Modelling Land Use and Land Cover Change in a Central Region of the Republic of Benin, Using a Markov Model
*Vincent Joseph Mama1 and Oloukoi, Joseph2
1Institut National des Recherches Agricoles du Bénin, 06BP1105, Cotonou Bénin Republic
Email: firstname.lastname@example.org. 2Regional Centre for Training in Aerospace Surveys (RECTAS) Obafemi Awolowo University Campus PMB 5545, Ile-Ife, Nigeria. E-mail: email@example.com, firstname.lastname@example.org
Accepted on December 15, 2010
In order to improve the understanding of the land use/cover change, an approach integrating GIS to Markov model was used to analyze vegetation dynamics. The model was built upon a set of land cover change trajectories over two periods of observation years. A time series data from 1986 to 1999 were analysed and the transition sequences and their associated probabilities for deforestation were estimated. The linear functional relationships between two response dependent variables – the probabilities that forested areas can be transformed into non-forested (FORA) and non-forested into forested (NFAR) – and the explanatory variables selected to describe the processes of land use/cover conversion was also analysed. In order to get thorough results, the land use/cover classes were merged in three categories: forested areas, built-up areas and farmlands. Whereas correlation between FORA and Cotton area, FORA and Built-up as well as between Built-up and Carrying capacity were found negative and significant at 0.05 level, relationship between NFAR and Cotton area, NFAR and Built-up area, Cotton area and the distance to farm as well as between the population density and the distance to farm have shown positive correlation. This study has shown that over an observation period of 13 years, farmlands and built-up areas have increased by 18.4% and 1.35% respectively whereas the forested areas have decreased by 19.76%. On the other hand, the model permits to demonstrate the fundamental land use/cover pattern: when built-up area and farmland are estimated to increase, forested lands would decrease. This indicates that further development in land use and land cover change would take place at the expense of forested areas. Overall, this study has highlighted how from developments in geographic information system and remotely sensed image analysis and Markov model, it is possible to predict changes in land use/cover classes.