Multi-source remote sensing-based monitoring of surface water storage in the Yangtze River middle and lower reaches

Abstract

Ponds and small reservoirs are widely distributed in the middle and lower reaches of the Yangtze River and provide important support for irrigation, rural water supply, ecological replenishment, and local drought mitigation. However, their storage dynamics are difficult to monitor because most facilities are small, scattered, and lack continuous water-level observations or site-specific water-level storage curves. This study develops a capacity-constrained area-change method for estimating regional storage dynamics of ponds and small reservoirs using multi-source remote sensing and statistical storage-capacity records. Landsat-8/9, Sentinel-2, and Sentinel-1 SAR observations were used to extract surface-water area, while lake, river, and reservoir masks were used to remove non-target water bodies. Small reservoirs were separated from ponds using inventory-based buffer matching, and regional storage was estimated by converting aggregated water-area changes into storage changes under statistical capacity constraints. The source dataset contained 44,019 valid satellite scenes, including 4,234 Landsat scenes, 36,256 Sentinel-2 scenes, and 3,529 Sentinel-1 SAR scenes. The total storage capacities of small reservoirs and ponds were 182.92 × 108 m3 and 85.30 × 108 m3, respectively. The results show that pond storage has a larger seasonal amplitude and a higher coefficient of variation than small-reservoir storage, indicating stronger sensitivity to rainfall deficits, evaporation, and water withdrawal. During the extreme 2022 flood-drought abrupt alternation event, annual mean pond storage decreased to 52.83 × 108 m3 and the pond coefficient of variation reached 0.225. Approximately 35%-45% of ponds became dry or nearly dry during September-October 2022. The proposed framework provides an interpretable regional approach for monitoring distributed small-water-body storage under limited in-situ observations.