Tropical precipitation changes in response to thermodynamic forcing

Mean precipitation and column-integrated water vapor (CWV) in tropical oceans (30S-30N) have been increasing since 1988 based on estimate by more reliable satellite measurements due to both global warming and climate oscillations at interannual-interdecadal scales. In this study, we apply the Partial Least Squared Regression method to the fields of precipitation and CWV against that of SST to regress out natural climate variability in an attempt to extract the forced response by global warming. Linear trends in the residual data show 0.33% and 0.4% per decade while raw data show 1.5% and 1.2% increase in mean precipitation and CWV, respectively (or multiply by 10 to % change per degree warming assuming 1 degree warming per 100 years). While the trends in regressed data are more consistent with the theoretical thermodynamic response and in line with the change obtained from global warming experiments in General Circulation Model, more physical evidence is required to attribute it to global warming. We adopt the cloud resolving model (CRM) developed by Jung and Arakawa (2008) to address this issue. Two pairs of experiments are carried out with interactive radiative and cloud microphysical processes. The first pair are designed to have uniform SST specified at 300K and 303K, respectively. The second pair have the same SST gradient [a quarter domain of warm pool that linearly decrease to a surrounding cold pool of a quarter domain] but a uniform difference of 3K. Each experiment is integrated for 100 days to reach quasi-equilibrium, and the last-50-day is analyze. Linear trends in both pairs show about 8% and 4% increase per degree warming in CWV and mean precipitation, respectively. Despite very different circulation, and associated stability and cloud organization in the two pairs of experiments, the model responses to uniform thermodynamic forcing is consistent with trends changes in the regressed water vapor and precipitation data.