Spatiotemporal scenarios for deforestation in Brazil's legal Amazon
Andrea Santos Garcia, Rafaella Almeida Silvestrini, Alvaro Maia Batista, Lais Ferreira, Marek Hanusch, Philipp Kollenda, Carla Cristina Solis Uehara, Dieter Wang
In light of the consequences of deforestation in the Brazilian Amazon for climate change and biodiversity erosion at global and regional scales, this study explores future deforestation scenarios, their application in a publicly available online dashboard [https://forestatrisk.ipam.org.br], and relationships with public policies. We move beyond deforestation projections, which use a constant rate or moving average assumptions to describe the historical reference level (HRL). Instead, we project deforestation using a novel business-as-usual (BAU) baseline, which predicts the amount of forest loss due to macroeconomic factors alone. We then develop a policy scenario with stronger conservation effort, where non-designated public forests are designated as protected areas (GOV). We estimate our model using data from 1999 to 2021, validate using data from 2022 and forecast deforestation and its spatial allocation from 2023 to 2025. Total observed deforestation area is 32% larger in 2022 compared to the BAU baseline, likely indicating weakened forest governance in 2018-2022. Still, we find a good spatial allocation match between modelled BAU deforestation areas and observed deforestation areas, with an overall mean spatial accuracy corresponding to 80% with a 12 x 12 km window size, and 90% with a 20x20 km. If deforestation would continue at the HRL rate, we would accumulate 35% more deforestation until 2025 than what we estimate for our preferred BAU baseline, indicating that macroeconomic conditions are projected to be conducive for reduced deforestation. All models show large areas of expected deforestation concentrated in central Pará (PA) and in southern Amazonas (AM), especially along the main roads. Another smaller deforestation patch is observed along the border between the Brazilian Amazon and the Cerrado Biome, an older deforestation frontier. In the GOV scenario we simulate that leakage to other areas mainly occurs in rural settlements and rural properties. The study contributes to a better understanding of the factors influencing the amount of deforestation and its distribution in time and space. The spatially explicit model can help identify risk areas for targeted policy responses as well as shed light on where leakage can be expected when local protection mechanisms are enforced.
Event: World Bank Land Conference 2024 - Washington
Only personal, non-commercial use of this document is allowed.