Modeling tropical deforestation in the southern Yucatan peninsular region:comparing survey and satellite data

This paper presents some initial modeling results from a large, interdiscip linary research project underway in the southern Yucatan peninsular region. The aims of the project are: to understand, through individual household s urvey work, the behavioral and structural dynamics that influence land mana gers' decisions to deforest and intensify land use; model these dynamics an d link their outcomes directly to satellite imagery; model from the imagery itself; and, determine the robustness of modeling to and from the satellit e imagery. Two complementary datasets, one from household survey data on ag ricultural practices including information on socio-economic factors and th e second from satellite imagery Linked with aggregate government census dat a, are used in two econometric modeling approaches. Both models test hypoth eses concerning deforestation during different time periods in the recent p ast in the region. The first uses the satellite data, other spatial environ mental variables, and aggregate socio-economic data (e.g., census data) in a discrete-choice (logit) model to estimate the probability that any partic ular pixel in the landscape will be deforested, as a function of explanator y variables. The second model uses the survey data in a cross-sectional reg ression (OLS) model to ask questions about the amount of deforestation asso ciated with each individual farmer and to explain these choices as a functi on of individual socio-demographic, market, environmental, and geographic v ariables, in both cases, however, the choices of explanatory variables are informed by social science theory as to what are hypothesized to affect the deforestation decision (e.g., in a von Thunen model, accessibility is hypo thesized to affect choice; in a Ricardian model, land quality; in a Chayano vian model, consumer-labor ratio). The models ask different questions using different data, but several broad comparisons seem useful. While most vari ables are statistically significant in the discrete choice model, none of t he location variables are statistically significant in the continuous model . Therefore, while location affects the overall probability of deforestatio n, it does not appear to explain the total amount of deforestation on a giv en location by an individual. (C) 2001 Elsevier Science B.V. All rights res erved.