Wetland degradation monitoring using time series of multi-sensor data: A case study of Yangtze River

Effective monitoring and assessment of land degradation are essential for the land manager and decision-makers to make optimal land management decisions. This work aims to present a methodology to monitor land degradation in the Chongming Dongtan wetland, where wetland degradation and plant invasions have been observed in the past few decades. Time series of multi-sensor imagery from 2013 to 2016 were selected to identify the rate and status of land degradation due to reclamation based on land cover classification and change detection techniques. Results showed that the overall accuracy of each image classification was higher than 80% and Kappa statistics of agreement more than 0.74. The distribution maps of land degradation in each period were generated respectively based on the classification and change detection results. Coastal wetlands decline, biodiversity loss and wetland hydrological regime change are the main degraded types. Although climate change plays an important role, human activities especially via large-scale reclamations are more directly impacted by land degradation in the study area.