Spatial Econometric Analysis of a Watershed Utilizing Geographic Information Systems: Water Quality Effects of Point and Non-Point Pollution Sources in the Neuse River Basin, NC.

Abstract

This study utilizes elements of several different fields of study to facilitate more effective and efficient policy development for water pollution control. In order to implement efficient environmental policy, spatial aspects of watersheds should be carefully incorporated into empirical analysis. The geographical attributes of a watershed induce various spatial stochastic processes, causing surface water quality data in streams to have a unique spatial structure. In this study, geographical data of watersheds are collected and manipulated to find a consistent basis for comparing measures of pollution sources with variations in water quality across hydrologic units in the Neuse River basin in North Carolina.

This research seeks to calibrate an empirical watershed model using available spatial (statistical) analytical techniques. Methods are demonstrated of utilizing Geographic Information Systems (GIS) to convert data from multiple sources to a common basis for water quality analysis. A spatial autoregressive response model is chosen considering spatial aspects of a regional watershed, and a corresponding structural watershed model is constructed. The empirical watershed model is designed to incorporate spatial effects and to produce accurate estimates. The model specifies that the spatially weighted sum of neighbor water qualities (total nitrogen [TN] concentrations) affects the TN concentration of each downstream monitoring unit, as do the standard covariates of local pollution sources and heterogeneous watershed characteristics. The completed standard econometric analysis includes cross-sectional estimation of several functions predicting TN concentration in streams conditional on watershed characteristics and potential sources of TN in the hydrologic unit.

Results show that a clear understanding of regional spatial capacity will help avoid overuse of water resources. Specific knowledge of spatial information and empirical relationships allows improved design of controls on economic activity across regions (e.g., Total Daily Maximum Daily Load [TMDL] and nutrient trading programs) to preserve environmental resources. The study concludes by recognizing that a more robust watershed analysis would require more spatial data refinement and the option of panel data analysis.