Application of modified cellular automata Markov chain model: forecasting land use pattern in Lebanon

Abstract

This article modifies the use of the Cellular Automata Markov Chain Model to predict future land use pattern in Lebanon, and compares it to the current developed model. LandSat images of years 2000, 2009 and 2018 are used to generate land use maps within the geographic information system. Current developed model was generated by integrating Population density data with land use classification maps to decompose the built-up development to three sub-classes: High, Medium and Low-density built-up land uses. Simulations of future land use pattern over the year 2018 based on these two prediction models reveal that the Modified Cellular Automata Markov Chain Modelling technique is more accurate than the Extended Markov Chain model. Spatial effects of built-up densities are validated in this study. Consequently, the extension of the Cellular Automata Markov Chain Model represents an innovative tool for regional and urban planning to forecast potential locative distribution of old and new urban agglomeration. The sequential shift of the urban areas among different density classes in addition to the interactions of urban agglomerations should be employed as a guiding tool for decision-makers and planners during the phase of developing new population and economic strategies, new urban Masterplan and during the process of enacting/developing new land-use policies. In the final part of the study, a simulation of land use pattern for the year 2036 is generated using TerrSet v.18 software and an analysis of the outcome for the forecasted map is discussed.