ABSTRACT Air quality in large cities is a global public health issue that has received increasing attention in recent decades. The dynamics of atmospheric pollutant concentrations form a complex system influenced by meteorology, topography, emission sources, and fleet composition, interacting in a nonlinear pattern. We monitored particulate matter (PM) in two transport modes (walking and bus) and a fixed point in Curitiba, Brazil, to understand human exposure. Real‐time and concentrations, with GPS data, were collected using low‐cost sensors and microcontrollers. Machine learning models identified key pollution drivers. Results showed that mean PM concentrations were higher inside buses. Afternoon bus samples frequently exceeded 45 / for , surpassing WHO daily limits, while the highest walking peak (approximately 34–35 /) occurred near a major highway. and were strongly correlated, especially for walking ( R = 0.91). Despite higher concentrations inside buses, cumulative exposure and inhaled dose were greater while walking, with average exposure of 0.18 (vs. 0.11 on buses) and average inhaled doses of 5.74 (walking) versus 3.05 (bus). Longer commute duration and increased inhalation rates explain this difference. Random Forest (RF) analysis identified boundary layer height as the most important predictor of , followed by relative humidity and the number of SUVs. These findings highlight how meteorology, fleet composition, and travel mode shape exposure and support strategies for healthier and sustainable urban mobility.
Chaves et al. (2026) studied this question.