Lucas Augusto Pereira da Silva, Cristiano Marcelo Pereira de Souza, Edson Eyji Sano, Taya Cristo Parreiras, Édson Luis Bolfe, Bartolomeu Israel Souza, Ramon Santos Souza, Marcos Esdras Leite, Mário Marcos Espírito-Santo, Carolina Cabral das Chagas-Reis, Ivanildo Costa da Silva, João Paulo Sena-Souza, Claudionor Ribeiro da Silva
Aridification driven by future climate change is expected to significantly change vegetation vigor in global drylands. However, its impacts on the Brazilian semiarid region, one of the world most populous and biodiverse drylands, remains poorly understood. In this study, we employed the Random Forest (RF) model trained with historical edaphoclimatic data to predict changes in vegetation vigor under future climatic scenarios (2061–2080), defined by the Shared Socioeconomic Pathways (SSPs), including an optimistic scenario (SSP1-2.6) and a pessimistic scenario (SSP5-8.5). Model performance was satisfactorily, with the Holdout Test exhibiting R 2 = 0.93, RMSE = 0.02, and MAE = 0.02. Seasonal forests are projected to face a decline in greenness (SSP1-2.6 = down to −3.23% and SSP5-8.5 = −7.23%). Similarly, ombrophilous forests are expected to lose between −0.72% and −6.04% in vegetative vigor in the optimistic and pessimistic scenarios, respectively. Despite its natural adaptation to water deficit conditions, the Caatinga biome is also expected to show a significant decrease in vegetation vigor (−2.90% to −3.40%). Ecotonal zones are also projected to lose vegetative vigor (−2.08% to −4.31%). In contrast, savanna formations are expected to undergo a greening trend (+0.78%). Our results provide valuable insights for environmental planning and management, highlighting the most vulnerable vegetation types to future climate-induced aridification.