. AGU, Water resource monograph, Washington DC. 78 pp.
Wohl, E. 2010. Mountain Rivers Revisited, Water Resour. Monogr. Ser., vol. 19, 573 pp.
Wohl, E. E. 1992a. Bedrock benches and boulder bars: Floods in the Burdekin Gorge of Australia, Geol.
Soc. Am. Bull., 104, 770–778.
Wohl, E. E. 1992b Gradient irregularity in the Herbert Gorge of northeastern Australia, Earth Surf.
Processes Landforms, 17, 69–84.
Wohl, E. E. 1995. Estimating flood magnitude in ungauged mountain channels, Nepal, Mt. Res. Dev., 15,
69–76.
Wohl, E. E., J. R. Goode, 2008.Wood dynamics in headwater streams of the Colorado Rocky Mountains,
Water Resour. Res., 44.
Wohl, E. E., Merritt, D. M. 2008.Reach-scale channel geometry of mountain streams, Geomorphology, 93,
168–185.
Wohl, E.E. 1992. Bedrock benches and boulder bars: Floods in the Burdekin Gorge of Australia. Geol. Sot.
Am. Bull., 104: 77& 778.
Wolman, M. G., and J. P. Eiler, 1958Reconnaissance study of erosion and deposition produced by the flood
of August 1955 in Connecticut, Trans. Am. Geophys. Union, 39, 1-14,
Wolman, M. G., Leopold, L. B. 1957. River flood plains: Some observations on their formation, U.S. Geol.
Surv. Prof. Pap.282-C.
Wolman, M.G. 1954. A method of sampling coarse river-bed material. Trans. Am. Geophys. Union, 35:
951-956.
Wolman, M.G., Gerson, R. 1978. Relative scales of time and effectiveness of climate in watershed
geomorphology. Earth Surf. Process., 3: 189-208.
Wolman, M.G., Miller, J.P. 1960. Magnitude and frequency of forces in geomorphic processes. J. Geol., 68:
54-74.
Wyżga, B., Zawiejska, J. 2005. Wood storage in a wide mountain river: Case study of the Czarny Dunajec,
Polish Carpathians, Earth Surf. Processes Landforms, 30, 1475–1494.
Yager, P.L., D.W.R. Wallace, K.M. Johnson, W.O. Smith, Jr., P.J. Minnett and J.W. Deming 1995. The
Northeast Water Polynya as an atmospheric CO2 sink: A seasonal rectification hypothesis. Journal
of Geophysical Research 100(C3): 4389-4398.
Yanosky, T y Jarret, R. 2002. Dendrochronologic Evidence for the Frequency and Magnitud of Palofloods.
En: Ancient Floods, Modern Hazards: Principles and Applications of Paleoflood Hydrology. House
PK, Webb RH, Baker VR, Levish DR (Eds) Water Science and Application Vol. 5, American
Geophysical Union: Washington, D.C.77-89.
Yanosky, T.M. 1983. Evidencs of flood on Potomac River from anatomical abnormalities in the wood of
flood plain trees. U.S. Geological Survey Professional Paper 1296: 1-42.
Zanon, F., Borga, M., Zoccatelli, D., Marchi, L., Gaume, E., Bonnifait, L., Delrieu, G., 2010. Hydrological
analysis of a flash flood across a climatic and geologic gradient: the September 18, 2007 event in
Western Slovenia. J. Hydrol. 394, 182–197, doi:10.1016/j.jhydrol.2010.08.020.
Zhu, T. and J.R. Lund, “Up or Out? - Economic-Engineering Theory of Flood Levee Height and. Setback,”
Journal of Water Resources
Zimmerman I 1980. Factors influencing the feed intake and live weight change of beef cattle on beef cattle
on a mixed tree savannah in the Transvaal. J. Range Manage., p. 33.
Zimmerman, A., Church, M. 2001. Channel morphology, gradient profiles and bed stresses during flood in
a step-pool channel, Geomorphology, 40, 311 – 327,
Zuhal, A., S. Aynur, S. Arda, Unal, S. A. 2007. Calibration/validation of satellite derived snow products
with in situ data over the mountainous eastern part of Turkey, Remote Sensing for Environmental
Monitoring and Change Detection, edited by M.Owe, C. M.Neale, IAHS Publ.316, 157–168.
Zuna J. 2008. Hrazení bystřin (in Czech). Česká technika –nakladatelství ČVUT, Praha. 180 p.
Figure 1.2: Stylized and simplified sketch of the three primary zones of a mountain
drainage basin based on hydrogeomorphic processes. Esquema idealizado y
simplificado de las diferentes zonas en que puede dividirse una cuenca hidrológica de
montaña en función de los procesos hidrogeomorfológicos.
Figure 1.3: Differences in disturbance processes among colluvial. confined alluvial, and
unconfined alluvial channels in mountain drainage basins (based on Montgomery
and Buffington, 1997). Procesos y morfologías asociadas a los diferentes canales de
una cuenca de montaña: canales coluviales, confinados aluviales y aluviales no
confinados (Tomado y traducido de Montgomery y Buffington, 1997).
Figure 1.4: Idealized long profile from hillslopes and unchanneled hollows downslope
through the channel network showing the general distribution of alluvial channel
types and controls on channel processes in mountain drainage basins (modified from
Montgomery and Buffington, 1997). Red box shows the type of channels studied in
this work. Clasificación de diferentes morfologías de ríos en función del sustrato y
distribución espacial en la cuenca (modificado y traducido de Montgomery y
Buffington, 1997). El cuadro rojo indica los tipos estudiado en este trabajo.
Figure 1.5: Hypothesized relative influence of wood on the world’s rivers throughout
geologic time (from Montgomery et al., 2003). Influencia relativa (hipotética) de la
presencia de material leñoso en los ríos a lo largo del tiempo geológico (Tomado y
traducido de Montgomery et al., 2003).
Figure 1.6. (A) Upstream view of Santa María stream crossing Canturral area in La Adrada
(Ávila, Gredos mountain range, Spain, January 2012). The boulder size is decimetre.
Vista hacia aguas arriba de la Garganta Santa María en La Adrada a su paso por la
urbanización Canturral (Ávila, Sierra de Gredos, enero 2012). El tamaño de los
bloques es decimétrico.
Figure 1.6: (B) Woody jam with tilted trees in the flow direction of the Teverga River in
Tuñón (Asturias, Picos de Europa mountain range, October 2011). Acumulación de
material leñoso entre árboles inclinados en la dirección de la corriente en el río
Teverga en las cercanías de Tuñón (Asturias, Picos de Europa, octubre 2011).
Figure 1.6: (C) Pull up trees in a plane bed reach of the College Burn River in the North of
NewCastle (England, November 2009). Árboles arrancados en la dirección de la
corriente en un tramo de lecho plano del río College Burn al Norte de NewCastle
(Inglaterra, noviembre de 2009).
Figure 1.6: (D) Wood and blocks and boulders in the Kaligandaki River in the Annapurna
region (Nepal, Hymalaya, February 2012). Restos de árboles entre los cantos y
bloques en un tramo del río Kaligandaki en la región de los Annapurna (Nepal,
Himalaya, febrero 2012).
Figure 1.7: Pattern of flood disposition in relation to altitude (from Weingartner et al.,
2003). Influencia de determinados factores en la generación de inundaciones en
función de la altitud (Tomado y traducido de Weingartner et al., 2003).
Figure 1.8: Headwaters of the Tajo River after an important flood, large quantities of wood
were mobilized and deposited in jams (1920-30´). Picture taken from the collection
of Complutense University of Madrid. Río Tajo en cabecera, tras una crecida
importante, que movilizó grandes cantidades de madera y la colocó formando una
represa (años 1920-30).Fotografía de la colección fotográfica histórica del
Departamento de Geodinámica de la Universidad Complutense de Madrid.
Figure 1.9: Sediment and wood deposits after the 1996 flash flood in the Arás stream in
Biescas (Spain), Las Nieves Camp Site. 86 people died (picture taken by Carlos
Sancho and courtesy of Virginia Rubio). Depósito de sedimentos y de madera en el
barranco de Arás en Biescas, Camping Las Nieves, tras la avenida de 1996.
Fallecieron 86 personas (fotografía de Carlos Sancho y cedida por Virginia Rubio).
CAPÍTULO/CHAPTER 2
Figure 2.1: Methodological flowchart. Numbers correspond with sections in this chapter.
Esquema metodológico seguido a lo largo del trabajo de investigación. Los números
se corresponden con los nueve apartados de este capítulo.
Figure 2.2: Physiographic units delineation in one of the study sites. Ejemplo de
delimitación de unidades fisiográficas homogéneas en el análisis geomorfológico en
una de las zonas de estudio.
Figure 2.3: (a) Main types of morphometric indices for the characterization of the shape,
size and height of the catchment basin and which are useful in flood hydrology; (b)
Morphometric parameters of the drainage network which affect flood hydrology.
(Taken from Díez-Herrero , 2009). Índices morfométricos para la caracterización de
la forma, el tamaño o la altura de la cuenca hidrográfica, útiles en hidrología de
avenidas; (b) Parámetros morfométricos de la red de drenaje con incidencia en
hidrología de avenidas. (Extraído de Díez-Herrero , 2008).
Figure 2.4: (a) Unit bars features found in gravel-bed channels; (b) Bar types and different
parts. cb: head bar; lb: tail bar; cp: main channel; cs: secondary channel; cl: lateral
channel. (From Church and Jones, 1982 and Ramos and Friend, 1982). Tipos de
barras en sistemas aluviales de gravas; (b) Esquema de tipos de barras y sus zonas. cb:
cabeza, lb: cola de barra; cp: canal principal o exterior; cs: canal secundario o interior;
cl: canales laterales. (Extraído de Church y Jones, 1982 y Ramos y Friend, 1982).
Figure 2.5: Dendrogeomorphological evidence useful in the study of torrential floods. 1.)
“candelabrum” growths; 2) Tilted with feet tipped over; 3) Decapitated; 4) Impacted
by sediment load; 5) Branches torn off; 6) Grazing by other falling boles; 7) Erosion
on the trunk; 8) Trunk bends and angles; 9) Sudden narrowing of trunk; 10)
Bifurcations; 11) Bends and angles in branches; 12) Regrowth from buried trunks;
13) Exposed roots with stripped bark and eroded surface; 14) Roots aloft without
contact with substratum; 15) New roots from buried trunks. (Díez-Herrero , 2009
and 2012). Tipologías de los principales indicios dendrogeomorfológicos
macroscópicos individuales útiles en el estudio de las avenidas torrenciales. 1)
Crecimientos en candelabro; 2) Inclinados y pies volcados; 3) Decapitados; 4)
Impactos por la carga sólida; 5) Arranques de ramas; 6) Rozamientos por caídas de
otros fustes; 7) Erosiones en el tronco; 8) Codos y ángulos del fuste; 9)
Estrechamientos bruscos del fuste; 10) Bifurcaciones; 11) Codos y ángulos en las
ramas; 12) Rebrotes desde troncos enterrados; 13) Raíces expuestas con
descortezados y con erosiones; 14) Raíces flotantes sin contacto con el substrato; 15)
Nuevas raíces desde troncos enterrados. (Díez-Herrero , 2008 y 2012).
Figure 2.6: Field work: (a) Tree sampling; (b) Using of increment borer; (c) obtained
samples; (d) scarred tree and stumps; (e) obtained cross-sections. Tareas de campo:
(a) muestreo de ejemplares; (b) uso de la barrena; (c) testigos obtenidos en el
muestreo; (d) Descortezados y tocones dejados por las tareas de tala de la
Confederación Hidrográfica del Tajo; (e) secciones completas.
Figure 2.7: Laboratory analysis: (a) samples preparation; (b) tree ring width measurements;
(c) tree-ring width curves; (d) samples analysis; (e) disk showing several external and
internal injuries; (f) sample showing reaction wood. Trabajo de laboratorio: (a)
preparación de muestras; (b) análisis y medida de anchura de anillos de crecimiento;
(c) curvas de crecimiento obtenidas en el análisis; (d) examen visual de secciones; (e)
sección con diversas heridas externas e internas; (f) muestra con madera de reacción.
Figure 2.8: (a) Example of a news item featuring a historical flood in the study site (taken
from ABC newspaper); (b) 1982 flood plaque placed on a memorial statue in Luis
Manuel Lopez camp site in memorian of a fatality. Ejemplo de noticia en la que se
describe un evento de avenida en la zona de estudio (extraído del periódico ABC).
(b) placa ubicada sobre un busto conmemorativo en el camping Luis Manuel López
en recuerdo a una víctima de una avenida en 1982.
Figure 2.9: Variogram and variogram model (modified from Guardiola, 2009). Variograma
y modelo de variograma (modificado y traducido de Guardiola, 2009).
Figure 2.10: Instrumental network installed and applied in the study sites; (a) rain gauges;
(b) stream gauges; (c) measuring flow velocity. Instrumentación instalada y utilizada
en varios puntos de la región de estudio; (a) pluviógrafos de cazoletas; (b)
limnímetros de presión; (c) medida de velocidad del flujo con molinete.
Figure 2.11: Evidence and marks found and analysed in one of the study sites. (picture
taken by Andrés Díez). Marcas y evidencias encontradas y analizadas en una de las
zonas de estudio (fotografía de Andrés Díez).
Figure 2.12: Places where deposits and markings useful for palaeohydrological purposes are
often found in mountain basins (based on Jarret , 2002, and Benito and
Thorndycraft, 2004).HWM means high water mark. Esquema ilustrativo de los
depósitos, marcas y evidencias de utilidad paleohidrológica en zonas de montaña
(adaptado de Jarret , 2002 y Benito y Thorndycraft, 2004).HWM significa marca de
aguas altas.
Figure 2.13: Topographical survey in Arenal River: (a) Total station: (b) GPS.
Levantamiento topográfico en el Río Arenal: (a) estación total (b) GPS.
Figure 2.14: (a) Digital elevation model of the Arenal River crossing Arenas de San Pedro
village; (b) Calculation mesh created for Cabrera stream. Modelo digital d
elevaciones del río Arenal a su paso por Arenas de San Pedro. (b) Malla de cálculo
generada para un tramo del arroyo Cabrera.
Figure 2.15: (a) Illustration of variables used for mass balance of wood; (b) Potential tree
fall area showing total tree height, Ht; effective tree height, He; and total arc
distance, 2π He; (c) Schematic illustrating the probability of woody debris falling
into a stream from a tree located (A) at the edge of the stream, (B) at a distance less
than the effective tree height, and (C) at a distance greater than the effective tree
height. (from Robison and Beschta 1990; Martin and Benda, 2001) Esquema de las
variables que se tienen en cuenta en los procesos de incorporación de madera a los
cauces; (b) área potencial de caída de un individuo; (c) ilustración esquemática de la
probabilidad de que una pieza de madera llegue al cauce en función de su
localización: (A) en el banco de orilla; (B) a una distancia menor que su altura
efectiva; (C) a una distancia mayor. (tomado y traducido de Robison y Beschta 1990;
Martin y Benda, 2001).
Figure 2.16: Schematic and body-force diagrams of some of the components of the force
balance acting on a log (from Braudrick and Grant, 2000). Diagramas en los que se
muestra de forma esquemática algunas de las fuerzas involucradas en el inicio de
movimiento de una pieza de madera (tomado y traducido de Braudrick y Grant,
2000).
Figure 2.17: Transport regimes based on wood density (ρw), flow density (ρ), water depth
(h) and log diameter (Dw). A, B and D, rolling or sliding; C, floating. Regímenes de
transporte en función de la densidad de la madera (ρw), la densidad del flujo (ρ), el
calado (h) y el diámetro de la pieza de madera (Dw). A, B y D, rodadura o arrastre; C,
flotación.
Figure 2.18:(a) Flume at the FLUMEN Institute (UPC); (b) water depth measurements
with limnimetre; (c) Flow velocity 3D field using a microacoustic Doppler
Velocimeter; (d) Round white beech wooden dowels used. Canal de experimentación
del Instituto FLUMEN (UPC); (b) medida de calados con limnímetro; (c) medida
del campo 3D de velocidad del flujo con un velocímetro microacústico Doppler; (d)
piezas de madera utilizadas.
Figure 2.19: Uncertainty bounds in the frequency analysis (1) and hydraulic modeling (2).
Limites de incertidumbre en el análisis de frecuencia (1) y en la modelización
hidráulica (2).
Figure 2.20: Illustration of the terms flood hazard, vulnerability and risk (from Merz and
Thieken, 2004) Esquema ilustrativo de los términos peligrosidad, vulnerabilidad y
riesgo por inundación (tomado y traducido de Merz y Thieken, 2004).
Figure 2.21: Illustration of the risk analysis process and the configuration of scenarios with
large wood. Esquema ilustrativo del proceso seguido en el análisis de riesgo y la
generación de escenarios con y sin carga leñosa.
CAPÍTULO/CHAPTER 3
Figure 3.1: Sierra de Gredos (Gredos mountain range) location. Localización de la Sierra de
Gredos
Figure 3.2: Geological and structural sketch of the Sierra de Gredos and surroundings
(from Tectonic and geologic map of Spain 1:200000 by IGME). Esquema geológico-
estructural de la Sierra de Gredos y alrededores. (tomado y traducido del Mapa
Tectónico de España 1:200000 del IGME).
Figure 3.3 (a): Uppermost surface in Los Galayos, Guisando, Ávila. Superficie de cumbres
formada por galayares en la zona de Los Galayos, Guisando, Ávila.
Figure 3.3 (b): High and medium steep slopes,;view from road N-502 in the Mombeltrán
area. Laderas escarpadas y débilmente escarpadas; vista de la zona de Mombeltrán
desde la carretera N-502.
Figure 3.3 (c): Granitic forms in the pathway from San Esteban del Valle to Cuevas del
Valle. Morfologías graníticas en el camino de San Esteban del Valle a Cuevas del
Valle.
Figure 3.3 (d): Fluvial system of the Arenal River crossing Arenas de San Pedro village.
Sistema fluvial del río Arenal a su paso por Arenas de San Pedro.
Figure 3.4: Köppen-Geiger Climate Classification for the Iberian Peninsula (AEMET,
2011). Clasificación climática de Köppen-Geiger en la península Ibérica (AEMET,
2011).
Figure 3.5: Average total annual precipitation for the Iberian Peninsula. (AEMET, 2011).
Precipitación media anual de la península Ibérica (AEMET, 2011).
Figure 3.6: 100-years return period of maximum daily precipitation in Spain (Libro Blanco
del Agua, MAGRAMA, 2000). Mapa de lluvias máximas diarias en la España
Peninsular para un periodo de retorno de 100 años (Libro Blanco del Agua,
MAGRAMA, 2000).
Figure 3.7: Pinus pinaster in Los Galayos (Guisando). Pinus pinaster en la zona de Los
Galayos (Guisando).
Figure 3.8: Cytisus scrubs in the high altitude range (Sierra del Valle, Venero Claro,
Navaluenga). Matorrales de Cytisus en la zona de alta montaña (Sierra del Valle,
Venero Claro, Navaluenga).
Figure 3.9: Riparian forest in the Arenal River crossing Arenas de San Pedro. Bosque de
ribera en el río Arenal a su paso por Arenas de San Pedro.
Figure 3.10: Schematic vegetation catena of two cross sections of the Sierra de Gredos (up)
western section; (down) eastern section. (Adapted from Génova , 2009). Catenas
esquemáticas de la vegetación de la Sierra de Gredos en dos cortes correspondientes al
sector oeste (arriba) y este (abajo) de la Sierra de Gredos (adaptado de Génova ,
2009).
Figure 3.11: Main predominant flood types: 1, flash floods; 2, torrential floods; 3, flooding
of big rivers; and 4, flooding of endorheic areas (from Díez-Herrero, 2009).
Principales tipologías de inundaciones: 1, avenidas súbitas; 2, avenidas torrenciales; 3,
crecidas en grandes ríos; y 4, anegamiento de zonas endorreicas (extraído de Díez-
Herrero , 2008).
Figure 3.12: Studied basins and rain gauge and stream gauge networks in the region.
Subcuencas estudiadas y localización de la red de estaciones de aforos y pluviómetros
de la zona.
CAPÍTULO/CHAPTER 4.1
Figure 1. (A) Location of the study area in the Gredos Mountain Range (Spanish Central
System); (B) overview of the Pelayo River catchment and the study site of this
analysis; (C) the Pelayo River at Guisando village during normal conditions and (D)
during the flash flood event in December 2008 (picture of 2008 flash flood event,
courtesy of Gloria Suarez García, used with permission).
Figure 2. Flow diagram illustrating the different analysis methods applied, from the data
sources to the results.
Figure 4. External evidence of flash flood activity in trees growing along the Pelayo River:
(A) abrasion scars; (B) bark erosion; (C) stem tilting; (D) stem burial; (E) floating
roots; (F) exposed roots. The blue arrows indicate the flow direction.
Figure 5. Geomorphological mapping of the upper part of the study reach. The legend
shows the classification of units and elements.
Figure 6. Relation between the energy of the external evidence and the geomorphic sides.
Ball size represents the percentage of trees showing a specific disturbance located in a
specific geomorphic side related to the total number of trees showing that
disturbance. The legend may be checked in Table 5.
Figure 7. Age structure of the sampled trees in the torrent. Left: upper sector of the study
reach. Right: lower sector.
Figure 8. Spatial distribution of living and damaged trees of four dated events.
Table 2: Weights used for the quantification of different growth disturbances with
different intensitiesa
Table 3: Types of gravel bars and area (absolute and relative) in the study reach
Table 5: Trees showing external disturbances and located in different geomorphic settingsa
Table 7: Calendar year, damaged trees, living trees, percentage of damaged trees, weight
values for the detected growth disturbances, weight for the spatial distribution along
the torrent, and the calculated coefficient of confidence.
CAPÍTULO/CHAPTER 4.2
Figure 1: Location of the four basins analyzed.
Figure 3: Regional flood event chronology based on documentary sources and tree ring
data from the Pelayo and Arenal Rivers.
Figure 5: Fitted GEV distributions: (a) using the instrumental records since 1973; (b) the
instrumental records and the 1936 estimates including uncertainty (h=70, X0=300).
Figure 6: Fitted GEV distributions based on instrumental data and (a) including the four
largest historical events (i.e. the highest magnitude from all estimates) with
uncertainty; instrumental records and (b) all estimates with uncertainty based on
historical records.
Figure 7: Fitted GEV distributions for the Pelayo River at Guisando (A and B) and the
Arenal River at Arenas de San Pedro village (C and D), with discharge estimates
treated as instrumental records (A and C) and as instrumental records with the
exception of the 4 (5) largest events which were considered historical events with
uncertainty (B and D).
Figure 8: (A) Regional gauged data set; (B) regional including the whole set of estimated
event discharges.
Figure 9: Catalogue of maximum observed floods in Spain (AHIS, 2004) and estimates of
the 1936 flood in the study region (four catchments analyzed). The black line
represents the regional envelope curve (REC) of the AHIS database.
Figure 10: (A and C) Discharge for different return periods for all FFA and RFFA tests in
the ungauged basins. μ1 and σ1 are mean and standard deviation for each return
period computing all tests, μ2 and σ2 are mean and standard deviation for at-site tests
and CAUMAX but without RFFA (see B and D).
Table 1: Main morphometric characteristics of the four basins analyzed in this study. Note
the existence of an automatic hydrological information system (AHIS) station for
the Arenal River and of a stream gauge at Candeleda for the Santa Maria River
(located in Figure 1).
Table 2: Discharge estimates (mean and standard deviation) for the four study sites (units
are in m3 s–1). No data was available for cells given in grey.
Table 4: Estimation of the discharge quantiles Q(ML) for Arenal River at AHIS station
corresponding to the return periods T = 10, 50, 100, 500. CI0.05 (CI0.95) is the 5%
(95%) confidence limit of the estimates Q(ML), ΔCI = CI0.95- CI0.05. Grey highlights
the lowest ΔCI and ΔCI/Qt values.
CAPÍTULO/CHAPTER 4.3
Figure 1: (A) Location of the study site within the Spanish Central System. (B) Drainage
basin to the bridge section. The Arroyo Cabrera watershed has been monitored since
2004, with 6 rain gauges and 1 stream gauge installed (shown on map). (C) Detail of
the study area, with main reaches described in the text. The study site (Reach 5) is
marked in red.
Figure 2: Images taken the day after the 1997 event. Photos correspond to Reach 5 in
Figure1.
Figure 3: (A) Clogging curves: water depth in the bridge section represented on the Y-axis,
and the reduced area in the hydraulic radius in percentage on the X-axis. (B) Three
types of behaviour can be differentiated in the curves: LT, CT and UT (explained in
the text).
Figure 4: Results of the hydraulic simulation (123 m3 s-1 peak discharge). Longitudinal
sections of the bridge reach with obstruction in the bridge (A) and without
obstruction (B). (C) Bridge sketch.
Figure 5: Rating curve for the bridge section with stream gauge installed since 2004.
Figure 6: Water depth for the estimated discharge of the 1997 flash flood, (A) without any
obstruction at the bridge, (B) taking into account the LWD and obstruction at the
bridge, (C) water surface with LWD (in red) and without LWD (blue). Gas deposits
and camping site are shown on the maps.
Figure 7: Flood frequency analysis. Continuous black line shows frequency distribution of
CEDEX data. Dotted grey lines are the different adjustments using other frequency
distributions; continuous grey line is the mean of these and the dashed area shows
the confidence intervals (5 and 95%).
Figure 8: Equivalent return period. Curve representing the effect of wood clogging in
critical sections in flood hazard analysis. The broken red line is the estimated return
period for the 1997 event based on the recorded short data series; the continuous
blue line is the return period provided by CEDEX. The dotted black line is the mean
and the grey shading shows the uncertainty area.
Figure 9: Experimental variograms for DEM and precipitation. Black line is the
exponential model applied. Numbers over squares are data pairs for precipitation
data (for DEM variogram more than 1000 pairs were used).
Figure 10: Example of stochastic realization which fulfilled the conditions and so is reliable
for use in hydrological simulation. Daily registered data is placed over the station
(blue squares).
Figure 11: Ratio between a given volume of total precipitation in the catchment of given
duration (12 h) and the discharge range generated. Red dots are storm events from
2004-2011 correlated with observed discharges. Black triangles and grey squares are
design storms and corresponding simulated discharges.
Table 1: Main physical characteristics of the Arroyo Cabrera catchment and reach studied
CAPÍTULO/CHAPTER 4.4
Figure 1: The three main wood recruitment processes on a catchment scale studied in this
analysis: (Ls) landslides; (Ft) fluvial transport during floods; (Be) stream bank
erosion.
Figure 3: Probability (in percentage) that wood will be recruited from the hillslopes based
on connectivity to the channel.
Figure 4: Probability in percentage of wood recruited within the fluvial corridor based on
flood frequency, flood severity and stream bank erosion capacity.
Figure 5: Direct and indirect wood recruitment and delivery evidence can be used in the
field work to interpret LWD dynamics. HWM: high water marks, PSI: palaeostage
indicators.
Figure 7: Land use and forested canopy cover of the study region.
Figure 8: Vegetation clasification based on the resistance to be recruited, and pie chart with
areas.
Figure 9: Number of potentially recruitable trees for the selected 14 basins for (A) Scenario
1, (B) Scenario 2, (C) Scenario 3.
Figure 10: Dendograms of hierarchical clustering using Ward's method. Dendograms for
(A) Scenario 1. (B) Scenario 2. (C) Scenario 3.
Figure 11: (A, B, C) Graphs of a correspondence analysis for each scenario. Basin: basin
area; Veg_Area: forested area; Lnd: landslide-prone area: H_Lnd: high severity
landslide-prone area; Ft: flood- and bank erosion-prone area; H_Ft: high severity
flood-prone area; H_Be: high severity bank erosion-prone area.
Figure 12: Model sensitivity analysis. Graph shows landslide-prone area reduction (%) and
recruitable trees reduction (%).
Table 2: Defined forested areas, areas prone to landslides, floods and bank erosion for
selected basins. Percentages indicate the ratios between the total forested area and the
delineated area defined in the previous column.
Table 3: Results of potentially recruitable trees obtained for the selected basins. SC1:
Scenario 1; SC2: Scenario 2; SC3: Scenario 3 and percentage of total.
Table 4: Summary of field observations in the selected basins: YES (with shading)=
observed: NO = not observed. In bold : Group 1 and 2 for Scenarios 1 and 2, and
group 2 for Scenario 3. *See comments.
CAPÍTULO/CHAPTER 4.5
Figure1: Schematic and body-force diagram of the force balance components acting on a
piece of wood.
Figure2: Cross section of a piece of wood in a stream in different conditions. (A): Log
density greater than water density, and water depth lower than log diameter. (B): Log
density greater than water density, and water depth higher than log diameter (C):
Log density lower than or equal to water density, and water depth higher than or
equal to log diameter. (D): Log density lower than or equal to water density, and
water depth lower than log diameter. Log diameter (Dw), water depth (h), submerged
log area (Asub) and submerged log diameter (y).
Figure 3: Schematic illustrations of log interactions with each other and with the river
banks. (A) In time interval 1, one piece of wood is moving in the flow direction and
meets another piece; they collide in time interval 2 and after the collision they
continue moving with different velocities. (B) Part of the log is out of the river (dry
area). (C) Log hits the bank and slides parallel to it before being reincorporated into
the flow. (D) Log hits the bank and bounces off. Different time steps are represented
as t1, t2 and t3,; α is the incidence angle.
Figure 4: Sketch illustrating the influence of woody debris (deposits or individual logs) on
hydrodynamics.
Figure 6: (A) Measurement points over velocity simulation of geometry 1, with flow from
left to right. (B) Velocity recorded with the Vectrino and simulated by the model for
the flume midpoint.(C) Velocity recorded with the Vectrino and simulated by the
model for the flume right side. (D) Velocity recorded with the Vectrino and
simulated by the model for the flume left side.
Figure 7: Numerical model results (geometry 3): three logs placed with different
orientations (on the left) rotated to a more stable hydrodynamic position, parallel to
the flow direction. Log trajectories are shown as black lines. Flow direction left to
right.
Figure 8: (A and B) Log (mass centre) trajectory in geometry 1 observed and simulated
with kinematic and dynamic methods (background is a flume image). (C) Log (mass
centre) location at each time step, observed (recorded and post-processed) and
simulated (calculated by the numerical model; K: kinematic and D: dynamic) with
turbulence (k-e method). (D) Absolute error in log position coordinates (meters);
(E) Squared error in log coordinates.
Figure 9: Flow (V_flow) and woody log (mass centre; V_log) velocity components (x and
y) simulated by the numerical model (A: kinematic and B: dynamic methods) with
turbulence. Without turbulence flow velocity and log velocity are equal.
Figure 10: Numerical model results of geometry 1: water depth, flow velocity (module),
Froude number and turbulence viscosity are shown together with log trajectory
(black lines). Flow direction left to right.
Figure 11: Log (mass centre) trajectory in Geometry 1 observed and simulated with
kinematic method, same case as Figure 8 (A) Log (mass centre) location at each time
step, observed: recorded and post-processed; simulated: calculated by the numerical
model (kinematic method) without turbulence. (B) Absolute error in log position
coordinates (meters) over time (C).
Figure 12: (A and B): Log (mass centre) trajectory in geometry 2 observed and simulated
with and without turbulence. Correlation coefficient. > 0.8. (C) Log position over
time in: observed and simulated. (D) Absolute error in log position coordinates
(meters) over time. (E) Squared error.
Figure 13: Numerical model results (kinematic method) of Geometry 2: water depth, flow
velocity (module), Froude number and turbulence viscosity are shown together with
log trajectory (black lines). Flow direction left to right .
Figure14: Numerical model results: Log interaction: (A) Log is moving and rotating
according to the velocity field; (B) a second log arrives and collides with log 1 and
changes its trajectory. Flow direction left to right.
Figure 15: Flume geometry 4: two central piers. Logs were entering the flume and most of
them were entrapped by the piers.
Figure 16: Geometry 4 simulation results. (a) Logs blocked (coloured lines) and trajectories
(black lines); (b) Water depth with and without LWD ; (c) Velocity module with
and without LWD ; (d) Froude number with and without LWD ; (e) Turbulence
viscosity with and without LWD .
Figure17: Simulation results for geometry 1 varying the drag coefficient. (A) log trajectory;
(B) log position.
Figure 18: LWD model transport simulation in Fluviá River (Catalonia). Water depth is
shown by colours and logs by brown lines.
Table 2: Statistics for the test case shown in Figure8. X_obs and Y_obs mean X, Y
coordinates observed in the flume and extracted from the records. Sim_K and sim_D
mean X, Y coordinates resulting from the simulation applying kinematic (K) and
dynamic (D) methods.
Table 3: Statistics for the test case shown in Figure12. X_obs and Y_obs are X, Y
coordinates observed in the flume and extracted from the records. Sim and sim_k-e
CAPÍTULO/CHAPTER 4.6
Figure 1: Location of the study area: Arroyo Cabrera stream basin with vegetation density
and species, and location and picture of the bridge affected by wood clogging.
Figure 2: Proposed methodology scheme for the study and simulation of a flood event with
large wood transport. Qi and Qw are the inlet flow discharge and inlet wood budget
respectively.
Figure 4: Scheme of the gate and weir internal contour conditions and wood.
Figure 3: Reconstructed event hydrograph (simulated) and uncertainty intervals and the
established scenarios for the wood budget.
Figure 4: Simulation result for scenario 2 and validation using post-event photographs.
Figure 5: Maximum water depth for the simulated hydrograph without wood, and for the
three scenarios (SC1, SC2, and SC3). The black line represents the bridge, and the
grey polygon the small white building with the high water mark (see Figure4).
Figure 6: Maximum flow velocity for the simulated hydrograph without wood, and for the
three scenarios (SC1, SC2, and SC3). The black line represents the bridge, and the
grey polygon the small white building with the high water mark (see Figure4).
Figure 7: Maximum transport capacity for the simulated hydrograph without wood, and
for the three scenarios (SC1, SC2, and SC3). The black line represents the bridge,
and the grey polygon the small white building with the high water mark (see
Figure4).
CAPÍTULO/CHAPTER 4.7
Figure 1: Location of the study site. There are four bridges along the studied reach, and the
Cuevas tributary stream joins the Arenal in mid reach.
Figure 2: (left) Flash flood in 1999; (right) Post-flood in 2009. Note deposited wood.
Figure 4: Probabilistic flood map (without wood transport) for 500-year flood.
Figure 5: Water depth (in meters) for the 90% probabilistic flooding map (A) without
wood; (B) scenario 1 of wood transport and deposition; (C) scenario 2; (D) scenario
3. Figure 6: Examples of the damage functions used in this analysis. (A) Structural
vulnerability (B) contents vulnerability (depth-damage functions) for one-storey
residential buildings with basement.
Figure 7: Potential damage to the structure and content of the three main types of
buildings in the area based on simulation without wood.
Figure 8: Total potential damage in the three main types of buildings in the area based on
simulations without wood (light blue bars) and for the three scenarios of wood
transport (woody brown bars).
Figure 9: (A) Flow transport capacity (diameter of sediment in meters) without wood and
(C) scenario 3 of wood transport; (B) Danger zone for people without wood and (D)
scenario 3 of wood transport. Maps are related to the 90% probability 500-year flood.
Figure 10: Potential damage to vehicles in the riverside parking area based on simulations
without wood (light blue bars) and for the three wood transport scenarios (woody
brown bars).
Table 1: Main morphometric characteristics of the Arenal river basin and the study reach.
Table 2: Arenal river at Arenas de San Pedro : estimated discharge quantiles Q(ML)
corresponding to return period T =500. CI0.05 (CI0.95) is the 5% (95%) confidence
limit of estimates Q(ML).
CAPÍTULO/CHAPTER 5
Figure 5.1: Conceptual scheme of the proposed methodology to estimate the flood
frequency. Esquema conceptual de la metodología propuesta para estimar la
frecuencia de avenidas
Figure 5.2: Proposed methodology for incorporating woody material transport in flood
risk analysis Propuesta metodológica para incorporar el transporte de material leñoso
en los análisis de riesgo por avenidas
Figure 5.3: (A) Woody jam and scarred trees during the 2008 flood in College Burn River
(England). (B) Scarred tree in Merced River (Josemite, USA), the injury height is
below the flood water level (Yanosky and Jarret, 2002; Jarret, 2007). Acumulaciones
de material leñoso en la base de árboles de ribera que presentan heridas producidas
durante una avenida ocurrida en 2008 en el río College Burn (Inglaterra). (B) La
altura de las heridas observadas en los árboles situados en el río Merced (Josemite,
Estados Unidos), se encuentran por debajo del nivel que alcanzó la inundación
(Yanosky y Jarret, 2002; Jarret, 2007).
Figure 5.4: Schematic and body-force diagrams of some of the components of the force
balance acting on a log with a root-wad (from Braudrick and Grant, 2000).
Diagramas esquemáticos del balance de fuerzas sobre una pieza de madera
suponiendo ésta como un cilindro y un disco (extraído y traducido de Braudrick y
Grant, 2000).
Figure 5.5: Main components of the fluvial dynamics could be modeled after this research
work. Componentes principales de la dinámica fluvial que podrían ser modelizados
tras los resultados de este trabajo de investigación en futuros trabajos.
Diseño de la cubierta:
Esaú Sánchez,
Con la colaboración de Mario Dequel y Elena González.
Fotografías: Virginia Ruiz Villanueva
New Methods in the Flash Flood Hazard and Risk Analysis in Mountain Basins
El objetivo de este trabajo de investigación ha sido desarrollar nuevas metodologías que permitan mejorar
la evaluación del riesgo por avenidas en zonas montañosas. En primer lugar, se han integrado y combinado
diferentes fuentes de datos (instrumentales, históricas o documentales y dendrogeomorfológicas) para
recopilar y caracterizar las avenidas pasadas, e incorporar esta información al análisis de frecuencia.
A continuación, se ha incluido la carga leñosa en los análisis de peligrosidad y riesgo por avenidas.
A pesar de la importancia que tiene su presencia en la dinámica fluvial, este tipo de carga raramente ha
sido tenida en cuenta en los estudios de inundaciones. Para poder incluirla, se ha desarrollado un modelo
The main objective of this research paper is therefore to develop new methods to improve flood risk evaluation
in mountain areas. First, the data from different instrumental, historical, documentary and dendrogeomorphologilcal
sources was integrated to compile and characterize past flash-floods and include this combined information in frequency
analysis. The woody load was then included in the flash-flood hazard and risk analyses. In spite of the significant impact
of the presence of this type of load on fluvial dynamics, it is seldom taken into account in flood studies. To allow this
factor to be included, a numerical model was developed to simulate the woody material transport in conjunction with
the hydrodynamics. Finally, the uncertainty at each stage of the risk analysis (flow rate estimate, frequency analysis,
hydrodynamic simulation and vulnerability assessment) was evaluated from a probabilistic viewpoint.This methodology
was implemented in various basins or river reaches in the Sierra de Gredos (Spanish Central System).
Directores:
Dr. José María Bodoque del Pozo Dr. Andrés Díez Herrero
Universidad de Castilla-La Mancha Instituto Geológico y Minero de España