Annual Maps of Surface Water Body, Paddy Rice, and Wetlands in Northeast China Using PALSAR, Sentinel-1, Sentinel-2, Landsat, and MODIS Imagery in 2020

Three water-related land cover types (surface water body, paddy rice, and natural wetlands) have been changing rapidly in Northeast China over the last few decades due to human activities and climate and hydrological changes. Accurate and up-to-date annual maps of water-related land cover types at high spatial resolution are essential for improving our understanding of biodiversity, climate change, food security, and transmission of infectious diseases. In this study, we used pixel- and knowledge- based algorithms and integrated datasets from ALOS PALSAR-2, time series MODIS land surface temperature (LST), Landsat (ETM+/OLI), Sentinel-2, and Sentinel-1 imagery in 2020 to map different types of water-related land cover types at 10-m spatial resolution in Northeast China. The resultant water-related land cover type maps had a high accuracy with an overall accuracy of 95.2 % and a Kappa coefficient of 0.91. There was a total of 230,011 km2 of water-related land cover types, which included 21,018 km2 of yearlong surface water bodies, 81,106 km2 of paddy rice, 26,881 km2 of seasonal open-canopy marshes, 63,607 km2 of yearlong closed-canopy marshes, and 37,399 km2 of deciduous forest wetlands. Heilongjiang Province had the largest area of water-related land cover types, followed by Inner Mongolia, Jilin, and Liaoning Provinces. The spatial distribution and area of the resultant water-related land cover types had high consistency with other public maps and statistical data. Our findings demonstrate the potential of knowledge-based algorithms and integrated multi-source image data for water-related land cover mapping, which could provide useful data for the planning of ecological civilization construction and ecological protection and restoration.