Object-based Evaluation of Precipitation Forecast Quality

V. Z. Kisel’nikova

The method of object-based diagnostic evaluation (MODE) is used to estimate the quality of forecasts of intensive precipitation zones using the WRF-ARW mesoscale nonhydrostatic model. The model is integrated on the horizontal 3-km grid without the cumulus parameterization. The prognostic fields of the reflectivity at the level of 700 hPa are compared with the observed radar reflectivity at the level of 3 km. Objects in the simulated and observed fields obtained by means of the convolution (smoothing) and thresholding (>20 dBZ) were matched according to the distance separating their centers of gravity (centroids). This distance was 50 km or lower in 12% of the total number of forecasts and 100 km or lower, in 30% of forecasts. It is revealed that the model overestimates the heavy precipitation areas.

Joomla templates by a4joomla