The influence of historical sea-surface temperature patterns on regional precipitation trends

Jaydeep R. Pillai, Kyle C. Armour, David S. Battisti

State-of-the-art coupled global climate models (GCMs) fail to simulate key features
of observed seasonal precipitation trends since 1980, including drying of the southwestern US,
the southeastern US, East Africa, and subtropical South America, as well as wetting of the
Maritime Continent and the Amazon. They also fail to simulate the sea-level pressure (SLP) trends
since 1980 associated with a poleward shift of the North Pacific storm track in the mid-latitudes
and a strengthened Pacific Walker Circulation. We show that state-of-the-art atmosphere-only
climate model ensembles driven by observed sea-surface temperatures (SSTs) simulate historical
precipitation and SLP trends that are more similar to those observed in the regions noted above,
suggesting that the observed pattern of SST changes has shaped regional precipitation and SLP
trends. Analysis of the coupled and atmosphere-only model ensembles reveals that multidecadal
SST patterns similar to those of the interannual El-Ni˜ no Southern Oscillation are responsible for
some of the regional trends simulated. The tropical Pacific zonal SST gradient is found to have
substantially contributed to observed drying over the southwestern and southeastern US, signifying
a key role for tropical Pacific warming patterns in future precipitation trends in these regions.