- A study looking at land and atmosphere interactions in the Amazon Basin across four decades found that 52-72% of the rainfall decline in the southern Amazon is due to large-scale deforestation.
- Between 1980 and 2019, annual precipitation in the southern Amazon declined by 8-11%, with most of the region losing on average 7.7% of its forest cover over largely the same period.
- The research also indicates that climate models might underestimate the contribution of deforestation to precipitation reduction by as much as 50%, which could mean that rainfall thresholds in the Amazon could be crossed earlier than expected.
Forest loss, along with climate change, is changing the resilience of the Amazon Rainforest. By disrupting the movement of moisture through the atmosphere, deforestation is reducing rainfall and extending the dry season, especially in the southern Amazon Basin. But according to recent research, the impacts of large-scale deforestation could be much bigger than climate models have estimated for the region.
A study published in Nature Communications found that between 52% and 72% of the rainfall decline in the southern Amazon Basin over the last four decades can be attributed to deforestation. Between 1980 and 2019, annual precipitation in the area has dropped by 8-11%. Additionally, researchers determined that rainfall decline was not just attributed to local forest loss, but to deforestation in upwind regions across South America.
“Many studies only focus on the local scale, and local land-atmosphere feedback,” Jiangpeng Cui, associate professor at the Institute of Tibetan Plateau Research and lead author of the study, told Mongabay in an interview. “We combined observational data, like precipitation and evapotranspiration, with moisture tracking across South America. … This way we can know how [deforestation] changes vapor movement from one place to another.”
Since 1985, natural forest cover in South America has declined by 16%, largely due to human-caused deforestation. In the Brazilian Amazon, which lost one-fifth of its forest cover between 1970 and 2019, primary forest is frequently converted to agricultural land or destroyed by wildfires.
According to the study, cutting down large swaths of forest reduces the available evapotranspiration that drives rainfall, contributing to a destructive climate feedback loop that threatens tropical ecosystems. In the southern Amazon Basin, the research showed that a 1% loss of forest cover resulted in a 6-millimeter (0.2-inch) per year drop in rainfall annually.
By combining satellite-derived estimates of forest loss and precipitation with a model that traces water’s movement through the atmosphere, Cui and his team determined that rainfall in the southern Amazon Basin decreased by 3.9-5.4 millimeters (roughly 0.2 inches) per year between 1982 and 2016, while the northern Amazon showed a slight, nonsignificant increase. At the same time, the region lost on average 7.7% of its forest cover, largely driven by human activity, such as timber harvesting, agriculture and mining.
The researchers found that between 52% and 72% of the rainfall decrease was associated with land use change in the Amazon, which contributes to lower evapotranspiration. They also pointed out that this decline is linked not only to forest loss within the southern Amazon, but also to land use changes in upwind regions, which alter the amount of moisture carried toward the southern region.
According to study co-author Chris Huntingford, climate modeler at the U.K. Centre for Ecology and Hydrology, these findings demonstrate that human activity influences rainfall patterns through more than just greenhouse gas emissions. “As we go into the future, it’s not just climate change that’s going to affect rainfall patterns,” he told Mongabay in an interview. “We also impact them by changing the land’s surface.”
Modeling uncertainty
Climate modeling has vastly improved since researchers operated the first models in the mid-1960s. Early models were relatively simple, relying on basic energy-balance equations and limited inputs such as CO2 concentrations and general estimates of climate sensitivity.
Retrospective analysis shows that the warming these low-resolution models predicted was consistent with observations. Over time, the predictive power of models increased further when researchers began to incorporate inputs such as aerosols, volcanic activity, atmospheric processes, solar radiation and land-ecosystem responses.
Despite these advances, this research finds that many models may not capture the full scope of deforestation’s impact on climate, particularly rainfall. The study points to a specific gap: Models may underestimate how strongly forest loss disrupts evapotranspiration and atmospheric moisture transport across South America.
By comparing their results to a meta-analysis of 44 previous studies examining how deforestation impacts rainfall in the Amazon, the researchers determined that climate models tend to underestimate the impact of forest loss on rainfall by up to 50%.
This discrepancy, Huntingford said, may stem from how models represent evaporation and atmospheric transport. “Earth system models are amazing things, but the thing about them is they have to be accurate absolutely everywhere,” he said. “We need to make sure that climate models are getting South American atmospheric water transport accurate, along with how deforestation changes evaporation from the land.”
To more accurately trace the atmospheric transport of water vapor, the authors used a vapor tracking model informed by observational, satellite-derived data on rainfall, evaporation and land use change collected between 1982 and 2016. Compared with using simulated data alone, this provided them with a more complete picture of the relationship between deforestation, atmospheric vapor transport and downstream rainfall patterns.
Climate models that underestimate how deforestation influences rainfall may also underestimate how much the Amazon will dry in the future. The study notes that future models could produce more accurate projections by incorporating additional observational data, including in situ measurements of vegetation, surface water flux and atmospheric processes.
“We want to dig deeper, to find new things that we don’t know,” he said. “This can improve our projections of future climate, land-atmosphere feedback and the trajectory of the Amazon Rainforest.”
Mitigation through reforestation
The study points to reforestation as one potential solution to pulling the Amazon back from the brink of a tipping point, by raising levels of recycled rainfall and putting more moisture back into the atmosphere.
Eva Nafiseh Goodarzi, an ecosystem resilience scientist at The Nature Conservancy, said she believes reforestation could help increase rainfall. In 2022, she led a modeling-based study on reforestation in the Amazon Basin and determined that seasonal rainfall increased by up to 5 mm (almost 0.2 in) per day in the areas where forests were restored.
Using a weather prediction system designed for atmospheric research and forecasting, Goodarzi manually converted all cropland within the Amazon Basin to mature tropical forest, then examined differences between the simulations to see how the climate responded.
“Because around 30% of the moisture over the Amazon is provided to the atmosphere by recycling through the system, it’s very important to bring back that moisture to the atmosphere through reforestation,” she told Mongabay. “It can change the behavior of the whole atmosphere, regionally and even globally.”
In reality, restoring degraded land is more complex than simply converting it to a mature forest in a model. If left to recover naturally, tropical forests can regain much of their structure and function, but returning to old-growth conditions, especially in terms of biomass and species composition, can take more than a century. Areas where the soil is too severely degraded for natural regeneration need to be replanted manually, which requires careful planning and coordination to ensure the right species composition.
Goodarzi said restoration requires a coordinated effort. “For reforestation, there are different strategies, and everyone should contribute — from policymakers, stakeholders, local communities and Indigenous people,” she said. “And there should always be incentives for local communities who are doing this work.”
According to Goodarzi, The Nature Conservancy is promoting agroforestry as one solution that’s compatible with both restoration and human livelihoods. When it comes to legislation, she said policymakers need to support adaptation strategies and get behind them.
If the Amazon’s future hinges in part on land use decisions, Cui said science must help guide them. He said that discovering new information about land-atmosphere feedbacks can help shape effective policy, better predict future changes and determine how deforestation or reforestation may exacerbate the impact of climate change.
For Huntingford, understanding the relationship between deforestation and rainfall is also tied to human well-being. Loss of biosphere and forest loss are changing rainfall patterns, which he said has “big implications” for people living in regions in and around the rainforest. During severe drought in 2023 and 2024, Indigenous communities in the Amazon that rely on waterways for livelihoods and transportation experienced food insecurity and isolation.
“Our role with this research is twofold: to inform policymakers on the resilience of the rainforest and to try and keep people safe,” Huntingford said. “This piece of work uniquely covers both.”
Banner image: Logging activity in the Brazilian state of Rondônia, with trees already tagged and waiting for transportation. Image courtesy of Vicente Sampaio/Imaflora.
Citations:
Cui, J., Piao, S., Huntingford, C., Wang, T., & Spracklen, D. V. (2026). Historical deforestation drives strong rainfall decline across the southern Amazon basin. Nature Communications, 17(1). doi:10.1038/s41467-026-68361-z
Zalles, V., Hansen, M. C., Potapov, P. V., Parker, D., Stehman, S. V., Pickens, A. H., … Kommareddy, I. (2021). Rapid expansion of human impact on natural land in South America since 1985. Science Advances, 7(14). doi:10.1126/sciadv.abg1620
Ayad, H., Hassoun, S. S., Abdelkader, S. B., & Sallam, O. A. (2024). Assessing deforestation in the Brazilian forests: An econometric inquiry into the load capacity curve for deforestation. Forest Policy and Economics, 159, 103135. doi:10.1016/j.forpol.2023.103135
M Naval, M. L., Bieluczyk, W., Alvarez, F., Da Silva Carvalho, L. C., Maracahipes-Santos, L., De Oliveira, E. A., … Feldpausch, T. R. (2025). Impacts of repeated forest fires and agriculture on soil organic matter and health in southern Amazonia. Catena. Retrieved from https://www.sciencedirect.com/science/article/abs/pii/S0341816225002267?via%3Dihub
Dominguez, F., Eiras-Barca, J., Yang, Z., Bock, D., Nieto, R., & Gimeno, L. (2022). Amazonian Moisture Recycling Revisited Using WRF With Water Vapor Tracers. JGR Atmospheres. Retrieved from https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2021JD035259
Alizadeh, O. (2022). Advances and challenges in climate modeling. Climatic Change, 170(1-2). doi:10.1007/s10584-021-03298-4
Hausfather, Z., Drake, H. F., Abbott, T., & Schmidt, G. A. (2020). Evaluating the performance of past climate model projections. Geophysical Research Letters, 47(1). doi:10.1029/2019gl085378
Haghtalab, N., Moore, N., & Nejadhashemi, P. (2022). Would forest regrowth compensate for climate change in the Amazon basin? Applied Sciences, 12(14), 7052. doi:10.3390/app12147052
Poorter, L., Craven, D., Jakovac, C., & Van der Sand, M. (2021). Multidimensional tropical forest recovery. Science. Retrieved from https://www.researchgate.net/publication/356913360_Multidimensional_tropical_forest_recovery
Poorter, L., A. Rozendaal, D. M., Bongers, F., & Westoby, M. (2021). Functional recovery of secondary tropical forests. PNAS. Retrieved from https://www.pnas.org/doi/10.1073/pnas.2003405118
FEEDBACK: Use this form to send a message to the author of this post. If you want to post a public comment, you can do that at the bottom of the page.
Credits
