Refining and densifying the water inundation area and storage estimates of Poyang Lake by integrating Sentinel-1/2 and bathymetry data

Abstract

Poyang Lake, the largest freshwater lake in China, is featured with high-frequency fluctuations and unique seasonal changes in water level and inundation areas, which plays a substantial part in regulating the runoff of the Yangtze River and maintaining the ecological balance of the region. Frequent observations enable more effectively reflecting the water budgets and related driving force of Poyang Lake. This study first utilized the Sentinel-1 SAR data and cloud-free Sentinel-2 images in conjugation with lake bathymetry to reconstruct the time series of the water inundation area and storage of Poyang Lake from 2017 to 2020. Further, a simple and straightforward method, named the Fitting Curves of each Block and Total area (FCoBT), is developed to reconstruct the complete inundation area of Poyang Lake from partially cloud-free Sentinel-2 images to densify the original time series. This method takes full advantage of the significant characteristics on the strong correlations of the block areas (one small part of lake area) varying with the complete lake areas, with a mean R2 and ρ value of 0.94 and 0.98, respectively. The complete inundation area of Poyang Lake can be reconstructed from 113 scenes of partially cloudless Sentinel-2 images acquired from 2017 to 2020 by using the FCoBT. After integrating the Sentinel-1 and Sentinel-2 images, a maximum of 13 measurements are available within one month, with the temporal intervals of 2–3 days. Results show that our reconstructed time series of water inundation area can finely depict significant seasonal fluctuation (wet season: from April to September, dry season: from October to the next March) and inter-annual variations (larger areas: 2017 and 2020, smaller areas: 2018 and 2019) of Poyang Lake during 2017–2020. Similar to the area change, the water storage of Poyang Lake also shows evident seasonal variation and inter-annual fluctuation, reaching the peaks in July and the lowest stages during November to February. This study cannot only provide an efficient and feasible remote sensing means of constructing long-term observations for highly variable lakes at high spatial–temporal resolutions, but also greatly refine our understanding on the intra-annual and inter-annual variations of Poyang Lake.