Modeling the Impacts of Deforestation: local drying of the atmosphere and potential effect on downwind precipitation.

The Amazon rainforest is a vital component of the hydrological cycle of South America. Its evapotranspiration is an essential supply of atmospheric moisture for precipitation in more southern regions of the continent. The potential impacts of deforestation on precipitation in these distant regions are yet not fully understood.

The present research project aims at quantifying the deficit of evapotranspiration occurring at the location of deforestation, focusing on the southern part of Amazonia, which has experienced intense deforestation since the 80s. We first use the ORCHIDEE land surface model forced with the reanalysis dataset CRU-JRA to simulate the impacts of an imposed land cover change: from observed states of vegetation cover to a massive extension of croplands. The ORCHIDEE model computes all the components of evapotranspiration, giving, in turn, the expected deficit of atmospheric moisture at the location of the land cover change.

Then, thanks to existing datasets connecting any place on Earth with the area that supplies its moisture through the atmosphere, we link this deficit with downwind locations highly dependent on this upwind evapotranspiration for its precipitation. From there, we draw hypotheses about the potential changes in precipitation amounts and seasonality.

In the project’s second phase, these hypotheses are tested against land-atmosphere coupled simulations produced with the IPSL global climate model, nudged to winds from the ERA5 reanalysis. The model grid is zoomed on the South American continent to better describe the atmospheric transport in the region. The land-atmosphere coupled simulations provide information on the atmospheric feedback induced by the land cover change, confirming or invalidating the hypotheses. Since land cover affects not only water fluxes but also energy fluxes, the coupled model experiments give us insights into the atmospheric processes at stake, the changes in cloudiness and local convection, and the potential shifts in precipitation location or timing.