Atmospheric circulation types controlling rainfall in the Central American Isthmus

Abstract

Rainfall mechanisms in the Central American Isthmus are controlled by complex physical interactions across spatial and temporal scales, which are reflected on the dynamics of atmospheric circulation patterns affecting the region. However, physical mechanisms and their relationships with thermodynamic distributions connected to overturning circulations remain elusive. Here, a set of six recurrent daily atmospheric patterns, or weather types (WT), is defined using a k-means++ clustering algorithm on standardized fields of Convective Available Potential Energy (CAPE) and winds at 925, 850, and 200 hPa. The relationships between these weather types, their temporal characteristics, and anomalous distributions of moisture flux divergence, equivalent potential temperature (saturated and unsaturated), and observed rainfall are used to describe physical processes controlling the latter, for all seasons. Regional observed rainfall is analysed from a set of 174 automatic stations from all countries from Mexico to Panama. By modulating vertically integrated moisture fluxes, these weather types, and the different climate drivers linked to them, control the temporal and spatial rainfall characteristics in the region, especially over the Pacific side of the isthmus. During some stages of the regional rainy season, described by two weather types, thermal anomalies in convective quasiequilibrium characteristic of the upward branch of the Hadley cell force westerly flow over Central America, enhancing rainfall. While during other stages, the enhancement of the trades and the displacement of convection to the ITCZ area over the eastern tropical Pacific, characteristic of the midsummer drought, diminishes rainfall. This study sets the stage for a better understanding of the mechanistic relationship between these weather types and rainfall characteristics in general, like onset, demise, and duration of rainy seasons. Hence, these results can inform process-based model diagnostics aiming at bias-correcting climate predictions at multiple timescales