This research is intended to demonstrate that climate dynamics understanding and climate predictability can be advanced from the study of long time series of atmospheric climate variables. Complementary microscopic and macroscopic approaches are to be pursued. In the former, the methods of modern non-linear dynamics will be applied to highly-sampled observations to investigate the limits of short-term predictability. In the macroscopic method, the classical methods of statistical mechanics will be applied to slowly-changing climatic variables. Both approaches potentially lead to quantitative estimates of the sensitivity of simulated or observed climates to small perturbations. A program of developing, testing, and applying these methods to a sequence of increasingly sophisticated models and to data is included under this grant. The research will involve the training of undergraduate and graduate students and postdoctoral fellows. This request is important because it examines alternative and potentially fruitful alternative approaches to modeling and predicting the climate system.
