National Oceanic and
Atmospheric Administration
United States Department of Commerce


FY 2018

The effects of external forcing and internal variability on the formation of interhemispheric sea surface temperature gradient trends in the Indian Ocean

Dong, L., and M.J. McPhaden

J. Climate, 30(22), 9077–9095, doi: 10.1175/JCLI-D-17-0138.1 (2017)

A striking trend of the Indian Ocean interhemispheric gradient in sea surface temperatures (SSTs) developed during the recent global warming hiatus. The contributions of external forcing and internal variability to this trend are examined in forced climate model experiments. Results indicate that the observed negative trend was strong by historical standards and most likely due to internal variability rather than to external forcing. Anthropogenic aerosol forcing favors negative gradient trends, but its effects are countered by greenhouse gas forcing, and both are weak relative to internal variability. The observed interhemispheric gradient trend occurred in parallel with a negative phase of the interdecadal Pacific oscillation (IPO), a linkage that is also found in climate models. However, the physical mechanisms responsible for these gradient trends in models differ from those in ocean reanalysis products. In particular, oceanic processes via an increased Indonesian Throughflow (ITF) transport into the Indian Ocean forced by stronger Pacific trade winds are the principal cause of the observed negative SST gradient trend during 2000–13. In contrast, atmospheric processes via changing surface wind stress over the southern Indian Ocean remotely forced by the IPO appear to play a dominant role in changing the interhemispheric SST gradients in climate models. The models underestimate the magnitude of the IPO and produce changes in the ITF that are too weak owing to their coarse spatial resolution. These model deficiencies may account for the differences between the simulations and observations.

Feature Publications | Outstanding Scientific Publications

Contact Sandra Bigley |