Changes in the Seasonal Cycle of Precipitation and Heat Extremes Over Tropical Land Regions: An Amazon Story

Abstract

We analyze observations and experiments from a super‐ensemble of Earth System Models (seven models, each with initial condition ensembles of three members or more) over tropical land regions, aiming at detecting delays in the onset of the rainy season and concurrent changes in the statistics of heat extremes. The expectation is that if the dry season is lengthening, the conditions would favor hotter temperatures in the last days of the already hot, dry season. After considering daily precipitation and maximum temperature behavior over the period from 1979 to present, both observed and modeled, we find that only the Amazon region shows a significant and consistent trend across the data sets toward a delay in the onset of precipitation and more numerous and intense hot days during the longer drier season. These days are warmer than what would be expected from the overall mean warming over the decades. None of the other regions shows such consistent and significant changes. Focusing on the Amazon, both the increase in greenhouse gas concentrations and the decrease in aerosol forcings are necessary to explain the observed behavior across models, with more of the same in store for the future. For most models however, aerosol appears as the dominant forcing to explain the historical behavior. We propose that the role of the two forcings, smaller or larger, depending on the model, in warming the Northern and Southern hemispheres differentially, and therefore shifting the ITCZ and altering its seasonal migration, may be key in explaining such model diversity.