Storm forecasting reaches new accuracy

For the second year running, scientists have been flying through violent winter weather over the North Pacific to improve the accuracy and timeliness of storm forecasts on the North American mainland.

Now, they are assessing the results from what may become a regular program to measure storms while they are thousands of miles away at sea.

"From a scientific point of view, we were trying to measure the butterfly effect," says Craig Bishop, an assistant professor of meteorology at Penn State University in College Park.

During the National Oceanic and Atmospheric Administration's Winter Storm Reconnaissance Program, which ran from mid-January to mid-February, National Weather Service (NWS) forecasters were asked to look at their four- to five-day outlooks and identify regions of the country that concerned them. Then modelers worked back to identify the Pacific weather system that was most likely to affect the areas forecasters were fretting over. Air Force Reserve C-130s and NOAA aircraft then flew to the areas from bases in Hawaii and Alaska, dropping instrument packages to gather data.

The accuracy of forecasts this season using Pacific data appear to have improved about 70 to 80 percent of the time, researchers say. Dr. Craig adds that he and his colleagues now are puzzling over the 20 to 30 percent that failed to improve.

Forecasters are trying to give winter weather the hurricane treatment in other ways. In Cheyenne, Wyo., the NWS forecast office is testing a winter-weather severity index. The 1 to 5 index is designed to give people a simple tool for gauging how a storm is likely to affect them. Unlike hurricane or tornado categories, the new index involves subjective judgments. A light snow in Cheyenne would shut down Washington, D.C., where it would earn a higher severity rating. The most life-threatening storms rate a 5.