Boyle95b Boyle, J. S., 1995b: Estimates of zonally averaged tropical diabatic heating in AMIP GCM simulations. PCMDI Report 25, Program for Climate Model Diagnosis and Intercomparison, Lawrence Livermore National Laboratory, 39 pp. 

The vertical distribution of zonally and seasonally averaged diabatic heating is estimated for 29 GCM AMIP decadal simulations using the thermodynamic equation. Since only the zonally averaged, monthly means were available, the transient and stationary wave components are not included in this budget. The exclusion of these terms limits the useful analysis to the Tropics. The vertically averaged values from the budget computation are compared to the vertically averaged diabatic heating computed directly from the sensible heat and radiative fluxes, and precipitation. The comparison is quite favorable in the Tropics, with the effects of the neglected heat fluxes becoming apparent at about 30° poleward from the Equator. The computations are carried out for the solstitial seasons. Based on the median heating distribution of the 29 models we find the following: (1) The model consensus of near equatorial heating is greater in magnitude and lower (~500 mb) than that computed from the ECMWF analysis by Hoskins et al. (1989). (2) The subtropical cooling tends to be greater in magnitude and higher than the Hoskins et al. computation, although this will be affected by the terms neglected in the budget computation. Consideration of the individual model fields show that (3) there is a large variation in the magnitude and distribution of the tropical diabatic heating amongst the models. The magnitudes in the northern summer vary by more than a factor of two. (4) the amount of seasonal asymmetry about the equator varies widely among the models. For some models the heating maximum remains on the northern side of the equator for both seasons. (5) It is evident that the interactions among the many parameterizations and model formulations obscure any systematic signature of a particular penetrative convective scheme. Finally, given the differences in the heating distributions among the models for this zonally-averaged, seasonally-averaged ten-year data set, it is clear that there is not yet a consensus on the proper parameterization suite to simulate this essential field.