Simulating Precipitation Reductions from Land-Use Changes in South America: A Novel Emulator Approach

The Amazon rainforest faces mounting pressure from deforestation, resource extraction, and infrastructure development, with approximately 20% of its forest cover lost in recent decades. These changes, alongside rising temperatures and shifting precipitation patterns, are severely impacting the forest’s resilience Deforestation not only reduces local evapotranspiration and alters surface energy balance—leading to declines in precipitation and increases in temperature—but also disrupts downstream rainfall through changes in water vapor transport, affecting regions dependent on Amazonian moisture.

While Earth System Models (ESMs) offer critical insights into these impacts, their high computational demands limit the range of scenarios they can assess. To overcome this, ESM emulators such as the IMOGEN system provide efficient, pattern-scaled projections. However, existing emulators often fail to incorporate essential local climate feedbacks, which are critical for understanding the Amazon’s resilience to climate change and land-use shifts.

This study enhances the IMOGEN/PRIME emulator to account for localized rainfall changes driven by upstream land-use alterations and deforestation. Using the WAM-2layers model with ERA5 data, we generate sensitivity matrices to quantify how evapotranspiration (ET) from different Amazon regions contributes to precipitation elsewhere. These are combined with ET anomalies simulated by the JULES land-surface model under various land-use scenarios. Scenarios are derived from the LuccME framework (Aguiar et al., 2016) and include: Sustainability, reflecting socio-economic and environmental advancements; Fragmentation, representing resource depletion and inequity.; Middle of the Road, a mix of both; Extreme cases, such as total South American deforestation, are also assessed.

By combining ET anomalies with water vapor transport sensitivities, precipitation change patterns are spatially mapped for each scenario and incorporated into IMOGEN. This integration allows for simulations of cascading effects from land-use changes on regional precipitation and climate.

The enhanced emulator offers a powerful framework to assess deforestation-driven climate impacts, including their effects on forest resilience and biogeochemical cycles. This approach provides a comprehensive evaluation of Amazon forest dieback risks under diverse CMIP6-aligned scenarios, delivering critical insights for conservation and sustainable land management strategies.