Algorithm application to improve weather radar snowfall estimates for winter hydrologic modelling

Algorithms were applied to weather radar data to improve the precipitation estimation for winter hydrologic modelling. The radar data were adjusted to consider the occurrence of mixed precipitation at above freezing air temperatures, the shape of snow particles, and a site specific scaling phenomenon. Radar data, uncorrected and corrected gridded gauge precipitation data were used as input to the linked WATFLOOD/CLASS hydrologic model for simulation of streamflow. WATFLOOD performed the horizontal water routing and CLASS performed the vertical energy and water budgeting. Modelling of the Grand River watershed that is within the coverage of the Atmospheric Environment Service C‐band radar in King City, Ontario, Canada for the five winters from 1993 to 1997 illustrated that on average the adjusted radar images produced ±15% of the observed runoff volumes whereas the corrected gauge precipitation yielded 35% less runoff than observed. Substantial seasonal variation was observed. Radar provided more realistic winter precipitation quantities for streamflow modelling than the corrected gauge data. Application of the algorithms improved upon the raw radar estimates. Copyright © 1999 John Wiley & Sons, Ltd.