Amazon's climate crossroads: analyzing air pollution and health impacts under machine learning-based temperature increase scenarios in Northern Mato Grosso, Brazil.

Abstract

Air pollution has long been a public health concern in South America, now increasingly linked to climate change. In Brazil, this issue is particularly acute in smaller cities with limited monitoring infrastructure. Sinop, located in the Amazon biome of Mato Grosso, exemplifies the intersection of agricultural expansion and environmental vulnerability. This study presents a comprehensive assessment of meteorological conditions, air pollutant levels, and related health impacts in Sinop, focusing on PM_2.5, PM₁₀, and ground-level ozone (O₃). Using 2022 data, we applied Health Impact Assessment (HIA) via the WHO's AirQ + tool and simulated temperature rise scenarios through a machine learning model. ur findings indicate that O₃ levels are highly sensitive to meteorological variation, while PM_2.5 and PM₁₀ concentrations are projected to increase by up to 50% and 70%, respectively, under extreme warming scenarios. These changes correspond to an estimated increase in non-external mortality of over 90% for PM_2.5 and more than 1000% for PM₁₀. Seasonal simulations revealed that the dry season, marked by intense biomass burning, significantly exacerbates pollutant concentrations. These results underscore the growing burden of air pollution on public health in rapidly urbanizing regions. They also highlight the urgent need for enhanced air quality monitoring and climate-adaptive public health strategies in vulnerable areas such as the southern Amazon.