Multisensor Data Fusion for Quantifying Agricultural Fire Impacts on Air Quality and Environmental Degradation

This study introduces a novel multisensor fusion methodology that enhances spatial–temporal resolution by 40% compared with single-sensor approaches, demonstrating scalable applications for ecosystem monitoring. The framework’s practical applications include real-time air quality assessment, agricultural policy formulation, and climate change adaptation strategies across similar agroecological zones in South Asia. Utilizing the Google Earth Engine platform, we integrate heterogeneous datasets from moderate-resolution imaging spectroradiometer (thermal anomalies and land surface temperature), fire information for resource management system (FIRMS—fire activity), Sentinel-2 [vegetation indices: enhanced vegetation index (EVI), normalized difference vegetation index, normalized burn ratio (NBR)/differenced NBR (dNBR)], and TROPOspheric Monitoring Instrument (CH4, CO, and NO2 concentrations) to evaluate prefire (March–April 2022) and postfire (May–June 2022) conditions. Key findings reveal a 35% increase in fire activity postfire (FIRMS values: 1006.4–1359.7), a 15% decline in vegetation health (EVI: 0.225–0.189), and elevated LST (29.2–44.0 °C) correlating with increased greenhouse gas emissions (CH4: 1881–1935 ppb; CO: 0.037–0.042 mol/m2). The dNBR analysis (−0.90 to 0.69) highlighted burn severity heterogeneity, underscoring the need for spatially adaptive mitigation strategies. Our approach provides a scalable model for assessing ecological vulnerabilities in region, where multisensor synergy is critical to capture complex interactions between terrestrial and marine systems. The results advocate for policy reforms to reduce crop residue burning and promote sustainable practices. © 2025 Elsevier B.V., All rights reserved.