 |
| Titel |
Application of a stochastic rainfall model in flood risk assessment |
| VerfasserIn |
M. A. Campo, A. Sordo, D. González-Zeas, J. C. Cirauqui, L. Garrote, J. J. López |
| Konferenz |
EGU General Assembly 2009
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 11 (2009) |
| Datensatznummer |
250024615
|
|
|
|
|
|
| Zusammenfassung |
| In Spain, daily series of precipitation data are widely available in many sites from the middle
of last century. However, neither the length nor the time resolution of those series is enough
for some hydrologic applications. For instance, in order to evaluate the probability of
occurrence of peak discharges in a small-size basin (100 km2), hourly precipitation data are
required. In this paper, preliminary results of the application of a stochastic disaggregation
model are presented. The model is used for the generation of precipitation time
series, which are then introduced into a rainfall-runoff model for hydrologic design
purposes.
The Alloz basin, located in the North of Spain with an area of 134 km2 and available
daily precipitation (from 1965) and hourly discharge (from 1964) data, was selected for the
study. The modified Bartlett-Lewis model was selected for stochastic precipitation
disaggregation, due to its conceptual simplicity and its ability to simulate storms of different
nature. The six model parameters were estimated every month from 24-hour and
48-hour precipitation data. A time series of 10000 years of hourly precipitation data
was generated using these parameter values. This time series of precipitation was
analyzed to extract the main storm events of every year, which were simulated with an
event-based rainfall-runoff model to obtain the maximum peak discharge of every
year.
First results show good model skill to capture the temporal structure of precipitation
events and its associated basin response. The simulated time series not only reproduces
correctly the observed basic statistics which were used in model calibration, but also shows a
good fit of the series of yearly maximum daily precipitation data and yearly maximum
discharge.
It can be concluded that the modified Bartlett-Lewis model is an adequate tool
for the generation of long time series of high temporal resolution of precipitation
data for hydrologic application. In future work, spatial variability will be included
in the Bartlett-Lewis model to asses its applicability to midsize and large basins. |
|
|
|
|
|
|