TOGA: Have We Kept the Promises We Made for Dynamical Seasonal Prediction

In the years before TOGA was launched, scientific research had shown some tantalizing evidence of predictability in the midst of chaos – certain aspects of climate system, for example monthly and seasonal averages, could be predictable even if daily weather could not be predicted beyond 1-2 weeks. This was found to be possible because of higher predictability of low frequency planetary waves, and significant predictable effects of the slowly varying boundary conditions at the earth's surface, especially tropical sea surface temperature (SST). It was thought at that time that even if we did not know how to predict the boundary conditions, the fact that tropical SST anomalies persist for several months, monthly mean and seasonal mean predictions should be possible using atmospheric general circulation models. Subsequent research on ENSO dynamics, and successful experimental predictions with simple coupled ocean-atmosphere models created great expectations for climate prediction.

It is in this context that TOGA was launched with great enthusiasm and expectation. The expectation was that with better observations of upper ocean, and much improved coupled ocean-atmosphere models, it should be possible to make accurate and reliable operational dynamical seasonal predictions. It was considered the dawn of a new era in climate prediction, an era which will witness a scientific revolution of predictability in the midst of chaos, and provide accurate and reliable climate information to society.

One of the major aims of TOGA was “to establish an operational climate (monthly, seasonal) prediction capability.” It is therefore entirely appropriate to refer to COLA as one of the most successful research programs of USGCRP and WCRP. This paper will show that while it is true that there are several centers in the world that are making operational dynamical seasonal predictions, serious deficiencies of climate models in simulating the observed mean climate and its variability have not been reduced at all – some have gotten worse. This paper will present some conjectures on possible causes for lack of progress and suggestions to meet the expectations raised by TOGA.