Wetland classification by time series satellite imagery

Wetlands are important ecosystems on Earth, yet the knowledge about its distribution is still limited, which mainly comes from wetland transient and dynamic characteristics. Satellite image time series may be promising for the long-term monitoring of wetland ecosystems. The present study firstly introduced the revised wetland classification system, which can be easily used to classify wetland using time series satellite images. Then, moderate resolution imaging spectroradiometer (MODIS) time series data collected in 2001 and 2013 was used to monitor the large 100 Ramsar sites across various continents. The results showed that specific seasonal wetland classes, such as flooded mud, permanent water, and seasonal marshes, can be identified using MODIS time series data with acceptable accuracy. The slight wetland loss (decreased by no more than 1%) noted in the 100 reserves evaluated herein could indicate the effectiveness of efforts to protect Ramsar sites. Due to the dual impact of climatic fluctuation and anthropic development activities, wetland types and the areas in Ramsar sites often show obvious inter-annual variation.