A review of the application of numerical weather predication for short-term streamflow forecasting

[摘要] This reports presents a review of Numerical Weather Prediction (NWP) model generated Quantitative Precipitation Forecasts (QPF) and their application in streamflow forecasting systems. The review is an initial phase of the WIRADA project 4.1 aimed to evaluate the potential benefits of NWP to the operational flood forecasting system of the Bureau of Meteorology, Australia. The current status of QPF applications in both research and operational fields is described. The major issues and challenges are identified and potential solutions are recommended as follows. For streamflow nowcasting (very short leadtimes, e.g., 6 hours), many applications are based on observed in situ or radar-based real-time rainfall and/or the extrapolation of recent rainfall patterns. NWP QPF has the most potential to extend streamflow forecast leadtime, particularly in the range of a few days. For short (2 days ahead) and medium (2-14 days ahead) range streamflow forecasting, the quality of the NWP output remains a major obstacle. Either statistical or dynamical downscaling is used to refine the QPF resolution. In some cases, short to medium range NWP output needs bias correction and post-processing. Ensemble NWP is used to investigate QPF uncertainty, improve forecast skill and extend forecast lead times. Future rainfall uncertainty is often among the largest sources of streamflow forecast uncertainty. Ensemble hydrological prediction has been an active research field but operational ensemble hydrological prediction lags behind. Conversely, relatively little study has been done on the hydrologic component (i.e. model, parameter) of uncertainty in the streamflow prediction system. For WIRADA 4.1 project, the research focus should be on the quantification of NWP QPF quality and uncertainty, hydrological model uncertainty, data assimilation and the seamless use of multiple sources of information (e.g. short and medium range NWP). Use of the combination of NWP and radar-based ensemble forecasts would be a novel avenue for research. Most studies and operational or pre-operational applications of NWP to short-term streamflow forecasting were done in Europe and the United States for the purpose of extending forecast lead time and improving forecast skill. Particular systems of interest are the United States Experimental Ensemble Forecast System and the European Flood Alert System (EFAS). The Hydrologic Ensemble Prediction EXperiment (HEPEX) is an international initiative to foster hydrologic ensemble prediction science and may be a useful community to engage.

[发布日期] 2010-12-31 [发布机构] CSIRO

[学科分类] 地球科学(综合)