Summary of Forecasts
Dynamic Models: Most of the dynamic models are producing fairly consistent forecast. The COLA, UCLA, NCEP, SIO, SNU, LDEO, IRD models are all predicting modest cold anomalies of the next two to three seasons. These forecasts produce a relatively large spread in terms of amplitude and the timing of the return to near normal in early 2004. One forecast (Neural Dynamic Hybrid) is predicting warm anomalies through July 2004. In summary, in the coming two to three seasons the largest probability is for modest cold anomalies, there is a modest probability for cold, but near normal anomalies and there is a very small probability of warm anomalies.
Statistical Models: The Contructed Analogue forecast calls for cold anomalies to develop through JFM03. The Linear Inverse model is predicted weak cold anomalies through boreal winter 2003-04 with the peak cold anomalies for JJA 2003. The Neural Network model is cold during though September 2003 and then predicts a significant warming through MAM2004 giving El Niño-like conditions. The Singular Spectrum Analysis-Maximum Entropy Method forecast calls for continued cooling through DJF03-04.
For Land Surface Temperature and Rainfall
The Sahel 2003 rainy season forecast from the UK Met. Office call for wet conditions in the three northern regions and very dry conditions in the extreme south. The ECPC global precipitation forecasts for the next one to two seasons call for call for the tropical land masses tending toward a dry state in contrast to the adjacent oceans. The CCA forecast for Canada calls for a warming trend of the central part of the country during JJA03 and drier than normal conditions over southern Alberta and southern Ontario.