Mymensingh Division District-Level CAMS–ERA5 PM2.5, PM10 and AQI Dataset (2015–2025)
Description
This dataset provides hourly, district-level atmospheric composition and meteorological estimates for Mymensingh, Jamalpur, Sherpur, and Netrokona districts in the Mymensingh Division of Bangladesh. It covers the UTC period from 1 July 2015 to 31 December 2025. The canonical dataset contains 368,352 observations and 113 variables, with 92,088 hourly records for each district. Atmospheric composition data, including PM2.5, PM10, and total-column gaseous quantities, were obtained from the Copernicus Atmosphere Monitoring Service (CAMS) global atmospheric composition forecasts. Temperature, dew point, relative humidity, wind speed, wind direction and related meteorological variables were derived from ERA5 hourly single-level data. District estimates were produced by independently intersecting the native CAMS and ERA5 grid cells with verified geoBoundaries ADM2 district polygons. Values were aggregated using polygon-intersection areas calculated with WGS84 geodesic methods. ERA5 observations were then joined to the CAMS hourly time backbone using exact UTC timestamps. The dataset includes trailing 24-hour and Bangladesh-local daily PM summaries, temporal and spatial completeness indicators, quality-control flags, and PM2.5/PM10 AQI estimates calculated using the U.S. EPA 2024 particulate-matter breakpoints. These AQI values are model-derived research proxies and are not official monitoring-station observations or complete regulatory AQI reports. A historical CAMS joint-zero source artifact affects 3,276 district-hour observations for each PM pollutant. These values were retained as explicitly flagged missing data rather than interpreted as clean-air measurements. No temporal interpolation was performed. High modelled PM values were retained but have not been validated against ground-monitoring stations. The release includes the canonical Parquet table, compressed full and compact CSV derivatives, a Bangladesh-local daily CSV derivative, an Excel convenience file, all 84 CAMS and ERA5 NetCDF inputs, administrative-boundary data, processing code, a column dictionary, methodology documentation, quality-control reports, and SHA-256 checksum manifests. The dataset is suitable for reproducible environmental research, exploratory analysis, methods development, and predictive modelling. Because CAMS is an operational forecast system that changed over time, the combined series should not be treated as a homogeneous reanalysis or used for unqualified long-term trend attribution.
Files
Institutions
- Bangladesh Army University of Engineering & TechnologyRajshahi Division, Rajshahi