MULTIVARIATE STATISTICAL METHODOLOGIES FOR TESTING HYPOTHESIS OF LAND-USE CHANGE AT THE REGIONAL LEVEL. A REVIEW AND EVALUATION

摘要

Various models of land-use and land-cover change have been developed for better understanding of existing relationships and interactions between human and natural phenomena. Most of these models concentrate on human behaviour (behavioural models) holding a strong explanatory power of the observed patterns of land-use change though very often lacking spatial explicitness. At the same time, there has been a need to study human activities and land-use conversion and modification, with a stronger spatial component and in more depth going into the basic elements that form the underlying causes of land metamorphosis. Combining techniques and methodologies developed in different disciplines might help to better understand and respond to issues related to the changing patterns of land-use. This paper selectively reviews current statistical methodologies for modelling land-use changes. The methodologies are presented and discussed in terms of their mathematical formula as well as the necessary transformations that the dependent and independent variables require in order to fit the models. Different statistical models hold different capabilities that can further help our understanding in human induced land-use and land-cover changes and subsequently promote better decision-making. These models can be used with a wide range of explanatory variables and address serious issues relevant to the level of analysis.