The Sea: The Science of Ocean Prediction

At the beginning of the 20th century Vilhelm Bjerknes defined the “ultimate problem of meteorology and hydrography” as the discovery of “the laws according to which an atmospheric or hydrospheric state develops out of the preceding one” and the “precalculation of future states” from gridded analyzed observations—that is, forecasting. The development of the electronic computer and the vision of several meteorologists allowed the transformation of meteorology into a sophisticated scientific discipline based on physics and mathematics. The first successful meteorological forecast was carried out in the 1950s. Meteorological forecasting became an operational activity at the end of the 1960s. The contributions to society of such operations have been tremendous. Ocean forecasting began in the 1980s with a joint venture between Harvard University and the Naval Postgraduate School in Monterey, California, that completed the first successful forecast of ocean mesoscales in a limited area of the ocean. Since then, computational ocean modeling and prediction have led to major discoveries across multiple time and space scales, from ocean turbulence to climate. In the first decades of the 21st century, ocean forecasting has become an operational activity. The rapidly increasing interconnectivity between humans and the Earth system suggests that ocean prediction will become ever more vital to society in the years to come. Ocean prediction is now part of the wider human endeavor to understand, monitor, and forecast whole-earth physical and biogeochemical dynamics and cycles. The science of ocean prediction is the systematic development of fundamental knowledge about ocean dynamics in the form of testable ocean models and estimation systems that relate to forecasting the ocean’s evolution. It includes governing laws, model equations, parameterizations, numerical implementation, data-driven computational integration, and evaluation of models and systems outputs against ocean observations. Central to the science is the acquisition of relevant observational data sets and the development of models and numerical techniques, together with data assimilation schemes that quantitatively meld all information sources in accord with the uncertainties of the observations and models. It is the sustained combination