The importance of the diabatic exchange of energy between the ocean and the atmosphere in regulating sea surface temperature in the tropical Pacific region is much stronger than at other latitudes. Understanding the coupled system in the western tropical Pacific in particular is dependent upon our understanding of the complex feedbacks between the ocean and the atmosphere in this region. The low wind speeds in this region, combined with the high precipitation amounts, imply that the upper ocean is very sensitive to changes in surface heat fluxes. A number of negative feedbacks have been suggested as explanations for the small range of observed sea surface temperatures. One of the important components in these hypotheses is the cloud/sea surface temperature feedback. A key factor in differences in climate feedbacks between general circulation models is the parameterization of cloud feedback processes. In the tropics the simulations of the mean climate system are substantially affected by the convection paramaterization used. In order to evaluate the overall ability of the model to reproduce the observed atmosphere/ocean system, it is desirable to evaluate the various feedbacks operating within the model itself. Similar results are possible from models with very different feedback components that compensate each other.

In order to address the thermodynamic coupling of the ocean-atmosphere system in the western Pacific, this study evaluates feedbacks between the tropical Pacific ocean and atmosphere from both a single-column coupled atmosphere-ocean model and data. A single-column model (SCM) provides a computationally inexpensive method to investigate various feedbacks within the model and compare with data. The model is however limited in its uses. An important feature of the coupled atmosphere-ocean tropical system is the effect of the gradient of SST from the western to the eastern Pacific. The resulting large-scale dynamics and feedbacks with the sea surface temperature cannot be studied within the confines of a single column model. This does not limit however our ability to determine the local feedbacks occurring within the model or within the natural system. The causality of values regulating the sea surface temperature and possible feedbacks are analyzed similarly between the model simulations and the data using Granger causality time series techniques. These techniques allow us to determine whether one component of the heat flux is causing variability in the sea surface temperature, and vice versa. In this talk we will describe the formal feedback analysis for evaluating the model response and comparison of the observed feedback strengths with model simulations. The time period and location of this study is the TOGA COARE intensive observation period in the western equatorial Pacific during late 1992 – early 1993.