LONG-LEAD SEASONAL FORECASTS - WHERE DO WE STAND

The National Weather Service intends to begin routinely issuing long-l ead forecasts of 3-month mean U. S. temperature and precipitation by t he beginning of 1995. The ability to produce useful forecasts for cert ain seasons and regions at projection times of up to 1 yr is attribute d to advances in data observing and processing, computer capability, a nd physical understanding-particularly, for tropical ocean-atmosphere phenomena. Because much of the skill of the forecasts comes from anoma lies of tropical SST related to ENSO, we highlight here long-lead fore casts of the tropical Pacific SST itself, which have higher skill than the U.S forecasts that are made largely on their basis. The performan ce of five ENSO prediction systems is examined: Two are dynamical [the Cane-Zebiak simple coupled model of Lamont-Doherty Earth Observatory and the nonsimple coupled model of the National Centersfor Environment al Prediction (NCEP)]; one is a hybrid coup red mode I (the Scripps In stitution for Oceanography-Max Planck Institute for Meteorology system with a full ocean general circulation model and a statistical atmosph ere); and two are statistical (canonical correlation analysis and cons tructed analogs, used at the Climate Prediction Center of NCEP). With increasing physical understanding, dynamically based forecasts have th e potential to became more skillful than purely statistical ones. Curr ently, however, the two approaches deliver roughly equally skillful fo recasts, and the simplest model performs about as well as the more com prehensive models. At a lead time of 6 months (defined here as the tim e between the end of the latest observed period and the beginning of t he predictand period), the SST forecasts have an overall correlation s kill in the 0.60s for 1982-93, which easily outperforms persistence an d is regarded as useful. skill for extratropical surface climate is th is high only in limited regions for certain seasons. Both types of for ecasts are not much better than local higher-order autoregressive cont rols. However, continual progress is being made in understanding relat ions among global oceanic and atmospheric climate-scale anomaly fields . It is important that more real-time forecasts be made before we rush to judgement. Performance in the real-time setting is the ultimate te st of the utility of a long-lead forecast. The National Weather Servic e's plan to implement new operational long-lead seasonal forecast prod ucts demonstrates its effectiveness in identifying and transferring '' cutting edge'' technologies from theory to applications. This could no t have been accomplished without close ties with, and the active coope ration of, the academic and research communities.