Forecasting Wetland Transformation to Dust Source by Employing CA-Markov Model and Remote Sensing: A Case Study of Shadgan International Wetland

Abstract

Wetlands are disappearing globally at alarming rates; since 1900, 71% of wetlands have changed into other forms of land cover. The CA-Markov model is one of the most effective methods for forecasting LULC change. In order to predict LULC changes of Shadegan wetland in 2050, images for the years 1980, 1990, 2000, 2010, and 2020 were classified based on segmentation and artificial neural networks (ANNs), and three classes were considered, including vegetation, bare land, and water. To assess accuracy of classification and prediction, the Kappa coefficient was calculated. Results indicate that CA-Markov has moderate predictive capability for future changes. Results of the image classification show that most of the changes occurred in vegetation from 2000 to 2020. So, about 170,000 hectares of this class have been converted to bar land. By comparing the LULC map in 2020 and 2050, if the current trend in the region is continued, in the 2050 year, 79.6% of the total area will be covered by the bare land. Increasing the amount of dry land in the area can create dust sources. During the last years, with the intensification and continuation of drought, dried parts of wetlands such as Shadegan became the most active dust sources in the southwest of Iran. The aerosol optical depth time series data were used to verify the model’s prediction findings. The result of the Mann-Kendall (MK) test shows the positive trend in the AOD time series, indicating an increasing trend in dust concentration.