| Title | Geothermal Resource Assessment: a Case Study of Spatial Variability and Uncertainty Analysis for the State of New York and Pennsylvania |
|---|---|
| Authors | Gloria Andrea AGUIRRE, Jery R. STEDINGER, and Jefferson W. TESTER |
| Year | 2013 |
| Conference | Stanford Geothermal Workshop |
| Keywords | geothermal resource assessment, exploratory data analysis (EDA), outlier detection and treatment, ordinary kriging, robust regression and LOESS, bootstrapping statistical inference |
| Abstract | Because of its low grade characteristics, geothermal resources in the Northeastern United States have not been carefully explored. Partly, this is because it is not feasible to utilize low grade resources for electricity generation in today’s energy markets. Advanced technologies, such as Enhanced Geothermal Systems (EGS), have shown potential for using low grade resources for both direct thermal use and combined heat and power applications (Tester et al., 2006; Tester et al., 2010). Recently published geothermal maps in New York and Pennsylvania have shown the availability of temperatures for direct thermal use applications at economical drilling depths shallower than 6 km (Shope et al., 2012; Stutz et al., 2012). However, uncertainties in resource assessment coupled with expected high drilling costs for deeper wells often restrict the feasibility of low grade geothermal resources. Our study focuses on utilizing new methodology developed in collaboration with Southern Methodist University (SMU) to quantify the uncertainties associated with the assessment of geothermal heat flow in New York and Pennsylvania, with the possibility of applying these techniques to thermal gradient and temperature-at-depth maps. A thorough statistical and spatial analysis of the variability and the uncertainty associated with the produced maps was undertaken to provide a robust and quantifiable estimate of the availability of geothermal heat flow in New York and Pennsylvania. Areas of higher resource potential and lower uncertainties were found to be along the eastern border of the Appalachian Basin in Pennsylvania. In New York, promising areas included various counties of central New York, as well as the southernmost border of the western and central portion of New York with the neighboring state of Pennsylvania. The methods and techniques applied to this case study are potentially useful for quantifying the uncertainty associated with selecting future exploration sites based on resource maps. |