 |
| Titel |
Rainfall estimates for hydrological models: Comparing rain gauge, radar and microwave link data as input for the Wageningen Lowland Runoff Simulator (WALRUS) |
| VerfasserIn |
Claudia Brauer, Aart Overeem, Remko Uijlenhoet |
| Konferenz |
EGU General Assembly 2015
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 17 (2015) |
| Datensatznummer |
250111323
|
| Publikation (Nr.) |
 EGU/EGU2015-11427.pdf |
|
|
|
|
|
| Zusammenfassung |
| Several rainfall measurement techniques are available for hydrological applications, each
with its own spatial and temporal resolution. We investigated the effect of differences in
rainfall estimates on discharge simulations in a lowland catchment by forcing a novel
rainfall-runoff model (WALRUS) with rainfall data from gauges, radars and microwave
links.
The hydrological model used for this analysis is the recently developed Wageningen
Lowland Runoff Simulator (WALRUS). WALRUS is a rainfall-runoff model accounting for
hydrological processes relevant to areas with shallow groundwater (e.g. groundwater-surface
water feedback). Here, we used WALRUS for case studies in the Hupsel Brook
catchment.
We used two automatic rain gauges with hourly resolution, located inside the catchment
(the base run) and 30 km northeast. Operational (real-time) and climatological
(gauge-adjusted) C-band radar products and country-wide rainfall maps derived from
microwave link data from a cellular telecommunication network were also used. Discharges
simulated with these different inputs were compared to observations.
Traditionally, the precipitation research community places emphasis on quantifying
spatial errors and uncertainty, but for hydrological applications, temporal errors
and uncertainty should be quantified as well. Its memory makes the hydrologic
system sensitive to missed or badly timed rainfall events, but also emphasizes the
effect of a bias in rainfall estimates. Systematic underestimation of rainfall by the
uncorrected operational radar product leads to very dry model states and an increasing
underestimation of discharge. Using the rain gauge 30 km northeast of the catchment
yields good results for climatological studies, but not for forecasting individual
floods.
Simulating discharge using the maps derived from microwave link data and the
gauge-adjusted radar product yields good results for both events and climatological studies.
This indicates that these products can be used in catchments without gauges in or near the
catchment.
Uncertainty in rainfall forcing is a major source of uncertainty in discharge predictions,
both with lumped and with distributed models. For lumped rainfall-runoff models, the main
source of input uncertainty is associated with the way in which (effective) catchment-average
rainfall is estimated. Improving rainfall measurements can improve the performance of
rainfall-runoff models, indicating their potential for reducing flood damage through real-time
control. |
|
|
|
|
|
|