Seasonal characterization and machine learning prediction of atmospheric pollutants in an agricultural–urban area of the Atlantic Forest biome

Márcio Teixeira

doi:10.1016/j.apr.2026.103101

Seasonal characterization and machine learning prediction of atmospheric pollutants in an agricultural–urban area of the Atlantic Forest biome

2026-06-22

0

OA

AI

M

Márcio Teixeira

M

María de Fátima Andrade

P

Prashant Kumar

M

Marco A. Franco *

DOI：10.1016/j.apr.2026.103101

Original

Original request for help

Save

Abstract

En 中文

• PM2.5 and PM10 peak in the dry winter season, with ∼30 % of days exceeding WHO guidelines, especially in 2024. • Machine learning models outperformed traditional statistical approaches, especially when temporal features were included, improving the representation of pollution persistence and episodes. • Combustion tracers and atmospheric moisture variables are the main drivers of PM variability, linking air quality deterioration to biomass-burning emissions and meteorological controls.

Keywords:

Urban air pollution

Biomass burning

Atlantic Forest

Seasonal characterization

Machine learning prediction

Seasonal characterization and machine learning prediction of atmospheric pollutants in an agricultural–urban area of the Atlantic Forest biome

Abstract

Journal

Organization

Cited Papers

Citing Papers