This study will develop prototypes of directly comparable observation and model derived products for routine, daily evaluation of regional model numerical weather forecasts of winter storms near Portland, OR. The Portland area was chosen for its high frequency of winter precipitation events and the orographic forcing by the Coastal Mountains and Cascade Mountains which constrain locations of precipitation development compared to flat terrain. A key aspect of this research is the comparison of 3D observations from the Portland WSR-88D NWS radar to the 3D model output. Analysis of comparison metrics from a large sample of storms will help define: What structural characteristics of storms are most important to reproduce in forecasts? What subset of characteristics are most reliably observed? How to distinguish between details unique to an individual storm versus those that are repeated among a group of similar storms? Routine 3D evaluation provides means of using a large sample of storms to determine: forecast model strengths and weaknesses for particular storm types, evaluation of proposed model changes, and applicability of new conceptual models based on case studies.
In collaboration with Brian Colle of Stonybrook University-SUNY, near-real time regional forecast modeling of the Portland, OR area and accompanying observation analysis will commence 1 November 2005 for the 2005-06 winter storm season.