- UW Department of Geology and Geophysics
Four elements of the long-enigmatic Heart Mountain detachment are the focus of this synthesis of new and earlier work. First, the geometry of the detachment system is more completely defined by our interpretation that the nearby South Fork thrust was the contractional toe of the Heart Mountain detachment during its early movement. Second, emplacement of the allochthon included phases with both catastrophic and slow rates of movement, driven by gravitational instability of active volcanoes above a dipping, pre-volcanic substrate. Third, initiation of displacement resulted from reduction of basal friction by elevated fluid pressure along the basal detachment, presumably beneath a critically stressed Coulomb wedge. Fourth, maintenance of low basal friction, to allow displacements of the allochthon in excess of 30 kilometers, was aided by endogenic formation of a gas suspension along the basal detachment. Thus, a unique combination of conditions and processes led to formation of the world's largest-known subaerial, detached extensional system. It represents a worst-case scenario for the magnitude of destruction resulting from sector collapse of an active subaerial volcano.
- Abiathar breakaway
- Absaroka volcanics
- Heart Mountain
- South Fork
- volcano collapse
Two major detachment systems have been described on the northeast flank of the Absaroka Mountains of northwest Wyoming and adjacent Montana (Fig. 1). The extensional Heart Mountain fault system (HMf) has been a well-known yet enigmatic staple of the literature of structural geology for many decades (e.g., Dake, 1916; Pierce, 1957, 1973; Suppe, 1985) and was the subject of research (e.g., Bucher, 1933, 1947; Voight, 1974; Melosh, 1983) and contemplation (e.g., Hubbert and Ruby, 1959; Davis, 1965; Goguel, 1969; Hsu, 1969) by a number of well-known structural geologists of the latter half of the 20th century. Nearby is the lesser-known, contractional South Fork fault (SFf) system (Dake, 1916; Bucher, 1933; Pierce, 1941, 1957, 1986; Pierce and Nelson, 1969; Blackstone, 1985; Clarey, 1990, 2008). Here we report that these two detachment systems may be linked in the subsurface, allowing two seemingly disparate geologic features to be integrated into a single coherent entity representing the multistage Eocene collapse of an active Absaroka volcanic pile. The remarkably different expressions of collapse associated with this volcanic edifice, as illustrated by these detachment systems, provide a rich end-member example of the disparate manifestations, both catastrophic and gradual, of the collapse of volcanoes.
HEART MOUNTAIN FAULT SYSTEM
The rootless Heart Mountain detachment fault (HMf; Fig. 1) in northwest Wyoming and adjacent Montana has generated some of the longer-lasting puzzles of geology. The Heart Mountain allochthon is more than an order of magnitude larger in volume (ca. 2,000–3,000 km3) than any other documented subaerial landslide or debris avalanche (Siebert, 1984). Between 48–50 Ma, a slab of Ordovician through Mississippian carbonate rocks and overlying Eocene, 2–4 kilometers in thickness and > 1,100 square kilometers in area, detached, thinned, and extended, translating up to 50 kilometers southeastward to cover > 3,500 square kilometers. The detachment area is crudely triangular in map view, with the New World intrusive/volcanic center at its up-dip apex (Fig. 1). The HMf is comprised of a rear-breakaway fault, a bedding-plane component inclined < 20 to the southeast (in the direction of transport) with sidewalls typically poorly exposed or removed by erosion, and a ramp component along which the fault rises stratigraphically to the level of Eocene rocks; and possibly moved onto the contemporaneous land surface (Figs. 2 and 3A). In the drainage of the Shoshone River's North Fork, the HMf subparallels footwall bedding in the Eocene Willwood Formation, which may have been the land surface at the time of Heart Mountain faulting, or may have been a detachment internal to the Willwood Formation. All workers on questions associated with the HMf owe a debt to the mapping by William G. Pierce, who spent much of his career creating 1:62,500 geologic quadrangle maps of most of the detachment area.
The Heart Mountain allochthon has been variously described through the years as a contractional thrust sheet, an extensional allochthon comprised of numerous free-sliding detached blocks, and a continuous extensional allochthon. The thrust-sheet model (e.g., Dake, 1916; and Pierce, 1957) was discounted by Bucher (1947), based on observations of extensional rather than contractional structures within the allochthon. It was further discredited by Pierce's (1960) discovery of what he interpreted to be the rear ‘breakaway’ fault, demonstrating the allochthon to be rootless. The model of numerous, free-sliding detached blocks (e.g., Bucher, 1947; and Pierce, 1973), with ‘tectonic denudation’ of detachment areas between the detached blocks, was negated by the observation that volcanic rocks in contact with the detachment are allochthonous and lie within detached graben, the ‘slide blocks’ being horsts within a laterally continuous allochthon (e.g., Hauge, 1982, 1985, 1990). Beutner and Craven (1996) showed that fault kinematics documented by Hauge were the same regardless of whether they were observed within Paleozoic carbonate or Tertiary volcanic rocks. Their analysis also discounted the model of Malone (1995), who argued for two successive catastrophic emplacement events, with different movement kinematics: free-sliding, dominantly carbonate masses were emplaced first (similar to Pierce's model of 1973), followed by emplacement of most of the volcanic rocks through sector collapse of the Sunlight volcano.
Many aspects of the Heart Mountain allochthon and its emplacement mechanics remain enigmatic. Foremost among these is the manner in which basal friction was reduced to allow long-distance transport of little-deformed, kilometer-scale masses within the allochthon along the shallowly dipping detachment. The rate of motion also has been controversial, with proponents favoring both slow (Hauge, 1985, 1990, 1993b) and catastrophic (Bucher, 1947; Pierce, 1973; Beutner and Craven, 1996; Anders et al., 2000; and Beutner and Gerbi, 2005) movement. Equally elusive has been an explanation as to why sliding initiated on a bedding surface 2–3 meters above the base of the Ordovician Bighorn Dolomite rather than within thick underlying Cambrian shales. And finally, if rapid movement occurred, the nature of the trigger that initiated catastrophic sliding has remained unresolved (Pierce, 1973; Voight, 1974; and Melosh, 1983).
SOUTH FORK FAULT SYSTEM
The South Fork fault system (SFf) is exposed in the drainage of the South Fork of the Shoshone River, to the south of the bedding-plane component of the HMf (Figs. 1 and 4). It forms classic decollement fold/thrust geometries with up to ca. 10 kilometers of shortening (Pierce and Nelson, 1969; and Clarey, 1990), ramping up-section southeastward from a decollement in the Jurassic Sundance Formation and possibly Gypsum Spring Formation to a decollement in the Upper Cretaceous Cody Shale, from which it ramps upward once again to offset the Tertiary Willwood Formation. Followed northeastward from where it appears from beneath a cover of younger volcanic rocks in the upper South Fork Valley, it is exposed along strike for ca. 35 kilometers to Buffalo Bill Reservoir. During a drawdown of that reservoir, the SFf system was mapped to bend 90° to join exposures that trend northwest along the steeply southeast-dipping forelimb of the basement-involved Rattlesnake Mountain anticline (Pierce, 1970). In the Rattlesnake Mountain forelimb, several exposed splays of the SFf system ramp between Jurassic and Lower Cretaceous strata and accrue less than a kilometer of apparent strike-slip. Farther to the northwest, these faults disappear beneath the HMf near the head of the valley. The principal strand or strands of the SFf that accommodate the bulk of the documented 8–10 kilometers of slip on the system must lie to the southwest, hidden beneath the Heart Mountain allochthon, younger volcanic rocks, and Quaternary deposits.
The South Fork fault variously has been interpreted to be the front of a gravity slide unrelated to the HMf (Blackstone, 1985; and Pierce, 1957, 1986), the easternmost expression of the Cordilleran overthrust belt (Clarey, 1990), or the toe of the HMf (Beutner et al., 2004). Here we develop the basis for and implications of the last interpretation.
AGES AND RELATIVE TIMING OF HMf AND SFf
Both the SFf and HMf are constrained as early to middle Eocene in age, but different workers have interpreted their relative ages differently. The movement of the HMf is dated by numerous Ar/Ar ages of volcanic and intrusive rocks of the Absaroka Supergroup. These indicate that faulting may have begun prior to ca. 50.1 Ma and ended by ca. 48.1 Ma. The HMf overlies middle Wasatchian strata (early Eocene in age) of the Willwood Formation (Gingerich, 1983). Faults of the SFf system tilt and offset strata of the Willwood Formation (Pierce and Nelson, 1969; Pierce, 1986; and Clarey, 1990), but the SFf lacks constraining Ar/Ar ages. Pierce (1957, 1973) interpreted the SFf as older than the HMf, “… but the time interval between them is not necessarily large” (Pierce, 1957, p. 624). Blackstone (1985) interpreted the SFf as older than the HMf but acknowledged the inferential rather than empirical basis for this conclusion. Clarey (2008) recently argued that the SFf is younger than the HMf, as had Sales (1983).
Cross-cutting relationships between the SFf and HMf have been viewed as ambiguous due to poor exposure, which has led to the disparate interpretations of their relative timing. Pierce and Nelson's (1969) cross section A–A′ suggests that the Heart Mountain detachment is not folded by, and therefore must be younger than, the South Fork thrust footwall ramp at Sheep Mountain. Clarey's (1990) cross section A–A′ (Fig. 4) shows the Heart Mountain fault folded above the South Fork footwall ramp in the same area. The stratal dips within the Heart Mountain allochthon above the South Fork ramp are variable, but they do not indicate a systematic northwestward tilting as would be expected if the SFf were younger than the HMf. We favor the interpretation by Pierce and Nelson (1969) that the SFf is older than the HMf.
In addition, a cross-cutting relationship is indicated (Pierce and Nelson, 1969; and Clarey, 1990) between an allochthonous carbonate mass within the HMf allochthon and one of the NW-striking faults within the SFf allochthon (i.e., the Castle fault; Fig. 1). We agree with Clarey (1990), who refers to it as a ‘tear fault,’ that the Castle fault is restricted to the SFf allochthon. For two reasons, we disagree with the suggestion by Blackstone (1985) and Pierce (1986) that it is a normal fault that offsets the SFf. First, the latter suggestion unnecessarily complicates the structural history by invoking a structural style (basement-involved normal faulting) not observed elsewhere in the area. Secondly, we agree with Clarey (1990) that no seismic reflections from rocks older than Jurassic are offset on the data provided by Brittenham and Tadewell (1985).
We describe the Castle fault as a dip-slip (probably normal) fault above a lateral ramp in the Sff. The Castle fault and the lateral ramp that underlie it project beneath the carbonate mass of the HMf allochthon on the southeast side of Sheep Mountain ca. 1000 m to the northwest (Fig. 1). Northeast of the Castle fault, the SFf is a frontal ramp across which the SFf cuts stratigraphically upward from Jurassic rocks to the Cretaceous Cody Shale; its hanging wall is tilted ca. 25° to the northwest (Fig. 4A). To the southwest of the Castle fault at this location, the SFf is a bed-parallel detachment within Jurassic strata, and its hanging wall is sub-horizontal (Fig. 4B). Along the Castle fault, strata in the hanging wall of the SFf within a few hundred meters of the overlying Heart Mountain allochthon are tilted ca. 60° to the southwest (Pierce and Nelson, 1969), to accommodate approximately 600 meters of throw across the Castle fault at this location. If the HMf were older than the SFf, the HMf and its allochthon would have been monoclinally down folded to the southwest along with the strata of the SFf allochthon that they overlie. According to published geologic maps (e.g., Pierce and Nelson, 1969), however, the HMf allochthon at this location is disrupted above the HMf, as is common in many parts of the allochthon, but the HMf climbs ca. 100m in elevation to the southwest: it is not folded as part of the South Fork allochthon. We find this to be strong evidence that the portion of the Heart Mountain allochthon that now lies within the Shoshone River drainage was emplaced after South Fork thrusting.
RATE OF EMPLACEMENT OF HEART MOUNTAIN ALLOCHTON
Since the work of Bucher (1933; also see Pierce, 1973; Voight, 1974; and Malone, 1995), the predominant view has envisioned catastrophic emplacement of the Heart Mountain allochthon. Catastrophic emplacement rates seemed required by the long-distance transport of individual slide blocks in Pierce's (1957) ‘tectonic-denudation’ model of Heart Mountain faulting (e.g., Voight, 1974; and Prostka, 1978). In the context of Hauge's ‘continuous-allochthon’ model, however, catastrophic emplacement rates have been associated with postulated pressurized fluids that engendered and/or maintained displacement.
Hughes (1970a, b, 1973) first proposed gas-supported emplacement. Despite repeated strong objections by Pierce and Nelson (1970) and Nelson et al. (1972, 1973), Hughes argued that such support could have been provided by volcanic gas injected into the fault horizon. He based this view on his observations of volcanic glass, small sub-spherical matrix-coated grains, and small vesicles within what he viewed as ‘fluidized’ detachment-fault breccia. Beutner and Craven (1996) provided an updated version of the volcanic gas model. More recently, Beutner and Gerbi (2005) described a distinctive, cement-like microbreccia layer, millimeters to meters in thickness, along many parts of the Heart Mountain bedding-plane detachment and in dikes injected into the upper plate. They argued that the microbreccia represents residue from a gas suspension that formed along the fault and facilitated catastrophic sliding.
The strongest evidence for such a suspension is the abundance of accreted grains in the microbreccia (Fig. 5). The grains are indistinguishable from those formed in volcanic and impact ejecta clouds and in gas-charged diatremes, and are incompatible with the presence of an aqueous phase along the fault during their formation. The accreted grains also rule out an explanation for the microbreccia as having formed simply by frictional comminution during faulting. A variety of sedimentary structures and fabrics within the microbreccia layer, including grain-size lamination (Anders et al., 2000), dropstone-like objects deforming laminations, groove casts, and shape-preferred orientations (Beutner and Gerbi, 2005), also are in accord with the gas-suspension model.
In contrast, Hauge (1982, 1985, 1993b) argued against catastrophic emplacement of the Heart Mountain allochthon. Hauge (1982, 1985) suggested that, whereas a model of the allochthon as numerous slide blocks seems to require catastrophic emplacement (e.g., Bucher, 1933), a continuous-allochthon model permits non-catastrophic emplacement rates driven by gravity spreading rather than gravity sliding. Hauge (1985) documented cross-cutting relationships among dikes and faults that require multiple faulting events within the allochthon rather than the single catastrophic-emplacement event as envisioned by previous workers.
Hauge (1993b) presented evidence for slow (mm–cm/yr) movement on the Heart Mountain fault, citing common occurrences of calcite slickenfibres on normal and normal-oblique faults within upper plate carbonate rocks and to a lesser extent in the volcanic rocks. These faults and fibres accord kinematically with movement of the Heart Mountain fault (Beutner and Craven, 1996) and thus suggest a phase of slow extension within the upper plate. Additionally, a listric normal fault with ca. 100 meters of displacement in the upper plate in Painter Gulch (Fig. 1) shows evidence of recurrent movement interacting with multiple phases of dike injection and calcite deposition. The well-known presence of hundreds of mafic dikes in the upper plate, compared to very few in the lower plate, also has been taken to indicate that they contributed to slow extension of the upper plate (Hauge, 1985).
In addition, the accreted grain shown in Figure 5 may provide evidence for a movement event with a ‘normal’ geologic rate (perhaps averaging a few cm/yr) preceding the catastrophic event. The core of this accreted grain is composed of a laminated-carbonate cataclasite. We interpret the cataclasite as having formed at a non-catastrophic geologic strain rate because of its textural similarity to typical fault breccias. After lithification, the cataclasite was fragmented during the beginning of subsequent catastrophic movement. Finally, radiometric age data, discussed below, suggest that Heart Mountain faulting took place over a time interval of at least 1.8 million years. Thus, there appears to be evidence for both slow and catastrophic movement on the Heart Mountain fault system.
RATE OF EMPLACEMENT OF SOUTH FORK ALLOCHTHON
In Dake's (1916) description of the SFf, normal rates (i.e., in the range of a few cm/yr) of thrust emplacement seem implicit. In Pierce's (1973) review of the mechanism problem of the ‘probably cataclysmic’ (p. 462) emplacement of the Heart Mountain fault, he suggested that a similar mechanism (and similar rates?) seem to have been indicated for the SFf. Clarey (2008) explicitly advocated catastrophic emplacement. Our recent field studies show that slow movement and high fluid pressures along the SFf are indicated by the abundance of cross-fibre, crack-seal veins along the fault. Indeed, the fault and subsidiaries can be traced in float in many locations by the presence of fragments of these veins. Where the fault is exposed along the South Fork of the Shoshone River, remarkable jigsaw breccias with crack-seal veins separating matching brecciated fragments indicate up to 40-percent expansion of the rock volume (Fig. 6). These jigsaw breccias demonstrate existence of greater-than-lithostatic fluid pressure along the fault. The source of these fluids is uncertain. They may have been transported along the fault by a downslope hydraulic gradient, or perhaps they sourced from the underlying, Tertiary Willwood Formation. Because of shallow burial depths, the Willwood Formation probably was not fully compacted before being overridden by the South Fork thrust sheet.
THE HMf–SFf CONNECTION
In concept, the detached contractional deformation of the SFf might have been accommodated in its hinterland either by the basement-involved contraction or the basement-detached extension that characterize its structural setting. The surface geology of the SFf hinterland, near the western projection of Pat O'Hara Mountain, could not have accommodated a zone of basement-involved contraction with 10 kilometers of shortening. Such displacement is comparable to that of the northeastern and eastern flanks of the Beartooth Mountains, which are characterized by up to 10 kilometers of structural relief and > 100 kilometers of strike length. We conclude that detached extensional deformation must have accommodated South Fork thrusting.
We view as admissible two models of detached extension to describe the hinterland of the SFf. Both require a zone of basement-detached extension, filled at least in part with Eocene volcanic rocks (Fig. 7). One model (Fig. 7B), in accord with suggestions of Blackstone (1985), envisions South Fork faulting as entirely predating Heart Mountain faulting. A South Fork breakaway fault was located along the south flank of the monocline extending southwestward from the junction of the Pat O'Hara and Rattlesnake Mountain uplifts. The other model (Figs. 3B and 7C), which we prefer, links the contractional SFf directly to the Heart Mountain extensional detachment. We prefer the latter model because it is geometrically and kinematically admissible, provides a simpler explanation for the spatial and temporal juxtaposition of the two detachments, and it would have more easily allowed the accumulating pile of volcanics of the New World, Crandall, and Sunlight vent areas to drive the detachment system.
In this model, the earlier SFf allochthon includes an exposed hinterland extensional zone (i.e., the Heart Mountain detachment), a concealed extensional-translational zone southwest of Rattlesnake Mountain, and an exposed contractional zone (i.e., the South Fork thrust). The HMf and SFf thus shared the bedding-plane component of the HMf as a slip surface. The earlier SFf ramped up-section stratigraphically southward from near the base of the gently south-dipping Ordovician (HMf) to the Jurassic Sundance Formation. It then remained in that gently south-dipping unit to form a broad flat, before again ramping up-section, forming the fold/thrust belt in the valley of the South Fork of the Shoshone River. In this model, up to 10 kilometers of shortening above the South Fork thrust is balanced by up to 10 kilometers of extension above the Heart Mountain detachment.
We conclude that a Heart Mountain–South Fork allochthon was emplaced as a detached, linked system, extensional in its up dip regions and contractional in its down dip regions. In this sense, it is kinematically analogous to detached, linked extensional–contractional systems that characterize passive margins worldwide (e.g., Rowan et al., 2004; and Bilotti and Shaw, 2005). The SF–HM detachment differs from typical linked-system detachments in that it formed across a terrane of basement-involved monoclonal structure. As a result, it locally cut downward in elevation as it cut upward stratigraphically—from an Ordovician dolomite horizon to presumably weaker Jurassic units of shale and gypsum.
Viewed in this manner, the early HMf–SFf system is remarkably similar in character and scale to the rootless extensional/contractional fault system formed in and south of the Bearpaw Mountains in northern Montana at approximately the same time (Reeves, 1946; Hearn, 1976; and Gukwa and Kehle, 1978). Both of the systems span approximately 100 kilometers from the rearmost normal faults to the frontal thrusts, both have an intrusive/volcanic complex at their rear, and both probably formed due to gravitational instability engendered by volcanism.
SEQUENCE AND RATES OF DETACHED FAULTING EVENTS
The arguments presented above make key points: (1) HMf and SFf share a hinterland (i.e., the northwestern, bedding-plane part of the HMf); (2) SFf predates the toe of the HMf (at least in the area southwest of Rattlesnake Mountain); (3) SFf was emplaced at a non-catastrophic rate; and (4) the shared hinterland detachment shows evidence for both non-catastrophic and catastrophic emplacement rates, the final event having been catastrophic. These arguments indicate a history that can be most simply envisioned as two phases (Fig. 8). The earlier phase was dominated by non-catastrophic, extensional spreading in the hinterland of the HMf. The spreading was directly linked to contractional emplacement of the SFf thrust sheet, with or without expansion of the HMf across Dead Indian Hill into the Bighorn Basin. The later phase included emplacement of the HMf allochthon to the south-southeast across the SFf thrust sheet and to the east-southeast into the Bighorn Basin, if the allochthon there had not already been emplaced. The later history was in part, or possibly entirely, characterized by catastrophic emplacement events.
Our favored model reconciles the structural field observations and inferences cited above, but the mechanism of initiation of the detachment system remains to be addressed. Since the work of Hubbert and Ruby (1959), enhanced fluid pressure along basal detachments has been the ‘usual suspect.’ The jigsaw breccia (Fig. 6) provides evidence for elevated fluid pressures on the South Fork detachment, and it supports recent isotopic data, discussed next, which provide concrete support for assertion of elevated fluid pressure in the case of the Heart Mountain detachment.
MECHANICS OF DETACHMENT INITIATION: FLUIDS ALONG HMf/SFf
The mechanical state of the HMf-SFf system is explained by Coulomb wedge theory (e.g., Dahlen et al., 1983), whereby gravitational instability of a (proto-) allochthon can result from an oversteepened surface slope, a reduction in cohesive strength, or a reduction of basal friction. Templeton et al. (1995) and Douglas et al. (2003) demonstrated that hydrothermal fluids were introduced into the bedding-plane component of the HMf. Veins along the HMf near the breakaway show 18O and 13C depletions characteristic of the New World intrusive/volcanic center (Douglas et al., 2003; Fig. 1). Isotopic depletion in veins adjacent to the fault decreases with distance to the southeast along the fault, becoming negligible at the ramp near the Dead Indian monocline. We suggest that these fluids played a critical role in initiation of the phase of slow volcanic spreading, which was floored by the bedding plane part of the HMf and extended to the south as the SFf system.
The fluids also may have played an important role in selecting the slip horizon. The HMf bedding-plane detachment lies within a limey horizon ca. 2–2.5 meters above the base of the Ordovician Bighorn Dolomite, between an underlying Cambrian shaley unit and overlying, massive carbonates. Calcite veins are present in both hanging wall and footwall rocks and also are found along some faults in the upper plate. We suggest that many of these veins formed during the earlier, noncatastrophic phase of movement on the HMf. A vein with fibres in the slip direction of the HMf, collected by one of the authors, lies along a bedding-parallel fault about 1.5 meter below the HMf at White Mountain. It does not reflect the Laramide shortening direction found in pre-HMf veins. Rather, it has a steeply inclined calcite-twin shortening signature that may reflect vertical loading resulting from emplacement of the allochthon (Craddock et al., 2000).
Where presently exposed, intrusions in the New World center are emplaced into the Cambrian sequence of shales and limestones and into the underlying Archean gneisses. Younger Paleozoic rocks and the Tertiary volcanic pile that presumably overlay them have since been removed by erosion. The presence of a cap of Ordovician Bighorn Dolomite at the time of intrusion is suggested by the previously cited observation (Douglas et al., 2003) that hydrothermal fluids isotopically characteristic of the New World center were funneled from the New World center into the relatively soluble horizon near the base of the Bighorn Dolomite. These fluids then moved to the southeast and south down the topographic and hydraulic gradient, trapped between underlying shales and overlying massive dolomites. Volcano spreading typically occurs along a basal tuff or shale horizon (e.g., Wooller et al., 2004). The movement along the carbonate slip surface of the bedding-plane part of the HMf/SFf system appears to be unique among collapsing volcanoes (but Burchfiel et al., 1982 reported that ‘hard basal-layer thrusting’ is observed within orogenic belts).
The location of the Heart Mountain detachment within carbonates, although admittedly enigmatic, may be explained by the role of high-pressure hydrothermal fluids along this surface. The Sunlight and Crandall intrusive/volcanic centers (Fig. 1) may have added to the load produced at the New World intrusive/volcanic center that induced spreading. There is, however, no sign of their influence on the movement pattern or on the areal pattern of isotopic depletion. Rather, given the presence of fluids derived from the New World center, the movement pattern seems to have been controlled by the slope of the detachment surface (as with other spreading volcanoes; Wooller et al., 2004), rather than by the specific location of extrusive centers. The true ‘breakaway’ probably was located in or just north of the New World center, striking to the northeast, now removed by erosion. The breakaway described by Pierce (1960, 1980) probably is a detachment sidewall rather than a breakaway, as suggested by Hauge (1985), based on shallowly plunging striae observed on the exposure east of Abiathar Peak (Fig. 1).
INITIATION OF LATE-PHASE CATASTROPHIC MOVEMENT
Initiation of catastrophic movement has been attributed to earthquake shaking (Bucher, 1947; and Pierce, 1960), a fluid wedge at the rear of the allochthon (Voight, 1974), and a ‘phreatomagmatic/hydraulic’ mechanism (Straw and Schmidt, 1981a, b). More recently, Aharanov and Anders (2006) provided a refined version of the last model. They invoked fluid pressures created through heating by extensive diking of pore fluids, which were trapped in Ordovician Bighorn Dolomite. However, the areas that they cite as typical (i.e., Cathedral Cliffs and White Mountain) have many more dikes in the upper plate than are usual, and the evidence of fluid flow along the HMf presented by Templeton et al. (1995) and Douglas et al. (2003) contradicts this model of fluid-pressure trapping.
Although the mechanism or event that triggered catastrophic collapse remains enigmatic, we suggest that the collapse may be recorded by the formation of the body of the Homestake phreatomagmatic breccia at the New World center. That body lies along the HMf breakaway near its northern preserved tip (Johnson and Thompson, 2006). Deposits of Pb, Ag, Cu, and Au have been mined on a small scale from hydrothermal deposits in the New World center since the 1880s, and the large subsurface extent of the Homestake breccia was documented in the 1980s and 1990s through renewed exploration. We propose that a pulse of magmatic and extrusive activity at the New World center both engendered transitory enhancement of fluid pressure locally beneath the HMf allochthon and increased the surface slope of the allochthon. As the gravitational state of the allochthon moved from critical to supercritical, the allochthon collapsed catastrophically, unroofing the magmatic system sufficiently to engender phreatomagmatic explosions recorded as the Homestake breccia pipe at the New World center.
If the phreatomagmatic explosion that produced the Homestake breccia was the immediate result of catastrophic sliding on the HMf, its age is key to pinning down the time of sliding. Although systematic dating of the detachment area has not been undertaken, available data provide the basis for a provisional radiometric chronology. Igneous bodies in the New World center have been dated at 50.1 ± 0.3 to 48.1 ± 0.5 Ma using Ar/Ar in biotite and hornblende, and potassic alteration associated with Au suggests mineralization at 48.0 ± 0.3 Ma (Douglas et al., 2003). Further age control comes from Ar/Ar-plateau ages of 48.34 ± 0.11 and 48.59 ± 0.06 (J. Colgan, personal communication, 2002) obtained from biotites in rootless plutons in the upper plate of the HMf, This requires that catastrophic movement was younger than ca. 48.4 Ma. A deformed dike at White Mountain has an Ar/Ar age of 48.21 ± 0.05 Ma (Hiza and Snee, 1999; and Hiza et al., 2000). In dating volcanics from Sunlight Basin southward, Feeley and Cosca (2003) bracketed an unconformity separating the lower (ca. 48.5 Ma) and upper Trout Peak Trachyandesite (ca. 48.1 Ma). We suggest those dates also may bracket catastrophic movement of the HMf. Taken together, these ages appear to constrain catastrophic movement to between ca. 48.2 and 48.1 Ma. However, a 50.01 ± 0.14 Ma date from an ashflow tuff that postdates the breakaway west of Hurricane Mesa (Hiza and Snee, 1999; and Hiza et al., 2000) indicates that spreading began at least 1.8 million years earlier.
A fundamental problem with determining the time of HMf movement based upon the ages of Absaroka volcanic rocks stems from the unwarranted assumption that Heart Mountain faulting comprised a single catastrophic event that defined a boundary between volcanic formations. Following this assumption, various authors (Pierce 1963; Nelson and Pierce, 1968; Pierce et al., 1973; Prostka et al., 1975; Pierce, 1997; and Nelson et al. 1980) implicitly and often explicitly (Pierce 1963; Nelson and Pierce, 1968; Pierce et al., 1973; Prostka et al., 1975; Pierce, 1997; and Nelson et al., 1980) defined the Cathedral Cliffs Formation as having developed exclusively prior to HMf movement and the Wapiti Formation as having developed entirely after its movement. However, large volumes of rock mapped as Wapiti Formation were subsequently found to be in tectonic contact with the HMf (Hauge, 1985, 1990, 1993a), and Heart Mountain faulting has been shown to have comprised multiple events (Hauge, 1985; this paper) that may have spanned ca. 1.8 m.y. Detailed remapping of rocks assigned to the Absaroka Supergroup in this area, together with a robust program of radiometric dating and geochemical analyses, will be required to sort out the complex volcanic/volcaniclastic stratigraphy of these rocks.
Rhodes et al. (2007) argued that a profound, ca. 49-Ma desiccation event in Lake Gosiute within the Green River Basin (ca. 175 km south of the HMf) resulted from blockage of southward-flowing drainages by catastrophic Heart Mountain faulting. They relate the subsequent flood of volcanic debris into the lake to breaching of this blockage. We suggest that it is more probable that the desiccation event was produced by drainage blockage as a consequence of the early phase of gradual volcano spreading on the HMf/SFf system. This argument is based partly on the Ar/Ar dates younger than 49.0 Ma of rocks transported by catastrophic HMf movement. Equally important is the presence of a beheaded channel remnant in the HMf footwall at Fox Creek (Fig. 1) that is filled with volcanic and volcaniclastic rocks (Hauge, 1983). The channel has an outcrop width of ca. 300 meters and thickness of ca. 60 meters in cliffs trending ca. S. 60° E. Presuming southward flow, the true width would have been ca. 250 meters. Older fragments of filled channels (including the Crandall Conglomerate) found in both the upper and lower plates of the HMf contain no volcanic clasts (Pierce and Nelson, 1973; and Beutner and DiBenedetto, 2003). Because the Fox Creek channel is filled with coarse volcanic debris and flows that are truncated upward by the HMf, southward transport of volcanic debris from the Absaroka province toward the Green River Basin seems likely to have preceded catastrophic HMf emplacement.
MAINTENANCE OF CATASTROPHIC MOVEMENT
The greatest enigma of the Heart Mountain detachment has long been the large magnitudes of displacement required by the kilometer-scale horsts of Paleozoic strata within distal reaches of the allochthon. At Heart Mountain, the preserved dimensions of allochthonous Paleozoic rocks (length × width × height in km) are 2 × 1 × 0.5, and at Sheep Mountain the preserved dimensions are 6 × 5 × 0.5 kilometers. Both of these horsts moved a minimum of 25–30 kilometers with little internal deformation. Regardless of the mechanism by which displacement was initiated along the Heart Mountain detachment, an unusual mechanism for reduction of basal friction seems to have been required to engender such displacements. Many such mechanisms have been suggested.
Since the mid 1970s, pressurized fluid of some sort has been the favored mechanism. Reduction of friction as a result of large vertical earthquake accelerations was a contending hypothesis (Bucher, 1947; and Pierce, 1973). Hydrothermal fluids clearly moved along the surface of sliding (Templeton et al., 1995; and Douglas et al., 2003), but the role of these fluids in maintaining movement is unclear. Support (Hughes, 1970a, b; and Beutner and Gerbi, 2005) or driving (Voight, 1974) of the slide sheet by gases also have been proposed, as has acoustic fluidization of rocks along the fault (Melosh, 1983). Beutner and Gerbi (2005) called upon frictional heating to dissociate limestone along the sliding surface. That mechanism provides a possible endogenous source of gas, an explanation for the unusual character and cementation of the fault rock, and a positive feedback mechanism to maintain low friction along the base of the allochthon, thus permitting anomalously great displacements of large blocks. By this model, a detachment within or at the base of carbonates is required for such large displacements to occur.
In the context of any proposed mechanism, we find the displacement magnitudes of the large coherent horsts within the Heart Mountain allochthon (e.g., Heart Mountain itself) to be astonishing. Comparison with volcanic debris avalanches, however, suggests that the Heart Mountain allochthon may have been preconditioned for large displacements by virtue of three factors: (1) its large volume (Hsu, 1975); (2) the pre-collapse development of fractures and normal faults during the earlier South Fork phase of gradual extension (e.g., Ui, 1983; and Voight, 1973); and (3) its shallowly southeast-dipping carbonate substrate. We favor the Beutner and Gerbi (2005) endogenic-gas model because of its strong empirical basis.
We recognize multi-phase, multi-mode collapse of the northeastern flank of the Absaroka volcanic field, as expressed by a linked Heart Mountain detachment and South Fork thrust. Early-phase collapse was characterized by non-catastrophic spreading that involved extension above the proximal component of the Heart Mountain detachment linked to distal contraction above the South Fork thrust. Late-phase collapse included a final catastrophic event, with extensional deformation of both the proximal and distal parts of the Heart Mountain allochthon. Late-phase collapse probably included other catastrophic and non-catastrophic spreading events that cannot be resolved using available data. The basal detachments lie within Ordovician, Jurassic, Cretaceous, and Eocene sub-volcanic strata. The detachment surface within Ordovician carbonates was favored for detachment by hydrothermal fluid migration, whereas detachments within younger strata were along inherently weak shaley and gypsum-rich horizons. Similar collapse events have been recognized at other volcanoes, both ancient and recent (Wooller et al., 2004). However, no other subaerial detachments on the scale of the Heart Mountain fault are known.
In our view, the long-standing enigma of Heart Mountain faulting is rendered significantly less enigmatic by recognition that three prominent geologic elements of northwest Wyoming—the Heart Mountain detachment, South Fork thrust, and Absaroka volcanic field—are best understood not as three distinct geologic phenomena unrelated except by proximity, but rather as three expressions of the same phenomenon: a gravitationally unstable, active volcanic field.
We thank Timothy Clarey, Kent Sundell, and Robert Thomas for improving this manuscript through their thoughtful reviews, as well as Gregory Davis, who critiqued an early version of the manuscript, and Jason Lillegraven and Brendon Orr for their thorough editing. Discussions with Joe Colgan, Don Fisher, Tom Oesleby, and Don Wise greatly helped clarify our thinking. Edward Beutner died on December 23, 2008, a few weeks after we received the reviewers' comments and RMG's indication that the paper likely would be accepted for publication pending revisions. Ed was delighted. To the junior author of this paper, he was an inspiring colleague, a brother, and a great friend. Thanks, Ed!
- Received October 7, 2008.
- Revision received March 4, 2009.
- Accepted July 11, 2009.