The response of tropical precipitation to extratropical thermal forcing

The response of the tropics to the cooling or heating of higher latitudes is interesting because of its importance for a variety of problems, including the effects of ice age ice sheets on tropical precipitation and the tropical effects of midlatitude aerosol forcing concentrated in one hemisphere. This issue also tests our understanding of climate dynamics on what constrains the tropical precipitation. Paleoclimate data and modeling studies suggest that cooling or heating confined to the extratropics can perturb tropical precipitation so that it is skewed towards the warmed hemisphere. To study the mechanism by which tropical precipitation is perturbed from the extratropics, we have performed a number of experiments using a comprehensive atmospheric GCM and an idealized moist atmospheric GCM with gray radiative transfer, both of which are coupled to a slab mixed layer ocean. Beneath the mixed layer, cross-equatorial oceanic flux is imposed by cooling the northern extratropics and warming the southern extratropics.;We first study the results from the idealized model and develop a simple theory that predicts the tropical precipitation response. First, given the diffusivity in the extratropics of the idealized model, the changes in poleward energy fluxes are predicted using a one-dimensional energy balance model. Second, this predicted energy flux change is converted into a mass flux and a moisture flux change assuming as given the total gross moist stability (the total atmospheric energy transport per unit mass transport) of the model tropics. The sensitivity of the tropical response to the convection scheme in the idealized model results from the change in the total gross moist stability, which measures the extent to which the Hadley circulation is efficient in transporting energy poleward.;We then consider a set of experiments with a comprehensive model, which suggest that the response of tropical precipitation can be understood in terms of the degree of compensation between the imposed oceanic flux and the resulting response in the atmospheric energy transport in the tropics. This degree of compensation varies widely as a parameter in the convection scheme is modified that controls dry air entrainment into convective plumes. Hence, the magnitude of the ITCZ displacement is very sensitive to the convection scheme. As confirmed by additional calculations with models in which clouds are not allowed to change, the change in the convection scheme affects the extratropical-tropical interactions in the model primarily by modifying the cloud response. From a comparison with the idealized model, it is argued that the degree of compensation consist of a relatively weak robust part (25%) that is dynamically controlled through eddy energy fluxes, and similar to that simulated in the idealized model that has no cloud or water vapor feedback, and a remainder that is feedback controlled that is more likely to be model dependent.;The mechanism described above focuses on energy balance at top of the atmosphere (TOA) and does not explicitly involve surface dynamics. We also study how tropical precipitation response is affected by surface processes such as wind-evaporation-SST (WES) feedback, of which an essential ingredient is that increase in wind speed increases evaporative cooling, leading to a reduction in local surface temperature. We modified both models by deliberately suppressing WES feedback in the bulk formulation for surface heat fluxes and performed the identical set of experiments. When WES feedback is suppressed, compared to the control model experiments, the SST response is weakened in both models whereas the precipitation response stays almost unchanged. These results suggest that constraints on the atmospheric energy budget are more fundamental than surface flux responses, such as those involved in WES feedback, for the communication between the extratropics and tropics.;We further extend the study by considering zonally asymmetric forcings to study the role of stationary eddies. The width of the forced region is varied, and the amplitude of the forcing is altered accordingly to keep the zonally averaged magnitude constant. In response to zonally asymmetric extratropical forcing, the tropical responses are highly zonally symmetric, especially in a comprehensive model. In addition, the zonally averaged responses of precipitation and SST vary little as the width of the forcing is changed. Neither the spatial structure nor the zonal responses are strongly affected by WES feedback.