Abstract

In order to evaluate the setting of the Humboldt–Rye Patch geothermal field, we carried out a program of hyperspectral and light detection and ranging (LiDAR) imaging of the Humboldt River basin to test (1) whether fault patterns, surface mineral alteration, and mud volcanoes in the Humboldt–Rye Patch district offer the potential for additional geothermal exploration sites; (2) whether mud diapirism in this region could be caused by seismic shaking; and (3) whether significant improvements in exploration can be made using these remote-sensing tools in addition to the more traditional techniques. In the southern (Rye Patch) region, a set of faults cuts the surface of the alluvial fans, and several faults cut shorelines of Lake Lahontan. These shorelines lie at an elevation of 1290 m, which corresponds with the elevation of the Lake 12,500 ± 500 yr ago. We find no signs of surface mineral alteration in the Rye Patch area in spite of the existence of these faults and known alteration at depth. Farther north, in the Humboldt House region, we find abundant evidence of alteration products, including siliceous sinter, carbonate, montmorillonite, hematite, and jarosite. This alteration is widespread, and corresponds to young faulting in only one location. The LiDAR data show at least two mud volcanoes and a large field of low-carbonate mounds. Some of these (apparently) diapiric features may have been associated with seismicity, and both active and paleoseismic events would have been sufficiently close and energetic to have initiated liquefaction in this region. Such liquefaction events would have been more likely, however, during the high stands of Lake Lahontan, when the ground would have been saturated, consistent with reported ages on rocks correlated with the carbonate mounds. We propose further geothermal exploration based on these results.

INTRODUCTION

Geothermal energy represents one of a number of “clean” energy sources, and development of such resources has taken on increased urgency as the effects of rapidly rising atmospheric greenhouse gas concentrations appear to be accelerating global warming (Solomon et al., 2007). Geothermal systems are associated with areas of magmatic activity, deep and permeable fault zones, vapor-dominated fields, geopressured fields, and areas of hot, dry rocks (Gupta and Roy, 2007). In addition to such high-grade geothermal resources, the recent Massachusetts Institute of Technology study on the future of geothermal energy suggested that enhanced geothermal resources (conduction-dominated systems) could grow to 100 GW by the year 2050 (Tester et al., 2006). Part of the growth of the total United States geothermal resource base would likely focus on Nevada, where high geothermal gradients dominate the Basin and Range province (Blackwell and Richards, 2004), and many young range-front faults provide potential access to deeply circulating, high-temperature fluids that could be mined for geothermal energy.

The first step in exploration for geothermal resources should be remote sensing. Remote sensing offers the capability of evaluating large areas in order to identify particular locations for more detailed study, including ground-based field work, geophysics, and drilling. Use of hyperspectral remote sensing has been shown to be very effective in locating surface outflow of hot fluids by mapping mineral alteration patterns. Examples of such studies in the Great Basin include Long Valley (Martini et al., 2003), Steamboat Springs (Vaughan et al., 2003, 2005), Dixie Meadows (Kennedy-Bowdoin et al., 2004), Humboldt Valley (MacKnight, 2005), Dixie Valley (Nash et al., 2004), Brady–Desert Peak (Kratt et al., 2006), Pyramid Lake (Faulds et al., 2003; Kratt et al., 2010), and Fish Lake Valley (Littlefield and Calvin, 2010).

Here we focus on a study of the Humboldt–Rye Patch geothermal district (Fig. 1), which has an existing but not producing power plant. We propose to test (1) whether fault patterns, surface mineral alteration, and mud volcanoes in the Humboldt–Rye Patch district offer the potential for additional geothermal exploration sites; (2) whether mud diapirism in this region could be caused by seismic shaking; and (3) whether significant improvements in exploration can be made using these remote-sensing tools in addition to the more traditional techniques. We approach these tests by integrating hyperspectral observations with detailed LiDAR topographic data to gain a much clearer view of the faults, fault timing, mineral alteration, and fluid venting in the Humboldt–Rye Patch district of the Humboldt basin. We utilize HyMAP hyperspectral imagery (MacKnight, 2005) with more recently acquired Optech LiDAR. We begin with an overview of the background information on the Humboldt–Rye Patch district and then discuss evidence for faults and timing, followed by evidence for fluid venting and mineral alteration. We speculate on a possible origin of the vent features, related to seismic shaking.

BACKGROUND

The Great Basin represents a zone of anomalously high heat flow (Blackwell and Richards, 2004), as a result of widespread middle Tertiary and younger volcanism and extensional faulting (Sonder and Jones, 1999 and references therein) and anomalously thin crust (Stauber and Boore, 1978; Priestly et al., 1982; Louie et al., 2004; Heimgartner et al., 2006). Coolbaugh et al. (2005a) pointed out that many geothermal systems in the Great Basin are associated with gold deposits younger than 7 Ma. They are also commonly associated with NE-trending belts of extension undergoing NW widening. Coolbaugh et al. (2005b) mapped the distributions of geothermal potential of the Great Basin, showing the locations of power plants, faults, and measured ground and subsurface temperatures of the region. Nevada's power plants and geothermal areas appear to lie on NE-trending zones (Faulds et al., 2005), and most are broadly encompassed by Walker Lane and the Humboldt shear zone.

Gritto et al. (2003) carried out a surface-to-borehole seismic experiment in the area of the Rye Patch geothermal field. They found a significant velocity increase at the base of the section of Tertiary sediments and volcanic rocks making up this part of the Humboldt River basin. The velocity transition corresponds to the depth where Tertiary materials are in contact with Triassic carbonate rocks with much higher seismic velocities, and in the area of the geothermal field, the carbonates lie at a depth of ∼700 m beneath the surface. At a depth of 880 m beneath the surface is a thin (60 m) clastic unit that acts as the reservoir for the geothermal field. Wells completed in this aquifer record temperatures of ∼200 °C. The seismic tomography of Gritto et al. (2003) indicated subsurface topography in the geothermal region, with a broad west-trending ridge in the shallow subsurface, possibly bounded on one or both sides by faults. The interpretation of east or ENE-striking faults is supported by results of a 3D seismic experiment that covers the Rye Patch geothermal anomaly (Feighner et al., 1999). The fault or faults are inferred to cut through the Triassic carbonates and clastic reservoir and continue into the overlying Tertiary sedimentary and volcanic rocks.

A detailed map of the region just west of the Humboldt Range by Davis (1983) delineated faults and linear shoreline features and deposits of Pleistocene Lake Lahontan. Davis also mapped a group of “travertine armored spring mounds” west of the Florida Canyon mine (Fig. 2) as part of the Wyemaha Formation, which he reported as intruding lacustrine and deltaic deposits of the Eetza Formation and underlying alluvial, deltaic, lacustrine, and shoreline deposits of the Sehoo Formation. Davis (1983) also reported three 14C ages on bone apatite and one tephra age, all taken from the Wyemaha Formation. These ages ranged between 23,000 and 30,000 yr. DePolo (2008) compiled a map of Quaternary faults in Nevada, and that map shows a set of faults along the west side of the Humboldt Range, considered to be younger than 15,000 yr. DePolo's map indicates faulting along the west base of the range and a set of NE-striking faults just to the west.

The northern part of the Humboldt Range consists largely of Triassic metasedimentary rocks, including limestones and dolomites of the Prida and Natchez Pass formations, and argillite and quartzite of the Grass Valley Formation. Triassic rhyolite is common within these units. An outcrop of Quaternary basalt occurs near the range front (Johnson, 1977).

The Humboldt House geothermal reservoir is one of two geothermal reservoirs located in the Humboldt–Rye Patch geothermal district. Exploration and development of geothermal prospects in this area began in the late 1970s by the Phillips Petroleum Company; currently Presco Energy LLC owns the geothermal leases in this area. Initial recognition of a possible geothermal reservoir at the Humboldt House came from the presence of surface siliceous sinter (opal and chalcedony), which may imply the presence of hot spring deposits, and from measurement of elevated temperatures in this region (Waibel et al., 2003). The Rye Patch geothermal area is sited close to one of the NE-striking faults, the Rye Patch fault. Phillips Petroleum found temperatures of 163 °C in a test well in 1977, and later wells found values up to 243 °C (http://www.nbmg.unr.edu/Geothermal/index.html). A power plant was constructed on this site but has not produced electricity. The Great Basin Center for Geothermal Energy (University of Nevada, Reno), Presco Energy, and Apollo Gold drilled a set of wells in the Rye Patch and Florida Canyon mine areas (labeled “D” in Figs. 3 and 4), and this exploration discovered higher temperatures in the latter region than at similar depths near the power plant (Johnson, 2005). Waibel et al. (2003) also reported temperatures of 350 °C to 400 °C at depths of 650 m.

METHODS

We used the HyMap airborne hyperspectral imager for this study that is flown in light, twin-engine aircraft at altitudes of 2000–5000 m above ground level (AGL). The HyMap sensor has 126 contiguous spectral bands spanning the visible and near infrared to the short-wave infrared (VNIR-SWIR, 0.46–2.5 μm) with an average bandwidth and spectral sampling interval of 15 nm. The pixel size varies between 3 and 5 m depending on flight and target elevation. The spectral and spatial resolution of HyMap allows for accurate mineral identification and mapping at scales appropriate for geothermal exploration. HyMap has been used previously to map alteration minerals associated with fluid flow along fault zones in tectonically active areas in the United States (Crowley and Zimbleman, 1997; Kruse, 2000, 2011; Martini, 2002; Martini et al, 2003; Kennedy-Bowdoin et al, 2003).

The HyMap data were acquired on June 1 and 2, 2003, over the Humboldt–Rye Patch field area (Fig. 2). The acquisition consisted of 19 north-south flight lines, which covered an area of ∼500 km2 and spanned the latitudes of 40°40′ to 40°28′N and longitudes of 118°08′ to 118°28′W. HyVista Corporation did the atmospheric correction of the raw spectral data. We processed the atmospherically corrected data using the ITT commercial software package ENVI. This software allows the user to discriminate and map the distribution of mineralogy within spectral images using a variety of different algorithms and is designed to focus spectrally on the most pure end members for identification and mapping.

On September 23, 2004, we collected 800 km of Optech ALTM 3100 LiDAR data, flown by Sky Research, Inc., from Ashland, Oregon. The LiDAR was flown on a Cessna Caravan aircraft. The plane flew at 1000 m AGL taking 100,000 measurements per second. Scan overlap was 30% with a frequency of 60.1 Hz. Spot spacing was 45 cm; vertical and horizontal accuracies were 15 cm and 50 cm, respectively. Data preprocessing filtered noise and assigned geographic coordinates to each point, based on differential global positioning system (DGPS) base station data and an inertial measurement unit to determine aircraft attitude. A triangular irregular network (TIN) was generated based on initial evaluation of the data, and extraneous points were visually removed. A final bare-earth model stripped most of the vegetation, and digital elevation models (DEMs) were created from this TIN.

We made three field excursions to the Humboldt–Rye Patch field area to obtain field spectra as a ground truth for the mineral mapping, to field check the fault maps, and to obtain geologic field observations. We checked mineral maps generated using an advanced spectral device (ASD) Field Spec Pro and a GPS unit for location. The ASD Field Spec Pro is a handheld spectrometer with 2150 bands spanning a spectral range of 0.35 to 2.5 μm. The instrument was calibrated before each use and after approximately three to five spectral collections. Calibration from radiance to reflectance was done using the reflectance from a diffusely reflecting white plate and a dark current correction. Variation in light reflecting from the white plate is assumed to be from atmospheric influence and is therefore subtracted from the data. Dark current correction works similarly but subtracts any interference in the signal from the instrument, such as changes in the temperature of the instrument. Field spectra were collected using an 8° lens at a distance of 0.3–1 m. We created a spectral library using the ENVI software package, and then identified each spectrum by comparison with the U.S. Geological Survey laboratory spectra. The resultant mineral identifications were used to assess the accuracy of the mineral maps.

RESULTS

Fault Systems

The west side of the Humboldt range is marked by a set of faults, considered by dePolo (2008) to be younger than 15,000 yr (Figs. 3 and 4). The locations of faults that we have mapped with LiDAR are very similar to the locations of Davis (1983) and dePolo (2008). What we can add using LiDAR is the relationship between these faults and the Lake Lahontan shorelines. The base of the mountains shows a sharp break, indicating range-front faulting (Figs. 3 and 4). This fault, named the Humboldt Mountain fault, can be followed essentially continuously northward through the Florida Canyon mine and to the north of the mine. A very prominent fault cuts irregularly across the alluvial plain between the mountains and the reservoir, which we name the Standard mine fault, because the fault is on line with that mine (Fig. 3). The fault is mapped in three segments, and its outcrop pattern is that of a normal fault dipping to the NW. Northwest of the Standard mine fault is the Rye Patch fault (Fig. 3), which we map in three segments. The northernmost segment is irregular, with sharp escarpments separating highly dissected older alluvium surfaces on the east from much smoother, younger alluvial fans to the west (Fig. 3). Our mapping of the Rye Patch fault location is the same as that of Davis (1983), who used air-photo coverage.

Largely to the west of the Standard mine and Rye Patch faults lies a series of curvilinear features that we interpret as shorelines from ancient Lake Lahontan (Fig. 3). These inferred shorelines lie precisely along topographic contours, whereas the mapped faults do not (Fig. 5). The southern part of the Rye Patch fault passes across a small alluvial fan and displaces its surface, with the northwest side down (Fig. 5). The fault also cuts shorelines from Lake Lahontan. The shorelines are in deposits mapped as Holocene by Davis (1983), and the fault separates these Holocene deposits from late Pleistocene Sehoo Formation in the south and from pre-Lahontan sediments toward the northeast. The elevation of the Rye Patch fault as it cuts the small alluvial fan and Lake Lahontan shorelines is 1290 m (Fig. 5). The lake was at this elevation 12,500 ± 500 yr ago, according to the work of Thompson et al. (1986), Lin et al. (1996), and Adams and Wesnousky (1998).

The southwest end of the Rye Patch fault displaces the ground surface ∼1 m (Fig. 5), and appears to die out farther to the southwest. To the northeast the fault scarp rises to a height of 5–6 m (Fig. 6). The scarp is not seen across a small drainage and then rises to 3–5 m height. After a larger gap of a few hundred meters, the fault is again seen crossing an older alluvial fan surface, but its height is diminished to about a meter (just to the east of the Rye Patch power plant; Fig. 3).

The Rec Area scarp, named because of its proximity to the Rye Patch recreational area, is a long, straight escarpment over 2 km long (Fig. 3). The scarp has a relief of up to 5 m, and contours follow it closely (Fig. 7). It can be traced northward to a zone of complexity in the Lahontan shorelines, and is not distinguishable from the other shorelines. We interpret this feature as a beach ridge, and numerous borrow pits along the scarp suggest local mining for sand. The scarp is cut by an alluvial stream channel (Fig. 7). The Standard mine fault scarp intersects the Rec Area scarp, but it does so in a region that has been recently worked by earth-moving efforts, so the nature of the intersection is unclear (Fig. 7). In addition, the Standard mine fault scarp is quite subdued in this area.

The Standard mine fault zone is composed of three segments that displace the surface of older alluvium. The southwestern segment displaces the surface about a meter, near where it intersects the Rec Area scarp. The central segment of the fault has a scarp height of 3–5 m, and the northeastern extent of the fault has relief of 5–8 m. Directly north of the latter segment, the Humboldt mountain fault has variable scarp height, ranging up to 12 m in relief (Fig. 3). This segment of the fault has been trenched (location T, Fig. 3) by Wesnousky et al. (2005). The vertical separation at the trench site was 3 m, and displaced charcoal found in the trench at the fault was dated at 4626 ± 181 calendar years (cal yr) B.P. Using the scarp analysis method of Hanks (2000), Wesnousky et al. (2005) concluded that the penultimate event that dominates the trench site occurred ∼35 k.y. ago, and was responsible for most of the displacement (2.7 m) of the scarp. Significant variability of scarp height is seen on the Humboldt Mountain fault, the Standard mine fault, and the Rye Patch fault. Such variability may be due in part to variable displacement history along the fault segments. However, some variation is due to the differences in erosion and burial as a result of older and younger alluvial processes.

The Humboldt Mountain fault cannot be traced through the Florida Canyon mine by LiDAR because of all of the earth-moving activity within the mine site, but the fault is clearly visible within the mine, where highly altered rocks are juxtaposed against unaltered alluvial fan deposits along the western pit walls (MacKnight, 2005). Unaltered alluvial deposits on the hanging wall of the range-front fault are evidence against recent high-temperature fluid flow to the surface along the fault. Fracturing is intense in the pits along the range-front fault but decreases away from the pits.

North of the Florida Canyon mine, we see a continuation of the mountain-front faulting (Fig. 4), and one example of faulting that cuts the alluvial fan surface (Fig. 8). In this region the Lake Lahontan shorelines are clearly delineated, and the fault cuts at an angle through them. While the fault can be seen to cut the older alluvial fans, it does not cut through the younger fans, which appear to have the peaks of their cones originating just above the top of the fault scarp. The older fans also originate at the Humboldt Mountain fault scarp.

Recent geodetic information based on permanent GPS stations of the Basin and Range Geodetic Network (BARGEN) indicate that western Nevada moves northwest (∼295°) with respect to a stable North America reference frame (Niemi et al., 2004). The closest site to the Rye Patch geothermal district is station TUNG, located 7 km to the south of the southern edge of Figure 3 (Hammond and Thatcher, 2005). The dashed arrow in Figure 3 shows this direction, and indicates that the Rye Patch and Standard Mountain faults are oriented at high angles to that direction of movement.

Mineralization and Mud Volcanoes

Mineral mapping in the Humboldt House area was focused over a 4 km × 8 km area located northwest of the Florida Canyon mine and east of the Humboldt River (Fig. 9). Farther to the south, in the Rye Patch area, illite and chlorite dominate the clay mineralogy mapped with the HyMap data, and these are likely derived by drainage from the adjacent mountains. Five mineral end members derived from analysis of the HyMap data in the Humboldt House region are sinter (opal and chalcedony), calcite, montmorillonite, gypsum, and iron oxides (hematite and jarosite) (Fig. 9). The distribution of these minerals is mapped over the Humboldt–Rye Patch area, and specific areas of interest were then targeted for field investigation. Most of the alteration minerals (sinter, calcite, montmorillonite, gypsum, and iron oxides) are found in abundance in the region of dune fields west of the Florida Canyon mine. Minerals such as illite and chlorite are found throughout the area, but these may indicate erosional products from the mountains. Drilling results (Johnson, 2005) have shown a change from smectite to illite at depth, indicating temperatures between 100 °C and 150 °C (Inoue et al., 2004).

Using the LiDAR data we have identified several possible hydrothermal features: two mud volcanoes and a field of low mounds less than 1 m tall (Fig. 9). These rocks were mapped as part of the Wyemaha Formation by Davis (1983), which he described as “travertine armored spring mounds.” Figure 10 shows a closeup of a field of dunes with a clearly identifiable mud volcano and a field of small mounds. The mud volcano is located 1.4 km northwest of the Florida Canyon mine and 1 km southwest of the Humboldt House. This feature is ∼25 m at the base and 4 m tall (Fig. 10). At the peak of the outcrop is a crater 3 m in diameter and 1.5 m deep. Along the eastern flank is a small tunnel that was dug at the base of the outcrop, in which we identified gypsum, jarosite, and hematite using the ASD spectrometer. Compositionally this outcrop grades from bulk sinter sheets at the base, into a sinter-travertine mixed bedding at the peak (MacKnight, 2005). The transition occurs a foot below the peak of the outcrop. Extending from the northern flank of the outcrop is an elongated deposit of white sinter that is oriented N-S, and its source appears to have been from the base of the outcrop. A set of small conjugate faults was found in the tunnel. The major fault identified in the tunnel is synthetic to the range-front fault, striking N-NE and dipping 60°–65° west. Minor offset (2–4 cm) along this fault is documented using the sinter bedding.

Thin sections taken from samples at the base and upper part of the volcano flanks (Figs. 11 and 12) show it to be composed of carbonate-cemented mudstone and sandstone. The photomicrograph in Figure 12 shows fine, arcuate banding and carbonate veins, implying flow structures, which would be consistent with mud diapirism. We also field checked the region of small mudrock mounds (Figs. 13 and 14) and found all of these small features to contain abundant carbonate. The section in Figure 13 shows small mudstone pellets surrounded by carbonate cement, and carbonate veins filling cracks in the rock. These features are also layered (Fig. 13), and a section from a different mound in this set showed the same structure.

We discovered a second possible mud volcano (MV2) in the LiDAR data just to the NW of the zone of small carbonate mounds (Fig. 15 top). We field checked this finding and discovered that the proposed mud volcano was indeed a mound ∼2 m tall, with flanking structures similar to that of MV1 and a small central crater (also visible on the LiDAR image). The rocks making up this mound are carbonate rich as well, and appear to have the same origin as that of MV1 and the small mounds. Davis's (1983) depiction of this set of rocks as travertine-armored spring mounds implies that he felt they were formed by fluid (springs) rising through the surface and injecting into the Eetza Formation. Our discovery of at least two of these mounds with central craters (MV1 and MV2), and macroscopic and microscopic evidence for soft sediment flowage (Fig. 12), supports our suggestion that these features are mud volcanoes with carbonate cement. The latter explains why these features are lithified, although they are only 23,000–30,000 yr old.

A puzzling aspect of this region is that a relatively small part of the Wyemaha Formation is identified as carbonate in the hyperspectral data, even though the rock exposed at the surface contains significant amounts of carbonate. We suggest that the weak carbonate signal in the airborne data is due to two potential causes. The first is that much of the rock is not carbonate; only the veins and pore filling are carbonate cement. A second reason may be that regional dust and surface alteration blanket the rocks, filtering the spectral signature in flown data.

On the southern part of Figure 9 there are two parallel north-trending ridges, and the eastern ridge shows displacement of beds in the field, indicating faulting (Figs. 9 and 16). The eastern ridge is associated with abundant sinter and some gypsum deposits. A very subdued trace of the west side of the easternmost ridge continues to the south, cutting across the contours (Fig. 16). Davis (1983) mapped this short segment as a fault but mapped the ridge as dune deposit. We examined the ridge and found rock outcrops exposed on its west side, suggesting to us that the fault mapped by Davis (1983) actually continues northward and may be responsible for the formation of this ridge. The western ridge may be a dune deposit. As the feature that Davis (1983) mapped as a fault continues to the south, it becomes exactly parallel to contours and thus indistinguishable from a shoreline feature.

DISCUSSION

Do the fault patterns and distribution of mineral alteration in the Humboldt–Rye Patch district suggest the potential for additional exploration sites? We feel that the answer to this question is yes, and the clearly mapped faults in the Rye Patch area are closely associated with the Rye Patch geothermal system. The Rye Patch fault, Standard mine fault, and the range-front faults all are oriented at high angles to the direction of movement indicated by regional GPS studies (Fig. 3), suggesting maximum rates of opening. Fluids circulating through these faults at depth could be heated to high temperature because of the high geothermal gradients in this region and evidence for mineral alteration at depth in the cores. Exploration for high-temperature fluids might well focus just to the west of the Standard Mountain and Rye Patch faults nearer to the range front, where these faults show the greatest vertical displacement. In addition, the intersections of these faults with the Humboldt Mountain fault would be expected to show abnormally high fluid conductivity because of the complex fracturing of such intersections.

Is there a relationship between the mineral alteration and the faulting that we see in the Humboldt River valley? At first glance the latter relationship is weak. The faults in the Rye Patch area, several of which cut the young alluvial deposits and Lake Lahontan shorelines, do not show signs of surface mineral alteration associated with them, although significant alteration was discovered at depth by drilling (Johnson, 2005). In the region to the west of the Florida Canyon mine, where we find abundant evidence of sinter, calcite, montmorillonite, gypsum, and hematite alteration products, their distributions are generally not associated with faults. We interpret one scarp associated with sinter and carbonate as a fault (Fig. 8), but it has also been interpreted by Davis (1983) as a dune structure. We note that the mud volcanoes and small outcrops of sinter and calcite-cemented sands contain abundant alteration products, suggesting that they may be the source of the altered minerals. Some of these altered minerals were transported by wind and runoff to attain a widespread distribution throughout the region.

A recent drilling program (sites labeled “D” in Fig. 4) undertaken near Humboldt House in May 2003, revealed silicified sediments at depth. One of the mud volcanoes (MV1) shows evidence of being vertically-built hot spring deposits, based on its layering and mineralogy, indicating that hydrothermal fluids flowed up through a vent and deposited material along the flanks (MacKnight, 2005). We didn't carry out detailed studies of the other mud volcano nor did we bore into the mounds. However, their surface structures and compositions are similar to that of MV1.

How were the fluids brought to the surface? One possibility is by earthquake shaking, destabilizing buried sand and mud under elevated fluid pressure to produce the mounds and mud volcanoes. The requisite size of an earthquake needed to produce liquefaction at a given location can be estimated by the empirical relationships of Galli (2000), who studied 317 instances of liquefaction features in Italy and matched them with historical and modern earthquakes. Galli found relationships of the form: 
graphic
where Io is earthquake intensity at the epicenter, Re is distance from epicenter to the given location, and a and b are constants that differ with the time clustering of the earthquakes. When all earthquakes from 1117 to 1990 are considered, their relationship is: 
graphic

When we apply this relationship to a list of all major earthquakes in Nevada since 1900, we find that one—the 1915 Pleasant Valley earthquake (Jones, 1915)—falls within the region (∼60 km away) in which liquefaction of material near Humboldt House could have occurred. This earthquake had an estimated magnitude of 7. In addition, trenching by Wesnousky et al. (2005) discovered evidence for 2.7 m of vertical displacement ∼35 k.y. ago, and such displacement would be consistent with a magnitude 7+ earthquake. Thus, recent events as well as late Pleistocene events would have been able to generate such liquefaction.

Several more recent studies have been reported by Wang et al. (2005) and by Manga and Brodsky (2006), who showed a relation between the maximum hypocentral distance (Rmax) and earthquake magnitude (M) for liquefaction: M = –5.0 + 2.26 log Rmax. Using this relation implies that mud volcanism could occur in the Humboldt Valley area from earthquakes as far as 200 km away, if they were M 7 or above. In addition, the numerous faults we have mapped might be capable of generating sufficient levels of shaking to generate the mud volcanoes with earthquake magnitudes as low as 5.0, assuming other factors are favorable, such as pore pressures and liquefaction susceptibility (Manga and Brodsky, 2006). Thus, as long as the conditions are appropriate for liquefaction, this region has sufficient seismicity to generate it.

Some fluid vents, such as mud volcano MV1 (Fig. 10), have gone through multiple injections and have developed a compositional layering from sinter to travertine to hematite and/or jarosite. This sequence suggests a longer lifetime than single events. If seismicity played a role in fluidization of materials, it would likely have been enhanced during late Pleistocene high stands of Lake Lahontan, allowing the ground to fully saturate. Because the cores show increasing temperature gradients approaching the range front (Johnson, 2005), exploration would be best focused there in the Humboldt House district.

Finally, can geothermal exploration be improved significantly by the use of remote sensing? A number of previous studies have shown a close relationship between mineral alteration and faulting, and successful geothermal sites in Long Valley, California; Dixie Valley, Nevada; and Brady–Desert Peak, Nevada, are associated with high-temperature mineralization along fault zones. Such regions have been mapped with hyperspectral remote-sensing tools. In the Humboldt–Rye Patch district, such relationships are not clearly presented, but intersecting fault systems mapped by LiDAR and earlier with air photos (Davis, 1983) offer distinct exploration possibilities. What our study provided beyond that of Davis (1983) is the LiDAR evidence showing that the Rye Patch fault cuts shorelines from ancient Lake Lahontan at elevations consistent with an age of ∼12,000 yr, providing some control on fault timing. The fault also cuts a small alluvial fan, which in turn has buried the 12,000-yr-old shorelines.

Previous work by Davis (1983) had mapped the distribution of “travertine armored spring mounds” as part of the Wyemaha Formation, which he dated at 23–30 k.y. Our LiDAR work showed clear locations of at least two mud volcanoes within the Wyemaha, and a set of small carbonate mounds having similar structure to that of the mud volcanoes. These structures and associated siliceous sinter deposits suggest the outflow of fluids sourced at some depth below the surface. However, based on the results of boreholes in this region, geothermal gradients increase toward the mountain front, suggesting that the better exploration sites for geothermal fields would be closer to the mountain front, possibly associated with either the Humboldt Mountain fault or one of several small faults just to the west of it.

CONCLUSIONS

We have mapped four recent faults in the Humboldt–Rye Patch geothermal region, using LiDAR imaging. These faults include the Humboldt Mountain fault, the Rye Patch fault, the Standard Mountain fault (new name), and a small fault to the northwest of the Florida Canyon mine. Most of these faults were known from previous work (Davis, 1983; dePolo, 2008). However, the LiDAR allowed us to see the detailed relationships between their escarpments and latest Pleistocene shorelines from Lake Lahontan, whose ages (∼12,000 yr) are reported by previous studies. One of the faults, the Rye Patch fault, also cut a small alluvial fan, which in turn had buried these shorelines. These faults are not associated with clear signs of surface mineral alteration, but our hyperspectral data showed widespread alteration in the Humboldt House district, including siliceous sinter, carbonate, montmorillonite, hematite, chlorite, and illite. Also in the Humboldt House district we mapped two mud volcanoes and a set of low mounds that are lithified with carbonate cement. We interpret these as intrusion features, possibly triggered by nearby seismic events, similar to events known to have occurred in the past century as well as from paleoseismic studies on the Humboldt Mountain fault.

In this paper we have explored whether faults, mineral alteration, and evidence of fluid venting offer additional geothermal exploration sites, whether signs of fluid venting can be explained by earthquake shaking, and whether use of remote-sensing tools improves geothermal exploration. The first question is answered affirmatively, and we suggest that additional exploration sites should be associated with intersections of NE-striking faults and the Humboldt Mountain fault. The locations of fluid vents may not be ideal sites for exploration, but sites closer to the mountain front are preferred, based on results of regional measurements of thermal gradients. For the second question, we have shown that this region receives sufficient seismic shaking from local and more regional seismicity to trigger the eruption of mud diapirism. Finally, much evidence already exists to show the utility of remote sensing in geothermal exploration. Here we have added LiDAR to demonstrate its great usefulness in mapping relationships between geomorphic features (faults, ancient shorelines, and small alluvial fans) and in recognizing mud volcanoes and small carbonate mounds. We recommend that such tools be used as the first line of exploration to narrow down potential sites for more detailed, on-site studies.

We are grateful for the excellent comments of two anonymous reviewers and the associate editor. We thank Ty Kennedy Bowdoin and Martha Jordan for field assistance, Sky Research for acquisition and processing of the LiDAR data, and Brigette Martini and Richard Ellis for very helpful discussions. Funding for data acquisition was provided by a Department of Energy grant (W-7405-Eng-48) to Bill Pickles, in whose memory we dedicate this paper. Funding for student and travel support was made from Lawrence Livermore National Laboratory to University of California, Santa Cruz (E. Silver), on an inter-university transfer (B52628).