Weather by the Numbers: The Genesis of Modern Meteorology (review)

The historiography of computing has long associated the Meteorology Project at the Institute for Advanced Study with physicist John von Neumann, relegating meteorologists themselves to the background. Kristine Harper’s fine monograph carefully documents how meteorologists, particularly Swedish-American Carl-Gustaf Rossby, drove the development of numerical weather prediction. By exploring the influence of operational users charged with making daily forecasts on a strict schedule, Harper expands our understanding of postwar scientific computing beyond research into practice.

Harper aims to explain “American meteorology’s transformation from a discipline more art than science in the early twentieth century to a sophisticated science by 1955” (p. 2). This transformation culminated in the Meteorology Project, which pioneered modeling atmospheric dynamics by solving nonlinear equations using digital computers. Numerical simulation has become central to atmospheric science in the decades since, elevating it from a “guessing science” in the 1920s into a Nobel-worthy activity that shapes global environmental policy today.

The book is divided into eight chapters. The first three survey American meteorology from the establishment of organized weather prediction in the United States through the end of World War II. Harper documents the Weather Bureau’s institutional resistance to research, meteorology’s tentative emergence as an academic discipline during the interwar years, and the transformative professional and scientific effects of training thousands of military meteorologists for World War II.

The next three chapters draw on underused archives to reconstruct the Meteorology Project. Chapter 4 shows how the project emerged from a confluence of interests: the military’s desire to control the weather, the Weather Bureau’s plans to influence meteorological research, von Neumann’s interest in proving the utility of digital computing, and Rossby’s career-long effort to describe the atmosphere’s general circulation numerically. The project thus combined “scientific goals, civilian manpower, and military funding” (p. 91). But in 1948 meteorological theory could not yet frame problems that could be solved by computers. Chapters 5 and 6 explore efforts by a “Scandinavian tag team” (p. 121) led by Rossby and his protégés between 1948 and 1952 to create a solvable, theoretically sound computer atmosphere that behaved like the real one. Unlike most American meteorologists, Scandinavians could create simulations because they combined research experience in dynamical meteorology with years of synoptic forecasting practice.

Harper does not herself make a connection to Peter Galison’s discussion of simulation as a path between experiment and theory, but her story expands that conversation nonetheless. She saves her sharpest words for Philip Thompson, “a man of extraordinary intelligence and unbridled ambition” (p. 110), whose effort to establish a separate air force numerical weather-prediction effort was aborted by a high-level agreement in 1952. Harper’s careful archival work revises Thompson’s later claims about his contributions to the project, claims to which she argues that subsequent histories have given too much credence.

The final two chapters explore an issue of particular value to historians of technology, the institutional and technical challenges of integrating research into routine operations. Just as numerical weather forecasting became operationally useful in 1956, efforts to run a unified system that served the Weather Bureau, air force, and navy equally began to break down. The air force preferred a baroclinic model that provided forecasts for high-altitude flight, while the navy preferred a barotropic model that worked best for the eastern Atlantic Ocean. By 1961, declining computer costs and increasing military budgets produced a typical cold war solution: three parallel systems. Sketching developments up to the present, Harper’s conclusion shows how the materiality of scientific practice, such as forecasters’ need to visualize the weather through synoptic maps, slowed the usage of numerical products even into the 1990s.

Harper’s exceptionally clear writing, strong narrative voice, and assumption of no prior knowledge of atmospheric science make the book ideal for classroom use. Centering her first book on Rossby was an ambitious choice; though clearly the most important figure...