Simulating moisture-vegetation feedbacks in the Amazon under drought and deforestation scenarios

The Amazon rainforest, a global biodiversity hotspot and home to over 40 million people—2.2 million of whom are Indigenous—plays a critical role in the global regulation of water and carbon cycles. However, its unique biocultural diversity is increasingly threatened by climate and land-use changes, which could shift vegetation in multi-stable forest areas to savannah- or grassland-like states. Satellite-based observations, Earth system models, and rainfall exclusion experiments provide evidence of the rainforest's critical dependency on precipitation and seasonality. Additionally, complex systems approaches suggest that forests in bistable areas are maintained by cascading moisture recycling, a process that is significantly reduced by regional deforestation.

This research employs  the dynamic global vegetation model LPJmL (version 5.9), incorporating variable tree rooting strategies and coupled with moisture network data derived from the Lagrangian moisture transport model UTrack. The observation-based monthly moisture networks for the period 2003–2014 proportionally redistribute evapotranspiration from LPJmL over the Amazon basin as precipitation, providing a partially dynamic representation of the moisture-vegetation feedback. Future scenarios, including increased drought frequencies (based on the major droughts of 2005 and 2010 as analogs for future extremes)and two deforestation projections (based on the Governance and Business as Usual scenarios from Soares-Filho et al. (2006)), are implemented to analyse rainfall changes and the forest's local and telecoupled moisture response in LPJmL. We also provide a first estimate of the collective contribution of Indigenous Peoples’ Lands to terrestrial precipitation in the Amazon, by explicitly accounting for atmospheric water flows originating from Indigenous territories as in the data provided by Garnett et al. (2018). 

These findings add to our understanding of forest-water interactions from a moisture recycling perspective, assessing the impacts of drought and deforestation while highlighting the role of Indigenous land management. Advances in modelling could support future assessments of forest resilience and tipping risks, providing critical inputs for forest management and underscoring the urgency of effective climate mitigation.