Exploring the driving forces behind deforestation in the state of Mexico (Mexico) using geographically weighted regression

The goal of this work is to explore diverse factors that have induced the loss of forest areas in the State of Mexico from 1993 to 2000, using a technique for local analysis known as Geographically Weighted Regression (GWR). The behavior of variables at a local level was analysed, and some of them were found to present significant spatial variability. This represents an improvement on the understanding offered by global analysis as, rather than producing an average coefficient for the entire territory, the technique yields an estimated coefficient for each location analysed; in other words, the type of relationship that exists in each portion of the territory is ascertained, and not only a general overview. This method reveals aspects of the relationships which do not emerge when using traditional global specifications, such as sign change in some of the parameter estimates.