Prediction Models for Landscape Development in GIS

Ameliorating the impacts of global change on the physical and socioeconomic environment is essential for the restoration and sustainability of our ecosystems. Landscape modifications have been discovered as one of the primary causes of the environmental change and has therefore gained reasonable attention in the modeling techniques, because understanding the land use-land cover change (LULCC), the drivers and processes provides the solution to the environmental challenge. Sequel to this, several empirical methods and software for modeling LULCC have been developed and applied such as the spatial-statistical based (regressions, Artificial Neural Networks, GISCAME), Markov Chain, Cellular automata, the hybrid (CA-Markov), Agent-Based, CLUE, Land Change Modeler (LCM), Dinamica EGO, GEOMOD, and Scenarios for InVEST. This paper reviews the implementations, prospects, and the limits of these modeling software packages. Comparative assessment review of the models including their capabilities, applications and output were also highlighted. Finally, two of the models (LCM and CLUE) were used to predict the LULCC in a municipal area in south-east, Nigeria (a case study), and this helps to illustrate the afore-mentioned explanations and variations about the outputs of different models in assessing the LULCC of same location in time. Different models can behave differently when applied in same location at the same time as demonstrated by the applications of LCM and CLUE in our study. In addition to other LULC type dynamics in the models outputs, we have prediction map from CLUE showing higher built-up areas (42.7 km²) compared with that of LCM result (35.2 km²) while, the LCM projection revealed more areas for light vegetation cover (29.5 km²) in comparison with the 16.5 km² from the CLUE model result.