Expanding Amazon dry-hot season under anthropogenic climate change

The Amazon rainforest, a crucial global carbon sink, plays a vital role in the global climate system. As ongoing climate change and local deforestation push the Amazon toward a critical tipping point, understanding the region's changing climate patterns becomes increasingly important. In this study, we reveal a significant expansion of the dry-hot season across the Amazon rainforest from 1980-2022, creating prolonged adverse climate conditions for the ecosystem and local communities. A machine learning clustering algorithm is used to define the dry-hot season automatically by considering the temperature, precipitation, and soil moisture simultaneously.

The land-atmosphere interaction predominates the dry-hot season expansion in the Amazon. During the dry season (Aug-Oct), the daily maximum temperature has warmed by ~1 degree per decade, much faster than that in the wet seasons (~0.4 degree per decade). By the surface heat budget analysis, we found that intensive dry-season warming is predominantly driven by reduced evapotranspiration, leading to decreased surface latent heat flux and increased shortwave radiation due to diminished cloud cover. The declining evapotranspiration rates stem from a combination of increasing soil moisture deficits and local deforestation.

By large-ensemble climate model simulations, we further demonstrate that this dry-hot season expansion is highly unlikely to occur without anthropogenic climate change and this expansion will exacerbate under future warming scenarios. By single-forcing experiment, we further confirm the critical role of local deforestation in amplifying this expansion. These findings emphasize the urgent need for targeted mitigation and adaptation strategies to protect this vital ecosystem from the compounding effects of climate change and deforestation.