Seven years of net surface solar irradiance (S) derived from cloud information provided by the International Satellite Cloud Climatology Project and 4 years of surface latent heat flux (E) derived from observations of the special sensor microwave imager were used to examine the relation between surface heat fluxes and sea surface temperature (Ts) in their global geographical distribution, seasonal cycle, and interannual variation. The relations of seasonal changes imply that evaporation cooling is significant over most of the ocean and that solar heating is the main drive for the change of Ts away from the equatorial wave guide where ocean dynamics may be more important. However, Ts is not the most direct and significant factor in the seasonal changes of S and E over most of the ocean; the solar incident angle may be more important to S, and wind speed and air humidity are found to correlate better with E. Significant local correlations between anomalies of Ts and S and between anomalies of Ts and E are found in the central equatorial Pacific; both types of correlation are negative. In this area, organized deep convection overlies the warm ocean, forms high clouds, and reduces S, while the low wind speed and high humidity that result from surface convergence reduce E. The negative correlation is not present in the surrounding areas where equally warm water and strong Ts anomalies are found under a subsiding atmosphere without similarly strong S and E anomalies.

Correlation between anomalies of temperature tendency and the fluxes is weak, indicating that other factors are more influential in changing upper ocean heat balance during El Ni¿o. The result shows that the relations between Ts and the flux components, in annual and interannual timescales, are not universal and not consistent with the local negative feedback postulations which require that an increase in Ts would result in an increase in local evaporative cooling and a decrease in local solar heating of the ocean. Large-scale atmospheric circulation changes clouds, winds, and humidity; they, in turn, influence the fluxes significantly. The influence of ocean dynamics in changing Ts in the tropical ocean cannot be ignored. ¿ American Geophysical Union 1994