Satellite-Derived Surface Air Quality Index Estimation and Spatio-Temporal Characterization of Formaldehyde Hotspots over India using a CNN-LSTM Framework

Ambient air pollution remains a major environmental health concern in India, and the Central Pollution Control Board (CPCB) National Air Quality Index (AQI) is the primary instrument used to communicate multi-pollutant exposure risk to the public. Because ground-based monitoring stations are spatially sparse, large parts of the country lack a locally representative AQI value. A separate atmospheric-chemistry problem concerns tropospheric formaldehyde (HCHO), a short-lived trace gas that serves as a proxy for volatile organic compound emissions and is closely associated with biomass-burning activity; HCHO is not one of the eight pollutants used in the CPCB AQI formulation and must therefore be treated as a distinct scientific variable rather than as an AQI input. This paper proposes a two-objective satellite-to-surface framework for India that (i) estimates a CPCB-consistent surface AQI by learning a mapping from INSAT-3D Aerosol Optical Depth, Sentinel-5P TROPOMI trace-gas columns, and ERA5/IMDAA reanalysis meteorology to CPCB ground-station pollutant concentrations using a CNN-LSTM architecture, and (ii) characterizes spatio-temporal HCHO hotspots during biomass-burning periods by computing a climatological standardized-anomaly score from TROPOMI HCHO columns and relating flagged hotspots to MODIS/VIIRS active-fire detections and reanalysis wind fields. The two objectives share a common data-acquisition and spatio-temporal alignment pipeline but remain scientifically and evaluatively separate. This paper presents the system architecture, the CPCB sub-index and AQI aggregation formulation, the CNN-LSTM mathematical formulation, the HCHO anomaly-detection formulation, and a reproducible experimental-evaluation protocol based on spatially and temporally held-out validation. No experimental results are reported, since model training and validation have not yet been performed; parameters such as dataset size, spatial/temporal resolution of the final fused product, and performance metrics will be determined during experimental validation. The intended contribution is a reproducible, dual-objective pipeline that keeps CPCB-regulated AQI estimation and trace-gas hotspot characterization scientifically distinct while sharing a common multi-source data infrastructure.