WASHINGTON — George P. Cressman, a former National Weather Service director who took the lead in applying computers to meteorology and helped change weather forecasting from a form of cloud-gazing guesswork to a codified science, died April 17 at the National Lutheran Home in Rockville, Md. He was 88 and had Alzheimer's disease.

In the 1950s, Cressman developed the first program that could produce accurate and reliable forecasts prepared by computer, which led to a monumental change in how weather is predicted and brought meteorology into the computer age.

As director of the Weather Service from 1965 to 1979, Cressman expanded the number of local weather radars, developed a nationwide weather radio network and introduced systems to provide early warnings of tornados and flash floods. He also made important contributions to cooperative meteorological efforts around the globe.

"He really, truly was a giant in meteorology," said Richard Hallgren, who succeeded Cressman as National Weather Service director. "Worldwide, he was extraordinarily well known."

In the late 1940s, meteorologists were among the first scientists attempting to harness the new technology of computers. They had little sustained success until Cressman was named director of the federal Joint Numerical Weather Prediction Unit in Suitland, Md., in 1954. Using an IBM 701 computer, he recorded the weather conditions at equally spaced points around the world, then devised a program that allowed the computer to produce forecasts derived from the cumulative data.

"At the time, that was a major breakthrough," said Ron McPherson, executive director emeritus of the American Meteorological Society. "Before that was done, forecasting was mostly an art" based on extrapolations from hand-drawn weather charts.

"When computers came in, forecasting became much more of a science," McPherson added. "It started, literally, a revolution in forecasting."

George Parmley Cressman was born Oct. 7, 1919, in West Chester, Pa. He and a boyhood friend, who also became a meteorologist, took an early interest in weather to determine when snowstorms would provide good opportunities for sledding.

After graduating from Pennsylvania State University, Cressman studied meteorology in a military course at New York University, then served as a forecaster with the Army Air Forces. In January 1943, he began teaching meteorology to military students at the University of Chicago, where he became a protege of Carl-Gustaf Rossby, a renowned meteorologist who identified the jet stream.

Cressman manned such distant outposts of the U.S. Weather Bureau as Lewistown, Mont., Homestead, Fla., and Mount Home, Idaho, before receiving his doctorate from the University of Chicago in 1949. He then worked at the Air Force's central weather command at Andrews Air Force Base outside of Washington, seeking ways to improve the scientific validity of weather predictions. In the early 1950s, he compiled forecasts for atomic bomb tests in the Nevada desert.

In 1955, one year after Cressman began to concentrate on applying computers to meteorology, the Weather Bureau (renamed the National Weather Service in 1967) was issuing its first computer-generated forecasts. The "Cressman Analysis" or "Cressman Method," as it became known, greatly refined forecasting methods and gave rise to what meteorologists call "numerical weather prediction."

"That is an achievement that, I think, 100 years from now he'll still be known for," Hallgren said.

In the early 1960s, when Cressman was director of the National Meteorological Center, which was the government's joint military and civilian forecasting unit at the time, he instigated efforts to share weather data with the Soviet Union. He continued to expand the use of computers in weather forecasting throughout the 1960s.

As director of the Weather Service, Cressman added 100 weather radars to the national network in the 1970s and established dozens of Weather Service branches to provide accurate local forecasts across the country. He stepped down in 1979 but continued to work as a consultant to weather services in China, Spain and Brazil for several years.

He was president of the American Meteorological Society and received the International Meteorological Organization Prize — the highest honor in his field — in 1978. He conducted research on gravity waves and other climatic phenomena until the late 1980s.

Cressman lived in District Heights, Md., before moving to Rockville in 1964. He was a skilled photographer, won awards in archery, enjoyed mountain hiking with his family and was a member of the Cosmos Club. He also enjoyed classical music, and throughout the 1950s was organist at the old Bethlehem Lutheran Church in Southeast Washington.

His first wife of 32 years, Nelia Hazard Cressman, died in 1974.

Survivors include his wife of 32 years, Frances Cressman, of North Bethesda, Md.; four children from his first marriage, Ruth Boyd of Kealakekua, Hawaii, George I. Cressman of Manchester, Mo., Catherine Beck of Grove City, Pa., and Florence Gardner of Mooresville, N.C.; a brother; nine grandchildren; and 10 great-grandchildren.