Change and prediction of the land use/cover in Ebinur Lake Wetland Nature Reserve based on CA-Markov model.

Abstract

Taking Ebinur Lake Wetland Nature Reserve as the study area, and based on the classified result of Landsat TM images of 1998, 2006, and Landsat OLI images of 2014, this paper simulated and forecasted land use/cover types of Ebinur Lake Wetland Nature Reserve by using the CA-Markov model. In the model-building process, the transition probability matrix and the transition area matrix were obtained through the Markov model, which determined the conversion rules and iterative times of the CA model. The land use/cover pattern of the study area in 2014, 2022 and 2030 was simulated and forecasted with the CA-Markov model. Then the forecast result was compared with the actual classified data of 2014 to verify the forecast accuracy. The prediction error of land use/cover type area was not more than 6.4% in 2014, and the spatial location accuracy was 76.0%. The results showed that the forest and grassland, saline-alkali area, dry lake bed and desert had an increasing trend in 1998-2014. The saline-alkali area increased the most by 37.4% and the water body decreased by 34.8%. In 2014-2030, the areas of the forest and grassland, saline-alkali soil and desert would increase, while that of the dry lake bed, water body and other objects would decrease. This study could provide a basis for dynamic monitoring land use/cover as well as sustainable development of Ebinur Lake Wetland Nature Reserve.