(212) 854-6580 Wednesday, Feb. 24, 1999

rjn2@columbia.edu 2 P.M. EST

Scientists can now predict with some degree of accuracy such extreme climate perturbations as El Ni–o. The next question, say two researchers at Columbia University and a third at the National Center for Atmospheric Research, is how to translate t hose scientific findings into benefits for societies affected by climate variability.

In a commentary that appears in the Feb. 25 issue of the British journal Nature, the authors examine the fishing sector in Peru, the country where El Ni–o was named and where its impact is often severe. In 1997, fishing generated more than $1 billion in foreign exchange earnings and at least 70,000 jobs. But despite advance warnings, Peru was battered by the 1997-98 El Ni–o: rain and floods affected thousands, crop yields fell and the fish catch fell sharply. Why didnât the climate forec asts head off the disaster?

In addition to the limitations of the forecasts themselves, the authors point to two basic issues that require study: first, how forecasts are disseminated in the tropical regions of the globe affected by El Ni–o, including which groups receive wh at news; and second, which groups benefit and which are harmed in the process. While the responsibility to decide who should benefit from the forecasts is not one that falls to scientists, both these issues can be clarified by analysis of societies, say the authors, who intend their work to contribute to more effective dissemination of climate forecasts.

They are Alex Pfaff, assistant professor of economics and international affairs at Columbiaâs School of International and Public Affairs, its Department of Economics and at the Center for Environmental Research and Conservation at Columbia; Kenneth Broad, senior staff associate and applications specialist of the International Research Institute for Climate Prediction at Lamont-Doherty Earth Observatory, Columbiaâs earth sciences campus in Palisades, N.Y., and Michael Glantz, senior scientist at the federal governmentâs National Center for Atmospheric Research in Boulder, Colo.

El Ni–o was named by Peruvian fishermen and described scientifically as early as 1892, but it took the strong 1972-73 event and the subsequent collapse of the Peruvian anchovy fishery to focus attention on the problem. Forecasts now reach Peru reg ularly, but since the chief medium for direct dissemination is the Internet, and the language is English, individual farmers and fishermen likely learn of them only through second or third-hand sources, usually the local media. Problems can result when m edia covering climate-related press releases remove caveats related to the probabilistic nature of climate forecasts, implying a level of certainty not warranted by the facts. Further, editors may derive their own conclusions regarding climate impacts, s uch as mortality or crop yields.

Large Peruvian fishing companies have greater access to climate forecasts than do most small-scale fishermen. They have used this advance information to anticipate fish migration patterns, which vary as a pool of warm water migrates across the Pac ific to Peruâs shores during El Ni–o events. A better estimate of fish migrations gives those companies an advantage over competitors, and may allow them to haul in large catches even though fish stocks are under greater stress.

In light of such issues of access, understanding and private reactions, Professor Pfaff said: ãClimate forecasts have been used to some success, but there are still and should be debates about how the information is being disseminated. Part of th e reason for writing this commentary is to draw the attention of both climate researchers and public policy analysts to the problem.ä

Erroneous local interpretation and inequitable use of climate forecasts occurs in many societies, including the United States, where many individuals still believe that El Ni–o has predictable, detailed effects on local weather, said Antonio Divino Moura, director of the IRI. The organization is participating in a

project undertaken jointly with the Tinker Foundation, New York, N.Y., and involving the work of local and American anthropologists, to develop a weekly bulletin in Spanish that will bring El Ni–o forecasts directly to local newspaper editors. IRI is conducting climate modeling with local partners in Brazil and in Indonesia, and has held training courses in Mali and Australia on using climate information. Such courses are scheduled in Argentina, Ecuador and Peru.

ãWe expect the issue of how climate forecasts are used to arise in every society that receives them,ä Dr. Moura said. ãThe scientific issue is not just one of climate modeling, but also of public policy and of social structure in these societies. We need to carefully consider the impacts these forecasts can have and how societies finally make use of them.ä

Information to avoid or reduce the human and economic devastation caused by extreme weather is available through several channels. The IRI was established in 1997 jointly by Columbia and the Scripps Institution of Oceanography at the University of California, San Diego, with funding provided by the National Oceanic and Atmospheric Administration (NOAA). Forecasts are also developed at a handful of government laboratories such as NOAAâs Climate Prediction Center in Washington, D.C., and at researc h universities such as the University of Hawaii and the University of British Columbia. (A list of such forecasting agencies is available on the NOAA website.)

The IRI, an important component of the Columbia Earth Institute, relies on computer models developed by the international climate modeling community, including one by climate researchers Mark Cane and Stephen Zebiak. This model has successfully pr edicted several climate events since 1986, and, for example, in the 1986-87 El Ni–o, helped Peruvian farmers avoid crop losses by taking steps such as planting rice, which grows well with more water, instead of cotton, which does not. But Dr. Cane, the G . Unger Vetlesen Professor of Earth and Climate Sciences, notes that although physical models now in use can accurately predict a climate event, they often canât detect its severity.

2.22.99 19,483