HR: 0800h
AN: GC41A-0389    [Abstracts]
TI: Improving the Streamflow Forecasting using Merged Precipitation Products
AU: * Ishak Boushaki, F
EM: fishakbo@uci.edu
AF: The University of California, Irvine, Center for Hydrometeorology & Remote Sensing University Research Park 5251 California Ave., Suite 200, Irvine, CA 926 United States
AU: Hsu, K
EM: kuolinh@uci.edu
AF: The University of California, Irvine, Center for Hydrometeorology & Remote Sensing University Research Park 5251 California Ave., Suite 200, Irvine, CA 926 United States
AU: Sorooshian, S
EM: soroosh@uci.edu
AF: The University of California, Irvine, Center for Hydrometeorology & Remote Sensing University Research Park 5251 California Ave., Suite 200, Irvine, CA 926 United States
AB: Accurate measurement of precipitation is essential for a range of hydrologic studies extending from severe flood forecasting to climatological studies of droughts. At present, there are essentially three systems for providing precipitation measurements: (1) rain gauges, (2) ground-based radar, and (3) satellite derived estimates. Variability of rainfall has been acknowledged as a key source of uncertainties in streamflow and flood forecasting (Droegemeier et al. 2000). The major objective of this study is two fold: 1) obtain more reliable precipitation product by optimally combining all precipitation sources instead of using one sensor measurement alone. This proposed study intends to use available multi-source precipitation data, along with a rainfall-runoff model in which runoff measurements are assumed to be the reference information, to trace back the uncertainty associated with every precipitation data source. 2) Improve the performance of streamflow simulation and forecast as a consequence of the merged precipitation product mentioned above that will be used as input to the hydrological model and accordingly will result in a more accurate simulation of the streamflow. The problem is complicated first by our poor understanding of the relationship between the measured quantities and the actual rainfall amounts on the ground. A second source of complexity is the common reliance on a single source for reference precipitation measurements (i.e., gauge rainfall measurements) which in most cases are not widely available. Addressing this inherent problem requires methodologies that make best use of all precipitation measurement systems. Thus this study suggests a merging of precipitation sources by optimizing the weighted average of gauge, radar, and satellite-based precipitation observations. Because both weighted average parameters and hydrologic conceptual model parameters are unknown variables, a global search algorithm, the shuffle complex evolution (SCE), is used to find the optimal set for all parameters.
DE: 1821 Floods
DE: 1854 Precipitation (3354)
DE: 1860 Streamflow
DE: 3354 Precipitation (1854)
DE: 3360 Remote sensing
SC: Global Climate Change [GC]
MN: Fall Meeting 2005