Sensitivity of Tropical Tropospheric Temperature to Sea Surface Temperature Forcing*

Abstract

Abstract During El Niño, there are substantial tropospheric temperature anomalies across the entire tropical belt associated with the warming of sea surface temperatures (SSTs) in the central and eastern Pacific. The quasi-equilibrium tropical circulation model (QTCM) is used to investigate the mechanisms for tropical tropospheric temperature response to SST forcing. In both observations and model simulations, the tropical averaged tropospheric temperature anomaly 〈T̂′〉 is approximately linear with the tropical mean SST anomaly 〈T′s〉 for observed SST forcing. Regional SST anomaly experiments are used to estimate regional sensitivity measures and quantify the degree of nonlinearity. For instance, SST anomalies of 3°C in the central Pacific would give a nonlinear 〈T̂′〉 response about 15% greater than a linear fit to small SST anomaly experiments would predict, but for the maximum observed SST anomaly in this region the response differs by only 5% from linearity. Nonlinearity in 〈T̂′〉 response is modest even when local precipitation response is highly nonlinear. While temperature anomalies have large spatial scales, the main precipitation anomaly tends to be local to the SST anomaly regions. The tropical averaged precipitation anomalies 〈P′〉 do not necessarily have a simple relation to tropical averaged tropospheric temperature anomalies or SST forcing. The approximate linearity of the 〈T̂′〉 response is due to two factors: 1) the strong nonlinearities that occur locally tend to be associated with the transport terms, which become small in the large-area average; and 2) the dependence on temperature of the top-of-atmosphere and surface fluxes has only weak nonlinearity over the range of 〈T̂′〉 variations. Analytical approximations to the QTCM suggest that the direct impact of climatological SST, via flux terms, contributes modestly to regional variations in the sensitivity α of 〈T̂′〉 to 〈T′s〉. Wind speed has a fairly strong effect on α but tends to oppose the direct effect of SST since cold SST regions often have stronger climatological wind, which would yield larger slopes. A substantial contribution to regional variation in α comes from the different reaction of moisture to SST anomalies in precipitating and nonprecipitating regions. Although regions over climatologically warm water have a slightly higher sensitivity, subregions of El Niño SST anomalies even in the colder eastern Pacific contribute substantially to tropospheric temperature anomalies.