Detection and predictive modeling of land use changes by CA-Markov in the northern part of the Southern rivers: From Lower Casamance to Gba river (Guinea Bissau)

The northern part of the Southern rivers, which is subject this study, lies between 18° 54' and 14° 48' north latitude 11° 30' 12° west longitude. It characterized by a dense hydrographic network dominated maritime marshes expanses salty lands mangrove mudflats that have developed on vast coastal plain. objective study to analyze spatio-temporal changes geomorphological units simulate their spatial behavior in 2035. methodology based processing Landsat TM OLI satellite images acquired 1986 2018, respectively. Change detection performed post-classification comparison two previously validated land cover states with overall accuracies 91% for image 97% 2018 image. prediction made using CA-Markov model yielded result 88% agreement. This it possible understand evolution use near future (2035). results 2035 show bare will highest gain 93% probability while tans decrease 60% other categories. Vegetation mangroves lose 47 44% areas respectively soil mudflats. A strong degradation vegetation predicted Markov chain change area covered water lowest during period.   Key words: CA-Markov, landsat, change, predictive modeling, rivers.