A simulation is also performed in which SST is fixed, but latent heat fluxes are set to climatological values taken from the control simulation, suppressing the wind-evaporation or “WISHE” feedback (No-WISHE simulation). The amplitude of intraseasonal precipitation variations in the No-WISHE simulation is greatly reduced from the fixed-SST control simulation, showing that WISHE is important to the intraseasonal variability in this model. This helps to explain the reduction in precipitation variability in the 2 meter SOM simulation. Although intraseasonal SST variations are stronger in the 2 meter SOM simulation than in any of the other simulations, these SST variations act in such a way as to diminish the amplitude of equatorial latent heat flux variations. Reducing the mixed layer depth is thus nearly equivalent to eliminating WISHE, which in this model has the effect of weakening equatorial convective variability. The control simulation, and the SOM coupled simulations with mixed layer depths greater than 2 meters, produce near-equatorial intraseasonal precipitation variations that more closely resemble observations than the 2 meter or No-WISHE simulations. Besides the amplitude differences, propagation of wintertime west Pacific equatorial precipitation anomalies in the No-WISHE and 2 meter SOM simulations is eastward at around 13 m/s, as compared to near-observed values of 4-5 m/s in the control simulation and deeper SOM simulations.
These results are broadly consistent with those of Watterson, who found an increase in intraseasonal atmospheric variability as mixed layer depths were decreased to 10m in a coupled model. The slope of the variability as a function of mixed layer depth changes sign at smaller mixed layer depths in the NCAR CAM2.0.1, however. This behavior was predicted by Sobel and Gildor using a highly idealized model. Further experiments are described here with the same idealized model, which help to interpret results derived from the modified NCAR CAM2.0.1.
Supplementary URL: