Dynamics of landscape patterns in an inland river delta of Central Asia based on a cellular automata-Markov model

The analysis of landscape pattern changes is of significant importance for understanding spatial ecological dynamics and maintaining sustainable development, especially in wetland ecosystems, which are experiencing indirect human disturbances in arid Central Asia. This study attempted to examine the temporal and spatial dynamics of landscape patterns and to simulate their trends in the Ili River delta of Kazakhstan through quantitative analysis and a cellular automata (CA)-Markov model. This study also sought to examine the effectiveness of using the CA-Markov model for investigating the dynamics of the wetland landscape pattern. The total wetland area, including the river, lake, marsh, and floodplain areas, and the area of sandy land have remained steady, while that of desert grassland has decreased slightly, and shrublands have increased slightly from approximately 1978 to 2007. However, the wetland and shrubland areas exhibited a trend of increasing by 18.6 and 10.3 %, respectively, from 1990 to 2007, while the desert grassland and sandy land areas presented the opposite trend, decreasing by 30.3 and 24.3 %, respectively. The landscape patterns predicted for the year 2020 using probabilistic transfer matrixes for 1990–2007 (Scenario A) and 1990–1998 (Scenario B), respectively, indicated that the predicted landscape for 2020 tends to improve based on Scenario A, but tends to degrade based on Scenario B. However, the overall Kappa coefficient of 0.754 for the 2020 predicted landscapes based on Scenarios A and B indicates that the differences in the predicted landscapes are not distinct. This research indicates that the applied CA-Markov model is effective for the simulation and prediction of spatial patterns in natural or less disturbed landscapes and is valuable for developing land management strategies and reasonably exploiting the wetland resources of the Ili River delta.