Transition from Late Jurassic rifting to middle Cretaceous dynamic foreland, southwestern U.S. and northwestern Mexico Open Access

A fundamental plate tectonic reorganization took place during middle Cretaceous (Albian–Cenomanian) time in the southwest U.S.–Mexico border region immediately following Jurassic–Early Cretaceous continental rifting (Bilodeau, 1982; Lawton and McMillan, 1999; Dickinson and Lawton, 2001b). Prior to the middle Cretaceous, an extensional tectonic setting of northern Sonora, southern Arizona, and southwestern New Mexico contrasted with regional tectonics to the north, where crustal shortening of the Sevier orogeny was well under way by Aptian time in southern Nevada, Utah, and Wyoming (Armstrong, 1968; Heller et al., 1986; Lawton, 1994; DeCelles, 2004; Yonkee and Weil, 2015). Flexural subsidence adjacent to the Sevier orogenic belt created an adjacent retroarc foreland basin (sensu Ingersoll, 2012), termed the Cordilleran foreland basin, whose foredeep lay along the western flank of the epicontinental Western Interior seaway (Jordan, 1981; Lawton, 1982, 1994; Robinson Roberts and Kirschbaum, 1995; Currie, 1997, 1998). Dynamic topography created by the subducted Farallon slab beneath western North America likely contributed to long-wavelength subsidence in the central and eastern parts of the basin (e.g., Cross, 1986; Pang and Nummedal, 1995; Nummedal, 2004). In comparison, development of a retroarc foreland basin in the U.S.–Mexico border region, as inferred from subsidence analysis of the Lower Cretaceous section in southwestern New Mexico (Mack, 1987a), did not begin until late Albian time, or at least 20 m.y. later than farther north.

Studies of foreland basin depositional history and tectonics have tended to focus on latitudes north of Las Vegas, Nevada, as demonstrated by extensive reviews of basin history adjacent to the Sevier orogenic belt (Lawton, 1994; DeCelles, 2004; Yonkee and Weil, 2015). In contrast, mechanisms of Late Jurassic to middle Cretaceous sedimentary basin development in the U.S.–Mexico border region of Arizona, Sonora, and southwestern New Mexico, and their relationship with the Cordilleran foreland basin, remain poorly understood because of less extensive study and uncertainty regarding age and depositional setting of the Upper Jurassic–Cretaceous section. Basin geometry and stratigraphic architecture are extensively concealed by younger sedimentary and volcanic rocks and overprinted by subsequent deformation and magmatism (e.g., Soreghan, 1998; McKee et al., 2005; Amato et al., 2009; Mauel et al., 2011; González-León et al., 2011; Clinkscales and Lawton, 2018). In addition, the middle Cretaceous foreland basin formed on continental crust only newly extended during Jurassic–Cretaceous rifting (Bilodeau, 1982; Dickinson et al., 1986; Lawton and McMillan, 1999; Lawton, 2000, 2004; Dickinson and Lawton, 2001b). Basin geometry, forebulge development, and the migration history of foreland basins formed on recently rifted lithosphere differ significantly from equivalent characteristics of basins developed upon old lithosphere that behaves like an unbroken elastic plate (Fildani and Hessler, 2005; Fosdick et al., 2014), potentially rendering inception of foreland history difficult to recognize.

The Bisbee basin, an important archive of the sedimentary history of the U.S.–Mexico border region, initially formed as part of the Late Jurassic Mexican Border rift system (Fig. 1; MBR, originally termed Mexican Borderland rift; Lawton and McMillan, 1999). The MBR extended from the McCoy basin of southern California and southwestern Arizona (Fig. 1) across southern Arizona, northern Sonora, and southwestern New Mexico, then southeastward to the Gulf of Mexico via the Chihuahua trough of northern Mexico (Dickinson and Lawton, 2001b; Haenggi, 2002; Lawton, 2004; Mickus et al., 2009; Stern and Dickinson, 2010; Spencer et al., 2011). The MBR began to form in Late Jurassic time as a northwest- and west-northwest–trending assemblage of rift basins created by the combined effects of rollback of an oceanic slab beneath southwestern North America (Lawton and McMillan, 1999; Dickinson and Lawton, 2001a; Fitz-Díaz et al., 2018) and separation of North and South America during the breakup of Pangea (e.g., Pindell and Kennan, 2009). Sedimentary basins of the MBR are alternatively interpreted as an array of pull-apart basins formed along a throughgoing Middle–Late Jurassic (Anderson, 2015) or Late Jurassic (Anderson and Silver, 2005) sinistral transcurrent fault, the Mojave-Sonora megashear, which is inferred to have accommodated the opening of the Gulf of Mexico (Anderson and Nourse, 2005), Importantly, Late Jurassic–middle Cretaceous deposition in the Bisbee basin continued beyond the extensional history of the MBR, a theme we develop in this paper.

West and southwest of the MBR in Mexico, Late Jurassic–middle Cretaceous magmatism took place without apparent interruption along the North American continental margin. Magmatic history is inferred on the basis of zircon grain ages in the adjacent retroarc foreland basin of northern Mexico, which are common in the range ca. 145–123 Ma, less common in the range ca. 122–110 Ma, and abundant in the range ca. 105–88 Ma (Juárez-Arriaga et al., 2019). In the southwestern-most U.S. and along western Mexico, Early Cretaceous arc magmatism occurred in the Guerrero composite terrane, which collided diachronously with North America during a debated time interval, probably sometime in the Aptian–Cenomanian (Dickinson and Lawton, 2001a; Centeno-García et al., 2008; Martini et al., 2014). Magmatism in the Sierra Nevada of eastern California, which is generally inferred to have been caused by east-dipping subduction of the Farallon slab beneath the North American plate, experienced a mid-Cretaceous magmatic lull, after which it became active beginning at ca. 123 Ma and peaked in volume between 100 Ma and 90 Ma (Ducea, 2001; DeCelles et al., 2009). A compilation of 430 U-Pb ages from the Sierra Nevada batholith south of 37.25°N demonstrates a nearly continuous range of ages from 248 Ma to 76 Ma but contains only 11 ages (2.6% of the total) in the range 145–124 Ma (Nadin et al., 2016). Ages in the range 140–120 Ma lie along the western flank of the batholith (Nadin et al., 2016). As a result of an evident low volume of pre-Aptian Early Cretaceous magmatism, and possibly because of its distribution within the batholith, ca. 143–124 Ma zircon grains are uncommon in Cretaceous strata of the Cordilleran foreland basin in the United States (Laskowski et al., 2013), in contrast with grain age distributions in the foreland basin of northern Mexico (Juárez-Arriaga et al., 2019).

In this paper, we integrate new and published stratigraphic, petrographic, and geochronologic data for Oxfordian–middle Cenomanian strata in southern Arizona, northern Sonora, and southwestern and northwestern New Mexico to interpret Late Jurassic through middle Cretaceous basin development of the southwestern U.S.–Mexico border region. Stratigraphic correlation, subcrop relations, petrography, and U-Pb detrital-zircon and ash-fall tuff ages permit an improved understanding of the sedimentary history of southwestern North America and the nascent Western Interior seaway and provide insight into the tectonics of foreland history that followed crustal extension. We use the new dates and previously published biostratigraphic ages to construct geohistory diagrams, or subsidence curves, for Upper Jurassic–Cenomanian strata in southwestern New Mexico and northern Sonora. Published subsidence histories for Bisbee basin strata in southwestern New Mexico (Mack, 1987a) and Sonora (González-León, 1994) included only Early Cretaceous strata because Upper Jurassic strata from New Mexico were not yet described; in addition, correlative Jurassic strata in Sonora (e.g., Rangin, 1977) had not yet been factored into the stratigraphic history of the basin. Improved biostratigraphic and geochronologic data published over the intervening 25 years and new data presented here yield similar subsidence histories from the two locations. The regionally consistent subsidence curves suggest that strata of the Bisbee basin, as defined above, record regional Late Jurassic–middle Albian crustal extension and passive thermo-tectonic subsidence followed by rapid late Albian–Cenomanian subsidence, likely of combined flexural and dynamic origin. We infer a tectonic history in southwestern North America of rift basin subsidence that accompanied Late Jurassic separation of the Guerrero composite volcanic terrane (abbreviated as Guerrero; e.g., Dickinson and Lawton, 2001a) from the continental margin during formation of a marginal backarc basin southward from the latitude of Ensenada, Baja California (in present coordinates; Fig. 1). Rifting was succeeded by broad foreland subsidence during progressive closure of the intervening marginal basin and middle to late Albian collision of Guerrero with the continental margin, combined with progressive southeastward reintroduction of the subducted Farallon slab beneath continental lithosphere of the southwestern corner of North America.

Most strata in this study pertain to the Bisbee Group, defined from the Bisbee mining district in southeastern Arizona (Fig. 1; Ransome, 1904). The type Bisbee Group includes the Glance Conglomerate and Morita, Mural, and Cintura Formations, which span latest Jurassic–late Early Cretaceous time (Ransome, 1904; Stoyanow, 1949; Bilodeau and Lindberg, 1983; Dickinson et al., 1986). Following recognition that the Glance Conglomerate records syndepositional normal faulting (Bilodeau, 1979, 1982), the Bisbee basin and its eponymous fill came to be regarded as products of continental rifting (Bilodeau, 1979, 1982; Dickinson et al., 1986; Lawton and McMillan, 1999). Thinning and pinch-out of the Glance Conglomerate in the Mule Mountains north of Bisbee, Arizona, indicate the presence of a topographic high of incompletely defined distribution within the greater Bisbee basin (Fig. 1; Bilodeau, 1982; Bilodeau et al., 1987). A part of this paleohigh, where it is expressed in the Mule Mountains, was termed the Bisbee block (McKee et al., 2005).

Discovery of Upper Jurassic marine, continental, and mafic volcanogenic strata in narrow basins in northern Sonora, southeasternmost Arizona, and southwestern New Mexico corroborated the previously posited extensional origin of the Bisbee basin (Lawton and Olmstead, 1995; Lawton and Harrigan, 1998; Lucas et al., 2001; Villaseñor et al., 2005; Mauel et al., 2011). Marine and volcanic strata in the Chiricahua Mountains of southeastern Arizona that overlie a thin (∼40 m) basal conglomerate interpreted as Glance Conglomerate were assigned to the Bisbee Group (Drewes et al., 1995), a stratigraphic convention that was retained when the section was recognized as Jurassic (Fig. 2) (Lawton and Olmstead, 1995). Bisbee Group nomenclature was later extended into southwestern New Mexico (Lawton and Harrigan, 1998; Lucas and Estep, 1998b), where the group presently includes the Upper Jurassic Broken Jug Formation and the Lower Cretaceous Hell-to-Finish, U-Bar, and Mojado Formations (Zeller, 1965, 1970; Lawton, 2004; Lucas and Lawton, 2000; Lucas et al., 2001). A well-exposed Jurassic–Lower Cretaceous section overlies Permian carbonate rocks in the Little Hatchet Mountains, northward of which Paleozoic and Mesozoic strata thin toward the Burro Mountains, where the middle Cretaceous Beartooth Member of the Mojado Formation directly overlies Mesoproterozoic granitoids and metamorphic rocks (Paige, 1916; Darton, 1917; Lasky, 1936; Chafetz, 1982). Thinning of the Phanerozoic section takes place by erosional truncation of upper Paleozoic strata beneath the Lower Cretaceous section and by northward stepwise thinning of Lower Cretaceous strata at syndepositional normal faults (Bayona and Lawton, 2003; Lawton, 2004; Machin, 2013). The Early Cretaceous paleogeographic high in the Burro Mountains is termed the Burro uplift (Elston, 1958), a southeastward extension of the Mogollon highlands (Fig. 1; Bilodeau, 1986). A resulting southeast-trending paleogeographic element, the Burro-Mogollon uplift, formed the Late Jurassic–Early Cretaceous rift shoulder of the Bisbee basin (Bilodeau, 1982; Dickinson et al., 1986; Mack, 1987a, 1987b). Bisbee Group strata are present on the international border between Chihuahua and New Mexico near El Paso, Texas (Fig. 1), where they consist of mudstone and subordinate sandstone intercalated with carbonate strata, which dominate the middle Cretaceous section of the Chihuahua trough (Monreal and Longoria, 2000; Haenggi, 2002; Lucas et al., 2010).

Lower Cretaceous strata of northern Sonora are likewise included in the Bisbee Group. The stratigraphic succession there resembles that of Arizona, and the formation names Glance, Morita, Mural and Cintura, together with local variants, are employed in Sonora (González-León, 1994; Jacques-Ayala, 1989, 1995). The Bisbee basin in Sonora lies south of a horst block termed the Cananea high, which lacks a Jurassic–Lower Cretaceous section (Fig. 1; McKee and Anderson, 1998; McKee et al., 2005; Page et al., 2010). South of the Cananea high, an Upper Jurassic deepwater section with interbedded mafic volcanic rocks termed the Cucurpe Formation underlies the Morita Formation locally and is equivalent in part to the Glance Formation (Figs. 1 and 2; Villaseñor et al., 2005; Mauel et al., 2011; Peryam et al., 2012). Upper Jurassic strata of Sonora are not included in the Bisbee Group (Mauel et al., 2011).

We divide Upper Jurassic–Lower Cretaceous strata of the broadly defined Bisbee basin into five regionally correlative tectonostratigraphic assemblages (TSAs; Fig. 2). Tectonostratigraphic assemblages “record phases of basin history during which the fundamental controls of tectonic setting, subsidence style, and basin geometry are relatively similar” (May et al., 2013, p. 1403). The TSAs of this study consist of one or more formations whose names in some cases vary among sub-basins within the broader basin: (TSA 1) Upper Jurassic volcanic-sedimentary strata that include deep marine, deltaic, fluvial, and alluvial deposits; (TSA 2) Valanginian–lower Aptian continental redbeds and local shallow marine deposits; (TSA 3) upper Aptian–middle Albian carbonate and siliciclastic strata; (TSA 4) upper Albian–lower Cenomanian siliciclastic continental and shallow-marine strata; and (TSA 5) middle–upper Cenomanian shale with interbedded ash-fall tuff, bentonite, and subordinate sandstone. The lithic types within these assemblages vary in detail, but each assemblage can be recognized throughout the region.

TSA 1 consists of marine shale, locally tuffaceous turbidites, subaqueous and subaerial mafic volcanic rocks and hyaloclastite, and feldspathic to volcanic-lithic sandstone (Fig. 2; Lawton and Olmstead, 1995; Mauel et al., 2011). Fossils of Tethyan affinity indicate an Oxfordian–Tithonian age range and a connection with the Gulf of Mexico (Olmstead et al., 1996; Olmstead and Young, 2000; Lucas et al., 2001; Villaseñor et al., 2005). Interbedded mafic volcanic rocks have ocean-island basalt chemistry, indicating asthenospheric derivation (Lawton and McMillan, 1999). In southern Arizona, the assemblage is instead represented by the locally thick alluvial Glance Conglomerate, which thins across syndepositional normal faults near the town of Bisbee (Bilodeau, 1982; Bilodeau and Lindberg, 1983; Bilodeau et al., 1987). Marine strata in Sonora grade to apparently alluvial conglomerate, which is also termed Glance, on the south flank of the Cananea high in northern Sonora (Fig. 1; Peryam et al., 2012), indicating local high relief and syndepositional fault movement (e.g., McKee and Anderson, 1998; McKee et al., 2005). TSA 1 defines initial development of three separate, southeast-trending rift basins, each with depositional connections via a system of extensional basins to the nascent Gulf of Mexico (e.g., Bilodeau, 1982; Lawton and McMillan, 1999). We use the terms Altar-Cucurpe basin for the sub-basin south of the Cananea high (Mauel et al., 2011), Huachuca basin for the sub-basin in southern Arizona, which contains the type Bisbee Group, and Bootheel basin for the sub-basin in southeastern Arizona and southwestern New Mexico. The Huachuca basin as defined here encompasses smaller basins previously termed the Hereford basin of Late Jurassic age and the upper Bisbee basin, in which strata of TSA 2–4, described below, accumulated (McKee et al., 2005). The general extent of the Altar-Cucurpe basin has also been termed the Sonora basin, which is inferred to have accumulated sediment from Late Jurassic through Eocene time (Rodríguez-Castañeda, 2002), an interval that includes post-Bisbee basin tectonics; therefore, we prefer the more temporally restricted term of Mauel et al. (2011). Because basin margins are not well defined everywhere, and normal faults created numerous local half-graben structures within the basins (e.g., Soreghan, 1998), sub-basin extents are depicted schematically in Figure 1.

TSA 2 unconformably overlies Permian to Jurassic strata in all depozones and consists of dominantly fluvial continental strata encompassing channel sandstone complexes that are locally conglomeratic and interbedded with thick red siltstone intervals containing calcareous paleosols (Dickinson et al., 1986; Mack et al., 1986; Mauel et al., 2011; Peryam et al., 2012). Interfingering alluvial fan and thick lacustrine deposits are present in the complexly faulted northwestern part of the Huachuca basin, which contained several half-graben structures during Aptian time (Soreghan, 1998). Deformation of subjacent strata of TSA 1 varies significantly with respect to that of TSA 2. In southwestern New Mexico and southeastern Arizona, local discordance with underlying Paleozoic or Jurassic strata is slight, but regional relations indicate abrupt changes of subjacent strata across high-angle faults of Jurassic and Early Cretaceous age (Lawton, 2000, 2004; Bayona and Lawton, 2003). Omission of older TSA 2 strata above the unconformity at neighboring localities in Sonora and the Huachuca basin indicates onlap of Lower Cretaceous strata onto beveled, tilted Paleozoic and Jurassic sections (Soreghan, 1998; Mauel et al., 2011). Elsewhere, Jurassic strata were tightly folded prior to deposition of TSA 2 (Peryam et al., 2012).

TSA 3 consists of interbedded carbonate and siliciclastic strata deposited in a carbonate ramp setting that included large rudistid reefs, skeletal carbonate platforms, delta front environments, and mixed siliciclastic-carbonate shorelines (Zeller, 1965, 1970; Archibald, 1987; Ferguson, 1987; Warzeski; 1987; Jacques-Ayala, 1989; Lawton et al., 2004; González-León et al., 2008). Abundant fossils provide the stratigraphically densest age control of the section (Lucas and Estep, 1998b; Lawton et al., 2004; González-León et al., 2008). TSA 3 strata have been interpreted as a record of post-rift thermal subsidence (Mack, 1987a, 1987b; González-León, 1994).

TSA 4 is a thick succession of fluvial and shallow-marine siliciclastic strata represented by the Mojado Formation of southwestern New Mexico and the Cintura Formation of Arizona and Sonora. These units overlie TSA 3 with a sharp contact. The Mojado Formation has been interpreted as recording inception of a flexural foreland basin in New Mexico (Mack, 1987a, 1987b). In Sonora, by contrast, the Cintura Formation is interpreted to record continued post-extensional, thermo-tectonic subsidence (González-León, 1994). TSA 4 also includes marine shale, limestone, and conglomerate of the informal La Juana formation, which overlies the Cintura Formation in northern Sonora (Mauel et al., 2011).

TSA 5 includes marine strata that overlie upper Albian (or, locally, earliest Cenomanian) strata on a sharp contact in southwestern New Mexico (Cobban et al., 1989; Lucas et al., 1988, 2000; Lucas and Lawton, 2005; Machin, 2013), but they are largely absent from southern Arizona and likely eroded before and during Laramide deformation. This assemblage includes the Mancos Formation, locally exposed in southwestern New Mexico, and the Dakota Formation in northwestern New Mexico, with a base that correlates with the uppermost part of TSA 4 of southwestern New Mexico. TSA 5 permits correlation of late-stage Bisbee basin history with early deposition in the Western Interior seaway and Cordilleran foreland basin.

Important to detrital zircon provenance interpretations, local basement rocks of northern Sonora, southeastern Arizona, and southwestern New Mexico consist of Paleoproterozoic continental crust of southwestern Laurentia assigned to the Yavapai and Mazatzal basement provinces (Fig. 1; ca. 1.79–1.64 Ga; Karlstrom et al., 2004; Anderson and Silver, 2005; Amato et al., 2008). Proterozoic granitoids with ages of ca. 1.46 Ga in the Burro Mountains of southwestern New Mexico intrude older (ca. 1.65 Ga) metamorphic rocks (Amato et al., 2009; Amato et al., 2011). In the Little Hatchet Mountains of southwestern New Mexico, basement rocks include granitoids with ages of ca. 1.1 Ga (Clinkscales, 2011; Amato and Mack, 2012; Clinkscales and Lawton, 2018).

The Peninsular Ranges and Sierra Nevada batholiths constitute the roots of Jurassic–Middle Cretaceous magmatic arcs of southwestern North America that were active during deposition in the Bisbee basin; thus, they represent possible sources of sediment in the basin. The Peninsular Ranges batholith (PRB) is zoned from east to west in composition and age (Gastil, 1975; Silver and Chappell, 1988; Todd et al., 1988; Kimbrough et al., 2001). The western zone includes the roots of an older arc system that is composed of the Santiago Peak and Alisitos arcs and exposed in the northern Baja California Peninsula and southern California. The Santiago Peak arc, represented by an Upper Jurassic–Lower Cretaceous volcano-sedimentary succession that has yielded Late Jurassic marine fossils and U-Pb zircon ages ranging from 138 Ma to 120 Ma, extended from the Transverse Ranges of southern California to the Agua Blanca fault directly south of Ensenada (∼32°N; Fig. 1; Wetmore et al., 2002, 2003). Inherited Precambrian cores in some zircon grains indicate a depositional connection with basement of southwestern North America (Wetmore et al., 2002).

Plutons of the Alisitos arc crop out from the Agua Blanca fault to the central part of the Baja California Peninsula (∼28°N); although covered by younger volcanic rocks farther south, the plutons create a strong magnetic anomaly that extends to the tip of the Baja California Peninsula but does not clearly continue onto the Mexican mainland (Fig. 1; Langenheim and Jachens, 2003). Associated strata contain Aptian–Albian fossils and a narrow range of zircon ages (116–115 Ma; Wetmore et al., 2002). Zircon grains lack inheritance (Wetmore et al., 2002), and detrital zircons in basinal sandstone older than 110 Ma lack evidence of continental derivation, suggesting isolation from North America prior to that time (Alsleben et al., 2012).

The Alisitos arc was likely developed on oceanic lithosphere and is alternatively interpreted as an exotic oceanic arc (Dickinson and Lawton, 2001a; Wetmore et al., 2003) or fringing arc separated from North America by an ocean basin (Busby et al., 2006; Marsaglia et al., 2016; Boschman et al., 2018). Farther south in western Mexico, the Guerrero terrane contains Upper Jurassic–Lower Cretaceous volcanic and intrusive rocks (Talavera-Mendoza et al., 2007; Mortensen et al., 2008; Centeno-García et al., 2008, 2011; Martini et al., 2009, 2011). Quartzarenite collected from deformed deep-marine rocks near the city of Zacatecas in north-central Mexico has a maximum depositional age of 109 ± 3 Ma (early Albian) and contains a dominant detrital zircon age group ranging over ca. 145–108 Ma, with age peaks at ca. 158 Ma, 137 Ma, 130 Ma, and 117 Ma, that is interpreted as derived from Guerrero (Ortega-Flores et al., 2016). The detrital ages indicate that igneous rocks with ages equivalent to the Early Cretaceous Santiago Peak and Alisitos arcs are likely present in mainland Mexico as part of the Guerrero composite terrane, consistent with interpretations of previous workers that include the Santiago and Alisitos arcs in Guerrero (e.g., Dickinson and Lawton, 2001a). The detrital record also demonstrates that Guerrero contains a suite of Lower Cretaceous magmatic rocks older than 117 Ma, in the range of ca. 145–108 Ma. Triassic metasedimentary rocks that make up the oldest known rocks of Guerrero contain detrital zircons that indicate a connection with continental Mexico prior to development of an oceanic basin, the Arperos basin, between Mexico and Guerrero, likely in latest Jurassic time (Centeno-García et al., 2008).

The eastern zone of the PRB is defined by tonalite and low-K granodiorite intrusions termed La Posta suite, named after an intrusion that spans the U.S.–Mexico border (Silver and Chappell, 1988; Walawender et al., 1990). The voluminous La Posta suite constitutes ∼47% of the surface exposures in the Peninsular Ranges batholith and has U-Pb ages ranging from 98 Ma to 92 Ma, which indicates voluminous magma emplacement during that time interval (Kimbrough et al., 2001). La Posta intrusions postdate an episode of shortening that deformed western zone igneous rocks during the time interval 118–105 Ma (Todd et al., 1988; Thomson and Girty, 1994; Johnson et al., 1999).

The California arc, represented by the Sierra Nevada batholith, had a somewhat different magmatic history than the arcs of the Baja California Peninsula (Ducea, 2001). Magmatism took place in two short-lived episodes, a Late Jurassic episode from 160 Ma to 150 Ma and a more voluminous Late Cretaceous episode beginning at ca. 121 Ma and that became particularly marked from 100 Ma to 85 Ma (Ducea, 2001; DeCelles et al., 2009; Nadin et al., 2016), with an intervening episode of reduced magmatism recorded mainly in the western part of the Sierra Nevada batholith (Nadin et al., 2016) that occurred during continuing magmatic activity in the Alisitos arc, which is described above. The Late Cretaceous part of the second episode is thus, in part, correlative with the La Posta event (Ducea, 2001; Kimbrough et al., 2001). Ash-fall tuff beds generated during La Posta eruptions are widely distributed in Cenomanian and Turonian marine deposits of the Western Interior seaway (Christiansen et al., 1994).

Zircons were separated from ∼5–10 kg of sample using standard crushing and mineral separation techniques, including magnetic separation and dense liquid settling in sodium polytungstate and methylene iodide. The high-density fraction was washed in acetic acid if carbonate minerals were present and in nitric acid if sulfide minerals were present. Zircons were hand-picked (for igneous samples) or randomly poured (for detrital samples) and mounted in epoxy for analysis.

Laser ablation–multi-collector–inductively coupled plasma mass spectrometry (LA–MC–ICPMS) and single-collector LA–ICPMS dating were conducted on detrital and igneous zircon grains at the Arizona Laserchron Laboratory at the University of Arizona. Beam diameter was generally 20–30 µm. Errors on spot ages of individual zircon grains are reported in the text and tables at 1σ, and we report weighted mean ages in the text and figures at the 2σ level. We report and plot ²⁰⁶Pb/²³⁸U ages for grains <1 Ga and ²⁰⁶Pb/²⁰⁷Pb ages for grains >1 Ga. Data are presented on concordia diagrams using Isoplot (Ludwig, 2012).

New and published stratigraphic ages were used to construct geohistory diagrams for southwestern New Mexico and north-central Sonora. The diagrams incorporate the decompaction algorithim of Angevine et al. (1990) and show subsidence history for complete stratigraphic sections in the two regions. Age inputs for the subsidence curves include biostratigraphic data, U-Pb ages on tuff beds, and maximum depositional ages calculated from the youngest zircon grain ages that overlap at 2σ error, using the weighted mean algorithm of Isoplot (Ludwig, 2012). Errors for these different data types are on the order of ±1 m.y. (∼1% 2σ); errors for stratigraphic thicknesses are approximately ± 5%. The analytical method is sensitive to the thickness of overburden inferred to have overlain the analyzed section, which is estimated from regional stratigraphic relations, but this factor consistently affects the degree of corrected compaction in a particular section and does not change the general shape of the subsidence curves. We employ the current geologic time scale of the International Commission on Stratigraphy (Cohen et al., 2013; updated).

We focus here on Lower and middle Cretaceous siliciclastic strata of TSAs 2, 4, and 5, although fossiliferous strata of TSA 3 are important in regional correlation. TSA 1 has been described previously (Lawton and Olmstead, 1995; Lawton and Harrigan, 1998; Busby et al., 2005; Bassett and Busby, 2005; Mauel et al., 2011).

In northern Sonora, siliciclastic marine strata of the Rancho La Colgada Formation unconformably overlie Upper Jurassic deepwater strata of the Cucurpe Formation and grade upsection to fluvial strata of the Morita Formation, which ranges from 700 m to 1200 m thick (González-León, 1994; Peryam et al., 2012). The Rancho La Colgada Formation is absent at some localities where fluvial Morita beds directly overlie the Jurassic section (Fig. 3; Mauel et al., 2011; Peryam et al., 2012). The contact between Jurassic and Cretaceous strata in northern Sonora has alternatively been considered transitional (Rodríguez-Castañeda, 1990, 1991).

In southern Arizona, fluvial channel complexes and red to purple siltstone, commonly with calcareous nodules, similarly dominate the Morita Formation (Fig. 3). At Bear Canyon in the southeastern Huachuca Mountains, red siltstone and sandstone—interpreted here as the basal part of the Morita Formation—overlie thick-bedded, clast-supported Glance Conglomerate on a sharp contact. Siltstone and sandstone beds contain irregular micrite nodules and interbeds of micrite-granule conglomerate 50–125 cm thick. This stratigraphic interval, ∼30 m thick, represents a succession of paleosols and conglomerate containing micrite nodules reworked from the paleosol horizons (McKee et al., 2005). Although measured bed attitudes do not unambiguously demonstrate discordance between Glance and Morita strata in Bear Canyon, the paleosol and granule conglomerate interval thickens to greater than 50 m in outcrops 2100 m to the northwest, which suggests angular discordance at the Glance-Morita contact. Previous workers have recognized the condensed pedogenic nature of the Glance-Morita transition interval; some have assigned it to the Glance (McKee et al., 2005) and others to the Morita (Vedder, 1984; Klute, 1991). Strata overlying the condensed interval contain laterally continuous beds of fine- to medium-grained, well-sorted quartzose sandstone 5–10 m thick with abundant trough cross-beds, planar lamination, and abundantly burrowed bed tops (Fig. 3). We interpret this part of the section as shoreface deposits approximately correlative with the Rancho La Colgada Formation. Continental fluvial strata continue to a stratigraphic level 300 m above the base of the Morita, where the sandstone composition abruptly changes to a volcanic compositional petrofacies (Klute, 1987). We collected samples for detrital zircon (DZ) analysis from the shoreface deposits (10LM11) and a channel sandstone bed of the volcanic petrofacies (11BC1).

The Hell-to-Finish Formation of southwestern New Mexico overlies Jurassic basaltic rocks of the Broken Jug Formation on a sharp but concordant contact in the central part of the Little Hatchet Mountains (Lawton and Harrigan, 1998), whereas elsewhere it overlies Paleozoic limestone (Gillerman, 1958; Zeller, 1965). The formation consists of red siltstone and thin, upward-fining beds of conglomerate and sandstone with scoured bases. A stratigraphically complete exposure of the unit in the central part of the Little Hatchet Mountains is 525 m thick (Lucas and Lawton, 2000). In the northern Little Hatchet Mountains, a section with faulted top and base is 580 m thick and consists of limestone- and chert-pebble conglomerate, feldspathic sandstone, siltstone, and shale that is interpreted as alluvial-fan and fluvial deposits (Mack et al., 1986). Interbedded conglomerate and fossiliferous sandy limestone in correlative strata of the Peloncillo Mountains on the Arizona–New Mexico border are interpreted as fan-delta deposits (Bayona and Lawton, 2003). The Hell-to-Finish Formation has been interpreted as continental deposits with transverse alluvial sediment dispersal from local bounding uplifts and axial, southeast-oriented fluvial systems (Lawton, 2004). A single sample of Hell-to-Finish sandstone (HF1) was collected from the lower part of the unit in the central Little Hatchet Mountains.

In northern Sonora and southeastern Arizona, the uppermost formation of the Bisbee Group is termed the Cintura Formation. In Sonora, it gradationally overlies the Mural Formation and consists of as much as 2000 m of interbedded quartzose sandstone and red-weathering siltstone, commonly with calcareous pedogenic nodules (Mauel et al., 2011). The entire formation represents deposits of fluvial systems with paleocurrent directions to the northeast and southeast (González-León, 1994). In Sonora, the informal La Juana formation of TSA 4, described below, gradationally overlies the Cintura Formation and correlates with marine intervals in the upper part of the Cintura and Mojado formations of the Bootheel basin.

In Arizona, the Cintura Formation contains strata composed of upward-fining sandstone that grades to red and gray siltstone (Dickinson et al., 1986). This formation is dominated by fluvial strata in the western part of the Bisbee basin, in the vicinity of Tucson, and in the Huachuca basin (Klute, 1991). It is considered to grade to mixed marine and fluvial strata in southeastern Arizona (Dickinson and Lawton, 2001b), but the presence of marine beds in the Huachuca basin, although expected, remains unconfirmed. The Cintura Formation in the Chiricahua Mountains of the Bootheel basin consists of laterally extensive sandstone bodies deposited in shoreface settings, with rounded quartzite clasts to 15 cm in diameter locally in the upper part of the section. Although the formation there is not well exposed, fluvial strata appear to be absent. The age of the Cintura Formation in Arizona is controlled by its stratigraphic position above the middle Albian Mural Limestone and below Upper Cretaceous Laramide strata (Dickinson and Lawton, 2001b).

Near Tuape, Sonora, red siltstone of the upper part of the Cintura Formation grades to calcareous shale, fossiliferous limestone, sandstone, and pebble conglomerate of La Juana formation. Pebbles consist of tan, gray, and laminated dolostone; fossiliferous, intraclastic, and oolitic limestone; 10%–30% chert, which is more abundant in the smaller pebble fraction; and white, fine-grained quartz arenite. Fossil molluscs recovered from La Juana formation near Arizpe (Fig. 1) indicate a late Albian age (R.W. Scott, written communication, 2001). As noted above, La Juana represents marine incursion into the Altar-Cucurpe basin near the end of the Albian, a transgression recorded in the Bootheel basin by siliciclastic strata.

The Mojado Formation of southwestern New Mexico is a succession of middle Cretaceous sandstone, siltstone, and shale in the Big Hatchet and Little Hatchet Mountains (Zeller, 1965, 1970) and adjacent ranges, including the Peloncillo Mountains and Cookes Range, where correlative strata were formerly termed the Johnny Bull and Sarten Formations (Gillerman, 1958; Mack et al., 1988; Lucas and Estep, 1998b; Lawton, 2004). The Mojado Formation generally thins northward from its thickest occurrences in southwestern New Mexico and overlies progressively older rocks to the north. In the northern part of the Little Hatchet Mountains, the formation is 1245 m thick (Galemore, 1986); in the Cookes Range, a composite section is ∼107 m thick (Fig. 4; Lucas et al., 1988). In southwestern New Mexico localities, it overlies marine carbonate strata of the upper Albian U-Bar Formation on a sharp, concordant contact variably interpreted as gradational (Mack et al., 1986, 1988) or disconformable (Lucas and Lawton, 2000). In the Cookes Range, the Mojado Formation overlies lower Permian red beds of the Abo Formation (Clemons, 1982); farther east, in the southern San Andres Mountains north of El Paso, Texas (Fig. 1), a thin section (∼6 m) of Mojado overlies lower Permian strata and is inferred to be overlain by ∼50 m of Upper Cretaceous Dakota Formation with large trough cross beds, although the relative proportion of the two units there remains debated (Lucas and Estep, 1998a, and references therein).

In the Little Hatchet Mountains, the Mojado Formation consists of three members (Galemore, 1986; Mack et al., 1986, 1988), formally termed the Fryingpan Spring, Sarten, and Rattlesnake Ridge members (Lucas and Estep, 1998b; Lucas and Lawton, 2000), the upper and lower of which represent shoreface deposits. The Sarten Member consists of 650 m of lenticular, trough cross-bedded sandstone beds and red or green shale deposited by meandering rivers that flowed eastward to a marine shoreline near El Paso, Texas (Galemore, 1986; Mack et al., 1986, 1988; Lucas et al., 2010). As the members pass into adjacent ranges, they either pinch out or change facies, with the result that correlations of fluvial strata with fossil-bearing marine facies are uncertain; therefore, we restrict our usage to the Mojado Formation, except where our new data permit better correlation between different facies tracts (Fig. 4). In the Cookes Range, the base of the Mojado contains 10–15 m of locally derived, limestone-clast conglomerate and cross-bedded sandstone of fluvial origin (Fig. 4). The remainder of the section consists of interbedded heterolithic feldspathoquartzose sandstone, siltstone, and shale and upward coarsening and thickening successions of burrowed sandstone, which we interpret as estuarine and tidally influenced shoreface deposits, respectively. Paleocurrent measurements in the upper part of the unit indicate sediment transport to the southwest, probably by ebb-tidal currents.

Fossil content and bracketing stratigraphic relations indicate that the Mojado Formation in the Cookes Range is late Albian to earliest Cenomanian. Upper Albian ammonites are present in lower part of the marine section (Fig. 4), and lower Cenomanian molluscs characteristic of the Del Rio and Buda Formations of Texas are present in the upper part of the section (Cobban, 1987; Lucas et al., 1988; Lucas and Estep, 1998b). Fossils suggest that the entire Mojado Formation in the Big Hatchet Mountains is late Albian (Lucas and Estep, 1998b). The dominantly marine section of the Cookes Range locality has been interpreted to represent the downdip marine shoreline equivalent of the fluvial system of the Little Hatchet Mountains (Galemore, 1986; Mack et al., 1986); alternatively, the fluvial succession of the Little Hatchet Mountains may predate deposition of the Cookes Range section, with the possible exception of the lower fluvial beds there.

We collected three detrital zircon samples from the upper part of the Mojado Formation, two from the upper member in the Little Hatchet Mountains (11BQ17, 11BQ19), and one in the Cookes Range nine meters below the top of the formation (12BQ37); we also review DZ data for two previously published samples, one from the lower part of the formation in the Little Hatchet Mountains (5.12.15.10; Clinkscales and Lawton, 2015) and one from a shoreface deposit in the upper part of the Cintura Formation of the Chiricahua Mountains (KBCR; Dickinson et al., 2009) that is probably equivalent to the Rattlesnake Ridge Member of the Mojado in New Mexico.

The Beartooth Quartzite, named by Paige (1916), is now termed the Beartooth Member of the Mojado Formation (Lucas and Estep, 1998b). Prior to geochronologic data presented here, Beartooth-Mojado equivalence was inferred based on stratigraphic position (“Beartooth problem” of Lawton, 2004). In the Burro Mountains, at Saddlerock Canyon, ∼20 km west of Silver City, New Mexico (Figs. 1 and 4), it is 30 m thick and directly overlies Proterozoic granodiorite (Fig. 4; Hedlund, 1980; Amato et al., 2011). The basal part of the Beartooth consists of pebble-to-boulder conglomerate 1.5 m thick that fines upward to fine- to medium-grained sandstone. Conglomerate clasts include angular to rounded quartzarenite and chert pebbles, as well as distinctive granodiorite boulders as much as 30 cm in diameter. Most of the overlying sandstone is laminated to apparently structureless with uncommon burrows and shale intervals. A pebble lag at the top of the unit is directly overlain by shale with thin interbeds of siltstone and very fine-grained sandstone of the Mancos Formation, as indicated by our composite section from the Burro Mountains (Fig. 4).

The Beartooth Member was deposited in a nearshore marine environment characterized by mixed tidal and wave energy. Heterolithic sandstone-siltstone indicates intermittent tidal currents (e.g., Nio and Yang, 1991). Hummocky cross-stratification low in the section indicates a shoreface setting. Although the Beartooth Member contains no fossils, correlation with some part of the Mojado Formation was inferred (e.g., Mack, 1987a, 1987b; Lucas and Estep, 1998b; Lawton, 2004) and now is confirmed by our new geochronologic data.

The Mancos Formation, or Mancos Shale, overlies the Mojado Formation in the Little Hatchet Mountains and Cookes Range (Fig. 4; Lucas et al., 1988; Lucas and Lawton, 2005; Cobban et al., 1989) and the Beartooth Member in the Burro Mountains, where the shale was previously termed Colorado Formation (Hedlund, 1980; Finnell, 1987). The basal contact at all localities is a sharp, transgressive unconformity overlain by a thin bed of poorly sorted sandstone with pebbles and shell fragments. An incomplete section of the Mancos is ∼100 m thick in the Little Hatchet Mountains (Lucas and Lawton, 2005), 60 m in the Cookes Range (Lucas et al., 1988), and from 115 m to 160 m thick in complete sections of the Burro Mountains and at exposures on the Gila River in westernmost New Mexico (Finnell, 1987; Lucas and Estep, 1998b).

The texture of the shale, which contains abundant thin beds of volcanic ash, varies among localities. In the Little Hatchet Mountains, the shale is dark gray and subfissile. A 15 cm tuff collected in the lower part of the shale (11BQ18) is a coarse-grained, vitric crystal ash-fall tuff with fresh crystals 0.15–2.0 mm in diameter that consist of 30%–35% biotite, 20% plagioclase, 15% inclusion-free quartz, and 5%–15% sanidine with Carlsbad twins. Vitric volcanic grains consisting of microcrystalline, low-birefringent crystal domains make up 20%–25% of the tuff and are interpreted as tuffaceous lapilli. The lower six meters of the Mancos in the Cookes Range consist of laterally continuous, tan-weathering platy beds, 5–15 cm thick, of dark brown shale with abundant white tuffaceous grains. Bedding surfaces contain abundant pits up to 0.25 mm in diameter where lapilli have weathered out. Ammonites and inoceramid bivalves are common on bedding surfaces (Cobban et al., 1989). The shale contains at least 11 conspicuous waxy orange bentonite beds 0.5–16 cm thick in the lower 6 m of the formation. The uppermost sampled bentonite (12BQ35) consists of altered tuffaceous or vitric material lacking obvious grain boundaries and containing ∼3% quartz and feldspar crystals to ∼0.03 mm in diameter and ∼10% opaque oxide grains to ∼0.04 mm in diameter. Above this bentonite, the shale changes in character to laminated, blocky weathering brown to light gray shale and silty shale. Uncommon thin sandstone beds with current-ripple cross lamination are present in the blocky shale. At Clyde Canyon in the Burro Mountains, the Mancos consists of gray to black fissile shale with thin interbeds of very fine-grained sandstone and siltstone (Fig. 4). An ash-fall tuff beneath a thin bed (∼30 cm) of gray, sandy limestone was sampled (12BQ47) 14 m above the base of the formation. We also collected a sample of Mancos sandstone (12BQ21) for detrital zircon analysis at Clyde Canyon.

Fossils indicate that the Mojado–Mancos contact is diachronous in southwestern New Mexico. Middle Cenomanian ammonites assigned to the Acanthoceras amphibolum zone were recovered during this study 4–5 m above the base of the Mancos at Clyde Canyon (Figs. 4 and 5) and are ∼70–90 m above the base in the Little Hatchet Mountains (Lucas and Lawton, 2005). Upper Cenomanian ammonites occur in the lower part of the Mancos Formation in the Cookes Range (Fig. 4; Cobban et al., 1989). Therefore, the lower part of the Mancos Formation is possibly as old as early Cenomanian in the Little Hatchet Mountains, middle Cenomanian in the Burro Mountains, and late Cenomanian in the Cookes Range. The Mancos Formation was deposited in an offshore setting below fairweather wave base.

The relationship of the Dakota Formation to the Mojado Formation is important in understanding the tectonic link between the Bisbee basin and Cordilleran foreland basin during the middle Cretaceous; therefore, an important Dakota locality is included in this study. In and adjacent to the San Juan basin of northwestern New Mexico, the Dakota Formation interfingers complexly with the Mancos Formation; together, the units represent broadly transgressive deposits of fluvial, estuarine, marginal marine, and offshore environments (Fig. 4; Owen and Owen, 2003; Owen et al., 2005, 2007). Spanning TSAs 4 and 5 (Fig. 2), the formation correlates, in part, with the upper part of the Mojado Formation and overlies progressively younger Paleozoic and Mesozoic strata northward across western New Mexico and eastern Arizona (Dickinson et al., 1989). At San Ysidro, near the southeastern flank of the San Juan basin, the Encinal Canyon Member of the Dakota Formation, which consists of discontinuous fluvial sandstone bodies (Owen et al., 2007), discordantly overlies strata assigned to the Jackpile Sandstone, a fluvial sandstone widespread in the eastern San Juan basin and interpreted alternatively as part of the Upper Jurassic Morrison Formation or the Lower Cretaceous Burro Canyon Formation (Aubrey, 1998; Owen and Owen, 2003; Owen et al., 2007). The basal contact of the Encinal Canyon Member represents a regional unconformity on Lower Cretaceous and Jurassic strata (Aubrey, 1986; Anderson and Lucas, 1996; Owen et al., 2007). The overlying Oak Canyon and Cubero members contain fossils of the lower middle Cenomanian Conlinoceras tarrantense ammonite zone, which is interpreted as 95.73 ± 0.49 Ma on the basis of calibrated biostratigraphy and ⁴⁰Ar/³⁹ tuff ages elsewhere (Cobban, 1977; Cobban et al., 2006), and a regionally persistent ash-fall tuff, termed “ash A” (Owen and Head, 2001), that is present in the Oak Canyon Member. We collected a sample of medium-grained sandstone (12BQ38) from the Encinal Canyon Member 1 m above the unconformity with the Jackpile Sandstone Member in the Hagan basin (Fig. 1) and a sample of ash A (12BQ41) in the middle part of the Oak Canyon Member at San Ysidro, New Mexico.

Sandstone composition of Lower Cretaceous siliciclastic strata varies greatly within each assemblage across the Bisbee basin (Fig. 6). Sandstone of the Rancho La Colgada and Morita Formations includes quartzo-lithic, feldspatho-quartzose, and litho-quartzose arenite (classification of Garzanti, 2016), with lithic sandstone dominated by volcanic lithic grain types (Table 1; Fig. 6A; González-León, 1994; Peryam et al., 2012). Conglomerate channel lags associated with quartzolithic sandstone in Sonora locally contain as much as 75% sedimentary clasts, including quartzite, chert, and limestone pebbles, some containing crinoid fossils (Peryam et al., 2012). Quartzo-lithic sandstones fall within the compositional field of the quartzose petrofacies of Klute (1987) that is defined in southeastern Arizona (Fig. 6A). The quartzose sandstones, with rounded to subrounded grains, contain abundant, inclusion-rich monocrystalline quartz (Qm). Plagioclase makes up 10%–15% of the framework; discontinuous twins in some grains suggest alteration of another grain type, probably potassium feldspar, to albite (e.g., Walker, 1984). Relative abundance of lithic fragments tends to increase upsection. One of our detrital zircon samples (11BC1) plots near the volcanic petrofacies field of Klute (1987). It contains a diverse suite of volcanic lithic fragments, including microlitic (Lvm), felsitic (Lvf), and uncommon lathwork (Lvl) volcanic grains. The Arizona volcanic lithic sample and many Sonora Morita samples plot in the magmatic arc field of Garzanti (2016). In contrast, the Hell-to-Finish Formation is quartzo-lithic, with Qm and chert dominating the lithic fraction, and plots in continental block and recycled orogen provenance fields.

Most sandstone of the Mojado, Beartooth, and Dakota Formations is more quartzose than older sandstone, but some sandstone intervals contain volcanic lithic grains. Quartzose samples are dominated by monocrystalline quartz and contain subordinate polycrystalline quartz and chert (Fig. 6B; Mack, 1987a). Potassium feldspar is more abundant than plagioclase, which is typically absent (Table 1), indicating that TSA 4 was not extensively albitized. An uncommon compositional variant consists of quartzo-lithic sandstone with abundant potassium feldspar and volcanic lithic fragments (Fig. 6B; Mack, 1987a). Some Qm grains are clear and inclusion-free, and some are euhedral with pyramidal terminations, although a few well-rounded spherical grains are present. Potassium feldspar, mostly sanidine, makes up 24% of the framework of the volcanic-lithic sandstone. Volcanic lithic fragments, including Lvf, Lvm, and vitric volcanic grains (Lvv), compose more than half of the framework. The Lvv grains are low-birefringence microcrystalline aggregates, some of which contain devitrified glass shards. We interpret these volcanic lithic fragments as neovolcanic grains generated by contemporary volcanism (e.g., Critelli and Ingersoll, 1995).

The uppermost part of the Mojado Formation in the Cookes Range and the Beartooth Member are quartzose sandstones, but variable proportions of grain-sized domains consist of finely crystalline mixtures of quartz and white mica or hematite and clay. These suggest the former presence of unstable lithic grains and feldspar, indicating that the quartzose composition is diagenetic.

Cintura sandstone of Sonora is consistently more lithic than Mojado sandstone, with varying feldspar content (Fig. 6B; González-León, 1994). Volcanic lithic grains make up 91% of the total grain population, followed by sedimentary lithic grains (7%) and metamorphic lithic grains (2%). Sandstone composition of the overlying La Juana Formation differs markedly from that of the Cintura Formation. Examination of two La Juana thin sections revealed feldspatho-quartzo-lithic composition with abundant chert and detrital carbonate grains consistent with pebble types observed in conglomerate beds. La Juana sandstone also contains 10%–15% feldspar, both plagioclase and alkali feldspar, and uncommon schistose white mica-quartz grains. The combination of pebbles and grain types suggests a primary source in Paleozoic sedimentary rocks and a subordinate basement source. Beds of volcanic lithic sandstone are interbedded apparently randomly with quartzose sandstone beds in the Cintura of southeastern Arizona (Klute, 1987).

Detrital U-Pb ages of zircon grains indicate that age populations changed from early to late Early Cretaceous time in our samples. Young zircon grains from some Lower Cretaceous sandstone samples proved useful in calculating maximum depositional ages (MDAs) that in many cases indicate near-depositional ages. U-Pb zircon ages of tuff beds and fine-grained bentonites in the Mancos Formation help correlate the unit from southwestern to northwestern New Mexico. A summary of MDAs and tuff ages for our new analyses and published data is shown in Tables 2 and 3. Sample locations from this study are in Table 4.

Most Laurentian basement age provinces and Cordilleran arc sources are represented by zircon age groups in the detrital sample set. Accordingly, observed age groups are assigned to eight time intervals defined by previous authors (e.g., Dickinson and Gehrels, 2009a; Gehrels et al., 2011) and augmented by additional citations in the text: (1) Archean grains (ca. 3000–2500 Ma) of the Wyoming craton; (2) Paleoproterozoic grains (ca. 2300–1800 Ma) of the Wopmay and Trans-Hudson orogens; (3) late Paleoproterozoic and Mesoproterozoic grains (ca. 1780–1630 Ma) of the Yavapai and Mazatzal basement provinces; (4) Mesoproterozoic grains (ca. 1480–1350 Ma) of the Laurentian granite-rhyolite province (Anderson, 1989); (5) late Mesoproterozoic and earliest Neoproterozoic grains (ca. 1250–980 Ma) of the Grenville orogen of eastern and southern Laurentia; (6) Neoproterozoic (ca. 730–510 Ma) Pan-African and Suwannee terrane grains; (7) Paleozoic grains (ca. 490–270 Ma) attributed to the Taconic, Acadian, and Alleghanian orogens of the Appalachian region (Thomas, 2011); and (8) Permian–Mesozoic grains (ca. 275–92 Ma) derived from Cordilleran arc sources (Barth and Wooden, 2006; Riggs et al., 2013, 2016).

TSA 2 contains zircon grains with ages that can be attributed to local Proterozoic basement, recycling of Paleozoic-Mesozoic sedimentary rocks, and arc sources (Fig. 7; Table DR1¹). Samples from the lower parts of the sections in Sonora and Arizona and the Hell-to-Finish sample of New Mexico contain prominent age peaks in the Yavapai-Mazatzal, granite-rhyolite, and Grenville time intervals (Fig. 7). The Hell-to-Finish includes only a few percent late Paleozoic and Mesozoic grains, but other units contain ages that span the late Paleozoic to earliest Cretaceous, ranging from ca. 278–133 Ma. Volcanic lithic samples from the upper part of the Morita in Sonora and Arizona are distinguished by uncommon pre-Jurassic grains, with the bulk of the grain ages forming a bimodal distribution of Jurassic and Early Cretaceous ages in the age range of ca. 191–113 Ma. The volcanic lithic sample from Arizona contains 46 grains (of 97 grains analyzed) with overlapping grain ages in the range of ca. 127–120 Ma.

Six of seven La Colgada and Morita samples with abundant volcanic lithic fragments yielded calculated Early Cretaceous MDAs ranging from 139 Ma to 113 Ma (Valanginian–latest Aptian; Table 2). Samples at the top of the Morita section in Sonora, collected directly beneath fossiliferous upper Aptian beds of the Mural Limestone (Fig. 2; Lawton et al., 2004), yielded late Aptian MDAs of 115 ± 1 Ma and 113 ± 3 Ma (Table 2). Previously published age data for the lower part of the section in Sonora (Peryam et al., 2012) yielded MDAs ranging from ca. 250 Ma to ca.134 Ma. Although no biostratigraphic data exist to confirm that the MDAs there represent depositional ages, their consistent Valanginian age (139 ± 3 to 134 ± 2 Ma; Table 2), apparent consistency between sections, and inferred proximity to a magmatic arc that began prior to those ages and continued beyond them imply that they represent approximate depositional ages. We infer that the Rancho La Colgada–Morita succession was deposited between ca. 136 Ma and 115 Ma (Valanginian–late Aptian) and that a hiatus of 10–15 m.y. exists between youngest preserved strata of the Cucurpe Formation (ca. 149 Ma; Mauel et al., 2011) and basal Lower Cretaceous strata. Physical stratigraphic evidence discussed above likewise indicates the presence of an unconformity. Calculated MDAs for the Morita Formation of the Huachuca basin are 198 ± 5 Ma for the quartzose sample (10LM11) at the base of the section and 123 ± 1 Ma for the volcanic lithic sandstone (11BC1) that is 300 m higher in the section (Table 2; Fig. 3). The MDA from the base of the section is pre-depositional based on stratigraphic relationships. Although unsupported by MDAs from multiple samples, the MDA of the upper sample likely approximates its inferred depositional age (early Aptian) based on physical correlation with the La Colgada and Morita sections and its compositional similarity to the volcanic-lithic Morita in Sonora. The MDA of the Hell-to-Finish sample, 1014 ± 30 Ma, indicates that young zircons were not supplied to the Bootheel basin during Aptian time.

Detrital zircon age populations of TSAs 4 and 5, represented by the Mojado, Cintura, Dakota, and Mancos Formations, fall into two categories (Fig. 8; Table DR3a): (1) Samples with diverse and abundant Proterozoic ages, subordinate Paleozoic ages, and intermediate numbers of Mesozoic ages; and (2) samples with fewer old grains and a high proportion of Mesozoic ages. Category 1 samples average 69% Archean and Proterozoic grains, with appreciable proportions of Yavapai-Mazatzal (10%), granite-rhyolite (10%), and Grenville (30%) ages, 11% Paleozoic grains, and 20% Mesozoic grains, whereas category 2 samples average 39% Proterozoic grains, 7% Paleozoic grains, and 53% Mesozoic grains. Archean grains are rare to uncommon in samples of both categories. Category 1 samples include the lower Mojado in the Little Hatchet Mountains (5.12.15.10), the Beartooth Member in the Burro Mountains (11BQ01), the upper part of the Mojado in the Cookes Range (12BQ37), and the upper Cintura Formation in the Chiricahua Mountains (KBCR). Category 2 samples include the upper Mojado samples in the Little Hatchet Mountains (11BQ17, 11BQ19), the Dakota Formation in northwestern New Mexico (12BQ38), and the Mancos Formation sandstone in the Burro Mountains (12BQ21).

Samples of the upper part of the Mojado Formation yielded young grains that provide MDAs of late Albian–early Cenomanian age (Table 2). The calculated MDA in the Cookes Range (12BQ37) is 98 ± 1 Ma. MDAs of the samples in the Little Hatchet Mountains, both of which contain abundant neovolcanic grains (Table 1), are 102 ± 1 Ma and 99 ± 1 Ma. The Beartooth Member (sample 11BQ01) yielded an MDA of 101 ± 2 Ma, and the Mancos sandstone from high in the formation exposed at Clyde Canyon yielded an MDA of 94 ± 2 Ma. The Encinal Canyon Member of the Dakota Formation in northwestern New Mexico yielded an MDA of 99 ± 1 Ma. The MDAs are consistent with published biostratigraphic ages from the same sections and indicate correlation of the upper part of the Mojado with the Dakota Formation.

Six tuffs in the lower part of the Mancos Formation and one from the Dakota Formation were analyzed with LA–ICPMS. U-Pb zircon ages range from 97.6 Ma to 94.3 Ma (Fig. 4; Tables 3, DR3b). The oldest tuff, ash A, in the Oak Canyon Member of the Dakota Formation (12BQ41) yielded an age of 97.6 ± 1.3 Ma (Fig. 9). The coarse-grained tuff near the base of the Mancos in the Little Hatchet Mountains (11BQ18) yielded an age of 97.2 ± 1.6 Ma. The bentonite bed in the Burro Mountains yielded an age of 96.3 ± 1.3 Ma.

The lower part of the Mancos Formation, which directly overlies the Mojado Formation in the Cookes Range, has at least 11 orange bentonite layers, which are interpreted as ash-fall tuffs, over a stratigraphic interval of 6 m and ranging in thickness from 1 cm to 15 cm. We dated three of these layers (Fig. 10). A sample of the lowermost layer (19BQ45), which directly overlies a 10-cm-thick transgressive lag at the base of the Mancos, yielded an age of 94.9 ± 0.9 Ma (mean square of weighted deviates [MSWD] = 1.7, n = 35). Only five grains were outside of this population, two around 100 Ma, one at 312 Ma, one at 425 Ma, and one Proterozoic grain at 1138 Ma. The next highest sample (19BQ44) yielded an age of 94.3 ± 1.0 Ma (MSWD = 3.0, n = 39). The high MSWD indicates that this group of ages likely includes multiple populations, but we were unable to determine whether the older or younger zircons are more representative of the eruption age, and thus we report the average for all of the grains younger than one clearly older grain at 290 Ma. The highest sampled tuff (12BQ35) returned an age of 94.3 ± 1.1 Ma (MSWD = 0.30, n = 18).

Geohistory diagrams constructed using stratigraphic data from southwestern New Mexico and north-central Sonora reveal broadly similar subsidence history through Late Jurassic–Late Cretaceous time (Fig. 11). Each geohistory plot (Fig. 11) has two curves, one for total subsidence of the decompacted section with water depth included that is termed the total subsidence curve, and another for subsidence with the isostatic effects of sediment and water load removed. The latter curve represents the tectonic subsidence caused by extrinsic mechanisms, such as lithospheric thinning or flexural loading (Heller et al., 1986; Angevine et al., 1990; Xie and Heller, 2009).

We analyzed sections in the two areas that maximize stratigraphic completeness and offer the most extensive biostratigraphic and geochronologic data for calibrating stratigraphic age. The New Mexico analysis derives from complete sections in the central and northern Little Hatchet Mountains (Table 5), spanning from Oxfordian through middle Cenomanian time, a time interval of ∼66 m.y., as compared to a 26 m.y. interval analyzed by Mack (1987a) (Aptian-Albian; Cohen et al., 2013; updated). The Sonora analysis represents a composite of sections near Tuape, Sonora, which lies 15 km south of Cucurpe (Fig. 1). Ages of stratigraphic intervals of both curves employ a combination of biostratigraphic data, MDAs from detrital zircon data, and U-Pb ages of ash beds from published sources and this study. In the Little Hatchet Mountains, the ages of Jurassic strata are based principally on correlation with a Kimmeridgian–Tithonian Jurassic section in the Chiricahua Mountains of southeastern Arizona (Lawton and Olmstead, 1995; Olmstead and Young, 2000), and local corals, which resemble Oxfordian forms from the subsurface of Arkansas (Lucas et al., 2001); therefore, the inception of the steep curve in New Mexico, where Jurassic strata directly overlie Permian strata, is poorly controlled. Deposition may have begun later in the Late Jurassic, as suggested by a late Oxfordian–Kimmeridgian age (ca. 159–156 Ma) for the base of the Jurassic section in the Chihuahua trough (e.g., Haenggi, 2002). This would shorten the time span represented by the New Mexico curve stated above by ∼4 m.y.

The Late Jurassic–earliest Cretaceous subsidence history of both localities shows initially rapid but decreasing rates of tectonic subsidence during Late Jurassic through middle Aptian (Fig. 11A) or middle Albian (Fig. 11B) time, followed by increased subsidence rates in late Albian time, recorded by TSA 4. As discussed earlier, an unconformity with a hiatus of as much as 15 m.y. separates Jurassic and Lower Cretaceous strata in Sonora, and a similar lapse in sedimentation can be inferred from reduced subsidence in the New Mexico curve. Given uncertainties that might result from lithostratigraphic correlation and errors in unit thicknesses, the two subsidence histories are remarkably similar. The concave-upward, declining Jurassic–middle Albian curves generally resemble subsidence histories that result from synrift stretching to post-rift thermotectonic subsidence of extensional basins or passive margins, where the amount of rapid initial tectonic subsidence and sediment accumulation is determined by the amount of crustal stretching (Steckler and Watts, 1978; Xie and Heller, 2009). In contrast, the rapid late Albian–Cenomanian subsidence resembles flexural subsidence curves from foreland basins (Xie and Heller, 2009). The inflection point between subsidence rates of TSA 3 and TSA 4 in Sonora is particularly well defined (Fig. 11B), whereas TSA 3 marks an intermediate increase in subsidence rate between TSA 2 and TSA 4 in New Mexico (Fig. 11A). Possible subsidence mechanisms are elaborated upon further in the Discussion.

U-Pb ages derived from detrital sandstone samples and tuffs across the U.S.–Mexico border region provide improved correlation of Upper Jurassic–middle Cretaceous stratal successions. Similar subsidence histories derived from correlative strata with contrasting provenance relationships in turn yield insight into a temporal transition from separate rift basins to a partitioned foreland integrated with the Cordilleran foreland-basin system. We discuss correlations among the three discrete basins of the greater Bisbee basin, interpret sandstone provenance and sediment dispersal systems during basin evolution, and propose a link between basin history and tectonics of the continental margin.

New and published U-Pb geochronological data augmented by new fossil identifications corroborate published biostratigraphic data and provide new ages for unfossiliferous continental and high-energy marine deposits (Figs. 2–4 and Fig. 11). Our geochronological ages are consistent with their stratigraphic positions in individual sections (Figs. 3 and 4), and despite the inherent problems of comparing biostratigraphic ages calibrated with ⁴⁰Ar/³⁹Ar ages (e.g., Scott, 2014), our age errors overlap with interpreted fossil zone boundary ages, as discussed below.

Ages of Upper Jurassic marine strata, generally of deepwater origin, in the Altar-Cucurpe and Bootheel basins, are well known from U-Pb and biostratigraphic data (Olmstead and Young, 2000; Lucas et al., 2001; Villaseñor et al., 2005; Mauel et al., 2011). The age range of the alluvial Glance Conglomerate of the Huachuca basin is less well controlled, as it is bracketed between 172 ± 2 Ma, the age of an ignimbrite at the base of the conglomerate in the Huachuca Mountains (Gilbert, 2012), and ca. 136 ± 3 Ma, the inferred age of the basal Morita Formation in Bear Canyon. The bracketing data permit a Late Jurassic age and are consistent with a post-170 Ma Jurassic age inferred for the Glance Conglomerate from exposures in the Santa Rita Mountains, 45 km northwest of the Huachuca Mountains, on the basis of regional stratigraphic relations (Bassett and Busby, 2005).

Although uncertainty remains regarding true depositional ages of the lower part of the Cretaceous section, discussed above, MDAs suggest that deposition began ca. 136 Ma in the Altar-Cucurpe basin, where initial Early Cretaceous deposition of the Rancho La Colgada Formation followed a hiatus of ∼10–15 m.y. and took place in shallower marine conditions. Local absence of the Rancho La Colgada at the base of the section suggests onlap of Cretaceous strata onto Jurassic rift-basin sections tilted during the hiatus. Initial Cretaceous marine incursion into the formerly continental Huachuca basin, recorded by shoreface deposits in the Morita Formation, took place at about the same time based on stratigraphic position above the Glance Conglomerate and correlation with the Sonoran stratigraphic sections. Early Cretaceous sediment accumulation similarly followed a prolonged depositional hiatus or slow sedimentation indicated by an interval of paleosols and granule conglomerate. In New Mexico, where a concordant contact with the subjacent Broken Jug Formation demonstrates a post-Jurassic age for the Hell-to-Finish Formation but does not reveal presence or absence of a hiatus, the age of earliest Cretaceous deposition in the Bootheel basin remains poorly known (Lucas and Lawton, 2000); nevertheless, superjacent Aptian carbonate strata indicate a general correlation with the TSA 2 units of the other basins (Fig. 2).

Cintura and Mojado strata of TSA 4 were deposited during late Albian–earliest Cenomanian time. Fossiliferous middle Albian beds of subjacent Mural and U-Bar strata in the region support the older age of TSA 4 at localities where fossils are not yet known from the lower part of TSA 4 (Lucas and Estep, 1998b; Lucas and Lawton, 2000; Lawton et al., 2004; González-León et al., 2008). Ammonites in the lower part of the Mojado Formation of the Cookes Range (Fig. 4) and bivalves in the La Juana Formation of Sonora demonstrate that middle Cretaceous marine incursion had taken place throughout the region by late Albian time. Lower parts of thick Mojado sections in the Bootheel basin did not yield near-depositional zircon ages, but upper Mojado samples from the Little Hatchet Mountains, Cookes Range, and and Burro Mountains have MDAs ranging from 102 ± 1 Ma to 98 ± 1 Ma, which indicate correlation with the basal part (Encinal Canyon Member) of the Dakota Formation at San Ysidro (MDA = 99 ± 1 Ma; Table 2; Figs. 4 and 11).

U-Pb tuff ages and biostratigraphic data from TSAs 4 and 5 indicate that the upper part of the Mojado, the Mancos of southwestern New Mexico, and lower part of the Dakota Formation in NW New Mexico constitute a diachronous but genetically related and widespread complex of upper Albian–Cenomanian estuarine and shallow-marine deposits (Tables 2 and 3; Figs. 4 and 11). U-Pb ages of 97.6 ± 1.3 Ma from ash A at San Ysidro, 97.2 ± 1.6 Ma from a tuff in the Mancos in the Little Hatchet Mountains, and 96.3 ± 1.3 Ma from a tuff in the Burro Mountains indicate equivalence of marine shelf deposits there with estuarine and shoreface deposits in the Cookes Range (Figs. 4 and 11).

The age of ash A at San Ysidro is consistent with its position 20 m below the mid-Cenomanian Conlinoceras tarrantense ammonite zone (Cobban, 1977), which has a 96.2 Ma lower age boundary (Scott, 2014). U-Pb and biostratigraphic ages of uppermost Albian–lowermost Cenomanian marine deposits of the region thus demonstrate not only previously posited northward stratal onlap onto the Burro uplift (e.g., Mack, 1987a), but also widespread marine deposition from southwesternmost to northwestern New Mexico (Fig. 12). The short time interval (94.9 ± 0.9 Ma to 94.3 ± 1.0 Ma) suggested by U-Pb ages of tuff beds in the lower Mancos deposition in the Cookes Range agrees with late Cenomanian ammonites interbedded with the bentonites (Fig. 4; Cobban et al., 1989). Rapid deposition during the transgressive part of the Mancos Formation may have resulted from voluminous volcaniclastic input to the region during the 98–92 Ma La Posta magmatic event of southern California and the Baja California Peninsula (Kimbrough et al., 2001; Ortega-Rivera, 2003), but we cannot presently discriminate between potential contemporary volcanic sources represented by the Peninsular and the Sierra Nevada batholiths (e.g., Ducea, 2001; DeCelles et al., 2009). Our Mancos tuff ages all fall within the range of ⁴⁰Ar/³⁹Ar tuff ages (97.5–91 Ma) reported from the Buda and Eagle Ford Formations in west Texas and lie in the age range of Oceanic Anoxic Event 2 (Eldrett et al., 2014).

Changing provenance relationships in Jurassic–middle Cretaceous strata accompanied basin evolution and provide evidence for local sediment sources that served as topographic barriers between the basins. The rapidly subsiding Late Jurassic Altar-Cucurpe basin received sediment rich in pyroclastic material from active volcanic centers in a westward-migrating magmatic arc (Fig. 13A; Mauel et al., 2011), but volcanic detritus did not reach the Huachuca basin from ca. 172 Ma (Gilbert, 2012) until it appeared in the Morita Formation at ca. 123 Ma; the Jurassic Bootheel basin never received arc-derived detritus. Sediment derived from recycled Lower and Middle Jurassic eolianite of the southwestern United States, which contains grain ages that include distinctive Grenville-aged Proterozoic (1180–980 Ma), Neoproterozoic-early Paleozoic (680–510 Ma), and Paleozoic (490–350 Ma) ages (Dickinson and Gehrels, 2009a), is present in many Upper Jurassic units of the region and suggests export of eolianite-derived sediment in several directions during Late Jurassic time (Fig. 13A). Important stratigraphic units containing the distinctive combination of grain ages include upper Kimmeridgian strata of the Salt Wash fluvial fan of the Morrison Formation (Owen et al., 2015), derived from the Mogollon highlands or the southern end of the Cordilleran fold and thrust belt (Dickinson and Gehrels, 2008a), and the basal part of the Cucurpe Formation in the Altar-Cucurpe basin, derived from subjacent Middle Jurassic eolian sandstone (Mauel et al., 2011). Similar grain-age distributions in deformed Tithonian turbidites of the forearc region, including the Peñasquitos and Eugenia Formations (Kimbrough et al., 2014), also suggest recycling of the southwestern edge of the former erg, likely uplifted on the southern end of the Central Nevada thrust belt as early as 160 Ma (Giallorenzo et al., 2018) and routed through a low-standing arc (e.g., Busby-Spera, 1988).

Valanginian–Aptian sandstone of TSA 2 varies in lithic content and detrital zircon ages among basins, demonstrating persistence of separate sediment sources and of the basins themselves. Although pebbles of quartzite, fossilifereous limestone, and chert in some conglomerate lags indicate local Paleozoic sources, Sonoran rocks contain an upsection increase in volcanic-lithic content (Fig. 6A). Zircon ages in Rancho La Colgada and lower Morita samples from Sonora and the basal Morita sample in Arizona suggest that most Phanerozoic zircon grains were derived principally from Permian, Triassic, and Jurassic rocks, including late Paleozoic–Triassic plutons and widespread Early–Middle Jurassic plutons and ignimbrites in Sonora and the Mojave Desert region of southwestern Arizona and southern California (Figs. 7 and 13B; Riggs et al., 1993, 2013, 2016; Barth and Wooden, 2006; Arvizu et al., 2009; Mauel et al., 2011; González-León et al., 2011). Local eolian quartzarenite interbedded with Middle Jurassic ignimbrites in Sonora (Leggett, 2009; Mauel et al., 2011) likely yielded Yavapai-Mazatzal, Granite-Rhyolite, Grenville, and Paleozoic ages observed in the older samples (Fig. 7). Zircon grains in the range of ca. 149–133 Ma and andesite clasts with ages ranging from ca. 143 Ma to 125 Ma (Peryam et al., 2012) in the lower part of the section in Sonora indicate that an Early Cretaceous continental margin arc was a principal source for the Altar-Cucurpe basin.

Volcanic-lithic composition and interpreted near-depositional Aptian zircon MDAs higher in the Morita section of southern Arizona demonstrate that volcanogenic sediment flooded the Huachuca basin by mid-Aptian time (ca. 123 Ma; Fig. 7). As in Sonora, abundant grains in the range of ca. 127–120 Ma imply that sediment was derived from the Santiago Peak arc, which occupied the continental margin beginning ca. 138 Ma (Wetmore et al., 2003). Abrupt influx of arc-derived sediment with a wide range of Early Cretaceous ages into the basin suggests extension of the fluvial catchment into the arc (Fig. 13B). In Sonora, the proportion of sediment derived from the Cretaceous arc likewise increased over time (Figs. 6A and 7). An Aptian influx of arc detritus did not take place in the Bootheel basin, where the Hell-to-Finish Formation contains no volcanic-lithic fragments and few zircon grains with Mesozoic ages. Rather, feldspathic sediment containing Proterozoic zircon grains was derived from basement exposed as early as Aptian time on the rift shoulder north of the basin (Fig. 7, sample HF-1; Mack et al., 1986; Mack, 1987a). Contrasting sandstone composition and lack of arc-derived sediment confirm that the Bootheel basin in southwestern New Mexico was separate from other sub-basins of the Bisbee basin.

Volcanic sources continued to dominate sediment of TSA 4 in Sonora and likely in southeastern Arizona, but composition and zircon age distributions in the Bootheel basin indicate a shift to sources in older sedimentary strata (Fig. 8; Dickinson et al., 2009; Clinkscales and Lawton, 2015). Quartzose composition, well-rounded quartz grains, and paleoflow directions in the Mojado Formation were interpreted to indicate dominantly sedimentary source rocks, probably part of a fold-thrust orogen that lay west of the basin (Mack, 1987a). Eastward sediment dispersal, despite apparently contradictory zircon age modes that suggest ultimate eastern Laurentian origin, require that the enormous volume of compositionally mature detritus in the Cintura and Mojado Formations of the Bootheel basin was routed from the west, likely from Jurassic eolianite strata in thrust sheets in southern Nevada, such as the Wheeler Pass thrust of the Central Nevada thrust belt (Fig. 13C; Giallorenzo et al., 2018). Late Triassic age peaks in the TS4 strata are likewise explained by recycling of Triassic units such as the Upper Triassic Chinle Formation (Chinle Group of Lucas, 2004), which contains abundant Late Triassic zircon grains (Dickinson and Gehrels, 2008; Riggs et al., 2013). The Triassic units were also uplifted in thrust sheets of southern Nevada. Alternatively, Late Triassic plutons in the Mojave Desert region (Barth and Wooden, 2006; Riggs et al., 2013) could have provided grains of the appropriate ages.

Recycling of Jurassic eolianite strata that lay on the rift shoulder of the Bisbee basin has been suggested as a source for sediment of the Cintura and Mojado Formations (Dickinson et al., 2009), but our analysis suggests alternative sources farther to the west, in the southern end of the Sevier orogenic belt (Fig. 13C). By Late Jurassic time, the eolian sandstones were already eroded well back from the rift shoulder, as is demonstrated by progressive overlap of the Upper Jurassic Morrison Formation onto Triassic strata in central New Mexico (New Mexico Bureau of Geology and Mineral Resources, 2003). By middle Cretaceous time the Dakota Formation had onlapped Permian strata along the Mogollon Rim in Arizona, the site of the former rift shoulder (Reynolds, 1988; Dickinson et al., 1989). Presence of zircon derived from local basement in the locally arkosic Hell-to-Finish Formation further indicates extensive unroofing of the rift shoulder prior to deposition of TSA 4 strata. The presence of Late Triassic grains in Cintura and Mojado samples is not well explained by derivation from the eolianites; instead, the early Mesozoic grains were likely derived from basement sources in the Mojave Desert or Triassic strata in the Cordilleran thrust belt.

Compositional and detrital zircon similarities of our Dakota sample to the Mojado Formation (Fig. 8) can likewise be explained by derivation from Triassic strata and Jurassic eolianite. However, the location of our Dakota sample far north of the former rift shoulder makes it more difficult to determine actual sediment-dispersal pathways, whether directly from the Cordilleran thrust belt or more local sources. Recycling of Dakota sediment by erosion of previously deposited foreland basin strata from a time-equivalent forebulge that ran northeast, parallel with the front of Sevier orogenic belt, as suggested by Aubrey (1998), is consistent with paleocurrent data and not precluded by the zircon data (Figs. 8 and 13C). An alternative possibility—that zircon grains of the Dakota Formation were recycled, in part, directly from subjacent units within the beveled Paleozoic–Triassic stratigraphic section on the north flank of the former rift shoulder—is suggested by local variability of grain-age distributions and locally abundant Triassic grains in Dakota samples from south-central New Mexico (Stopka, 2017).

Ash plumes transported zircon grains directly to the Bootheel basin and northwestern New Mexico to be deposited in the Mojado and Dakota Formations in late Albian–early Cenomanian time (Fig. 13C). This conclusion is derived from the following observations: (1) Abundant neovolcanic grains are present in strata interbedded with otherwise quartzose sandstone in the Mojado and Dakota Formations; (2) neovolcanic sandstones yield calculated MDAs consistent with published biostratigraphic age assessments where biostratigraphy is available (Table 2, Fig. 4); (3) sandstone MDAs and tuff ages young stratigraphically at all localities (Fig. 12); and (4) younger U-Pb tuff ages in superjacent Mancos strata are likewise consistent with ammonite biostratigraphy. Neither a developing forebulge nor a relict rift-flank source can explain the presence of syndepositional zircon ages in the Dakota sample. A mixture of euhedral quartz, sanidine, tuffaceous fragments, and young zircon grains suggests that the volcanic-lithic sandstones were reworked directly from intrabasinal ash beds. Following transgression of low-energy shelfal conditions into the region, the ash falls were preserved as tuffs and bentonites of the Mancos Formation and younger parts of the complexly intertongued Mancos and Dakota Formations at San Ysidro, New Mexico (Figs. 4 and 13D).

Subsidence of the Late Jurassic–Middle Cretaceous sedimentary basins followed broadly similar historical trajectories in Sonora and southwestern New Mexico (Figs. 11A and 11B). As described above, both basins experienced rapid Late Jurassic subsidence rates that waned during Berriasian–early Aptian time and accelerated during the Albian; however, neither of the subsidence curves follows trends diagnostic of standard basin histories, and this likely indicates superposition of simultaneous tectonic factors (e.g., Xie and Heller, 2009). We infer that rapid Late Jurassic subsidence in the Altar-Cucurpe and Bootheel basins (Fig. 11) resulted from crustal extension that accompanied westward migration of the Late Jurassic arc and subduction zone (Lawton and McMillan, 1999; Dickinson and Lawton, 2001a; Fitz-Díaz et al., 2018). In this scenario, initial rapid subsidence resulted from synrift crustal stretching, with slightly greater tectonic subsidence in Sonora than in New Mexico a likely result of greater crustal thinning with increased proximity to the arc. Initial crustal extension resulted in local basins and uplifts typical of rifted settings (e.g., Gawthorpe and Leeder, 2000).

Several aspects of the Jurassic–Aptian transition suggest that subsidence was not a consequence of simple rifting: (1) Neither subsidence curve (Fig. 11) has a smooth transition from synrift stretching to post-rift thermal subsidence; (2) evidence for a depositional hiatus exists between Upper Jurassic and Lower Cretaceous stratigraphic sections in the Altar-Cucurpe and Huachuca basins and can be inferred in the Bootheel basin; and (3) at some localities in Sonora, Upper Jurassic strata were folded prior to deposition of superjacent Lower Cretaceous strata. These characteristics of the Late Jurassic–Early Cretaceous stratigraphic transition support inferences of Late Jurassic transcurrent movement along basin-bounding faults (e.g., Anderson and Nourse, 2005; Bassett and Busby, 2005; Busby et al., 2005; McKee et al., 2005, and references in those papers); nevertheless, we do not view such transcurrent deformation as resulting from a throughgoing transform fault such as the Mojave–Sonora megashear, but rather as a widely distributed response to northwest–southeast separation of North and South America during opening of the Atlantic and Gulf of Mexico basins (e.g., Pindell and Kennan, 2009). Long-distance separation of sediment sources and sinks is not supported by the provenance data described above.

The middle- to late-Albian increase in subsidence rate likely resulted in part from onset of flexural subsidence following emplacement of a tectonic load (e.g., Mack, 1987a). Although we still cannot specify a particular load, pebbles of limestone, dolostone, chert, and granite in La Juana formation at the top of TSA 4 in Sonora suggest nearby late Albian uplift of the local Paleozoic section, probably on the Caborca block (Figs. 1 and 13C). The shape of the decompacted subsidence curves at both localities resembles those that result from flexure of an elastic plate (Fig. 11; e.g., Xie and Heller, 2009), but the distribution of subsidence across the resulting foreland is quite broad, extending >300 km from the Caborca block (Fig. 13C), and it does not conform well with predictions of a narrow basin developed on a recently broken plate (Fosdick et al., 2014). We suggest that broadly distributed subsidence across the older basin array and burial of the former rift shoulder resulted from dynamic subsidence after emplacement of the subducted Farallon slab beneath the U.S.–Mexico border region, as discussed further below.

Concomitant evolution of sedimentary basins of the southwestern U.S.–Mexico border region and the continental margin arc in southernmost California and the Baja California Peninsula suggests a common link with the history of the Mexican Guerrero composite volcanic terrane (Fig. 13). The extensional phase of basin development was coeval with Late Jurassic–Early Cretaceous retroarc shortening that formed the Central Nevada and Sevier thrust belts north of Las Vegas, Nevada (DeCelles, 2004; Yonkee and Weil, 2015; Giallorenzo et al., 2018). As shortening to the north created the retroarc Cordilleran foreland basin (Currie, 1998; DeCelles, 2004), a migrating Jurassic magmatic arc sweeping westward across northern and central Mexico had just passed the site of the Late Jurassic Bisbee rift basins (Fig. 13A; Mauel et al., 2011). Recent models for the origin of Guerrero posit the opening of a marginal oceanic Arperos basin between mainland Mexico and the westward-migrating arc in latest Jurassic–Aptian time (Fig. 13B) (Martini et al., 2014; Fitz-Díaz et al., 2018). The opening phase of the Arperos basin thus corresponded in time with synrift extension and decreasing Early Cretaceous subsidence rates in at least two of the Bisbee sub-basins. Although the likelihood of regional transcurrent faulting that might have resulted from continental separation of North and South America (e.g., Pindell and Kennan, 2009) cannot be discounted, the proximity of a westward migrating arc (e.g., Fitz-Díaz et al., 2018) that provided sediment to Jurassic and Early Cretaceous backarc basins is here regarded as the primary basin-forming mechanism.

The Arperos basin subsequently closed in central and southern Mexico in Cenomanian time (Martini et al., 2014), and limited paleomagnetic and geochronological data suggest no lateral translation of Guerrero relative to North America between the opening and closing of the marginal basin (Boschman et al., 2018). Ongoing magmatism from ca. 145–124 Ma in the Mexican segment of the arc system, consisting of the Alisitos arc of the Guerrero terrane and the Santiago Peak arc in southern California and the northern Baja California Peninsula, is recorded in detrital zircon analyses of Upper Cretaceous sandstones of northern and central Mexico (Lawton et al., 2009; Juárez-Arriaga et al., 2019). Detritus in the Morita Formation of Sonora records time-equivalent magmatism of the Santiago Peak arc in southern California (Fig. 7; e.g., Wetmore et al., 2002, 2003). The same time interval marked a magmatic decline in the California arc of the Sierra Nevada (Ducea, 2001; DeCelles et al., 2009) and a shift of pluton emplacement to the western flank of the batholith (Nadin et al., 2016). The contrasting along-strike magmatic histories of Guerrero and the western United States mirrored the respective, contemporary along-strike transition from Late Jurassic–Early Cretaceous backarc extension to retroarc shortening.

Sedimentary basin development and provenance relationships of the U.S.–Mexico border region outlined above thus evolved together with opening and closing phases of the marginal basin model, and this provides a mechanism for observed patterns of basin subsidence and sediment sources. Pre-Neogene restoration of the Lower Cretaceous arc rocks of southern California and the Baja California Peninsula, according to estimates of extension in northern Sonora (Gans, 1997) and dextral offset across the Gulf of California (Fletcher et al., 2007), places the continental Santiago Peak arc ∼100 km from the northwestern end of the Altar-Cucurpe basin. The Santiago Peak arc could have provided sediment, including andesite pebbles, to the nearby extensional backarc basin even as the Alisitos arc pivoted away from North America at the Agua Blanca fault to become progressively more isolated from the continental rift basins as the marginal basin opened. Continued Valanginian–middle Aptian extension accompanied marginal basin development; this is indicated by the subsidence curves from Sonora and New Mexico (Figs. 11 and 13B).

The marginal basin model predicts that closure of the Arperos basin began sometime between mid-Aptian and mid-Albian time. Diachronous closure of the marginal basin and progressive suturing of Guerrero may explain apparent inconsistencies in the timing of the post-rift increased subsidence (Fig. 11). We infer that the well-defined, late Albian subsidence event in Sonora resulted from collision of Guerrero with North America (e.g., Martini et al., 2013; Palacios-García and Martini, 2014). Collision-induced shortening of basinal rocks and North American platformal strata yielded locally derived pebbles in La Juana formation of the Sonora section, which occupied the flexural foredeep (Figs. 11B and 13C). In contrast, a specific tectonic load has yet to be identified in southwestern New Mexico adjacent to the thick Mojado Formation; instead, that section is part of a broad subsiding region that preserved the former basins as separate depocenters with their own provenance and sediment routing characteristics. Widespread sedimentation eventually buried the former rift margin in southwestern New Mexico (Fig. 12) and accompanied resumed sediment accommodation following a hiatus of at least 20 m.y. in northwestern New Mexico (Fig. 2). The Albian–Cenomanian subsidence event was a long-wavelength phenomenon that resembles modeled dynamic topography resulting from initial emplacement of a slab subducted into the asthenosphere (e.g., Gurnis, 1992). Indeed, inversion of the local basins to form potential tectonic loads by crustal shortening, as recorded by Bisbee Group clasts in synorogenic conglomerate, did not take place until Turonian time in northern Sonora (González-León et al., 2011) and Campanian time in southwestern New Mexico (Lawton, 2000; Clinkscales and Lawton, 2015). The later inversion-related deposition is indicated by the Laramide subsidence events in Figure 11. We therefore interpret emplacement of the Farallon slab beneath southwestern North America following closure of the Arperos basin as the principal mechanism of regional late Albian to middle Cenomanian dynamic subsidence of the U.S.–Mexico border region.

New U-Pb zircon ages on tuffs, combined with subsidence analysis and consideration of provenance history based on sandstone petrography and U-Pb zircon detrital ages, elucidate Late Jurassic–middle Cretaceous sedimentary evolution of the Bisbee basin and its link to magmatic arc systems on the continental margin of the southwestern U.S.–Mexico border region. Three rapidly subsiding back-arc rift basins, with subsidence possibly augmented by a strike-slip component of faulting during separation of North and South America, formed in Late Jurassic time. In Valanginian–middle Aptian time, subsidence rates decreased as the three basins persisted as separate depocenters. Sandstone petrography and U-Pb detrital zircon ages indicate that sediment sources of the individual basins varied from south to north. A continental margin arc provided abundant sediment to the Altar-Cucurpe basin of northern Sonora from latest Jurassic through Aptian time, whereas arc-derived sediment did not arrive in the Huachuca basin of southern Arizona until ca. 123 Ma and failed to arrive at all in the Bootheel basin of southwestern New Mexico and southeastern Arizona. Beginning in the latter part of a late Aptian–middle Albian period of carbonate deposition, subsidence rates increased across the region to form a foreland basin partitioned by the depocenters that persisted on the sites of the older basins. Foreland basin culmination in late Albian–earliest Cenomanian time was marked by rapid subsidence across the former rifted region and burial of the rift flank.

Sediment provenance demonstrates that the broad foreland basin remained structurally partitioned. Arc-derived detritus continued to reach the proximal Sonoran part of the basin, which likely constituted a foredeep adjacent to an arc-proximal thrust belt. At the same time, quartzose detritus eroded from Jurassic eolianite strata in the southern part of the Cordilleran fold and thrust belt was delivered southeastward along the strike of the partitioned foreland to southwestern New Mexico. Foreland basin subsidence in Sonora likely had a flexural component, but long-wavelength subsidence across the foreland region reaching as far into the continent as northwestern New Mexico is interpreted as a dynamic topographic effect induced by introduction of the Farallon slab beneath northwestern Mexico and the southwestern United States. Basin evolution from rift to foreland basin paralleled Early Cretaceous opening and closing of a back-arc marginal basin that extended the length of modern Mexico east of the Guerrero composite volcanic terrane.

Our stratigraphic analysis demonstrates that strata historically included in the Bisbee Group contain at least one significant region-wide unconformity that separates Jurassic and Lower Cretaceous strata. In addition, individual formations or combinations of formations that constitute five tectonostratigraphic assemblages defined here record deposition in the U.S.–Mexico border region as it evolved from lithospheric extension to crustal shortening. Changing basin genesis from Jurassic through earliest Late Cretaceous time suggests a need to reevaluate the formal definition of the Bisbee Group and its stratigraphic components. A possible nomenclatural solution would be to raise the Bisbee Group to supergroup rank, a “formal assemblage of related or superposed groups, or of groups and formations. Such units have proved useful in regional and provincial syntheses” (North American Commission on Stratigraphic Nomenclature, 2005, p. 1570). Jurassic and Lower Cretaceous strata, separated by a readily identifiable regional unconformity, could be assigned to different groups within the Bisbee Supergroup.

George Gehrels and the staff at the Arizona LaserChron Center, supported by EAR-1649254, helped acquire zircon data. Lawton thanks Chris Clinkscales for discussions of Bisbee Group stratigraphy and structure. This study originated as part of the Master’s thesis work of Sarah E.K. Machin at New Mexico State University; she was supported in part by grants from the Geological Society of America and the New Mexico Geological Society. The manuscript was improved by insightful, thorough reviews by Carl Jacobson and an anonymous reviewer.

ERRATUM: Transition from Late Jurassic rifting to middle Cretaceous dynamic foreland, southwestern U.S. and northwestern Mexico

Timothy F. Lawton, Jeffrey M. Amato, Sarah E.K. Machin, John C. Gilbert, and Spencer G. Lucas

ORIGINAL ARTICLE: 2020, https://doi.org/10.1130/B35433.1. First published 9 April 2020

ERRATUM PUBLICATION: 2020. First published 15 July 2020

In Figure 1, "Fig. 11 Fence diagram" should instead read "Fig. 12 Fence diagram."