Air-Sea interactions:Observations and models

Abstract:

The interaction between the atmosphere and ocean consists of intense and episodic exchanges of heat, momentum, and moisture. However, the way in which these episodic events, and higher-scale frequency events in general, influence the coupled atmosphere/ocean system on longer time scales is not well understood. Models have predicted anthropogenic changes in the mean state and variability of various aspects of the climate which are closely tied to surface turbulent fluxes, but disagree on the nature of the changes. The competing influences of increased humidity, which should reduce fluxes, and increased wind speed and variability, which should increase fluxes, in determining the overall flux under a warming climate is not known. One motivating factor for this research is that there are enormous difficulties in understanding the coupled atmosphere-ocean system, given the complexity of two fluids with large-scale circulations that differ in time/space scales and are coupled through relatively small-scale surface fluxes and cloud processes. In order to make a tractable problem, such statistical techniques as joint distribution analysis, clustering, and variabilities of extremes in time and space must be used to simplify these complex relationships. In order to understand how the climate responds to variations in forcing, one necessary component is to understand the full distribution of variability of exchanges of heat and moisture between the atmosphere and ocean. A number of studies recognize the important role of surface heat and moisture fluxes in the generation and decay of important coupled air-sea phenomena. These mechanisms operate across a number of scales and contain significant contributions from interactions between the anomalous (i.e. non-mean), often extreme-valued, flux components. It is important to have a characterization and understanding of these processes for the development of accurate modeling efforts. In this talk I will make comparisons of some of these analyses between satellite data sets, reanalysis products, and CMIP5 models. By focusing to a large extent on distributional characteristics, and particularly to extremes, comparisons of in situ observations and satellite data sets with distributions from coupled models (e.g. CMIP5) can be made directly.