Monitoring surface moisture changes using NDWI and extended indices: a case study of Suyahu Reservoir

Based on the differential reflectance of surface moisture in the green and near-infrared bands observed via remote sensing, this study explores the use of the Normalized Difference Water Index (NDWI) and its extended indices to monitor surface moisture changes. Multi-temporal Sentinel-2 optical imagery from 2018 to 2022 was utilized to compare the performance of water body extraction methods using a fixed NDWI threshold of 0 and Otsu's adaptive thresholding method. Furthermore, two novel indices—the Binary NDWI Difference Index (NDWI_b-del) and the Normalized NDWI Difference Index (NNDWI_del)—were proposed for identifying dynamic water transitions and surface moisture changes. The study analyzed these indices using Suyahu Reservoir as a case study. Results indicate that the fixed NDWI threshold of 0 outperformed Otsu's thresholding method, which had an effective period of less than 50%. Water distribution closely aligned with dredging and expansion activities. In 2018, natural changes in water bodies were mainly observed in the northwest and western areas. Post-construction, the water area in the western reservoir decreased in 2019, while increases in the deep-water zones, areas near Artificial Island No. 1, and post-dam dredging zones were observed in 2020–2021. By 2022, water bodies around Artificial Islands No. 2 and No. 3 continued to expand. The NDWI_b-del index identified significant water transitions in construction fill zones, deep-water zones, and post-dam dredging zones from 2019 to 2022. The NNDWI_del index revealed extreme changes corresponding to transitions between water and non-water states, while moderate drying and wetting changes corresponded to precursor phases of extreme changes, including water body reductions due to earthworks and water accumulation from silt deposition in less apparent transition zones. This study extends the NDWI framework by proposing two remote sensing indices, providing a novel and efficient method for automated monitoring of surface moisture and environmental changes.