Cellular automata Markov chain model based deforestation modelling in the pastoral and agro-pastoral areas of southern Ethiopia

Abstract

Permanent conversion of woodland to large-scale commercial agriculture, pastures or urban areas and temporary or partial removal of indigenous trees for shifting cultivation and selective logging remained major environmental challenges in the tropical region. Cognizant of the environmental changes prevailing in the pastoral and agro-pastoral areas of Southern Ethiopia, we have examined the past conversion of woodland to other land uses through the analysis of Landsat Multi-spectral scanner (MSS) 1973, Thematic Mapper(TM) 1986, Enhanced Thematic Mapper (ETM+) 2003, Operational Land Imagery (OLI) 2017 and then projected the future change in land use/cover (LUC) as well. We have employed Cellular Automata Markov chain model to simulate and predict LUC changes between 2017 and 2060. Four spatial driver variables such as distance to road and settlement, slope and elevation were used to run the simulation. Prior to the prediction, we have simulated the LUC of 2017 using transition potential maps of 2003 and transition matrix between 1973 and 2003. The predictive power of the model was then examined by comparing the reference and simulated LUC maps of 2017and also using the kappa index. A good correlation was obtained between the reference and simulated LUC maps of 2017. In addition, the computed kappa index was above 0.9, which implies that the model is effective in predicting change in LUC. The analysis result revealed that in the entire monitoring period (1973–2017) the area lost 89,875 ha of woodland. The loss is expected to continue during the period 2017–2060, with an estimated loss of 32,423 ha of woodland, if a proper measure is not taken against the continuous loss of woodland. Thus, relentless efforts are needed to rehabilitate the already degraded land and also minimize the potential loss of woodland in the future through the implementation of conservation – livelihood approach, REDD + project, and sustainable land use management strategies.