Global uncoupled storm-resolving simulations using the ICOsahedral Non-hydrostatic (ICON) model with prescribed sea surface temperature show a double band of precipitation in the Western Pacific, a feature explained by reduced precipitation over the warm pool. Three hypotheses using an energetic framework are advanced to explain the warm pool precipitation bias, and they are related to 1) high-cloud radiative effect, 2) too-frequent or highly efficient precipitating shallow convection, and 3) surface heat fluxes in light near-surface winds. Our results show that increasing surface heat fluxes in light near-surface winds produce more precipitation over the warm pool and a single precipitation band in the Western Pacific. Further improvements were stronger near-surface winds over the Western Pacific and a moister and warmer tropical troposphere, but these improvements were not enough to fully overcome the existing biases. Simulations with an increased high-cloud radiative effect did not affect precipitation over the warm pool, and according to the energetic framework, due to compensation between the radiative effect and both, surface heat fluxes and circulation. Moreover, the representation of shallow convection did not affect warm pool precipitation. Thus, our results show the importance of the feedback between winds, surface heat fluxes, and convection for a correct representation of the oceanic tropical rainbelt structure in regions of weak sea surface temperature gradient as the warm pool.

Volcanic aerosol forcing has been reported in the literature to be less effective in changing the earth’s surface temperature than CO2 forcing. This implies a different feedback strength, and therefore different contributions from individual feedback mechanisms. We employ the CMIP6 version of MPI-ESM to understand the reasons for these apparent differences in the ability to change the surface air temperature. Using a highly idealized eruption scenario and comparing it to a doubling and a halving of CO2 concentration, we identify key reasons for changes in the magnitude of the feedback parameter. We show that the “efficacy” [Hansen et al. 2005] of volcanic aerosol forcing depends strongly on the method and the time scale used to calculate it. We argue that the seemingly established result of a lower-than-unity efficacy of volcanic aerosol forcing might only hold under the specific methodological choices other authors have made, but not in general. Furthermore, we find qualitative differences between the cooling and warming simulations, but strong similarities between the 0.5xCO2 and the idealized eruption cases. This hints towards processes, which are not forcing agent-specific, but specific to the sign of the forcing. A pronounced curvature in the N(T) plot (“Gregory plot”) for the cooling scenarios makes the computation of feedback through regression even more sensitive to subjective choices than in the 2xCO2 case. We disentangle the role of ocean heat uptake efficacy and atmospheric feedback processes in the framework of the pattern effect.