Non-linear impact mechanisms of multi-modal urban traffic on air quality: An interpretable machine learning study for sustainable policy making

Abstract

Urban air pollution, specifically Nitrogen Dioxide (NO 2 ), presents a multifaceted challenge that is intricately coupled with the stochastic, multi-modal, and non-linear dynamics of mega-city traffic systems. This study systematically investigates the non-linear impacts of mixed traffic flow—comprising motorcycles (MC), private cars (PC), and heavy vehicles (BT)—on local air quality at the iconic Bundaran HI intersection in Jakarta, Indonesia. Leveraging a high-resolution, year-long longitudinal dataset, we developed a robust Random Forest (RF) modeling framework integrated with Permutation Importance and Partial Dependence Analysis (PDP) to decipher the environmental footprint of urban transport under tropical conditions. Our results reveal that private car volume and the Volume-to-Capacity (V/C) ratio act as the primary catalysts for NO 2 spikes, significantly outweighing the contribution of heavy vehicles in this specific urban corridor. Crucially, a distinct non-linear threshold effect was identified: NO 2 concentrations undergo a regime shift, rising exponentially once the V/C ratio exceeds a critical “elbow” of 0.65. This non-linearity indicates that traditional linear mitigation strategies and average-speed-based emission models significantly underestimate pollution risks during saturated traffic states. Policy scenario simulations demonstrate that a 30% reduction in private vehicle volume yields a 5.8% reduction in mean NO 2 , offering nearly six times the environmental utility of heavy vehicle restrictions. Furthermore, the study explores the role of road surface materials—specifically Stone Mastic Asphalt (SMA)—and meteorological interactions in exacerbating localized pollution. This research provides a data-driven, interpretable framework for urban planners to transition from generic traffic bans toward precision-based, sustainable management strategies that align with the core principles of cleaner production, urban resilience, and UN Sustainable Development Goal 11.