Forecast of soil salinization change trend based on Markov chain

To reveal the law of soil salinization evolution in Yinchuan plain, In this study, the information of soil salinization of Yinchuan Plain was extracted from TM image in April 2000, CBERS images in April 2004,and China’s environmental satellite image in April 2009, which acquired the distribution figure of soil salinization of the Yinchuan Plain in three periods. The 2000 data was regarded as the model’s initial state. The data in 2000-2004 was used for establishing probability transfer matrix of soil salinization of Yinchuan Plain that was used for simulating dynamic trend of soil salinization in Yinchuan plain and testing its accuracy. Test for model X2 shows that there exists no significant difference between the model predictions and the actual values, these two being in good agreement. Model results show that under the current driving force, the saline soil area will be stable, the area of severe salinization, moderate salinization and mild salinization will be 450.23, 451.66, and 1469.84 square kilometers, respectively, accounting for 30.42% of total area, which will decrease by 5.86% than in 2000 (36.28%), 100 years later. Using markov model to predict the soil salinization in Yinchuan plain is feasible, and the prediction results can be used for providing the theoretical basis for soil salinization management.