Analysis and evaluation of surface water changes in the lower reaches of the Yangtze River using Sentinel-1 imagery
Zongqiao Yu, Qing An, Wensong Liu, Yuhong Wang
Abstract
The lower reaches of the Yangtze River, China. A novel framework for analysis and evaluation of surface water changes in the lower reaches of the Yangtze River was proposed in this study, which was carried out using the Google Earth Engine (GEE) cloud computing platform. The Sentinel-1A images of dry and wet seasons in the study area during 2016 and 2020 were initially synthesized based on the median value of the backscatter coefficient. Then, water extraction results were obtained from the combination of the Random Forest (RF) algorithm, digital elevation model (DEM), and mathematical morphology model according to preferred textural features. Verification regions were selected to evaluate the performance of proposed framework; the overall accuracy (OA) and kappa coefficient (Kappa) of the proposed method was 94.13% and 0.89, respectively. Meanwhile, the proposed water extraction framework can provide a better performance when compared with state-of-the-art methods (Otsu threshold method, k-nearest neighbors (KNN), and support vector machine (SVM) algorithms) regarding qualitative and quantitative evaluations. The mountain shadows could be eliminated and water continuity improved when the DEM and mathematical morphology model were introduced in the proposed framework. The results indicate that frequent floods and droughts occur in the study area. Additionally, this area is experiencing because of the increases in the water levels and a decline in the regulation and storage capacity of lakes.