A kinematic model linking the Sevier and Laramide belts in the Idaho-Montana fold-thrust belt, U.S. Cordillera

Significant spatial and temporal overlap of thin- and thick-skinned thrusts has been described for many global orogenic belts, with structural style defining a continuum between thin- and thick-skinned end members (Lacombe and Bellahsen, 2016; Pfiffner, 2017; Butler et al., 2018; Horton and Folguera, 2022). Thin-skinned thrusts generally have low-angle décollements that follow weak stratigraphic contacts within sedimentary rocks of the upper crust (e.g., Bally et al., 1966; Dahlstrom, 1970; Boyer and Elliott, 1982). Thin-skinned thrusts rarely carry thin slivers of basement rocks to form structural culminations (e.g., DeCelles, 2004; DeCelles and Coogan, 2006). Thick-skinned thrusts cut across primary lithologic contacts (i.e., mechanical basement) at moderate to high angles, with décollements in the middle crust or deeper (e.g., Blackstone, 1940; Smithson et al., 1979; Erslev, 1993; Lacombe and Bellahsen, 2016; Pfiffner, 2017). In part based upon these observations, many workers have recognized that a primary global control on orogenic structural style is the variable crustal rheology into which continental fold-thrust belts propagate (e.g., Allmendinger et al., 1983; Kulik and Schmidt, 1988; Allmendinger and Gubbels, 1996; Kley et al., 1999; Pearson et al., 2013; McGroder et al., 2015; Lacombe and Bellahsen, 2016; Pfiffner, 2017; Butler et al., 2018; Parker and Pearson, 2021).

In the U.S. Cordillera (part of the North American Cordillera), early workers recognized a marked contrast between the Sevier and Laramide belts and interpreted that the two orogenic domains are distinct in style, with distinct spatial distributions, durations, and tectonic causes (e.g., Armstrong, 1968). This “fundamental” distinction led many workers to interpret episodes of flat-slab subduction as the primary driver of thick-skinned deformation (Lipman et al., 1971; Burchfiel and Davis, 1975; Coney and Reynolds, 1977; Dickinson and Snyder, 1978; Jordan and Allmendinger, 1986; Saleeby, 2003; Yonkee and Weil, 2015; Heller and Liu, 2016; Copeland et al., 2017). While alternative drivers, such as increased convergence rates, have been proposed, the primary focus has been on the influence that the subducting oceanic plate may have on the structural style of the fold-thrust belt in the overriding plate (e.g., Burchfiel et al., 1992; Yonkee and Weil, 2015). As a consequence, the potential kinematic continuity between end-member thin- and thick-skinned thrusts has been largely unexplored. As a result, this “lower-plate” perspective has obscured the role that crustal rheology may have had on the style and tempo of mountain building in the North American Cordillera.

The Late Jurassic to Paleogene North American Cordillera is a classic example of an ancient ocean-continent subduction system that displays end-member thin- and thick-skinned structural styles (e.g., Yonkee and Weil, 2015). While much work supports the interpretation that thin-skinned Sevier belt deformation is generally west of the thick-skinned Laramide belt and began much earlier at most latitudes, some segments of the Cordillera—such as within the Idaho-Montana and Mexican fold-thrust belts—are characterized by a spatial overlap and concurrent timing of thrusting within the Sevier and Laramide belts (e.g., Kulik and Schmidt, 1988; Yonkee and Weil, 2015; Fitz-Díaz et al., 2018; Parker and Pearson, 2021). The significant overlap in space and time suggests that the range of structural styles observed in the Idaho-Montana segment of the Sevier-Laramide fold-thrust belt—similar to those of other global orogenic belts (e.g., Lacombe and Bellahsen, 2016; Lescoutre and Manatschal, 2020; Tavani et al., 2021)— formed during progressive deformation, with kinematic linkages between thin- and thick-skinned structures (Kulik and Schmidt, 1988; Parker and Pearson, 2021; Parker et al., 2022). Though kinematic models linking thin- to thick-skinned thrusting end members are common in global orogenic belts, such as the Zagros (e.g., Koshnaw et al., 2020) and Andes (e.g., Fuentes et al., 2016), only a few workers (Kulik and Schmidt, 1988; O’Neill et al., 1990; McClelland and Oldow, 2004; Erslev et al., 2022) have attempted to integrate the Sevier and Laramide thrust systems into a kinematically linked fold-thrust belt. The Idaho-Montana fold-thrust belt is an ideal study location not only because of its overlapping structural styles, but also because it contains potentially the oldest documented thick-skinned thrusts in the Cordilleran foreland (e.g., Carrapa et al., 2019; Garber et al., 2020) and broad (~75 km) basement-involved folds and thrusts in the hinterland that may share a mid-crustal detachment with the foreland (Parker and Pearson, 2021; Parker et al., 2022).

In this contribution, we present the first balanced and sequentially restored cross section that spans from the interior, thin-skinned fold-thrust belt of the hinterland to the end-member thick-skinned uplifts of the foreland, encompassing the full range of structural domains in the Jurassic to Paleogene retroarc fold-thrust belt of the North American Cordillera (Fig. 1). The shortening estimate provided by our cross section also helps fill a long-standing data gap within the North American Cordillera (Elison, 1991). We present an internally consistent kinematic model that suggests that variations in structural style were largely controlled by the prior distribution of sedimentary cover rocks that limited the development of detachment horizons in the upper crust. From this model, we offer hypotheses regarding the relative timing of activity and shortening magnitudes of particular thrusts and the Idaho-Montana fold-thrust belt as a whole. Our results fill a significant gap in our understanding of the North American Cordillera and test recent tectonic-inheritance models (e.g., Lacombe and Bellahsen, 2016; Parker and Pearson, 2021; Tavani et al., 2021) that describe the process by which strain is transferred from the upper to the middle crust during shortening of former rift margins, thereby producing self-organized changes in structural style and irregular space-time patterns. Our model builds upon decades of work in the Sevier and Laramide belts by linking them together with a kinematically balanced section. This allows us to discuss the role that rheology plays in determining not only the style but also the spatiotemporal patterns of orogenic wedges without requiring changes in plate-boundary geodynamics.

The study area lies within the Idaho-Montana segment of the North American Cordillera, between the Helena and Wyoming salients, in a region generally referred to as the southwestern Montana recess (Fig. 2). Like all segments in the orogenic belt, the southwestern Montana recess accommodated major horizontal shortening during Late Jurassic to Eocene ocean-continent (Andean-style) subduction (e.g., Armstrong, 1968; Burchfiel and Davis, 1972, 1975; DeCelles, 2004; Yonkee and Weil, 2015). The structural style of the Sevier fold-thrust belt is widely described as thin skinned because, in general, décollements are bedding-parallel within pre-orogenic strata of the region (Armstrong, 1968; Royse et al., 1975; Price, 1981; Royse, 1993; Yonkee and Weil, 2015), whereas the structural style of the Laramide belt is widely described as thick skinned (e.g., Erslev, 1993; Lacombe and Bellahsen, 2016; Pfiffner, 2017).

Early work on the Idaho-Montana fold-thrust belt within and adjacent to the southwestern Montana recess recognized that some thick-skinned structures of the Laramide belt were active prior to terminal thrusting in the more thin-skinned Sevier belt to the west (e.g., Ryder and Scholten, 1973; Perry et al., 1983; Nichols et al., 1985). This led workers to hypothesize buttressing of structures of the Sevier belt against the uplifted Laramide belt (e.g., Beutner, 1977; Perry et al., 1983, 1989; Schmidt et al., 1988; Skipp, 1988a). The boundary between these two thrust systems cannot be described easily in two dimensions, as a line on a map. The boundary is more three dimensional, with the thin-skinned Sevier belt structurally overlying the thick-skinned Laramide belt in general (Kulik and Schmidt, 1988; Tysdal, 1988; Parker and Pearson, 2021). Even in three dimensions, locating a precise boundary between the Sevier and Laramide belts is difficult within the Idaho-Montana fold-thrust belt. For example, southeast of the Snowcrest Range (Plate 1), industry well logs and unpublished seismic data suggest the presence of a blind, basement-involved but thin-skinned thrust system (Perry et al., 1983, 1989), whereas to the northwest, thick-skinned thrust systems carried basement rocks to high structural levels in the hanging wall of the Blacktail-Snowcrest uplift (Perry et al., 1983; Kulik and Perry, 1988).

Basement uplifts in the Montana Laramide belt are interpreted as thick-skinned thrusts that are relatively deeply detached, commonly with evidence of reactivation of Precambrian structures (e.g., Schmidt et al., 1988; O’Neill et al., 1990). For example, the Blacktail-Snowcrest uplift (Perry et al., 1983; McBride et al., 1992) and Hawley Creek thrust (Parker and Pearson, 2021) exploited Paleozoic or older normal fault zones. About 80 km north, the southern boundary of the Helena salient is interpreted to be a Mesoproterozoic, basement-involved normal fault that accommodated subsidence of the Belt Basin (e.g., Harrison et al., 1974; Winston, 1986; Schmidt et al., 1988; Sears, 2007, 2016). Though most prior workers have considered the Sevier and Laramide belts as kinematically separate thrust systems with a distinct boundary, O’Neill et al. (1990) proposed that the Dillon cutoff is a thick-skinned transfer zone that linked the thin-skinned Helena salient and southwestern Montana recess by reactivating structures related to the Paleoproterozoic Big Sky orogen in the northwestern corner of the Archean Wyoming province. This body of evidence suggests that preexisting weaknesses in metamorphosed Precambrian basins have influenced the detachment geometry of thrust faults in the Montana Laramide belt.

Unmetamorphosed sedimentary rocks above the crystalline basement may also have influenced the geometries of thrust faults in the Idaho-Montana fold-thrust belt. The pre-orogenic stratigraphy of the study area mostly consists of Neoproterozoic and Paleozoic rift and passive-margin strata that thin eastward onto the Lemhi arch (Sloss, 1954; Scholten, 1957) and continental shelf of southwestern Montana (Figs. 2 and 3; Scholten, 1957; Ruppel, 1986; Lund et al., 2010; Grader et al., 2016; Link et al., 2017a; Brennan et al., 2020, 2023). The Lemhi arch was inherited from the Neoproterozoic to Ordovician rift margin, consisting mostly of Archean to Paleoproterozoic crystalline basement on its inboard (northeastern) flank (DuBois, 1982; M’Gonigle, 1993, 1994), with ~1.5–3.0 km of overlying Belt Supergroup–equivalent quartzite near its apex (Skipp, 1988a) and predominantly Belt quartzite on its western flank with crystalline basement rocks at depth (Link et al., 2017b). A sub–Middle Ordovician angular unconformity (e.g., Scholten and Ramspott, 1968; Ruppel, 1986; Hansen and Pearson, 2016; Pearson and Link, 2017) marks the top of laterally discontinuous Neoproterozoic to Cambrian rocks on the flanks of the Lemhi arch (Montoya, 2019; Pearson and Link, 2021). This ~2.5-km-thick rift-to-drift succession can be correlated with thicker and more stratigraphically complete sections along strike in central and southeastern Idaho (Trimble, 1976; Link et al., 1987; Brennan et al., 2020, 2023), where rift-to-drift rocks lie on crystalline basement and served as the décollement to the fold-thrust belt in the Wyoming salient (Royse et al., 1975; Coogan, 1992). The sub-Ordovician unconformity omits weak Neoproterozoic and Cambrian layers that host the basal décollement in adjacent portions of the Cordillera (e.g., DeCelles and Coogan, 2006; Yonkee and Weil, 2015), given that the dominantly carbonate Ordovician to Devonian passive-margin succession pinches out against the flanks of the Lemhi arch from both sides (Fig. 3; Scholten, 1957; Brennan et al., 2020). A regionally extensive intra-Devonian unconformity defines the top of the Lemhi arch (Grader et al., 2016), marking the stratigraphically lowest rocks that are laterally continuous across the study area. In the absence of deeper Neoproterozoic to Cambrian rocks that host the basal décollement elsewhere in the Cordillera, the shales and silty carbonate rocks above the Lemhi arch acted as a regional décollement within the study area (Skipp, 1988a; Parker and Pearson, 2021).

Immediately above the intra-Devonian Lemhi arch unconformity, the Devonian section is as thin as 50–100 m (Grader et al., 2016; Parker and Pearson, 2020) but laterally continuous across the study area. The overlying Mississippian section is more than ~400 m thick (Huh, 1967), consisting of a carbonate sequence constructed on the edge of the craton (Rose, 1976). Mississippian rocks become nearly 1.5 km thick (Huh, 1967) and are dominated by fine-grained siliciclastic rocks in the lower part of the section in the Lost River Range, northeast of the former non-collisional Antler “orogeny” highlands (e.g., Sandberg, 1975; Wilson et al., 1994; Link et al., 1996; Beranek et al., 2016). Detachment and fault-propagation folds (Fig. 4) commonly exploited fine-grained shales and silty limestones at the base and near the top of the Mississippian section (e.g., McGowan Creek, Middle Canyon, Lodgepole, Railroad Canyon, and Lombard formations; Messina, 1993; Fisher and Anastasio, 1994; Anastasio et al., 1997). Pennsylvanian and Permian passive-margin strata are widespread throughout the study area, but Mesozoic and earliest Cenozoic rocks are nearly absent from the Idaho segment (Rodgers and Janecke, 1992; Parker et al., 2022). Based on available drill logs in the foreland, the thickness of Triassic and Jurassic strata was likely insignificant at the scale of the cross section (Fig. 3). Despite their limited thickness, these units were likely important décollements in southwestern Montana (Gorder, 1960; Perry et al., 1981, 1983), as were the overlying relatively thin (as much as ~900 m thick) Early Cretaceous foreland basin strata of the Kootenai and Blackleaf formations (e.g., Brumbaugh and Hendrix, 1981; McBride, 1988; McDowell, 1997; Skipp et al., 2017). The overlying Late Cretaceous units (Frontier Formation and Beaverhead Group) are considerably thicker, forming kilometers-thick foreland basins adjacent to uplifts throughout the foreland (e.g., Haley, 1986; DeCelles et al., 1987; Haley et al., 1991; Vuke, 2020). Though widespread erosion has removed Mesozoic and earliest Cenozoic rocks throughout much of the Idaho-Montana fold-thrust belt, anomalously high maximum burial temperatures of ~240–270 °C suggest that an extensive wedge-top basin, at least 1.5–5 km thick, overlapped much of the fold-thrust belt in Early Cretaceous time (Fig. 5; Parker et al., 2022).

The influence of the Lemhi arch basement high on later fold-thrust geometry has been inferred based on the spatial coincidence between structural style and the pre-thrusting stratigraphy (e.g., Parker and Pearson, 2021). Crosscutting relationships between thin- and thick-skinned thrusts are well documented throughout the Idaho-Montana fold-thrust belt (e.g., Perry et al., 1988; Schmidt et al., 1988; Tysdal, 1988; McDowell, 1997; Parker and Pearson, 2021). Archean and Paleoproterozoic gneisses and schists of the Dillon block of the Wyoming craton (e.g., Condie, 1976) as well as Mesoproterozoic quartzites of the Lemhi subbasin in east-central Idaho (Ruppel, 1975) lack sub-horizontal, weak lithological layers and thus behaved as mechanical basement. In some locations, there is convincing evidence that older structures have been reactivated (e.g., Schmidt and Garihan, 1983; Parker and Pearson, 2021). Thrusts and folds that involve the basement high of the Lemhi arch are characterized by variable fault dips and folds with wavelengths of tens of kilometers (Lonn et al., 2016), which suggest detachment levels in the middle crust; these large-wavelength fault-propagation folds and faults are best described as exhibiting a thick-skinned structural style (Parker and Pearson, 2021).

In the study area, end-member thin- and thick-skinned thrusts do not delineate distinct domains in map view. Instead, structural domains constitute a double-decker system with thin-skinned thrusts above the intra-Devonian Lemhi arch unconformity and thick-skinned thrusts below it (Parker and Pearson, 2021). Similar geometries have been documented in many other fold-thrust belts globally (e.g., Lacombe and Bellahsen, 2016; Pfiffner, 2017; Tavani et al., 2021). Some of the type localities from which critical taper theory was originally developed, including Taiwan and the Zagros (e.g., Davis et al., 1983), are now interpreted as having structurally underlying, thick-skinned components (Hung et al., 1999; Yang et al., 2001; Lacombe and Mouthereau, 2002; Molinaro et al., 2005; Mouthereau et al., 2006, 2007, 2012; Sherkati et al., 2006; Allen et al., 2013; Lacombe and Bellahsen, 2016; Le Garzic et al., 2019; Tavani et al., 2021). This double-decker geometry in the study area suggests kinematic compatibility among a range of structural styles and a model of progressive deformation that is controlled by the distribution of weak sedimentary cover rocks (Parker and Pearson, 2021) and the rheology of the crust inherited from its earlier development as a rifted continental margin (Lacombe and Bellahsen, 2016; Lescoutre and Manatschal, 2020; Tavani et al., 2021).

The ca. 135–110 Ma Kootenai Formation represents the oldest definitive foredeep deposit in the study area. The Kootenai Formation is up to 300 m thick, with detritus interpreted to have been shed from exhumed Ordovician to Mississippian rocks carried by active thrusts west of the Lemhi arch in central Idaho (Rosenblume et al., 2021, 2022). The provenance of the overlying ca. 110–105 Ma Blackleaf Formation, as much as 300 m thick, suggests continued shortening and unroofing down to Cambrian stratigraphic levels (Rosenblume et al., 2022), with a ca. 105–100 Ma episode of thrusting that exhumed Triassic to Permian sources in east-central Idaho as the upper, thin-skinned fold-thrust belt propagated into middle Paleozoic and younger strata above the Lemhi arch (Gardner et al., 2022). The Pioneer and other thrusts west of the Lemhi arch (Plate 1) are older than crosscutting ca. 97–91 Ma plutons (Montoya, 2019; Porter, 2021). The ca. 95–83 Ma Frontier Formation is up to 2 km thick, made of predominantly fine-grained volcaniclastic sediments, recording increased sedimentation rates (Dyman et al., 2008) as increased volumes of ash from the magmatic arc associated with the Idaho batholith filled the foreland basin, overwhelming detritus shed from the exhuming thrust belt (Finzel et al., 2023).

There is evidence of deformation across most of the Idaho-Montana fold-thrust belt after ca. 90 Ma (e.g., Perry et al., 1988; Garber, et al., 2020; Finzel et al., 2023). The provenance of the Frontier Formation suggests both distal western sources from the hinterland and more proximal sources in the foreland (Finzel et al., 2023). Deposition proximal to exhumed Pennsylvanian and Permian strata constrains early movement on the basement-involved Blacktail-Snowcrest uplift prior to ca. 85–87 Ma, which was contemporaneous with continued exhumation of Lemhi subbasin strata and associated ca. 1.38 Ga intrusions (Finzel et al., 2023). The ca. 83–66 Ma Beaverhead Group records continued unroofing of the Blacktail-Snowcrest uplift from ca. 83 to 81 Ma (Wilson, 1970; Ryder and Scholten, 1973; Nichols et al., 1985; Perry et al., 1988; Garber et al., 2020). Zircon (U-Th)/He (ZHe) data from quartzites within the Patterson culmination constrain cooling of the hanging wall below ~180 °C (Reiners, 2005) to 87.7 ± 5.4 Ma (Fayon et al., 2017). The nearby Poison Creek and Hawley Creek thrusts, which have commonly been interpreted as defining the western trailing edge of the Sevier belt (e.g., DeCelles, 2004; Yonkee and Weil, 2015), yield younger dates with calculated mean ZHe ages of ca. 68 and 85 Ma, respectively (Hansen and Pearson, 2016; Kaempfer, 2021). Distinctive ca. 498 Ma plutons in the hanging wall of the Hawley Creek thrust provided zircons to the Beaverhead Group starting ca. 68 Ma (Garber et al., 2020). Nearby, the basement-involved Cabin thrust postdates the Medicine Lodge–Four Eyes Canyon thrust, which cuts the ca. 83–66 Ma Beaverhead Group (Perry et al., 1988; Skipp, 1988a; Garber et al., 2020). Near what is generally considered the leading edge of the Sevier belt, the Tendoy thrust deformed the Beaverhead Group, with maximum depositional ages of ca. 83 Ma in its hanging wall and 70 Ma in its footwall (Perry et al., 1988; Garber et al., 2020). The Johnson thrust system crosscut the Medicine Lodge–McKenzie thrust system and deformed Beaverhead Group strata with maximum depositional ages ca. 72 and 68 Ma (Garber et al., 2020), suggesting that the Johnson thrust system postdates all other deformation features. A regional unconformity beneath the ca. 53 Ma Challis Volcanic Group signifies the end of contractional deformation in this region (Rodgers and Janecke, 1992).

Deformation in the foreland is generally younger than exhumation of the Blacktail-Snowcrest uplift, which is likely the oldest foreland uplift in the study area (Perry et al., 1988; Garber et al., 2020; Orme, 2020; Finzel et al., 2023). In the Madison Range, crosscutting igneous rocks and depositional ages of the Livingston Formation bracket the age of the Scarface-Hilgard thrust system between ca. 79 and 69 Ma (Tysdal et al., 1986; DeCelles et al., 1987; Kellogg and Harlan, 2007). This agrees with ZHe data from the Gravelly-Madison (Scarface-Hilgard thrust) and Madison-Gallatin (Spanish Peaks thrust) uplifts that suggest thrust-related cooling initiated by ca. 80 Ma (Kaempfer et al., 2021). Deformation of the Madison-Gallatin uplift continued until ca. 56 Ma (DeCelles et al., 1987). Depositional ages of strata that pre- and postdate folding (Fort Union Formation and Absaroka-Gallatin volcanic field) in the Crazy Mountains Basin and Bridger Range (Harlan et al., 1988; Lageson, 1989; Vuke, 2020) bracket deformational ages to ca. 60–53 Ma near the deformation front.

In contrast to age constraints within other segments of the North American Cordillera (e.g., Wiltschko and Dorr, 1983; Harlan et al., 1988; Burtner and Nigrini, 1994; Kellogg and Harlan, 2007) and previous interpretations of in-sequence deformation (DeCelles, 2004), widespread crosscutting relationships (e.g., Skipp, 1988a; Perry et al., 1988; Tysdal, 1988; McDowell, 1997; Parker and Pearson, 2021) and overlapping age constraints demonstrate significant out-of-sequence deformation within the Idaho-Montana fold-thrust belt. The available age constraints summarized above suggest that out-of-sequence deformation was particularly widespread following initiation of the Blacktail-Snowcrest uplift after ca. 85 Ma.

To better constrain the relative timing and magnitude of shortening in the Idaho-Montana fold-thrust belt and develop a viable kinematic model, we created a regional-scale, area-balanced, and sequentially restorable cross section that spans most of the retroarc orogenic wedge, from the hinterland to the foreland. We compiled available maps at 1:250,000 scale (see Plate 1 for references). Using Petroleum Experts’ kinematic modeling software MOVE, we projected attitudes, contacts, available wells, subsurface data, and published cross sections onto the cross-section line (see Plates 1 and 2 and the Supplemental Material¹). We also projected depths of burial estimated from maximum temperature results using Raman spectroscopy of carbonaceous material (RSCM), conodont color alteration index, and vitrinite reflectance data (Parker et al., 2022) onto the cross-section line and used them to construct an approximate burial envelope to further constrain the cross section. Figure 3 shows the stratigraphic thicknesses we used for segments of the cross section.

Using accepted principles for creation of balanced cross sections (Dahlstrom, 1969), we initially hand drafted the cross section at 1:250,000 scale before digitizing, area balancing, and progressively restoring the cross section in the software program MOVE. We used a flat décollement for three reasons. First, while adjacent segments of the Cordillera have gently dipping (~3°–5°) basement-cover contacts that hosted the later décollement (e.g., Bally et al., 1966; DeCelles and Coogan, 2006; Fuentes et al., 2012), map relationships and stratigraphic disparities demonstrate that the basement-cover contact in the study area is discontinuous and varies in orientation due to faulting related to the Lemhi arch and possible inherited structural relief from rifting in the Mesoproterozoic Belt Basin. Second, paleothermometry data, isopachs of foreland basin rocks, and flexural modeling suggest that the early (upper) décollement Idaho-Montana fold-thrust was broad and low relief, with very little flexural subsidence and a low taper angle (Parker et al., 2022). Third, the substantial final width of the orogenic belt suggests an essentially flat décollement, detached in the middle crust, and a near-zero taper angle (Erslev et al., 2022).

During restoration, we used the simple shear module in MOVE for all normal faults and for thrust faults with interpreted fault-bend folds. We used the fault-parallel flow module for thrust faults without fault-bend folds. For approximate restorations of fault propagation folds, we used a combination of forward and inverse modeling using the tri-shear module (Erslev, 1991). We schematically restored detachment folds by pinning and unfolding contacts. During forward modeling of the cross section, we used simple shear and fault-parallel flow modules but did not model folds in detail where independent constraints on folding variables were unavailable.

We began by restoring Cenozoic normal faults first. When available, wells constrained the minimum thickness of hanging-wall basin fill, allowing for normal fault geometries to be precisely modeled in MOVE so that they agree with contact geometries in the hanging wall. Normal fault geometries were modeled in MOVE from hanging-wall contacts when appropriate. For post-Eocene normal faults without independent evidence of normal reactivation of thrust faults (e.g., Janecke et al., 1999), we used an inferred décollement at 15 km below sea level (bsl) and a listric geometry. This is consistent with the observed lowest extent of active seismicity in the region (e.g., Smith et al., 1985; Richins et al., 1987), which implies a sub-horizontal extensional detachment in the middle crust. Based on the lowest elevation of the basement-cover contact inferred from drill logs (Perry et al., 1981, 1983; Dyman and Nichols, 1988), seismic reflection data (Lopez and Schmidt, 1985; McDowell, 1997; Johnson et al., 2005), and gravity models (Kulik and Schmidt, 1988), we used a regional basement-cover contact elevation of ~4.5 km bsl in southwestern Montana in our restorations. A detailed description of criteria for constraining the regional basement-cover contact is included in the Supplemental Material.

To gain a more complete understanding of the series of deformation events, we present our balanced and restored cross section at the 1:500,000 scale (Plate 2). Figure 6 shows the most complicated portion of the cross section at the time of maximum shortening, with normal faults restored. Following convention, structures are grouped into structural domains based on structural style generalities. These domains are described below from hinterland (southwest) to foreland (northeast) (Plates 1 and 2). Specific descriptions and justifications for our interpretations are listed in the notes on Plate 2.

Between the Pioneer and Hawley Creek thrusts of east-central Idaho is a >100-km-wide zone consisting of folded Mesoproterozoic to Pennsylvanian strata that we will refer to as the central fold belt (Plates 1 and 2; Fig. 4). Though this region exhibits significant structural relief of as much as ~10 km, it lacks surface exposures of thrust faults with significant (kilometer-scale) stratigraphic offset. The principal structures are the SE-plunging Patterson culmination (Janecke et al., 2000), which exposes Mesoproterozoic quartzites of the Lemhi arch in the Lemhi Range, and a broad synclinorium that exposes fold trains of primarily Mississippian and minor Pennsylvanian rocks in the Lost River Range (Fig. 4). The wavelength of this first-order fold pair is at least ~70 km, suggesting a décollement in the middle crust (Parker et al., 2022). Smaller-wavelength folds ranging from 1–2 km to 5–10 km formed within overlying passive-margin strata, suggesting that this upper region of deformation was decoupled from lower thrusts along the upper décollement (i.e., roof thrust) for the Patterson culmination near the intra-Devonian and sub-Ordovician unconformities (Fig. 6; Hait, 1965; Mapel et al., 1965; Beutner, 1968; Messina, 1993; Fisher and Anastasio, 1994; Anastasio et al., 1997). The Patterson culmination is therefore interpreted as a hinterland-dipping duplex (e.g., Boyer and Elliott, 1982) with the upper and lower décollements linked by a system of SW-dipping thrusts that cut across bedding of quartzites of the Lemhi subbasin. We infer a linked upper and lower décollement, at the base of the overlying sedimentary cover rocks and within the middle crust of the Lemhi arch, respectively, for the central fold belt.

Both the upper and lower décollements appear to have been kinematically linked with the neighboring fold-thrust belt and foreland uplift domain (Fig. 6). Detachment folds related to the upper décollement (Fig. 4) are upright to NE verging in the Lost River Range, where they accommodated ~20% shortening (Messina, 1993; Anastasio et al., 1997). In the southern Lemhi Range and Beaverhead Mountains, folds above the upper décollement are recumbent (White Knob fold belt of Hait, 1965; Beutner, 1968; M’Gonigle, 1982) at the same structural and stratigraphic level as the Thompson Gulch–Fritz Creek thrust (Dry Canyon thrust of Lucchitta, 1966; Skipp, 1988a; Parker and Pearson, 2021). This establishes lateral continuity of the upper décollement following the intra-Devonian Lemhi arch unconformity. The lateral transition from open or upright folds to E- to NE-verging and recumbent folds suggests increasing components of shear strain along the flat décollement, ultimately becoming the well-developed mylonite that defines the Thompson Gulch thrust (structure 7 on Plate 1; Parker and Pearson, 2021). Maximum temperatures along this datum constrain this décollement to a depth of at least ~6 km (Parker et al., 2022). These results demonstrate that shortening in the central fold belt was kinematically linked to the adjacent fold-thrust belt along the upper décollement, near the intra-Devonian unconformity which marks the stratigraphically lowest, laterally continuous strata above the Lemhi arch.

The lower décollement of the central fold belt fed slip to the upper décollement in the form of a duplex and was contemporaneous with foreland uplifts that are interpreted to have shared a mid-crustal décollement. The more steeply SW-dipping backlimb of the Patterson culmination suggests progressive rotation of hanging walls as deformation advanced in sequence, with the spacing of the thrusts exceeding the magnitude of slip (e.g., McClay, 1992). In this interpretation, slip was fed from a deep mid-crustal décollement within the Lemhi arch to bedding-parallel detachment horizons above the Lemhi arch. Contemporaneous exhumation of the Patterson culmination and the Blacktail-Snowcrest uplift suggests that while most of the shortening along the lower décollement was progressively linked to the upper décollement and ultimately the neighboring fold-thrust belt, a portion was transmitted >175 km laterally through the middle crust of the Lemhi arch to the foreland uplift domain.

To the east or northeast of and in contrast to the central fold belt, numerous tightly spaced thrust faults with significant stratigraphic offset occur in the Beaverhead and Tendoy Mountains. Across the region, thrust faults (e.g., Thompson Gulch, Fritz Creek, Medicine Lodge, and McKenzie) shared a common décollement just above the intra-Devonian Lemhi arch unconformity, signifying kinematic linkage with the central fold belt (structures 7, 9, 11, and 15 on Plate 1). Unlike in the central fold belt, where deeper and shallower thrusts merged, mechanical basement–involved thrusts (Baby Joe Gulch–Hawley Creek, Cabin, and Johnson) in the fold-thrust belt domain crosscut the upper décollement (Fig. 6). The two classes of structures highlight upper and lower décollements that linked the adjacent central fold belt and foreland uplift domains.

The Thompson Gulch, Fritz Creek, Medicine Lodge, and McKenzie thrusts shared an upper décollement that followed the intra-Devonian Lemhi arch unconformity with a frontal ramp exposed in the localized foreland basin of the Beaverhead Group between the Tendoy Mountains and Blacktail Range. Restoring slip on the later Hawley Creek, Cabin, and Johnson thrusts restores these thrusts to a continuous upper décollement, described previously. Field observations establish structural continuity between the Fritz Creek (Dry Canyon thrust of Lucchitta, 1966) and Thompson Gulch thrusts, with both thrusts characterized by carbonate mylonite that cut gently up-section to the east in its footwall from the intra-Devonian unconformity to the Triassic (Parker and Pearson, 2020).

Whereas hanging-wall stratigraphy, field mapping, and maximum temperature constraints (Parker et al., 2022) suggest continuity between the Thompson Gulch, Fritz Creek, Medicine Lodge, and McKenzie thrusts, the stratigraphic positions of footwall cutoffs vary for the different thrusts, suggesting significant footwall ramps. Unlike in the central fold belt, the basal Devonian contact is undeformed in the immediate footwall of the Thompson Gulch thrust (Parker and Pearson, 2020). This suggests that at the Thompson Gulch thrust, the upper décollement diverges from the intra-Devonian unconformity and cuts up-section above the Lemhi arch, defining a major footwall ramp beneath the correlative Fritz Creek and Medicine Lodge thrusts. Footwall cutoffs are in Permian to Triassic rocks for the Fritz Creek thrust and Pennsylvanian rocks for the Medicine Lodge thrust, suggesting that the thrust cut across a large, pre-thrust fold in Mesoproterozoic quartzites, hence cutting locally down-section in the direction of transport (Ruppel, 1994). Alternatively, the Fritz Creek and Medicine Lodge thrusts may represent two distinct ramps that branch off a shared décollement at the intra-Devonian unconformity. This alternative interpretation is not favored because no shear zones were observed within any Devonian rocks in the footwall of the Thompson Gulch, Fritz Creek, or Medicine Lodge thrusts. From the Medicine Lodge thrust, continuity with the McKenzie thrust is demonstrated by their shared and continuous footwall stratigraphy (Four Eyes Canyon thrust plate; Plate 1). Beneath the Four Eyes Canyon thrust (structure 12 on Plate 1), which is an imbricate beneath the Medicine Lodge–McKenzie thrust segments, a large frontal footwall ramp cuts from Pennsylvanian to uppermost Cretaceous strata in the foreland (Perry et al., 1988).

From these observations, we interpret the Thompson Gulch, Fritz Creek, Medicine Lodge, and McKenzie thrusts as a formerly integrated thrust system (referred to herein as the Medicine Lodge–McKenzie thrust system): The décollement of the Medicine Lodge–McKenzie thrust system was linked to the sub-horizontal upper décollement of the neighboring central fold belt, likely cut across prior folds in the Beaverhead Mountains, and ultimately carried rocks above the intra-Devonian Lemhi arch unconformity to the surface along a frontal ramp in the Tendoy Mountains.

Unlike in the central fold belt, where the lower décollement transferred slip to an upper décollement above a duplex, mechanical basement–involved thrusts rooted in the lower décollement of the fold-thrust belt domain (Baby Joe Gulch–Hawley Creek, Cabin, and Johnson thrusts) truncated the overlying upper décollement (Fig. 6). Kinematic modeling and maximum temperature data are consistent with the interpretation that the Johnson thrust offset the older Medicine Lodge–McKenzie thrust system (Parker et al., 2022), suggesting that the Baby Joe Gulch–Hawley Creek, Cabin, and Johnson thrusts likely deformed in sequence and in succession from ca. 80 to 68 Ma. The geometry of this complicated region, at the time of maximum shortening, is shown in Figure 6.

The large wavelengths of uplifted hanging walls, minimal shortening required for restoration, evidence of reactivation on the Hawley Creek thrust (Parker and Pearson, 2020, 2021), moderate fault dips, and the piggy-back relationship between the Cabin and Johnson thrusts suggest that these thrusts are characterized by a thick-skinned structural style. Rocks in the hanging walls of the Baby Joe Gulch–Hawley Creek, Cabin, and Johnson thrusts (thrusts a, c, and e on Plate 2) are far above the regional elevation (see Supplemental Material, where constraints behind interpretation of regional elevation are discussed) with structural lows in the Lost River Range and Beaverhead and Tendoy Mountains, defining a wavelength of ~75 km (Parker et al., 2022). The hanging-wall anticline of the Baby Joe Gulch–Hawley Creek thrust restores on the ~35°-dipping fault with ~3–4 km of heave, and changes in stratigraphy across the fault suggest reactivation of a normal fault (Parker and Pearson, 2021). The Johnson thrust dips ≥10° based on map relationships, observed dips, and the fact that it is structurally beneath the Medicine Lodge–McKenzie and Cabin thrusts (Fig. 6). These observations suggest a thick-skinned style, compatible with a mid-crustal décollement that is shared with the neighboring Patterson culmination and Blacktail-Snowcrest uplift (McBride et al., 1992). The alternative interpretation of a thin-skinned fault geometry, similar to that proposed by Skipp (1988a), requires >70 km of slip on the Hawley Creek or Cabin thrust and an entirely blind footwall flat that is not supported by any independent evidence. Our preferred interpretation of the Hawley Creek, Cabin, and Johnson thrusts as thick-skinned thrusts suggests that the Sevier and Laramide belts of this region (Fig. 1) represent a continuum in thin- and thick-skinned structural styles (Parker and Pearson, 2021); our interpreted kinematic linkage of these thrusts by a décollement in the middle crust is similar to other models integrating the Sevier and Laramide belts (O’Neill et al., 1990; McClelland and Oldow, 2004; Erslev et al., 2022).

The foreland uplift domain consists of widely spaced (>50 km) uplifts of variable orientation that exposed crystalline basement rocks of the Dillon block of the Wyoming craton. Our interpretations in this region are largely summaries of previous work (e.g., Schmidt and Garihan, 1983; Sheedlo, 1984; McBride, 1988; Schmidt et al., 1988; Lageson, 1989). The foreland basin domain basically consists of two overlapping thrust and backthrust systems or “pop-ups” (e.g., Erslev, 1993) that appear to have advanced in sequence: the Blacktail-Snowcrest and Gravelly-Madison uplifts and the Madison-Gallatin uplift (Plate 2; Schmidt et al., 1988).

The cross section cuts obliquely through the Blacktail-Snowcrest and Gravelly-Madison uplifts, which may have had a roughly east-west transport direction according to available slickenside data (Schmidt and Garihan, 1983; Schmidt et al., 1988). For transport-parallel cross sections of the Blacktail-Snowcrest uplift, refer to McBride (1988), Sheedlo (1984), and Guthrie et al. (1989). A mid-crustal detachment depth of ca. 23–25 km bsl is suggested by the wavelength of the uplift, the observed dip of the thrust, and stratigraphic relationships that argue for and were used to constrain the geometry of a reactivated normal fault (McBride et al., 1992). The western backlimb of the Blacktail-Snowcrest uplift consists of a moderately dipping basement-cover contact with ~7 km of structural relief, which is cut by the sub–Beaverhead Group unconformity. The Greenhorn-Snowcrest thrust system exhumed the Blacktail-Snowcrest uplift and transferred slip to numerous backthrusts (Tysdal, 1981, 1988; Schmidt and Garihan, 1983; Schmidt et al., 1988) and the underlying Scarface-Hilgard thrust system to the east (Gravelly-Madison arch of Schmidt et al., 1993).

The Madison-Gallatin uplift marks the front of the fold-thrust belt and is apparently younger than the Blacktail-Snowcrest and Gravelly-Madison uplifts because the back-verging Spanish Peaks thrust (of the Madison-Gallatin uplift) truncated the Scarface-Hilgard thrust system (of the Gravelly-Madison arch of Schmidt and Garihan, 1983). Where best exposed, the Spanish Peaks thrust shows as much as 5 km of stratigraphic offset (McMannis and Chadwick, 1964; Garihan et al., 1983; Tysdal et al., 1986). The Cherry Creek fault (structure 22 on Plate 1) is smaller, with ~1.5 km of stratigraphic separation (McMannis and Chadwick, 1964) and smaller-wavelength folds that suggest its detachment flattens at a few kilometers depth. The large wavelength of the Madison-Gallatin uplift, between the Spanish Peaks and sub-Bridger thrusts, is compatible with a mid-crustal décollement at ~25 km bsl, similar to that of the Blacktail-Snowcrest uplift (McBride et al., 1992). At the leading edge of the Madison-Gallatin uplift, an east-verging overturned anticline-syncline pair with >7.5 km of relief is observed in the Bridger Range north of Bozeman, Montana, with no corresponding thrust mapped at the surface (e.g., Schmidt et al., 1988; Lageson, 1989). Only 10 km away, the Sohio Moats 1-3 well penetrated a blind thrust that placed Devonian rocks over Mississippian rocks, ~2 km above the regional elevation, requiring at least two strands on the inferred sub-Bridger thrust (Lageson, 1989). The sub-Bridger thrust likely fed slip into the folded, syn-deformational Crazy Mountains Basin. The Battle Ridge backthrust (structure 23 on Plate 1) accommodated tightening in the leading syncline and balanced bed-parallel thrusting of the sub-Bridger thrust, acting as a triangle zone (e.g., McClay, 1992; MacKay et al., 1996; von Hagke and Malz, 2018) or tectonic wedge (Price, 1986). This interpretation follows that of previous workers (McMannis, 1955; Schmidt et al., 1988; Skipp et al., 1999; Lageson, 1989), who had slightly different subsurface interpretations. Out-of-sequence deformation at ca. 90–70 Ma was kinematically linked from the central fold belt to the pop-up trailing the Gravelly-Madison uplift by a mid-crustal décollement; deformation propagated in sequence to the Madison-Gallatin pop-up, with a frontal ramp linking the décollement to the surface by a triangle zone in the deep Crazy Mountains foreland basin.

The presented balanced cross section restores along two principal detachment horizons, with an overall shortening magnitude of 244 km, or 36%. The length of the restored cross section is 550 km, while the length of the cross section after shortening is 306 km. The upper décollement mostly followed the intra-Devonian unconformity atop the Lemhi arch and was fed slip from a ramp to the southwest. The thrust system associated with this upper décollement accommodated the bulk (~155 km or ~60%) of the total shortening. Incorporation of observed maximum temperatures and burial estimates along the same transect (Parker et al., 2022) suggests that a thicker (~6–8 km thick) southwestern section was thrust over a thin (~3–4 km) stratigraphic section overlying the Lemhi arch to the northeast. Our hypothesized linked Medicine Lodge–McKenzie thrust system, with a single thrust that acted as the upper décollement in the balanced restoration, requires a high magnitude of slip because the line length of Mississippian strata in the hanging wall of the Medicine Lodge–McKenzie thrust system must restore west of the footwall ramp of the Fritz Creek thrust that marks the top of the Lemhi arch (correlative with Thompson Gulch thrust of Parker and Pearson, 2021). Elsewhere in the Cordillera, individual, large-slip thrusts are observed linking folded hinterland regions to the foreland. For example, the Willard (Yonkee et al., 2019), Canyon Range (DeCelles and Coogan, 2006), and Lewis thrusts (Sears, 2001; Fuentes et al., 2012) have shortening magnitudes of ~50, 100, and 150 km, respectively. Our total shortening estimates of the upper system above the Lemhi arch (~155 km) and the system as a whole (244 km) are within the range documented elsewhere within the North American Cordillera (e.g., Elison, 1991; DeCelles and Coogan, 2006; Evenchick et al., 2007; DeCelles, 2012).

The balanced and restored cross section in conjunction with available age and depth constraints (Parker et al., 2022) are the basis for a forward model that recreates the observed map patterns and interpreted subsurface geometries. The presented forward kinematic model (Fig. 7) will be described in four relative time steps.

Early deformation (phase 1) resulted in a moderate amount of shortening (~40 km) accommodated on the upper décollement. A moderately dipping (~20°) ramp fed slip from detachments in Mesoproterozoic and Neoproterozoic–Cambrian strata west of the Lemhi arch to the intra-Devonian unconformity that overlies the Lemhi arch. First-order folds were produced by fault-bend folding, whereas second-order folds were formed by duplexes linking an intra-Devonian and sub-Mississippian detachment horizon. Smaller, third-order detachment folding along the sub-Mississippian detachment contributed to the overall thickness shown (not shown at this scale). Slip was transferred from these two detachment horizons to a single, long thrust flat and then to a small ramp near the top of the intra-Devonian Lemhi arch unconformity (Medicine Lodge–McKenzie thrust system), essentially doubling the thickness of the section above the thinnest part of the Lemhi arch (Parker et al., 2022). A frontal ramp brought the hanging wall to the surface in the active foreland basin. The décollement geometry of the system avoided the Lemhi arch basement high by ramping through the thick package of Neoproterozoic to Cambrian strata west of the arch then following the intra-Devonian unconformity along the western flank and over the top. As a result of the frictional décollement with this geometry, the wedge had a very low taper angle, consistent with the observed maximum temperature profile (Parker et al., 2022). Extensive wedge-top deposition may have buried much of the low-relief fold-thrust belt at this time (Parker et al., 2022).

A transitional phase (phase 2) began ca. 90 Ma as the advancing retroarc wedge activated a deep décollement at ~25 km bsl. A ramp linked this deep décollement to the active upper décollement above the Lemhi arch and fed slip to the front of the system in the form of a duplex. The first-order fold created above the blind ramp of the duplex represents the early Patterson culmination. In effect, this accommodated internal thickening of the sub-critical wedge and increased taper angle as the décollement moved forward, cutting across bedding in quartzites within the western flank of the Lemhi arch. Internal thickening and increased basal slope angle resulted in a fold-thrust belt that resembled a super-critically tapered orogenic wedge, detached in the viscous middle crust. The viscous detachment efficiently transmitted the strain a great distance (consistent with modeling of Ruh et al., 2012; Borderie et al., 2018) through the Lemhi arch. As a result, exhumation of the Lemhi arch, above the Patterson culmination, coincided with early exhumation of the Blacktail-Snowcrest uplift in the foreland. We hypothesize that a lack of potential stratigraphic décollement geometries promoted activation of a viscous mid-crustal décollement, which provided the most efficient way to advance the retroarc orogenic wedge. The shift from a shallow frictional to a deep viscous décollement transferred strain through the Lemhi arch, giving the system an apparent bifurcation from one to two wedges as the wavelengths of the structures increased in response to the increased depth to detachment.

Once strain was linked through the Lemhi arch by exploiting the viscous middle-crust, considerable (~85 km) shortening of the Lemhi arch itself occurred (phase 3, ca. 90–70 Ma) along with continued shortening (~60 km) and exhumation of the overlying thin-skinned thrust system. After initiation of the Blacktail-Snowcrest uplift, the taper angle of the sub-critical wedge increased in the same way as before: by advancement of the upper décollement, thereby increasing the basal slope, and by internal shortening of the wedge during out-of-sequence thrusting. As the Patterson culmination grew and the décollement migrated forward through the quartzite-dominated western flank of the Lemhi arch, hanging-wall horses of formerly active thrusts were rotated toward the hinterland as slip was fed to the upper, thin-skinned décollement above the Lemhi arch. After the duplex shortened the predominantly quartzitic western flank of the Lemhi arch, linking the upper and lower décollements, crystalline basement rock of the edge of the Dillon block was shortened by a series of tightly spaced, in-sequence imbricate thrusts. Unlike in the Patterson culmination, where slip on the deeper thrust system was fed into the roof thrust of the upper thrust system to form a duplex, these thrusts (Hawley Creek, Cabin, and Johnson) truncated the upper, thin-skinned décollement. It is important to note that slip continued to be fed to the active foreland basin in front of the Medicine Lodge–McKenzie thrust system, which also had the effect of thickening the wedge and exhuming the overlying upper thin-skinned system.

Meanwhile, active thrusts of the Blacktail-Snowcrest uplift also thickened the wedge. When super-critical conditions were achieved, the wedge propagated forward, sharing slip between the Blacktail-Snowcrest and Gravelly-Madison uplifts, which effectively acted as one pop-up. This sequence reproduced the two, mutually crosscutting and overlapping, hybrid thin- and thick-skinned systems documented by previous workers (e.g., Kulik and Schmidt, 1988; Kulik and Perry, 1988; Perry et al., 1988; Schmidt et al., 1988; O’Neill et al., 1990; McDowell, 1997; Parker and Pearson, 2021).

During the final phase of deformation (phase 4, ca. 70–55 Ma), the orogenic wedge advanced well into the foreland, as far as the Crazy Mountains Basin. A considerable amount of shortening (~50 km) was accommodated on the leading structure, the sub-Bridger thrust. This structure has the geometry of a hybrid thick- to thin-skinned triangle zone and fed slip from the deep, mid-crustal viscous décollement at the base of the system to bedding-parallel blind thin-skinned thrusts in the thick foreland basin. Slip on formerly active structures ceased as those structures were completely abandoned and the newly active Spanish Peaks thrust crosscut the Scarface-Hilgard thrust system; basically, the pop-up of the Blacktail-Snowcrest–Gravelly-Madison uplift was abandoned in favor of the pop-up of the Madison-Gallatin uplift. Thickening of the Lemhi arch increased the wedge beyond critical, allowing the system to advance forward as a single, thick-skinned wedge.

In summary, this forward kinematic model links the variety of structural styles that define the central fold belt, fold-thrust belt, foreland uplift, and foreland basin domains to a single progressive history of a forward- and downward-progressing wedge of deformation through a basement high. While there was a transition from a generally thin-skinned to a generally thick-skinned style, the transition was gradual in time (tens of millions of years) and heterogeneous in space, resulting in a large region of overlapping styles (across a distance of hundreds of kilometers); this makes it impossible to precisely separate thin-from thick-skinned domains in space and time in this portion of the Sevier-Laramide fold-thrust belt.

The two-dimensional forward kinematic model proposed in Figure 7 describes both the geometric relationship and kinematic continuity between end-member thin- and thick-skinned orogenic wedges. The cross section also illustrates the double-decker geometry described by Parker and Pearson (2021) at a more regional scale, defining a single composite orogenic wedge with the local structural style depending on the availability of sub-horizontal detachment horizons (Kulik and Schmidt, 1988). We hypothesize that the Lemhi arch and its relatively thin sedimentary cover rocks determined the structural style of particular thrusts by limiting the distribution of the early upper thin-skinned décollement, thereby predetermining where a lower thick-skinned décollement would later emerge as the orogenic wedge grew.

Using our forward model, we can improve upon previous models of the Idaho-Montana fold-thrust belt that described along-strike changes in structural style and kinematic connectivity between the Sevier and Laramide belts. O’Neill et al. (1990, p. 1110) also interpreted a common mid-crustal décollement for the Sevier and Laramide belts, emphasizing that “the frontal thrust zone of the Cordilleran [Sevier] thrust belt in southwestern Montana fundamentally involves basement.” O’Neill et al. (1990, p. 1107) proposed a “thick-skinned displacement transfer zone that cuts basement rocks of the Lima [southwestern Montana] recess,” termed the “Dillon cutoff.” In map view (Fig. 2), the transfer zone links the Helena salient to the southwestern Montana recess. Within this transfer zone, lateral (southwestern Montana transverse zone) and oblique (Dillon cutoff) thrust ramps cut off the corner of the Dillon block basement by reactivating a portion of the NW-dipping suture of the Paleoproterozoic Big Sky orogen (e.g., O’Neill and Lopez, 1985; Harms et al., 2004; Condit et al., 2015).

South of where O’Neill et al.’s (1990) description ends (Armstead anticline; structure 16 on Plate 1), our cross section shows how the imbricated Hawley Creek, Cabin, and Johnson thrusts are short and segmented fault splays that “cut off” the edge of the craton (i.e., the Dillon block), likely defining the southern continuation of the Dillon cutoff. Furthermore, our forward model suggests that these thrusts were activated during the final stage (phase 3, ca. 90–70 Ma; Fig. 7) of out-of-sequence deformation that characterized the gradual transition from thin- to thick-skinned thrusting. Recent thermal modeling of ZHe and apatite (U-Th)/He age data from basement rocks of the Armstead anticline suggests rapid cooling between ca. 100 and 80 Ma (Mosolf, 2021), consistent with our interpreted timing of the transition. We thus interpret that closely spaced, imbricated thick-skinned structures of the Dillon cutoff record the geometry that resulted from the incremental (ca. 90–70 Ma) transition from mostly thin- to thick-skinned thrusting.

Comparing spatial and temporal patterns of structural style domains near the Dillon cutoff also allows us to visualize the three-dimensional (3-D) décollement geometry of the greater North American Cordillera and investigate why the structural style evolved from predominantly thin to thick skinned. North of the Dillon cutoff, the Disturbed Belt (Mudge, 1970, 1982) consists of ramp-flat geometries (e.g., Fuentes et al., 2012) that generally characterize a thin-skinned structural style, similar to the early deformation features shown in our forward model (phase 1, Fig. 7). However, west of the Dillon cutoff, near the Idaho-Montana border, the large-wavelength folds in the Lemhi subbasin (Tysdal, 2002; Lonn et al., 2016)—which we interpret to be younger than structurally overlying thin-skinned structures—suggest a more thick-skinned structural style near the trailing ramp. We hypothesize that strain was ultimately transferred from these structures near the trailing ramp northeastward to the leading ramp (phase 2, Fig. 7) near the Dillon cutoff by activation of a viscous mid-crustal décollement in the footwall of the former thin-skinned décollement. Farther south, within the central Wyoming salient, the distance between the trailing ramp near the Cordilleran hingeline and the leading ramp of the Laramide belt increases dramatically (Fig. 2), suggesting a décollement geometry similar to the late deformation features shown in our forward model (phase 4, Fig. 7). Along-strike changes in structural style within the North American Cordillera, from north to south, may mimic the progressive change from a predominantly thin- to thick-skinned style shown in our forward model as a mid-crustal viscous décollement was activated.

Along-strike changes in structural style (Dillon cutoff) occurred where the deformation front crossed segment boundaries of the former rift margin (Lemhi arch; e.g., Lund, 2008); across-strike changes (Laramide belt) occurred where the deformation front crossed the rift margin (Cordilleran hingeline) and encountered unattenuated crust. The transition from thin- to thick-skinned thrusting coincided not only with deformation of thinner stratigraphy but also with thicker crust that we infer developed viscous mid-crustal weak layers. Progressive strain transfer from thin-skinned translation of basin fill constructed on thin brittle crust to thick-skinned shortening outside of basin margins where thick crust contained a viscous middle crust explains the final décollement geometry described above (Fig. 8B; O’Neill et al., 1990; McClelland and Oldow, 2004). These interpretations are similar to predictions of tectonic inheritance models (Horton and Folguera, 2022, their figure 1.3), a class of models that relate structural style to intrinsic properties of the crust. Specifically, deformation of the Idaho-Montana fold-thrust belt resembles the model of Lescoutre and Manatschal (2020), which predicts that a shift from thin- to thick-skinned thrusting occurs when the deformation front reaches continental crust that is thick and hot enough to have a viscous middle crust (Fig. 7), particularly at segment boundaries (e.g., Lemhi arch) separating adjacent basins (e.g., Belt Basin, rift and passive margin basin). Simply by advancement of a wedge-shaped deformation front forward and structurally downward, through lithosphere with discontinuous crustal weak layers at various depths, spontaneous changes in the style and tempo of mountain building may emerge in the absence of changes in plate-boundary geodynamics (e.g., Lacombe and Bellahsen, 2016; Tavani et al., 2021). While the limited availability of weak stratigraphic layers may have increasingly impeded shortening on thin-skinned structures, the viscous mid-crustal weak layers, which likely became abundant in the foreland, may have increasingly promoted thick-skinned shortening as the deformation front advanced through the basin margin and into the continent.

The emergence of a viscous mid-crustal décollement may have a twofold effect. First, by establishing a link between once-isolated viscous detachment horizons in the now-thickened hinterland and the always-thick foreland, the emergence of a through-going mid-crustal décollement allows for strain to be transmitted from the plate boundary to the continental interior (Tavani et al., 2021). Secondly, the transition from a frictional to a viscous décollement may increase the rate at which strain is transferred through the orogen because a viscous décollement is much weaker than a frictional décollement. Numerical and analog models demonstrate that—compared to frictional detachments—viscous detachments result in a lower critical taper angle and are much more effective at rapidly transmitting strain to the front of the orogen (Williams et al., 1994; Ruh et al., 2012; Borderie et al., 2018). We propose that as viscous, mid-crustal weak layers arise and become interconnected during progressive thickening and heating of the crust, not only do they link thin- and thick-skinned thrusts to a single orogenic wedge, they also prompt a switch from a frictional to a much weaker viscous décollement, giving the deformation front the ability to suddenly “jump” in front of (and beneath) its former thin-skinned system (Fig. 8).

This combined top-down stratigraphically inherited and bottom-up rheology-inherited strain propagation model may explain the apparent bifurcation (phase 2, Fig. 7) and later integration (phase 4, Fig. 7) of the wedge within the context of critical taper theory as applied to brittle-plastic wedges (e.g., Williams et al., 1994). We propose that several classes of tectonic inheritance models work together to explain how feedbacks among preexisting weaknesses, temperature, and rheology modulate orogenesis. Stratigraphic thickness and preexisting faults may predict the 3-D geometry of the décollement and the structural style of specific structural domains (Parker and Pearson, 2021). These upper-crustal weaknesses may also provide the most efficient way to internally thicken the wedge and transmit strain from the viscous middle crust to the surface by advancing the former décollement of the trailing system in a top-down fashion. The resultant crustal thickening and burial heating during foreland basin sedimentation may heat and weaken the crust, promoting formation of through-going viscous mid-crustal weak layers that allow emergent thick-skinned thrusts to propagate forward into the continent in a bottom-up fashion. Feedbacks between these inherited properties of the crust and the evolving rheological profile of growing orogenic belts may explain why the transition between end-member thin- and thick-skinned styles is so gradual (tens of millions of years) and distributed over much of the width of the orogen. Our proposed kinematic model may be a useful tool for understanding not only how self-organized changes in structural style may emerge simply as a consequence of progressive shortening but also how the changing rheology of the middle and upper crust may itself govern the behavior of growing orogenic belts.

Our balanced and restored cross section of the Idaho-Montana fold-thrust belt demonstrates that a spectrum of structural styles, ranging from thin- to thick-skinned, linked the Sevier and Laramide belts. Early (pre–ca. 90 Ma) shortening was accommodated on a stratigraphically controlled décollement, following strata above the intra-Devonian Lemhi arch unconformity, immediately above the basement high. Thin stratigraphy over the Lemhi arch limited the depth, basal angle, and geometry of this frictional upper décollement (Medicine Lodge–McKenzie thrust). We hypothesize that in response to the lack of structurally deeper weak layers, the wedge advanced by activating a lower décollement within the weak, viscous middle crust, thereby steepening the basal décollement and allowing the wedge to propagate. Emergence of a viscous mid-crustal décollement, possibly triggered by heating due to thickened crust and burial heating by the adjacent foreland basin, efficiently transmitted strain through the rift margin (Lemhi arch) and into the foreland at ca. 90–70 Ma. The weaker, quartzite-dominated western flank of the Lemhi arch was shortened by advancement of the basal slope of the upper décollement, progressively linking the upper and lower décollements as a duplex (Patterson culmination). After the upper décollement was completely abandoned, imbricated thick-skinned thrusts of the Dillon cutoff truncated the former upper décollement, allowing deformation to propagate forward. In-sequence deformation ultimately fed slip into the Crazy Mountains foreland basin by a thick- to thin-skinned triangle zone. A minimum of 244 km of shortening was accommodated from ca. 100 to 55 Ma during this prolonged transition from thin-skinned deformation of sedimentary cover rocks to thick-skinned deformation of continental crust.

In this segment of the North American Cordillera, the gradual transition from thin- to thick-skinned thrusting and corresponding jump of the deformation front was self-organized, not requiring a sudden change in plate-boundary conditions. Stratigraphic inheritance limited potential frictional décollements, hindering top-down growth of the thin-skinned wedge. Crustal rheology, specifically the evolving distribution of viscous mid-crustal weak layers, facilitated rapid late-stage, bottom-up growth of the wedge as a viscous décollement emerged. These feedbacks between inherited structural and stratigraphic weaknesses in the upper crust and the evolving rheology of the middle crust may more accurately predict the spatiotemporal deformation patterns observed in orogens that propagate into continental interiors than competing models that call on discrete tectonic events at the margin.

This project was completed as a component of Parker’s Ph.D. dissertation at Idaho State University. We gratefully acknowledge Petroleum Experts for use of MOVE software, funding from U.S. National Science Foundation EAR grant 1728563 to Pearson, and conversations with Jim Coogan, Paul Link, Emily Finzel, Justin Rosenblume, Leslie Montoya, Chase Porter, and Cole Gardner. We appreciate comments by Todd LaMaskin, Sean Long, and Kurt Constenius, which improved the manuscript.