Linear and polynomial regression analysis of the relationship between vehicle volume and PM₁₀ concentration on road sections in Makassar City

¹ Departement of Environmental Engineering, Faculty of Engineering, University of Hasanuddin, St. Poros Malino KM.6, Bontomarannu, Gowa, South Sulawesi, 92172, Indonesia
² Transportation and Air Quality Research Group, University of Hasanuddin, St. Perintis Kemerdakaan No. KM 10, Makassar South Sulawesi, 90245, Indonesia
³ Centre for Transportation Research, Department of Civil Engineering, Faculty of Engineering, Universiti Malaya, 50603 Kuala Lumpur, Malaysia

^* Corresponding author: marni_hamidaly@yahoo.com

Abstract

Urban air quality has become a critical concern in the context of climate change, as increasing emissions from the transportation sector contribute not only to greenhouse gases but also to particulate pollution that directly affects public health. In rapidly developing cities like Makassar, understanding the factors influencing PM10 concentrations is essential for supporting climate-responsive urban planning. This study aims to analyze the predictive relationship between vehicle volume and PM10 concentration on 6/2D arterial roads and 2/2UD undivided two-way roads using linear and polynomial regression models. Vehicle volume and PM10 data were collected from eight measurement points on each road type across four time intervals. PM10 concentrations were measured using a High Volume Air Sampler (HVAS), while vehicle volume was recorded simultaneously using a traffic counter application. Results show that vehicle volume on 2/2UD roads was lower than on 6/2D roads, with motorcycles dominating traffic composition (71.42–82.54%). All points recorded PM10 concentrations exceeding the national standard of 75 µg/m³. Linear regression produced a moderate relationship (r = 0.519; R² = 26.9%), whereas the polynomial model showed stronger predictive ability (r = 0.696; R² = 48.5%). These findings indicate that vehicle volume alone cannot reliably predict PM10 levels, highlighting the need to incorporate variables such as vehicle speed and wind speed in future climate-related air quality modelling.