A comprehensive method of estimating the influences of local versus remote sea surface temperatures (SSTs)
on precipitation is developed. The method was applied to two ten-year simulations made with a general circulation model (GCM) and forced with prescribed SSTs. The simulation period spanned from 1 January 1982 to 31 December 1991. The first simulation (called Cs) was forced with naturally varying SSTs taken from the analysis of observations and the second simulation (called Cc) was forced with monthly mean climatology of SSTs used in Cs. Monthly data of evaporation, precipitation, mean vertical velocity and atmospheric moisture convergence were binned by 1 ° C SST intervals and plotted as bin means and within-bin standard deviations. The plots showed that (i) binning captured the averaged trend of SST influences on the monthly fields, but with large standard deviations; (ii) all bin-averaged SST dependences were remarkably similar in the two simulations as well as in the single El Nĩno Southern Oscillation year of 1987; (iii) evaporation increased monotoni -cally with SST up to about 27 ° C after which it plateaued; and (iv) precipitation correlated much more with the vertical velocity than with the local SST.

Monthly precipitation fields were doubly binned with respect to SSTs and boundary-layer moisture convergence (BLqC); data binned in this way were used to compute the partial derivatives of precipitation with respect to SST and BLqC. Together with the total rate of change of BLqC with local SSTs, the rate of change of precipitation with local SST was computed. The remaining precipitation differences were lumped together as all other remote effects. Simulation Cc was used to infer the natural variability of the precipitation required for the statistical significance of the local SST and/or remotely caused changes in precipitation. This analysis categorized all precipitation anomalies into four types: (i) where the local
SST influences were significant, (ii) where the remote SSTs influences were significant, (iii) where both influences were
significant, and (iv) where the natural variability was larger than both influences. Most of the precipitation responses to SSTs were as expected, while their seasonal behaviour revealed that local SSTs contribute to a number of features of the Intertropical Convergence Zone (ITCZ) including the double ITCZ in March – April – May.

Published in 2008 by John Wiley & Sons, Ltd.