Weekly Sentinel-5P TROPOMI NO₂ and CH₄ over 63 Vietnam Provinces with Cropland and Paddy Rice Area, 2019–2024
Description
Research Hypothesis Tropospheric NO₂ and CH₄ concentrations measured by Sentinel-5P TROPOMI can capture seasonal agricultural activities, particularly paddy rice cultivation, across Vietnam. We hypothesize that NO₂ follows recurring seasonal patterns associated with rice growing cycles, CH₄ increases during flooded paddy seasons, and both gases exhibit distinct spatiotemporal behaviors across the six administrative regions of Vietnam, from the Red River Delta in the north to the Mekong River Delta in the south. What the Data Show This dataset provides weekly province-level summaries of tropospheric NO₂ and CH₄ for all 63 provinces of Vietnam from 2019 to 2024, covering 313 ISO weeks. Each record includes descriptive statistics of both gases, pixel counts, a six-region classification, rice season labels, and land-area estimates from three sources: Google Dynamic World annual cropland, ESA WorldCover static cropland, and NESEA-Rice10 paddy rice. Companion GeoTIFFs provide weekly NO₂ and CH₄ maps at 1,113 m resolution, annual cropland maps at 30 m, and a paddy rice binary map at 10 m resolution. Notable Findings NO₂ data are available for all 19,275 province-week records, while CH₄ coverage reaches 27.3% due to satellite retrieval constraints. NO₂ ranges from 0 to 309 µmol/m² with clear dry-season peaks in the Red River Delta. CH₄ concentrations range from 1,673 to 2,026 ppb. WorldCover estimates substantially higher cropland area (5.2 million ha nationally) than Dynamic World (2.0 million ha), while NESEA-Rice10 paddy rice (2.3 million ha) aligns closely with Dynamic World, suggesting Dynamic World is a reasonable rice proxy despite its broader cropland classification. Data Collection Sentinel-5P NO₂ and CH₄ processing, Dynamic World cropland, and ESA WorldCover extraction were conducted in Google Earth Engine using OFFL Level-3 products with ESA-recommended quality filters. Weekly composites follow ISO 8601 definitions and province statistics use FAO GAUL 2015 Level-1 boundaries. NESEA-Rice10 paddy rice tiles were downloaded from Zenodo, merged with rasterio, and processed locally using Python rasterstats zonal statistics. All cropland and rice areas are reported in hectares. How to Interpret and Use The weekly CSV is the primary analysis file, joinable by 'province' and 'week_start'. The 'region' column enables aggregation across six administrative regions; province_data_quality.csv details per-province CH₄ and NO₂ coverage rates by year. The 'data_quality' field indicates whether both gases are available. Cropland columns (cropland_area_ha_dw, cropland_area_ha_worldcover) are general land-cover proxies; rice_area_ha_nesea represents paddy rice specifically. Rice season labels follow the Vietnamese crop calendar: Dong_Xuan, He_Thu, and Thu_Dong. The dataset supports air quality monitoring, agricultural emission studies, and remote sensing research.