Sasaki’s Pivotal Contribution: Calculus of Variations Applied to Weather Map Analysis

Yoshikazu Sasaki developed a variational method of data assimilation, a cornerstone of modern-day analysis and prediction in meteorology. Fundamentally, he formulated data assimilation as a constrained minimization problem with equality constraints. The generation of this idea is tracked by analyzing his education and research at the University of Tokyo in the immediate post–World War II (WWII) period. Despite austere circumstances—including limited financial support for education, poor living conditions, and a lack of educational resources—Sasaki was highly motivated and overcame these obstacles on his path to developing this innovative method of weather map analysis. The stages of his intellectual development are followed where information comes from access to his early publications, oral histories, and letters of reminiscence. It has been argued that Sasaki’s unique contribution to meteorological data assimilation stems from his deterministic view of the problem—a view founded on the principles of variational mechanics. Sasaki’s approach to the problem is compared and contrasted with the stochastic view that was pioneered by Arnt Eliassen. Both of these optimal approaches are viewed in the context of the pragmatic–operational objective analysis schemes that were developed in the 1950s–1960s. Finally, current-day methods [e.g., three- and four-dimensional variational data assimilation (3DVAR and 4DVAR)] are linked to the optimal methods of Eliassen and Sasaki.