Faulty foundations: Early breakup of the southern Utah Cordilleran foreland basin

Anomalous features of Upper Cretaceous strata in southern Utah challenge existing tectonic and depositional models of the Cordilleran foreland basin. Extreme thickness variations, net to gross changes, and facies distributions of nonmarine to marginal marine strata of the Turonian–early Campanian Straight Cliffs Formation are documented across the Southwestern High Plateaus. Con-trary to most traditional models of foreland basin architecture, regional correlations demonstrate abrupt stepwise thickening, with a punctuated increase in average grain size of key intervals from west to east, i.e., proximal to distal relative to the fold-thrust belt. Except in the most proximal sections, fluvial drainage systems were oriented predominantly subparallel to the fold-thrust belt. Combined, these results suggest that modern plateau-bounding faults may have had topographic expressions as early as Cenomanian time, and influenced the position of the main axial river system by creating northeast-trending paleotopography and sub-basins. Laramide-style tectonism (e.g., basement-involved faults) is already cited as a driver for sub-basin development in latest Cretaceous–Cenozoic time, but new data presented here suggest that this part of the foredeep was “broken” into distinct sub-basins from its earliest stages. We suggest that flexural foundering of the lithosphere may have caused early stage normal faulting in the foredeep. Regional implications of these new data indicate that both detachment-style and basement-involved structures were simulta-neously active in southern Utah earlier

Evolution of the Sevier fold-thrust belt and foreland basin system in southern Utah (Fig. 1B) is poorly understood relative to other parts of the Cordillera. Major thrust structures appear to have been emplaced by the Cenomanian (DeCelles, 2004), with little evidence of major basinward propagation over time. Preliminary data from foreland basin strata suggest unusually high accommodation, sediment supply, and coarse-grained deposition in distal parts of the foredeep, with relatively stationary (nonmigrating) depozones (Eaton and Nations, 1991;Goldstrand et al., 1993;Mulhern and Johnson, 2016). Although flexure adjacent to the Sevier fold-thrust belt is thought to be a primary control on subsidence and sedimentation patterns in northern Utah and Wyoming before ca. 83 Ma (DeCelles, 2004), previous studies of southern Utah foreland basin stratigraphy indicate a more complicated history from at least Cenomanian time, long before onset of Laramide-style tectonics (Gustason, 1989;Eaton and Nations, 1991;Laurin and Sageman, 2001).
Cenozoic normal faults bound a series of plateaus in the study area (Southwestern High † gaestpierre@gmail.com. Plateaus, SWHP; Fig. 2), exposing relatively undeformed Upper Cretaceous strata ∼200 km west to east (proximal to distal relative to the fold-thrust belt), and ∼50 km of outcrop from south to north. Whereas fluvial architecture along the western Kaiparowits Plateau has been studied in some detail (Gooley et al., 2016;Benhallam et al., 2016;Primm et al., 2018;Koch et al., 2019), much less is known about fluvial systems to the west in the Markagunt and Paunsaugunt Plateaus, which preserve more proximal deposits of the Cretaceous foreland basin. Regional correlations to date focus mainly on biostratigraphy and marine deposits, but do not detail regional fluvial interactions (Eaton and Nations, 1991;Eaton et al., 1993;Goldstrand, 1992Goldstrand, , 1994Moore and Straub, 2001;Biek et al., 2015).
We completed a regional correlation across the southern margin of the Sevier foreland basin to investigate controls on foreland basin development, and in particular their effect on fluvial systems during Late Cretaceous time. Targeted measured sections document stratigraphic architecture, proximal-to-distal changes in thickness and average grain size, and paleocurrent indicators. Ultimately, these data allow us to infer controls on sedimentation and to test prevailing models for foreland basin development.

Structural Evolution of Southern Utah
The Sevier fold-thrust belt trends northeast across Utah, deforming and exhuming Paleozoic and Mesozoic strata which provided sediment to the adjacent foreland basin to the east (DeCelles and Coogan, 2006). At its southwestern extent, the Sevier fold-thrust belt formed a high-angle junction with the Mogollon Highlands ( Fig. 1) in northwestern Arizona. The geometry of these features allowed for incursion of the Utah Bight, an embayment filled by the Cretaceous Western Interior Seaway (Fig. 2 inset; Zapp and Cobban, 1960;McGookey et al., 1972).
The Sevier fold-thrust belt in northern and central Utah shows evidence for eastward propagation of thin-skinned thrust sheets, punctuated periods of out-of-sequence thrusting, and growth of structural culminations, all of which influenced subsidence and sediment supply (Fig. 1B;DeCelles et al., 1995). The Canyon Range thrust system represents the first Figure 1. Simplified maps of key Precambrian to Late Cretaceous structures across the Utah (UT) Cordilleran orogenic system (with parts of Nevada (NV) and Arizona (AZ)). The field area for this study is marked with a red box outlining the Southwestern High Plateaus, including the Gunlock and Pine Valley areas (G and PV), the Markagunt Plateau (MP), the Paunsaugunt Plateau (PP), and the Kaiparowits Plateau (KP). (A) Relevant basement structures include the Wasatch Hinge Line (Picha and Gibson, 1985); the suture zone between the Mojave and Yavapai provinces (Karlstrom and Shaw, 1999); and Neoproterozoic normal faults and lineaments (Picha, 1986;Currie, 2002 et al., 1995). This culmination included reacti-vation and incorporation of the Canyon Range and Pavant thrust sheets, which created a topographic high, providing sediment to the foreland basin through Paleocene time (DeCelles et al., 1995;Stockli et al., 2001). The structure and timing of major thrust faults in southwestern Utah are poorly known relative to central Utah, partly due to Cenozoic exten-sional overprinting. The Wah Wah thrust sheet was active from ca. 120 to 97 Ma and is primarily composed of back-breaking imbricate thrusts (Fig. 1B;Miller, 1963;Fillmore and Middleton, 1989;Friedrich and Bartley, 2003;DeCelles, 2004). The nearby Blue Mountain thrust system ∼10 km to the southeast was emplaced by latest Cenomanian-Turonian time (ca. 95-85 Ma;  Miller, 1966;DeCelles, 2004;Fig. 1B). The Iron Springs thrust system is the youngest and easternmost expression of the Sevier fold-thrust belt in southern Utah, and it is characterized by three east-vergent, forward-breaking thrust sheets. It is estimated to have been active in Santonian-Campanian time (ca. 84-70 Ma;Goldstrand, 1994), but drag folds in Paleocene-Eocene conglomerates indicate minor late stage displacement as well (Anderson and Dinter, 2010).
In contrast with the central Utah fold-thrust belt (e.g., Canyon Range, Pavant, Paxton-Gunnison thrusts), the southern thrust systems did not consistently propagate eastward, and surface expressions stalled ∼100 km farther west of the easternmost central Utah thrust sheet (Fig. 1B). Younger faults in the area include the Ruby's Inn and Paunsaugunt thrust faults, but these structures are attributed to the emplacement of the Oligocene-Miocene Marysvale Volcanic Field (Merle et al., 1993;Davis, 1999;Biek et al., 2019). Proprietary industry data and a regional gravity survey (e.g., Upper Valley Oil field near Cedar City; Cook and Hardman, 1967;Van Kooten, 1988) indicate no evidence for subsurface thrust faults east of the Iron Springs thrust system. Basement-involved features in southern Utah began to form during the Late Cretaceous as asymmetric folds with varying orientations ( Fig. 1B; Davis and Bump, 2009). The Kanarra, East Kaibab, and Circle Cliffs monoclines are the most prominent Laramide-style features in this region (Fig. 1B). Movement along the East Kaibab monocline during deposition of the Campanian Wahweap Formation is the earliest evidence for Laramide-style deformation in this area (Fig. 1B;Tindall and Davis, 1999;Hilbert-Wolf et al., 2009).
The nature and orientations of Cretaceous structures in southern Utah are influenced by basement features underlying the Colorado Plateau (Picha and Gibson, 1985;Schwans, 1995;Neely and Erslev, 2009). Beneath Phanerozoic strata is a northeast-trending shear zone of the Paleoproterozoic Yavapai-Mojave basement, which overlaps with the southwestern edge of the Colorado Plateau ( Fig. 1A; Shaw and Karlstrom, 1999;Davis and Bump, 2009). Proterozoic normal faults overprinted the Yavapai-Mojave shear zones during breakup and rifting of Rodinia (Dickinson, 2004). Basement-cored faults of the Laramide orogeny align with these Proterozoic normal faults (e.g., the East Kaibab monocline; Fig. 1), suggesting that basement heterogeneities influenced the location of Cretaceous structures (Picha, 1986;Huntoon, 1993;Stone, 1993;Tindall and Davis, 1999;Marshak et al., 2000;Bump and Davis, 2003;Neely and Erslev, 2009). Present-day structure and topography in the study area is reflected in the Markagunt, Paunsaugunt, and Kaiparowits Plateaus, which together comprise the Southwestern High Plateaus. These modern features are bounded by Cenozoic normal faults (from west to east: the Hurricane, Sevier, and Paunsaugunt faults), which mark the transition between the Basin and Range and Colorado Plateau provinces (Fig. 2). These faults have offsets of up to ∼1000 m, and are thought to have initiated during the early Miocene, with the Hurricane fault showing modern active seismicity (Maldonado et al., 1997;Davis, 1999;Lund et al., 2008;Biek et al., 2015).

Stratigraphic Evolution of Southern Utah
The SWHP preserve relatively undeformed Upper Cretaceous strata of the Cordilleran foreland basin system. These deposits include marine to fluvial-alluvial strata of the Naturita (formerly Dakota), Tropic, Straight Cliffs, Wahweap, and Kaiparowits Formations (Fig. 3;Peterson, 1969;Gustason, 1989;Eaton and Nations, 1991). The Straight Cliffs Formation, of primary interest to this study, overlies the marine Tropic Shale and records the initial regression of the Western Interior Seaway (Fig. 3, Peterson, 1969). Depositional environments of the Straight Cliffs Formation transition from proximal fluvial-alluvial systems west of the Markagunt Plateau to coalbearing marginal marine strata in the eastern Kaiparowits Plateau (Eaton and Nations, 1991;Little, 1997). Changes in fluvial and marginal marine stacking patterns within the Straight Cliffs Formation have been attributed to sealevel fluctuations  and/or interactions between axial and transverse drainages, which suggest tectonic and possibly climatic influences on deposition (Lawton et al., 2014;Szwarc et al., 2015;Gooley et al., 2016;Primm et al., 2018). Recent geochronologic data from the Straight Cliffs Formation indicate that deposition spans Turonian-early Campanian time (ca. 94-81 Ma; Primm et al., 2018;Fig. 3).
The Coniacian-Santonian John Henry Member comprises the thickest part of the Straight Cliffs Formation and preserves fluvial-alluvial deposits in the west, with marginal marine and offshore deposits in the eastern Kaiparowits Plateau (Peterson and Waldrop, 1965;Eaton and Nations, 1991;Shanley and McCabe, 1995;Hettinger et al., 1996). Marine deposits record shoreface and barrier island-lagoon systems with initially mainly transgressive then regressive stacking patterns throughout the Santonian (Hettinger et al., 1993;Shanley and McCabe, 1995;Allen and Johnson, 2010;Mulhern and Johnson, 2016). Extensive coal zones occur within the John Henry Member of the central Kaiparowits Plateau, which are correlated to fluvial sections in the west that vary in terms of fluvial style and tidal influence Hettinger et al., 2009). The upper boundary of the John Henry Member records a transition to higher-energy, coarser grained fluvial sheetdeposits of the basal Drip Tank Member (Lawton et al., 2003(Lawton et al., , 2014Gooley et al., 2016).
Unconformities throughout the Cretaceous sections are important considerations for interpreting thickness variations. Unconformities within the Straight Cliffs Formation are considered relatively minor sequence boundaries, with minimal erosion into underlying sections (Peterson, 1969;Biek et al., 2015;Primm et al., 2018). More significant unconformities across the SWHP occur in latest Campanian-Maastrichtian successions, between deposition of the Kaiparowits, Canaan Peak, and Claron Formations, and are principally located in the Paunsaugunt and Kaiparowits Plateaus ( Fig. 3; Goldstrand, 1994;Biek et al., 2015;Beveridge et al., 2020). These unconformities are largely attributed to Laramide-style deformation and formation of structures like the East Kaibab and Dutton monoclines (Roberts, 2007;Hilbert-Wolf et al., 2009;Lawton and Bradford, 2011). To date, there is little evidence that sub-Claronaged erosion significantly affected the Straight Cliffs Formation in this area (Lawton et al., 2003;Biek et al., 2015).

METHODS
Measured sections across the SWHP document stratigraphic architecture (Figs. 4, 5, and 6) as well as proximal-to-distal changes in thickness, grain size, and paleocurrent indicators (Fig. 7). Six new measured sections of the John Henry Member are combined with previous work from Tilton (1991) in the Paunsaugunt Plateau to understand basin-wide stratigraphic and thickness trends. Net to gross (NTG) is calculated by measuring sandstone thickness (net) relative to total section thickness (gross) and reported as a decimal number (0.0-1.0, net thickness/gross thickness). Paleocurrent data were collected from trough axes, ripples, and barforms where possible, and combined with previous work of Tilton (1991) to create a more holistic picture of paleoflow directions (Fig. 2).
Correlation across the SWHP from proximal to distal (west to east) is based on the formation descriptions of Peterson (1969) from the Kaiparowits Plateau, and extrapolated to the west with the aid of publications, theses, and regional reports (Tilton, 1991;Eaton and Nations, 1991;Moore and Straub, 2001;Lawton et al., 2014;Biek et al., 2015), as well as new measured sections presented herein. Regional marker beds used for correlation within the Straight Cliffs Formation include the Tibbet Canyon Member shoreface and the Calico Bed and Drip Tank Member gravelly sheet sands (Fig. 3). Although these widespread deposits are likely to be at least partly time-transgressive, previous studies and available age-control suggests they are reasonable datums relative to the age of the entire section and are suitable to guide regional correlations (Peterson, 1969;Lawton et al., 2014;Biek et al., 2015).

Lithofacies and Facies Associations
Thirteen lithofacies were identified through field observations (Table 1), and these were grouped into four facies associations (Table 2: FA-1 to FA-4) that are distinguished by distinct channel architectural patterns and recurring associations of lithofacies. Channel facies associations were distinguished based on channel architecture, bed boundary relationships, and grain-size distributions (e.g., Miall, 1996;Slingerland and Smith, 2004). Specific channel-fill elements include bedforms and individual bed boundary relationships that fill a single channel (e.g., scour-and-fill structures). Channel stories are defined as a single channel unit with related channel-fill elements, separated by conspicuous basal scours that are laterally continuous. Channel belts are defined as laterally and/or vertically coalescing channel stories that form a single elongate sandstone body (Gibling, 2006;Ford and Pyles, 2014).

Floodplain Facies Associations (FA-1 and FA-2)
Description. Facies association 1 (FA-1) consists of fine-grained lithofacies, including massive (Fm) and laminated mudstones (Fl) and siltstones with minor lenticular to tabular sandstones (Table 2; Fig. 4). Mudstones are expressed as structureless beds ranging from 0.1 to 30 m thick and are commonly light gray, but can be yellow, red, and tan in color (Fig. 4B, 4C). Locally, mudstones have blocky structures, Facies association 2 (FA-2; Fig. 4) consists of organic-rich fine-grained lithofacies (Fc) and coal (C) with minor components of isolated sandstones (Table 1: Fc and C; Fig. 4A, 4E). Fine-grained lithofacies of FA-2 are expressed as structureless beds ranging from red, yellow, light gray, and black to charcoal gray in color (Fm, Fl, Fc, C). Coal seams (subbituminous grade) are common and are up to   1 m thick in the northern Kaiparowits Plateau (Fig. 4A). Darker carbonaceous shales are present, with laterally variable oxidized lamina, and siderite concretions (Fig. 4A). Plant, coal, and woody fragments exist throughout the carbonaceous mudstones.
Both FA-1 and FA-2 contain distinctive fineto medium-grained, isolated sandstone units/ bedsets (Fig. 4D). These sandstone bedsets are up to 4 m thick and less than ∼10 m wide laterally. FA-1 and FA-2 sandstones commonly exhibit a lenticular geometry (both symmetric and asymmetric) but can also be tabular, and are no more than 2 stories (Fig. 4D). Individual beds are composed of very fine-to medium-grained sand. Bedsets thin upwards from 0.1 to 0.5 m thick. Sedimentary structures within FA-1 typically include soft sediment deformation (Sc), low-angle cross-lamination (Sl), and asymmetric ripples, particularly at the tops of sandstone bedsets (Fig. 4D). Sedimentary structures within FA-2 include asymmetric ripples, planar to lowangle cross-bedding (Sl), with no visible structures in the thinnest beds (Sm). Bioturbation is variable and infrequent for both FA-1 and FA-2, and where present, individual Skolithos traces are observed. Vertical successions of sandstones within FA-1 and FA-2 include basal beds with erosive bases, and trough cross-stratification A A¢  that transitions upwards to planar or low-angle cross stratified beds, capped by ripple-laminated sand beds.

Fine-grained Facies Associations (FA-1 and FA-2) Interpretation.
FA-1 represents welldrained floodplain deposition between active fluvial channels. Siltstone and mudstone units represent overbank fines deposited during flooding events when bankfull discharge is exceeded and channel levees are breached (Slingerland and Smith, 2004). The blocky mudstone structures, mottled coloring, and plant rootlets represent weakly developed paleosols and indicate brief periods of subaerial exposure between active channels and well drained floodplains (Kraus and Wells, 1999;Kraus, 2002). Sandstone lithofacies represent crevasse splay deposits with crevasse-channel development in some areas; these formed as the suspended and bedload material within a channel spills into the floodplain during levee-breaching flood events (Smith et al., 1989;Perez-Arlucea and Smith, 1999).
FA-2 represents poorly drained floodplain development between active fluvial channels (Davies-Vollum and Smith, 2008). Carbona-ceous material within mudstone and siltstone lithofacies of FA-2 suggests there was abundant organic matter present during floodplain deposition. Coal lithofacies indicate the substantial growth of vegetation and an elevated water table, interpreted as deposition in peat swamps during times of limited clastic input (Fielding, 1987;Shanley et al., 1992;Hettinger et al., 1996). This rapid plant accumulation occurred under humid tropical conditions, as was common during deposition of the John Henry Member (Wolfe and Upchurch, 1987;McCabe and Shanley, 1992). Sandstone lithofacies are interpreted as overbank splay deposits in which channels were breached during times of high discharge or flooding events (Slingerland and Smith, 2004;Donselaar et al., 2012). Lenticular and tabular sandstone lithofacies are interpreted as minor channel-fill successions, abandoned during channel migration or avulsion (Slingerland and Smith, 2004).

Channel-Fill Facies Associations (FA-3 and FA-4)
Two channel facies (FA-3 and FA-4) were distinguished based on the geometry of the sandstone body, specifically narrow and discrete channel belts versus broader sheet-like units.
Description FA-3. Facies association 3 (FA-3) consists of sandstone-rich lithofacies (Fig. 5), including very fine-to coarse-grained sandstones (St, Sr, Sc, Sp; Table 1) with infrequent gravels (Gt). Sandstone beds are 0.5-3 m thick, and form bedsets that are 5-7 m thick. Observed sedimentary structures include trough cross-beds that have a cross-bed height of 5-25 cm (St) and frequent soft sediment deformation, more commonly present in the thicker sandstone beds, including convolute bedding and dewatering structures (Sc; Figs. 5D). The margins of the larger channel belts (>3 m) preserve lateral accretion elements. Bioturbation is scarce and varies by location, with the most common identifiable traces being Skolithos and Asthenopodichnium (Moran et al., 2010). Areas with higher bioturbation intensity are typically unlined burrow clusters within sandstone channels (Fig. 5E).
Overall, channels within FA-3 exhibit lensshaped architecture and are composed of a sandstone-rich core with wings that thin laterally Very fine-to medium-grained sandstone; tabular-lenticular beds range from 1-3 m thick; 5-7 m thick bed sets; trough cross-beds and soft sediment deformation common; mud clasts up to 4 cm wide at base of erosive scours; channel fill includes barforms and channel lags. into adjacent overbank and floodplain deposits (Kjemperud et al., 2008). Channels extend laterally from 40 to 100 m (Figs. 5A, 5B). Channel-fill style varies from simple single channels to complex channel belts that comprise 6-10 stories separated by erosive scours (Ford and Pyles, 2014). Larger channel-fill elements can be strongly vertically amalgamated, incising into lower stories and preserving incomplete channel fill and barforms (cf. Chamberlin and Hajek, 2019). Mud rip-up clasts up to 4 cm in diameter are common at the base of these scour surfaces.
In general, there is a decrease in grain size, bed thickness, and bedform height, with an increase in mudstone content moving from the channel axis toward the channel wings (Fig. 5B). Lateral accretion sets occur along the margins of channels (Hassanpour et al., 2013). Description FA-4. Facies association 4 (FA-4; Fig. 6) typically consists of coarser-grained bedforms with fine to gravelly sandstone lithofacies (Table 2). Locally, gravel lithofacies are more common (Gt, Gm; Fig. 6D). Gravel clast sizes range from 2 mm to 4 cm, and commonly consist of subangular to angular gray to black chert, quartzite and carbonate lithic clasts, and mud rip-ups (Fig. 6D). These beds are often normally graded with larger grains and clasts at the base of trough cross-beds and erosional scours (Se, St). Sandstone beds are 0.1-2.5 m thick and bedsets (greater than 2 m) typically exhibit a fining upwards trend. Sedimentary structures observed within the sandstone and gravel-prone bedsets include trough cross-beds and convolute bedding (Se, St; Fig. 6B). Many bedsets contain thicker beds at the base that have trough-cross and convolute bedding, which are overlain by thinner, or "flaggy" bedding that is commonly ripple-laminated. Mud rip-up clasts, pebbles, and coarse-grained sandstones are common at the base of erosive scours and trough cross-beds, particularly at the base of channel bedsets. Siltstones and mudstones are infrequently preserved between channel stories (Fig. 6E).
Bioturbation is variable, ranging from scarce to common, with occurrences of Skolithos, Asthenopodichnium (Moran et al., 2010) and escape traces. Bioturbation intensity is variable across the plateaus and can be relatively high locally (∼15% bioturbated). Wood casts (cmscale to >1 m in length) are commonly preserved at the base of erosive scours. Asthenopodichnium traces are visible within wood casts. Casts of therapod and ornithopod tracks are preserved on the bases of some channel deposits (e.g., Parowan Canyon; Milner et al., 2006). Crocodilian, dinosaur, and turtle bone fragments erode out of sandstone and siltstones of FA-4 (Eaton, 1999). Concretions 0.3-1.0 m thick are commonly found in the larger sandstone beds. Sandstone architecture is dominated by laterally extensive tabular sandstones, with erosive scours between stories (Figs. 6A, 6E). Tabular sandstones are often vertically and laterally amalgamated forming laterally extensive sheet complexes, but lateral and vertical channel amalgamation varies by location. Some sheet deposits form channel belts that are laterally extensive and can be mapped for 100s of meters with >5 stories identifiable, e.g., the Calico Bed (Fig. 6E). In contrast, some channel belts have a reduced lateral extent (10s of meters), but instead have high vertical amalgamation (typically between 1 and 3 stories) with rare finegrained deposits preserved (Fig. 6A). Tabular sandy channel belts are commonly 1-3 m thick but can be up to ∼5 m thick.

Channel-Fill Facies Associations (FA-3 and FA-4) Interpretation
FA-3 Interpretation. Channel-fill successions of FA-3 represent deposition of discrete channel belts within well-developed floodplain, similar to the "steers-head" channels of Kjemperud et al. (2008). Channel deposits of FA-3 represent fluvial systems with a higher sediment suspended load and lateral bank stability, interpreted as distributary channels draining into floodplain deposits (Figs. 5A, 5B). Distinct channel belt forms suggest bank stability, allowing for lateral and vertical aggradational deposition within the channel margins (Gibling, 2006).
Channel-fill facies are locally adjacent to lateral-accretion elements, although complete barform preservation is rare. The lateral wings of channels are typically thinner-bedded and finer grained with a higher preservation of interbedded mudstone and siltstones. These are interpreted as channel levees, where sediment spilled onto overbank regions (Allen, 1978;Friend et al., 1979). Channel levee development and preservation indicates a high level of bank stability that is concurrent with the higher frequency of vegetated floodplain deposition seen in FA-3 (Slingerland and Smith, 2004). Additionally, preservation of overbank and inter-channel fine-grained material suggests there was a higher volume of suspended sediment load relative to the sandrich sheet deposits that comprise FA-4. Lateral accretion bedsets and barforms are infrequently preserved and vary by location, but where present indicate some level of reworking within channel margins (Mohrig et al., 2000;Hassanpour et al., 2013). These fixed-location channels could indicate a less mobile river system, with less frequent avulsion and stable, well-developed river channels (Straub and Esposito, 2013).
FA-4 Interpretation. Sheet-like sandstone deposits of FA-4 represent unconfined flow within poorly developed fluvial channels dominated by sandy to gravelly bedload. The multi-story character of these deposits records episodes of cutting and filling, likely a result of fluctuations in discharge causing incision during high discharge events, and subsequent deposition during waning flow (Holbrook, 2001). A lack of distinct channel boundaries and laterally extensive tabular geometries indicate lateral instability and a dominance of unconfined flow (Smith et al., 1989;Owen et al., 2017). The coarse-grained nature of the channel fills, numerous mud-rip-ups, and gravelly lags suggests high discharge with frequent bankfull flows (Gibling, 2006). The cut-and-fill patterns combined with coarser grained channel fill may also indicate some level of internal reworking, suggesting a fluvial system that migrated or avulsed more frequently (Straub and Esposito, 2013) or during relatively slow accommodation creation (Miall and Arush, 2001).

Regional Stratigraphic Patterns
Figure 7 shows a regional correlation composed of six new measured sections of the John Henry Member combined with two sections from Tilton (1991;Glendale and Tenny Canyon). Sections range from the northern Kaiparowits Plateau (Table Cliffs) region, to the east and southern flanks of the Paunsaugunt Plateau (PP; Fig. 2), and into the central and western Markagunt Plateau (MP; Fig. 2).

Markagunt Plateau
Measured sections from the northwestern edge of the Markagunt Plateau (Parowan Gap and Parowan Canyon; Fig. 7) represent the most proximal strata in the study area. Near the Wah Wah and Blue Mountain thrust faults (Fig. 8), the Iron Springs Formation thickens from at least 350 m as measured in this study (Parowan Canyon), and up to ∼900 m thick to the southwest (Fig. 8, Gunlock and Pine Valley Mountains; Cook, 1957;Hintze, 1986;Fillmore, 1991). The base of the Straight Cliffs Formation (Tibbet Canyon and Smoky Hollow Members) is not exposed in Parowan Gap. There, the partially exposed upper John Henry Member is ∼80 m thick (Fig. 7), and this section thickens to ∼360 m in Parowan Canyon (∼15 km southeast; Fig. 7). In the central part of the Markagunt Plateau, the Smoky Hollow Member ranges from ∼110-117 m thick (Orderville Gulch, Fig. 7). The Calico Bed has a well-developed ∼17-m-thick lower bed but pinches out along the western margin of the Markagunt Plateau. The John Henry Member has an average thickness of ∼200 m along the south-central margin of the Markagunt Plateau.

Paunsaugunt Plateau
Thinning of the John Henry and Smoky Hollow Members occurs across the Sevier and Paunsaugunt faults (Fig. 7). Within the Paunsaugunt Plateau, the Smoky Hollow Member is roughly half as thick as it is to the west (∼30 m). John Henry Member strata thin across the Sevier fault from ∼200 m to ∼160 m (Glendale, Fig. 7; Tilton, 1991). This thickness is maintained along the western and southern flanks of the Paunsaugunt Plateau ( Fig. 7; Tilton, 1991). Localized thickening occurs along the northeastern edge of the Paunsaugunt Plateau, in both the John Henry Member (∼160 to ∼270 m) and the Smoky Hollow Member (∼30 m to ∼60 m; Heward Creek, Fig. 7).

Kaiparowits Plateau
Across the Paunsaugunt fault, strata from the Smoky Hollow Member and John Henry Member strata thicken across the Paunsaugunt fault, in the northern Kaiparowits Plateau-  Fig. 7). These thickness changes continue into the paralic and shallow-marine John Henry Member strata to the east, which are ∼460 m thick along the northeastern margin of the Kaiparowits Plateau in Buck Hollow (Mulhern and Johnson, 2016). The John Henry Mem-ber in the southern portion of the Kaiparowits Plateau is only ∼200 m thick (Gallin et al., 2010;Allen and Johnson, 2010;Dooling, 2013;Pettinga, 2013;Gooley et al., 2016).

Markagunt Plateau
Fluvial architecture in the most proximal John Henry Member sections at Parowan Gap and Parowan Canyon is characterized by sheet deposits of FA-4 intercalated with minor components of FA-1 and some FA-3 (Fig. 7). These strata are typically composed of medium-to coarse-grained sandstones with pebble lags at the base of scours. Channel belts are vertically amalgamated with minor floodplain development; NTG in these proximal sections is ∼0.64 and ∼0.48 (Fig. 7).
In contrast, across the rest of the Markagunt Plateau, the John Henry Member consists of well-drained floodplain material (FA-1) with isolated sandstone channels of FA-3 ( Fig. 7; Orderville Gulch). Instances of FA-4 within the middle John Henry Member can be traced laterally for 100s of meters. Sheet deposits of FA-4 within the John Henry Member are composed of fine-to medium-grained sandstone and lack pebble lags. NTG estimates of the John Henry Member decrease from ∼0.56 along the western margin of the plateau to ∼0.37 in the central plateau. Overall paleocurrent indicators are oriented north/northeast (average paleoflow ∼055; n = 120; Fig. 7).

Paunsaugunt Plateau
Along the western and southern margins of the Paunsaugunt plateau (Glendale and Tenny Canyon; Fig. 7), the John Henry Member consists of laterally and vertically amalgamated sheet-like deposits ( Fig. 6; FA-4) with minor components of isolated channel belts within well-developed floodplain deposits (FA-1 and FA-3). This architecture changes dramatically in the northeast margin near the Paunsaugunt fault ( Fig. 7; Heward Creek), where the John Henry Member consists primarily of FA-3 with minor components of FA-1. Here, channel belts of FA-3 are typically ∼8 stories high and situated within well-developed gray to red mudstone deposits (Fig. 5A, B). Typical bedforms are difficult to identify in many of the thickest channels due to the abundance of convolute and slumped bedding disrupting bedding structures (Fig. 5D). Overall, the average NTG of the John Henry Member in the south and southwest of the Paunsaugunt Plateau is ∼0.64 (average of Tenny Canyon and Glendale), but drops to ∼0.32 at Heward Creek. The average paleocurrent direction is to the northeast, with no significant changes in trends across the plateau (Tilton, 1991;Lawton et al., 2014; Fig. 7).

Kaiparowits Plateau
At Shakespeare Mine in the northern Kaiparowits Plateau (Table Cliffs)  amalgamation increases both laterally and vertically up-section, transitioning from small, isolated channels of FA-2, to larger defined channels of FA-3, and to sheet-like sandstones of FA-4. Fluvial John Henry Member NTG in the northern Kaiparowits region is ∼0.87. Overall paleocurrent direction is to the east-northeast (Fig. 7).

DISCUSSION
Stratigraphic data presented here show increasing thickness and NTG to the east in distal parts of the foredeep, which raises questions regarding controls on basin architecture and evolution of the Late Cretaceous southern Utah foreland. This discussion focuses primarily on thickness and facies variations of the fluvial John Henry Member of the Straight Cliffs Formation, and factors controlling sediment supply and accommodation within the foreland basin.

Thickness Trends
Regional stratigraphic correlations of Upper Cretaceous strata across the SWHP of southern Utah indicate abrupt, stepwise thickness variations in key parts of the section that coincide with modern plateau-bounding faults (Figs. 8,9). The most proximal sections of the Iron Springs Formation (e.g., Parowan Canyon) are at least 350 m thick, whereas the age-equivalent Straight Cliffs Formation thins by 50% to the east across the Markagunt and Paunsaugunt Plateaus (average thickness ∼175 m). This proximal eastward-thinning trend is reversed across the Paunsaugunt Plateau where John Henry Member sections thicken from ∼270 m at Heward Creek to over ∼450 m at Shakespeare Mine over a distance of ∼25 km (Fig. 7). This thickening into the distal foredeep also occurs from south to north across the Kaiparowits Plateau in the John Henry Member, as well as the overlying Wahweap and Kaiparowits Formations ( Fig. 9B; Lawton et al., 2003;Gallin et al., 2010;Gooley et al., 2016).
Present-day stratigraphic thicknesses are reported here, and thus are minimum sediment accumulation estimates. The observed eastward-and northward-thickening trends in the Kaiparowits Plateau (Fig. 9) would be exaggerated by decompaction, given the concentration of mudstone and coal deposits in paralic and marine sections within the John Henry Member (Hettinger, 2000;Hettinger et al., 2009). Mulhern and Johnson (2016) estimated the entire Straight Cliffs Formation could be expanded from 456 m to over 800 m at Buck Hollow (Fig. 2), even with conservative decompaction estimates.
As mentioned previously, intraformational unconformities and inferred sequence boundaries in Upper Cretaceous sections are thought to be relatively minor in terms of duration, potential eroded section, and angular stratigraphic relationships (Peterson, 1969;Gustason, 1989;Tilton, 1991;Shanley and McCabe, 1995;Lawton et al., 2003Lawton et al., , 2014Szwarc et al., 2015;Primm et al., 2018). While the sub-Claron unconformity is regionally significant, available data suggest that the Kaiparowits Formation was mainly deposited in an isolated depocenter located in the northern Kaiparowits Plateau ( Fig. 9; Eaton et al., 1993;Roberts, 2007;Sampson et al., 2013;Biek et al., 2015) and that variations in thickness are not entirely a function of subsequent erosion. Thus, observed stratigraphic thicknesses of Cretaceous sections in the SWHPs likely represent paleo-accommodation variations.

Facies Trends
Using fluvial architecture and NTG as proxies for grain size and facies trends across the foreland basin (Fig. 9A), results show consistent "distal fining" only from the most proximal Iron Springs Formation (NTG 0.64) to the Straight Cliffs Formation in the Markagunt Plateau, with its well-developed floodplain deposits and isolated fluvial channel belts (NTG 0.37). In the Paunsaugunt Plateau, distal fining trends become more complex. Across the Sevier fault in the western and southern Paunsaugunt Plateau, channel belt amalgamation increases, becoming more sheet-like with less floodplain deposition (NTG 0.64). In contrast, to the northeast along the Paunsaugunt fault (e.g., Heward Creek; Fig. 7), the John Henry Member section becomes thicker with much higher floodplain deposition and well-developed isolated channels (NTG 0.32). Coincident with dramatic thickening across the Paunsaugunt fault, the John Henry Member sections in the northern Kaiparowits Plateau (Shakespeare Mine; Fig. 7) have a notably high NTG (∼0.87) with highly amalgamated fluvial channel belts and a few well-developed coal zones primarily in the lower parts of the section ( Fig. 4; Hettinger, 2000;Hettinger et al., 2009).
These Late Cretaceous facies trends can be partly explained by fluvial drainage networks that were aligned subparallel to the trend of the fold-thrust belt in southern Utah, thus implying an axial fluvial system (Fig. 10). Direct evidence for orogen-transverse drainage patterns is essentially limited to the most proximal deposits of the Iron Springs Formation (Gunlock and Pine Valleys; Fig. 8; Fillmore, 1991;Goldstrand, 1991). In addition, episodic progradation of transverse fluvial systems across the foreland is recorded in extensive "sheet-sands" that punctuate the Cretaceous stratigraphy in the area (e.g., Calico Bed, Drip Tank Member, and Wahweap Formation;Lawton et al., 2003Lawton et al., , 2014Lawton and Bradford, 2011;Szwarc et al., 2015;Primm et al., 2018). These distinctive and widespread units typically have east-directed paleocurrent indicators and provenance signatures that indicate a Sevier fold-thrust belt source.
Thus, evidence for orogen-transverse drainage systems appears to be localized and episodic relative to most of the Cretaceous foreland basin succession in southern Utah. Previous interpretations of major axial fluvial systems are partly based on facies distributions combined with paleocurrent analyses, which show dominantly northeast transport in Upper Cretaceous fluvial sections of the Kaiparowits Plateau (Szwarc et al., 2015;Gooley et al., 2016). Our data compilation confirms overall northeast-directed paleocurrent directions for the John Henry Member across the Markagunt and Paunsaugunt Plateaus (Fig. 7). Similarly, U-Pb detrital zircon geochronology of the Straight Cliffs Formation in the Kaiparowits Plateau (Szwarc et al., 2015;Primm et al., 2018) indicates relatively minor detrital zircon input from the Sevier fold-thrust belt. Rather, major zircon source areas lie to the south (e.g., 1.4 and 1.7 Ga zircon populations sourced from the Mogollon Highlands) and southwest (e.g., ca. 147 Ma zircon populations likely sourced from the Mojave Dike Swarm in southeastern California; Fig. 1).
In sum, fluvial sections in the Straight Cliffs Formation show distinct facies and thickness variations by location; these features are oriented sub-parallel to the fold-thrust belt and coincide with modern plateau-bounding faults. The interpretation of these basin-axial corridors suggests some allogenic control on local accommodation rather than a continuous foredeep throughout the Late Cretaceous.

Flexural Loading
Previous studies have interpreted the Upper Jurassic-Lower Cretaceous Morrison, Carmel, and Cedar Mountain Formations to record forebulge migration and deposition, with the forebulge lying just east of the Kaiparowits Plateau by Cenomanian or Turonian time (Royse, 1993;Currie, 1998Currie, , 2002DeCelles, 2004). An important implication of this interpretation is that high Late Cretaceous sediment accumulation rates of the Kaiparowits Plateau appear to have formed in the distal foredeep to forebulge depozones, 150-200 km east of the Wah Wah and Blue Mountain thrust systems. This pattern is particularly odd given thinning of time-equivalent strata to the west in more proximal parts of the foredeep, where thicker sediment accumulation would be expected (Fig. 9).
Differential loading could have increased flexural accommodation in the northern Kaiparow-its Plateau due to movement along the Pavant/ Paxton thrust sheets of central Utah, ∼75 km to the northwest (Fig. 1B). The thrusting style of the Pavant/Paxton system consisted of duplexing and culmination growth during the Coniacian-Santonian, simultaneous with deposition of the John Henry Member (DeCelles et al., 1995;Coogan and DeCelles, 1996). During this time, sediment supply into the central Utah foreland basin was likely high, but provenance data and evidence for dominant paleoflow to the east and northeast suggest the majority of sediment was being transported away from the Kaiparowits Plateau (Lawton, 1983;DeCelles et al., 1995;Fielding et al., 2010). Furthermore, a 47% expansion of the John Henry Member occurs over just 15 km between Main Canyon and Buck Hollow, too short of a distance to relate to regional flexural patterns ( Fig. 2; Chentnik et al., 2015;Mulhern and Johnson, 2016).

Blind Thrust Faults
Given evidence for early emplacement and a relatively stationary fold-thrust belt in the Wah Wah and Blue Mountain thrust systems, the possibility of buried Sevier-related thrust faults may be considered to explain observed foredeep sedimentation patterns. No subsurface thrusts have been reported across the SWHP during Late Cretaceous time (Cook and Hardman, 1967;Van Kooten, 1988), nor is there outcropping evidence of significant detachment horizons. Late-breaking thrust faults at Parowan Gap and Ruby's Inn likely occurred after deposition of the Straight Cliffs Formation (Merle et al., 1993). Furthermore, we note that studies of wedge-top basins indicate very different stratigraphic patterns than those documented here, including common progressive unconformities and coarse-grained alluvial deposits with local provenance (Lawton and Trexler, 1991;Horton, 1998;Quinn et al., 2018).

Dynamic Load
Subsidence mechanisms across the southern Utah foreland can be attributed to flexural loading of the Sevier fold-thrust belt and dynamic processes (Jordan, 1981;Mitrovica et al., 1989;Pang and Nummedal, 1995;Liu and Nummedal, 2004;. During the Coniacian to Santonian (i.e., during deposition of the John Henry Member), subsidence within the foreland is thought to transition from a narrow flexural profile to a longer-wavelength foreland basin (Roberts and Kirschbaum, 1995). Dynamic subsidence likely contributed 10s of meters of subsidence across the entire southern Utah foreland basin, a wavelength of 100s of kilometers (Painter and Carrapa, 2013;Heller and Liu, 2016). This effect is on a much broader scale than the evidence for localized variable accommodation presented in this study (10s-100s of m variable thickness over <25 km distance). These abrupt thickness changes suggest that dynamic subsidence alone cannot explain evidence for discrete sub-basins in the distal foredeep.

Laramide-style Tectonics
The idea of sub-basin formation within the Cordilleran foreland basin is commonly cast in the context of the Laramide orogeny, or "broken foreland" of Dickinson et al. (1988). Laramidestyle, basement-involved uplifts exist on the margins of the SWHP, including the Kanarra Fold to the west, and the East Kaibab monocline and Circle Cliffs uplifts to the east (Fig. 1). The earliest evidence of this foreland break-up in southern Utah includes growth faults in the Campanian Wahweap Formation in the northern Kaiparowits Plateau (e.g., Hilbert-Wolf et al., 2009;Simpson et al., 2014). One of the main drivers for Laramide tectonism is flat-slab subduction, which is thought to have initiated at the latitude of southern Utah during the earliest Campanian (Simpson et al., 2014). Inverse mantle convection models show the flat-slab segment, likely the conjugate of the Shatsky Rise, entering the subduction zone along the Mojave area of California at ca. 90 Ma, followed by northeast underthrusting beneath Arizona to Wyoming during the latest Cretaceous Liu and Currie, 2016). Although this is a plausible driving mechanism, timing estimates would place onset of Laramide-style deformation up to ∼20 m.y. later than the suspected formation of sub-basins within the Straight Cliffs Formation and thickness changes in the Naturita Formation. This timing is similar to the emergence of basement cored structures of the northern foreland in southwest Montana, where thermochronologic evidence shows exhumation of Laramide structures occurred prior to ca. 80 Ma, and as early as ca. 100-120 Ma (Schwartz and DeCelles, 1988;Carrapa et al., 2019;Garber et al., 2020;Orme, 2020). The early (Aptian-Albian) onset of Laramide-style deformation is similar to our earliest estimates of foreland partitioning from the Cenomanian Naturita Formation (Gustason, 1989), and argues against migration of the Shatsky rise as a driver of Laramide deformation (Carrapa et al., 2019;Garber et al., 2020). Furthermore, classic Laramide-style "broken foreland" basins (e.g., the Uinta basin) are generally much larger than the sub-basins now exposed along the SWHP (Picard and High, 1968;Dickinson et al., 1988).

Structural Inheritance and Normal Faults
The location of modern plateau-bounding faults was likely influenced by preexisting crustal weaknesses. Inherited rift-related basement faults across southern Utah (Fig. 1A) are known to have been reactivated during multiple tectonic events throughout the Phanerozoic (Schwartz, 1982;Picha and Gibson, 1985). Evidence for Late Cretaceous-early Cenozoic activity along the Hurricane fault (Cook and Hardman, 1967) suggests that other SWHP-bounding faults may have been prone to structural inheritance controls and reactivation as well. We suggest that reactivation of Proterozoic basement features, possibly as normal faults, was a factor in early partitioning of the Cretaceous southern Utah foreland basin. Structural variations within the crust of foreland basins, like those in southern Utah, have been shown to influence crustal deflection and evolution of the foredeep (e.g., Romans et al., 2010). In some cases, they may cause segmentation of the foreland (Waschbusch and Royden, 1992) or rapid migration of the foredeep depocenter (DeCelles et al., 1995). For example, eastward propagation of the central Utah thrust sheets may have been aided by the Ancient Ephraim fault and other basement structures (Fig. 1). The northeast-trending Paragonah lineament separates the southern and central Utah fold-thrust belt and is speculated to have offset the southern Utah thrust sheets, inhibiting their further propagation to the east (Picha, 1986). In the Magallanes foreland basin in Argentina, pre-foreland attenuated lithosphere led to atypical foredeep subsidence, amplified by sediment loading (Romans et al., 2010;Fosdick et al., 2014;Gianni et al., 2015).
Multiple mechanisms have been suggested for normal faulting associated with flexure. Extension observed along the outer rise in subduction zones may provide a simple conceptual analog (Ludwig et al., 1966). The Taconic peripheral foreland basin was accompanied by motion along normal faults due to inelastic extensional deformation on the convex side of the flexed plate, which suggests bending of the lithosphere beyond its elastic limit (Bradley and Kidd, 1991). A possible expression of flexural foundering has also been documented in the Apulian foreland basin in Italy, where systematic joints formed along the forebulge due to flexure-related fiber stresses (Billi and Salvini, 2003). This flexural foundering within the forebulge can alter typical basin depocenters by stopping forebulge migration (Waschbusch and Royden, 1992). Normal faulting due to extensional stresses within flexural foreland basins is likely more common than is recognized because it may present as subtle shifts in stratigraphic architecture that cannot be explained by other allogenic or autogenic signals (Bradley and Kidd, 1991;Waschbusch and Royden, 1992;Londoño and Lorenzo, 2004;Gianni et al., 2015).

Syndepositional Tectonic History-An Integrated Model
Traditional models of foreland basin evolution conflict with aspects of stratigraphic architecture and depocenters across the southern Utah Cretaceous foreland basin. Specifically, thickening and increases in NTG in the distal foredeep or forebulge area (the northern Kaiparowits Plateau) is unusual for classic foreland basin models. We propose that inherited structural heterogeneities within the crust restricted typical foredeep migration and may have favored the formation of normal faults on the bending plate (Fig. 10), which ultimately triggered formation of isolated sub-basins within the southern Utah foreland (Fig. 9). Restricted foreland basin migration could have increased local stress within the plate allowing for faulting to occur along preexisting heterogeneities (e.g., Bradley and Kidd, 1991;Hudson, 2000).
Even given uncertainty regarding their orientation and kinematics, we infer that movement along inherited structures during Late Cretaceous time influenced sediment transport and accumulation within the foreland basin (Fig. 10). In some cases, changes in thickness and fluvial architecture are abrupt on either side of modern plateau-bounding structures such as the Sevier and Paunsaugunt normal faults . Sub-basin formation partitioned the axial fluvial system into distinct northeast-trending fluvial facies belts ( Fig. 10) with variable NTG and channel belt architectures preserved. Thinning of foreland basin strata across the Markagunt and Paunsaugunt Plateaus is reversed in the most distal parts of the foredeep, particularly in the northern Kaiparowits Plateau. Similar evidence for foredeep partitioning is reported as early as Cenomanian time (Gustason, 1989;Eaton and Nations, 1991), and we interpret these signals to have intensified by Santonian time during deposition of the John Henry Member.

CONCLUSIONS
New stratigraphic data from the Straight Cliffs Formation of the SWHP, including sediment thickness, facies architecture, and net to gross estimates, indicate that the Cretaceous southern Utah foredeep was divided into discrete subbasins at least 10-20 million years prior to onset of known local Laramide-style deformation. Flexural foundering at the edge of the Colorado Plateau may have favored sub-basin formation within the foreland, thinning strata across the Paunsaugunt Plateau, and dramatically thickening strata in the northern Kaiparowits Plateau. We suggest that the Paunsaugunt, Sevier, and Hurricane faults are the modern expressions of these inherited features, which ultimately reflect lithospheric heterogeneities formed in the Proterozoic. These findings demonstrate that "broken forelands" evolve different basin geometries, sedimentary facies, and stratigraphic stacking patterns than those predicted by traditional models of foreland basin systems.

ACKNOWLEDGMENTS
This work would not be possible without the technical guidance, logistical support, and immense insight into southern Utah stratigraphy provided by Jeff Eaton. This research was supported by the National Science Foundation Graduate Research Fellowship Program, AAPG student grants, and industry supporters of the Rocks2Models consortium, in particular ConocoPhillips. Thank you to the Bureau of Land Management, the Grand Staircase-Escalante National Monument, and the Dixie National Forest for permits to conduct research in the Kaiparowits, Paunsaugunt, and Markagunt Plateau region. We would like to acknowledge that although we are using state and local names for location markers, it is important to remember that these places are known in many ways, and this research was done on the land of the Southern Paiute, Pueblo, Dine, Western Shoshone, and Ute peoples, and others that we may be unaware of. We want to recognize that in western and especially academic and STEM spaces, indigenous people's ability to describe and name places have been diminished if not fully taken. Last, we are also grateful to the University of Utah Basins Research Group, Magdalena Curry, and an anonymous reviewer, as well as John Byrd for their thoughtful feedback and suggestions.