Efficient water management is crucial for enhancing crop productivity and ensuring food security, especially with growing global water scarcity. To support data-driven agricultural decision-making, this research develops a spatial decision support system using satellite-based Big Earth Observation Data. The system leverages MODIS time-series products to automatically retrieve and analyze Evapotranspiration (ET) and Gross Primary Production (GPP), which are used to calculate Water Use Efficiency (WUE)—a key metric for assessing crop water productivity.The core innovation is a scalable architecture that integrates multi-temporal satellite data with a Web GIS platform. The system uses PostgreSQL with PostGIS for efficient spatial database management and optimized SQL queries. Its backend, powered by Node.js, delivers RESTful APIs, while GeoServer publishes spatial layers as Web Map Services (WMS). These layers are visualized on an interactive, React-based interface using Leaflet.js, allowing users to dynamically analyze historical water use patterns by region and time via a standard browser.This system's ability to process and visualize large-scale spatiotemporal data empowers agricultural stakeholders to identify inefficiencies, adapt irrigation practices, and optimize resource allocation. The platform demonstrates how advanced geospatial information and Big Data processing contribute to sustainable agriculture and climate resilience, ultimately helping to improve crop yields and secure long-term food stability. Future enhancements may integrate ground-based or farmer-contributed data to further strengthen its decision-making capacity.

Fine particulate matter (PM_2.5) remains a major public-health concern in Thailand, with seasonal peaks amplified by terrain-induced stagnation. We delineate nationwide airshed regimes by integrating topography with multi-variable meteorology in an object-based workflow coupled to K-means clustering. Monthly composites for 2003–2024 were assembled for digital elevation (DEM), planetary boundary-layer pressure (PBL; proxy for PBL height), 10-m wind speed (WS10), surface pressure (SUR), surface shortwave radiation (SRAD), TMAX/TMIN, vapor pressure (VAP), and precipitation (PPT), harmonized to a common grid and aggregated to objects. Candidate K=2–10 was evaluated; the Elbow criterion supported K=4. Four physically interpretable clusters emerge—lowland, two transition belts, and highland—monotonically ordered by elevation and accompanied by coherent atmospheric gradients. From lowland to highland, SUR, TMAX, TMIN, VAP, and PPT decline, winds weaken in transition/highland zones, and PBL pressure decreases (consistent with higher PBL tops); SRAD is modestly lower aloft. Seasonality is systematic: November–February features expansion of transition/highland classes into basins under cooler, more stable conditions, whereas during the Southwest monsoon (May–October) transition belts retract as ventilation and wet scavenging strengthen. Despite boundary shifts, cluster identity at fixed locations is highly repeatable. The zoning provides an operational scaffold for airshed-aware management: targeting monitoring and advisories in transition belts and basin interiors, supporting cross-provincial coordination, and informing regime-conditioned forecasting. Limitations include a dispersion-focused design and reliance on gridded products. Planned extensions will incorporate ventilation capacity (VC=WS10×PBLH), stagnation frequency, and expanded validation (Silhouette/Gap, bootstrap stability, external/temporal holdouts with independent PM_2.5).