Validation and adjustment of precipitation simulated by CCM2/BATS over the continental United States

Abstract

This study compares frequency, intensity and amount of precipitation simulated by NCAR CCM2/BATS with those observed over the continental United States. The emphasis is placed on both spatial and temporal variations. The analysis indicates that: a) The model simulates the seasonal variation of daily intensity reasonably well, while the seasonal variation of frequency is poorly produced. b) The model underestimates the amplitude of intensity, overestimates the amplitude of frequency, and distorts the distribution of phase in diurnal variations. c) The simulated maximum hourly precipitation has almost the same range as is observed. However, the model fails to simulate the overall pattern of maximum hourly precipitation, maximum wet and dry periods from observations. d) The model overestimates the frequency of hourly precipitation by a factor of 13-48 and underestimates intensity by about 90%. These results demonstrate a necessity to adjust simulated precipitation for realistic surface simulation. After adjustment, the simulated intensity is in much closer agreement to the observed value.