Comparative analysis of remote sensing water indexes for wetland water body monitoring using Landsat images and the Google Earth Engine Platform0 (A Case study: Meighan Wetland, Iran)

Wetlands are dynamic and complex aquatic ecosystems that play an important role in the survival of many plant and animal species. This study modeled the spatio-temporal changes of the Arak Meighan wetland during 1985–2020 using the multi-temporal Landsat images. In doing so, the applicability of different satellite-derived indexes including NDVI, NDWI, MNDWI, AWEIsh , AWEInsh , and WRI was investigated for the extraction of surface water from Landsat data. In addition, the correlation coefficient between the images extracted by the indexes is also estimated. For comparison purposes SVM classification was applied on some selected images.  Because the Meighan wetland has a very shallow water depth and contains saline water, the capability of water indexes to extract surface water areas of the wetland has been investigated on this area. The results of the study indicate that climate change such as increasing rainfall has a direct relationship with changes in the surface water area of wetland extracted by various indexes in the last 35 years. Also, the AWEIsh  and AWEInsh  indexes are not able to properly and accurately extract surface water area when there is a drought and the water depth is reduced. Overall, it is not recommended to rely on the results of one index to monitor surface water area during drought condition. Since there is a combination of low water and high water years in time-series analyses, it is seriously recommended that in similar studies, several indexes be used simultaneously for detecting and extracting surface water. However, classification can be used as a robust method for water body extraction in all times.