Recent Developments in Operational Long-Range Prediction at CPC

The skill and coverage of NWS Official forecasts of 3-month average temperature (T) and total precipitation (P) over the continental United States (CONUS), made operationally from 1995 through 2004 at the Climate Prediction Center (OFF), are compared and contrasted with forecasts for the same period, made using an objective statistical consolidation process (CON), which used the same tools as were available for the subjective official forecasts. The objective retrospective (CON) forecasts for both T and P are greatly improved over the operational (OFF) ones, as measured by the percentage by which they improve upon random forecasts (heidke skill score), and by the area covered by non-random probabilities. These are both qualities which users of the forecasts have demanded. The CON technique was implemented into CPC 3-month forecast operations in 2006, and has resulted in a large increase in the skill of operational forecasts made since then. This technique has also been used to improve NWS's L3MTO downscaled versions of CPC official 3-Month Outlooks. Regional and seasonal variations of the skill and coverage of OFF and CON forecasts, and future plans to upgrade the verification system are discussed.