Projection of future land use/cover change in the Izeh-Pyon Plain of Iran using CA-Markov model

Abstract

Detection and prediction of land use and land cover (LULC) changes in natural resource management and environmental monitoring provide the regional and national decision-makers with useful information. Izeh-Pyon Plain as one of the important wildlife habitats in Khuzestan Province, Iran was selected to detect LULC changes in the past three decades (1985–2017), and LULC in 2033 was also predicted. The LULC maps were obtained using the maximum likelihood classification and Landsat images for (TM) 1985, (ETM+) 2001, and (OLI) 2017. The LULC mapping for 2033 was done using cellular automata and Markov chain (CA-Markov) model and validating the model in the 2017 map simulation. Suitability maps were prepared for each LULC class using weighted linear combination method and applying constraint maps, whereas the weight of each criterion was determined in analytical hierarchy process and standardized based on fuzzy theory. Furthermore, CA-Markov validation was performed using three measures of quantity disagreement, allocation disagreement, and figure of merit. The results showed that from 1985 to 2017, wetlands, forests, and grassland areas decreased by 43.7%, 9.21%, and 8.43%, respectively. In contrast, agricultural lands and residential areas increased by 26.38% and 129.3%, respectively. This decreasing/increasing trend will continue up to 2033, so that one of the wetlands will completely dry out by 2033 and compared with 1985 and 2017, total wetland area will decrease by 68% and 44%, respectively. Since these wetlands are home to many birds and aquatic animals and are considered the tourist attraction of the region, their destruction and the increase of crop production will seriously threaten the ecosystem of the region in the future.