Monitoring the spatiotemporal change of Dongting Lake wetland by integrating Landsat and MODIS images, from 2001 to 2020

Dongting Lake wetland provides critical ecological functions in the Yangtze River Basin, however its landscape pattern has changed during the past 20 years due to both climate change and the human activities such as the operation of the Three Gorges dam (TGD). Numerous studies have used remote sensing technology to monitor such changes. However, most of studies were conducted at low spatial resolutions (250–1000 m) or low temporal resolutions (few images per year), which can introduce a degree of unreliability as regards the associated conclusions. To thoroughly analyze the spatiotemporal characteristics of Dongting Lake wetland, high spatiotemporal resolution images from 2001 to 2020 have been produced by a modified version of the flexible spatiotemporal data fusion method (cuFSDAF 2.0) developed in this paper. This method is able to retrieve both phenology and land-cover changes both quickly and accurately. In addition, a tailored category by category extraction mapping method (CCEMM) is proposed to reliably classify wetland cover types. The mapping results show that the water area expanded by 66.43 Km 2 in the wet season, and shrank by 132.86 Km 2 in the dry season, between and including the years, 2001 to 2020. It appeared that the water area in the dry season declined obviously from 2003 to 2010 due to the operation of the TGD and is gradually stabilized after 2010. The reed area increased by 279.59 Km 2 and the grass area decreased by 220.46 Km 2 in the dry season, from 2001 to 2020. The shrinkage of the water area in the dry season may be one of the main factors driving reeds to occupy the previous grass growth area. • A new spatiotemporal fusion method is developed to produce high spatiotemporal resolution images of the Dongting Lake. • A tailored mapping method is proposed to reliably classify the wetland. • The spatiotemporal characteristics of the wetland from 2001 to 2020 are analyzed.