Dynamic spatial simulation of residential development in the Hudson River Valley, New York State.

摘要

This dissertation projects residential development in Hudson River Valley watersheds within Dutchess County in New York State using an integrated modeling framework consisting of a novel spatial econometric model, a geographic information system (GIS), and Monte Carlo simulation. The spatial econometric model is used to project residential development at the tributary watershed level, sub-watershed level, and census block level. The GIS is employed to extract socio-economic and county-level tax parcel data to be used in conjunction with bio-geophysical attributes, such as slope, soil, and location attributes, to calculate and project growth trends on a residential level for undeveloped land parcels. Monte Carlo simulation is used to distribute these projections into the GIS to display dynamic development scenarios using regional maps to help policy-makers in their decision-making processes.; Detailed case studies, consisting of several related scenarios, are used to present stakeholders the economic, social, and environmental implications of a possible course of action. The scenario analyses indicated that new development will likely occur in the agricultural and forest land regions confirming the trend of decentralized development. Other scenarios demonstrate possible courses of action that policy-makers could take to prevent decentralized development as well as how to protect the Wappinger Creek Watershed.