A Dynamic Model of Land Use Change With Spatially Explicit Data

100 p.Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2004.In this study, we use a discrete choice dynamic model of land-use where the agent's choices are regarded as the solution to a dynamic optimization problem. The first result of using of a dynamic framework is that the assumption of observing the system at a stationary state, one that characterizes all earlier studies, is no longer necessary. Secondly, multitemporal analysis makes it possible to better account for the forces that propagate through space and time and are affected by time lags and spatial diffusion processes. Finally, our model introduces more of the complexities that characterize the decision process. Many of these complexities are assumed away in limited dependent variable models and only partially captured in survival analysis models. In particular, the irreversibility of some decisions (e.g. when a primary forest is cleared away, it is not an available choice in the next time period), expectations about future prices, and forward-looking behavior of the land operator are accounted for. For the estimation of the parameters of interest we use a pseudo-maximum likelihood estimator, the Nested Pseudo-Likelihood algorithm. We tested our model using satellite images and other ancillary data for an area in Panama. We calibrated the model using three time periods (1985, 1987, and 1997) and the parameter estimates were used to predict land use change in the year 2000. Our results show that this model improves upon the existing literature in several ways. First, prediction accuracy of land use change is superior to any of the existing models. Second, we demonstrate how simpler models of land use change, models that do not account for friction in moving in and out uses, overestimate the effects of changes in transportation costs. Third, the incorporation of output prices and expectations regarding future states of the system allow to simulate the effects of policies that would otherwise remain unexplored.U of I OnlyRestricted to the U of I community idenfinitely during batch ingest of legacy ETD