Vegetation-atmosphere-soil nutrient feedbacks in the Amazon for different deforestation scenarios

In recent decades, large areas of the Amazon forest have been deforested and the rainforest's future may be dependent on climate and soil nutrient feedbacks associated with deforestation. This is a two-way biosphere-atmosphere interaction problem: the response of the regional climate system to the land cover varies with the forest growth, which, in turn, depends on climate and nutrient stress. Nutrient stress also varies with forest age, being most severe for young forests and declining as forests mature. Here we use a coupled climate-biosphere model to investigate how these feedbacks interact to control the secondary forest recovery after different deforestation scenarios, looking for a threshold of deforestation that could cause dangerous interference on the Amazon recovery. Results show that the reduction in rainfall is proportional to the amount of deforestation and is more drastic when the deforested area is higher than 40% of the original forest extent. In addition, this simulated precipitation reduction alone is not sufficient to prevent the rainforest regrowth. However, when the precipitation reduction is associated with a soil nutrient stress, a savannization process may start over southern Amazonia ( northern Mato Grosso state), no matter how much is deforested. In this region, a large precipitation reduction in the transition from the dry to the rainy season and an increase in the dry season duration are favorable to the savanna maintenance on nutrient-limited simulations. These results may be a valuable tool for prioritizing forest conservation in this region, which presently has the highest clearing rate in Amazonia.