| Groundwater resources have become more vulnerable to pollution due to rapid growth and urbanization. Regional-scale groundwater vulnerability assessment models based on multi-criteria decision-making (MCDM) approach have been widely used to define intrinsic aquifer vulnerability. While regional-scale aquifer vulnerability delineation has been an active area of research, some of the fundamental and application issues have not been completely addressed. These limitations arise due to spatial and temporal variability of input data (scale and resolution), data processing methods (sampling and interpolation methods), subjectivity in assigning weights and ratings by decision makers, non-integration of intrinsic and specific vulnerability, and non-linear relationships between the hydrogeological parameters. The overall goal of this dissertation is to develop a framework addressing the above questions and enhance the applicability of regional-scale MCDM-based intrinsic aquifer vulnerability assessment methods.;In particular, a statistical framework is illustrated to evaluate the effects of scale, resolution, sampling and interpolation under differing granularization of output. An innovative methodology based on fuzzy multi-attribute theory is proposed to address the issues of imprecision in weights and ratings specification. Several new probability-based metrics for aquifer vulnerability unifying both intrinsic and specific vulnerability characterizations are developed. A new Aquifer vulnerability index based on the principle of conservation of mass that accounts for multi-species nitrogen transformation in the vadose zone is developed and implemented in GIS.;Specific application results obtained for the 18 county study area in South Texas include -- (1) low resolution data appear to provide less conservative depiction of intrinsic vulnerability; (2) incorporation of decision makers imprecision with regards to weighting and ratings lead to more conservative estimates for the intrinsic vulnerability and are consistent with precautionary principle; (3) intrinsic vulnerability and specific vulnerability characterizations do not necessarily coincide and the use of combined indices are weighted more towards specific vulnerability characteristics; (4) nitrate-nitrogen concentrations are strongly correlated to land use characteristics and can be reasonably modeled with the developed multi-species model. In conclusion, the proposed research outcomes address major sources of uncertainty in regional-scale aquifer vulnerability characterization including those arising from data acquisition scale, differences in decision makers' preferences as well as model structure. |
