|
Abstract:
|
The study of climate change and global-scale fluids in the climate system (ocean and atmosphere) raises many statistically challenging problems. These range from the difficulties of separating nonstationarity from non-normal behavior---and the significance of apparent trends in phenomena with multiple memory scales---to time-series analysis in the presence of unknown time-errors, to the display error fields in model results, where the dimensionality is overwhelming. One must do statistical inference with short, noisy, records; the interactions of complex physics, chemistry, and biology; and where inferences that are important to society are being drawn. Examples will be given of these various issues, including inferences about the possibility of abrupt climate change, "resonances" in the climate system, nonlinear phase entrainment, and state estimates with order 10^9 dof.
|
