Secrecy of Data Said to Limit Climate Forecasting

CLIMATE modeler Michael Schlesinger looks at improved East-West relations and says he hopes for ``sea ice glasnost.'' Lack of climatological data on the month-by-month variation of sea ice thickness headed a list of global warming forecast limitations that the University of Illinois geophysicist discussed during the recent American Association for the Advancement of Science annual meeting. Yet some 40 years' worth of just such data - gathered by sonar-probing submarines - are locked up in United States and Soviet secret naval archives.

Dr. Schlesinger said that if those data were released, he and colleagues at the Main Geophysical Observatory in Leningrad could do a better job of fulfilling their 1987 mandate from Presidents Reagan and Gorbachev to come up with an improved forecast of future climate.

Yet even with those coveted sea ice data, forecasting the climate effects of global warming, which is driven by rising atmospheric concentrations of carbon dioxide and other heat-trapping gases, would still be an uncertain business.

Syukuro Manabe of the National Oceanic and Atmospheric Administration's Geophysical Fluid Dynamics Laboratory explained, ``Since the model is always a simplified version of the real climate system ... you can never authoritatively say how large its uncertainty is.'' Decisionmakers, he said, will have to work with ``fudgy information.''

Dr. Manabe added that, one way to make the information less fudgy, is to continue to work to improve the computer forecasting models while simultaneously closely monitoring global climate. Then, as model short-term predictions more closely approximate what the real climate is doing, confidence in the models' general forecasting ability will grow.

Two main types of factors limit that confidence today.

One is raw computing power. Schlesinger said that modelers do not have the computing power even to make the most of their present mathematical programs that simulate the climate system. They cannot, for example, handle a combination of the best simulation program for atmospheric circulation coupled with a program that simulates circulation throughout the full depth and range of the ocean. So they simplify the ocean program.

However, they cannot neglect ocean currents altogether, especially in the top water layer. When they do not include the currents, Schlesinger said, the resulting model does not even simulate present climate well, let alone forecast the future.

This also illustrates the second type of limiting factors - lack of knowledge of relevant geophysical elements. Cloud cover and how to model it are poorly understood. There are few data on precipitation over the oceans. And there is the lack of sea-ice thickness data.

Schlesinger explained that this makes it difficult to take proper account of ocean currents. Modelers can do it for the open sea. But they cannot do it under the ice without the thickness data. They also cannot take proper account of changes in the ice cover itself, which is an important climatic factor.

Manabe cited recent work in his laboratory that illustrates the importance of including ocean circulations in computer climate projections. Models that do not take this into account forecast that a doubling of atmospheric carbon dioxide concentration will induce warming all over the globe. But the new model with ocean circulation shows two areas where warming is significantly retarded.

One region is the circumpolar sea that rings Antarctica and the windy region to the north of this - around south latitudes 40 to 50 degrees. Here, circulation extending from the surface to great depth mixes heat deeply into the ocean and slows surface warming. This prospect may force climate forecasters to reevaluate the danger that warming might cause collapse of the Antarctic ice pack, Manabe said. The other region where deep mixing slows surface warming is in the northern North Atlantic.