Noise Mapping and Air Quality Assessment in a New Industrial Area: Evaluation, Correlation and Environmental Impact Analysis

Abstract

Noise and air pollution are among the most pressing environmental concerns associated with industrial expansion at the urban fringe, yet they continue to be assessed as separate phenomena despite shared emission sources and co-exposure health consequences that exceed either stressor alone. This study addresses that gap by conducting an integrated environmental assessment of noise and air quality conditions in a newly activated industrial zone in Debrecen, Hungary, during the initial operational phase of a large-scale manufacturing facility. The study followed a three-stage design: pre-operational baseline data collection at three continuously operating monitoring stations; computational noise propagation modelling using IMMI software (CNOSSOS-EU methodology) across a range of emission scenarios; and statistical analysis of 31-day post-operational measurements (22 January – 21 February 2026) against both the baseline and model predictions. The statistical framework comprised a Before-After-Control Impact (BACI) design, Pearson, Spearman, and Kendall correlation analysis across 480 variable pairs, partial correlation for meteorological decompounding, and 22 families of multiple linear regression yielding 396 equations — incorporating a novel Upwind Factor for directional industrial source attribution. Models were validated against independent March 2026 data. BACI analysis revealed divergent noise dynamics: the background control station recorded +2.04 dB (p = 0.003); the traffic-dominated station showed no significant change (−1.04 dB, p = 0.104); and the primary downwind residential station recorded −1.60 dB (p = 0.006), consistent with IMMI predictions of full industrial noise attenuation at receptor distances. Air quality analysis found inter station concentrations strongly coupled through regional pollution episodes rather than local industrial emissions. The best-performing regression models achieved R² = 0.69 (KER14/PM₂.₅) and R² = 0.826 (KER04/NO₂); at KER14, the Upwind Factor was statistically significant (β = +2.31, p = 0.021), with out-of-sample validation confirming generalisation capability (R² = 0.734, RMSE = 3.1 µg/m³)