Monthly CAPE generation rates predict tropical precipitation.

Predicting the spatial distribution, intensity, and variability of tropical precipitation is important in the context of present and future climate change. However, climate models have consistently failed to simulate tropical precipitation correctly, even as their resolution has progressively improved (Tian and Dong, 2020). Convective Available Potential Energy (CAPE) is a measure of the amount of buoyant energy usable by convection. Inspired by convective quasi-equilibrium theory (Arakawa, 1974), we test whether the rate of CAPE generation is a good indicator of tropical precipitation in the past four decades of the Japanese Reanalysis for Three Quarters of a Century. We find that CAPE generation predicts the spatial distribution and intensity of observed tropical precipitation significantly better than CAPE itself, as well as precipitation trends and extreme seasonal precipitation. CAPE generation is therefore a good proxy to study convective events which are too small to be directly simulated at the resolution of climate models. Further, we decompose the physical sources of buoyancy generation and find that local evaporation is the main energy source in the tropical rainbands, and surprisingly, heat and moisture convergence play a minor role in providing buoyancy for convection. Based on these conclusions, it may be more useful to study air-sea fluxes and local evaporation as a key to improving climate precipitation simulations.