Verification of precipitation in weather systems: determination of systematic errors

摘要: Abstract An object-oriented verification procedure is presented for gridded quantitative precipitation forecasts (QPFs). It carried out within the framework of “contiguous rain areas” (CRAs), whereby a weather system defined as region bounded by user-specified isopleth in union forecast and observed fields. The horizontal displacement determined translating field until total squared difference between fields minimized. This allows decomposition error into components due to: (a) location; (b) volume (c) pattern. Results are first Monte Carlo simulation 40,000 synthetic CRAs order to determine accuracy when systems only partially presence domain boundaries. Verification then operational 24-h from Australian Bureau Meteorology LAPS numerical prediction model over four-year period. Forty-five percent all events were well model, with small location intensity errors. Location was generally dominant source QPF error, directions most frequent varying region. extreme rainfall (>100 mm d−1) forecast, but this case model's underestimation error.