Title

Using GIS to Analyze Tracer Test Data at the Geysers Geothermal Power Plant

Author

Karly McCrory
American River College, Geography 350: Data Acquisition in GIS; Fall 2003

Abstract

Operators at the Geysers Geothermal Power Plant in Northern California were faced with the challenge of processing a large volume of tracer test data on a daily basis. Three methods were evaluated for processing this data: hand-plotting, CAD, and GIS. Researchers quickly discovered that GIS had value that the other methods did not: functionality. Utilizing basic GIS features and combining them with customized user interfaces automated a previously manual process—greatly reducing the amount of time and money spent on analyzing tracer test data at the Geysers.

Introduction

The Geysers is part of the Clear Lake volcanic range in Northern California. As an active volcanic area, Geysers is well suited for geothermal energy. Geothermal energy is harnessed through steam and hot water directly from the Earth, and can be used to generate electricity (Figure 1).

Figure 1. How a geothermal power plant works.

In 1960, a geothermal power plant was established at Geysers. Gradually, the Geysers plant (below) grew to become the largest geothermal development in the world (Kious and Tilling, 1996). 

Photograph by Julie Donnelly-Nolan, USGS

At first, steam production began only in the north-central region of Geysers, but by 1980 production had gradually spread to the entire stream field. In the mid-1970s, concerns about the abundance of Geysers’ reserves began to emerge. In 1975, operators began tracer tests at Geysers to track both the dry-out of the reservoir, and the rate at which the reservoir naturally re-saturates itself (Beall et al, 2001).

When tracer testing began at the Geysers, tests were only run every two to four years, so there wasn’t a need for a mapping tool other than hand-plotting. Over the years, the frequency of tracer tests increased dramatically. Currently, the return time for a tracer test is often less than one day for nearby wells (Nash and Adams, 2001). The results from a day of testing must be analyzed to determine which wells are tested the following day. Each tracer test generates a large amount of data that must be mapped and interpreted efficiently. This paper discusses the different methods used for visualizing tracer test data, and how GIS reduced costs and increased efficiency at the Geysers.

Background

GIS has many capabilities and a wide variety of applications, such as the ability to tie a location on a map to a table of attributes, and the ability to query said table. Yet given the obvious benefits and applications of GIS in the geothermal industry, it has not been widely adapted. The Energy & Geoscience Institute (EGI) in Salt Lake City, Utah has been at the forefront of bringing GIS to the geothermal industry.

The data used in this project was not hard to find, however it was also not plentiful. The literary and data sources used for this project were available over the Internet. The two main publications used came from the Geothermal Resources Council Transactions, a collection of papers for a given year.

I also tried to find information about other geothermal plants that may use GIS to map tracer test results, but I could not find any. It seems that although the geothermal industry has a great need for this tool, they are far behind in actually utilizing it to its fullest extent.

Methods

The Geysers operators tasked the researchers at EGI with finding a way to effectively manage tracer test data. The researchers decided to evaluate three approaches to solve the Geysers problem: (1) hand-plotting, (2) computer aided design (CAD) software, and (3) GIS.

First they evaluated the approach that the Geysers operators were already using: hand-plotting. The first drawback of the hand-plotting method was an obvious one—the large amount of data from a single tracer test was overwhelming to conquer with such a slow and tedious process. Researchers also found that many times the map maker would try to draw the map to a degree of accuracy that goes beyond the definition of an isoline, adding even more time to an already lengthy process (Nash and Adams, 2001).

Researchers then decided to look to computers to solve their problem. They decided to evaluate two types of computer mapping programs: AutoCAD LTâ 2000i with Autodesk CAD Overlayâ versus ArcView 8.1 with ArcView Spatial Analystâ. While both programs provided a method for digital mapping, thereby decreasing the amount of time spent on map creation, ArcView offered functionality that CAD did not.

With ArcView, not only could the user create and isarithm map quickly, they could also link points on the map to an attribute table. Even further, they could superimpose the isarithm map on satellite, topographic, or geologic maps of the area (Nash and Adams, 2001). By utilizing the query function in ArcView, operators were also able to quickly and efficiently analyze the results of the tracer tests.

Results

Using ArcGIS Spatial Analystâ to create isolines, researchers found that with some initial experimentation, good statistical surfaces could be generated very quickly (Figure 2). Numeric values are given next to sampled wells for comparison to the computer-generated contours. Because the contouring parameters can be reset, several new maps can be created for comparison in a matter of a few minutes. The user only has to repeat the process a few times to get a feeling for the proper parameters, enabling them to generate new maps even more quickly (Nash and Adams, 2001).

Figure 2. Isarithm map.

Figure 3. Classified graduated circle map.

Researchers also experimented with classifying attributes in different ways, along with different visual representations of those classifications. They found that if a high degree of spatial accuracy was needed, graduated circles were preferred (Figure 3). Because the human eye is able to easily differentiate between the sizes of the circles, visual comparisons of the data were made much easier. Researchers also experimented with combining graduated circles with other shapes and graduated color to represent multiple data sets per sample point (Figure 4) (Nash and Adams, 2001).

Figure 4. Combination of multicolored graduated circles and graduated symbol map.

GIS was proving to be a useful tool for analyzing and mapping the tracer test data. However, researchers soon discovered that the Microsoft Excelâ spreadsheet they were using was not the best method for importing data into ArcView. Because the Excel format was not favorable, the data had to be reformatted and exported into a dBase table or comma delimited text file for import into ArcView. Researchers decided to develop a new database format that would be used as an input for a new bridge interface (Figure 5) (Nash and Adams, 2001).

Figure 5. Bridge interface.

The bridge interface is then used to create a new database in Microsoft Accessâ. Both raw tracer and steam flow data are needed for the bridge interface to generate the new database and tables. The bridge interface tool makes the process of database generation very fast and easy to use. In addition, the interface can be used with any GIS software that facilitates an SQL connection to Access. The format for input of the data is:

Well_Names

Date1

Date2

Date3

Well_1

Data1

Data2

Data3

Well_2

Data1

Data2

Data3

 

The Well_Names field is used as the unique identifier for any table relations or joins. When the Access database is generated, it contains the following seven tables: 1) raw tracer data, 2) normalized data, 3) log transformed normalized data, 4) interpolated data, 5) cumulative values of interpolated data, 6) normalized interpolated data, and 7) log transformed normalized interpolated data. The user defines the normalization value, and the interpolation process estimates missing values.

The bridge interface is linked to the Tracer Data Analyst (Figure 6), a customized object oriented user interface within ArcView. This interface can call up the bridge interface, rapidly create SQL connections to tables that have been generated, and facilitated automatic joins to generate isarithmic maps. To make map generation possible, a table consisting of well names and XY coordinates must be added.

Figure 6. ArcView Tracers Data Analyst user interface.

Conclusions

In the end, GIS proved a far better solution for the Geysers Geothermal Power Plant than either hand-plotting or CAD did. The main reason behind this is the functionality that GIS brings to cartography. By giving the users the ability to join and query tables, use different attribute classifications to increase visual effectiveness, and the capability to overlay maps, GIS became a very effective solution to the Geysers problem. Researchers furthered this effectiveness by customizing user interfaces; thereby greatly reducing the man-hours needed to process tracer test data.

References

Beall, J.J., M.C. Adams and J.L.B. Smith, 2001. Geysers Reservoir Dry Out and Partial Resaturation Evidenced by Twenty-Five Years of Tracer Tests. Geothermal Resources Council Transactions, v. 25, p. 725-729.

Kious, W.J. and R.I. Tilling, 1996. This Dynamic Earth: The Story of Plate Tectonics. U.S. Geological Survey General Interest Publication.

Nash, G.D. and M.C. Adams, 2001. Cost-Effective Use of GIS for Tracer Test Data Mapping and Visualization. Geothermal Resources Council Transactions, v. 25, p. 461-464.

Data Reference: http://www5.egi.utah.edu/Geospatial_Data/The_Geysers__California/the_geysers__california.html