Winter Quarter 2005. (2009)

Instructor: Greg Hakim

Class meets: MWF 2:30-3:20. ATG 310C.

|  syllabus  |  resources  | lecture log  |

Course description:

Predictability involves the capacity for estimating the future state of a system. Perhaps the most successful illustration of predictability is numerical weather prediction, which is one of the great achiements of 20th century science. Similarly, climate predictions are now considered reliable enough to motivate decision makers to consider future impacts on the environment in present-day policy.

This class is about understanding predictability, and how it may be exploited through a synthesis of data from numerical models and observations. Data assimilation involves the fusion of information from these two sources. In addition to building a foundation for modern approaches to predictability based on dynamical systems theory, students will gain practical experience by using DART (the Data Assimilation Research Testbed), which provides a friendly environment to test ideas on a built-in hierarchy of models. For example, idealized models such as the classic 3-variable model of Lorenx (1963) allow rapid testing of ideas, which may then be explored in more complex models such as the CAM GCM (Community Atmosphere Model), or WRF (Weather Research and Forecast model).