Prediction and Predictability in the Atmosphere and Oceans
MIT OpenCourseWare offers this advanced course on the science of forecasting atmospheric and oceanic systems. It examines why even a physically perfect model cannot guarantee an accurate forecast, covering chaos, error growth, and the sources of uncertainty that limit predictability. Students learn methods for estimating forecast skill in advance, essentially predicting how predictable a given situation is, and survey the operational systems meteorological and oceanographic centers actually use to generate forecasts. Course materials include lecture notes, readings, and problem sets drawn from current research and operational practice, following MIT's standard OpenCourseWare format under a Creative Commons license. The course suits students with a physical science or engineering background who want to understand the mathematical and physical limits on prediction, rather than just how to build a forecasting model. No enrollment or certificate is offered; materials are free to use for self-study.