Idealized hotspot experiments with a general circulation model

Idealized tropical hotspot experiments are conducted using a GCM coupled to a 20-meter slab ocean model. Hotspots are imposed upon an aquaplanet with globally-uniform 28°C SST, insolation, and trace gas concentrations designed to mimic tropical warm pool conditions. No boundary condition or external parameter other than the Coriolis parameter varies with latitude. A 15-member GCM ensemble is initiated using random atmospheric initial conditions. After a spin-up period with fixed 28°C SST, a 2°C equatorial warm anomaly is imposed along with ocean coupling. Enhanced deep convection quickly develops near the hotspot, forcing a response in the large-scale circulation that resembles that in the Gill model. The strength of convection, the anomalous large-scale circulation, and associated positive wind speed anomalies peak at about Day 15. Negative tropical-averaged wind speed anomalies caused by suppression of eddies occurs after Day 30. The eddies have dominant spatial structure resembling TD-type disturbances, and are suppressed in regions of anomalous low-level easterly flow.

Atmosphere-ocean exchange results in a 60-70 day oscillation in tropical mean (30°N-30°S) oceanic heat content, accompanied by a compensating out-of-phase oscillation in vertically-integrated atmospheric moist static energy. The exchange of energy between the ocean and atmosphere is primarily regulated by wind-driven latent heat flux anomalies forced by variability in both vector mean wind and eddy variance. Beyond Day 70 of the simulation, positive SST anomalies occur across much of the tropical belt and slowly decay toward the 28°C background state.

A west Pacific hotspot experiment with realistic February SST, radiative forcing, and continents produces results qualitatively similar to those in the idealized case. A 15-member ensemble is too small to generate statistically significant results because of an increase in noise relative to the idealized case, however. An idealized 2°C cold spot experiment is also conducted. The response of the ocean-atmosphere system is much weaker than in the hotspot case and is non-oscillatory. The cold anomaly simply spreads slowly after t=0 and gradually decays with time.