Regional land salinization assessment and simulation through cellular automaton-Markov modeling and spatial pattern analysis

• DOI: 10.1016/j.scitotenv.2012.09.013
• Publication Date: 2012-10-17T09:03:54Z
• Authors: De Zhou; Zhulu Lin; Liming Liu
• Keywords: China; Conservation of Natural Resources; Salinity; Spatial Analysis; 15. Life on land; 01 natural sciences; Markov Chains; Soil; 13. Climate action; Soil Pollutants; Computer Simulation; Ecosystem; Environmental Monitoring; 0105 earth and related environmental sciences

Land salinization and desalinization are complex processes affected by both biophysical and human-induced driving factors. Conventional approaches of land salinization assessment and simulation are either too time consuming or focus only on biophysical factors. The cellular automaton (CA)-Markov model, when coupled with spatial pattern analysis, is well suited for regional assessments and simulations of salt-affected landscapes since both biophysical and socioeconomic data can be efficiently incorporated into a geographic information system framework. Our hypothesis set forth that the CA-Markov model can serve as an alternative tool for regional assessment and simulation of land salinization or desalinization. Our results suggest that the CA-Markov model, when incorporating biophysical and human-induced factors, performs better than the model which did not account for these factors when simulating the salt-affected landscape of the Yinchuan Plain (China) in 2009. In general, the CA-Markov model is best suited for short-term simulations and the performance of the CA-Markov model is largely determined by the availability of high-quality, high-resolution socioeconomic data. The coupling of the CA-Markov model with spatial pattern analysis provides an improved understanding of spatial and temporal variations of salt-affected landscape changes and an option to test different soil management scenarios for salinity management.