Boundaries of research: Civilian leadership, military funding, and the International network surrounding the development of numerical weather prediction in the United States

American meteorology was synonymous with subjective weather forecasting in the early twentieth century. Controlled by the Weather Bureau and with no academic programs of its own, the few hundred extant meteorologists had no standing in the scientific community. Until the American Meteorological Society was founded in 1919, meteorologists had no professional society. The post-World War I rise of aeronautics spurred demands for increased meteorological education and training. The Navy arranged the first graduate program in meteorology in 1928 at MIT. It was followed by four additional programs in the interwar years. When the U.S. military found itself short of meteorological support for World War II, a massive training program created thousands of new mathematics- and physics-savvy meteorologists. Those remaining in the field after the war had three goals: to create a mathematics-based theory for meteorology, to create a method for objectively forecasting the weather, and to professionalize the field. Contemporaneously, mathematician John yon Neumann was preparing to create a new electronic digital computer which could solve, via numerical analysis, the equations that defined the atmosphere. Weather Bureau Chief Francis W. Reichelderfer encouraged von Neumann, with Office of Naval Research funding, to attack the weather forecasting problem. Assisting with the proposal was eminent Swedish-born meteorologist Carl-Gustav Rossby. Although Rossby returned to Stockholm to establish his own research school, he was the de facto head of the Meteorology Project---providing personnel, ideas, and a publication venue. On-site leader Jule Charney provided the equations and theoretical underpinnings. Scandinavian meteorologists supplied by Rossby provided atmospheric reality. Six years after the Project began, meteorologists were ready to move their models from a research to an operational venue. Attempts by Air Force meteorologist Philip D. Thompson to co-opt numerical weather prediction (NWT) prompted the academics, Navy, and Weather Bureau members involved to join forces and guarantee that operational NWP would remain a joint activity not under the control of any weather service. This is the story of the professionalization of a scientific community, of significant differences in national styles in meteorology, and of the fascination (especially by non-meteorologists) in exploiting NWP for the control of weather.