Ensemble-based mapping and trophic characterization of lentic water bodies in the Middle Ganga Basin under climate...

Lentic small water bodies (LSWBs), such as ponds and small lakes, play an important role in freshwater ecosystems by supporting biodiversity, groundwater recharge, and rural subsistence. However, their size and ecological integrity are being impacted by changing climates and land-use change, and they are still under-represented in large-scale water assessments. This study examines the decadal dynamics (2013-2024) of LSWBs in the Haridwar watershed of the Middle Gangetic Basin, determining how their extent and trophic status have changed as a result of environmental and land-use changes. The watershed, located in the middle Ganga Basin, was chosen for its extensive pond network that is greatly impacted by urbanisation, agriculture, and monsoonal variability. To overcome the constraints of single-sensor techniques, multi-sensor satellite datasets (Landsat-7/8, Sentinel-1/2) and important spectral indices were combined in an ensemble machine-learning algorithm to accurately map yearly LSWBs (training = 96.89 %, testing = 87.27 %). Compared to conventional classifiers, the ensemble technique improved accuracy, stability, and cloud resilience, enabling continuous long-term monitoring. Trophic State Index (TSI) mapping, confirmed with field data (accuracy = 78.57 %, kappa coefficient = 0.74), showed that eutrophic and hypereutrophic LSWBs dominated the region (>70 %), indicating chronic nutrient enrichment. Low rainfall and elevated temperatures were associated with increasing trophic levels, while agricultural expansion and urbanization contributed to land degradation and declining water quality. These findings show that climate and land-use change have a synergistic effect on LSWBs, and the suggested ensemble multi-sensor algorithm provides a scalable, cost-effective tool for continuous monitoring and management of small freshwater ecosystems in data-scarce regions.