Highly Conductive Horizons in the Mesoproterozoic Belt-Purcell Basin: Sulfidic Early Basin Strata As Key Markers of Cordilleran Shortening and Eocene Extension

12.

OLD G

The Geological Society of America Special Paper 522 OPEN ACCESS

Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin: Sulfidic early basin strata as key markers of Cordilleran shortening and Eocene extension

P.A. Bedrosian* U.S. Geological Survey, Denver Federal Center, West 6th Ave. Kipling St. MS 964, Lakewood, Colorado 80225, USA

S.E. Box* U.S. Geological Survey, 904 West Riverside Ave., Room 202, Spokane, Washington 99201, USA

ABSTRACT

We investigated the crustal structure of the central Mesoproterozoic Belt Basin in northwestern Montana and northern Idaho using a crustal resistivity section derived from a transect of new short- and long-period magnetotelluric (MT) stations. Two- and three-dimensional resistivity models were generated from these data in combination with data collected previously along three parallel short-period MT profiles and from EarthScope MT stations. The models were interpreted together with coincident deep seismic-reflection data collected during the Consortium for Continental Reflec- tion Profiling (COCORP) program. The upper-crustal portion of the resistivity model correlates well with the mapped surface geology and reveals a three-layer resistivity stratigraphy, best expressed beneath the axis of the Libby syncline. Prominent features in the resistivity models are thick conductive horizons that serve as markers in reconstructing the disrupted basin stratigraphy. The uppermost unit (up to 5 km thick), consisting of all of the Belt Supergroup above the Prichard Formation, is highly resistive (1000–10,000 W⋅m) and has relatively low seismic layer velocities. The intermediate unit (up to 7 km thick) consists of the exposed Prichard Formation and 3+ km of stratigraphy below the deepest stratigraphic exposures of the unit. The intermediate unit has low to moderate resistivity (30–200 W⋅m), relatively high seismic velocities, and high seismic reflectivity, with the latter two characteristics resulting from an abundance of thick syndepositional mafic sills. The lowest unit (5–10 km thick) is nowhere exposed but underlies the intermediate unit and has very high conductivity (4–8 W⋅m) and intermediate seismic velocities. This 17–22-km-thick three- layer stratigraphy is repeated below the Libby syncline, with a base at ~37 km depth. Seismic layer velocities indicate high mantle-like velocities below 37 km beneath the Libby syncline. The continuous high-conductivity layer in the lower repeated section

*E-mails: [email protected]; [email protected].

Bedrosian, P.A., and Box, S.E., 2016, Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin: Sulfidic early basin strata as key markers of Cordilleran shortening and Eocene extension, in MacLean, J.S., and Sears, J.W., eds., Belt Basin: Window to Mesoproterozoic Earth: Geological Society of America Special Paper 522, p. 305–339, doi:10.1130/2016.2522(12).

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 306 Bedrosian and Box

is apparently displaced ~26 km to the east above a low-angle normal fault inferred to be the downdip continuation of the Eocene, east-dipping Purcell Trench detachment fault. Reversal of that and other Eocene displacements reveals a >50-km-thick crustal section at late Paleocene time. Further reversal of apparent thrust displacements of the three-layer stratigraphy along the Lewis, Pinkham, Libby, and Moyie thrusts allows construction of a restored cross section prior to the onset of Cordilleran thrusting in the Jurassic. A total of ~220 km of Jurassic–Paleocene shortening along these faults is indicated. The enhanced conductivity within the lowest (unexposed) Belt stratigraphic unit is primarily attributed to one or more horizons of laminated metallic sulfides; graphite, though not described within the Belt Supergroup, may also contribute to the enhanced conductivity of the lowest stratigraphic unit. A narrow conductive horizon observed within the Prichard Formation in the eastern part of the transect correlates with the stratigraphic position of the world-class Sullivan sedimentary exhalative massive sulfide deposit in southern British Columbia, and it may represent a distal sulfide blanket deposit broadly dispersed across the Belt Basin. By analogy, the thick conductive sub–Prichard Formation unit may represent repeated sulfide depositional events within the early rift history of the basin, potentially driven by hydrothermal fluids released during basaltic underplating of attenuated continental crust.

INTRODUCTION overlying thrust belt geometries. Estimates of total shortening from these and other studies of the entire width of the thrust belt The Cordilleran thrust belt in the southern Canadian Rockies near the international border are variable, ranging from 100 km and Glacier National Park of the northern U.S. Rockies (Fig. 1) (Price and Sears, 2000) to over 200 km (Price, 1981), with most is a classic area for viewing the structure of a foreland fold-and- estimates being somewhere between. thrust belt (Bally et al., 1966; Dahlstrom, 1969, 1970; Price, 1981; Farther west, in the hinterland of the Jurassic to Paleocene Boyer and Elliott, 1982). These and other studies have docu- Cordilleran thrust belt, exposures of midcrustal metamorphic mented significant Jurassic to Paleocene shortening (50–150 km) core complexes in the United States and Canada, which were of Phanerozoic strata in the frontal ranges of the eastward-verging uplifted in early Eocene time, are bounded to the east by Eocene thrust belt above a gradually westward-deepening basal décolle- east-dipping low-angle normal faults. Significant displacement ment that occurs at or above the contact with basal Precambrian (tens of kilometers) on the bounding low-angle normal faults is crystalline basement. Additional, but less well-constrained, short- required to account for uplift and exposure of these midcrustal ening occurs west of the frontal ranges, where a thick sequence of rocks. These eastward-deepening faults must displace features Mesoproterozoic sedimentary rocks (Belt and equivalent Purcell within the Cordilleran thrust belt, but they have not been consis- Supergroups) occurs structurally above the imbricated Phanero- tently imaged in the seismic-reflection data (Yoos et al., 1991; zoic strata along the Lewis thrust. Spectacular exposures for over Cook and van der Velden, 1995). 200 km along strike of the Lewis thrust in Glacier and Water- In order to provide new information relevant to the crustal ton National Parks (Willis, 1902; Dahlstrom, 1970; Mudge and structure in the hinterland of the Cordilleran thrust belt, we Earheart, 1980; Boyer and Elliot, 1982; Fermor and Price, 1987; imaged the crustal conductivity structure along a 143 km tran- Yin et al., 1989; Fermor and Moffat, 1992; Whipple, 1992) have sect from just east of the Rocky Mountain Trench near Whitefish, provided clear, well-exposed examples of thrust belt geometries Montana, to the Priest River metamorphic core complex (herein over a broad area of the foreland region. Priest River complex) near Bonners Ferry, Idaho (Figs. 1 and 2). The fate of the Lewis thrust as it deepens farther west and Parts of this transect coincide with legs of the Idaho-Montana presumably joins the westward-deepening basal décollement of COCORP deep seismic-reflection profile (Yoos et al., 1991). We the Phanerozoic front ranges was beyond the reach of early indus- deployed paired long- and short-period magnetotelluric (MT) try seismic-reflection data (Bally et al., 1966), but it has since instruments at 34 stations spaced roughly 5 km apart along the been pursued by government-funded deep seismic-reflection transect that allowed us to generate a continuous model of electri- programs in the United States (Consortium for Continental cal resistivity along the transect down to a depth of ~50 km. We Reflection Profiling [COCORP]; Potter et al., 1986; Yoos et al., combined that data with seismic-reflection profiles and seismic 1991) and in Canada (LITHOPROBE; Cook et al., 1992; van der interval velocities derived from the COCORP data, as well as Velden and Cook, 1994; Cook and van der Velden, 1995). These preexisting MT profiles to the north and south and recent Earth- studies have had variable success in tracking the basal thrust belt Scope MT data. We used this new data set to model and interpret décollement into the deeper crust to the west and in elucidating the two- and three-dimensional crustal resistivity structure over

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin 307

Figure 1. Geological map of Mesoproterozoic strata of the Belt-Purcell Basin, showing the locations of Figures 2 and 9. Locations of two major sedimentary exhalative massive sulfide deposits (Sullivan and Sheep Creek) within the Prichard and equivalent Al- dridge Formations are shown, along with two deep drill holes. Figure is modified from Box et al. (2012). ALB—Alberta; BC—Brit- ish Columbia; ID—Idaho; MT—Montana; OR—Oregon; WA—Washington; WY—Wyoming.

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 308

Figure 2. Location of magnetotelluric (MT) stations from this study shown on a base of (A) geologic map (simplified from Miller et al., 1999; Harrison et al., 1992; Harrison and Cressman, 1993), and (B) isostatic residual gravity map (from Mankinen et al., 2004), shown along with MT stations from Phoenix USA (KCU-1, KCU-5, and KCU-9 from Gupta and Jones, 1995) and from EarthScope (http://www.earthscope.org), and seismic reflection lines ID-2 and MT-2 from the COCORP program (Yoos et al., 1991), and line WTF-82-1 from Signal of Montana (Yoos et al., 1991). Also shown is the location of the deep oil exploratory well Gibbs #1 (Boberg et al., 1989). COCORP—Consortium for Continental Reflection Profiling.

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin 309

a 20,000 km2 area centered on the main transect, revealing new in British Columbia (East Kootenay orogeny of McMechan and images of the geometry of the hinterland thrust belt. Herein, we Price, 1982) and in east-central Idaho, ~350 km to the southeast explore the implications of this thrust belt geometry for palin- (Doughty and Chamberlain, 1996). Although McMechan and spastic reconstruction of the Belt-Purcell Basin geometry, and we Price (1982) suggested the East Kootenay orogeny was com- consider the implications of the imaged resistivity structure on pressive in nature, Anderson and Parrish (2000) suggested it our understanding of early Belt Basin evolution. may have been an extensional event akin to that in east-central Idaho documented by Doughty and Chamberlain (1996). A REGIONAL SETTING widespread static thermal event between 1.05 and 1.15 Ga has been documented 100 km to the north near the Sullivan mine Belt Supergroup (Anderson and Parrish, 2000), within the Priest River complex in northern Idaho (Doughty and Chamberlain, 2008), and The transect area is underlain primarily by folded and ~170 km to the south in north-central Idaho (Zirakparvar et al., faulted strata of the Mesoproterozoic Belt (Purcell in Canada) 2010). Multiple late Neoproterozoic to Late Cambrian mafic Supergroup, fill of the Belt (or Belt-Purcell) basin (Figs. 1 and alkalic intrusive events (780, 660, and 500 Ma) are recorded by 2). The Belt Supergroup has an estimated stratigraphic thickness small intrusive bodies throughout the Belt Supergroup outcrop of between 15 and 20 km in the central part of its exposure belt area, and these are associated with the late Neoproterozoic rift- (Cressman, 1989; Lydon, 2000; Hoy et al., 2000), including ing of western North America (Cecile et al., 1997; Harlan et al., 5–7 km of section below the stratigraphically deepest exposures, 2003; Lund et al., 2010). This was followed by deposition of a which have only been imaged seismically (Cook and van der thin (<2 km) sequence of Paleozoic carbonate and clastic rocks. Velden, 1995). The basin is considered to have developed in an We infer that any deformation of Belt strata in the study area aborted intracontinental rift setting (Chandler, 2000), with basin prior to Paleozoic deposition was local and minor, as Cambrian fill deposited between 1500 and 1400 Ma (Anderson and Davis, strata are only known to depositionally overlie the youngest for- 1995; Evans et al., 2000; Lydon, 2000). Basin strata (from bottom mation in the Belt Supergroup (Harrison et al., 1992), indicating upward) consist of (1) the basal 5–7-km-thick unit of unknown little if any folding and erosion prior to Paleozoic deposition. Sig- lithology that is seismically imaged as weakly reflective, (2) the nificant deformation of Belt strata occurred during Late Jurassic– Prichard Formation, a southwest-derived, deep-water, generally Paleocene activity of the Cordilleran thrust belt and during major argillaceous turbidite section with locally abundant syndeposi- Eocene extensional faulting. Granitic plutonism occurred locally tional mafic sills (5–7 km thick) that produce prominent seismic in Cretaceous and Eocene time. reflectivity, (3) the Ravalli Group, a southwest-derived fluvial- deltaic complex (2.5 km), (4) the Piegan Group, a cyclic carbon- Structural Framework from Surface Geology ate and siliciclastic sequence of either lacustrine or marine setting (formerly the middle Belt carbonate) capped by a sequence of Northwestern Montana and northern Idaho are part of the mafic to felsic lavas (2 km), and (5) the Missoula Group (2.5 km), Sevier fold-and-thrust belt (DeCelles, 2004; Dickinson, 2004; a south- and east-derived fluvial succession (Fig. 1; Winston and Burchfiel et al., 1992), the eastern front of the Late Jurassic– Link, 1993; Link et al., 2007; Winston, 2007; for correlation Paleocene part of the Cordilleran thrust belt that extends from of the U.S. unit nomenclature with Canadian unit nomenclature Canada to Mexico. This reach of the Sevier belt differs from to the north across the border, see Box et al., 2012). The Prich- other parts of the belt to the north and south by the presence of ard Formation argillaceous strata are characterized by abundant the thick Mesoproterozoic Belt strata in the hanging wall west of carbonaceous seams and by abundant iron sulfide (Huebschman, the front ranges of the belt. The Sevier belt here consists of the 1973; Cressman, 1989), which can result in higher electrical con- frontal “Montana disturbed belt” (Mudge, 1970) of imbricated ductivities if physically interconnected. Typical thicknesses indi- Paleozoic and Mesozoic strata, structurally overlain by the Lewis cated for these units are quite variable, owing in part to internal and associated thrusts with Proterozoic strata in their hanging thickness variations across the geographic outcrop belt as well as walls. From east to west, these thrusts are the Hefty, Whitefish, unseen structural thickening or thinning. The depositional base Pinkham, Libby, and Moyie thrust systems (Fig. 2; Harrison of the Belt Supergroup is only exposed along the eastern flank et al., 1992; Harrison and Cressman, 1993; DeCelles, 2004). of the basin in central Montana. There, the thin Neihart Quartzite These thrust systems and intervening major folds are discussed rests nonconformably on Paleoproterozoic basement and grades in the following. upward through shelf and slope strata into a greatly compressed Regional outcrop patterns (Fig. 2) are controlled by broad section of Prichard Formation–equivalent strata, with the entire north-northwest–trending folds. From east to west, these are the Belt section ~5 km thick. synclinal Rocky Mountain Trench, the Purcell anticlinorium, the The Belt Supergroup experienced a number of deforma- Libby syncline, and the Sylvanite anticline. Most workers inter- tional and thermal episodes through its long postdepositional pret folding to be thrust-related (Yoos et al., 1991; Cook and history. Deformation, magmatism, and metamorphism have been van der Velden, 1995; DeCelles, 2004). However Harrison et al. documented at 1.37 Ga ~120 km to the northwest of our transect (1992) and Harrison and Cressman (1993) suggested that the

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 310 Bedrosian and Box

major folds are Precambrian in age and are unrelated to, and cut The Libby syncline is narrow and steep sided, with maxi- by, Cordilleran thrusts. Precambrian folding with slaty cleav- mum dips greater than 70° on each flank. Harrison et al. (1992) age has been documented near the Sullivan Pb-Zn-Ag deposit and Harrison and Cressman (1993) showed several west-dipping (Leech, 1962; McMechan and Price, 1982), and an unconfor- thrust faults within the axial portion of the Libby syncline, which mity between Belt and Cambrian strata in western Montana is are simplified in Figure 2. As mapped, the thrust faults have locally observable (Deiss, 1935). However, in the area of Fig- limited stratigraphic throw, although south of our transect, Cam- ure 2, all of the very small occurrences of Cambrian strata rest brian strata depositionally overlying upper Missoula Group are on the stratigraphically highest formation of the Missoula Group preserved beneath upper Missoula Group in the footwall of the (Harrison et al., 1992), suggesting to us that any folding and ero- Libby thrust. West of the town of Libby, the east-dipping, west- sion prior to Cambrian deposition was minor and is not reflected directed Snowshoe thrust has a displacement of ~4 km (Fillipone in the strongly folded outcrop pattern. and Yin, 1994). Fillipone and Yin suggested that the Snowshoe Beginning on the east of our transect (Fig. 2A), the White- thrust is most likely a back thrust within the upper plate of the fish Range consists of middle and upper Belt Supergroup units Libby thrust. (Ravalli, Piegan, and Missoula Groups), cut by several south- The Sylvanite anticline (Moyie anticline in Canada) is a west-dipping thrust faults (Harrison et al., 1992). The gently broad, doubly plunging anticline (northern nose in Canada) with southwest-dipping Hefty thrust system occurs on the east flank a wide gentle crest and sharp, steeply dipping flanks. The gen- of the Whitefish Range near the Canadian border, placing Mis- tle crest of the anticline exposes Prichard Formation, while the soula Group over Paleozoic carbonates and quartzose sandstones. more steeply dipping flanks expose strata from higher in the Belt Fritts and Klipping (1987a, 1987b) used seismic data to image Supergroup stratigraphy. A well on the crest of the anticline in other southwest-dipping thrust systems of lesser stratigraphic off- Canada (Duncan Energy Moyie #1) penetrated 3500 m of Prich- set within the Belt Supergroup (Wigwam, Whitefish, Ten Lakes ard Formation strata heavily intruded by basaltic sills. The paral- thrust faults, simplified in Figure 2 as Whitefish thrust) that are lelism between the axis of the Sylvanite anticline and the Libby mapped within the western Whitefish Range. The western flank thrust system suggests it is a hanging-wall anticline developed of the Whitefish Range is marked by a mostly inferred set of during east-directed thrust motion. down-to-the-west Tertiary normal faults under Quaternary cover The Moyie thrust is a west-dipping, east-directed thrust of the Rocky Mountain Trench, though inferred stratigraphic off- with Prichard Formation in its hanging wall and upper Missoula sets are small. Group in its footwall, which represents the greatest stratigraphic The Rocky Mountain Trench is a broad (15-km-wide) top- separation of any thrust fault in the area of Figure 2. At the lati- ographic low exposing mostly Quaternary glacial gravels with tude of the MT transect, the footwall strata dip steeply (70°–80°) scattered bedrock knobs of Belt strata. The trench is manifest as under the fault, while the hanging-wall strata dip steeply toward a discontinuous, possibly en-echelon series of Tertiary half gra- the fault. As the Moyie thrust is followed south of the MT tran- bens above a discontinuous set of down-to-the-west listric nor- sect, it bends to the west before it intersects the subvertical right- mal faults. Across the border in Canada, the Rocky Mountain lateral Hope fault. In that bend area, a north-northwest–trending Trench fault has up to 10 km of displacement (van der Velden syncline of Piegan and Missoula Groups projects under the and Cook, 1994), while to the south near Whitefish (Fig. 2), a Moyie thrust, while a north-northeast–trending syncline parallels well with 500 m of Tertiary sediment suggests a correlative fault the fault in the upper plate. with ~3 km of displacement (Fritts and Klipping, 1987a, 1987b). The hanging wall of the Moyie thrust consists of Prichard However, no Tertiary sediment is known at the latitude of our Formation riddled with mafic sills, dipping moderately eastward MT transect, nor does the geologic mapping suggest significant toward the thrust (Miller and Burmester, 2004). It is intruded normal displacement of stratigraphy. A gentle syncline of Pie- by mid-Cretaceous granitic plutons of the Kaniksu batholith gan Group is exposed in the western part of the Rocky Moun- (Gaschnig et al., 2012), with U-Pb zircon ages of 105–120 Ma tain Trench. and biotite and hornblende K-Ar (cooling) ages of 90–100 Ma. The southwest-dipping Pinkham thrust separates exposures These plutons predate at least some of the displacement on the of Prichard Formation and Ravalli Group strata within the Purcell Moyie thrust, indicated by 70 Ma mylonite formation along the anticlinorium to the west from Piegan Group exposures in the fault (Fillipone and Yin, 1994). Rocky Mountain Trench. The Gibbs #1 well (Fig. 2) was drilled The only major Eocene and younger normal fault exposed into the crest of the Purcell anticlinorium to test for possible within the MT transect area is the east-dipping Purcell Trench petroleum-bearing Paleozoic strata beneath the Pinkham thrust. detachment fault (Rhodes and Hyndman, 1984; Rehrig et al., 1987; The well is interpreted to have penetrated the gently dipping Harms and Price, 1992; Doughty and Price, 1999, 2000), bound- Pinkham thrust at ~5.2 km below the surface (calculated dip = ing the east side of the Eocene Priest River metamorphic core 18°), but Belt Supergroup quartzite, not Paleozoic strata, underlie complex. The Priest River complex consists of interleaved foli- the thrust (Boberg et al., 1989). The top of the anticlinorium is a ated Cretaceous granitic rocks and Belt metasedimentary rocks. very gentle and broad structure. However, the western flank of the Peak metamorphism occurred at 15–30 km depths at ca. 72 Ma Purcell anticlinorium steepens sharply into the Libby syncline. (Doughty et al., 1998). The fault marks a sharp break to high

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin 311

metamorphic grade and younger metamorphic cooling ages in the wells (Gibb #1 in Montana—Boberg et al., 1989; Moyie#1 in footwall below the fault (Miller and Engels, 1975; Doughty and southern Canada—Cook and van der Velden, 1995) were drilled Price, 1999, 2000). The actual fault trace is mostly hidden within on major lower Belt Supergroup anticlines (Purcell anticlinorium a wide valley beneath Quaternary glacial deposits. Doughty and and Sylvanite anticline, respectively). Geologic mapping at the Price (2000) described an exposure of the fault just north of our scale of 1:250,000 was completed over most of our study area transect, consisting of mylonitic rocks dipping 57° to the east and in the 1980s (Harrison et al., 1992; Miller et al., 1999). Deep capped by a 3–6 m zone of chloritic microbreccia. Normal offset seismic-reflection surveys penetrating the entire crust were com- across the combined Purcell Trench–Newport fault system is esti- pleted soon after as part of the COCORP program in western mated at 35–68 km (Harms and Price, 1992), while Doughty and Montana and northern Idaho (Yoos et al., 1991) and as part of the Price (1999) estimated 66 km of displacement across the southern LITHOPROBE program in adjacent southern Canada (Cook and Purcell Trench. van der Velden, 1995; van der Velden and Cook, 1994). Upper- The Priest River complex has exposures of Archean ortho- crustal (<15 km) MT studies in western Montana (Fig. 2, lines gneiss ~45 km to the south-southwest of the westernmost MT KCU-1, KCU-5, and KCU-9) and southern Canada occurred station in our profile (Doughty and Price, 1999). Those Archean during the same period (Gupta and Jones, 1995). Significant rocks occur beneath the gently north-dipping Spokane Dome interpretations of crustal structure in western Montana have mylonite zone, which is overlain by the Selkirk Crest portion occurred since that time utilizing some or all of the previous data of the Priest River complex. The Spokane Dome mylonite zone (Kleinkopf, 1997; Price and Sears, 2000; Fuentes et al., 2012). has top-to-the-east kinematics and has been interpreted to have The most comprehensive studies of crustal structure in formed in the deep crust either along an east-vergent thrust of the region have been the deep seismic-reflection studies by the the Sevier orogenic belt (Rhodes and Hyndman, 1984) or along LITHOPROBE program along strike in southern British Colum- an east-vergent detachment fault during Eocene extension, possi- bia (Cook and van der Velden, 1995; van der Velden and Cook, bly overprinting an earlier thrust-related mylonite (Doughty and 1994). These studies presented 11 detailed interpretations of full Price, 1999). crustal seismic-reflection profiles over an area from 50 to 225 km A major top-to-the-east detachment fault (Kettle fault) is along strike to the northwest of our MT transect; one of these exposed along the flooded Columbia River ~115 km west of the interpreted sections is reproduced here as Figure 3A. Broadly, west end of the MT profile (Rhodes and Cheney, 1981; Parrish these profiles show a 50-km-thick crust to the east of the Rocky et al., 1988), but it may be present in the deep crust in the western Mountain Trench, with ~12 km of thrust-imbricated Proterozoic part of the MT profile. Parrish et al. (1988) suggested that the and Paleozoic strata overlying about a 38-km-thick crystalline Kettle and related faults to the north in British Columbia may Archean and Paleoproterozoic basement. Their seismic profile have offsets on the order of 30–40 km. Hurich et al. (1985) used array shows a dramatic thinning of the crystalline basement to seismic reflection to track the mylonitic rocks along the fault for less than 20 km over a 40 km swath centered on the Rocky Moun- 10 km east of their surface exposure, dipping at ~15.5° eastward tain Trench, coinciding with an overall crustal thinning down to toward the MT profile. Potter et al. (1986) used the COCORP 35–40 km to the west. East of the Rocky Mountain Trench, the deep seismic-reflection profiles to track the gently dipping (12°) basal décollement of the Cordilleran thrust belt is interpreted to Kettle fault to depths of 12–15 km at a position ~55 km west of be either atop crystalline basement or within the thin (1–2 km) the west end of our MT profile. If the Kettle fault maintains that Paleozoic sequence that rests depositionally on the crystalline gentle dip farther east, it would enter the west end of our profile basement. West of the Rocky Mountain Trench, the profiles are at a depth between 23 and 27 km. interpreted to show the basal décollement atop crystalline basement or atop a thin veneer of lower Belt strata that rests deposi- Crustal Structure: Previous Work tionally on crystalline basement. The structure of the thrust belt above the basal décollement Overthrusting of Belt Supergroup over Mesozoic and Paleo- and west of the Rocky Mountain Trench was interpreted rela- zoic rocks has long been recognized in the geologic mapping of tively simply by Cook and van der Velden (1995). The thick Belt the front ranges of the northern Rockies (Willis, 1902; Camp- Supergroup sequence is basically shown to have been transported bell, 1914), and it was recognized that the thrust surfaces dipped bodily eastward with an attached 5–10-km-thick slab of crystal- below the surface to the west. Seismic-reflection studies in sup- line basement. The Belt strata are shown as mildly deformed in port of petroleum exploration in the foothills of the Canadian a broad, gentle anticline (northern continuation of the Sylvanite Rockies began in the 1940s (Link, 1949; Bally et al., 1966) and anticline) draped over a similarly shaped crystalline basement greatly clarified the crustal structure of the upper 5–10 km of the thrust slab, with a gentle syncline to the east. Later, more detailed frontal part of the thrust belt east of our study area and in south- analysis by Ainsworth (2009) reinterpreted the crystalline base- ern Canada. Industry seismic surveys were extended west into ment slab to consist of Aldridge and equivalent Prichard Forma- our study area in the 1980s with recording depths down to 15 km tions strata with abundant mafic sills. The Moyie thrust west of (Fritts and Klipping, 1987a, 1987b; Yoos et al., 1991), and deep the Sylvanite anticline is shown to have moderate (20 km) top-to- (5.2 and 3.5 km, respectively) wildcat petroleum exploration the-east displacement.

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 312 . Abbreviations: FF—Flathead fault; HT—Hefty thrust; KF—Kettle fault; LT—Lewis thrust; LiT—Libby thrust; MT—Moyie MT—Moyie thrust; LiT—Libby thrust; LT—Lewis fault; KF—Kettle thrust; HT—Hefty fault; FF—Flathead Abbreviations: B . part 3. Comparative crustal-scale cross sections through the Belt-Purcell Basin in southern Canada and northern Montana-Idaho. (A) Cross section at Figure 3. Comparative et al. (1991). (C) Cross section from this study along Yoos WTF-82-1 from and (1995). (B) Cross section along seismic lines ID-2, MT-2, Velden der ~49°15 ′ N from Cook and van as corridor same of thirds two western the RMT— fault; Trench PTF—Purcell complex; core metamorphic River PRC—Priest thrust; PT—Pinkham anticlinorium; PA—Purcell basement; American NAB—North thrust; WT—Whitefish thrust. anticline; SA—Sylvanite fault; Trench Mountain RMTF—Rocky topographic low; Trench Mountain Rocky

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin 313

Cook and van der Velden (1995) recognized several west- (20–30 W⋅m), due to the presence of a minor conducting phase dipping normal faults, but they did not recognize an east-dipping (Hyndman and Shearer, 1989; Haak and Hutton, 1986; Shank- Purcell Trench low-angle normal fault cutting the thrust belt land and Ander, 1983). structures. The Rocky Mountain Trench fault, on the east side MT studies seek physical explanations for regions of anoma- of the physiographic trough, was interpreted to be a listric west- lous conductivity. Certain mineralogies such as metallic sulfides dipping normal fault with ~10 km of post-thrust displacement. and graphite are highly conductive (s ≥ 1 S/m), giving rise to Several west-dipping normal faults were interpreted west of the enhanced bulk conductivity when these minerals are present even Moyie fault with relatively small (<2 km) displacements. in small percentages. Within regions of active or recent tectonics, Interpretation of the COCORP deep seismic-reflection data both aqueous fluids and partial melts are common causes of ele- in western Montana and northern Idaho (coincident with our cen- vated crustal conductivity (e.g., Hyndman and Hyndman, 1968; tral MT profile) by Yoos et al. (1991) is broadly similar to that Wei et al., 2001; Li et al., 2003; Wannamaker et al., 2008). In of Cook and van der Velden (1995); their interpreted cross sec- both these cases, rock conductivity is dominated by ionic con- tion is reproduced in Figure 3B. The crust thins from ~45 km to duction within the fluid phase. 35 km over 40 km horizontally, centered on the Rocky Mountain MT is a natural-source electromagnetic technique that Trench. Crystalline basement thins from 35 km to 10 km in thick- involves surface measurements of orthogonal electric- and mag- ness over that same interval. The basal thrust décollement occurs netic-field variations (Tikhonov, 1950; Cagniard, 1953). Inter- within the lower Belt Supergroup section west of the Rocky action of the solar wind (a stream of charged particles ejected Mountain Trench, overriding 3–7 km of basal Belt strata in from the Sun) with Earth’s magnetic field distorts Earth’s ambi- depositional contact with the thinned crystalline basement. They ent magnetic field and produces low-frequency electromagnetic also interpreted an allochthonous slab of crystalline basement to energy that penetrates Earth. Relative to the incident magnetic override autochthonous Belt strata at depth under the core of the field, the amplitude, phase, and polarization of the measured Sylvanite anticline. The deep structure of the thrust belt east of electric field depend upon the conductivity of the medium it the Sylvanite anticline is significantly imbricated in the interpre- travels through. At the frequencies of interest (104–10-4 Hz), the tation of Yoos et al. (1991), in contrast to essentially preserved propagation of electromagnetic energy in Earth is described by coherent stratigraphy 50 km to the north in Canada, in the inter- a diffusion equation, of the same form as that governing heat pretation of Cook and van der Velden (1995). Yoos et al. (1991) diffusion through a solid. Within this frequency range, lower- also interpreted the Purcell Trench low-angle normal fault to cut frequency energy penetrates deeper within Earth and averages thrust belt structure for up to 50 km east of its surface expression, over a larger volume. Bulk resistivity also affects the diffusion of penetrating the crust to a depth of 10–12 km. electromagnetic energy, with greater subsurface resistivity giving We present a structural interpretation of that portion of the rise to a greater depth of penetration for a given frequency. Typi- Belt Basin centered on the Montana and Idaho COCORP profiles cal depths of exploration in MT range from hundreds of meters (Fig. 3C for comparison). Our interpretation relies primarily on to tens or even hundreds of kilometers. magnetotelluric data but draws heavily upon the existing seismic MT data can be inverted, with varying degrees of complexity data. In contrast to previous studies, our interpretation gives more for one-, two-, or three-dimensional (1-D, 2-D, or 3-D) resistivity significance to Eocene extension along the Purcell Trench and models. The data used in inversion of MT data are frequency- Kettle faults. dependent transfer functions between the horizontal electric (E) and magnetic (H) field vectors. This complex, four-element METHODS impedance tensor, Z = E–1H, contains information about the heterogeneous and sometimes anisotropic subsurface resistivity

MT and Electrical Conductivity structure. Vertical magnetic-field transfer functions, Tz, are also commonly estimated and inverted alongside MT impedance data

Electrical conductivity within Earth spans many orders of to help further constrain subsurface resistivity structure. Tz data magnitude, reflecting variations in lithology, mineralogy, and are particularly sensitive to lateral changes in resistivity. Further hydrology (Guéguen and Palciauskas, 1994; Palacky, 1987; Haak background on the MT method can be found in Chave and Jones and Hutton, 1986; Keller, 1987). As such, spatial variations in (2012, and references therein). conductivity record structural information inherited from past tectonic events and the current physico-chemical state. Because the MT Data Acquisition conductivity of crustal materials is typically quite low (<<1 S/m), it is common to work interchangeably between conductivity and This paper describes the analysis and interpretation of MT its reciprocal, resistivity. On average, the Phanerozoic section in measurements along a series of subparallel profiles within the the western United States averages 100 W⋅m, while crystalline Belt Basin (Fig. 2). Broadband (102–10-3 Hz) and long-period continental upper crust is quite resistive (104 W⋅m), reflecting the (1–10-4 Hz) MT data were collected at a total of 34 sites along a low conductivity of silicate and carbonate rocks (Gough, 1986). 150 km transect crossing the central Belt Basin from the Priest In contrast, continental lower crust is commonly less resistive River metamorphic core complex (Selkirk Crest) in the west to

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 314 Bedrosian and Box

the Whitefish Range in the east. MT data collected along this these studies had sufficient penetration to image the base of these central profile (hereafter the main profile) closely followed high-conductivity zones. COCORP deep seismic lines ID-2 and MT-2, both collected in The most encompassing previous study of the Belt Basin the mid-1980s (Yoos et al., 1991). Additionally, industry broad- was that of Gupta and Jones (1995), who interpreted commer- band MT data collected in 1984 by Phoenix USA along three cial MT data across the Purcell anticlinorium between 48°N and subparallel profiles (KCU-1, KCU-5, and KCU-9; Fig. 2) were 49.5°N. As part of the current study, we reanalyzed and inter- reanalyzed. The central of the three profiles overlaps the east end preted the southern three profiles from Gupta and Jones (1995). of the main profile, and these data were incorporated into our Through primarily 1-D inversion, the authors imaged a resistive analysis of this profile (Fig. 2). 1-D and limited 2-D analysis of upper crust of 2–6 km thickness above a widespread electrically this industry data was carried out by Gupta and Jones (1995). conductive “basement” extending to their depth of investigation Regional data used in this study include six long-period MT sites at midcrustal depths. They further concluded that rocks of the recorded in 2008 as part of the EarthScope MT Transportable upper and middle Belt Supergroup are more electrically conduc- Array (http://ds.iris.edu/spud/emtf). tive than the underlying Prichard Formation, an interpretation Together, this rich data set permits detailed characteriza- that we revisit in the section on “Identifying Belt Stratigraphy in tion along the main transect while maintaining the wide aperture the 2-D Resistivity Section.” needed to look at basin-scale structural variability. Figure 2B shows station locations overlain upon the intermediate-wave- COCORP Deep Seismic Data length isostatic residual gravity map of Mankinen et al. (2004) and illuminates major structural features including gravity highs Two 96 channel, deep seismic-reflection profiles (ID-2, over the Purcell and Sylvanite anticlines and lows over the Pur- MT-2 in Fig. 2) were previously collected (Yoos et al., 1991) cell and Rocky Mountain Trenches and the Libby syncline. At along two segments of the central MT profile. The data acqui- the 23 western sites, both long-period and broadband data were sition parameters for those profiles were given in Yoos et al. recorded. At the 11 eastern sites, which overlap with broadband (1991). A depth-migrated version of one of those profiles (MT-2) stations in Phoenix USA line KCU-5, only long-period data were and a line interpretation of it was illustrated in Yoos et al. (1991). collected. All sensors were oriented in geomagnetic north and We acquired the original seismic data package for both lines from east directions, and electric dipoles were on average 100 m in COCORP. The data for ID-2 were processed by the U.S. Geo- length. The broadband data along the main profile were collected logical Survey (USGS) Seismic Data Processing (SDP) group, at a sampling rate of 500 (8) Hz using induction coil (fluxgate) and a depth-migrated section was generated for ID-2. We sub- magnetometers and nonpolarizable lead-lead chloride electrodes. sequently developed a line interpretation of that depth-migrated Broadband stations recorded for 1–2 d, while long-period stations profile. Yoos et al. (1991) also presented a line interpretation of recorded for up to 2 wk; the resulting MT and vertical magnetic- an unmigrated seismic-reflection profile previously collected by field transfer functions are at most sites well defined to -10 4 Hz. the petroleum industry (Fig. 2; Signal of Montana WTF-82-1). Time-series processing was carried out using the multistation In addition to the reflection sections, seismic interval veloci- remote-reference transfer-function estimation program of Egbert ties were calculated by the USGS SDP group from the veloc- (1997). The MT and vertical magnetic field transfer functions for ity models picked by COCORP and used to stack the data. The all stations collected along the main profile, including time series stacking velocities were first smoothed using a horizontal aver- for the long-period data, are archived and available through the age over 50 common depth points (CDPs) and a vertical average Incorporated Research Institutions for Seismology (IRIS) Data over 500 ms of two-way traveltime; the resulting average stack- Management Center (http://www.iris.edu/dms/products). ing velocities were calculated at an interval of 50 CDPs. The Dix equation (Dix, 1955) was finally used to convert average stack- Previous MT Studies ing velocities into interval velocities, which can be more directly related to subsurface velocity variations. These seismic interval Early electrical studies identified high conductivity within or velocities, as well as line interpretations of reflection profiles below the lowest exposed Belt Supergroup in western Montana ID-2, MT-2, and WTF 82-1, are discussed further in the section and southern British Columbia. Audio-frequency magnetotellu- on “Geologic Interpretation of the MT Model.” ric (AMT) data, sensitive to the upper few kilometers, were collected by Wynn et al. (1977) along a 45 km transect crossing the MT Data Dimensionality Purcell anticlinorium at a location between the central and southern MT profiles (Fig 2). Those authors identified a 10–15-km- The dimensionality of MT data provides important con- wide zone of high conductivity just east of the hinge line, the top straints on the complexity of the subsurface. This information is of which was between 1 and 2 km below the surface exposure reflected in the symmetry of the impedance tensor,Z , as well as

of Prichard Formation. AMT studies atop the Sylvanite anticline in the vertical magnetic-field transfer function, Tz. Dimension- just west of the northern MT profile revealed similar high con- ality is important because it determines the level of complexity ductivity at a depth of just over 1 km (Long, 1988). Neither of required in modeling and inversion, which in turn determines

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin 315

the computational resources needed. Defining the dimensional- The PT and IV information can finally be used to extract ity of MT data and determination of so-called geoelectric strike quantitative information about dimensionality and geoelectric go hand in hand and are commonly determined statistically by strike direction. As illustrated in the inset (Fig. 4), each individ- examining phase tensors (Booker, 2014; Caldwell et al., 2004) ual PT and IV diagram, pertaining to a single period at a single and induction vectors (Schmucker, 1970) at multiple sites and station, contains directional information in map view. If suffi- frequency bands within the survey area. ciently 2-D, the direction of geoelectric strike is both perpendicu- In the case of a 1-D or layer-cake Earth model, where con- lar to the IV direction and parallel to one of the PT ellipse axes.

ductivity varies only with depth, phase tensors are circular, Tz Whether each diagram is considered 2-D is a function of y, the vanishes, and there exists no well-defined geoelectric strike. In a normalized skew angle (not shown), which is a measure of the 2-D Earth with a well-defined geologic strike, Z can be decom- asymmetry of the phase tensor. In determining geoelectric strike, posed into transverse magnetic (TM) and transverse electric (TE) we take the recommendation of Booker (2014), and consider modes in which electric currents flow across and along strike. |y| < 6° to be a necessary condition for a 2-D phase tensor. We Over a 2-D Earth, phase tensors plot as ellipses with geo- also require the maximum and minimum principal phase angles, electric strike parallel to one of the ellipse axes and the normal- which are related to the lengths of the major and minor ellipse ized skew angle y is, to within measurement errors, zero. The axes, respectively, to differ by at least 5°. This latter condition

real and imaginary parts of the complex vector Tz can be plotted ensures that circular, or 1-D, phase tensors do not bias any statis- as induction vectors, and, in the case of 2-D structure, they are tical determination of geoelectric strike. Finally, we apply these oriented perpendicular to geoelectric strike. In 3-D, the phase- conditions to only a subset of the data, ignoring the four shortest tensor skew parameter is statistically different from zero, and and four longest periods, as considerable scatter is evident in the the induction vectors do not exhibit a consistent direction either PTs and IVs at these periods (Fig. 4). across the survey or as a function of frequency; in such cases, Using this subset of the data, a histogram of y (Fig. 5A) the concept of geoelectric strike breaks down, and a 3-D analysis shows more than half the data fall outside the 2-D criteria of ±6°. becomes necessary. Based on this and the phase split criteria, less than one quarter Figure 4 shows phase tensors (PT) and real induction vec- of the data are consistent with a 2-D resistivity structure; 1/10

tors (IV) for all long-period stations along the main profile. This are consistent with 1-D structure (|y| < 6° and Fmax – Fmin < 5°); information is shown in pseudosection format, in which period and the majority of the data, including almost all data at periods serves as a (nonlinear) proxy for depth. The ellipses are colored greater than 100 s, are considered 3-D. Of those data considered according to the maximum principal phase, which is related to 2-D, a geoelectric strike direction between 0° and 10° east of subsurface resistivity. Phase values approaching 90° indicate the north (not shown) is found at short periods, while longer periods presence of anomalous conductive material in the subsurface. We reveal no distinct geoelectric strike direction. used the information contained in Figure 4 both qualitatively and The IVs, however, reveal a more consistent picture quantitatively to examine structure, dimensionality, and geoelec- (Fig. 5B). At all but the longest periods (1000–10,000 s), rose tric strike. histograms of IV direction, taking only IVs with magnitude Taken as a whole, the real IVs, which as displayed point greater than 0.1, suggest a geoelectric strike direction between toward regions of high electrical conductivity (Parkinson, 1962), N30°W and N50°W. This direction is consistent with both the show a strong east-west grain, particularly from 10 to 1000 s surface geologic strike, as defined by faults and folds, and the period, and point toward the center of the basin. This, together trend of potential field anomalies, as shown in Figure 2. At with a broad region where the maximum principal phase is above longer periods, the IVs rotate to N15°E. The smooth counter- 80°, indicates a large region of high conductivity beneath the cen- clockwise rotation of IVs with increasing period toward this ter of the profile. High phases extending to periods of 1000 s or direction suggests volumetric averaging at the longest-period more indicate that a portion of this high conductivity is deep, data between a region of high conductivity beneath the profile likely within the middle to lower crust. A decrease in the maxi- and a separate region of anomalous conductivity to the west or mum principle phase at the longest periods, particularly near the southwest of the profile. We do not speculate on the source of ends of the profile, may reflect igneous basement beneath the this latter high-conductivity region, but we note that 3-D inver- Belt Supergroup metasediments. sion of EarthScope MT data imaged a zone of high conductivity Abrupt changes in the IVs and PTs provide indications in the lower crust beneath the Coeur d’Alene mining district, of structural changes along the profile. In this way, absent any 100–150 km southwest of the central profile (Bedrosian and modeling assumptions, we directly interpret several such bound- Feucht, 2014). Given our analysis, is 2-D modeling and inver- aries to mapped faults or folds. Changes in IV direction, many sion appropriate? The induction vectors suggest that a 2-D accompanied as well by changes in the PT ellipticity, orientation, approach, with a geoelectric strike between N30°W and N50°W, and maximum principal phase, occur at or near the surface trace is valid at all but the longest periods. The phase tensors, how- of the Purcell Trench fault, the Moyie thrust, and the Pinkham ever, indicate a predominantly 3-D data set. Undoubtedly, this thrust. Additional changes in the PT parameters are seen along data set exhibits 3-D behavior; we are faced, however, with the the hinge lines of the Sylvanite and Purcell anticlines. oft-encountered question of whether the 3-D character is strong

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 316 Bedrosian and Box

Figure 4. Phase tensor ellipses (PT) and real induction vectors (IV) in pseudosection format. Individual PT + IV diagrams display information in geographic coordinates (inset). PTs are colored according to the maximum principal phase, and IVs point toward regions of enhanced conductivity (Parkinson, 1962). Abrupt changes in the orientation of IVs or the orientation, ellipticity, or principal phase of PTs are evident near fault zones and in areas that have undergone significant deformation. Colored geologic strip map is shown at top with lines above showing stratal dips. Abbreviations: PRC—Priest River metamorphic core complex; PA—Purcell anticlinorium; SA—Sylvanite anticline; PTF—Pur- cell Trench fault; MT—Moyie thrust; PT—Pinkham thrust.

enough to invalidate a 2-D inversion. We have chosen to take ing the data, we focus on the TM-mode data and have excluded a “trust, but verify” approach, in which we proceed with 2-D the TE-mode data, which are more sensitive to off-profile struc- inversion assuming a N30°W strike, but we take certain steps ture in the case of quasi–2-D resistivity models (Wannamaker to ensure that off-profile structure is not being mapped onto our et al., 1984). Finally, we carried out a series of 3-D inversions 2-D model section. First, we do not apply tensor decomposition on this data, including data from profiles KCU-1 and KCU-9 methods to this data set. The prevalence of data with |y| > 6° is and regional EarthScope MT data, in order to verify that basic inconsistent with these methods, which are based upon a 2-D model features are consistent between the 2-D and 3-D inver- distortion model with y = 0 (Booker, 2014). Second, in invert- sions. Note that computational limitations and the nonuniform

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin 317

Figure 5. (A) Histogram of normalized skew angle, y, determined from phase tensors in Figure 4. Red lines indicate |y| < 6°, which is taken as an indication of 2-D resistivity structure. Much of the data have |y| > 6°, suggesting 3-D character. (B) Rose histograms of induction vector direction at four different period ranges. At all but the longest periods, induction vectors align along a northeast/southwest corridor (gray bars), consistent with a geologic strike of N30°W to N50°W (green bars).

distribution of MT stations limit the spatial resolution of our model and a highly nonuniform data fit due to limitations in the 3-D inversions; it is for this reason that we invest in 2-D inver- data set and our ability to model it. Measured MT data, for exam- sion of this data set. ple, are sensitive to small-scale resistivity structure that we cannot model due to mesh size limitations, cannot constrain due to MT Modeling and Inversion limited data bandwidth and site spacing, and cannot recover due to the regularization required to solve the overparameterized 2-D 2D Inversion inverse problem. Furthermore, assumptions about the MT source We carried out 2-D inversion of the data along each profile fields are sometimes violated, giving rise to precise but inaccu- following rotation to N30°W, the predominant geologic strike of rate data. Following a series of tests with various error schemes, the region (e.g., Fig. 2). We focus our discussion upon our new 5% relative error was applied to the TM-mode apparent resistiv- MT profile, but the same inversion approach and final settings ity data, and 1.4° error was applied to the TM-mode phase data. were also applied to profiles KCU-1 and KCU-9. The 2-D finite- An absolute error of 0.01 was applied independently to the real

difference conjugate-gradient inversion RLM2DI (Rodi and and imaginary parts of Tz. To suppress 3-D effects during the Mackie, 2001), as incorporated within the Winglink software inversion, the TE-mode apparent resistivity and phase data were package, was employed. Within the profile area, an average cell excluded from the inversion. width of 800 m was used. Topography was included in the inver- Over 100 individual inversions were run with various sub- sion mesh; cell thickness started at 50 m at the land surface and sets of data, applied errors, regularization, and inversion param- increased logarithmically with depth. Where present, six decades eters. The final 2-D resistivity model was generated along the of data from 100 Hz to 10,000 s were inverted. Regularization main profile from the combined broadband and long-period MT was included using a Laplacian model structure function with a data as follows: An initial inversion was started from a 100 W·m trade-off parameter of 30. All inversions were run until conver- half-space using the aforementioned data and applied errors. gence, defined as the point at which the iterative change in data The resulting inverse model was fed into a subsequent inver- misfit fell below a threshold value. sion in which static-shift parameters were allowed to vary dur- The statistically derived data errors show a clear dependence ing the inversion. Finally, this inversion model was modified to on period, with larger errors at long periods, reflecting the small include a 1000 W⋅m resistor at 35–45 km depth. The modified number of transfer-function estimates available for averaging. model was used to start a final inversion to test whether the Direct application of these errors results in a very coarse inverse measured data are consistent with a resistive upper mantle or

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 318 Bedrosian and Box

whether high conductivity is required at these depths in order 3-D Inversion to fit the measured data. The a priori model was in all cases a The 3-D inversion was carried out to assess whether struc- homogeneous half-space. This final 2-D inverse model is given tures identified within the 2-D model section can be attributed to in Figure 8B. artifacts that stem from the assumptions inherent in 2-D inversion. We furthermore examined the 3-D inversion model to deter- Model Assessment mine the spatial extent of high conductivity within the Belt Basin. Pseudosections of the measured data and corresponding We inverted data using the 3-D method of Siripunvaraporn model response are shown in Figure 6. Note that even though et al. (2005), which seeks the smoothest, minimum structure the TE-mode data were excluded from the inversion, the model models subject to an appropriate fit to the data. A data-space response is shown. The final model represents a 78% reduction approach, where matrix dimensions depend on the size of the in normalized root-mean-square (n.r.m.s.) relative to the starting data set rather than the number of model parameters, surmounts 100 W⋅m half-space model (n.r.m.s. = 29.1). The model does the computational demands of construction and inversion of not, however, fit the data to within the specified errors. Given model-space matrices. The data used for inversion include the the applied errors, the n.r.m.s. statistic taken over all the inverted full impedance tensor at 12 periods ranging from 0.01 s to 6000 s;

data (static-shifted TM-mode apparent resistivity, ra, and TM- Tz data were not inverted. In total, 91 sites were used, includ-

mode phase, Tz) is 6.4, split evenly between TM-mode and the ing wide-band data from profiles KCU-1, KCU-5, and KCU-9

Tz data. It is our experience, however, with this implementation and six long-period EarthScope stations (Fig. 2). Computational of the RLM2DI inversion, that artificially small errors must be resources limited the size of the model mesh to 70 × 70 × 40

applied, particularly in Tz, to produce models with appreciable cells (model extent is shown in Fig. 1), resulting in a horizontal structure. cell size of 3 × 3 km. The thickness of cells increases logarithmi- Rather than discuss the n.r.m.s. statistic at these artificially cally with depth starting from 75 m at the surface. Both the data low error levels, we examine the “whiteness” of the fit, that is, and model grid were rotated to N30°W to ensure that regulariza- how uniformly the data are fit as a function of period and dis- tion was applied in a quasi–2-D coordinate system, as described tance along the profile. For this, we examine the n.r.m.s. assum- by Tietze and Ritter (2013). We applied errors of 2.5% to the ing more realistic errors of 20% in apparent resistivity (0.2 in magnitude of the off-diagonal (principal) impedance elements

ln[ra]), 5.7° in phase, and 0.05 in Tz. A plot of n.r.m.s. versus site and 5% to the magnitude of the diagonal tensor elements. Apply- (Fig. 7A) illustrates a uniform misfit across the profile. A few ing larger errors to the diagonal impedance elements is common sites are characterized by elevated n.r.m.s. in TM-mode apparent practice, given their generally smaller amplitude in comparison resistivity but not in phase. This is the result of static shifts in to the off-diagonal principal impedances. the measured data that are not well estimated during the inver- Based on the specified errors, the final 3-D inversion model sion. Similarly, small numbers of sites show elevated n.r.m.s in has a n.r.m.s. misfit of 11, which represents a 93% reduction in TM-mode phase but not in apparent resistivity. These correspond data misfit over the 100W ·m homogeneous half-space starting to sites where phases are greater than 90° at longer periods; as model (n.r.m.s. of 152). The same half-space model was also TM-mode phases are constrained to fall within the range of 0° used as a prior model. In this manner, the inversion seeks devia- to 90° (Weidelt and Kaikkonen, 1994), these data violate 2-D tions from the prior model only where required by the data, while assumptions and are hence poorly fit. Looking at n.r.m.s. as a regions poorly constrained by the data tend toward 100 W·m. function of period (Fig. 7B), we see that misfit is largely uni- Practically, this leads to model structure only in the vicinity of form with respect to frequency. An exception to this is TM-mode measurement locations. phase at the longest periods. We attribute the elevated n.r.m.s. to 3-D effects in the data, as evidenced by high values of y and out- DISCUSSION of-quadrant phases at the longest periods (Fig. 6). We note that our choice of error levels result in n.r.m.s. val- The 2-D model is characterized by strong variations in elec- ues that are close to 1 for the different data subsets. Thus, we con- trical resistivity both laterally and vertically. Two subhorizontal, clude that our model fits the measured data to 20% in TM-mode highly conductive horizons (Fig. 8B, C1 and C2) stand out in the

ra, 5.7° in TM-mode phase, and 0.05 in Tz. It is worth noting that model section. A deep highly conductive, concave-up layer (C1) even though we have excluded the TE-mode data from the inver- occurs at 25–35 km depth from just west of the crest of the Pur- sion, the calculated TE phase response of the final model fits the cell anticlinorium and abruptly ends ~20 km east of the Purcell TE-mode phase data to within 5.7° errors over much of the fre- Trench. From this point west to the Selkirk Range, the crust is quency range and at approximately half of the 54 sites (Fig. 7). very resistive (R1), save for some dipping conductive structures The largest mismatch between the measured and modeled TE- within the upper 5 km. A compact resistive feature (R4) is evident mode response is at long periods, where the high observed TE- from the surface to 5 km depth at ~20 km distance along line; this mode phase data are underfit by the 2-D model response. This feature is in the hanging wall of the Purcell Trench fault. Atop is consistent with a finite strike length to the conductive basin C1, there is a pronounced east-dipping resistor (R2) that reaches (Wannamaker et al., 1984). the surface along the western limb of the Sylvanite anticline and

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 319 ), shown as distance east from westernmost station. White regions denote White regions as distance east from westernmost station. response of final central line model ( Fig. 8B ), shown Figure 6. Pseudosections of measured data (left) and forward the and phase data were not inverted; electric (TE) apparent resistivity Only long-period data were collected east of km 120. Note that transverse from the inversion. data excluded magnetic-field transfer function. Tz—vertical magnetic mode; TM—transverse for comparison. TE-mode response of the final model is shown

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 320 Bedrosian and Box

Figure 7. Normalized root-mean-square (n.r.m.s.) statistic averaged over (A) all frequencies and (B) all sites, including 20 broadband sites from Phoenix USA, line KCU-5 where broadband data were not collected for this study (east of km 120). TE phase data were excluded from the inversion, but the forward response of the final mode is shown (dashed line). TE—transverse electric mode; TM— transverse magnetic mode; Tz—vertical magnetic-field transfer function.

extends to 25 km depth. R2 shows some internal variation, with stations shown in Figure 2; however, Gupta and Jones (1995) higher resistivity from the surface to ~10 km depth. A midcrustal identified high crustal conductivity in MT data to the north of horizon at 10 km depth (C2) begins west of the Whitefish Range our study area. We will first discuss the 3-D extent of the major front (120 km along line) and continues to the west for ~60 km, conductive features C1 and C2 and then look in more detail at the where it ultimately fades away. A more compact conductive fea- comparison between the 2-D and 3-D model profiles. ture (C3) is imaged at similar depth beneath the Libby syncline, C2 is prominent along all three cross sections, but its along- which at the surface is highly resistive (R3). A more subdued and section extent narrows to the northwest. We interpret C2 to be possibly discontinuous conductive layer (C4) is imaged at ~5 km continuous between the station profiles; the disconnected nature depth along the eastern third of the model section and is over- of C2 in the along-strike direction is a result of nonuniform sta- lain by resistor R5 along the Rocky Mountain Trench. The lower tion spacing. At upper-crustal depths, those portions of the model crust beneath the eastern half of the model section is moderately domain far from stations are unconstrained by the data and resistive (R6), though significantly less resistive than R1–R5. remain at or near the starting model (100 W·m). This is confirmed from a series of 3-D inversions with different starting models, Comparison of the 2-D and 3-D Resistivity Models where the resistivity between profiles is observed to scale with the resistivity of the start model. The 3-D inverse model both confirms and builds upon C2 as imaged in the 3-D inverse model is subhorizontal, with the 2-D resistivity model. The 3-D model cross section B–B′ the highest conductivity imaged between 10 and 20 km depth. (Fig. 9C) is coincident with the 2-D model profile and can be The top of C2 traces out the broad synclines and anticlines, com- directly compared to the 2-D model (Fig. 8B). Other cross sec- ing closest to the surface beneath the hinge line of the Purcell tions and depth slices through the 3-D model reveal the mini- anticline. An analysis of the mean conductivity of C2 is presented mum 3-D extent of the prominent conductive features C1 and in a later section, but we note here that the conductivity of C2 as C2 revealed by the 2-D model (Fig. 9). Our modeling cannot recovered from the 3-D model is significantly higher than that address whether these conductive features extend beyond the recovered in the 2-D model. This is to be expected, however,

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin 321

Figure 8. Geologic, resistivity, and seismic data along the magnetotelluric (MT) profile of this report (located in Fig. 2). (A) Geologic cross section down to ~5 km, adapted from Miller et al. (1999), Harrison et al. (1992), and Harrison and Cressman (1993), with units colored as in Figure 2A. Abbreviations above cross section are as in Figure 3 caption; ST—Snowshoe thrust. (B) Two-dimensional resistivity model down to 50 km depth (relative to sea level), with labeled conductors (C1–C4) and resistors (R1–R6) discussed in the text; dashed outlines with lowercase letter labels identify locations of examples of resistivity stratigraphy shown in Figure 10. MT stations are shown by small circles over profile: red—this report; white—Phoenix USA data from line KCU-5 of Gupta and Jones (1995). (C) Seismic constraints from Consor- tium for Continental Reflection Profiling (COCORP) lines ID-2 and MT-2, and Signal of Montana line WTF-82-1. Colors depict contoured seismic interval velocities obtained during depth migration calculations from original COCORP data. Short lines are reflection interpretations for depth-migrated seismic section of upper 20 km of line ID-2 (migration by U.S. Geological Survey; line interpretation, this report), depth- migrated seismic section of line MT-2 (migration and interpretation from Yoos et al., 1991), and unmigrated line drawing for the eastern half of line WTF 82-1 (interpretation from Yoos et al., 1991).

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 322 Bedrosian and Box

Figure 9. (A, B) Horizontal depth slices through the 3-D resistivity model. Note that the model grid is rotated to align with the regional strike direction (N30°W; location shown in Fig. 1). The interpreted minimum areal extents of C1 and C2 are denoted by black dashed lines. (C, D) Cross sections through the (C) 3-D resistivity model and (D) enlarged portion of cross section C–C′, with labeled conductors (C1, C2, and C4) and resistors (R1–R3 and R5–R6) discussed in the text. Model features are labeled as in Figure 8. White regions outlined in A and B, and white lines on B–B′ define averaging regions and are described in the text. White line in D denotes Sullivan horizon in Gibbs well as identified by Boberg et al. (1989).

as TE-mode data are most sensitive to the conductivity of such subhorizontal, concave-up geometry with a bulbous, very high- elongate structures and were included in the 3-D inversion but conductivity “knot” at its western end. In the 3-D model section excluded from the 2-D inversion. (B–B′), C1 has a similar though broader subhorizontal, concave- C1, as imaged in the 2D model, is distinctly deeper than up geometry, as well as a similar though shallower high-conduc- C2 and lies to the southwest of it. In the 2-D model, C1 has a tivity knot at its western end. The 3-D model section B–B′ has

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin 323

an additional bulbous high-conductivity knot in its central part, in Figure 8C, along with interpreted seismic reflections from which lies at roughly 25 km depth, but may extend downward those lines and from Signal of Montana line WTF82-1. Several close to the base of the crust. C1 is elongated in the along-strike features from the geologic section have recognizable expressions direction and is most prominent along cross sections A–A′ and in the MT and seismic data. In the following discussion, spe- B–B′ (both with two very high-conductivity knots), while fading cific features are referred to by their depth (with “km” following somewhat to the south. This variation may reflect the absence the number; e.g., 10 km) and horizontal location as measured on of long-period data along the southern line. The northern profile the x-axis scale in Figure 8 (with “km” listed first; e.g., km 27). does not extend far enough to constrain the northern limit of C1; The resistivity stratigraphy developed in this section is shown however, data from an isolated EarthScope MT station (Fig. 9B, in Figure 10, while the interpreted geology of the 2-D resistiv- northernmost station) suggest C1 may continue to the north. ity section is presented in Figure 11. The discussion in this and The connection between C1 and C2 varies along strike. C1 the following section can best be understood by referring to both and C2 are distinct within the northern part of the model and are Figures 8 and 11. separated from one another by roughly 20 km along profile A–A′. Units within the Belt Supergroup stratigraphy can be corre- Along profile B–B′, C1 and C2 are connected; however, the loci lated with distinctive resistivities, seismic velocities, and reflec- of high conductivity (<5 W·m) remain distinctively separated. tivities in Figure 8. Areas with Prichard Formation strata exposed Along profile C–C′, C1 is located beneath and partly overlap- at the surface are underlain by intermediate resistivities (~100 ping C2. This along-strike variation may reflect differences in the W·m), while exposures of Belt units stratigraphically higher than amount of shortening and/or extension that has occurred through- the Prichard Formation are underlain by very high resistivities out the region. Both the extent and configuration of C1 and C2 in (1000–10,000 W·m). The stratigraphic relationship between these the 3-D model are consistent with 3-D resistivity models derived two units is best seen in the upper 10–15 km below the Libby from EarthScope MT data. The studies of Bedrosian and Feucht syncline (km 70) and the Rocky Mountain Trench (km 115). In (2014) and Meqbel et al. (2014) imaged two highly conductive each case, the surface syncline of middle and upper Belt Super- tabular bodies beneath the Belt Basin, with the western conduc- group strata (Yb3—all formations above the Prichard Formation) tor at greater depth than the eastern conductor. The connection is reflected with an apparent syncline of highly resistive (1000– between these conductors also varies from north to south. 100,000 W·m) material. Exposures of Prichard Formation in the Examining cross section B–B′ in greater detail, resistive fea- Purcell anticlinorium and the Sylvanite anticline are underlain tures R3 and R5 beneath the Libby syncline and Rocky Moun- by more intermediate-resistivity material (100–200 W·m). Where tain Trench, respectively, are similar in depth and lateral extent seismic interval velocities are available across the Sylvanite anti- between the 2-D and 3-D models. Both features can be traced cline and the Libby syncline, this intermediate-resistivity layer some distance along strike. Conductive feature C3 is distinct from (Yb2) also correlates with relatively high seismic velocity and C2 in both the 2-D and 3-D models, but it is slightly shallower seismic reflectivity (Figs. 8A and 11A), reflecting the abundance in the 3-D model. R6 has a resistivity of ~100 W·m and occupies of mafic sills within the exposed Prichard Formation (Cook and a similar region in both 2-D and 3-D models. East-dipping fea- van der Velden, 1995). Yb2 is interpreted to pass under Yb3 in the ture R2 is imaged in both the 2-D model and 3-D cross section Libby syncline as a continuous layer. At 10 km below the Libby B–B′. It is somewhat more pronounced in the 2-D model, extend- syncline and the Rocky Mountain Trench, highly conductive lay- ing to 70 km along-line distance and to 25 km depth; in the 3-D ers (C3 and C2, respectively in Fig. 8B) are encountered. Simi- model, it extends to 60 km along-line distance and between 15 lar conductive layers (between 5 and 10 km thick) are imaged and 20 km depth. The greatest discrepancy between the 2-D and beneath the Rocky Mountain Trench and Purcell anticlinorium 3-D models, however, is observed at the west end of cross section on cross sections near the northern and southern MT lines KCU-1 B–B′, where resistor R1 extends throughout the crustal column and KCU-9 (Fig. 9C). We interpret these deep high-conductivity in the 2-D model. In the 3-D model, high resistivity (>1000 W·m) layers to represent strata in the lower part of the Belt stratigra- is limited to the upper 15 km, below which resistivity drops to phy (Yb1—strata below the mapped Prichard Formation). Pos- ~100 W·m, comparable to the resistivity of R6. Independent sup- sible causes of their high conductivity will be discussed in a later port for the validity of the 3-D model in this area is discussed section. This three-layer resistivity stratigraphy and its apparent later in the section “Projection of Surface Faults to Depth.” correlation with established Belt Supergroup stratigraphy are highlighted in Figure 10. Geologic Interpretation of the MT Model We note that our resistivity stratigraphy is in disagreement with that of Gupta and Jones (1995), who concluded that the Identifying Belt Stratigraphy in the 2-D Resistivity Section upper and middle Belt rocks (Yb3) are more electrically con- A geologic cross section based on surface geology (Miller ductive than the underlying Prichard Formation (Yb2). Those et al., 1999; Miller and Burmester, 2004; Harrison et al., 1992; authors based their conclusion upon an apparent correlation Harrison and Cressman, 1993) is shown above the 2-D resis- between surface geology and determinant phase pseudosections. tivity profile inFigure 8A. Profiles of contoured seismic inter- We take issue with this correlation for several reasons: First, in val velocities from COCORP lines ID-2 and MT-2 are shown all but a 1-D environment, the determinant phase mixes TE- and

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 324 Bedrosian and Box

Figure 10. Apparent resistivity stratigraphy and its correlation with Belt Supergroup stratigraphy, interpreted from three parts of the 2-D resistivity model profile. Location of each example is shown onFigure 8B, along with resistivity scale. Seismic interval velocity contours (m/s) from Consortium for Continental Reflection Profiling (COCORP) seismic line MT-2 (Fig. 8C) are shown on a and c. Sullivan horizon in b is discussed in text.

TM-mode phase data, confounding any simple relation between, represent a doubled-up section of unit Yb1—an upper Yb1 layer for example, high phase and low resistivity. Second, the regions continuous with the western “tail” of feature C2, faulted over a of high and low determinant phase, upon which those authors lower, autochthonous Yb1 layer in the lower part of the bulbous based their correlation, are at periods sensitive to ~5 km depth, part of C2. and thus cannot be directly related to surface geology, particu- In the deep crust (below 20 km) along the central part of larly in an area where significant tectonic transport has occurred, the MT model (between km 30 and km 80), a very distinct as our 2-D and 3-D models show. Finally, given the dimension- 5–10-km-thick synclinal layer of very high conductivity (C1) ality of the MT data (Figs. 3 and 4), our explicit 2-D and 3-D and intermediate seismic velocity is interpreted to be Yb1. It is inversions are taken to be more realistic than heuristic displays overlain by a 5–6-km-thick layer of intermediate resistivity and of measured data. high seismic velocities characteristic of Yb2. That in turn is over- At the east end of our MT profile, a broad area of the crust lain by a 5-km-thick layer (deeper part of R2) with very high below 8–10 km shows intermediate resistivities (R6, 50–500 resistivity and intermediate seismic velocities, characteristic of W·m) down to the projected base of the crust (~39 km). We inter- Yb3. This 15–20-km-thick three-layer sequence thus has similar pret that to represent Archean plutonic and metasedimentary resistivities and seismic velocities to those of the three-layer Belt rocks of the Medicine Hat Province (Xmh; Ross, 2002). That Supergroup sequence beneath the Libby syncline (Fig. 10). On intermediate-resistivity body is bounded to the west at km 120 its east flank near km 80, it contacts a broad intermediate-resis- by the eastern flank of a high-conductivity body (C2), with the tivity block, similar to Archean basement (R6) as interpreted east contact dipping west ~45° from 10 to 25 km depths. We inter- of km 120. We interpret this block similarly as Archean basement pret that contact to be the depositional nonconformity of high- nonconformably beneath the basal very conductive layer of the conductivity basal Mesoproterozoic Belt strata (Yb1) on a ramp three-layer Belt Supergroup. of Archean crystalline basement (Xmh). Cook and van der Velden (1995) observed a similar but less steep basement ramp under the Projection of Surface Faults to Depth Rocky Mountain Trench in seismic data 50 km to the north in Several faults of major (>4 km) interpreted dip-slip displace- Canada. The more bulbous eastern half of C2 is interpreted to ment are mapped at the surface along our MT profile (Fig. 2). In

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin 325

Figure 11. Interpreted 2-D resistivity model (see text for discussion). (A) Resistivity model (from Fig. 8B) overlain by contours of seismic interval velocities (from Fig. 8C) and by geologic line-work interpretation. Depositional and intrusive contacts are shown as white solid lines (dashed lines show inferred bedding); fault contacts are shown as heavy lines with movement arrows (red—Eocene normal fault, black—Jurassic–Paleocene thrust fault). Fault abbreviations are as in Figure 3 caption; LeT—Lewis thrust. Geologic unit abbreviations: Xmh—Archean gneiss of the Medicine Hat Province; Yb1—basal Belt Supergroup unit (underlies the Prichard Formation); Yb2—Prichard Formation of Belt Supergroup; Yb3—Ravalli, Piegan, and Missoula Groups of Belt Supergroup; Kg—Cretaceous unfoliated granitic rocks; Kgg—Cretaceous orthogneiss. (B) For clarity, geologic units in geologic interpretation of Figure 11A shown as colored polygons.

this section, we discuss the constraints provided by the resistivity mylonitic fabrics produced by the faults at depth. This discus- and seismic data on their apparent projections to depth. Since sion is based primarily on the data given in Figure 8, including we associate prominent resistivity layers with particular inter- the shallow geologic cross section, the 2-D resistivity profile, the vals within the Belt Supergroup stratigraphy (units Yb1, Yb2, contoured seismic interval velocities, and the seismic reflectivity. and Yb3), their terminations may represent fault offsets. Seismic The geologic interpretation of the main profile is presented as interval velocity lows may also represent highly fractured zones line work over the resistivity model together with the contoured along major faults. Discontinuities in patterns of seismic reflec- seismic interval velocities in Figure 11A. For clarity, the same tions can further indicate fault displacements, and in some cases, interpreted model is depicted in Figure 11B as a cross section of prominent seismic reflections can be produced by faults or by colored geologic strata.

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 326 Bedrosian and Box

The Pinkham thrust crops out at km 113, and, based on the 8 km below sea level. We interpret this contact as the subsurface intercept in the Gibbs #1 well (Boberg et al., 1989), it dips gently projection of the Eocene Purcell Trench detachment fault. We (~20°) to the west. It has no obvious expression near the surface would project this fault east to at least km 25, where the sharp in the resistivity model but is marked by changes in phase tensor vertical resistivity contrast ends at 10 km below sea level. At this and induction vector orientation (Fig. 4). We interpret it to project point, the Eocene Purcell Trench detachment fault is interpreted above the very conductive feature at the west end of conductive to cut and abruptly end the subsurface trace of the older Late feature C4 (Fig. 8B), and to continue with a gentle dip to the Cretaceous–Paleocene Moyie thrust. The trend of the gently east- western end of the long synclinal tail of feature C2, intercepting a dipping resistivity gradient continues into a similarly dipping pronounced low in the seismic interval velocities at ~12 km depth seismic velocity gradient, continuing east to km 33. at km 75. Similar relationships are seen in cross sections through Interpretation of the eastward continuation of the Purcell the 3-D MT model along the northern and southern MT profiles. Trench fault east of km 33 is less certain. The fault appears to We infer the Pinkham thrust to continue farther west under the trend into, but not offset, the vertically continuous R2 resis- C3 conductor to ~20 km depth at km 55, where it seems to end tive feature. However, we suggest the Purcell Trench fault does against the deep continuous eastern tail of the resistive R2 feature. continue across the narrow neck region of the R2 feature at The Libby thrust surfaces at km 70 in the middle of resistive 13–16 km depth, following a subtle resistivity gradient between feature R3, which we have identified as the Libby syncline. We the broad, very resistive upper half of the R2 feature and its nar- infer that the fault projects through the prominent seismic inter- rower, less-resistive lower half. At km 50, the fault projects along val velocity low at 10 km depth at km 64, passing slightly west of the sharp, east-dipping resistivity gradient at 17 km depth and the prominent C3 conductor and its gently west-dipping tail. The follows it eastward to ~20 km depth at km 70, where the R2 fea- moderately dipping east limb and near-vertical west limb of the ture pinches out. This position of the Eocene Purcell Trench fault Sylvanite anticline are interpreted to be abruptly cut by the Libby would explain the apparent downdip terminations of the Pinkham thrust dipping 30° to the west. The Libby thrust is interpreted to and Libby thrusts against the lower R2 resistive feature, offset project to 17 km depth at km 49, where it abuts against the deep by the younger detachment fault. The Purcell Trench fault trend continuous tail of the resistive R2 feature. continues eastward along a zone of strong seismic reflectivity The Moyie thrust occurs at the surface at about km 40, and through a seismic interval velocity low between km 70 and with Prichard Formation strata thrust eastward over upper Mis- 78 (Fig. 8C) to ~20 km depth at km 80, where the fault proj- soula Group units (Miller et al., 1999; Harrison and Cressman, ects across the abrupt updip termination of the east limb of the 1993). The resistivity profile shows a prominent contact between synclinally folded layer C1. Continuing that trend eastward, the intermediate-resistivity and very high-resistivity (R2) rock bod- fault projects to km 110 at 24 km depth across the abrupt down- ies dipping to the west at 30° down to at least 7 km, which we dip termination of the highly conductive layer (part of C2) that interpret as the subsurface expression of the Moyie thrust. Pro- we interpreted to represent a layer dipping 45° to the west of jection to greater depths is uncertain. The wide section of high lower Belt strata nonconformably overlying Archean basement. resistivity (R2) below the Moyie thrust and west of the Yb2-Yb3 The Purcell Trench fault can thus be traced in the subsurface for contact on the western flank of the Sylvanite anticline is thought ~100 km east of its surface exposure to a depth of ~24 km in the to consist entirely of Yb3. At the surface, where last seen beneath middle-lower crust. the Moyie thrust, bedding dips are near vertical with tops to the Above the Purcell Trench fault at 15–20 km depth between km west. If bedding is subvertical throughout that high-resistivity 60 and 100, an intermediate-resistivity horizon occurs below fea- block, then unit Yb3 has about twice the thickness that it does in tures C2 and C3 that are interpreted as Yb1 horizons. Since it the Libby syncline, or in the deep synclinal fold 20 km below the sits below Yb1, could it be a displaced sliver of its depositional Sylvanite anticline. We suggest that the Yb3 unit is folded into Precambrian crystalline basement (Xmh)? Cook and van der a relatively tight syncline, which is paired with a similarly tight Velden (1995) interpreted an allochthonous slab of crystalline Sylvanite anticline to the east. Such a syncline projects toward basement in a similar position in their seismic line 100 km to the this point below the Moyie thrust ~20 km south of our transect north (Fig. 3A), although later, more detailed work by Ainsworth (Fig. 2A). (2009) reinterpreted those rocks to be equivalent to our Yb2 unit The Eocene east-dipping Purcell Trench detachment fault (Prichard-Aldridge Formations). Along the COCORP seismic occurs at the surface at about km 11, dropping upper-plate sub- lines coincident with parts of our MT profile, Yoos et al. (1991) greenschist-facies Prichard Formation strata on the east, intruded interpreted a thinner basement slab at similar depth but slightly to by unfoliated granitic rocks, against lower-plate foliated granitic the west of the sub-Yb1 slab cited here (Fig. 3C). rocks and amphibolite-facies Belt Supergroup metasedimentary How do the resistivity and seismic-reflection data compare rocks on the west (Doughty et al., 1998). The resistivity profile between this sub-Yb1 region between km 60 and 100 and regions shows an east-dipping contact at that location between very high- we interpret as Archean crystalline basement? Their intermediate resistivity rocks (R1) to the west against rocks of intermediate resistivities are characteristic of both the shallow Yb2 in our pro- resistivity (Yb2) to the east. The contact is listric (downward- file and of the deeper Archean crystalline basement (R6) below flattening) with a dip of near 60° in the upper 3 km and ~15° at the Belt Supergroup along the east side of our profile. However,

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin 327

the sub-Yb1 region has seismic interval velocities similar to those complex. The strong reflector band is interpreted to represent the of Yb2 on the east limb of the Sylvanite anticline, and somewhat Spokane Dome mylonite zone, overlain by Hauser Lake Gneiss higher than those in the interpreted wedge of crystalline base- and underlain by a 5-km-thick zone of few seismic reflectors and ment below the Yb1 layer at 30–37 km depth between km 70 intermediate seismic velocities that we interpret as Archean crys- and 80 (Fig. 8C). The western end of the sub-Yb1 region shows talline basement. The Spokane Dome mylonite zone and under- prominent, closely spaced reflectors (Fig. 8C), a characteristic of lying Archean rocks crop out ~45 km to the south, dipping gently Yb2 here and in the work of Cook and van der Velden (1995). northward toward the MT and COCORP lines. Whether the Spo- On balance, the data support interpretation of this sub-Yb1 slab kane Dome mylonite zone represents an exhumed deep thrust as Prichard Formation (Yb2). Since it is overlain by the highly fault (Rhodes and Hyndman, 1984) or a deep-seated Eocene conductive layer (Yb1), we infer a thrust fault along that over- detachment fault (Doughty et al., 1998) is not resolved. The con- lying resistivity gradient, apparently connecting the subsurface tact at ~15 km depth between the nonreflective Archean crystal- traces of the Pinkham and Lewis thrusts (discussed later). line basement rocks above and more reflective, higher-velocity A broad area under the Purcell Trench fault below 10 km rocks below (interpreted above as Prichard-equivalent Yb2) is and west of km 25 consists of very resistive crust (R1) with interpreted to be a gently west-dipping thrust fault. little internal resistivity variation. Are the rocks there different than the layered strata to the east (e.g., plutonic?), or are they Restoration of the Deformed Cross Section the same strata that have somehow lost their conductive character? We suggest the latter, as crust there does maintain some In this section, we attempt to restore the interpreted crustal- seismic velocity layering (Fig. 8C) that can be interpreted as a scale geologic cross section (Fig. 11B), derived from interpre- westward continuation of the deep Belt stratigraphy below the tation of the 2-D resistivity model and seismic-reflection data Purcell Trench fault farther east. The intermediate-resistivity (Fig. 11A), to its predeformation state in two steps: (1) to its Yb2 layer above the western end of the deep Yb1 layer at km state at the end of the Paleocene, after Jurassic to Paleocene com- 30 appears to continue to the west as a similar layer thickness of pressional fold-and-thrust deformation, but prior to Eocene and high-seismic-interval-velocity material (20–26 km depth) over- younger extensional deformation, and (2) to its state in Jurassic lying a lower-velocity layer with thickness similar to Yb1. The time prior to the onset of compressional fold-and-thrust defor- Yb2 layer is arched into an anticline, the western limb of which mation. Figure 12A shows the restoration at end of Paleocene continues to almost 30 km depth at km 10, where it either ends time, while Figure 12B shows the restoration at the start of or bends upward farther west. We interpret it to end against the Jurassic time. For ease of discussion, all the blocks separated by east-dipping Kettle detachment fault, which, based on COCORP pre-Eocene faults are given labels in Figures 12A and 12B, from data to the west discussed earlier, projects into our profile at B1 (west) to B10 (east). about that position. The high-seismic-velocity Yb2 layer is overlain by a wedge of lower-velocity material at 17–20 km depth Reversal of Eocene and Younger Extensional Faulting between km 27 and 33 that is interpreted as the westward con- As discussed already, the geophysical expression of the Pur- tinuation of layer Yb3. That layer is overlain by higher-seismic- cell Trench fault is interpreted for at least 100 km to the east to velocity material along a contact that dips west and continues a depth of ~24 km at km 110 (Fig. 11). At that point, the fault is along the seismic velocity gradient along the top of the west- inferred to mark the downdip cutoff of the west-dipping, highly ward continuation of the Yb2 layer. We interpret this contact as conductive lowermost Belt unit (Yb1), interpreted as being in a west-dipping fault with Yb2 above it, ending downdip against depositional contact with Archean crystalline basement farther the east-dipping Kettle fault. A wedge of lower-seismic-velocity east. At about km 80, the fault is inferred to mark the updip cutoff material at ~20 km depth west of km 8 is interpreted as a wedge of a similar, deeper, west-dipping Yb1 layer. If we infer that the of Yb1 above the fault and below Yb2. Supportive evidence for interpreted Archean–lower Belt contact above the deep Purcell the continuation of Yb2 layers in the lower crust west of 25 km Trench fault is offset from the similar contact below the fault, is given by the coincident 3-D profile (B–B′ in Fig. 9C), which, then the Purcell Trench fault records top-to-the-east displacement while showing a significant increase in resistivity west of km 25 of ~26 km, which is significantly less than that inferred previ- and below 10 km, does suggest material of intermediate resistiv- ously (Harms and Price, 1992; Doughty and Price, 1999). In ity like Yb2 below 15 km. Figure 12A, we reverse the 26 km of top-to-the-east displace- Within the shallower Priest River complex, a prominent ment along the Purcell Trench fault and realign Yb1 below the band of seismic reflectors occurs at 7 km depth at the west end fault with Yb1 above the fault. This reconstruction approxi- of the profile (Fig. 8C), gradually deepening east to 10 km at mately aligns two thrust faults above the Purcell Trench fault about km 25, where it intersects the resistivity gradient that we with the inferred deep continuations below the fault: the Libby interpret to reflect the Purcell Trench fault. Hauser Lake Gneiss, and the Pinkham thrusts with interpreted faults above and below the metamorphosed equivalent of the Prichard-Aldridge Forma- the sliver of Archean basement within the Priest River complex. tions of the Belt Supergroup (Doughty et al., 1998), occurs at The reconstruction moves the Yb2 layer in block B7 above the the surface within the Selkirk Crest portion of the Priest River deep layer Yb3 in the core of the deep Belt Supergroup syncline,

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 328 Bedrosian and Box

Figure 12. Palinspastic reconstruction of interpreted geologic profile (see text for discussion). Fault and geologic unit abbreviations are the same as Figure 11, with two additional faults as noted. Fault-bounded blocks are given arbitrary labels (B1, B2, etc.). Whitefish Range (block B9) and Glacier National Park (B10) east of our profile were added to allow complete restoration of upper plate of Lewis thrust. (A) Restoration of geologic profile of Figure 11B to the end of Paleocene time, with motion on Eocene normal faults reversed. (B) Restoration of Paleocene-restored geologic profile (part A) to the beginning of Jurassic time, by reversal of motion of thrust faults in part A, and unfolding of strata on both flanks of Sylvanite anticline in block B5. Fault abbreviations are as in Figure 3 caption; geologic unit abbreviations as in Figure 11 caption.

suggesting that a thrust fault is still required to account for this basement underlies the Kettle fault at these depths. Reversal of older-over-younger contact. The east end of this thrust fault top-to-the-east motion on the Kettle detachment near the base of strand would connect into the downdip projection of the Lewis the crust would put Archean crystalline crust below the deep Yb1 thrust above the Purcell Trench fault. The west end of the thrust layer that now sits directly on the mantle between km 40 and 65 fault below this Yb2 layer connects to the thrust fault below block (Fig. 11). In Figure 12A, we reverse ~55 km displacement on B4, allowing us to identify that deep fault as the continuation of the deep Kettle detachment fault to put at least 3 km thickness of the Lewis thrust. Archean crystalline crust below the deep Yb1 layer. Reversal of 26 km of top-to-the-east displacement on the Pur- A major Tertiary west-side-down listric normal fault (Flat- cell Trench fault also returns a considerable thickness (10–12 km) head fault) bounds the Tertiary Kishenehn Basin between the of crust over the top of the Selkirk Crest region of the Priest River Whitefish Range and Glacier National Park ~35 km east of the complex prior to Eocene detachment faulting. This is compatible MT profile (Constenius, 1996), and these features are shown with peak metamorphic pressures of 500 MPa (~15 km burial) in Figure 12A. The fault was active from 49 to 20 Ma and is during peak metamorphism at ca. 72 Ma as indicated by meta- interpreted to record ~15 km of normal offset. Some authors morphic mineral assemblages in metapelites under the Purcell (Dahlstrom, 1970; Yoos et al., 1991; Constenius, 1996) have Trench fault ~15 km south of our MT transect (Doughty et al., interpreted the fault to flatten into the preexisting Lewis thrust 1998; Doughty and Price, 1999). Rapid uplift and cooling of the plane at 7–8 km below the surface dipping 3°–6° west. How- Priest River complex at ca. 50 Ma have been known for some ever, other authors (Mudge and Earheart, 1980; Whipple, 1992) time (Miller and Engels, 1975; Harms and Price, 1992; Doughty have interpreted the Flathead fault to cut and displace the Lewis and Price, 2000) and document the age of detachment faulting on thrust. Both interpretations result in upward displacement of the the Purcell Trench fault. Lewis fault east of the Flathead fault. We leave that relationship As mentioned earlier, the gently east-dipping Kettle fault unresolved in Figure 12A. We assume that the Flathead fault projects into the west end of our profile at a depth between 23 flattens into the Lewis thrust plane and continues downdip into and 27 km (Fig. 11). We added the gently east-dipping (12°) the Purcell Trench detachment fault. Then, of the 26 km of off- Kettle detachment fault to our interpretation beneath the Priest set suggested for the deep Purcell Trench fault, only ~13 km of River complex, suggesting it intersects the base of the crust at top-to-the-east displacement would be required on the upper end a depth of 34 km at km 40. We infer that Archean crystalline of the Purcell Trench fault west of that junction. This alternate

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin 329

interpretation would also require us to account for an additional Yb2 of B9. Block B8 is restored west of block B4, such that their 11 km of shortening in the prethrusting palinspastic reconstruc- Yb1 and Yb2 layers align. The restored separation of blocks B8 tion discussed in the following section. and B9 results from reversal of movement on thrust faults in the Whitefish Range (Fig. 2A). Reversal of Jurassic to Paleocene Folding and Thrusting The block overlying the Pinkham thrust (B6) restores Once we have reversed the offset of the Eocene and westward along the Pinkham thrust until its Yb2-Yb1 contact younger normal faulting from our interpreted geophysical profile is aligned with that along the west side of block B8 (~24 km). (Fig. 12A), the crustal section is restored approximately to its Archean crystalline rocks at the east end of block B3 are inter- configuration prior to Eocene extension (although post-Eocene preted to be nonconformable beneath the west end of Yb1 of erosion is not restored). Pre-Eocene crustal thickness was over block B6, so it is restored westward with the B6 block. 50 km west of the ramp in the crystalline basement at about km West of the Libby thrust, Belt strata are folded into west- 100 in Figure 12A, resulting from Jurassic–Paleocene folding vergent folds of increasing tightness to the west within block B5. and thrust duplication during Cordilleran shortening. In this sec- Unfolding and realigning of the Yb2 layer of the Libby syncline, tion, we attempt to remove the effects of this Jurassic to Paleocene Sylvanite anticline, and the syncline under the Moyie thrust shortening to generate a palinspastically restored cross section, restore ~25 km of shortening. The unfolded strata restore to a which assumes that the area of the section prior to shortening was nonconformable position above the Archean rocks of block B3. equal to that area after shortening (Dahlstrom, 1969; Hossack, The metamorphosed Yb2 within the Priest River complex (block 1979). In practical terms, this presumes that original bed lengths B2) sits below the Moyie thrust before Eocene extension in our have also been preserved. Since the structural styles of folds and interpretation. The upper contact of unit Yb2 above the Moyie thrust faults are well known in the adjacent Canadian Rockies thrust is eroded, but realigning it with that contact below the fault (Bally et al., 1966; Dahlstrom, 1970; Price, 1981; Boyer and requires a minimum of 12 km displacement. Elliott, 1982), we presume that the downdip continuation of this Total eastward displacement along the Lewis thrust, adding thrust system is deformed in a similar style. In this case, we have the displacements above, is ~221 km. A comparison of the far- interpreted that deformation to have deformed the three-layer thest west point of the transect above the Lewis thrust (west end resistivity stratigraphy of the Belt Supergroup and its noncon- of Yb2 of block B1) with an eastern point below the thrust (say, formably underlying Archean crystalline depositional basement. the easternmost autochthonous Yb1-Xmh contact) after thrusting We restore the Belt Supergroup to its prethrusting/prefolding (km 112 in Fig. 12A) and before thrusting (km 334 in Fig. 12B) geometry by reversing top-to-the-east displacement on the inter- yields a similar value of 222 km of shortening, or ~66% shorten- connected Lewis, Whitefish, Pinkham, Libby, and Moyie thrusts ing. Obviously, the reconstruction is greatly simplified and con- and by unfolding folded strata to a flat, subhorizontal geometry tingent on a number of uncertainties of interpretation. However, with correct stratigraphic stacking. Figure 12B was constructed our estimate of shortening compares favorably with several other by slicing up Figure 12A along faults or along fold axial planes, authors over this transect (~175 km for entire transect—Price and rotating and moving them as rigid blocks (numbered B1, and Sears, 2000; 135 km vs. our 160 km for shortening east of B2, etc., in Fig. 12) into predeformation positions as geometri- Pinkham thrust—Fuentes et al., 2012). It also compares well cally tight as possible with minimal overlap of blocks. It can be with shortening estimates in the Sevier thrust belt farther south in seen that the pieces cannot be fit tightly together, and we have central Utah-Nevada (DeCelles, 2004). not tried to deform the pieces to produce a better fit. We review Several features of the palinspastically restored basin and the restoration process and rationale moving from east to west, subsequent thrust belt are worth mentioning. The Belt Basin at from foreland to hinterland. As with most thrust belts (Boyer and this latitude was originally at least 410 km wide west to east, Elliott, 1982), it is presumed here that thrusts propagated to the adding Belt exposures west of the MT profile to the 345 km of east through time, younging from the hinterland to the foreland. restored width of the MT profile. This is roughly the size of the At the east end of the MT profile, the block (B8) beneath present-day Black Sea. Prior to shortening, the Belt strata are the Pinkham thrust (combined with the area east of our profile restored westward roughly to the present edge of continental in the Whitefish Range [block B9] and Glacier National Park basement, marked by the present-day initial Sr = 0.706 line 20 km [block B10] above the Lewis thrust) must be moved westward west of the Okanagan valley near Omak (Carlson et al., 1991). along the Lewis thrust so that the Yb2 layer in eastern block B10 Archean rocks represented by those in the Priest River complex is aligned against the west side of the deep autochthonous Yb2 are allochthonous, and they may have originated ~190 km to the layer (minimum of 160 km of displacement). The Yb2 of block west (at a crustal ramp in the Lewis thrust?) beneath the present- B7 must also be moved back to the west (along the Lewis thrust) day Okanagan metamorphic core complex. but less than block B10 (~70 km), fitting nicely under the thin The Lewis thrust system is dominated by a large-displace- Yb2 layer of block B10 and also against the deep autochthonous ment basal décollement (Lewis thrust) from which rise imbricate Yb2 layer. Block B4 is interpreted to have been contiguous with faults (from west to east, the Moyie, Libby, Pinkham, and several block B7 prior to Eocene displacement on the Purcell Trench Whitefish Range thrust faults) of lesser displacement that shorten fault (Fig. 12A), so it also moves 70 km to the west, overlain by the upper plate. The basal décollement climbs up section in the

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 330 Bedrosian and Box

direction of transport, as illustrated in Figure 12B. In the west, depth for ~30 km east of its surface exposure, we see evidence for the block of Archean crystalline rocks in the Priest River com- offsets along the projected subsurface trend of this fault for more plex (B3) is interpreted to overlie the westernmost segment of than 100 km east of its surface exposure. the Lewis thrust (km 35, Fig. 12B). The thrust appears to climb above the Archean basement into Yb1 at the base of block B6 Source of Enhanced Conductivity (km 65, Fig. 12B) and continues to propagate within Yb1 under blocks B8 and the western part of B4. The thrust climbs into Yb2 We turn now to the source of high conductivity within unit under the eastern part of block B4 (km 200, Fig. 12B) and con- Yb1. Enhanced conductivity within the Belt Basin was recog- tinues within Yb2 under block B7. West of block B4, a second nized early by Wynn et al. (1977) and Long (1988), which they branch of the fault climbs up section off the lower thrust into a attributed to metallic sulfides within the thick Belt sequence. décollement higher in Yb2, above blocks B4 and B7 and below However, these audio-frequency MT studies were quite limited blocks B9 and B10. The relative timing of motion on those two in areal and vertical extent, sensitive only to a depth of a few kilo- branches is not constrained. Those branches reunite at the east meters. Gupta and Jones (1995) used broadband MT data cover- end of block B7 at km 265 (Fig. 12B), where the thrust climbs ing a wider area than the earlier studies and with deeper response out of Yb2 and through Yb3, into or above the basal contact with (down to ~10 km) but provided only 1-D modeling and inversion the overlying Paleozoic section. save for a rudimentary 2-D inversion of profile KCU-1 (Fig. 2). In light of the long-period MT data presented here and of our 3-D Comparison with Previous Regional Crustal-Scale MT inversion model, we revisit the issue of the source of high Cross Sections electrical conductivity within the Belt Basin. We start by char- A comparison of our geologic cross section with that of Yoos acterizing the high-conductivity zones imaged and then examine et al. (1991) along the same transect, and with that of Cook and potential conductivity mechanisms in relation to the geologic van der Velden (1995) along a parallel transect 50 km north in and tectonic setting. In the following section, we speculate on the Canada is given in Figure 3. Several differences between the implications of our preferred interpretation regarding the early interpreted geologic section in this report from those earlier ones evolution of the Belt Basin. are readily apparent. Three aspects of the resistivity model shed light on the (1) All three profiles show autochthonous lowermost Belt cause of the high conductivity—the geometry of C1 and C2, strata below the basal Cordilleran thrust at and somewhat west of their mean and minimum conductivity, and their vertically a pronounced thinning of the crystalline basement (at ~125 km in integrated conductivity, or conductance. The high conductiv- each section of Fig. 3), resting nonconformably on Archean crys- ity of C1 and C2 continues for ~150 km along geologic strike; talline basement. In the other two profiles, those autochthonous this represents a minimum extent given our station coverage strata consist of 3–4 km of Yb1, while in our profile, this autoch- (Fig. 9). C2 is between 40 and 60 km wide in the across-strike thonous section consists of the full 20 km Belt section consisting direction and between 5 and 10 km thick. Note that the seg- of units Yb1, Yb2, and Yb3. mented nature of C1 and C2, in particular between the MT pro- (2) The other two sections show an allochthonous slab of files, stems from our limited ability to resolve structure away Archean crystalline basement of variable thickness ~15–20 km from stations. We presume the high conductivity of C1 and C2 below the Sylvanite anticline, while we find no evidence for that is in fact continuous throughout the dashed regions defined in in our resistivity model. However, Ainsworth (2009) provided Figure 9. In the 2-D model, C1 is comparable in geometry to evidence that the interpreted Archean slab in the Cook and van C2 but is more tube-like in the 3-D model. Regardless of which der Velden profile is actually imbricated Aldridge Formation with model is chosen, C1 is at least 40 km wide in the across-strike mafic sills (Yb2). We do interpret a thrust slab of Archean base- direction and ~10 km thick. These estimates suggest that C1 and ment farther west in the Priest River complex, where outcropping C2 cover minimal spatial extents of 3500 km2 and 8000 km2, Archean rocks south of our line appear to project into our profile respectively. As we interpret C1 and C2 to be offset segments (Doughty et al., 1998). of an originally continuous horizon, this suggests a minimum (3) Archean crystalline basement rocks are interpreted to original extent of the conductive horizon of nearly 12,000 km2. continue in the lower crust across both of the other two sections, Given the range of layer thickness estimates, this constitutes a while we have strong evidence that the basal strongly conduc- volume of 60,000 km3 to 120,000 km3 of electrically conductive Belt unit (Yb1) sits directly atop the seismically inferred tive metasediments at mid- to lower-crustal levels. By means of mantle with no intervening Archean basement in the middle of comparison, this volume is equivalent to 10%–20% of the water our transect. We interpret that to have resulted from low-angle in the present-day Black Sea. extensional displacement of that Belt section from the west off of The estimated resistivity of C1 and C2 provides additional its original Archean depositional basement. constraints on the cause of high conductivity. We used the resis- (4) While Cook and van der Velden (1995) did not image the tivity model to calculate the mean and minimum resistivity of C1 east-dipping Purcell Trench normal fault less than 50 km north and C2. For this calculation, we chose the 3-D model over its 2-D of our line, and Yoos et al. (1991) imaged that fault continuing at counterpart because the 2-D model does not use the TE-mode

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin 331

data in the inversion, so it will not accurately reflect the true sub- tectonic environments. For example, lower-crustal conductance surface resistivity. We must, however, take care when averaging within the extending Basin and Range province falls between model resistivity, as not all parts of the 3-D model are equally 750 and 3000 S (Wannamaker et al., 1997, 2008), while the well constrained. We limited our calculation to those model cells lower crust beneath the Tibetan and Altiplano Plateaus has been that fall within the broad confines of C1 and C2 (black dashed estimated at 6000 S (Li et al., 2003) and 20,000 S (Brasse et al., lines in Fig. 9), and that are also within 5 km of an MT station. 2002), respectively. The high conductance in all these cases is We further restricted this calculation to stations along our MT attributed to some combination of partial melt and saline fluids profile, where long-period data provide the most reliable con- exsolved from partial melt. straints on deep-crustal structure. Finally, in order to examine the With these constraints on the geometry, conductivity, and resistivity of C1 and C2 separately, we attributed that portion of conductance of C1 and C2, we examined the source of high con- the model from 5 km to 15 km depth to C2 and from 15 km to ductivity within the Belt Basin. Within continental crust, the most 25 km depth to C1. The spatial extents of averaging for C2 and common causes of anomalous bulk conductivity are aqueous flu- C1 are shown in white on Figures 9A and 9B, respectively; the ids, partial melt, graphite, and metallic sulfides. We examine each depth intervals chosen to separate C1 and C2 are shown sche- of these in turn. matically on profile B–B′ (Fig. 9C). The mean resistivity of C2, averaged over 655 model cells, Aqueous Fluids is 4.6 W⋅m, and, at one standard deviation, ranges from 3.3 W⋅m In considering aqueous fluids as an explanation for high to 6.4 W⋅m. The minimum resistivity within C2 is two orders of crustal conductivity, the conductivity of the fluids, the distribu- magnitude less (0.04 W⋅m), corresponding to a maximum con- tion of fluids, and the stability of fluids over geologic time scales ductivity of 24 S/m. C1, in contrast has a mean resistivity of 14.4 must all be considered. For aqueous fluids, the conductivity of the W⋅m averaged over 245 model cells, and ranges between 8.2 W⋅m fluid phase depends primarily on temperature and ionic strength. to 25.3 W⋅m at the one standard deviation level. The minimum For salt solutions, fluid conductivity increases with concentra- resistivity within C1 is again orders of magnitude lower than the tion until very high concentrations (30–40 wt%) are reached. At mean (0.08 W⋅m), with a peak conductivity of 12.6 S/m. In addi- a fixed concentration and (upper-crustal) pressure, fluid conduc- tion to the mean and maximum conductivity, we calculated the tivity also increases with temperature until 300–400 °C (Quist conductance of C1 and C2. In contrast to conductivity, which is and Marshall, 1968). As the onset of wet partial melting occurs an intrinsic physical property, conductance is an extrinsic prop- around 650 °C, this is likely the highest equilibrium tempera- erty and scales with the amount of conductive material present ture at which free aqueous fluids can exist. These data can be and its distribution. Conductance is furthermore one of the best- combined to estimate peak fluid conductivities of 1–30 S/m at resolved parameters in MT models (Jiracek et al., 1995) and per- upper- and midcrustal conditions. Archie’s law is commonly mits comparison of this study to MT modeling results from other used to relate fluid conductivity to bulk conductivity. For porosi- geologic settings. Using the same geometric boundaries to define ties expected at mid- to lower-crustal depths, bulk conductivity C1 and C2, we calculated a mean conductance for C2 at 12,100 is likely 1–2 orders of magnitude less than the pure fluid con- S, with over 70% of the area defined in Figure 9A (825 km2) ductivity. For example, to explain a 20 W·m lower crust, Hynd- having >1000 S conductance. Similarly, the mean conductance man and Shearer (1989) estimated between 0.5% and 3% pore of C1 is 5200 S, with over 60% of the area defined in Figure 9B waters, assuming seawater salinity. In the case of C1 and C2, (625 km2) having >1000 S conductance. We note that the esti- with a mean resistivity of 4–14 W·m, the required porosity would mated conductance likely underrepresents the true conduc- significantly higher. tance, as we have limited our calculation to 5–15 km for C2 and The source and stability of such fluids at depth are other 15–35 km for C1, and thus may not capture the full conductance considerations. Within active tectonic regimes, the explanation of these features. of conductivity anomalies by way of aqueous fluids is tenable. While we are confident in attributing C1 and C2 to metased- Fluids may, for example, derive from amphibolite to granulite iments of deep-basin origin, the estimated conductance values metamorphism in continental crust or directly from the mantle in are far greater than that of typical continental-basin deposits. A an extensional environment (Glover and Vine, 1995). Following global conductance map assembled by Everett et al. (2003) sug- the cessation of metamorphism or tectonic activity, however, the gests that conductance of the sedimentary section in continental residence time of deep-crustal water is on the order of 100 m.y. regions is almost always less than 500 S. While this map is based (Bailey, 1990; Thompson and Connelly, 1990). An additional only on sediment thickness data (and assumes a conductivity of argument against aqueous fluids in stable lower continental crust 0.03 S/m for continental sediments), it highlights the difference comes from petrology. Yardley and Valley (1997) argued that between our conductance estimates and more typical values. Our retrograde mineral reactions, occurring as rocks cool from their conductance estimates are also high in comparison to average original equilibrium temperatures, will consume any traces of lower-crustal conductance, which a compilation of MT models pore fluids remaining from peak metamorphic recrystallization. from Precambrian crust suggests is on the order of 20 S (Hynd- These arguments suggest a dry lower crust outside of active or man and Shearer, 1989). Notable exceptions are found in active recently active tectonic regimes.

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 332 Bedrosian and Box

Taken together, the mean conductivity of C1 and C2, as con- While we may have discrete fault zones where TOC has constrained by the 3-D resistivity model, would require several vol- verted to graphite, C1 and C2 are interpreted to be stratigraphi- ume percent of hypersaline fluids distributed in crack-like pore cally bound and of broad regional extent. We therefore find it geometry at depths in excess of 25 km. Any less-extreme fluid unlikely that metamorphic graphite is the sole source of their conductivity or pore geometry would only serve to increase the high conductivity. required porosity. This is at odds with petrology, which suggests In addition to metamorphic graphite, fluid-deposited graph- that beyond 0.1%, fluid storage in the lower crust over geologic ite may form from precipitation of reducing, carbon-bearing time spans is implausible (Yardley and Valley, 2000). (C-O-H) fluids. Fluid-deposited graphite requires a source of carbon, a means of mobilizing it, and a mechanism for pre- Partial Melt cipitating it. Possible carbon sources include carbon-bearing Neglecting compositional differences, the onset of dry metamorphic fluids, mantle-derived fluids, or incorporation of melting requires temperatures of ~1200 °C, while water-satu- sedimentary carbon within igneous fluids; cooling is the most rated rocks begin to melt around 650 °C (Lebedev and Khi- efficient means of graphite precipitation (Luque et al., 1995, tarov, 1964). The conductivity of the melt phase varies slightly 1998). With few exceptions, fluid-deposited graphite is thought with rock composition but typically falls within the range of to form in high-temperature (>550 °C) environments, where it 1–10 S/m (Tyburczy and Waff, 1983). This represents the maxi- occurs as vein-type deposits in granulite-facies metamorphic mum conductivity to expect for a pure melt phase. The bulk rocks or mafic-ultramafic rocks (Luque et al., 1998). Fluid-rock conductivity of partially molten rock additionally depends on interactions may also facilitate graphite formation by the replace- its distribution within the supporting rock matrix. Numerous ment of silicate minerals, typically in igneous rocks. While mixing models exist (e.g., Hashin and Shtrikman, 1962); how- exposed rocks of the Prichard Formation outside of the Priest ever, it is difficult to attain bulk conductivity in excess of 1 S/m River complex are typically greenschist to amphibolite facies, at lower-crustal pressures and temperatures without invoking the older sub-Prichard rocks (Yb1) have likely experienced water-saturated melt fractions in excess of 30% (Li et al., 2003). pressure-temperature conditions similar to metamorphosed Attributing C1 and C2 to partial melt would further suggest high Prichard Formation strata (Hauser Lake Gneiss; 675–930 °C heat flow and a reduction in seismic velocity, neither of which and 600–1100 MPa) within the Priest River complex (Doughty is observed. It would also give rise to a rheologically weak- and Price, 1999). Reduced C-O-H fluids may have existed early ened crust, a prediction at odds with the supported topography. in the basin history. Submarine-vent fluids, such as those that Finally, the lack of volcanism since the Eocene is inconsistent fed the Sullivan massive sulfide sedimentary exhalative deposit with a present-day magmatic system; the sheer extent of C1 and in Prichard time (Yb2), may have deposited graphite in older C2, if attributed to even a few percent partial melt, would put it sub–Prichard Formation sediments. Fluids derived from the on par with the largest magmatic systems on Earth. Therefore, intrusion of mafic sills into the Prichard Formation may also we rule out partial melt as the cause of high conductivity within have deposited graphite at deeper stratigraphic levels. Whether the Belt Basin. such fluid systems could deposit graphite over the basinwide scales suggested by the resistivity models is unknown. While we Graphite cannot rule out fluid-deposited graphite as an explanation for C1 Graphite is a highly conductive mineral; if interconnected, and C2, there remains no direct evidence of graphite within the less than 1% graphite can give rise to bulk conductivity in the Belt Supergroup. range of 1–10 S/m. Graphite is common in metasedimentary rocks, where it forms through the conversion of organic car- Metallic Sulfides bon via metamorphism. It is not uncommon for marine basin Similar to graphite, trace amounts of metallic sulfides can sediments to have a significant amount of total organic carbon dramatically lower bulk resistivity. The bulk conductivity of (TOC). While the TOC content of the sub-Prichard (Yb1) layer is sulfide-bearing rocks is highly variable, with pyrite ores ranging unknown, cuttings from the Gibbs well (Yb2) show TOC at levels from 0.1 S/m to 104 S/m and pyrrhotite ores ranging from 103 of 1%–3% (Boberg et al., 1989). If this carbon were converted S/m to 105 S/m (Parasnis, 1956). Additionally, under reducing into graphite, it could dramatically increase bulk conductivity. Is conditions, laboratory measurements on low-grade metamor- graphite a viable explanation for the high electrical conductivity phic black shales at simulated midcrustal conductions (300 °C in the deep sub-Prichard metasediments? and 250 MPa or ~10 km depth) show an order-of-magnitude The formation of metamorphic graphite requires extreme enhancement in conductivity over measurements at ambient tem- temperatures (>1000 °C), unless strain energy is present, which perature and pressure (Raab et al., 1998). We note that C1, C2, can catalyze graphitization at midcrustal conditions (>450 °C; and C3 are all at depths at or below 10 km, similar to conditions Ross and Bustin, 1990; Nover et al., 2005). Graphite observed in where laboratory studies showed conductivity enhancement in the deeper levels of fold-and-thrust belts, as well as in sutures and black shale. Raab et al. (1998) found the increase in conductiv- fossil shear zones, is assumed to have formed under such condi- ity at elevated temperature and pressure (~1 W⋅m) to be irrevers- tions but is typically limited in spatial extent to discrete zones. ible and attributed it primarily to elongation of the newly formed

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin 333

sulfides and to a lesser extent from the transformation of pyrite pyrrhotite, laid down in a reducing marine environment along a to pyrrhotite. Fabric is also speculated to play an important role series of discrete stratigraphic horizons, as the primary source of in the conductivity of carbon- and sulfide-bearing rocks. Folia- high conductivity in the sub-Prichard metasediments (Yb1). tion developed during burial metamorphism increases carbon and sulfide connectivity through the alignment, flattening, and Implications of High Conductivity for Early Basin elongation of minerals. The presence of metallic sulfides requires Evolution a method of emplacement (sources of metals and sulfur) and a reducing environment for their preservation. While the lack of exposure of Yb1 limits our understanding In a marine deep-water environment, sulfides and carbon of the basin in pre-Prichard time, studies of the Sullivan Zn-Pb- are often found together, where anoxic conditions can lead to Ag sedimentary exhalative massive sulfide deposit shed light on large-scale deposition of both organic carbon and sulfides from conditions within the Prichard Formation (Yb2) at the time of blue-green algae hosting sulfate-reducing bacteria (Goodfellow, this syndepositional mineralization. The Sullivan deposit formed 2000). The Prichard Formation was laid down in such an envi- in and around the vent field of a submarine hydrothermal system ronment, most likely in a stratified water column with anoxic around 1470 Ma. Emplacement of sills forced the expulsion of bottom conditions, as indicated by both sulfur/carbon ratios and large volumes of heated formation waters along synsedimentary sulfur isotope values (Lyons et al., 2000; Goodfellow, 2000). faults within the rift basin. Expelled from below, the ore-forming Disseminated sulfides and carbon are present in the exposed fluids were both chloride-rich and metalliferous. Most of the sul- Prichard Formation (Yb2), typically at the 1–3 wt% level based fur within the Sullivan deposit is thought to have been sourced upon cuttings from the Gibbs well (Boberg et al., 1989). In the from local reduction of seawater sulfate (Goodfellow, 2000). Canadian Belt, laminated sulfides, predominantly pyrrhotite, are Before examining C1 and C2, it is instructive to consider described within the Prichard and Aldridge Formations (Yb2). conductor C4 within Yb2 higher in the section and its associa- Hoy et al. (2000) described a series of thin-bedded to laminated tion, if any, with metallic sulfides. Cuttings from the Gibbs #1 pyrrhotite intervals within the Prichard Formation. One of these well (Fig. 2) reveal 1–3 wt% disseminated pyrite throughout the intervals, the so-called Sullivan horizon, correlates temporally sampled Prichard Formation (Boberg et al., 1989). This corre- with the Sullivan sedimentary exhalative massive sulfide deposit sponds to unit Yb2 in our resistivity stratigraphy, which is gener- and is up to 100 m thick proximal to the deposit. At a distance ally characterized by resistivities of 10–100 W·m, but which also of 5 km from the Sullivan deposit, Goodfellow (2000) examined contains the thin highly conductive interval C4 east of km 100 this interval and found 5–15 wt% pyrrhotite aligned along bed- (Fig. 8B). ding planes and with a laminated appearance. The surrounding The position of conductor C4 in the 3-D resistivity model sedimentary rock, in contrast, contained between 1% and 6% along the southern MT profile (line KCU-9 onFig. 2) can be pyrrhotite, largely disseminated in the rock mass. We conclude compared to resistivity at depth from geophysical logs of the that there is ample evidence throughout the basin for both dis- nearby Gibbs #1 well. Within the 5.4-km-deep well, the lowest- seminated and laminated pyrite and pyrrhotite within the Prichard resistivity values recorded occur over a 150 m interval centered at (Yb2) strata. The mean resistivity of C1 and C2 (sub-Prichard; 3.1 km depth, where the resistivity logs drop to 3 W·m. This inter- Yb1), however, is significantly lower than that of the Prichard val corresponds closely with the depth of C4 along cross section Formation (10–100 W·m), leading us to speculate that the sulfide C–C′ (Fig. 9D), which passes through the location of the Gibbs content of the lowest (unexposed) Belt Supergroup strata is either #1 well. Conductor C4 can be traced in the upper 5 km through- significantly higher or has a greater degree of connectivity than out the eastern part of the 3-D model area (below the Pinkham the sulfides within Yb2. thrust in the northern MT lines; above the Pinkham thrust in the In summary, the mean bulk conductivity values of C1 and southern MT lines), suggesting that, like C1 and C2, it repre- C2 are considerably higher than can be produced by aqueous sents a stratigraphic horizon. Can the position of C4 be related fluids given reasonable mid- to lower-crustal porosities. Like- to stratigraphy established elsewhere in the basin? Boberg et al. wise, we dismiss partial melt as an explanation for the high (1989) tied the stratigraphy penetrated by the Gibbs #1 well to conductivity given the high melt fractions required and the lack the regional stratigraphy of Finch and Baldwin (1984), based of elevated heat flow or recent magmatism expected to accom- on drill cuttings and borehole logs. This low-resistivity interval pany such an extensive region of crustal melt. We cannot rule out corresponds to the top of the lower Prichard Formation, which interconnected graphite seams as a cause, and some conductivity is also the stratigraphic position of the “Sullivan horizon.” The enhancement, particularly in proximity to fault zones, may be Sullivan horizon is a stratigraphic interval rich in laminated pyr- attributed to metamorphic graphite. Metamorphic temperatures rhotite that correlates with the synsedimentary Sullivan massive may have been high enough to form graphite in the deep crust sulfide deposit and has been identified up to 30 km south of the after Cordilleran deformation had thickened the crust, but con- Sullivan deposit (~120 km NW of the Gibbs #1 well). The asso- ductive bodies now in the middle or upper crust (C2, C3, and ciation between the thin conductor within Yb2 and the Sullivan C4) are unlikely to have ever experienced temperatures that high. deposit is intriguing, and it suggests that C4 may reflect a distal We favor laminated metallic sulfides, predominantly pyrite and (laminated?) sulfide blanket deposit associated with the Sullivan

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 334 Bedrosian and Box

or a related exhalative event. Subsequent burial metamorphism The conductivity and geometry of C1 and C2, in compari- is likely to have increased sulfide connectivity, forming an elec- son to C4, suggest either proximity to one or more source vent trically conductive horizon. The apparent folded and segmented fields or much larger or longer-lived exhalative events. Within nature of C4 (Figs. 8B, 9C, and 9D) is attributed to subsequent the study area, the shallower of the two conductive layers (C2) tectonic activity recorded by this horizon. We note, however, that remains beyond the reach of exploration, but its upper horizon although low resistivities are recorded in the geophysical logs, no comes within 5 km of the surface throughout parts of the 3-D increase in sulfide content is noted in the Gibbs well cuttings at model area (Fig. 9). A more regional MT survey may potentially this interval. This apparent discrepancy may reflect the sampling enlarge the mapped extent of C1 and C2, particularly in the north method or coarse sampling interval (10 ft); given the continu- and south directions. We suggest that future MT investigations ous sampling and larger averaging volume of the resistivity log, of the deepest Belt Supergroup strata and their underlying base- we interpret the low resistivity at the Sullivan interval to reflect ment should be focused on regions where C2 comes closest to the an increase in sulfide content or connectivity relative to the sur- surface. Such a location, if within the reach of a deep drill hole, rounding formation. would permit further in situ investigation into the cause of high We posit that an electrically conductive Sullivan horizon conductivity. may extend more widely across the northern part of the Belt- In suggesting that C1 and C2 are the signatures of mas- Purcell Basin than previously known. Support for this interpreta- sive sulfide exhalative events in the early basin history, we ask tion comes from the 3477-m-deep Duncan Energy Moyie #1 well the question of whether this is consistent with age constraints located north of the study area near the hinge of the Sylvanite on basin formation and estimated accumulation rates. As C2 is (Moyie) anticline (Fig. 1). A resistivity log from this well reveals higher in the crustal column, we are more confident in its thick- at least five discrete high-conductivity zones between 1 and 2 km ness estimate (5–10 km) than that of the deeper conductive depth. Similar to the association noted in the Gibbs #1 well, this horizon. Lydon (2000) extrapolated pre-Sullivan accumulation depth interval is coincident with the Sullivan horizon, as defined rates of between 25 cm/k.y. and 57 cm/k.y. based on a series of by marker sills identified within the Moyie #1 well. Cook and stratigraphically controlled age dates—the oldest constraints Jones (1995) further correlated this horizon and the interval of being the 1468 Ma age of the Moyie sills (Anderson and Parrish, high borehole conductivity with a conductor at 1 km depth in 2000) and the 1470 Ma age of the Sullivan deposit (Jiang et al., an MT resistivity model that intersects the Moyie #1 well. We 2000). From Cook and van der Velden (1995), the base of highly did not, however, image C4 within the western part of the 3-D reflective crust (which we interpret as the transition from Yb2 to resistivity model (Fig. 9), suggesting that the Sullivan horizon is Yb1) is ~3 km below the Sullivan horizon, as inferred by those somewhat limited in its westward extent. This observation is con- authors based upon correlation of their seismic-reflection section sistent with that of Hoy et al. (2000), who suggested that the Sul- and stratigraphy in the Moyie #1 well. Using the accumulation livan horizon may be limited to, or thicker within, basin-parallel rates from Lydon (2000) and the minimum thickness of 5 km grabens within the broader Belt Basin. for C2, we extrapolated minimum estimates of the age of the We conclude that the moderately conductive Yb2 (10–100 base of Yb1 (the onset of Belt sedimentation) between 1502 and W·m) contains a few percent disseminated sulfides, but that the 1484 Ma. The higher C2 thickness estimate of 10 km results in an more conductive C4 is attributed to laminated sulfide stringers older estimate for the beginning of Belt Supergroup deposition at at the Sullivan horizon (Hoy et al., 2000). This interpretation is 1522–1493 Ma. These dates are compatible with other estimates consistent with Goodfellow (2000), who noted the difference in for the onset of Belt-Purcell Basin sedimentation (1543–1497 sulfide content between laminated pyrrhotite at the Sullivan hori- Ma—Anderson and Parrish, 2000; 1510–1485 Ma—Lydon, zon and a lesser degree of disseminated sulfide in the surround- 2000) and consistent with the maximum age for onset for Belt ing rock. Stratiform or stringer sulfides are recognized at several sedimentation at 1576 Ma, as constrained by basement gneiss of intervals within the Prichard Formation and are believed to be the Priest River complex (Evans and Fischer, 1986). concentrated at the transitions between cycles of tectonic exten- Regional MT studies (Bedrosian and Feucht, 2014; Meq- sion and subsequent basin infilling (Hoy et al., 2000). Four such bel et al., 2014) provide a coarse image of crustal conductivity periods during Prichard Formation deposition are recognized and and suggest that the highest crustal conductivity, of which C1 are thought to reflect times of elevated fluid flow and geother- and C2 are part, occurs only west of the Rocky Mountain Trench mal gradients, accompanied by increases in up flow of structur- and the ramp in Archean basement below it (Xmh, Fig. 3). We ally controlled fluids into basin waters. Within this framework, speculate that the present location of the Rocky Mountain Trench we speculate that C1 and C2 reflect a series of laminated pyr- was the eastern edge of Belt deposition prior to deformation and rhotite intervals deposited by a series of pre-Sullivan exhalative that sub-Prichard deposition occurred over an attenuated crustal events during early extensional episodes. We emphasize that the section prior to significant magmatism, as evidenced by the low 5–10-km-thick conductive zones defined as C1 and C2 need not reflectivity of Yb1. Basaltic underplating during this time may be uniformly conductive; the measured MT response can be pro- have served to thermally weaken the crust, facilitating crustal duced by a series of thin (meter scale) discrete sulfide horizons attenuation and feeding hydrothermal fluids to submarine vent within otherwise resistive metasediments. systems that disseminated metallic sulfides within the early basin

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin 335

deposits (Yb1). Within this framework, high reflectivity within east-dipping continuation of the Eocene Kettle detachment fault the Prichard (Yb2) marks the onset of magmatic addition to the to be present in the deep crust beneath the Priest River meta- crust through a then-hot and mechanically weakened crust. morphic core complex and to continue to the base of the crust to the east. Farther east, the Kettle fault drops the entire Belt strati- CONCLUSIONS graphic section directly onto the underlying mantle, excising its original depositional basement of Archean crystalline rocks. We combined new broadband and long-period MT data col- Reversal of motion on these and other Eocene normal faults lected from 34 sites in northern Idaho and northwestern Montana reveals a crustal section greater than 50 km thick at the end of with previous broadband and EarthScope long-period MT data to Paleocene time and returns the Cordilleran thrust geometries to generate 2-D and 3-D resistivity models of the 35–40-km-thick their positions at the end of Jurassic–Paleocene shortening. Fur- crust in the region. These data were combined with previous ther reversal of apparent thrust displacements of the three-layer seismic-reflection profiles to constrain a geologic interpretation stratigraphy along the Lewis, Whitefish, Pinkham, Libby, and of the crustal section. The shallow portion of the resistivity model Moyie thrusts allows construction of a restored cross section correlates well with the mapped surface geology of the Belt prior to the onset of Cordilleran thrusting in the Jurassic. A total Supergroup. The uppermost unit (up to 5 km thick) consists of all of ~220 km of Jurassic–Paleocene shortening along these faults of the Belt Supergroup above the Prichard Formation, is highly is indicated. Allochthonous Belt Supergroup strata restore west resistive (typically 1000–10,000 W⋅m), and has relatively low of underlying autochthonous Belt strata that depositionally over- seismic layer velocities. The intermediate unit (up to 7 km thick) lie thinned crystalline basement along and west of a prominent consists of the exposed Prichard Formation and 3+ km of stratig- ramp in crystalline basement thickness beneath the present-day raphy below the deepest stratigraphic exposures of the unit. The Rocky Mountain Trench. Given the thickness of the sub-Prichard intermediate unit has low to moderate resistivity (30–200 W⋅m), unit and Prichard sedimentation rates, we suggest that crustal relatively high seismic velocities, and high seismic reflectivity, thinning and the onset of Belt Basin sedimentation occurred at with the latter two characteristics resulting from an abundance roughly 1503 ± 19 Ma. of thick syndepositional mafic sills. Locally, a narrow, highly conductive horizon occurs within this intermediate unit and cor- ACKNOWLEDGMENTS relates in stratigraphic position with the Sullivan horizon, a distal sulfide-rich blanket deposit broadly dispersed across much of the Louise Pellerin (Green Geophysics) and Victor Labson (U.S. northern Belt-Purcell Basin. This horizon is considered to be the Geological Survey [USGS]) were instrumental in conceiving result of hydrothermal venting that formed the world-class Sul- and championing this work. Bedrosian was initially funded as livan sedimentary exhalative massive sulfide deposit in southern a USGS Mendenhall Fellow. Long-period MT data were col- British Columbia. lected with instrumentation made available by John Booker Thick conductive horizons occur deeper in the resistiv- (University of Washington). Regional MT data were collected ity section and are interpreted to represent a single, structur- and made available as part of the National Science Founda- ally disrupted stratigraphic unit (5–10 km thick) with very low tion–funded EarthScope program. Reprocessing of COCORP resistivity (4–8 W⋅m), intermediate seismic velocities, and little seismic lines MT-2 and ID-2 was carried out by John Miller seismic reflectivity. This unit underlies the intermediate unit (but (USGS). We gratefully acknowledge the support of numerous is nowhere exposed) and is interpreted to sit nonconformably land owners, both public and private, who provided access to atop Archean crystalline basement of the Medicine Hat Province. their land. This work would not have been possible without By analogy with the Sullivan horizon, the thick conductive sub- enthusiastic field support of Louise Pellerin, Chrissy Ricks, Prichard unit may record repeated sulfide depositional events Rick West, and Jay Sampson. This manuscript was improved as within the early rift history of the basin. The lack of seismic a result of constructive reviews from Jared Peacock, Phil Wan- indications of mafic sills in this sub-Prichard unit suggests that, namaker, and an anonymous reviewer. unlike at the Sullivan deposit, causative hydrothermal fluids were not associated with interlayered mafic sills but may instead have REFERENCES CITED been associated with high heat flow due to basaltic underplating of attenuated continental crust. Ainsworth, H.L., 2009, Structure of the Moyie Anticline Delineated on a Grid of Reflection Seismic Profiles in Southeastern British Columbia [Master’s Identification of this three-layer Belt Supergroup resistiv- thesis]: Calgary, Alberta, Canada, University of Calgary, 71 p. ity stratigraphy allows us to recognize the structurally disrupted Anderson, H.E., and Davis, D.W., 1995, U-Pb geochronology of the Moyie stratigraphy in the resistivity section and map out the subsurface sills, Purcell Supergroup, southeastern British Columbia: Implications for the Mesoproterozoic geological history of the Purcell (Belt) Basin: continuations of surface faults, as well as positions of unexposed Canadian Journal of Earth Sciences, v. 32, p. 1180–1193, doi:10.1139/ faults. The downdip continuation of the Eocene, east-dipping e95-097. Purcell Trench detachment fault displaces the high-conductivity Anderson, H.E., and Parrish, R.R., 2000, U-Pb geochronological evidence for the geological history of the Belt-Purcell Supergroup, southeastern Brit- sub-Prichard layer ~26 km to the east in the deep (24 km) sub- ish Columbia, Chapter 7, in Lydon, J.W., Hoy, T., Slack, J.F., and Knapp, surface over 100 km east of its surface exposure. We interpret the M.E., eds., The Geological Environment of the Sullivan Deposit, British

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 336 Bedrosian and Box

Columbia: Geological Association of Canada, Mineral Deposits Division Cressman, E.R., 1989, Reconnaissance Stratigraphy of the Prichard Formation Special Publication 1, p. 113–126. (Middle Proterozoic) and the Early Development of the Belt Basin, Wash- Bailey, R.C., 1990, Trapping of aqueous fluids in the deep crust: Geophysical ington, Idaho, and Montana: U.S. Geological Survey Professional Paper Research Letters, v. 17, p. 1129–1132, doi:10.1029/GL017i008p01129. 1490, 80 p. Bally, A.W., Gordy, P.L., and Stewart, G.A., 1966, Structure, seismic data, and Dahlstrom, C.D.A., 1969, Balanced cross-sections: Canadian Journal of Earth orogenic evolution of southern Canadian Rocky Mountains: Bulletin of Sciences, v. 6, p. 743–757, doi:10.1139/e69-069. Canadian Petroleum Geology, v. 14, p. 337–381. Dahlstrom, C.D.A., 1970, Structural geology in the eastern margin of the Cana- Bedrosian, P.A., and Feucht, D.W., 2014, Structure and tectonics of the dian Rocky Mountains: Bulletin of Canadian Petroleum Geology, v. 18, northwestern United States from EarthScope magnetotelluric data: p. 332–406. Earth and Planetary Science Letters, v. 402, p. 275–289, doi:10.1016/ DeCelles, P.G., 2004, Later Jurassic to Eocene evolution of the Cordilleran j.epsl.2013.07.035. thrust belt and foreland basin system, western U.S.: American Journal of Boberg, W.W., Frodesen, E.W., Lindecke, J.W., Hendrick, S.J., Rawson, R.R., Science, v. 304, p. 105–168, doi:10.2475/ajs.304.2.105. and Spearing, D.R., 1989, Stratigraphy and tectonics of the Belt Basin of Deiss, C., 1935, Cambrian-Algonkian unconformity in western Montana: Geo- western Montana: Evidence from the Arco-Marathon No. 1 Paul Gibbs logical Society of America Bulletin, v. 46, p. 95–124, doi:10.1130/GSAB well, Flathead County, Montana, in French, D.E., and Grabb, R.F., eds., -46-95. Montana Geological Society 1989 Field Conference Guidebook, Mon- Dickinson, W.R., 2004, Evolution of the North American Cordillera: Annual tana Centennial Edition: Billings, Montana Geological Society, Geologic Review of Earth and Planetary Sciences, v. 32, p. 13–45, doi:10.1146/ Resources of Montana, Volume 1, p. 217–229. annurev.earth.32.101802.120257. Booker, J.R., 2014, The magnetotelluric phase tensor: A critical review: Sur- Dix, C.H., 1955, Seismic velocities from surface measurements: Geophysics, veys in Geophysics, v. 35, p. 7–40, doi:10.1007/s10712-013-9234-2. v. 20, p. 68–86, doi:10.1190/1.1438126. Box, S.E., Bookstrom, A.A., and Anderson, R.G., 2012, Origins of mineral Doughty, P.T., and Chamberlain, K.R., 1996, Salmon River Arch revisited: New deposits, Belt-Purcell Basin, United States and Canada: An introduction: evidence for 1370 Ma rifting near the end of deposition in the Middle Pro- Economic Geology and the Bulletin of the Society of Economic Geolo- terozoic Belt Basin: Canadian Journal of Earth Sciences, v. 33, p. 1037– gists, v. 107, p. 1081–1088, doi:10.2113/econgeo.107.6.1081. 1052, doi:10.1139/e96-079. Boyer, S.E., and Elliott, D., 1982, Thrust systems: American Association of Doughty, P.T., and Chamberlain, K.R., 2008, Protolith age and timing of Pre- Petroleum Geologists Bulletin, v. 66, p. 1196–1230. cambrian magmatic and metamorphic events in the Priest River complex, Brasse, H., Laezaeta, P., Rath, V., Schwalenberg, K., Soyer, W., and Haak, V., northern Rockies: Canadian Journal of Earth Sciences, v. 45, p. 99–116, 2002, The Bolivian Altiplano conductivity anomaly: Journal of Geophysi- doi:10.1139/E07-067. cal Research, v. 107, no. B5, 2096, doi:10.1029/2001JB000391. Doughty, P.T., and Price, R.A., 1999, Tectonic evolution of the Priest River Burchfiel, B.C., Cowan, D.S., and Davis, G.A., 1992, Tectonic overview of the complex, northern Idaho and Washington—A reappraisal of the Newport Cordilleran orogen in the western United States, in Burchfiel, B.C., Lip- fault with new insights on metamorphic core complex formation: Tecton- man, P.W., and Zoback, M.L., eds., The Cordilleran Orogen: Contermi- ics, v. 18, p. 375–393, doi:10.1029/1998TC900029. nous U.S.: Boulder, Colorado, Geological Society of America, Geology Doughty, P.T., and Price, R.A., 2000, Geology of the Purcell Trench rift val- of North America, v. G-3, p. 407–480. ley and Sandpoint Conglomerate: Eocene en echelon normal faulting and Cagniard, L., 1953, Basic theory of the magnetotelluric method of geophysical synrift sedimentation along the eastern flank of the Priest River metamor- prospecting: Geophysics, v. 18, p. 605–635, doi:10.1190/1.1437915. phic complex, northern Idaho: Geological Society of America Bulletin, Caldwell, T.G., Bibby, H.M., and Brown, C., 2004, The magnetotelluric phase v. 112, p. 1356–1374, doi:10.1130/0016-7606(2000)112<1356:GOTPTR tensor: Geophysical Journal International, v. 158, p. 457–469, doi:10 >2.0.CO;2. .1111/j.1365-246X.2004.02281.x. Doughty, P.T., Price, R.A., and Parrish, R.R., 1998, Geology and U-Pb geo- Campbell, M.R., 1914, The Glacier National Park: A Popular Guide to Its Geol- chronology of Archean basement and Proterozoic cover in the Priest River ogy and Scenery: U.S. Geological Survey Bulletin 600, 54 p. Carlson, D.H., Fleck, R., Moye, F.J., and Fox, K.F., 1991, Geology, geochem- complex, northwestern United States, and their implications for Cordille- istry, and isotopic character of the Colville igneous complex, northeastern ran structure and Precambrian continent reconstructions: Canadian Jour- Washington: Journal of Geophysical Research, v. 96, p. 13,313–13,333. nal of Earth Sciences, v. 35, p. 39–54, doi:10.1139/cjes-35-1-39. Cecile, M.P., Morrow, D.W., and Williams, G.K., 1997, Early Paleozoic (Cam- Egbert, G.D., 1997, Robust multiple-station magnetotelluric data processing: brian to Early Devonian) tectonic framework, Canadian Cordillera: Bul- Geophysical Journal International, v. 130, p. 475–496, doi:10.1111/j.1365 letin of Canadian Petroleum Geology, v. 45, p. 54–74. -246X.1997.tb05663.x. Chandler, F.W., 2000, The Belt-Purcell Basin as a low-latitude passive rift: Evans, K.V., and Fischer, L.B., 1986, U-Pb geochronology of two augen gneiss Implications for the geological environment of Sullivan type deposits, terranes, Idaho—New data and tectonic implications: Canadian Journal of Chapter 6, in Lydon, J.W., Hoy, T., Slack, J.F., and Knapp, M.E., eds., Earth Sciences, v. 23, p. 1919–1927, doi:10.1139/e86-179. The Geological Environment of the Sullivan Deposit, British Columbia: Evans, K.V., Aleinikoff, J.N., Obradovich, J.D., and Fanning, C.M., 2000, Geological Association of Canada, Mineral Deposits Division Special SHRIMP U-Pb geochronology of volcanic rocks, Belt Supergroup, west- Publication 1, p. 82–112. ern Montana: Evidence for rapid deposition of sedimentary strata: Cana- Chave, A.D., and Jones, A.G., 2012, The Magnetotelluric Method—Theory dian Journal of Earth Sciences, v. 37, p. 1287–1300, doi:10.1139/e00-036. and Practice: Cambridge, UK, Cambridge Press, 552 p., doi:10.1017/ Everett, M.E., Constable, S., and Constable, C.G., 2003, Effects of near-surface CBO9781139020138. conductance on global satellite induction responses: Geophysical Journal Constenius, K.N., 1996, Late Paleogene extensional collapse of the Cordille- International, v. 153, p. 277–286, doi:10.1046/j.1365-246X.2003.01906.x. ran foreland fold and thrust belt: Geological Society of America Bulletin, Fermor, P.R., and Moffat, I.W., 1992, The Cordilleran collage and the foreland v. 108, p. 20–39, doi:10.1130/0016-7606(1996)108<0020:LPECOT>2.3 fold and thrust belt, in MacQueen, R.W., and Leckie, D.A., eds., Foreland .CO;2. Basins and Fold Belts: American Association of Petroleum Geologists Cook, F.A., and Jones, A.G., 1995, Seismic reflections and electrical conductiv- Memoir 55, p. 81–105. ity: A case of Holmes’s curious dog?: Geology, v. 23, p. 141–144, doi:10 Fermor, P.R., and Price, R.A., 1987, Multiduplex structure along the base of .1130/0091-7613(1995)023<0141:SRAECA>2.3.CO;2. the Lewis sheet in the southern Canadian Rockies: Bulletin of Canadian Cook, F.A., and van der Velden, A.G., 1995, Three-dimensional crustal struc- Petroleum Geology, v. 35, p. 159–185. ture of the Purcell anticlinorium in the Cordillera of southwestern Can- Fillipone, J.A., and Yin, A., 1994, Age and regional tectonic implications of Late ada: Geological Society of America Bulletin, v. 107, p. 642–664, doi:10 Cretaceous thrusting and Eocene extension, Cabinet Mountains, north- .1130/0016-7606(1995)107<0642:TDCSOT>2.3.CO;2. west Montana and northern Idaho: Geological Society of America Bulle- Cook, F.A., Varsek, J.L., Clowes, R.M., Kanasewich, E.R., Spencer, C.S., Par- tin, v. 106, p. 1017–1032, doi:10.1130/0016-7606(1994)106<1017:AAR rish, R.R., Brown, R.L., Carr, S.D., Johnson, B.J., and Price, R.A., 1992, TIO>2.3.CO;2. Lithoprobe crustal reflection cross-section of the southern Canadian Cor- Finch, J.C., and Baldwin, D.O., 1984, Stratigraphy of the Prichard Forma- dillera: 1. Foreland thrust and fold belt to Fraser River fault: Tectonics, tion, Belt Supergroup, in Hobbs, S.W., ed., The Belt: Montana Bureau of v. 11, p. 12–35, doi:10.1029/91TC02332. Mines and Geology Special Publication 90, p. 5–7.

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin 337

Fritts, S.G., and Klipping, R.S., 1987a, Structural interpretation of northeastern Jiang, S.Y., Slack, J.F., and Palmer, M.R., 2000, Sm-Nd dating of the giant Sul- Belt Basin: Implications for hydrocarbon prospects: Oil & Gas Journal, livan Pb-Zn-Ag deposit, British Columbia: Geology, v. 28, p. 751–754, v. 85, no. 39, p. 75–79. doi:10.1130/0091-7613(2000)28<751:SDOTGS>2.0.CO;2. Fritts, S.G., and Klipping, R.S., 1987b, Belt Basin needs further exploration: Jiracek, G.R., Haak, V., and Olsen, K.H., 1995, Practical magnetotellurics in a Oil & Gas Journal, v. 85, no. 39, p. 93–95. continental rift environment, in Olsen, K.H., ed., Continental Rifts: Evo- Fuentes, F., DeCelles, P.G., and Constenius, K.N., 2012, Regional structure and lution, Structure, Tectonics: Amsterdam, Elsevier, Elsevier Developments kinematic history of the Cordilleran fold-thrust belt in northwestern Mon- in Geotectonics 25, p. 103–129. tana, USA: Geosphere, v. 8, p. 1104–1128, doi:10.1130/GES00773.1. Keller, G., 1987, Resistivity characteristics of geological targets, in Naibighian, Gaschnig, R.M., Vervoort, J.D., and Lewis, R.S., 2012, The Kaniksu and Idaho M., ed., Electromagnetic Methods in Applied Geophysics: Tulsa, Okla- batholiths, northern U.S. Cordillera: Close relatives or a case of mistaken homa, Society of Exploration Geophysicists, p. 13–51. identity?: Geological Society of America Abstracts with Programs, v. 44, Kleinkopf, M.D., 1997, Geophysical Interpretations of the Libby Thrust Belt, no. 7, p. 383. Northwestern Montana: U.S. Geological Survey Professional Paper 1546, Glover, P.W.J., and Vine, F.J., 1995, Beyond KTB—Electrical conductiv- 28 p. ity of the deep continental crust: Surveys in Geophysics, v. 16, p. 5–36, Lebedev, E.B., and Khitarov, N.I., 1964, Dependence on the beginning of melt- doi:10.1007/BF00682710. ing of granite and the electrical conductivity of its melt on high water Goodfellow, W.D., 2000, Anoxic conditions in the Aldridge basin during for- vapor pressure: Geochemistry International, v. 1, p. 193–197. mation of the Sullivan Zn-Pb deposit: Implications for the genesis of mas- Leech, G.B., 1962, Metamorphism and Granitic Intrusions of Precambrian Age sive sulfides and distal hydrothermal sediments, in Lydon, J.W., Hoy, T., in Southeastern British Columbia: Geological Survey of Canada Paper Slack, J.F., and Knapp, M.E., eds., The Geological Environment of the 62-13, 8 p. Sullivan Deposit, British Columbia: Geological Association of Canada, Li, S., Unsworth, M.J., Booker, J.R., Wei, W., Tan, H., and Jones, A.G., 2003, Mineral Deposits Division Special Publication 1, p. 218–250. Partial melt or aqueous fluid in the mid-crust of southern Tibet? Con- Gough, D.I., 1986, Seismic reflectors, conductivity, water and stress in the constraints from INDEPTH magnetotelluric data: Geophysical Journal Inter- tinental crust: Nature, v. 323, p. 143–144, doi:10.1038/323143a0. national, v. 153, p. 289–304, doi:10.1046/j.1365-246X.2003.01850.x. Guéguen, Y., and Palciauskas, V., 1994, Introduction to the Physics of Rocks: Link, P.K., Fanning, M.K., Lund, K.I., and Aleinikoff, J.N., 2007, Detrital-zir- Princeton, New Jersey, Princeton University Press, 392 p. con populations and provenance of Mesoproterozoic strata of east-central Gupta, J.C., and Jones, A.G., 1995, Electrical conductivity structure of the Pur- Idaho, U.S.A., in Link, P.K., and Lewis, R.S., eds., Proterozoic Geology cell anticlinorium in southeast British Columbia and northwest Montana: of Western North America and Siberia: Society for Sedimentary Geology Canadian Journal of Earth Sciences, v. 32, p. 1564–1583, doi:10.1139/ (SEPM) Special Publication 86, p. 101–128, doi:10.2110/pec.07.86.0101. e95-127. Link, T.A., 1949, Interpretations of foothills structures, Alberta, Canada: Amer- Haak, V., and Hutton, V.R.S., 1986, Electrical resistivity in continental lower ican Association of Petroleum Geologists Bulletin, v. 33, p. 1475–1501. crust, in Dawson, J.B., ed., The Nature of the Lower Continental Crust: Long, C.L., 1988, Two Audio-Magnetotelluric Profiles across the Sylvanite Geological Society [London] Special Publication 24, p. 35–49. Anticline, Idaho and Montana: U.S. Geological Survey Open-File Report Harlan, S.S., Heaman, L.M., LeCheminant, A.N., and Premo, W.R., 2003, Gun- 88-658, 13 p. barrel mafic magmatic event: A key 780 Ma time marker for Rodinia plate Lund, K., Aleinikoff, J.N., Evans, K.V., duBray, E.A., Dewitt, E.H., and Unruh, reconstructions: Geological Society of America Bulletin, v. 31, p. 1053– D.M., 2010, SHRIMP U-Pb dating of recurrent Cryogenian and Late 1056. Cambrian–Early Ordovician alkalic magmatism in central Idaho: Impli- Harms, T.A., and Price, R.A., 1992, The Newport fault: Eocene listric normal cations for Rodinian rift tectonics: Geological Society of America Bul- faulting, mylonitization, and crustal extension in northeast Washington and letin, v. 122, p. 430–453, doi:10.1130/B26565.1. northwest Idaho: Geological Society of America Bulletin, v. 104, p. 745– Luque, F.J., Rodas, M., and Barrenechea, J.F., 1995, Fluid-deposited graphite 761, doi:10.1130/0016-7606(1992)104<0745:TNFELN>2.3.CO;2. mineralization in different geologic settings, in Pasava, J., Kribek, B., and Harrison, J.E., and Cressman, E.R., 1993, Geology of the Libby Thrust Belt of Zak, K., eds., Mineral Deposits: From Their Origins to Their Environ- Northwestern Montana and Its Implications to Regional Tectonics: U.S. mental Impacts: Rotterdam, The Netherlands, A.A. Balkema, p. 597–600. Geological Survey Professional Paper 1524, 42 p. Luque, F.J., Pasteris, J.D., Wopenka, B., Rodas, M., and Barrenechea, J.F., Harrison, J.E., Cressman, E.R., and Whipple, J.W., 1992, Geologic and Struc- 1998, Natural fluid-deposited graphite: Mineralogical characteristics and ture Maps of the Kalispell 1° × 2° Quadrangle, Montana, and Alberta and mechanisms of formation: American Journal of Science, v. 298, p. 471– British Columbia: U.S. Geological Survey Miscellaneous Investigations 498, doi:10.2475/ajs.298.6.471. Series Map I-2267, scale 1:250,000. Lydon, J.W., 2000, A synopsis of the current understanding of the geological Hashin, J., and Shtrikman, S., 1962, A variational approach to the theory of environment of the Sullivan deposit, Chapter 3, in Lydon, J.W., Hoy, T., effective magnetic permeability of multiphase materials: Journal of Slack, J.F., and Knapp, M.E., eds., The Geological Environment of the Applied Physics, v. 33, p. 3125–3131, doi:10.1063/1.1728579. Sullivan Deposit, British Columbia: Geological Association of Canada, Hossack, J.R., 1979, The use of balanced cross-sections in the calculation of Mineral Deposits Division Special Publication 1, p. 12–31. orogenic contraction: A review: Journal of the Geological Society [Lon- Lyons, T.W., Luepke, J.L., Schriber, M.E., and Zieg, G.A., 2000, Sulfur geo- don], v. 136, p. 705–711, doi:10.1144/gsjgs.136.6.0705. chemical constraints on Mesoproterozoic restricted marine deposition: Hoy, T., Anderson, D., Turner, R.J.W., and Leitch, C.H.B., 2000, Tectonic, mag- Lower Belt Supergroup, northwestern United States: Geochimica et Cos- matic, and metallogenic history of the early synrift phase of the Purcell mochimica Acta, v. 64, p. 427–437, doi:10.1016/S0016-7037(99)00323-3. Basin, southeastern British Columbia, Chapter 4, in Lydon, J.W., Hoy, T., Mankinen, E.A., Hildebrand, T.G., Zientek, M.L., Box, S.E., Bookstrom, A.A., Slack, J.F., and Knapp, M.E., eds., The Geological Environment of the Carkson, M.H., and Larsen, J.C., 2004, Guide to Geophysical Data for the Sullivan Deposit, British Columbia: Geological Association of Canada, Northern Rocky Mountains and Adjacent Areas, Idaho, Montana, Wash- Mineral Deposits Division Special Publication 1, p. 32–60. ington, Oregon, and Wyoming: U.S. Geological Survey Open-File Report Huebschman, R.P., 1973, Correlation of fine carbonaceous bands across a 2004-1413, 31 p. Precambrian stagnant basin: Journal of Sedimentary Petrology, v. 43, McMechan, M.E., and Price, R.A., 1982, Superimposed low-grade metamor- p. 688–699. phism in the Mount Fisher area, southeastern British Columbia—Implica- Hurich, C.A., Smithson, S.B., Fountain, D.M., and Humphreys, M.C., 1985, tions for the East Kootenay orogeny: Canadian Journal of Earth Sciences, Seismic evidence of mylonite reflectivity and deep structure in the Kettle v. 19, p. 476–489, doi:10.1139/e82-039. dome metamorphic core complex, Washington: Geology, v. 13, p. 577– Meqbel, N.M., Egbert, G.D., Wannamaker, P.W., Kelbert, A., and Schultz, A., 580, doi:10.1130/0091-7613(1985)13<577:SEOMRA>2.0.CO;2. 2014, Deep electrical resistivity structure of the northwestern U.S. derived Hyndman, R.D., and Hyndman, D.W., 1968, Water saturation and high electri- from 3D inversion of USArray magnetotelluric data: Earth and Planetary cal conductivity in the lower continental lower crust: Earth and Planetary Science Letters, v. 402, p. 290–304, doi:10.1016/j.epsl.2013.12.026. Science Letters, v. 4, p. 427–432, doi:10.1016/0012-821X(68)90017-4. Miller, F.K., and Burmester, R.F., 2004, Geologic Map of the Bonners Ferry 30′ Hyndman, R.D., and Shearer, P.M., 1989, Water in the lower continental crust: × 60′ Quadrangle, Idaho and Montana: U.S. Geological Survey Miscel- Modelling magnetotelluric and seismic reflection results: Geophysical laneous Field Studies Map 2426, scale 1:100,000, 30 p. Journal International, v. 98, p. 343–365, doi:10.1111/j.1365-246X.1989 Miller, F.K., and Engels, J.C., 1975, Distribution and trends of discordant ages .tb03357.x. of the plutonic rocks of northeastern Washington and northern Idaho: Geo-

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 338 Bedrosian and Box

logical Society of America Bulletin, v. 86, p. 517–528, doi:10.1130/0016 Shankland, T.J., and Ander, M.E., 1983, Electrical conductivity, temperature, -7606(1975)86<517:DATODA>2.0.CO;2. and fluids in the lower crust: Journal of Geophysical Research, v. 88, Miller, F.K., Burmester, R.F., Powell, R.E., Miller, D.M., and Derkey, P.D., p. 9475–9484, doi:10.1029/JB088iB11p09475. 1999, Geologic Map of the Sandpoint 1° × 2° Quadrangle, Washington, Siripunvaraporn, W., Egbert, G., Lenbury, Y., and Uyeshima, M., 2005, Three- Idaho, and Montana: U.S. Geological Survey Open-File Report 99-144, dimensional magnetotelluric inversion: Data-space method: Physics of the scale 1:250,000, 77 p. Earth and Planetary Interiors, v. 150, no. 1–3, p. 3–14, doi:10.1016/j.pepi Mudge, M.R., 1970, Origin of the Disturbed Belt in northwestern Montana: Geo- .2004.08.023. logical Society of America Bulletin, v. 81, p. 377–392, doi:10.1130/0016 Thompson, A.B., and Connelly, J.A., 1990, Metamorphic fluids and anomalous -7606(1970)81[377:OOTDBI]2.0.CO;2. porosities in the lower crust: Tectonophysics, v. 182, p. 47–55, doi:10 Mudge, M.R., and Earheart, R.L., 1980, The Lewis Thrust Fault and Related .1016/0040-1951(90)90341-5. Structures in the Disturbed Belt, Northwestern Montana: U.S. Geological Tietze, K., and Ritter, O., 2013, Three-dimensional magnetotelluric inversion in Survey Professional Paper 1174, 18 p. practice—The electrical conductivity structure of the San Andreas fault in Nover, G., Stoll, J.B., and von der Gönna, J., 2005, Promotion of graphite central California: Geophysical Journal International, v. 195, p. 130–147, formation by tectonic stress—A laboratory stress experiment: Geo- doi:10.1093/gji/ggt234. physical Journal International, v. 160, p. 1059–1067, doi:10.1111/j.1365 Tikhonov, A.N., 1950, On the determination of the electric characteristics of -246X.2005.02395.x. deep layers of the Earth’s crust: Doklady Akademii Nauk SSSR, v. 73, Palacky, G.J., 1987, Resistivity characteristics of geological targets, in Naibig- p. 295–297. hian, M., ed., Electromagnetic Methods in Applied Geophysics: Tulsa, Tyburczy, J.A., and Waff, H.S., 1983, Electrical conductivity of molten basalt Oklahoma, Society of Exploration Geophysicists, p. 53–129. and andesite to 25 kbar pressure: Geophysical significance and impli- Parasnis, D.S., 1956, The electrical resistivity of some sulphide and oxide cations for charge transport and melt structure: Journal of Geophysical minerals and their ores: Geophysical Prospecting, v. 4, p. 249–278, doi: Research, v. 88, p. 2413–2430, doi:10.1029/JB088iB03p02413. 10.1111/j.1365-2478.1956.tb01409.x. van der Velden, A.J., and Cook, F.A., 1994, Displacement of the Lewis thrust Parkinson, W., 1962, The influence of continents and oceans on geomagnetic sheet in southwestern Canada: New evidence from seismic reflection data: variations: Geophysical Journal International, v. 6, p. 441–449, doi:10 Geology, v. 22, p. 819–822, doi:10.1130/0091-7613(1994)022<0819:DOT .1111/j.1365-246X.1962.tb02992.x. LTS>2.3.CO;2. Parrish, R.R., Carr, S.D., and Parkinson, D.L., 1988, Eocene extensional tecton- Wannamaker, P.E., Hohmann, G.W., and Ward, S.H., 1984, Magnetotelluric ics and geochronology of the southern Omineca belt, British Columbia and responses of three-dimensional bodies in layered earths: Geophysics, Washington: Tectonics, v. 7, p. 181–212, doi:10.1029/TC007i002p00181. v. 49, no. 9, p. 1517–1533, doi:10.1190/1.1441777. Potter, C.J., Sanford, W.E., Yoos, T., Prussen, E., Keach, I.I.R., Oliver, J.E., Wannamaker, P.E., Doerner, W.D., Stodt, J.A., and Johnston, J.M., 1997, Sub- Kaufman, S., and Brown, L.D., 1986, COCORP deep seismic reflection dued state of tectonism of the Great Basin interior relative to its eastern traverse of the interior of the North American Cordillera, Washington and margin based on deep resistivity structure: Earth and Planetary Science Idaho: Tectonics, v. 5, p. 1007–1025, doi:10.1029/TC005i007p01007. Letters, v. 150, p. 41–53, doi:10.1016/S0012-821X(97)00076-9. Wannamaker, P.E., Hasterok, D.P., Johnston, J.M., Stodt, J.A., Hall, D.B., Price, R.A., 1981, The Cordilleran foreland thrust and fold belt in the southern Sodergren, T.L., Pellerin, L., Maris, V., Doerner, W.M., Groenewold, Canadian Rocky Mountains, in McClay, K.J., and Price, N.J., eds., Thrust K.A., and Unsworth, M.J., 2008, Lithospheric dismemberment and mag- and Nappe Tectonics: Geological Society [London] Special Publication 9, matic processes of the Great Basin–Colorado Plateau transition, Utah, p. 427–448, doi:10.1144/GSL.SP.1981.009.01.39. implied from magnetotellurics: Geochemistry Geophysics Geosystems, Price, R.A., and Sears, J.W., 2000, A preliminary palinspastic map of the Meso- v. 9, no. 5, Q05019, doi:10.1029/2007GC001886. proterozoic Belt-Purcell Supergroup, Canada and USA: Implications for Wei, W., Unsworth, M., Jones, A., Tan, H., Nelson, D., Booker, J., Chen, L., the tectonic setting and structural evolution of the Purcell anticlinorium Li, S., Solon, K., Bedrosian, P., Jin, S., Deng, M., Ledo, J., Kay, D., and and the Sullivan deposit, Chapter 5, in Lydon, J.W., Hoy, T., Slack, J.F., Roberts, B., 2001, Widespread fluids in the Tibetan crust: Science, v. 292, and Knapp, M.E., eds., The Geological Environment of the Sullivan p. 716–719, doi:10.1126/science.1010580. Deposit, British Columbia: Geological Association of Canada, Mineral Weidelt, P., and Kaikkonen, P., 1994, Local 1-D interpretation of magnetotellu- Deposits Division Special Publication 1, p. 61–81. ric B-polarization impedances: Geophysical Journal International, v. 117, Quist, A.S., and Marshall, W.L., 1968, Electrical conductivity of aqueous p. 733–748, doi:10.1111/j.1365-246X.1994.tb02466.x. sodium chloride solutions from 0 to 800° and at pressures to 4000 bars: Whipple, J.W., 1992, Geologic Map of Glacier National Park, Montana: U.S. Journal of Physical Chemistry, v. 72, p. 684–703. Geological Survey Miscellaneous Investigations Map I-1508F, scale Raab, S., Hoth, P., Huenges, E., and Müller, H.J., 1998, Role of sulfur and 1:100,000. carbon in the electrical conductivity of the middle crust: Journal of Geo- Willis, B., 1902, Stratigraphy and structure, Lewis and Livingston Ranges, physical Research, v. 103, p. 9681–9689, doi:10.1029/97JB03530. Montana: Geological Society of America Bulletin, v. 13, p. 305–352, Rehrig, W.A., Reynolds, S.J., and Armstrong, R.L., 1987, A tectonic and geo- doi:10.1130/GSAB-13-305. chronologic overview of the Priest River crystalline complex, northeast- Winston, D., 2007, Revised stratigraphy and depositional history of the Hel- ern Washington and northern Idaho, in Schuster, J.E., ed., Selected Papers ena and Wallace Formations, Mid-Proterozoic Piegan Group, Belt on the Geology of Washington: Washington Division of Geology and Supergroup, Montana and Idaho, in Link, P.K., and Lewis, R.S., eds., Earth Resources Bulletin 77, p. 1–14. Proterozoic Geology of Western North America and Siberia: Society for Rhodes, B.P., and Cheney, E.S., 1981, Low-angle faulting and the origin of Sedimentary Geology (SEPM) Special Publication 86, p. 65–100. Kettle dome, a metamorphic core complex in northeastern Washington: Winston, D., and Link, P.K., 1993, Middle Proterozoic rocks of Montana, Geology, v. 9, p. 366–369, doi:10.1130/0091-7613(1981)9<366:LFATO Idaho, and Washington: The Belt Supergroup, in Reed, J., Bickford, M.E., O>2.0.CO;2. Houston, R.S., Link, P.K., Rankin, R.W., Sims, P.K., and Van Schmus, Rhodes, B.P., and Hyndman, D.W., 1984, Kinematics of mylonites in the W.R., eds., Precambrian of the Conterminous United States: Boulder, Priest River metamorphic core complex, northern Idaho and northeastern Colorado, Geological Society of America, Geology of North America, Washington: Canadian Journal of Earth Sciences, v. 21, p. 1161–1170, v. C-2, p. 487–521. doi:10.1139/e84-122. Wynn, J.C., Kleinkopf, M.D., and Harrison, J.E., 1977, Audio-frequency mag- Rodi, W., and Mackie, W.L., 2001, Nonlinear conjugate gradients algorithm netotelluric and gravity traverse across the crest of the Purcell anticlino- for 2-D magnetotelluric inversion: Geophysics, v. 66, p. 174–187, doi:10 rium, northwestern Montana: Geology, v. 5, p. 309–312, doi:10.1130/0091 .1190/1.1444893. -7613(1977)5<309:AMAGTA>2.0.CO;2. Ross, G.M., 2002, Evolution of Precambrian continental lithosphere in Western Yardley, B.W.D., and Valley, J.W., 1997, The petrologic case for a dry lower Canada: Results from Lithoprobe studies in Alberta and beyond: Cana- crust: Journal of Geophysical Research, v. 102, p. 12,173–12,185, doi:10 dian Journal of Earth Sciences, v. 39, p. 413–437, doi:10.1139/e02-012. .1029/97JB00508. Ross, J.V., and Bustin, R.M., 1990, The role of strain energy in creep graphi- Yardley, B.W.D., and Valley, J.W., 2000, Reply to “Comment on ‘The pet- tization of anthracite: Nature, v. 343, p. 58–60, doi:10.1038/343058a0. rologic case for a dry lower crust’ by Bruce W.D. Yardley and John Schmucker, U., 1970, Analysis of Geomagnetic Variations in the Southwestern W. Valley”: Journal of Geophysical Research, v. 105, p. 6065–6068, United States: Scripps Institution of Oceanography Bulletin 13, 165 p. doi:10.1029/1999JB900323.

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Highly conductive horizons in the Mesoproterozoic Belt-Purcell Basin 339

Yin, A., Kelty, T.K., and Davis, G.A., 1989, Duplex development and aban- Zirakparvar, N.A., Vervoort, J.D., McClelland, W., and Lewis, R.S., 2010, donment during evolution of the Lewis thrust system, southern Glacier Insights into the metamorphic evolution of the Belt-Purcell Basin; evi- National Park, Montana: Geology, v. 17, p. 806–810, doi:10.1130/0091 dence from Lu-Hf garnet geochronology: Canadian Journal of Earth Sci- -7613(1989)017<0806:DDAADE>2.3.CO;2. ences, v. 47, p. 161–179, doi:10.1139/E10-001. Yoos, T.R., Potter, C.J., Thigpen, J.L., and Brown, L.D., 1991, The Cordilleran foreland thrust belt in northwestern Montana and northern Idaho from COCORP and industry seismic reflection data: American Association of Manuscript Accepted by the Society 8 December 2015 Petroleum Geologists Bulletin, v. 75, p. 1089–1106. Manuscript Published Online 6 April 2016

Printed in the USA

Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021 Downloaded from http://pubs.geoscienceworld.org/books/book/chapter-pdf/980412/spe522-12.pdf by guest on 25 September 2021