Integrating Geographic Information Systems and Remote Sensing with Spatial Economic and Mixed Logit Models for Environmental Valuation

This research focuses on the Emory and Obed Watersheds in the Cumberland Plateau in Central Tennessee and the Lower Hatchie River Watershed in West Tennessee. A framework based on market and nonmarket valuation techniques was used to empirically estimate economic values for environmental amenities and negative externalities in these areas. The specific techniques employed include a variation of hedonic pricing and discrete choice conjoint analysis (i.e., choice modeling), in addition to geographic information systems (GIS) and remote sensing. Microeconomic models of agent behavior, including random utility theory and profit maximization, provide the principal theoretical foundation linking valuation techniques and econometric models. The generalized method of moments estimator for a first order spatial autoregressive function and mixed logit models are the principal econometric methods applied within the framework.

The dissertation is subdivided into three separate chapters written in a manuscript format. The first chapter provides the necessary theoretical and mathematical conditions that must be satisfied in order for a forest amenity enhancement program to be implemented. Such a program is possible and would yield an efficient outcome under three conditions: (1) contributors are willing to pay an amount that maximizes the utility they derive from forest amenities; ( 2) an intermediary party sets a compensation price based on contributor aggregate willingness to pay such that the social value of the program is maximized; and (3) a participating landowner maximizes profit given this incentive. The second chapter evaluates the effect of forest land cover and information about future land use change on respondent preferences and willingness to pay for alternative hypothetical forest amenity enhancement options. Land use change information and the amount of forest land cover significantly influenced respondent preferences, choices, and stated willingness to pay. Hicksian welfare estimates for proposed enhancement options ranged from $57.42 to $25.53, depending on the policy specification, information level, and econometric model. The third chapter presents economic values for negative externalities associated with channelization that affect the productivity and overall market value of forested wetlands. Results of robust, generalized moments estimation of a double logarithmic first-order spatial autoregressive error model (inverse distance weights with spatial dependence up to 1500m) indicate that the implicit cost of damages to forested wetlands caused by channelization equaled -$5,438 ha^-1.

Collectively, the results of this dissertation provide economic measures of the damages to and benefits of environmental assets, help private landowners and policy makers identify the amenity attributes preferred by the public, and improve the management of natural resources.