Mean and extreme tropical precipitation changes caused by the uniform and spatially varying components of anthropogenic forcing

Anthropogenic aerosol and greenhouse gas forcings affect both the mean and spatial pattern of surface temperatures differently. These surface temperature components, in turn, are known to affect tropical precipitation patterns -- both extreme and time mean -- via different mechanisms. We explore this behavior by imposing SST anomalies representative of either historical aerosol forcing, its tropical mean (TM) value at all ocean gridpoints (i.e. a uniform component), or its TM-subtracted values (i.e. a spatially varying component) and an analogous triplet for historical well-mixed greenhouse gas forcing in two atmospheric general circulation models (AGCMs) from the NOAA Geophysical Fluid Dynamics Laboratory (GFDL), AM2.1 and AM3.

The mean precipitation response to the full SST anomalies is well captured by the linear sum of the responses to the TM and TM-removed cases. The ITCZ shifts southward in both the full and TM-removed aerosol cases, unlike the aerosol TM case or any of the greenhouse gas cases; this is consistent with prior studies of the ITCZ response to asymmetric forcing in the two hemispheres. Outside the ITCZ latitudes, zonal mean precipitation change in the full aerosol and greenhouse gas cases is well captured by a simple thermodynamic scaling (c.f. Held and Soden 2006, J. Climate) using only the TM change.

Regarding extreme precipitation, we focus on changes to wet season rainfall in the Sahel as a case study. Prior studies have established model dependence of the abatement (or lack thereof) in the 21st century of the observed 20th century Sahel drought. We find coherent responses in AM2 and AM3 to the TM-removed cases associated with shifts in the Atlantic ITCZ, but model-dependence to the TM cases. Preliminary results from a new GFDL high-resolution AGCM, HiRAM, shed further light on the mechanisms responsible for this model dependence.