Tipping the Amazon Rainforest: Regional deforestation and land-climate interactions

The Amazon rainforest is globally relevant and is considered a tipping element in the global climate system. Studies suggest that deforestation in the Amazon by around 30% may disturb regional convective rain patterns, which could increase drought frequencies and intensities locally and, may activate a cascade of tipping elements in the global climate system. Here, we aim to assess the relationship between deforestation and climate responses at a convection-permitting scale by employing a non-hydrostatic mesoscale Weather Research and Forecasting (WRF) model. For this, we first developed a spatially explicit deforestation model for the South-Western Amazon to see an effect of deforestation intensity ranging from 10% to 60%, and then based on 30% deforestation, we further see the role of deforestation pattern (e.g. deforestation alongside the roads, as a large single or small multiple circular plots, and their geographical positions), and shifts to anticipated land use. Then for each deforestation map, we simulate the land-atmosphere interactions and responses in the regional rainfall and temperatures by dynamically downscaling the ERA5 reanalysis using WRF for the year 2020 at 5km spatial resolution and by explicitly resolving convection. We assess non-linearity in the land-climate interaction to different combinations of deforestation quantities and deforestation patterns. Our preliminary results show a general pattern of decreasing mean and extreme rainfall with deforestation where the geographical location and the pattern of deforestation also play a role. The study will provide an insight into whether the employed quantitative methods are able, or good enough, to simulate relevant processes between the biosphere and the atmosphere that could promote assessing tipping points in the Amazon.