Masters Thesis

Nevada geothermal favorability mapping, weights of evidence smoothed functions and refined methodology

Geothermal energy provides clean renewable electrical power in the United States and around the world. It is most efficiently harnessed from active hydrothermal systems, which only occur in very specific tectonic environments with naturally occurring permeability. Most known systems have surface manifestations such as hot springs or fumaroles that led to their discovery, however a significant resource potential resides in blind systems that are not expressed at the surface. One of the largest barriers to successful exploration for blind systems stems from uncertainty in characterizing resource potential. This study develops new methodology and applies it to arrive at a new geothermal favorability assessment to predict where best to explore for active hydrothermal systems. While drilling is ultimately necessary to verify a resource, regional assessment tools can help reduce uncertainties, refine the identification of prospective areas, and significantly reduce costs associated with drilling and other exploration efforts. The approximately 286, 000 km2 study area covered the entire state of Nevada, which has high potential for containing undiscovered geothermal resources. Nevada has relatively comprehensive regional coverage of publicly available geological and geophysical data directly related to processes that control hydrothermal systems including: Quaternary faults, crustal heat flow, crustal strain rates, and seismicity. These data, known as evidence data were compiled and assessed to determine which were the most useful for predicting resource occurrence and to determine which processing techniques best transform raw geologic data into forms useful for statistical analysis. Locations of fifty-five moderate and high temperature geothermal systems served as training data, the benchmarks used to calculate weights for evidence data. The general framework of this study was adopted directly from prior geothermal assessments that used Weights of Evidence (WofE), a Bayesian data-driven technique to quantify spatial relationships between training and evidence data. This study sought to apply the best practices seen in prior assessments and to further refine those approaches when possible by developing and applying new techniques. New exploratory techniques were assisted by the development and application of customized Python-based tools that automated different data processing approaches and subsequent visualization of the resulting WofE statistics. Additionally, a technique was developed and applied in this study to calculate "smoothed" weights by fitting mathematical functions to weights generated using the traditional WofE approach. Generating smoothed weights eliminates a geologically artificial relic of the modeling process, the abrupt changes in favorability at category boundaries seen in prior WofE-based assessment maps. Smoothing may also provide greater accuracy in some parts of the map. Weight functions calculated in this study could potentially be applied to future studies using identical evidence in analogous tectonic environments. Ultimately, data-driven geothermal favorability maps were produced that could be used to guide future exploration for undiscovered geothermal resources, and several prospective target areas were identified. Three evidence layers were used to create final predictive surfaces, resulting in a conservative view of favorability driven by layers with relatively strong correlations and consistent distribution across the study area. Approaches and results from this study were compared with the most relevant prior assessments to contextualize decisions made and results discovered. Techniques developed in this project could be applied in other geographic regions or applied to investigations of other types of natural resources.

Relationships

In Collection:

Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.