Assessment of radar in flood forecasting using hydrologic models

Weather radar has been used for several years as a method of measuring precipitation. The distributed data produced by radar systems provides better spatial resolution than the point data produced by rain gauges. This study is a survey of the suitability radar for hydrological applications.

As part of the Mesoscape Alpine Progamme (MAP) in the European Alps the Numerical Weather Prediction (NWP) model MC2 (Mesoscale Compressible Community Model) and weather radar data were used in conjunction with the hydrologic model WATFLOOD to improve flow forecasts and to assess the validity of weather forecasts. For the duration of MAP a series of real-time flow forecasts were made once per day using forecasted precipitation from MC2 and radar data to produce flows.

The goal of the MAP study was to determine the effectiveness of NWP models for providing accurate precipitation estimates in comparison to radar data. It was also important to determine whether better flood forecasting results could be achieved through the use of NWP models compared to historical radar data and thereby to improve current forecasting capabilities. Radar data underestimated precipitation on the Toce-Ticino watershed and overestimated precipitation on the Ammer watershed. MC2 forecast data produced better flood forecasts than the real-time radar data. A post-facto re-analysis of the radar data provided improvements over the real-time radar data. On two occasions, MC2 forecasts resulted in good streamflow forecasts while for a third event, MC2 underestimated the precipitation.

Eight years of radar precipitation data for Ontario, Canada have been archived and used in the hydrological model WATFLOOD, to calculate flows. Radar data, when used directly in the model, has only recently been able to provide adequate precipitation estimates. This has been attributed primarily to signal attenuation, wind effects and the adjustment of operating parameters. Streamflow is a means to access radar data. A comparison between streamflow values calculated using radar data in the hydrologic model WATFLOOD and measured streamflows will be presented. These comparisons will provide new insight into the value of radar for hydrology and in particular, will show if the new data is an improvement.