A Possible Root Cause of the Equatorial Atlantic Systematic Biases in Global Climate Models

Most Global Climate Models (GCMs) suffer from systematic biases of surface westerlies and a reversed zonal gradient in sea surface temperature in the equatorial Atlantic. These biases have persisted through years of considerable model improvement. In this study it is hypothesized that unrealistic vertical profiles of diabatic heating over South America is a root cause of the equatorial Atlantic systematic biases. This hypothesis is based on diagnoses of diabatic heating profiles in observations, global reanalysis, GCM simulations, and numerical experiments using a regional atmospheric mesoscale model.

This investigation suggests that weaker vertical heating gradients in the mid Amazonian troposphere, due to weaker low-level heating, are likely to induce more severe spring equatorial Atlantic westerly biases. This issue could be related to difficulties of models to reproduce realistic convection in the deep Amazonia. Also, although most GCMs display more severe heating profile biases in equatorial Africa than in South America, the equatorial Atlantic wind regime seems to be more sensitive to the Amazonian heating forcing.