Forecast Methodologies That Improved Probability of Precipitation Forecasts at WFO (Weather Forecast Office) Billings

摘要

Through the 2005-2006 winter, awareness of a dry forecast bias was raised across NWS Western Region (WR) forecast offices. A local study at WFO Billings (BYZ) validated WFO Billings also exhibited a dry bias for multiple years. A large part of the problem was zero probability of precipitation (PoP) being forecast too often in medium range forecasts. Another problem was PoP forecasts were often lower than the GFS MOS or MEX at the medium range when measurable precipitation was observed. So in many cases when the weather pattern was forecast to be more favorable for precipitation than climatology and the MEX provided a PoP near climatology, a zero forecast (or near zero) PoP was still provided because of lower forecaster confidence. This lower confidence was often a result of model run to run discontinuity and/or differing solutions among numerical models.