NO₂ mapping of Perth bushfire utilizing Sentinel-5P TROPOMI

Abstract

In the face of escalating environmental concerns, effective management of air quality remains critical. This study focuses on Perth, Australia, a region impacted by frequent bushfires and industrial emissions, necessitating precise monitoring of atmospheric pollutants like NO2. Leveraging advanced remote sensing technologies, including Sentinel-2 and Sentinel-5P satellites, this research assesses the spatial and temporal dynamics of NO2 before, during, and after the 2021 Wooroloo bushfire. A key objective was to convert satellite-derived NO2 data from mol/m2 to μg/m3 to enable accurate environmental assessment. This conversion utilized a unit conversion method, improving accuracy metrics substantially, with a correlation coefficient (r) increasing from 0.59 to 0.82 and root mean square error (RMSE) decreasing from 7.58 μg/m3 to 3.20 μg/m3. A regression model, validated with ground-level measurements, demonstrates robust predictive capability (R2 = 0.76, RMSE = 2.58 μg/m3), aiding in the creation of NO2 distribution maps across Greater Perth. Comparison with six ground stations revealed varying accuracy (RMSE: 2.9249 to 7.2705 μg/m3), likely influenced by proximity to the fire and prevailing wind directions. Spatiotemporal analysis depicted distinct NO2 patterns: stable levels pre-fire, dramatic increases during, and gradual post-fire recovery. Maximum NO2 concentrations peaked during the fire (up to 79.227 μg/m3), exceeding air quality guidelines. Post-fire, concentrations normalized, yet sporadic peaks persisted, indicating an ongoing environmental impact. Furthermore, analysis of environmental parameters such as Land Surface Temperature (LST), precipitation, and Normalized Difference Vegetation Index (NDVI) during the study period revealed significant correlations with NO2 levels. LST showed a positive correlation (r = 0.64) with NO2 concentrations during the fire, suggesting temperature influences on atmospheric stability and pollutant dispersion. Precipitation exhibited a negative correlation (r = −0.52), indicating its role in scavenging NO2 from the atmosphere post-fire. NDVI displayed a weak negative correlation (r = −0.30), reflecting vegetation recovery trends post-fire. This comprehensive study integrates advanced remote sensing with statistical modelling to enhance air quality monitoring and inform decision-making in bushfire-prone regions. By elucidating NO2 dynamics and their environmental implications, this research contributes essential insights for mitigating air pollution and safeguarding public health amidst climate-induced challenges.