Are We Really That Bad at Predicting the Weather?

Oct. 18, 2016
Sw Jk Blog

In the wake of Hurricane Matthew, a New York Times article claims that US government weather forecasters are woefully inadequate at predicting the course of hurricanes and other meteorological events. For example, as late as September 29, the National Hurricane Center predicted that Matthew—then a Category 1 hurricane—would strengthen only slightly within the next 24 hours, but instead it turned into a Category 5 storm the next day, and it was still a Category 4 hurricane when it made landfall in Haiti on October 4.

The article quotes meteorologist Cliff Mass, a professor of atmospheric sciences at the University of Washington, who maintains a blog about the weather—analyzing after each major storm the accuracy of government forecasts, specifically those of the National Weather Service, which is part of the National Oceanic and Atmospheric Administration. His general criticism is that the models NOAA uses are outdated, and that models used in parts of Europe and commercially available models here would better serve us. He favors what is known as ensemble forecasting, in which many changes are made to the variables and the model is run repeatedly. The various results are then compared. “When all the forecasts in an ensemble agree, it’s a reasonably sure bet that the predictions will pan out,” the NYT article says, but it also points out that ensemble forecasting requires much greater computing power.

The consequences of inaccurate predictions are serious, not only those that underestimate a storm—and therefore might not give people enough time to evacuate—but also those that overestimate it. Too many evacuation warnings in areas that turn out to be only marginally affected, or in areas that storms pass by completely, tend to make people take such warnings less seriously the next time around. Besides the potential for loss of life, the article points out that many industries—shipping and agriculture, for example—lose money because of inaccurate predictions, and “even slightly more precise wind-speed projections would help airlines greatly reduce fuel costs.”

The remarkable thing is that some within NOAA seem to agree. The head of NOAA’s National Centers for Environmental Prediction says in the article that Mass’s criticisms after Hurricane Sandy may have prompted Congress to fund computer upgrades for the National Hurricane Center. The director of a regional National Weather Service office says NOAA has developed better prediction technologies than those widely in use, but they’re not employed, in part because many agencies still don’t have the computing power required to use them.

Has your area been affected by predictions that tremendously over- or underestimated a hurricane, tornado, blizzard, flash flood, or other weather event?

About the Author

Janice Kaspersen

Janice Kaspersen is the former editor of Erosion Control and Stormwater magazines.