Detecting natural fractures with ground penetrating radar and airborne night-thermal infrared imagery around Old Faithful Geyser, Yellowstone National Park, USA
Old Faithful Geyser is located in the Upper Geyser Basin (UGB) of Yellowstone National Park (Fig. 1, Fig. 2). Many of the approximately 4 million visitors to Yellowstone National Park (https://irma.nps.gov/Stats/Reports/Park/YELL, accessed 5/8/2019) travel to the Old Faithful area and watch the eruption of the iconic Old Faithful Geyser. The impact of visitation and subsequent infrastructure development within the UGB hydrothermal system prompted the National Park Service (NPS) to convene a panel of experts to discuss the shallow subsurface geology and hydrology of the Old Faithful area. The report resulting from this scientific panel, Hydrogeology of the Old Faithful area, Yellowstone National Park, Wyoming, and its relevance to natural resources and infrastructure, (Old Faithful Science Review Panel, 2014), recommended that Ground Penetrating Radar (GPR) be used to study the Old Faithful area. This paper presents results of a 2015 GPR study near the cone of Old Faithful Geyser that detected “blind” fractures in siliceous sinter and compares the GPR technique to TIR imagery and mapped surface fractures.
Geologic and hydrologic setting of the Old Faithful area
Within Yellowstone National Park, the Upper Geyser Basin (UGB) hydrothermal area (Fig. 3), including Old Faithful Geyser, occurs in a topographically low basin. This topographically low basin (gray area in Fig. 3), contains obsidian-rich sands, siliceous sinter and the Firehole River flowing north-northwest through the UGB hydrothermal area. Topographically high volcanic plateaus surround the UGB hydrothermal area and probably provide cold water recharge for the UGB hydrothermal system. As shown in Fig. 3, the volcanic plateaus consist of the following rhyolite lava flows (Christiansen, 2001; Christiansen and Blank, 1974): the Mallard Lake (Qpm) along the east side of the basin, the Scaup Lake (Qpul) on the southeast, the Spring Creek (Qpcs) on the southwest and the West Yellowstone (Qpcy) on the northwest. The Biscuit Basin rhyolites (Qpub) occur near Biscuit Basin and within the UGB (Muffler et al., 1982; Bindeman and Valley, 2000; Christiansen, 2001).
Fractured bedrock, obsidian-rich sediments and layered siliceous sinter permit an intricate upflow of hydrothermal fluids from subsurface volcanic reservoirs, through overlying sediment and onto the ground surface (see Fig. 2 in Old Faithful Science Review Panel, 2014). During 1929–1930, Fenner (1936) drilled a hole in the UGB to a depth of 129 m and was the first to explore the subsurface stratigraphy. Based on U.S. Geological Survey (USGS) boreholes in the UGB, White et al. (1975) recognized that there can be as much as 65 m of sediment overlying rhyolite lava flows. Obsidian-rich sediments (Waldrop, 1975) and fractured sinter may form shallow aquifers while underlying rhyolite lava flows form deeper aquifers (White et al., 1975; Keith et al., 1978). Bottoms of rhyolite lava flows, brecciated fronts of lava flows or fractured tuffs are permeable and allow lateral and vertical circulation of meteoric water (Jaworowski et al., 2016). Hydrographs show a low peak-flow and high base-flow in these fractured volcanic rocks, indicating a highly permeable aquifer that may source local cold-water recharge for the UGB (Gardner et al., 2009).
Fournier et al. (1994) describe hydrothermal fluid flow pathways in the UGB as follows: (1) hydrothermal water flows from a 270 °C reservoir that may or may not boil before reaching the ground surface; (2) hydrothermal water may move from the 270 °C reservoir, boil, enter a 215 °C reservoir with more boiling, flow through a rhyolite lava flow and discharge at the surface; or (3) hydrothermal water may move from the 270 °C reservoir, enter a 205 °C reservoir without boiling or losing steam, flow through sediments and discharge at the ground surface.
An overview of fractures in the Old Faithful area
Fracture trends are evident on the faulted Mallard Lake resurgent dome and on the surrounding volcanic plateaus (Fig. 2, Fig. 3). The trends of mapped faults on the Mallard Lake resurgent dome include the following: northwest (NW), north (N), and east-west (E-W). West of UGB hydrothermal area, a north-northwest (NNW) trend of volcanic vents occurs on the Madison Plateau (Christiansen, 2001). This NNW trend of volcanic vents is approximately parallel to the NNW trend of the Firehole River and the western boundary of the 640 ka Yellowstone caldera (Fig. 2).
Fracture trends in the UGB hydrothermal area have been documented by airborne remote sensing since the 1960’s. In the Old Faithful area, White (1969) recognized thermal infrared (TIR) anomalies and “a linear belt” on September 1967 night-TIR (8−12 micron) imagery. Marler and White (1974) noted that many hot springs and geysers align with regional trends, such as NW, NE, and N. Muffler et al. (1982) mapped a few fractures on and near the Old Faithful geyser cone. USGS researchers using 1998 low-altitude airborne visible and infrared imaging spectrometer (AVIRIS) data documented a NW linear trend of the UGB sinter deposits (Livo et al., 2007). Livo et al. (2007) utilized the 0.4–2.5 micrometer visible and near-infrared (NIR) spectroscopic data (6-m and 20-m spatial resolution) to identify and map hydrothermally altered rock (kaolinite and other clays) and hot spring deposits (siliceous sinter and/or cristobalite) in YNP. Using 8–12 micron, airborne night-TIR imagery for 2007–2009, Jaworowski et al. (2010, 2012) recognized TIR linear trends (NW and NE) in the UGB hydrothermal area.
A study of the ground temperature patterns surrounding Old Faithful Geyser began during a Yellowstone National Park-sponsored September 2007 airborne, night-TIR acquisition. Calibrated night-TIR temperature mosaics (Neale et al., 2016) show similar spatial patterns from September 2007 through March 2015. Fig. 4a and b show the spatial patterns captured during night-TIR September 2010 and 2011 image acquisitions, respectively: Fig. 4a clearly shows TIR trends whereas Fig. 4b shows substantial hydrothermal discharge from Old Faithful Geyser. The most prominent “hot area” near Old Faithful Geyser is a former geyser cone northeast of Old Faithful Geyser (area B in Figs. 4a and b, and 5 ). This former geyser cone has higher radiative ground temperatures than Old Faithful Geyser. During winter months, hydrothermal activity on this former geyser cone is evident on the brown-colored, warm ground that does not accumulate snow.
March 2012 and March 2015 night-TIR image acquisitions are spring image acquisitions prior to the April 2015 GPR survey of the Old Faithful area. Spring image acquisitions are important because the solar insolation approximates the winter minimum. On the calibrated March 2012, night-TIR mosaic (Fig. 5), similar TIR anomalies (A-F) first noted by White (1969) are recognized. Other night-TIR anomalies (G and H) are associated with infrastructure (buried water and sewer) in the Old Faithful area (Heasler and Jaworowski, 2011). The “linear belt” documented by White (1969) and identified as clay minerals by Livo et al. (2007) is evident as a 38 m long WNW linear trend (C) on the calibrated 2012 airborne, night-TIR temperature mosaic (Fig. 5), the helicopter air-oblique images (Fig. 6a and b) and the uncorrected, single-frame 27 March 2015 airborne, night-TIR image (6c).
Yellowstone National Park (YNP) helicopter air-oblique, day-TIR images acquired on 27 March 2015 at 10:26 a.m. (Fig. 6a) and fixed-wing night-TIR (Fig. 6c) show similar TIR trends and areas of hydrothermal activity seen prior to the 9–12 April 2015 GPR investigation. The 38 m long and approximately 5 m wide WNW-thermal trend seen on the 2010, 2011, and 2012 night-TIR mosaics (Figs. 4a, b and 5, label C) corresponds to the alignment of hydrothermal features in the 27 March 2015 helicopter air-oblique, day-TIR image (Fig. 6b) and the single-frame 27 March 2015, airborne night-TIR uncorrected image (Fig. 6c).
It is important to emphasize the attenuation of airborne TIR radiative ground temperatures by water vapor in the air above high temperature hydrothermal features (see discussion and Fig. 4 in Neale et al., 2016). Fig. 6a and b demonstrate the attenuation of TIR radiative ground temperatures due to water vapor associated with a minor eruption of Old Faithful Geyser. In Fig. 6a, the water vapor cloud associated with steam over the Old Faithful Geyser cone is evident in the center of the visible image, immediately to the right of the arrow labeled Old Faithful Geyser vent. In Fig. 6b, the helicopter air-oblique, day-TIR image shows the water vapor plume over the geyser cone attenuating the radiative TIR ground temperatures under the water vapor cloud. In contrast,the fractured and layered siliceous sinter of the former geyser cone (B in Fig. 6b and c) is not obscured by water vapor and contains higher temperatures over a greater area than Old Faithful Geyser. Thus, fracture patterns in siliceous sinter can be delineated with ground radiative temperatures.
In the Old Faithful area, fracture trends also are evident at the ground surface. U.S. Geological Survey researchers (White, 1969; Christiansen and Blank, 1974; Muffler et al., 1982;) mapped fractures on geyser mounds such as Split Cone, Old Faithful Geyser cone, and other geyser cones. White (1969) noted the following fracture trends on geyser cones in the Old Faithful area: a N to NNE-trending fracture on Split Cone, an E-W trending fracture on the geyser cone northeast of Old Faithful Geyser, and both NW and NE trending fractures on a geyser cone southeast of Old Faithful Geyser. Muffler et al. (1982) mapped similar fracture trends on the cone of Old Faithful Geyser and the surrounding geyser cones, including N-S and E-W trends on a small topographic rise northwest of Old Faithful Geyser.
Around Old Faithful Geyser, the following natural fractures were evident during the 9–12 April 2015 GPR survey of the area surrounding Old Faithful Geyser (Fig. 6): NNW, NW, NNE, and WNW. During the April 2015 GPR investigation, authors visually noted hydrothermal activity north of the Old Faithful Geyser cone and along a WNW-trending fracture on the Old Faithful Geyser cone. During the GPR survey, an alignment of hydrothermal features in hydrothermally altered red ground (C) also was observed and measured by a Trimble GPS receiver. On Spilt Cone (shown in Fig. 5), the following fractures were observed and measured with a Brunton compass: (1) an active NNE-trending fracture with steam and water and (2) an inactive pebble-filled NW trending fracture that was cross cut by the active NNE-trending fracture.
During the 9–12 April 2015 GPR survey, field measurements of fracture trends involved using a Brunton compass and a Trimble Pro-XH GPS receiver with a Zephyr antenna. Differential correction of the GPS points showed a range of horizontal error between 0.4 m and 1.1 m. The Brunton compass fracture measurements on Split Cone are estimated to be within 1°.
Thermal infrared images
TIR images used to construct Fig. 3, Fig. 4, Fig. 5, Fig. 6 were acquired using a FLIR Systems ThermaCAM SC660 with an 8–12 micron filter that was flown at night using a Utah State University fixed-wing aircraft, 1800 m above the ground level (AGL). Temperatures on the TIR images approximate ground surface temperatures within 1–3 °C. Neale et al. (2016) discuss the methods used to acquire, calibrate, and georectify the 2007–2012 night-TIR images.
Neale et al. (2016) also discuss the use of an ENVI change detection difference map algorithm for identifying change along natural fractures in Hot Spring Basin and Norris Geyser Basin within Yellowstone National Park. This previous use of the ENVI change detection algorithm (simple difference) in hydrothermal areas clearly identified changing hydrothermal activity along natural fractures (see Figs. 8a–c, 12, and 13a–b in Neale et al., 2016). In this study, the ENVI change detection difference produced an ENVI classification image (5 classes). This simple difference algorithm’s pixel-for-pixel comparison between the calibrated 25 September 2010 and 9 September 2011 temperature mosaics of the Old Faithful area generated a large magnitude, positive change map. The 2010–2011 temperature, difference image (Fig. 7) highlights hydrothermal trends in fractured siliceous sinter and changes that may not be obvious in a single night-TIR temperature mosaic or single, day-TIR scene. The September 2010-September 2011 difference map (Fig. 7) generated by the ENVI change detection algorithm also clearly shows an approximately 77 m long NW-thermal trend north of Old Faithful Geyser. On Fig. 7, there are approximately 25 m long NE-thermal trends (B in Fig. 5) northeast of the Old Faithful Geyser cone. Both the 25 September 2010 and the 9 September 2011 image acquisitions occurred under similar environmental conditions (Table 1) and each temperature mosaic also was georectified to the 2008 NSF EarthScope LiDAR base map.
It is clear that the spatial ground temperature patterns have been recognized by White’s (1969) initial TIR study of the Old Faithful area and by recent night-TIR image acquisitions (2007-2015; Jaworowski et al., 2010, 2012). This study utilizes a previously applied change detection technique in Yellowstone hydrothermal areas (Neale et al., 2016) and compares the derived night-TIR information with subsurface GPR imaging of the area surrounding Old Faithful Geyser.
Ground penetrating radar
Several papers have been published about the 2015 GPR study around Old Faithful Geyser. These manuscripts focused on initial GPR results (Lynne et al., 2017a), analyses of siliceous sinter (Lynne et al., 2017b), the hydrothermal fluid-filling sequence during an eruption of Old Faithful Geyser (Lynne et al., 2018b) and SEM analysis of a geyser egg (Smith et al., 2018).
A GSSI 200 MHz ground-coupled monostatic antenna with a 2014 GSSI SIR 4000 controller was used to collect GPR data. Raw data were post-processed using Radan 7 software. Post-processing was minimized to optimize preservation of raw data. The post-processing sequence involved adjusting the groundwave to zero on the vertical depth scale. A generic Radan 7 display gain setting of 6 was applied to all transects to enhance visualizations of low-amplitude reflections. Topographic correction was not necessary as transect slope gradients were minimal. The dielectric constant of 6 has been determined by field testing over vertically exposed sinter outcrops at other sites (see Fig. 2, Lynne et al., 2018b), where the velocity profile has been ground-truthed to a known depth.
Subsurface imaging data were collected along 35 individual transect lines (Fig. 8) during the 2015 GPR survey. Processing of the 2015 GPR data permitted an exploration of subsurface fractures permeability around Old Faithful Geyser. The combination of GPR and night-TIR made possible the identification of natural fractures that were not easily seen at the ground surface.
GPR results show interconnected hydrothermal fracture networks that extend to ∼3 m depth near the Old Faithful Geyser. Discrete fractures are shown on the GPR profiles among strong to weak amplitude reflections (Fig. 9). Strong amplitude reflections represent unaltered sinter. Fracture networks along these GPR profiles (i.e., purple dots on Fig. 9), generally occur where there is minimal heat discharging at the ground surface as shown on the 2010–2011 night-TIR difference image (Fig. 9). Partially altered sinter is shown as medium amplitude reflections. Fracture networks along these profiles (green dots, Fig. 9) usually occur close to areas where the 2010–2011 night-TIR difference image shows positive TIR anomalies on the ground surface.
Individual fractures located in zones with weak amplitude reflections (i.e., orange dots on Fig. 9) which are interpreted as highly altered sinter or clay, generally occur in areas where the 2010–2011 night-TIR difference image shows a positive TIR anomaly at the ground surface.
Fracture network orientations
A graphical summary of our results (Fig. 10) shows 27 GPR subsurface hydrothermal fracture networks and surface natural fractures as well as major TIR trends. The subsurface NNE, NE, ENE, NW and WNW trends clearly are evident around the cone of Old Faithful Geyser. Most of the subsurface GPR fractures identified do not have a surface expression.
NE and NNE trends
It is interesting that the active NNE trending fluid-filled fracture observed on Split Cone during April 2015 also was observed at ∼3 m depth on Split Cone (Fig. 10, line 4). Other locations surrounding the Old Faithful Geyser cone (Fig. 10, lines 10, 11, 25) and on the former Geyser/Hot spring Cone area (line 17) also show NNE trends.
At 3-m depth, similar NNE-trending GPR fractures (Fig. 10, lines 10 and 25) and NE-trending GPR fracture networks (line 23) intersect a NW trending night-TIR zone identified by applying an ENVI difference algorithm to the 2010 and 2011 night-TIR imagery. These intersections of active NNE – NE trends (Fig. 10, lines 10, 23 and 25) and NW night-TIR (heat) trends occur both on and near the cone of Old Faithful Geyser.
Northeast of Old Faithful Geyser at 3-m depth, the GPR analysis also identified a NNE-trending fracture network (Fig. 10, line 17) and a NE-trending fracture network (Fig. 10, lines 18, 19, 20, 27) on a former geyser/hot spring cone.
NW and NNW trends
Subsurface NW to NNW trends occur at Split Cone (Fig. 10, lines 2, 3), near Old Faithful Geyser Fig. 10, lines 9, 24, 26), between Split Cone and Old Faithful Geyser (Fig. 10, lines 5, 6, 22) and between Old Faithful Geyser and the former geyser/hot spring cone (Fig. 10, line 12, 14, 15, 21). Between Old Faithful Geyser and the former geyser cone (Fig. 10, lines 14, 21) the subsurface NW trends are adjacent and parallel to a night-TIR trend identified using a difference algorithm on 2010–2011 night TIR images.
NE – NW trends
The modeled GPR subsurface fracture network generally is consistent with surface faults and fractures identified by field work and analysis of night-TIR images around Old Faithful Geyser. However, the subsurface intersection of NE and NW trends at Old Faithful Geyser is perhaps the most interesting result of our study (Fig. 10, lines 24 and 25, 23 and 26, 9 and 10).
WNW and ENE trends
West of Split Cone, a subsurface WNW fracture trend (Fig. 10, line 1) also runs parallel along a WNW trending night-TIR zone associated with a surface exposure of red clay and hot ground. A subsurface WNW trend (Fig. 10, line 8) also runs parallel to WNW night-TIR trend extending through the cone of Old Faithful Geyser. A third subsurface WNW-trend occurs between the Old Faithful Geyser cone and the former geyser/hot spring cone (Fig. 10, line 13).
Located between Split Cone and Old Faithful Geyser, an ENE subsurface fracture (Fig. 10, line 7), runs perpendicular to two TIR anomalies shown on the 2010–2011 night-TIR difference image. On a former geyser/hot spring cone, another ENE subsurface fracture (Fig. 10, line 16) occurs with a TIR anomaly on the 2010–2011 night-TIR difference image.
Around Old Faithful Geyser, the hydrothermal fracture network identified by surface and subsurface techniques is noteworthy. These natural fractures were identified initially in the 1960’s during field mapping of the Old Faithful area. Additional geologic mapping and field work (2002-2015) provided relevant field observations to constrain interpretations of the subsurface GPR images and interpretations of the night-TIR images. These interpretations of GPR and night-TIR images are consistent with regional trends and local fractures originally mapped during the 1960’s field studies.
It is interesting that the 2010–2011 night-TIR difference image (Fig. 7) best showed the surface fractures around Old Faithful Geyser. Earthquake swarms and ground deformation are common in Yellowstone and the Old Faithful area. Prior to the September 2010 night-TIR image acquisition, there was a major earthquake swarm (Yellowstone Volcano Observatory, https://volcanoes.usgs.gov/observatories/yvo, accessed 12 February 2019) centered on the Madison Plateau (See Fig. 1). This volcanic plateau has a series of volcanic vents aligned along a NW trend and its rhyolite lava flows cover the western edge of the 640 ka Yellowstone caldera. The 2010 earthquake swarm during the 2010 winter season (January and February 2010) caused concern among residents and visitors in the Old Faithful developed area and prompted discussion of a winter evacuation plan. In response to these concerns, Yellowstone National Park geologists conducted a helicopter reconnaissance of the Madison Plateau to determine if there were any visible fractures on the volcanic plateau; only snow-covered trees and pressure ridges were observed.
North-northwest fractures were noted in the UGB after the 1959 Hebgen Lake earthquake (Marler and White, 1974). Marler and White (1974) wrote that new hot areas in the UGB also appeared “as linear distributions of dead or dying lodgepole pine, generally trending northwest”. Further, they noted that the Old Faithful vent is a fracture-controlled vent. Thus, it is possible that the 2010 Madison Plateau earthquake swarm affected fractures and the hydrothermal system so that important fractures around Old Faithful Geyser were detected at the ground surface during the 25 September 2010 night-TIR acquisition.
Additionally, it is also noteworthy that the 2010–2011 night-TIR trends and surface fractures in sinter had similar orientation to subsurface fractures detected by the 2015 GPR study. This similar orientation over time indicates that both are robust techniques and important to understanding the hydrothermal system around Old Faithful Geyser. A continuation of the 2015 GPR study that detects subsurface fractures at depths >3 m is a reasonable future endeavor around Old Faithful Geyser.
The combination of GPR on fractured siliceous sinter and high-spatial resolution night-TIR (1 m or less) around Old Faithful Geyser can be used for hydrothermal monitoring and protection of this iconic hydrothermal area. The hydrothermal monitoring technique developed and implemented at Norris Geyser Basin (Heasler and Jaworowski, 2018) can be applied to the UGB hydrothermal system and specific areas such as the cone of Old Faithful Geyser. In the Old Faithful developed area, the combination of field observations, GPR, laboratory analyses and high spatial resolution (1 m), night-TIR permits the separation of natural changes in the UGB hydrothermal system from anthropogenic influences.
Around Old Faithful Geyser, shallow (∼3 m depth) subsurface natural fractures may or may not be evident at the ground surface. Combining GPR, night-TIR, and physical observations of the ground surface can illustrate the connection between surface fractures and shallow subsurface natural fractures that permit shallow fluid flow around Old Faithful Geyser. Night-TIR imagery, GPR and field studies identified the following important natural fractures around the cone of Old Faithful Geyser: (1) active water-filled NNE and WNW surface fractures and (2) the intersection of surface TIR WNW trends with an ENE shallow subsurface trend (GPR-defined) between Old Faithful Geyser’s cone and Split Cone. Perhaps, the most significant result of this study is the GPR-defined, subsurface intersection of “blind” NE and NW natural fractures around Old Faithful Geyser.
Authors: Cheryl Jaworowski, Bridget Y.Lynne, Henry Heasler, Duncan Foley, Isaac J.Smith, Gary J.Smith