A vegetation‐atmosphere interaction study for Amazonia deforestation using field data and a ‘single column’ model

Abstract

AbstractThe recent high deforestation rates in Amazonia have raised the question of how climate, mainly the precipitation patterns, might be affected as pasture land replaces the primary forest. This work initially shows how the dry season soil moisture and water stress in typical Amazonian pasture modifies the energy partitioning at the surface. Low‐level moisture convergence is a primary mechanism to trigger convective cloud formation and precipitation. The relationship between moisture convergence and local surface evaporation in generating precipitation in Amazonia is studied utilizing a one‐dimensional ‘single column’ model (SiB‐1D). SiB‐1D couples a surface‐vegetation model (SiB) to a physical parametrization of deep convection (Kuo scheme), radiation, turbulent diffusion and large‐scale precipitation. Model simulations for short periods (2 days) show its ability to calculate the Amazonian surface energy‐balance components and boundary‐layer dynamics when compared with field observations from the Rondǒnia Boundary‐Layer Experiment and the Anglo‐Brazilian Amazonian Climate Observation Study. The model was further used to test the sensitivity of the deep convection scheme to a range of typical low‐level moisture convergence situations in a second experiment, and a third test investigated the local convective precipitation generated over forest and grass vegetation as a function of available soil moisture. Results showed that the rainfall over forest vegetation appeared to be almost insensitive to soil water stress whereas reduced precipitation was generated over pasture. When available soil water fell below a threshold of 60% the calculated precipitation over the pasture sites rapidly declined. Although these results are confined by the short integration period and the initial atmospheric profiles, they help to strengthen the notion that deforestation reduces evaporation and convective precipitation, especially during the dry season: a result already indicated by some previous general‐circulation model experiments.