Subtask 3.1.1: Hyperspectral Remote Sensing

Transcription

1 Subtask 3.1.1: Hyperspectral Remote Sensing Investigation of 1) geothermal-indicator minerals with hyperspectral thermal infrared remote sensing data and 2) the relationship of gravity data with a shallow temperature thermal anomaly Final Report September 2010 December 2010 Christopher Kratt Desert Research Institute 2215 Raggio Parkway Reno, NV ckratt@dri.edu PREPARED FOR THE UNITED STATES DEPARTMENT OF ENERGY Under Award No: DE-EE

2 Table of Contents Investigation of 1) geothermal-indicator minerals with hyperspectral thermal infrared remote sensing data and 2) the relationship of gravity data with a shallow temperature thermal anomaly... 1 List of Figures... 3 Executive Summary... 5 Introduction... 6 Study area location and geologic background... 7 Structural setting... 7 Geothermal evidence... 9 Stratigraphy Data sets and methods TIR remote sensing Background Processing and analysis methods Gravity Survey Results TIR remote sensing Gravity survey Publications / Presentations References Cited Appendix I Appendix II Appendix III Appendix IV

3 List of Figures Figure 1. Red outline shows the area covered by the airborne hyperspectral survey, a.) shaded relief map of Great Basin, Nevada state border outlined in blue, Walker Lane Belt approximated by black lines b.) major faults of the Mina Deflection shown in black, arrows show sense of motion c.) mountains and relevant features in the study area: CSM Columbus Salt Marsh, EP Emigrant Peak, FLV Fish Lake Valley, EPFZ Emigrant Peak Fault Zone, RM Rhodes Marsh, TM Teels Marsh, EMFS Excelsior Mountain Fault System, BSF Benton Spring Fault (faults modified from Wesnousky, 2005). BLM geothermal leases were granted at Rhodes and Teels Marshes after the 2m temperature data were published in Figure 2. Two-meter-deep temperature survey and VNIR-SWIR hyperspectral mineral maps, T = tincalconite, 71-1 is explained in the next figure Figure 3. Temperature gradient measured at Figure 4. Green dots mark field sample locations discussed in the Results section, red outline shows boundary of hyperspectral TIR data collection, and generalized geologic map (modified from Albers and Stuart, 1972), all displayed on top of shaded relief Figure 5. Spatial overlap of gravity survey, TIR survey, VNIR-SWIR survey and 2m temperature survey Figure 6. Figure 6. RGB display corresponding to channels 9.97 µm, 9.77 µm, and 9.46µm, respectively. Flightlines that appear blue displayed spectral artifacts between 9.2 and 10.2 µm (upper right), whereas flightlines that appear reddish brown lacked these artifacts (upper left) Figure 7. Montmorillonite (blue) and illite (green) library spectra above compared with commonly observed montmorillonite/illite mixed image spectra below Figure 8. Tincalconite library reference spectra (green) compared with mean image spectra (blue) in an area of high-purity tincalconite evaporite crust Figure 9. Library reference spectra for calcite Figure 10. Example mineral maps. Kaolinite is below the resolution of the map scale but station e-roi-3 is a kaolinite match and is represented by field sample spectra shown in Appendix Figure 11. Color ramp applied to the CBA, white circles are gravity stations. The relationship of the CBA to geology is discussed in the text Figure 12. Contour of the CBA and relationship to the 2 m anomaly, contour interval is 0.2 mgal

4 Figure 13. CBA horizontal gradient and relationship to the 2m anomaly, red colors indicate steeper gradients, blue colors indicate gentle gradients

5 Executive Summary Columbus Salt Marsh (CSM) is located in Esmeralda County, Nevada and exhibits evidence for a potential blind geothermal systems. Favorable geothermal indicators include spring water geothermometry, borate and sulfate evaporite crusts, structural setting, and a shallow temperature anomaly. Aside from the work done to identify the aforementioned characteristics CSM has otherwise received little attention as an exploration target; therefore, CSM was selected to investigate the potential of mapping geothermal-indicator minerals with the Spectrally Enhanced Broadband Array Spectrograph System (SEBASS) airborne hyperspectral thermal infrared (TIR) imaging data. More than 260 km 2 were covered with 5 m spatial resolution and 128 narrow wavelength channels across the µm wavelength region. Outcrops displaying endmember spectra were sampled and compared with laboratory and reference spectra. Potential geothermal-indicator mineral maps were created for calcite, kaolinite, and gypsum. One of the main research objectives was to see if the TIR data elucidated evaporite minerals that were not identified with previously acquired airborne hyperspectral visible and short-wave infrared data. New evaporite minerals were not identified in the area covered by SEBASS at CSM, likely because they had not yet formed by the evaporation of groundwater when, or that they simply don t exist at CSM. One interesting observation was that the borate mineral, tincalconite, exists in at least two, large high-purity locations but did not display strong spectral features in the SEBASS data that are apparent in library reference spectra. The results of this survey provide exploration geologists some guidance for using and interpreting SEBASS data. In addition to analysis of the SEBASS data a gravity survey, centered on a shallow temperature anomaly, was conducted at CSM. The purpose of the gravity survey was to explore the relationship of the shallow anomaly and the underlying structure. The gravity data clearly show the shallow temperature anomaly residing in a gravity saddle. Forward modeled twodimensional cross sections of the gravity data did not suggest the presence of a major fault that would facilitate direct upwelling of thermal fluids that would explain the shallow thermal anomaly. The shallow anomaly may instead be a result of lateral outflow that is forced closer to the surface by paleotopography. Suggestions to test for lateral outflow or direct upwelling are made in the Results section. Shallow temperature surveys have been responsible for detecting seven previously unidentified blind geothermal systems in the Great Basin during the past five years. In areas where shallow temperature survey data exists nearly all of the corresponding deeper sub-surface data is not in the public domain. The gravity survey results reported herein begins to provide the geothermal exploration community with insight to the relationship of shallow temperature anomalies and sub-surface geologic data. 5

6 Introduction Over the past decade several studies have explored the use of aerospace mulit and hyperspectral visible - near-infrared (VNIR) and short-wave infrared (SWIR) data for geothermal exploration; very limited attention, on the other hand, has been given to the utilization of thermal infrared (TIR) remote sensing data for the same purpose (e. g. Coolbaugh et al., 2000; Vaughan et al., 2003). The VNIR and SWIR cover the µm wavelength region where reflected solar energy exhibits maximum signal strength. Due to molecular electronic and vibrational processes, sulfates, carbonates, borates, and hydrated silicates display unique diagnostic spectral signatures in the VNIR and SWIR (Hunt, 1980). The TIR covers the 5-25 µm wavelength region where radiant energy emitted from rocks and minerals exhibits maximum signal strength (Hunt, 1980). Vibrational processes are also at work in the TIR where non-hydrated silicates, such as feldspars, quartz, and opal, in addition to the same mineral groups mentioned for the VNIR- SWIR, display differentiable spectral emissivity features (Christensen et al., 2000). Geothermal-indicator minerals are formed by several different processes. Upwelling geothermal fluids along faults and fractures can alter the host rock to clay and sulfate minerals. Silica-sinter (opal) and carbonate deposits (travertine and tufa) are chemical precipitates that can be deposited by geothermal fluids in and proximal to faults and fractures. Even when geothermal fluids have cooled they may still carry sulfate and borate ions that can form the associated evaporite minerals (e.g. gypsum, tincalconite) when these waters reach or come near to the surface. Geothermal-indicator minerals, like those mentioned above, can be identified and mapped with high spatial resolution (< 10m 2 per pixel) and spectral resolution (> 100 channels) remote sensing data sets (e.g. Kratt et al., 2006; Kratt et al., 2009; Kratt et al., 2010a) that in turn provide clues to where hidden geothermal resources may exist. Our goal was to identify and asses mineral spectral emissivity endmembers with high spatial resolution airborne hyperspectral TIR data in an area of known potential geothermal resource. Particular attention is given to geothermal-indicator minerals and their mapped distributions were compared with high spatial resolution VNIR-SWIR airborne hyperspectral reflectance mineral maps from Kratt et al In case the remote sensing results did not reveal anything especially new or useful for geothermal exploration a gravity survey, centered on a shallow thermal anomaly, was conducted as a hedge. The shallow thermal anomaly was detected with a technique described by Coolbaugh et al. (2007a) and reported in Kratt et al. (2009). During the past five years seven previously undiscovered geothermal systems have been identified with 2m-deep surveys (Kratt et al., 2008; Kratt et al., 2010b). The survey method also outlined upwelling zones at six additional locations where only single-point anomalies existed. Yet, uncertainty about how to interpret 2m survey data exists within the geothermal exploration community. Much of this uncertainty is due to the lack of publicly available geophysics and drilling data to corroborate 6

7 shallow temperature anomalies. The gravity survey in this study helps to address this last point, in addition to refining a target for temperature gradient drilling. Study area location and geologic background Structural setting Columbus Salt Marsh (CSM) is located in Esmeralda County, NV immediately to the northwest of Coaldale Junction where U. S. Highway 95 and U. S. Route 6 meet (Figure 1). CSM playa is bound to the north by the Candaleria Hills, to the west by Davis Mountain and to the northeast by the southwestern end of the Monte Cristo Range. The Volcanic Hills bound the southwest margin of the playa and the Emigrant Hills bound the southeast playa margin. Rhodes, Teels, and CSM playas (Fig. 1) are situated in a broad right-stepping transfer zone in the Walker Lane Belt referred to as the Mina deflection (Oldow, 1994). Here, right-lateral strike-slip strain is transferred from the southern to the central Walker Lane along a series of east-northeast-striking left-lateral faults. This structural configuration has led to the development of extensional, pullapart basins at all three playas, the associated faults in each of these basins show evidence of Quaternary, if not Holocene, displacement (Wesnousky, 2005). Northeast-striking Quaternary faults are spatially associated with geothermal activity in much of Nevada (Coolbaugh et al., 2002; Faulds et al., 2006); thus, from a regional perspective, Teels, Rhodes, and Columbus Marshes offer favorable settings for geothermal potential. 7

8 Figure 1. Red outline shows the area covered by the airborne hyperspectral survey, a.) shaded relief map of Great Basin, Nevada state border outlined in blue, Walker Lane Belt approximated by black lines b.) major faults of the Mina Deflection shown in black, arrows show sense of motion c.) mountains and relevant features in the study area: CSM Columbus Salt Marsh, EP Emigrant Peak, FLV Fish Lake Valley, EPFZ Emigrant Peak Fault Zone, RM Rhodes Marsh, TM Teels Marsh, EMFS Excelsior Mountain Fault System, BSF Benton Spring Fault (faults modified from Wesnousky, 2005). BLM geothermal leases were granted at Rhodes and Teels Marshes after the 2m temperature data were published in

9 Geothermal evidence Borate and sulfate mineral deposits are associated with geothermal activity around the world and Coolbaugh et al. (2006) discuss the relationship of Quaternary borate deposits to geothermal systems in Nevada. Borate salts were mined at CSM during the 1870s, and in fact, Teels, Rhodes, and CSM playas once led the world in borate production during that decade (Barker and Lefond, 1985). Many warm spring and hot spring waters have high concentrations of sulfur and boron; however, in some cases cold springs share this characteristic and where supported by favorable geothermometer temperatures may suggest the presence of a nearby geothermal reservoir. Kratt et al. (2009) used VNIR-SWIR hyperspectral data to map borate and sulfaterich evaporite crusts at CSM. Both gypsum (CaSO 4 H 2 O) and tincalconite (Na 2 B 4 O 7 5H 2 O) appear to have precipitated on the CSM playa by diffuse capillary evaporation and have not required flowing springs (Figure 2). There is, however, a cold spring adjacent to some of these deposits that yielded Mg-corrected Na-K-Ca and quartz geothermometers subsurface temperatures estimates of 137º C and 115º C, respectively (Coolbaugh et al., 2006b). The gypsum evaporite deposits are mostly found about 500m to the west of the spring and parallel to the toe of an alluvial fan for a distance of at least 1.7km. This same location is also found directly down the hydrologic gradient from the 2m temperature anomaly (Figure 2). 9

10 Figure 2. Two-meter-deep temperature survey and VNIR-SWIR hyperspectral mineral maps, T = tincalconite, 71-1 is explained in the next figure. Figure 3. Temperature gradient measured at

11 Stratigraphy Cambro-Ordivician formations of highly folded and faulted metasedimentary lithologies comprise basement rock in the study area (Albers and Stuart, 1972). Several highly faulted Miocene ash flow tuff units exhibiting various degrees of welding and overlie basement. Plio- Pleistocene basalts, with two intact vents are next in the sequence (Albers and Stuart, 1972). Alluvial fans and playa deposits make up the majority of study area deposits. Figure 4 illustrates the stratigraphic sequence with more detailed unit descriptions and provides a generalized geologic map. Figure 4. Green dots mark field sample locations discussed in the Results section, red outline shows boundary of hyperspectral TIR data collection, and generalized geologic map (modified from Albers and Stuart, 1972), all displayed on top of shaded relief. 11

12 Data sets and methods TIR remote sensing Background The TIR airborne hyperspectral data was collected on August 30, 2009 with the Spectrally Enhanced Broadband Array Spectrograph System (SEBASS). The Aerospace Corporation owns and operates SEBASS; the data were collected in collaboration with the Spectir Corp., Reno, NV. Spatial resolution was set to 5 m and the spectral resolution is 128 contiguous across the µm region. The detector elements are cooled to liquid helium temperatures, thus achieving a nominal signal-to-noise ratio of 2000:1 (Hackwell et al., 1996; Kirkland et al., 2002). Twenty-seven flightlines imaged more than 260 km 2 (Figure 5). The flightlines were between km in length and 550 m wide. Geo-registration files for each flightline were provided by the data vendor and registration accuracy was always within 40 m and commonly within 20 m. Two geo-registration files were found to be corrupt and these flightlines were not processed. Molecular vibrational motions occur within a crystal lattice at specific frequencies that are directly related to mineralogy (Christensen et al., 2000). Fundamental vibration modes occur within the TIR and show a systematic shift in wavelength with rock type (Hunt, 1980). Spectral mixing in the TIR is linear, therefore, a pixel s emission spectra is the sum of the endmembers weighted by their areal abundance (Ramsey and Christensen, 1998). Processing and analysis methods Aerospace Corp. used in-house processing algorithms to convert radiance to emissivity and surface temperatures data that were delivered to Desert Research Institute in late August of The process to convert radiance to emissivity is described in Kirkland et al. (2002). Analysis was performed on a subset of the data consisting of 82 channels between 8-12 µm, channels outside of the high and low end of this range were noisy due signal attenuation by atmospheric gasses. The signal in the µm region contains components of both reflected solar radiance and surface emitted radiance, therefore, spectral interpretation is ambiguous and we did not use this portion of the data set. Factors such as albedo, slope aspect, and thermal inertia confound the detection of potential geothermally influenced surface temperature anomalies, as discussed in Coolbaugh et al. (2007b); for these reasons we did not use analyze the surface kinetic temperature data that were also provided by the Aerospace Corp. ENVI image processing software was used in a rigorous statistically based multi-step process (Kruse et al., 1996 and 1999) to reduce the dimensionality of the 82 channel emissivity sub-set. Each flightline was individually processed using this procedure with the goal being to identify spectral endmembers. Some endmembers, not surprisingly, appeared in multiple flightlines, where this was the case a representative spectra with the strongest emissivity features were selected for each endmember. Arizona State University reference library spectra (Christiansen et 12

13 al., 2000) and the USGS spectral library (Clark et al., 2007) were then used to identify the endmember spectra, when possible. Once all of the flightlines were mosaicked, mineral maps were created for endmember spectra that were matched to reference spectra on the basis of absorption band positions and overall spectral shape. To create the mineral maps a mixture tuned matched-filter algorithm described by Boardman et al. (1995) was used. The matched-filter result consists of two images that are used to identify minerals within each hyperspectral image file. First, Matched Filter score (MF) is a grey scale image where brighter pixels more closely fit the input class spectra. Second, the Infeasiblility Image is a grey scale image that represents root-mean-square (RMS) error, where darker pixels correspond to low RMS errors, suggesting a more significant mineral identification. The two images are interactively linked in the display to visually identify pixels with high match and low error values that form optimum fits to input spectra. We used the MF and RMS images along with the Albers and Stuart (1972) geologic map, the image data spectra, and 1 m Nationwide Select data from ESRI World Imagery ( to select an MF threshold value, above which all pixels would be classified as the endmember mineral. Endmember spectra that did not provide clear matches to reference spectra were also mapped, along with selected library mineral spectra, using a similar method as described above. The difference in the mapping technique was that much more restrictive thresholds were used so that field samples would be collected from outcrops with spectra that closely matched respective endmember spectra. However, only sample locations, and not mineral maps are reported. In some cases, there were no image spectra that offered a high-confidence match to the mineral maps created with the reference library spectra. Hand samples were collected at 17 locations to either validate or determine the mineral composition of endmember spectra. Laboratory spectra were acquired with a Thermo/Nicolet Nexus 6700 FTIR spectrometer located at the University of Nevada, Reno and made available by Dr. Wendy Calvin. Spectra were measured on both powdered and chipped portions of the hand samples (Appendix I). 13

14 Figure 5. Spatial overlap of gravity survey, TIR survey, VNIR-SWIR survey and 2m temperature survey. 14

15 Gravity Survey The location of the gravity survey is encompassed by the yellow box in Figure 5. The goal of the survey was to explore the relationship of the 2 m anomaly to geologic structure. Observed gravity values were recorded at 260 stations on 300 m centers. The data were collected by a University of Nevada, Reno hydrogeology graduate student with a Lacoste and Romberg gravimeter combined with a Trimble Ranger ProXRT GPS receiving Omnistar real-time differentially corrected signal. The vertical accuracy of the GPS data was typically within 50 cm and often within 20 cm. Observed gravity values were also recorded at the Tonopah and Candelaria International Gravity Standardization Network stations during the beginning and end of each survey day. The precision of the gravity readings is ~0.05 mgal. Twenty-eight repeat readings at seven stations showed a mean repeat error of 0.11 mgal. Dr. Robert Karlin of the University of Nevada, Reno reduced the observed gravity data to a Complete Bouguer Anomaly (CBA) with a regional trend removed, in addition to creating data derivatives. The Geosoft program Oasis Montaj was used to process the data. Dr. Karlin has more than 30 years of experience as a geophysicist. The CBA corrects for instrumental drift, latitude, tidal effects, and local topography that collectively influence the observed gravity reading that is recorded in the field (Telford et al., 1991). After reduction to the CBA, CBA contour maps, and total horizontal gradient maps were created as data derivatives. Densely spaced contours of CBA values can elucidate steeply to moderately dipping fault planes or sharp lithologic contacts. The steepest horizontal gradient of a gravity anomaly will lie at the edges of the body. The total horizontal gradient (THG) of the CBA is very useful in delineating the margins of units with strong density contrast such as lithologic boundaries, faults, and fault zones. Gravity modeling was done using the GM-Sys module of the Oasis Montaj program from Geosoft Inc. Measured gravity models of unknown shape were forward modeled by trial and error adjustment of density and polygon vertices. Once the fit was considered close, XZ positions were optimized using inverse methods. The objective was to minimize the RMS error between observed and computed values. A fit was considered acceptable if the misfit M was less than 1% (M=100*RMS error/profile gravity data range). As described in the GM-Sys manual, 2-D models may be visualized as a number of tabular prisms with their axes perpendicular to the profile; blocks and surfaces are presumed to extend to infinity in the strike direction. 2¾-D modeling, as implemented in GM-SYS, allows the prisms to be truncated at some distance in the plus and minus strike directions (± Y). It also allows the strike direction to be skewed relative to the profile azimuth. The methods used to calculate the gravity and magnetic model response are based on the methods of Talwani et al., 1959, and Talwani and Heirtzler, 1964, and make use of the algorithms described in Won and Bevis, Two-and-a-half dimensional calculations are based on Rasmussen and Pedersen, The GM- SYS inversion routine utilizes a Marqardt inversion algorithm to linearize and invert the 15

16 calculations (Marqardt, 1963). Gravity and magnetics models are non-unique, i.e., several model families can be created to match the data. It is up to the interpreter to assess whether the model(s) are geologically reasonable. The principles of successful modeling are to create the simplest model with the fewest number of densities, blocks and vertices. Two-dimensional modeling takes the interpretation of the horizontal gradient a step further by fitting CBA values along-transect with a sub-surface depiction of depth-to-bedrock and/or juxtaposed lithologies. A value of 2.35 g/cm 3 was assumed for the background value in the 2D modeling; this value is a reasonable estimate for the density of unconsolidated alluvial deposits that blanket most of the gravity survey area. Three northsouth and two east-west 2D transects were modeled. Results are presented in the next section along with the other CBA derivatives. Results TIR remote sensing During initial inspection of the emissivity data provided by Aerospace Corporation spectral artifacts were identified in random flightlines. Figure 6 shows a color composite image created with RGB channels corresponding to 9.97 µm, 9.77 µm, and 9.46 µm. The flightlines that appear blue in this band combination exhibit sharp spectral artifacts between 9.2 and 10.2 µm, as illustrated by playa spectra in the upper right of Figure 6. The playa spectra shown in the upper left of Figure 6 is from the adjacent reddish-brown flightline that lacks the spectral artifacts. The artifacts did present some challenges with regard to the continuity of mineral maps, particularly for clay minerals. Therefore, mineral maps for most endmembers and most clay minerals were not created; however, field samples were collected for validation and example spectra are presented in figures that follow. Illite and montmorillonite are clay minerals that were commonly observed in the VNIR-SWIR hyperspectral data, especially on the playa. These minerals were also readily apparent in the emissivity spectra and Figure 7 compares representative image spectra with reference library spectra. 16

17 Figure 6. Figure 6. RGB display corresponding to channels 9.97 µm, 9.77 µm, and 9.46µm, respectively. Flightlines that appear blue displayed spectral artifacts between 9.2 and 10.2 µm (upper right), whereas flightlines that appear reddish brown lacked these artifacts (upper left). 17

18 Figure 7. Montmorillonite (blue) and illite (green) library spectra above compared with commonly observed montmorillonite/illite mixed image spectra below. Figure 8 compares reference spectra with image spectra for gypsum and tincalconite. The image spectra are averages from more than 50 pixels in each of two separate areas that were validated in 2009 for gypsum and tincalconite. Gypsum displays a strong absorption feature at 8.65 µm and was mapped with high confidence using the MF result that was in close agreement with the VNIR-SWIR mineral map shown in Figure 2. Tincalconite, on the other hand, displays several narrow absorption features in the reference spectra (Figure 8) but these features were only weakly resolved in the image spectra. One possible explanation for this lack of agreement is that borate minerals are highly soluble and any rain events in the weeks prior to the data acquisition could have dissolved the tincalconite crust. Another possible explanation is that grain size effects attenuated the absorption features in the tincalconite spectra. 18

19 Figure 8. Tincalconite library reference spectra (green) compared with mean image spectra (blue) in an area of high-purity tincalconite evaporite crust. The Palmetto Formation is partially composed of limestone units that crop out in much of the study area. The mineral calcite displays a strong absorption feature near µm (Figure 9) and the limestone units of the Palmetto were easily mapped with the MF result. Chert beds in the Palmetto Formation were strongly correlated with distinct quartz spectra. Appendix III provides example emissivity spectra for comparison with field sample laboratory spectra, along with sample descriptions, shown in Appendix I. Field sample locations are shown in Figures 4 & 10. Figure 10 also provides example mineral maps. Figure 9. Library reference spectra for calcite. 19

20 Figure 10. Example mineral maps. Kaolinite is below the resolution of the map scale but station e-roi-3 is a kaolinite match and is represented by field sample spectra shown in Appendix 1. Gravity survey Figure 10 shows the CBA and gravity stations overlain on the ESRI image data. The dark rocks in the north and south central part of the image are Quaternary basalts. The basalt outcrop in the northern part of the figure dips gently to the southeast and is bound on the north by a NE-striking normal fault. The basalt along the southern portion of the figure is bound to the north by the E- W-striking Coaldale fault that exhibits up to 2 km of left-lateral offset (Stockli et al., 2004). As would be expected, the basalt outcrops show a general correspondence to higher gravity values, especially in the southern portion of the figure. The high gravity values in the southwest corner of the figure overlaps with the eastern edge of the Black Horse mine where tungsten deposits are 20

21 hosted by siltstone of the Harkless Formation (Albers and Stuart, 1972). Light colored areas on the ESRI image data in the west-central portion of Figure 10 represent welded and non-welded outcrops of tuffaceous units and generally exhibit lower gravity values; however, the gravity pit in this area may be a result of greater basin fill thickness associated with a graben structure. The area in the central portion of the survey is covered by alluvial fan deposits that presumably thicken towards the basin s center to the east, hence lower gravity values on the eastern border of the survey. Figure 11. Color ramp applied to the CBA, white circles are gravity stations. The relationship of the CBA to geology is discussed in the text. The CBA is contoured in Figure 11 and shown in relationship to the 2 m temperature anomaly. The center of the 2m anomaly appears to be situated in a saddle between the major gravity highs and lows that were previously discussed. Immediately to the south of the 2m anomaly the horizontal gravity gradient is changing rapidly where the contours strike NE. Figure 12 compares the 2m anomaly with the true horizontal gradient. Figures presented in Appendix IV show the locations and results for the modeled 2D gravity cross-sections. The results generally suggest that depth to basement is shallow and that no major fault offsets are revealed. The steep horizontal gradient and gravity high immediately to the south of the 2m anomaly suggests at least a depositional and/or fault contact between basalt to the south and basin fill overlying tuffaceous deposits that host the 2m anomaly. The Line 10 cross-section shows an abrupt change in gravity between stations 3 and 4, thus providing evidence for a possible NE-striking fault that bounds the southern side of the 2m anomaly. 21

22 Figure 12. Contour of the CBA and relationship to the 2 m anomaly, contour interval is 0.2 mgal. Figure 13. CBA horizontal gradient and relationship to the 2m anomaly, red colors indicate steeper gradients, blue colors indicate gentle gradients. 22

23 Line 14, however, also transects the NE-striking gradient but shows a more gradual change from basalt into basin fill and tuffaceous deposits. Line 14 is therefore more consistent with a depositional contact. A separate source of funding is currently supporting structural mapping work on the west side of CSM. It is expected that the structural data will help with the interpretation of the gravity data, particularly for the major gravity low in the southwest portion of the survey. It is also expected that the gravity survey and structural data will aid in developing a conceptual structural model for the 2 m anomaly. Even without structural data to support the interpretation of the gravity survey one thing is still clear. The 2 m anomaly occupies a unique location within the gravity survey. It is still unclear if the 2m anomaly is a manifestation of lateral outflow from a more distant, uphill, upwelling zone, or, is the 2m anomaly situated directly above the upwelling zone? Shallow temperature measurements are most sensitive to where geothermal fluids are nearest to the surface. The gravity saddle that hosts the 2m anomaly could be a result of buried paleotopography that channels and forces lateral outflow from a more distal, uphill, source closer to the surface as it moves down the hydrologic gradient. There are at least three approaches that could be used to test further explore the question of lateral outflow or direct upwelling. A micro-gravity survey, costing a few tens of thousands of dollars, could potentially elucidate a gravity-high bullseye feature in the area of the 2m anomaly as a result of fluid mixing that over thousands of years would produce localized silicification or calcification of alluvial deposits. Similar in cost to micro-graivty, an electrical resistivity survey would be another method to test for lateral outflow, in that, a tongue of low resistivity extending away from the 2m anomaly and up the hydrologic gradient would not support a model for direct upwelling. The least expensive approach ($5-10k) would be to collect more 2m data up the hydrologic gradient where alluvium thickness decreases. 23

24 Publications / Presentations Kratt, C., Sladek, C., and Coolbaugh, M. F., Boom and bust with the latest 2m temperature surveys: Dead Horse Wells, Hawthorne Army Depot, Terraced Hills and other areas in Nevada. Geothermal Resources Council Transactions, v. 34, p Presented at the Geothermal Resources Council annual meeting on Oct in Sacremento, CA. The gravity data along with the shallow temperature survey data are publicly available on the Great Basin Center for Geothermal Energy website at: References Cited Albers, J. P., and Stewart, J. H., Geology and mineral deposits of Esmeralda County, Nevada. Nevada Bureau of Mines and Geology bulletin, 78, p. 80. Barker, J. S. and Lefond, J. S., Borates: economic geology and production. Proceedings of a symposium at the fall meeting of SME-AIME, Denver, Colorado, Oct. 24. Bradley, D. B., Stockli, D. F., Lee, J., and Winters, N. D., Constraints on the magnitude and rate of Pliocene to recent slip along the left-lateral Coaldale Fault, central Walker Lane Belt, Nevada. Abstracts With Programs Geological Society of America, v. 35, n. 6, p. 25. Boardman, J. W., Kruse, F. A., & Green, R. O. (1995). Mapping target signatures via partial unmixing of AVIRIS data. In: Proceedings of the Fifth JPL Airborne Earth Science Workshop, JPL Publication 95-1 (1), pp Christensen, P. R., Bandfield, J. L., Hamilton, V. E., Howeard, D. A., Lane, M. D., Piatek, J. L., Ruff, S. W. and Stefanov, W. L., A thermal emission spectral library of rock-forming minerals. Journal of Geophysical Research, v. 105, n. E4, p Clark, R.N., Swayze, G.A., Wise, R., Livo, E., Hoefen, T., Kokaly, R., Sutley, S.J., USGS digital spectral library splib06a: U.S. Geological Survey, Digital Data Series 231. Christiansen, P. R., Bandfield, J. L., Hamilton, V. E., Howard, D. A., Lane, M. D., Piatek, J. L., Coolbaugh, M. F., Sladek, C., Faulds, J. E., Zehner, R. E., and Oppliger, G. L., 2007a. Use of rapid temperature measurements at a 2-meter depth to augment deeper temperature gradient drilling. Thirty-Second Workshop on Geothermal Reservoir Engineering, Stanford University, Stanford, California, January 22-24, SGP-TR-183. Coolbaugh, M., Kratt, C., Fallacaro, A., Calvin, W., and Taranik, J. V., 2007b. Detection of geothermal anomalies using Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) thermal infrared images at Bradys Hot Springs, Nevada; USA. Remote Sensing of Environment, v. 106, p

25 Coolbaugh, M. F., Kratt, C., Sladek, C., Zehner, R. E., Shevenell, L., Quaternary borate deposits as a geothermal exploration tool in the Great Basin. Geothermal Resources Council Transactions, v. 30, p Coolbaugh, M. F., Taranik, J., Raines, G., Shevenell, L. A., Sawatzky, D. L., Bedell, R., Minor, T. B., A geothermal GIS for Nevada: defining regional controls and favorable exploration terrains for extensional geothermal systems. Geothermal Resources Council Transactions, v. 26, p Coolbaugh, M. F., Taranik, J. V., and Kruse, F. A., Mapping of surface geothermal anomalies at Steamboat Springs, NV. using NASA Thermal Infrared Multispectral Scanner (TIMS) and Advanced Visible and Infrared Imaging Spectrometer (AVIRIS) data; In Proceedings, 14th Thematic Conference, Applied Geologic Remote Sensing, Environmental Research Institute of Michigan (ERIM), Ann Arbor, MI., p Faulds, J. E., Characterizing structural controls of geothermal fields in the Northwestern Great Basin: a progress report. Geothermal Resources Council Transactions, v. 30, p Gm-Sys manual ( Hackwell, J.A., Warren, D.W., Bongiovi, R.P., Hansel, S.J., Hayhurst, T.L., Mabry, D.J.,Sivjee, M.G., and Skinner, J.W., LWIR/MWIR Imaging Hyperspectral Sensor for Airborne and Ground-based Remote Sensing. SPIE, 2819, pp Hunt, G. R., Electromagnetic radiation: the communication link in remote sensing. Ch. 2 in: Remote Sensing Geology, B. S. Siegal and A. R. Gillespie editors, John Wiley, New York, p Kirkland, L., Herr, K., Keim, E., Adams, P., Salisbury, J., Hackwell, J., and Treiman, A., First use of an airborne thermal infrared hyperspectral scanner for compositional mapping. Remote Sensing of Envrionment, v. 80, p Kratt, C., Calvin, W. M., and Coolbaugh M. F., Geothermal exploration with Hymap hyperspectral data at Brady-Desert Peak, Nevada. Remote Sensing of Environment, v. 104, p Kratt, C., Coolbaugh, M. F., Sladek, C., Zehner, R., Penfield, R., and Delwiche, B., A new gold pan for the west: discovering blind geothermal systems with shallow temperature surveys. Geothermal Resources Council Transactions, v. 32, p Kratt, C., Coolbaugh, M. F., Peppin, B., and Sladek, S., Identification of a new blind geothermal system with hyperspectral remote sensing and shallow temperature measurements at Columbus Salt Marsh, Esmeralda County, Nevada, Geothermal Resources Council Transactions, v. 33, p

26 Kratt, C., Calvin, W. M., and Coolbaugh, M. F., 2010a. Mineral mapping in the Pyramid Lake basin: Hydrothermal alteration, chemical precipitates and geothermal energy potential. Remote Sensing of Environment, v. 114, p , doi: /j.rse Kratt, C., Sladek, C., and Coolbaugh, M. F., 2010b. Boom and bust with the latest 2m temperature surveys: Dead Horse Wells, Hawthorne Army Depot, Terraced Hills and other areas in Nevada. Geothermal Resources Council Transactions, v. 34, p Kruse, F. A., Hungtington, J. F., & Green, R. O. (1996). Results from the 1995 AVIRIS Geology Group Shoot. In: Proceedings, 2nd International Airborne Remote Sensing Conference and Exhibition: Environmental Research Institute of Michigan (ERIM), (pp. I-211 I-220), Ann Arbor, MI. Kruse, F. A., Boardman, J. W., & Huntington, J. F. (1999). Fifteen Years of Hyperspectral Data: Northern Grapevine Mountains, Nevada. Proceedings of the 8th JPL Airborne Earth Science Workshop. JPL publication 99-17, pp Marqardt, D.W., 1963, An algorithm for least-squares estimation of non-linear parameters: J.SIAM, v.11: Oldow, J. S., Kohler, G., and Donelick, R. A., Late Cenozoic extensional transfer in the Walker Lane strike-slip belt, Nevada. Geology, v. 22, p Ramsey, M. S., and Christensen, P. R., Mineral abundance determination: quantitative deconvolution of thermal emission spectra. Journal of Geophyiscal Research, v. 103, n. B1, p Rasmussen, R., and Pedersen, L. B., 1979, End corrections in potential field modeling: Geophysical Prospecting, 27, Talwani, M., and Heirtzler, J. R., 1964, Computation of magnetic anomalies caused by twodimensional bodies of arbitrary shape, in Parks, G. A., Ed., Computers in the mineral industries, Part 1: Stanford Univ. Publ., Geological Sciences, 9, Talwani, M., Worzel, J. L., and Landisman, M., 1959, Rapid gravity computations for twodimensional bodies with application to the Mendocino submarine fracture zone: J. Geophys. Res., 64, Telford, W. M., Geldart, L. P., and Sheriff, R. E., Applied geophysics, second edtion. Cambridge University Press, New York, New York. Vaughan, R. G., Calvin, W. M., and Taranik, V., SEBASS hyperspectral thermal infrared data: surface emissivity measurement and mineral mapping. Remote Sensing of Environment, v. 85, p Wesnousky, S. G., Active faulting in the Walker Lane. Tectonics,v. 24, pp

27 Winters, N. D., Stockli, D. F., and Bradley, D. B., Geochemistry and geochronology of Neogene basalts along the Coaldale Fault, central Walker Lane Belt, Nevada. Geological Society of America Abstracts with Programs, v. 36, n. 4, p. 37. Won, I. J., and Bevis, M., 1987, Computing the gravitational and magnetic anomalies due to a polygon: Algorithms and Fortran subroutines: Geophysics, 52,

28 Appendix I % emissivity is shown on the y-axis and wavelength in micrometers are shown on the x-axis. Sample descriptions are presented at the end of this appendix. w-1-1 w-1-1 v2 w-1-1 v3 28

29 w-2-1 w-3-5 w-3-5 v2 29

30 w-3-5 v3 w-3-5 v4 w

31 w-7-5 v2 w-7-5 v3 w

32 w-8-1 w-8-3 w

33 w-8-4 v2 w

34 w-8-7 v2 e-1-1 e-1-1 v2 34

35 e-1-1 v3 e-4-4 pseudo in situ e-4-4 powder 35

36 e-4-5 e-4-5 v4 e-4-6 v6 36

37 e-roi-1 e-roi-1 v2 e-roi-2 37

38 e-roi-2 v2 e-roi-3 e-roi-6 v2 38

39 e-roi-7 e-roi-8 e-roi

40 e-roi-18-1 v2 Sample descriptions and spectral interpretations sample emissivity spectra sample description e-roi-1 opal, clay, minor feldspar whitish grey siltstone, hematitic stain on small fractures e-roi-2 mainly opal, some quartz and unknow at 8.5 silicified light grey-tan siltone and breciatted re-worked siltstone e-roi-3 75% qtz 25% sphene quartzite e-roi-6 high purity kaolinite white clay, breadcrumbs e-roi-7 finer grained calcite, sharp abs at 8 oragnish-red and dark grey clasts of limestone or limey siltstone e-roi-8 secondary calicite and unknowns orangish welded tuff, crystal rich, san and qtz w-8-7 qtz, feldspar, unknown abs at 8.19 greenish siltstone and lgight grey quartzite w-8-3 qtz, sphene, minor albite white quartzite w-7-6 geothite w/sphene and minor? quartzite with hematitic stain on weathered, light reddish grey on fresh w-8-1 qtz, sphene, minor feldspar, calcite xtl-rich welded AFT float, phenos <1 mm, reddish-orange, minor caliche e-1-1 alunite, qtz, minor calcite xtl-rich welded AFT, phenos < 1mm, reddish-orange weathered, white-grey fresh, minor caliche e-3-5 qtz chert w/secondary qtz veins, reddish-brown weathered, whitish fresh, e-4-4 montmorillonite w/qtz, minor calcite khaki playa clay e-18-1 mica w/minor qtz phyllite-schist, greenish light grey w/ minor hematitic stain, micaceous schist w-8-4 montmorillonite/illite/kaolin, some qtz white breadcrumbs, non-welded AFT, also partially silicified or vitric ash e-4-5 qtz, minor oligoclase and poss sphene basalt, chert and qtz vein float w % mont 40% kaolin white breadcrumbs, non-welded tuff w-1-1 qtz black chert w-2-1 qtz black-grey chert 40

41 Appendix II Sample SEBASS emissivity spectra for sample locations shown in Figure 9 and corresponding to field sample laboratory spectra shown in Appendix 1. 41

42 42

43 Appendix III Library reference spectra corresponding to endmember spectra identified in the SEBASS imagery. 43

44 44

45 45

46 Appendix IV Locations of 2D cross-sections presented in figures that follow. 46

47 47

48 48