By Weidong Guo and Jun Ge's team, School of Atmospheric Sciences
In recent years, global forest cover has been declining continuously, with Brazil in the equatorial region showing the most pronounced loss. In 2003 alone, Brazil lost over 40,000 km² of forest, the highest in the world. Tropical forests are critical for global biodiversity and terrestrial carbon storage, and they are also experiencing the most rapid habitat changes. Therefore, curbing deforestation and forest degradation is vital for mitigating climate change and biodiversity loss. Reductions in forest cover significantly alter the surface energy balance and regional water cycle, affecting agriculture, water resources, and ecosystems. However, the mechanisms by which precipitation responds to forest changes remain uncertain, especially under real-world deforestation conditions, where systematic studies are lacking.
Recently, the team of Weidong Guo and Jun Ge from the School of Atmospheric Sciences collaborated with Professor Dominick Spracklen from the University of Leeds, UK. They employed a regional climate model WRF-WVT embedded with a water vapor tracer module to evaluate the extent and mechanisms of forest cover change impacts on dry season precipitation and the water cycle in the Brazilian Amazon. Unlike previous studies that mainly used idealized land cover changes, this research reconstructed real vegetation changes based on satellite-observed surface properties from 2002 to 2015.
The study showed that from 2002 to 2015, forest cover in the Brazilian Amazon decreased by an average of 3.2%, resulting in a 5.4% reduction in dry season mean precipitation, closely matching satellite observations. Further analysis found that evapotranspiration during the dry season decreased by 4.2 mmžmonth-1, and net water vapor flux decreased by 2.2 mmžmonth-1. Notably, reduced non-local water vapor input significantly amplified the impact of deforestation on the water cycle, suggesting that models assuming a direct proportionality between precipitation and evapotranspiration may underestimate the actual impact of deforestation.
Figure 1: Schematic diagram of the impacts of forest loss on the atmospheric water cycle. Green arrows represent the control simulation, orange arrows represent the deforestation simulation. Components such as available water, evapotranspiration, vapor inflow/outflow, and tagged/untagged precipitation are indicated in mmžmonth-1. Italicized numbers show precipitable water and its tagged and untagged components. Blue (red) indicates water cycle reduction (increase) caused by deforestation.
Through water vapor tracer analysis, the team found that 76.9% of the precipitation reduction in the dry season was attributable to decreased non-local water vapor input, and only 23.1% to reduced local evapotranspiration. Additionally, atmospheric responses to local water cycle changes amplified precipitation reduction: deforestation reduced surface evapotranspiration and latent heat flux, leading to surface drying and warming, reduced low-level specific humidity, and decreased Convective Available Potential Energy (CAPE), thereby suppressing convective activity and precipitation. Quantitative analysis showed that reduced precipitation efficiency accounted for 84.5% of the dry season precipitation decline, while cooling in the upper troposphere further limited high-altitude vapor input.
Moreover, the study highlighted that under high humidity conditions, deforestation may actually enhance daily precipitation, supporting recent theories about deforestation intensifying storm activity. Analyses at multiple temporal scales revealed that changes in precipitation efficiency at seasonal scales were driven by changes in atmospheric moisture, emphasizing the need to consider multi-timescale dynamics in water cycle research.
Figure 2: Vertical profiles of responses to deforestation in the Brazilian Amazon: (a) mean flow convergence (black; kgžm-2žs-1žhpa-1), temperature (red; °C), (b) vertical velocity (mmžs-1), (c) CAPE (black; Jžkg-1), specific humidity (blue; kgžkg-1).
This study systematically revealed the nonlinear impacts of Amazon deforestation on precipitation and regional water cycles, emphasizing the key role of atmospheric feedbacks in precipitation changes. The findings support more scientific forest protection and climate policies, highlighting the importance of sustainable forest management in maintaining regional climate stability.
The research, titled "Recent forest loss in the Brazilian Amazon causes substantial reductions in dry season precipitation," was published in AGU Advances, a high-impact journal of the American Geophysical Union (AGU). Yu Liu, a 2020 Ph.D. student from the School of Atmospheric Sciences, is the first author. Professor Weidong Guo is the corresponding author. Co-authors include Professors Dominick Spracklen, Joseph Holden, and Douglas Parker from the University of Leeds, and Assistant Professor Jun Ge from the School of Atmospheric Sciences.
The work was supported by the National Natural Science Foundation of China (Key Project 42130602, General Project 42375115), the EU Horizon 2020 Programme (771492), and the UK Natural Environment Research Council (NE/Z00005X/1).
This research was featured in AGU's Eos as a "Research Spotlight", highlighting its significant contribution to climate change and forest hydrology feedback studies. This column selects research with broad scientific influence and social relevance, widely recognized in the international climate science community. It's worth noting this is a continuation of the team's work on regional water cycle and land cover interaction mechanisms. Their previous study on the Loess Plateau's ecological restoration and rainfall feedbacks in China was also selected as an "Editor's Highlight" by Eos (link), demonstrating the broad application potential of multi-scale climate-ecology coupled modeling in tackling environmental and climate challenges.
Paper DOI: https://doi.org/10.1029/2025AV001670
Eos Coverage: https://eos.org/research-spotlights/deforestation-is-reducing-rainfall-in-the-amazon
Related Paper: https://doi.org/10.1029/2023GL102809
(Provided by the School of Atmospheric Sciences)
June 23, 2025