Integrating Remote Sensing and GIS for Irrigated Area Mapping in the Betwa River Basin, India

Abstract

Accurate mapping of irrigated and non-irrigated croplands is crucial for sustainable water resource management and agricultural planning in semi-arid regions. This study mapped irrigated areas in the Betwa River Basin (43,469 km²), a major tributary of the Yamuna, using MODIS Terra surface reflectance (MOD09Q1) data for the 2020–2021 agricultural year. A total of 45 NDVI composites (250 m resolution) were generated, and a Maximum Value Composite (MVC) approach was applied to minimize atmospheric and cloud-related noise, producing a 46-layer NDVI time-series dataset. Unsupervised classification (ISODATA with ISOCLASS clustering) was refined with ground-truth GPS points and high-resolution imagery, differentiating nine land use/land cover classes, including various irrigation regimes, rainfed agriculture, forests, and water bodies. Distinct NDVI phenological signatures revealed double-cropped irrigated systems with two peaks (July–August and February–March) and rainfed croplands with a single monsoon peak (~0.55). The classification achieved 86% overall accuracy with a kappa coefficient of 0.82. Approximately 1.28 million hectares were identified as irrigated, with spatial heterogeneity reflecting canal-fed irrigation in the northern plains, groundwater-dependent irrigation in southern uplands, and minor/tank irrigation in plateau margins. The results demonstrate that MODIS NDVI time-series provides a cost-effective, scalable approach for basin-scale irrigation mapping, supporting agricultural planning, water allocation, and sustainable water resource management. Integration with higher-resolution datasets and hydrological modeling can further enhance accuracy and decision-making.