Journal of Climate, 15, 2702-2706.
Adam H. Sobel
Department of Applied Physics and Applied Mathematics and Department of Earth and Environmental Sciences,
Columbia University, New York, NY.
Isaac M. Held
Geophysical Fluid Dynamics Laboratory, Princeton, NJ.
Christopher S. Bretherton
Department of Atmospheric Sciences, University of
Washington, Seattle, WA.
Abstract
Interannual anomalies in tropical tropospheric temperature have been shown to be related to interannual anomalies in tropical mean sea surface temperature (SST) by a simple moist adiabatic relationship. On physical grounds, it is less obvious than it might at first seem that this should be the case. We expect that the free tropospheric temperature should be sensitive primarily to SST anomalies in regions where the mean SST is high and deep convection is frequent, rather than to the tropical mean SST. The tropical mean also includes nonconvecting regions where the SST has no direct way of influencing the free troposphere. However, interannual anomalies of SST averaged over regions of high monthly mean precipitation are very similar to interannual anomalies of tropical mean SST. Empirical orthogonal function analysis of the monthly SST histograms for the period 1982-1998 reveals a leading mode, well separated from the others, whose structure is very similar to a simple shift of the annual and climatological mean histogram, without change of shape. Consequently, many different ways of sampling the histogram will yield similar anomaly time series, and the adequacy of the mean SST for predicting the tropospheric temperature appears coincidental from the point of view of the uncoupled atmospheric problem with given SST. There is a suggestion in the results that changes in the histogram shape may be significant for the tropospheric temperature anomalies associated with some large El Ni\~{n}o events, and that in those events it is indeed the SST anomalies in the convective regions which are most important in controlling the tropospheric temperature.