Up to 36% of the Amazon rainforest may resist severe drought far better than scientists previously assumed, according to a growing body of peer-reviewed research linking shallow water tables and species diversity to forest survival. The finding reframes a long-running debate about the world’s largest tropical forest: rather than a uniformly fragile system sliding toward collapse, the Amazon appears to contain a patchwork of resilient zones that could buffer against the worst effects of climate-driven dry spells. For the roughly 30 million people who live in and around the Amazon basin, and for global climate stability that depends on the forest’s carbon storage, the distinction between vulnerable and resilient zones carries real consequences.
What the research shows
The resilience picture rests on several independent lines of evidence published across top-tier journals. A peer-reviewed study in the Journal of Ecology found that forests sitting above shallow water tables function as “hydrological refugia,” areas where accessible groundwater allows palms and trees to maintain canopy health even during extreme drought events. The researchers used transect-based field measurements across multiple forest types to demonstrate that trees in these zones showed significantly less drought stress than those on higher, drier ground. Accessible groundwater, in short, acted as a lifeline.
The headline 36% estimate comes from a separate, basin-wide analysis published in Nature that mapped resilience and vulnerability across the entire Amazon using ecotope-level drivers: water table depth, soil composition, and canopy height. An author correction later updated the computational workflow and some spatial outputs needed to reproduce the study’s maps, while confirming the underlying data pipeline through a publicly available Code Ocean capsule. That transparency step strengthened confidence in the original results, though it also highlighted how sensitive the percentage is to modeling assumptions.
A study in Nature Communications offered a mechanistic explanation for why some forest patches outperform others under water stress. By modeling functional hydraulic-trait diversity, the range of water-transport strategies among coexisting tree species, the researchers found that forests with greater trait variety lost measurably less biomass under both sudden and prolonged drought scenarios. Species-level diversity, the paper argued, acts as a form of biological insurance.
Long-running field experiments add a critical ground-truth dimension. At the Caxiuanã National Forest in eastern Amazonia, researchers have tracked how rainforest responds to artificially imposed drought sustained over more than a decade. Their work, published in Nature Ecology & Evolution, showed that biomass in experimental plots did eventually stabilize, but through a costly trade-off: large trees died, forest composition shifted toward smaller, more drought-tolerant species, and carbon accumulation patterns changed. The forest survived, but it emerged structurally different, shorter, and with fewer of the giant trees that store the most carbon per individual.
What remains uncertain
The 36% figure, while grounded in peer-reviewed modeling, has not been validated with basin-wide field measurements of real-time groundwater levels. The Nature study relied on modeled estimates of water table depth and satellite-derived indicators rather than direct hydrological monitoring across the Amazon’s roughly 5.5 million square kilometers. Whether those modeled refugia perform as predicted during actual multi-year droughts remains an open and urgent question.
Scaling presents a second challenge. The Nature Communications findings on trait diversity come from modeling exercises, not from large-scale observational plots. Small experimental sites like Caxiuanã confirm that forests can adjust to drought, but the compositional changes they undergo raise hard questions about what “resilience” actually means for carbon storage. A forest that survives but stores significantly less carbon offers a very different climate outcome than one that maintains its full structure.
Research from the opposite direction complicates the picture further. A study published in the Proceedings of the National Academy of Sciences quantified Amazon drought vulnerability using satellite-derived greenness data and multiple drought indices, finding that vulnerability varies sharply by region and that repeated drought exposure compounds the damage. Meanwhile, a study in Science estimated that human activity and drought together have degraded roughly 2.5 million square kilometers of the Amazon, approximately 38% of the remaining forest. The near-overlap between that 38% degradation estimate and the 36% resilience figure is coincidental, but it captures a real tension: different analytical lenses applied to the same forest produce strikingly different assessments of its health.
How to weigh the evidence
Three tiers of evidence feed this debate, and they carry different levels of confidence. The strongest comes from controlled field experiments like the Caxiuanã drought manipulation, where researchers directly measured tree responses to reduced rainfall over more than a decade. These experiments are geographically narrow but high in causal certainty. The second tier consists of peer-reviewed modeling studies that extrapolate from field data and remote sensing to make basin-wide predictions. The Journal of Ecology work on shallow water tables and the Nature Communications analysis of trait diversity both fall here: rigorous in method, but dependent on assumptions about how local findings scale across a continent-sized forest. The third tier includes satellite-based remote-sensing analyses that track canopy greenness or biomass across the full Amazon. These provide unmatched spatial coverage but measure proxies for forest health rather than direct biological outcomes.
Taken together, the research points to a practical conclusion: the Amazon is not a single system with a single fate. Forests above shallow water tables and those with high species diversity appear to carry built-in defenses against drought. Identifying and protecting these zones from deforestation and fire could preserve the Amazon’s most resilient patches while buying time for more vulnerable areas. For policymakers and conservation organizations working in the basin, the priority that emerges from the data is clear: shield the hydrological refugia that multiple studies consistently flag as drought-resistant, because losing those zones would strip away the Amazon’s best natural buffer against a drying climate.
The tension between resilience and degradation will likely sharpen as drought frequency increases. The evidence assembled so far suggests that up to a third of the Amazon has meaningful natural defenses, but those defenses operate through trade-offs, including the loss of large trees and shifts in species composition, that reduce the forest’s value as a carbon sink. Whether the resilient fraction can hold, or even expand through species dispersal if deforestation slows, depends on decisions being made now about land use, fire management, and protected-area enforcement across the basin.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
