BOX 1.1

One of the greatest societal benefits provided by Earth sciences in the past 30 years has been the steady improvement of weather forecasts. Figure 1.1.1 shows the monthly moving average of the correlation (a perfect forecast is 100 percent) between observed and forecast weather features for 3-day, 5-day, and 7-day forecasts. The accuracy of forecasts of large-scale weather patterns in both hemispheres has been increasing steadily since 1980. The Southern Hemisphere forecast (bottom curve), which in 1980 was significantly worse than the Northern Hemisphere forecast (top curve), has caught up in accuracy in recent years. This dramatic improvement has been due largely to more and better global satellite data.

FIGURE 1.1.1 Anomaly correlation of 500 hPa height forecasts. SOURCE: Adapted from A.J. Simmons and A. Hollingsworth, 2002, “Some Aspects of the Improvement in Skill of Numerical Weather Prediction,” Q.J.R. Meteorol. Soc. 128:647-678. Copyright Royal Meteorological Society. Reprinted with permission.

that similar improvements are possible in seasonal climate forecasts, which are needed for a variety of agriculture decisions.5


The 1997-1998 El Niño created torrential rains in California and led to rapid increases in food costs. See Kenneth Howe, “El Nino’s Costly Crops,” San Francisco Chronicle, February 26, 1998. Changes in sea temperature and sea surface topography observed by satellites and ocean moorings are critical for forecasting the strength and timing of impending El Niño events. Improved forecasts of the 1997-1998 El Niño event are estimated to have saved California residents on the order of $1 billion compared to the costs of a similar event in 1982-1983, which was not forecast. See “The Economic Impacts of an El Nino,” Space Daily, March 18, 2002, at <>.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement