Although the Aguja Formation (West Texas, southwestern USA) and its fossil vertebrate fauna have been known for over a century, its basic stratigraphic requisites (type area and type section) have not been formally documented. The formation is herein subdivided into a series of formal members, and a lectostratotype section is proposed. Lithostratigraphic and biostratigraphic subdivisions are documented and integrated with geochronologic data to provide an age model for the formation. Four terrestrial vertebrate biozones are proposed. There are at least four major depositional intervals represented in the Aguja and intertonguing Pen Formations. An initial progradational deltaic succession is recorded by the La Basa Sandstone and lower part of the Abajo Shale Members of the Aguja Formation. A second phase of deposition resulted in a retrogradational shoreface succession that includes the upper part of the Abajo Shale, overlying Rattlesnake Mountain Sandstone Member, and lower part of the McKinney Springs Tongue of the Pen Formation, up to a skeletal phosphate bed interpreted to represent the maximum flooding surface. The third phase of deposition comprises a progradational deltaic succession that includes the upper part of the McKinney Springs Tongue, Terlingua Creek Sandstone Member of the Aguja Formation, and lower part of the Alto Shale Member of the Aguja Formation. This third succession records eastward migration of the strandline and withdrawal of the Western Interior Seaway from the Big Bend region. The fourth phase of deposition comprises a series of aggradational fluvial channel and floodplain successions that form the upper part of the Alto Shale Member and is coincident with redirection of stream flow to the southeast. This interval is much thicker in the central part of the Big Bend region, thins to the southwest and northeast, and likely records initial subsidence in the Laramide Tornillo Basin. The upper part of this succession was also contemporaneous with a series of basaltic pyroclastic eruptions, the westernmost expression of the Balcones igneous province. A dramatic constriction in the southern entrance to the Western Interior Seaway through the Gulf of Mexico occurred during this final phase in deposition of the Aguja Formation and corresponds to a shift of stream flow southeastward and to an outbreak of local pyroclastic eruptions. Regional uplift associated with this episode of magmatism is likely responsible for closing the southern aperture of the Western Interior Seaway.

Upper Cretaceous paralic and continental strata of the Aguja Formation in the Big Bend region of West Texas (southwestern USA) are significant in preserving a diverse Campanian fossil vertebrate fauna that is one of the southernmost known in North America. This fauna is probably best known for including the giant crocodylian Deinosuchus (Colbert and Bird, 1954; Cossette and Brochu, 2020) and dinosaurs such as Aquilarhinus (Prieto-Márquez et al., 2019) and Agujaceratops (Lehman et al., 2017). Exposures of the Aguja Formation comprise a remnant of the paralic and continental deposits that accumulated along the western margin of the Western Interior Seaway in Texas, deposits that are much more extensive elsewhere to the north in the Rocky Mountains and Great Plains of North America (Lehman, 1997). Hence, the Aguja Formation provides key constraints on the paleogeography and paleoenvironments of the southwestern shoreline of the Western Interior Seaway (e.g., Lehman and Busbey, 2007). Although the stratigraphy and depositional settings for similar correlative terrestrial and paralic strata farther to the north in New Mexico (e.g., Ambrose and Ayers, 2007) and to the south in Mexico (e.g., Eberth et al., 2004) have been well documented, the basic stratigraphy and sedimentology of the Aguja Formation remain poorly known. The fossil fauna and flora of the Aguja Formation have figured prominently, however, in broader studies of Late Cretaceous biodiversity and biogeography in the Western Interior (e.g., Gates et al., 2010, 2012; Fowler, 2017).

Background and Purpose

The fundamental purpose of the present study is to document the stratigraphic framework, formal subdivisions, and depositional history of the Aguja Formation and so provide a foundation for future, more detailed studies. These strata were first described by Udden (1907) who designated them the “Rattlesnake Beds” in recognition of the typical exposures around Rattlesnake Mountains in what is today Big Bend National Park (Fig. 1). However, the name “Rattlesnake Formation” was already in use for Pliocene strata in Oregon (Merriam, 1901), and so Adkins (1932) substituted the name Aguja Formation, in honor of Sierra Aguja (a.k.a. “Needle Peak”) ~6 km southwest of Rattlesnake Mountain.

Maxwell et al. (1967) provided the first thorough account of the stratigraphy of the Aguja Formation and mapped its distribution across the entire park area. Lehman (1985a) subsequently proposed a series of informal stratigraphic subdivisions of the Aguja Formation; however, at the time of that work, it was uncertain whether these informal units could be correlated throughout the Big Bend region or mapped effectively. Outcrops of the Aguja Formation are isolated because these strata occur in multiple fault-bounded exposures, and intervening areas are interrupted by late Paleogene intrusive rocks or covered by Neogene and Quaternary alluvium (Fig. 1). Consequently, it is not possible to physically trace units recognized in one outcrop belt to another.

During the 1980s and 1990s, detailed local studies of the Aguja Formation and mapping of key areas documented the stratigraphic relationships in areas beyond those initially investigated by Lehman (Davies, 1983; Bohanan, 1987; Record, 1988; Schroeder, 1988; Mosley, 1992; Weil, 1992; Rowe et al., 1992; Macon, 1994; Cifelli, 1994; Straight, 1996; Tomlinson, 1997; Sankey, 1998). As a result of these studies, the extent of each of the informal members was better delineated, and their distribution was shown on a series of geologic maps.

Beginning in 2002, a joint U.S. Geological Survey and National Park Service effort was undertaken to produce a new geologic map of Big Bend National Park (Turner et al., 2011). As part of that effort, all exposures of the Aguja Formation within the park were investigated and mapped, and the formation subdivisions were found to be applicable to all exposures of these strata (e.g., Lehman, 2002, 2004, 2007), although at the scale of the resulting map (1:75,000) the Aguja Formation was shown undivided. Regardless, during the 2000s, the informal subdivisions became more widely cited and referred to in numerous other studies (e.g., Wheeler and Lehman, 2000, 2005, 2009; Wagner, 2001; Anglen, 2001; Waggoner, 2006; Breyer et al., 2007; Befus et al., 2008; Lewis, 2011; Fowler, 2017).

Basic goals of the present report include the formal recognition of heretofore informal stratigraphic subdivisions of the Aguja Formation, mapping of their distribution, and summarization of the depositional history of these strata. Documentation of the Aguja stratigraphic subdivisions allows for accurate correlation between separate exposures, details the relative positions of significant paleontological sites, and provides a consistent framework for future studies. Mapping of the individual members is also useful in determining the stratigraphic levels of numerous shallow intrusive rocks emplaced within the formation throughout the Big Bend region (e.g., Kovschak, 1973; Turner et al., 2011). The maps presented in this report illustrate key exposures of the Aguja Formation and are excerpted from previously unpublished maps cited by Lehman (2002, 2004, 2007), Collins et al. (2006), and Cooper (2011) that were the basis for parts of the geologic map of the park (Turner et al., 2011).

Further, although the Aguja Formation has been the focus of many individual local stratigraphic and paleontological studies, much of this work remains in unpublished theses and dissertations. Hence, an additional aim of the present report is to review results of these unpublished works and integrate those findings into a broader summary of the stratigraphy and depositional history of the formation. Finally, available biostratigraphic and geochronologic data are placed within the proposed lithostratigraphic framework and used to develop an age model for the Aguja Formation. A biostratigraphic zonation for the terrestrial vertebrate faunas is also presented.

General Character and Distribution of the Aguja Formation

Although the Aguja Formation varies lithologically, it includes a series of prominent well-indurated sandstone beds that hold up extensive cuestas and hogback ridges wherever exposed. As a result, these strata are physiographically distinct from the underlying and intertonguing recessively weathering marine shale of the Pen Formation. The distinctive ridge-forming Aguja sandstone beds are light gray, tan, or reddish brown and alternate with drab yellow, olive, and gray mudstone intervals that form intervening slopes and badlands. The overlying Javelina Formation is likewise distinct, including brightly colored beds of red, purple, and blue mudstone as well as conglomerate beds with chert pebbles that are absent in Aguja sandstones (Lehman et al., 2018).

Strata of the Aguja Formation are, for the most part, restricted to two Neogene graben—the Delaho Bolson and Estufa Bolson of Stevens and Stevens (1989), or together, the “sunken block” of Udden (1907). This structural depression lies between the Santiago Mountains and Sierra del Carmen ranges on the east and the Mesa de Anguila on the west (Fig. 1) and occupies much of the region encompassed by the “Big Bend” of the Rio Grande in Texas as well as adjacent areas to the south in Coahuila, Mexico (Shiller, 2017). A few outlying outcrops also identified as Aguja Formation lie beyond this region in the Providencia Carbón district in northeastern Chihuahua, Mexico (e.g., Brown et al., 2004; Westgate et al., 2006), although strata in this area have alternatively been identified as part of the San Carlos Formation (e.g., Alexandri-Rionda et al., 2008a). Isolated exposures are also found farther to the east in the Pico Etereo district in northern Coahuila (e.g., Daugherty and Powell, 1963), although strata in that area are alternatively shown as the San Miguel Formation (e.g., Alexandri-Rionda et al., 2008b).

The Aguja Formation is widely exposed throughout Big Bend National Park and the surrounding vicinity, mainly in the valleys of Tornillo and Terlingua Creeks, the two major tributaries of the Rio Grande in this region (Fig. 1). There are three primary outcrop belts of the Aguja Formation. A western outcrop belt extends from Sierra Aguja northward to Study Butte, and from there to Nine Point Mesa (Fig. 1). Uplift of the entire section along the Burro Mesa fault results in a central to northeastern outcrop belt extending from Chisos Pen eastward to Grapevine Hills and northward to Persimmon Gap. A southeastern outcrop belt winds around Cow Heaven Mountain, eastward to Mariscal Mountain and Sierra San Vicente, and into the lower reaches of Tornillo Creek (Fig. 1). This southeastern outcrop belt extends in places further southward into Coahuila, Mexico (Fig. 2; Shiller, 2017).

Most outcrops of the Aguja expose only part of the formation, and due to interruption by faults, intrusive igneous rocks, and alluvial cover, it is not possible to physically trace individual stratigraphic units within the Aguja Formation between the major outcrop belts. Moreover, there are no subsurface stratigraphic data available for the park or surrounding area. As a result, the regional correlation proposed here is based entirely on outcrops and lithologic characteristics. For the present report, stratigraphic sections were measured at 34 sites to show regional variation across all three of the outcrop belts and to record the stratigraphic positions of all significant vertebrate fossil localities and paleobotanical collection sites (Item S1 and Tables S1 and S2 in the Supplemental Material1). The proposed correlation of stratigraphic sections in each outcrop belt is shown here in a series of panel diagrams (Figs. 36). Portions of selected stratigraphic sections for each member of the Aguja are also shown in the lithostratigraphy section in greater detail to illustrate the typical sedimentary facies of each unit, and several outcrop drawings show examples of the typical facies relationships.

Subdivisions of the Aguja Formation

Several authors have informally subdivided the Aguja Formation. Knebusch (1961) indicated that the Aguja comprised “lower”, “middle”, and “upper” members. Maxwell et al. (1967) described the Aguja as consisting of generally a lower “marine phase” and an upper “non-marine phase,” and Kovschak (1973) subdivided the Aguja into “marine”, “fluvial”, and “lacustrine” intervals. Lehman (1985a) found that these broad facies subdivisions could not be readily correlated, mapped, or applied to exposures of the formation outside the limited areas where they had been utilized.

In establishing basic informal subdivisions of the Aguja Formation, Lehman (1985a) instead employed the paradigm established for other intertonguing paralic and marine Upper Cretaceous strata in the Western Interior Basin of North America; that is, each of the paralic sandstone units was identified with its own name (the basal sandstone, Rattlesnake Mountain sandstone, and Terlingua Creek sandstone members), and the intervening continental units comprising predominantly mudstone were designated separately (the lower shale and upper shale members).

As with other paralic strata that accumulated along the shore of the Western Interior Seaway, the Aguja Formation exhibits an intertonguing relationship with marine shelf deposits—in this case, the Pen Formation. The thickest and most widespread of these intervening marine shale tongues had in earlier works been considered part of the Aguja Formation; however, Lehman (1985a) included it instead as an informal member of the Pen Formation (the McKinney Springs tongue). Lehman (1985a) also described a “middle shale member” of the Aguja Formation in the western outcrop belt of the formation that included the thin interval of marine shale intervening between the Rattlesnake Mountain sandstone and Terlingua Creek sandstone members. However, it was subsequently discovered that this unit is simply the thin landward equivalent of the McKinney Springs tongue, which is much thicker and better exposed in the eastern outcrop belt. The two units are one in the same, and so strata formerly identified as the “middle shale member” were later included instead within the McKinney Springs tongue (Mosley, 1992).

In the present report, the major sandstone intervals of the Aguja Formation, previously referred to informally as the “basal sandstone”, “Rattlesnake Mountain sandstone”, and “Terlingua Creek sandstone”, are formally designated La Basa Sandstone, Rattlesnake Mountain Sandstone, and Terlingua Creek Sandstone Members (Figs. 26). The intervening mudstone-dominated intervals, previously referred to informally as the “lower shale” and “upper shale”, are herein formally designated the Abajo Shale Member and Alto Shale Member, respectively. The McKinney Springs Tongue is formally designated part of the Pen Formation. In the following descriptions for each member of the Aguja Formation, an account is given of its basic features, its extent is delineated, a typical section is established, the sedimentary facies typical of each are briefly described, and an interpretation of the depositional environment is presented.

A formal type section has not been properly designated for the Aguja Formation. When Udden (1907, p. 41) first described these strata, he recognized the outcrops around Rattlesnake Mountains as typical of the formation but did not measure the stratigraphic section there. He did, however, measure several sections in the vicinity of Chisos Pen that Lehman (1985a) suggested could serve as a composite type section (Fig. 2; Udden, 1907, p. 46–50; Lehman, 1985a, figure 11). When Adkins (1932, p. 505) first applied the name Aguja Formation to these strata, he designated the type area as instead near Sierra Aguja “…in front of the Santa Elena fault scarp, six miles south of Terlingua” but did not measure a section there or anywhere else. Later, in their review of the Aguja Formation, Maxwell et al. (1967, p. 79–87) described several additional stratigraphic sections but did not address the lack of a type section. As a result, there is no formal stratotype section for the Aguja Formation. The exposures at Rattlesnake Mountains, Chisos Pen, and Sierra Aguja all suffer from inadequacies that make it difficult to fully characterize the formation in any of these areas. Sections measured in these areas for the present study are herein designated auxiliary reference sections, and in the absence of a satisfactory original stratotype, a lectostratotype section is proposed instead on Dawson Creek where a complete section is exposed, easily accessed, and relatively free of structural complexity. Geologic maps and measured sections are provided herein for each of these four areas (Rattlesnake Mountains, Sierra Aguja, Chisos Pen, and Dawson Creek) to demonstrate the mappability of proposed formation subdivisions, and key aspects of the exposures in each area are briefly described.

Rattlesnake Mountains

Exposures on the southwestern flank of Rattlesnake Mountain provide a nearly complete section through the Aguja Formation, and this is presumably where Udden (1907) originally characterized the formation (Fig. 7; Item S1, this is auxiliary reference section 1). Exposures on the northern and eastern sides of the mountain do not include the upper half of the formation (Fig. 4, section 14). However, a thick gabbroic sill that caps Rattlesnake Mountains intrudes the Aguja at about the middle of the Alto Shale Member, and the contact with the overlying Javelina Formation is not exposed in the immediate area. Nevertheless, much of the lower part of the formation is very well exposed in this area (Fig. 3, section 7). A composite section that spans the entire formation can be assembled here if the section exposed at Rattlesnake Mountains is properly correlated with exposures nearby in the valley of Alamo Creek (Fig. 3, section 8, beds 1–11) and from there to Peña Mountain (Fig. 3, section 8, beds 12–21) ~3 km to the southeast.

Sierra Aguja

Exposures in the vicinity where Adkins (1932, p. 505) designated the type area “six miles south of Terlingua” could include any outcrops on the northern, eastern, or southern side of Sierra Aguja (Fig. 7; Item S1, measured section 3, auxiliary reference section 2). The formation is well exposed on the northwestern side of Sierra Aguja, but this area is remote and nearly inaccessible. Moreover, much of the northeastern flank of Sierra Aguja is on private ranchland and thus difficult for the general public to access. Most of the formation is, however, well exposed on the southeastern side of Sierra Aguja, although a fault interrupts the section here and cuts out most or all of the Abajo Shale Member in this area (Fig. 3, section 3, beds 6–17). The lower part of the formation is, however, well exposed nearby on the south side of Terlingua Creek just 0.5 km to the southeast (Fig. 3, section 3, beds 1–5), and so a composite section encompassing the entire formation can be assembled by correlation here as well.

Chisos Pen

Exposures of the Aguja Formation in the Chisos Pen area form a prominent cuesta that extends more than 4 km north of Burro Mesa and is bisected by Cottonwood Creek (Fig. 8; Item S1, auxiliary reference section 3). The section here is interrupted by several mafic sills; however, assuming that the sills are completely concordant with the stratification, it is possible to determine the thicknesses of all units except the Alto Shale in this area (Fig. 5, section 17). The only stratigraphic sections of the Aguja Formation presented by Udden (1907, p. 46–50) were measured in the Chisos Pen area, although none of the sections he measured there extend through the entire formation (Fig. 2). A section measured for the present report in the same area (Fig. 5, section 17) includes all of the members of the formation, but the contact with the overlying Javelina Formation cannot be reached in this vicinity. The conglomerate bed that Lehman (1985a, figure 11; Udden, 1907, p. 50, stratigraphic section “east of Chisos Pen”) interpreted as the base of the Javelina Formation is instead now believed to be correlative with the pyroclastic interval of the Aguja (see description of Alto Shale Member) and is likely at least 30 m below the upper formation contact. The contact with the Javelina Formation is instead exposed ~3 km southwest of the Chisos Pen area, near Gano Spring (Lewis, 2011), but there is extensive alluvial cover over the area intervening between these outcrops. At least one fault displaces the Alto Shale Member here as well (Fig. 8). Therefore, although Lehman (1985a) and Schiebout et al. (1987) suggested that the section here could serve as the type section, the full thickness of the Aguja Formation cannot be determined in the vicinity of Chisos Pen.

Dawson Creek

The Aguja Formation is well exposed at two sites along Dawson Creek near the western border of Big Bend National Park; one site is on the north bank of the creek 2 km west of the park entrance (Figs. 4 and 9, section 16), and the other is on the south bank near the headwaters of the creek 3 km east of the park entrance (Figs. 4 and 10, section 15). Contacts with the underlying Pen and overlying Javelina Formations are well exposed in both areas. However, at the western exposure, the Alto Shale Member is almost entirely covered by alluvium and interrupted by several dikes and sills (Fig. 9). Nevertheless, a composite section can be assembled here if the Alto Shale section is measured near the park highway, 2 km east of the main exposure, and correctly correlated with the lower part of the section measured further west (Fig. 4, section 16; Item S1, measured section 16, the auxiliary reference section 4).

In contrast, the exposure in the headwaters of Dawson Creek (Fig. 10; Item S1, measured section 15, the lectostratotype section) is relatively free of alluvial cover and shows both lower and upper contacts clearly. All of the members of the formation are present here. This area is also readily accessible and relatively free of complexities introduced by intrusive rocks or faulting. Even here, however, a minor fault slightly displaces the Aguja-Javelina contact, requiring proper correlation to attain the full thickness of the Alto Shale. Because this outcrop provides a complete uninterrupted exposure of all named members of the formation, we designate the section in the Dawson Creek headwaters as the lectostratotype section for the formation (Fig. 4, section 15).

La Basa Sandstone Member (New)

The lowermost stratigraphic unit in the Aguja Formation is equivalent at least in part to the “basal sandstone” mentioned by Maxwell et al. (1967, p. 79) and later referred to informally as the “basal sandstone member” by Lehman (1985a, p. 42). To preserve continuity with previous literature, the basal sandstone is herein designated La Basa Sandstone Member (“la basa” being Spanish for “the base”) in reference to its original title given by Maxwell et al. (1967).

A type section for the La Basa Sandstone Member is designated in the southwestern foothills of Rattlesnake Mountain (Fig. 3, section 7) where this unit is easily accessed and contacts with underlying and overlying units are well exposed. Because this unit varies in its stratigraphic relationship with overlying strata, a reference section for this member is designated near McKinney Springs (Fig. 5, section 21) where the La Basa Sandstone is overlain by the McKinney Springs Tongue of the Pen Formation.

The lower contact of the La Basa Sandstone is also the lower boundary of the Aguja Formation. The upper part of the underlying Pen Formation consists of easily eroded, slope-forming, gray or tan-yellow calcareous shale. In the uppermost 5–20 m of the Pen Formation, thin sandstone beds interfinger with the shale, and these sandstone beds are thicker and more closely spaced ascending to the contact with the La Basa Sandstone Member (Fig. 11H). The contact is placed arbitrarily within this zone of gradation at the base of the lowest prominent ridge-forming sandstone bed. In many areas, particularly along the western outcrop belt, there are one or more significant (i.e., >1 m thick) sandstone beds also in the underlying Pen Formation 5–20 m below the base of the La Basa Sandstone (Fig. 12). Typically, these beds cannot be traced beyond the limits of a given outcrop.

The La Basa Sandstone Member is a prominent thick sandstone bed or succession of sandstone bedsets ranging from 5 to 20 m in total thickness. Contrary to the report of Maxwell et al. (1967), this unit rests in gradational, not unconformable, contact with the underlying Pen Formation. It is exposed intermittently along the western Aguja outcrop belt, where it is overlain by the Abajo Shale Member. In the central outcrop belt, the La Basa Sandstone is exposed at the Chisos Pen anticline (Fig. 5, section 17) and is also overlain there by the Abajo Shale, but elsewhere in the central outcrop belt, the base of the Aguja Formation is not exposed. In the southeastern outcrop belt, exposures of the La Basa Sandstone are extensive and throughout most of this region overlain by the McKinney Spring Tongue of the Pen Formation.

The La Basa Sandstone pinches out within the Pen Formation at the La Clocha outcrop (Figs. 6 and 13, section 30), and so east of that outcrop the Terlingua Creek Sandstone Member is mapped as the base of the Aguja. Similarly, the La Basa Sandstone is absent at all exposures north of McKinney Springs, from Dagger Flat (Fig. 13, section 23) to Bone Spring Draw and Persimmon Gap (Fig. 5, section 25), and so it is inferred that it also pinches out in the intervening covered area. Isolated outcrops of Aguja Formation overlain and underlain by marine shale of the Pen Formation in northern Coahuila at Pico Etereo (Daugherty and Powell, 1963) and at La Unión (Shiller, 2017) indicate that the La Basa Sandstone likely extends south and east into those areas (Fig. 2).

In the western and central outcrop areas, the La Basa Sandstone is not well indurated and so does not typically form a prominent cliff or cuesta (e.g., Fig. 11E). In most areas, the La Basa Sandstone is also typically light gray or white due to high kaolinite content (e.g., Fig. 11H) rather than tan, yellow, or dark reddish brown as is typical of sandstone beds higher in the formation. However, at the eastern outcrops of the La Basa Sandstone, for example at Cow Heaven Mountain (Fig. 6, section 26), near La Clocha (Fig. 6, section 30), or at McKinney Hills (Fig. 5, section 21), the sandstone is well cemented and in those areas weathers to form more prominent physiographic features.

Abajo Shale Member (New)

The unit previously referred to informally as the “lower shale member” (Lehman, 1985a, p. 46) is herein formally designated the Abajo Shale Member (“abajo” being Spanish for “below”). The name preserves continuity with previous literature usage of the term “lower shale” and also refers to Terlingua Abajo—an abandoned village near the mouth of Terlingua Creek (U.S. Geological Survey, 1971a) where these strata are well exposed.

A type section is designated for the Abajo Shale Member in the southwestern foothills of Rattlesnake Mountain (Fig. 3, section 7), where this unit is easily accessed and contacts with underlying and overlying units are clearly evident. The contact between the La Basa Sandstone Member and the recessively weathering Abajo Shale is placed at the top of the uppermost prominent sandstone bed in the La Basa Sandstone; this contact coincides with the base of the lowermost carbonaceous mudstone, lignite, or coal bed in the Abajo Shale.

The Abajo Shale Member consists of dark gray and olive carbonaceous mudstone with beds of lignite and coal near the base (Figs. 11A, 11C, and 11F). Because the Abajo Shale is a recessive, slope-forming unit, it is covered by alluvium in many areas. Where it is exposed, it forms striking dark badlands terrain (Fig. 11G). Local sections of the Abajo Shale have been described by Record (1988), Wick et al. (2015), Brink (2016), and Lehman et al. (2019).

Where the Abajo Shale Member is thickest, it can generally be subdivided into three units (Fig. 14). Unit 1 comprises the lower 10–30 m and consists of dark gray carbonaceous mudstone, lignite, and coal interbedded with thick local lenses of fine-grained white sandstone (estuarine deposits described in Sedimentary Facies section). This part of the Abajo Shale contains the most extensive and highest-rank bituminous coal seams found in the Aguja Formation (Evans, 1974). Unit 2 typically comprises the central 10–15 m of the Abajo Shale and consists of thin beds of light gray fine-grained sandstone interbedded with carbonaceous mudstone and prominent red or purple fossiliferous ferruginous concretion horizons. Unit 3 comprises the upper 5–15 m of the section and consists of thickly bedded, more uniformly drab olive-gray or yellow mudstone and is less carbonaceous with fewer, less-extensive lignite beds than in unit 1 (Fig. 14).

The Abajo Shale varies markedly in thickness. For the most part, this reflects regional thinning from west to east across the Big Bend region (Fig. 15). The Abajo Shale is thickest along the western outcrop belt where it ranges from 70 to nearly 100 m thick near Sierra Aguja and Rattlesnake Mountain to 40 or 30 m along Dawson Creek. Typically, all three subdivisions of the Abajo Shale are present in the western outcrop belt. In the central outcrop belt, the Abajo Shale is thinner (~30 m at Chisos Pen, section 17) and only the lower part (unit 1) is present, but at Croton Spring (section 18) and Grapevine Hills (section 19), the base of the Abajo Shale is not exposed and its entire thickness is unknown (Fig. 5). It is absent entirely in the eastern outcrop belt and must therefore pinch out within the subsurface region beneath the Chisos Mountains between the central outcrop belt and the eastern outcrop belt, although the pinch-out is nowhere observed in outcrop (Fig. 15). There are also marked local variations in thickness along strike resulting predominantly from intertonguing with the overlying Rattlesnake Mountain Sandstone Member. As a result of this reciprocal relationship, where the Abajo Shale is thickest, the Rattlesnake Mountain Sandstone is typically its thinnest.

Rattlesnake Mountain Sandstone Member (New)

The Rattlesnake Mountain Sandstone Member was previously recognized informally by Lehman (1985a, p. 49), and the name was subsequently used by Macon (1994), Lehman and Tomlinson (2004), and Schubert et al. (2016). The name refers to Rattlesnake Mountains in the southwestern part of Big Bend National Park (U.S. Geological Survey, 1971a). Although this geographic feature is known as Rattlesnake Mountains (plural), the rock unit described herein has been referred to informally as the Rattlesnake Mountain (singular) sandstone member for several decades, as has the sill that caps the mountain (e.g., Turner et al., 2011). To preserve continuity with existing literature, the singular usage is retained in the formal name for this rock unit (e.g., see article 7 in North American Commission on Stratigraphic Nomenclature, 2005). A type section for the Rattlesnake Mountain Sandstone Member is designated on the southwestern foothills of Rattlesnake Mountains (Fig. 3, section 7), where this unit is easily accessed and contacts with underlying and overlying units are well exposed.

The Rattlesnake Mountain Sandstone is typically weakly consolidated and generally forms a low cuesta, and colluvium typically obscures the contact with the underlying Abajo Shale Member (Fig. 16A). Where the contact between the two is not covered, in most places it is an abrupt erosional scour surface overlain by conglomerate composed of reworked bivalve shells and mudstone rip-up clasts (e.g., Fig. 4, section 16; Figs. 16A, 16G, and 17). In such areas, the Rattlesnake Mountain Sandstone is also its thinnest (~5 m). However, in other areas, the lower contact is instead an alternating series of interfingering sandstone and mudstone beds, gradational with the underlying Abajo Shale, and the contact is placed at the base of the lowermost of these sandstone beds (e.g., Fig. 3, section 6). In these areas, the Rattlesnake Mountain Sandstone is also at its thickest (as thick as ~35 m; Fig. 4, section 14).

The Rattlesnake Mountain Sandstone is typically a friable, very fine-grained, yellow or tan series of sandstone beds with abundant shells of ostreid bivalves, particularly near its top (Fig. 17). In most areas, the Rattlesnake Mountain Sandstone is not a prominent cliff-forming unit, in contrast to the overlying Terlingua Creek Sandstone Member. These two sandstone units are separated by the McKinney Springs Tongue of the Pen Formation. This intervening, recessively weathering marine shale is readily recognized in the eastern and central outcrop belts (see description of McKinney Springs Tongue; e.g., Fig. 16B) but is very thin (from 10 m to <1 m) in the western outcrop belt. In several areas (e.g., Fig. 4, section 10), this intervening shale has nearly or completely pinched out and the two sandstone members are in depositional contact with one another; in these areas, the contact between the two is not easily delineated and so they must be mapped together as a single unit (e.g., Schubert et al., 2016).

Terlingua Creek Sandstone Member (New)

The Terlingua Creek Sandstone Member was named informally by Lehman (1985a, p. 58) and subsequently described by Bohanan (1987), Schroeder (1988), and Anglen (2001). The name refers to Terlingua Creek, a major tributary of the Rio Grande that flows along the western boundary of Big Bend National Park and has its mouth near the entrance to Santa Elena Canyon (U.S. Geological Survey, 1971a). A type section is designated for the Terlingua Creek Sandstone on the north side of Dawson Creek (Fig. 4, section 16), where this unit forms a prominent hogback ridge, is easily accessed, and has well-exposed contacts.

The lower contact of the Terlingua Creek Sandstone is gradational and interfingering with shale of the underlying McKinney Springs Tongue of the Pen Formation, which is a recessive slope-forming unit (Figs. 16D and 18B). The contact is covered by colluvium in many areas but where exposed comprises a series of thin sandstone beds that interfinger with the shale; these sandstone beds are thicker, more closely spaced, and amalgamated ascending stratigraphically to the contact (Fig. 19). The contact is placed at the base of the lowest prominent, laterally extensive, or significant (~>1 m thick) sandstone bed.

The Terlingua Creek Sandstone is typically the most prominent ridge-forming sandstone bed in the entire formation, and in many areas it forms a striking cliff or hogback ridge (Figs. 18A18C). The sandstone is generally tan or yellow with well-cemented dark reddish brown upper beds that weather differentially to create picturesque hoodoos and “cannonball” concretions (Fig. 16F). In most areas along the western outcrop belt, the Terlingua Creek Sandstone comprises a single discrete sandstone interval ~5 m in thickness (e.g., Fig. 19, section 21). Throughout much of the central and eastern outcrop belts, however, there are instead a series of successive thick sandstone beds separated by substantial intervening mudstone intervals. In these areas, only the lowermost of these sandstone beds is mapped as the Terlingua Creek Sandstone (e.g., Fig. 6, sections 31 and 32). The Terlingua Creek Sandstone generally preserves few invertebrate fossils and in most areas is not as fossiliferous as the underlying Rattlesnake Mountain Sandstone. The uppermost sandstone beds, in the contact interval with the overlying Alto Shale Member, however, typically have abundant petrified logs (the “conifer acme zone” of Wheeler and Lehman, 2005).

Alto Shale Member (New)

The uppermost subdivision of the Aguja Formation was previously referred to informally as the “upper shale member” (Lehman, 1985a, p. 60) and is herein formally recognized as the Alto Shale Member (“alto” being Spanish for “upper”). The name preserves continuity with older literature references to the upper shale and also refers to the Alto Relex, a distinctive plateau in the eastern part of Big Bend National Park (U.S. Geological Survey, 1971c). The Alto Shale is the thickest and most widespread of all members of the Aguja Formation. Exposures of the Alto Shale are found in all three outcrop belts throughout the park and adjacent areas as well as to the south in Coahuila, Mexico, where it is much thicker (as thick as 360 m; Shiller, 2017).

A type section for the Alto Shale Member is designated for the exposures in the unnamed tributary of Alamo Creek west of Rattlesnake Mountain extending southward to Peña Mountain (Fig. 3, sections 7 and 8). Although this section is partly covered and the contact with the underlying Terlingua Creek Sandstone is not well exposed, the unit is easily accessed in this area and all of its typical subdivisions (units 1–4 described in the following paragraphs) are present here.

The contact with the underlying prominent ridge-forming Terlingua Creek Sandstone Member is gradational and intertonguing and lies within an interval of recessively weathering thick carbonaceous mudstone beds separated by thin sandstone beds. The contact is placed at the base of the lowest thick mudstone bed above the Terlingua Creek Sandstone. The overlying Javelina Formation is also typically a ridge-forming unit, with sandstone beds that hold up prominent cuestas or hogback ridges. The Alto Shale and the Javelina Formation are therefore physiographically distinct in most areas, and the contact between them is placed at the base of the lowest of the prominent ridge-forming sandstone bed, above which the mudstone beds exhibit continuous bright red or purple color bands. Conglomerate layers within these sandstone beds also typically contain chert pebbles that are lacking in any Aguja Formation conglomerates. The nature of the contact between the Aguja and Javelina Formations was examined in detail by Lehman et al. (2018) as part of their investigation of the Tornillo Group. Their revised placement of that contact resulted in places in the transfer of a substantial section of strata formerly mapped as part of Aguja Formation (e.g., Maxwell et al., 1967) to the Javelina Formation.

The Alto Shale is predominantly a recessively weathering mudstone interval with intercalated weakly consolidated sandstone beds, and as a result, in most areas this is a slope-forming unit, covered or partially covered by alluvium. Exposures typically form badlands with intervening cuestas (Fig. 18). Because the Alto Shale preserves most of the known vertebrate fauna of the Aguja Formation, stratigraphic sections have previously been illustrated for many exposures of this unit (e.g., Sankey and Gose, 2001; Lehman and Wick, 2010; Lehman et al., 2017). The Alto Shale varies markedly in thickness across the Big Bend region (Fig. 20). It is thickest in the central outcrop belt, from ~220 m (Peña Mountain; Fig. 3, section 8) to 280 m (Cow Heaven Mountain; Fig. 6, section 26), and thins to the southwest (100 m at Sierra Aguja; Fig. 3, section 3) and to the northeast (80 m at Dagger Flat; Fig. 5, section 23). The marked thickness variation in the Alto Shale reflects in part erosional relief at the base of the overlying Javelina Formation, but also syndepositional thinning to the northeast and southwest due to initial uplift of the Mesa de Anguila and Sierra del Carmen monoclines that would later form the margins of the Tornillo Basin (see Depositional History section; Lehman et al., 2018).

The Alto Shale Member exhibits several internal subdivisions that are generally recognizable but not everywhere present or easily delineated (Fig. 3). The lower 50–100 m (unit 1) of the Alto Shale consists of drab, dark gray carbonaceous mudstone, lignite, and coal, very similar in appearance to uppermost strata (unit 3) in the underlying Abajo Shale Member (Fig. 3); however, intercalated with these deposits are coarsening-upward successions capped by thin sandstone beds with conglomerates composed of abraded mollusc shells, reworked concretions, wood, and bones (Fig. 21, unit 1).

Above unit 1 and in most areas comprising the middle 50–100 m of the Alto Shale is generally yellow, tan, and light gray mudstone with intercalated thick, cross-stratified, tan or yellow sandstone beds in fining-upward successions with basal conglomerates composed of reworked calcareous concretions (Fig. 21, unit 2). Several of the sandstone intervals in unit 2 are very thick (>20 m) and laterally extensive and can be traced over most or all of a given outcrop belt. In the upper part of unit 2, the mudstone intervals include lenticular brightly colored purple and red beds with calcareous concretions. The upper part of the Alto Shale (unit 3, also referred to here as the “pyroclastic interval”) includes intercalated olive-green, rhythmically bedded, basaltic pyroclastic deposits that fill local maar craters (Fig. 22; described in the following paragraph). These pyroclastic deposits are limited to small areas, but interstratified with and extending beyond the limits of the maar craters, there are distinctive rhythmically bedded limestone and siltstone deposits at the same stratigraphic level. Unit 3 includes the most unusual and varied deposits in the Alto Shale and is locally of significant thickness (e.g., ~15 m at Peña Mountain), but this interval is not everywhere present or identifiable. Above the pyroclastic interval (unit 3) in a few areas are brightly colored yellow, tan, and red or purple mudstone beds with calcareous concretions and intercalated tan-yellow conglomeratic sandstone beds (unit 4). These deposits are similar in appearance, facies characteristics, and likely origin to those in unit 2 and comprise the uppermost 10–40 m of the Alto Shale.

Unit 3 in the Alto Shale Member is composed of several types of deposits, referred to herein collectively as the pyroclastic interval. The most distinctive of these consists of basaltic lapilli tuff deposits that fill maar craters excavated into the surrounding fluvial floodplain deposits (Fig. 22, section 8). These strata were formerly mistaken to be sedimentary epiclastic or volcaniclastic deposits (e.g., Lehman, 1985a) but later recognized as pyroclastic in origin (Breyer et al., 2007; Befus et al., 2008). Two major accumulations of pyroclastic deposits have thus far been recognized. One of these is in the headwaters of Star Creek on the southeastern side of the Rosillos Mountains (Fig. 5, section 22; Breyer et al., 2007). However, this exposure is fault bounded, and although the pyroclastic beds are intercalated with fluvial floodplain deposits at the top of the exposure, the stratigraphic relationship with the underlying and overlying Alto Shale is obscured.

The second exposure of basaltic lapilli tuff surrounds the northeastern flank of Peña Mountain, is much more extensive, and is relatively free of structural complexity (Fig. 3, section 8; Befus et al., 2008). Although this exposure is interrupted by a thick gabbroic sill, the relationship of the pyroclastic deposits to underlying and overlying Alto Shale is more clearly evident (Fig. 23). The vent for the pyroclastic deposits is exposed at the northeastern end of the outcrop, and three successions of rhythmically bedded lapilli tuff, separated by angular discordances, extend beyond the vent to the southwest and rest on fluvial overbank floodplain deposits (facies association F; see Sedimentary Facies section). The tuff deposits have interbedded laminated limestone layers, apparently deposited in lacustrine environments during quiescent interludes between eruptions (facies association G; see Sedimentary Facies section; Fig. 23). Interstratified with the pyroclastic beds are fine-grained clastic fluvio-deltaic deposits (facies association G) near the central part of the outcrop; these deposits extend further to the southwest and are gradationally overlain by fluvial overbank floodplain sediments (Fig. 23).

Kovshak (1973) and Lehman (1985a) described similar but much more extensive lacustrine and fluvio-deltaic deposits exposed on the southwestern side of Grapevine Hills, where they locally comprise a significant (as thick as 20 m) separate accumulation (Fig. 5, section 20; Fig. 18G). A direct relationship with the pyroclastic deposits is not evident at this outcrop, but the pyroclastics at Rosillos Mountains are exposed a short distance to the northeast (Breyer et al., 2007). In other areas where pyroclastic deposits are absent and no discrete tuff beds are apparent, the fluvial floodplain mudstone deposits (facies association F; see Sedimentary Facies section) at a comparable stratigraphic level exhibit an unusual pea-green or blue coloration due to high chlorite content, a likely alteration product of the mafic tuff (Lehman et al., 2018). These beds also contain abundant “agatized” logs, some preserved rooted in growth position, and typically with unusual white, thoroughly silicified interiors (Lehman and Wheeler, 2001).

Wick (2023) described a highly silicified mafic tuff bed also intercalated with these facies in the Alto Shale, 2 m below the contact with the overlying Javelina Formation at Rooneys Place (Fig. 6, section 28). This likely represents a distal ash fall from one of the known eruptive centers or perhaps one not yet recognized. Also associated with the pyroclastic interval are local beds of very coarse-grained, matrix-supported boulder breccia composed of sandstone, mudstone, and carbonate clasts derived from underlying units in the Alto Shale and redeposited (e.g., Fig. 22, section 17). These beds are of uncertain origin but are interpreted as debris-flow and rock-fall (talus) deposits due perhaps to local mass-wasting events around the margins of the maar craters or otherwise related in some way to the pyroclastic eruptions. Collectively, these varied deposits combine to result in a distinctive stratigraphic interval (unit 3) within the upper part of the Alto Shale.

McKinney Springs Tongue (New) of Pen Formation

The marine shale interval intercalated within the Aguja Formation was originally recognized by Udden (1907, p. 46–47), who noted that the “thin stratum of sandstone” that marks the base of the formation (La Basa Sandstone Member of this report) is in places “overlain by 50 or more feet of clays of the same kind we find below it” (i.e., as in the underlying Pen Formation). Maxwell et al. (1967, p. 79, the “second unit” of their report) described a “fossiliferous marine clay,” 175–500 ft [50–150 m] thick that overlies the basal sandstone, also clearly referring to the marine shale unit herein identified as the McKinney Springs Tongue.

Although earlier authors included this marine shale interval within the Aguja Formation, it is herein formally removed from the Aguja Formation—these strata are contiguous with the main body of the Pen Formation along the eastern border of Big Bend National Park, exhibit similar sedimentary facies and marine invertebrate fauna, and are herein recognized formally as a member of the Pen Formation. In areas north of Dagger Flat (Fig. 13) and east of San Vicente (Fig. 13), where the La Basa Sandstone Member pinches out, the McKinney Springs Tongue is included within and mapped as the uppermost part of the main body of the Pen Formation (Turner et al., 2011) and likewise east of the park at Black Gap (St. John, 1965).

The name McKinney Springs Tongue was proposed informally by Lehman (1985a, p. 31) and subsequently used by Mosley (1992) to identify this marine shale interval and refers to McKinney Springs, a series of springs that empty into a tributary of Tornillo Creek along the northwestern flank of McKinney Hills (U.S. Geological Survey, 1971b). A type section for this unit is designated northwest of McKinney Springs (Fig. 5, section 21) where the McKinney Springs Tongue is easily accessed and well exposed and all four of the internal units described in the following paragraphs are present (Fig. 24).

The McKinney Springs Tongue is composed chiefly of weakly consolidated, recessively weathering shale and is both underlain and overlain by well-consolidated ridge-forming sandstone units (Fig. 16). As a result, its lower and upper contacts are physiographically distinct in most places. The lower contact, with either the La Basa Sandstone (in the eastern outcrop belt) or the Rattlesnake Mountain Sandstone (in the western outcrop belt), is typically abrupt and placed at the base of the lowermost marine shale bed. The upper contact with the Terlingua Creek Sandstone is instead gradational over a thickness of as much as 10 m and is placed at the base of the lowermost prominent (>1 m) thick sandstone bed.

In the eastern outcrop belt, where the McKinney Springs Tongue is thickest (~90 m), it consists of a lower interval of light gray calcareous clay shale ~15 m thick with two fossiliferous sandy marlstone beds near its base (unit 1; Fig. 25), above which is an interval of dark gray clay shale ~10 m thick with sandy coquinite and a thin phosphorite bed at the top (unit 2)—in many areas encased in cone-in-cone concretion (Figs. 16H and 25). The upper half of the McKinney Springs Tongue consists of light gray mud shale ~25 m thick with widely separated siltstone or fine-grained sandstone beds and several prominent septarian concretion horizons (unit 3; Fig. 25) and an upper interval ~5 m thick composed of rhythmically interbedded shale, allochthonous (i.e., transported and redeposited) lignite, and thin sandstone beds that are increasingly numerous and thicker upward to the contact with the Terlingua Creek Sandstone Member (unit 4; Fig. 16I).

The marlstone beds in unit 1 contain a diverse marine invertebrate fauna dominated by Exogyra ponderosa and inoceramid bivalves, with rare rudistid bivalves and ammonites. Unit 1 thins to the west and is absent from exposures in the central and western outcrop belts. In the western outcrop belt, unit 3 contains abundant “allochthonous” lignite beds and together with unit 4 comprises the entirety of the McKinney Springs Tongue in those areas, although the phosphorite bed (part of unit 2) is present in all areas.

The McKinney Springs Tongue thins to the west-northwest (Fig. 26). The gradational contacts above and below indicate that this is original depositional thinning and not a result of erosional truncation. In the eastern outcrop belt, the unit is 75–90 m thick. In the central outcrop belt, the McKinney Springs Tongue is thinner than it is to the east in most areas (20–30 m), the phosphorite bed (unit 2) is at or near the contact with the underlying Rattlesnake Mountain Sandstone, and the marlstone interval (unit 1) is absent. In the western outcrop belt, the McKinney Springs Tongue is very thin (10 m to <1 m), has nearly pinched out, and is absent in some areas (e.g., Steep Draw; Fig. 4, section 10) where the overlying Terlingua Creek Sandstone and underlying Rattlesnake Mountain Sandstone are in depositional contact with one another. The phosphorite bed (unit 2) remains recognizable even here (at the “Ten Bits Microsite” of Schubert et al., 2016) marking the contact between the two sandstone intervals.

The sedimentary facies represented in the Aguja Formation are similar to those recognized and described in detail for many Upper Cretaceous paralic and continental strata elsewhere in the Western Interior of North America (e.g., Eberth, 1990; Slingerland et al., 1996; Ambrose and Ayers, 2007; Li et al., 2015). As a result, detailed descriptions of these facies are unnecessary here, and the reader is referred to more thorough accounts cited below. Seven broad sedimentary facies associations are recognized in these strata, denoted herein as facies associations A through G, each of which is subdivided into an enumerated series of component facies (Table 1).

Facies Association A

Facies association A consists of laterally extensive, uniformly bedded shale and mudstone. These facies are interpreted as muddy marine shelf sediments accumulated in water of normal marine salinity and deposited with little current agitation below fair-weather wave base under inner to middle shelf water depths (Table 1). This facies association is found primarily in the McKinney Springs Tongue of the Pen Formation and in the main body of the Pen Formation 5–10 m below its contact with the Aguja Formation (Fig. 25). Local examples have been described by Lehman (1985b) and Mosley (1992). Similar Upper Cretaceous facies have been recognized throughout the Western Interior (e.g., Li and Schieber, 2018; LaCroix and Gingras, 2021). These deposits consist primarily of: facies A1, thinly bedded marlstone with whole shells of Exogyra, rudistid, and inoceramid bivalves and ammonites (interpreted as marine middle shelf deposits); facies A2, thin beds of sandy coquina composed of abraded mollusc shells, skeletal and pelloidal phosphate, and glauconite (interpreted as “sediment-starved” middle shelf deposits); and facies A3, dark gray silty calcareous shale with interspersed horizons of calcite septarian concretions (interpreted as marine inner shelf deposits). The marine invertebrate fauna found in this facies association is typical of inner- to middle-shelf facies in Campanian strata throughout the Gulf Coast and Western Interior (e.g., Kauffman, 1969, faunal assemblages 10 and 13).

Interrupting the typical muddy marine shelf deposits in a few places are thick (~10 cm to 1 m), very fine-grained sandstone beds interpreted as lower shoreface or inner shelf sand sheets—the seaward extensions of facies associations B and C (Fig. 12; see below). These beds exhibit a spectrum of primary and biogenic structures, from parallel-laminated and current ripple cross-laminated shelf turbidites to hummocky cross-stratified and oscillation-rippled tempestites (e.g., Dott and Bourgeois, 1982; Li et al., 2015) and storm-generated shell beds that preserve a diverse nearshore shallow marine fauna (e.g., Kidwell et al., 1986; Kauffman, 1969, faunal assemblages 27–30). Much of the marine invertebrate fauna known from intertonguing parts of the Pen and Aguja Formations has been obtained from these thin sandstone beds (Eley, 1938; Lehman, 1985a; Wick, 2021c).

Facies Association B

Facies association B is interpreted as a progradational marine deltaic succession comprising primarily prodelta, delta front, distributary mouth bar, and distributary channel deposits (Table 1). This facies association is found primarily in the La Basa Sandstone and Terlingua Creek Sandstone Members (Figs. 12 and 19). Local examples have been described in detail by Hopkins (1965), Lehman (1985a), Bohanan (1987), Schroeder (1988), and Anglen (2001). Similar Upper Cretaceous facies associations have been recognized throughout the Western Interior (e.g., Edwards et al., 2005; Allen and Johnson, 2011; Painter et al., 2013; Li et al., 2015). This facies association typically comprises a coarsening-upward succession and begins at the base with: facies B1, dark gray shale rhythmically alternating with thin beds of siltstone (interpreted as prodelta deposits); grading upward to facies B2, siltstone interstratified with a thickening-upward series of thin, laterally extensive, very fine-grained sandstone beds (interpreted as delta front deposits) with structures indicative of density current deposition (e.g., flute marks, groove marks, normal particle-size gradation, parallel lamination, and current ripple cross-lamination); overlain by facies B3, very thick beds (as thick as several meters) of fine-grained sandstone with parallel lamination, soft-sediment deformation (e.g., convolute lamination, dish structures, load casts), and/or bioturbation (interpreted as distributary mouth bar deposits); and facies B4, thick (decimeter-scale) lenticular beds of fine-grained sandstone with thick sets of broad subcritically climbing trough and tabular cross-stratification (interpreted as distributary channel deposits). The cross-stratification typically exhibits highly unimodal orientation, indicative of relatively low channel sinuosity. These facies (primarily facies B1 and B2) exhibit a gradational and intertonguing relationship with muddy marine shelf deposits of the Pen Formation (facies association A; Fig. 25) and are in general poorly fossiliferous. Thin beds of abraded mollusc-shell coquina are, however, interspersed sporadically within the succession (in facies B3 and B4), and large silicified logs are locally abundant (in upper part of facies B4). The marked lateral variation in thickness and internal structure within this facies association and lateral intertonguing with facies associations D and E suggest that these are a product of fluvial-dominated deltas (e.g., reviewed by Ainsworth et al., 2011).

Facies Association C

Facies association C is interpreted as a retrogradational shoreface succession (Table 1). This facies association is found primarily in the Rattlesnake Mountain Sandstone but also comprises a lesser local component of both the La Basa Sandstone and Terlingua Creek Sandstone (Figs. 12 and 17). Examples have been described by Lehman (1985a, 1985b) and Macon (1994). Similar Upper Cretaceous facies associations have been recognized throughout the Western Interior (e.g., Olsen et al., 1999; Sixsmith et al., 2008; Kieft et al., 2011). The typical facies succession is indicative of deepening upward, rests on a scoured ravinement surface cut on underlying strata, and begins at the base with: facies C1, thickly bedded sandy coquina composed of abraded mollusc shells in a matrix of fine-grained sandstone, interspersed with lenticular beds of whole cardid, inoceramid, or ostreid bivalves (interpreted as upper shoreface deposits); grading upward to facies C2, thickly bedded siltstone and muddy very fine-grained sandstone with horizons of whole ostreid bivalves commonly preserved articulated and in life position (interpreted as middle shoreface deposits); and facies C3, pervasively bioturbated very fine-grained sandstone, typically with boxwork Ophiomorpha (interpreted as lower shoreface deposits).

In most areas, this facies association is highly fossiliferous, in particular with dense biostromic accumulations of ostreid bivalves, many articulated in life position and representing primarily three species: Flemingostrea subspatulata, Flemingostrea pratti, and Crassostrea cf. C. cusseta (Fig. 17). In addition, there are extensive lenticular shell beds composed of cardid bivalves (Ethmocardium sp., Granocardium sp.), gastropods (Turritella sp.), bryozoans, crab carapace and pincer parts (Vega et al., 1997, 1998), and ammonites (Waggoner, 2006). Many of the Aguja Formation marine invertebrate species described by Eley (1938) and Lehman (1985a) were collected from these facies. In addition, a diverse fauna of sharks, sawfishes, rays, and bony fishes is also found (Schubert et al., 2016). The diverse marine fauna is one of the most distinctive features of these facies and is indicative of deposition under shallow wave-agitated brackish to marine water with active currents (e.g., Kauffman, 1969, faunal assemblages 1 and 19).

This typical facies succession is locally, however, underlain by or laterally gradational with thick beds of poorly fossiliferous cross-stratified sandstone interpreted as tidal inlet deposits (Fig. 17). These deposits compose facies C4, thickly bedded sandy conglomerate comprising mud galls and abraded mollusc shells, overlain by trough cross-stratified and thickly bedded low-angle cross-stratified fine-grained sandstone (Table 1). Cross-stratification in this facies exhibits highly varied orientations. Macon (1994) interpreted facies association C as recording the landward retreat of a barrier island complex and its subsequent submergence resulting in sandy inner shelf shoals.

Facies Association D

Facies association D is interpreted as an aggradational coastal marsh succession (Table 1; Fig. 14). This facies association predominates in the Abajo Shale and is a minor component in the lowermost part (unit 1) of the Alto Shale. Examples have been described by Lehman (1985a, 1985b), Record (1988), and Wagner (2001). Similar Upper Cretaceous facies associations have been recognized throughout the Western Interior (e.g., Cavaroc and Flores, 1985; Olsen et al., 1999; Ambrose and Ayers, 2007; Kieft et al., 2011). This facies association comprises a shallowing-upward succession interpreted as subaqueous bay-fill deposits, grading upward to subaerial poorly drained marsh deposits, and begins at the base with: facies D1, dark gray, organic-rich thinly bedded mudstone with concretion horizons of ferromagnesian oxides and carbonates; grading upward to facies D2, highly carbonaceous, crudely bedded, bioturbated mudstone; and capped by facies D3, lignite or coal. This succession is interpreted to culminate in a poorly drained wetland (hydric) soil with subsoil accumulation of pedogenic ferromagnesian oxides and carbonates (e.g., similar to those described by McCarthy, 2003; Flaig et al., 2013). Many of the lignite and coal beds that cap each succession exhibit a clastic texture suggesting that these are at least in part allochthonous organic deposits (i.e., transported and redeposited). Abundant carbonized leaves preserved in these facies are primarily those of herbaceous monocotyledonous angiosperms, suggesting vegetative cover was primarily marsh-like in habit (Record, 1988). However, also preserved are large fragments of wood with abundant Teredolites borings and rare silicified logs of conifer and dicot woods, indicating that arborescent vegetation was also present (Wheeler and Lehman, 2000, 2005). Baghai (1994, 1996) described a diverse angiosperm palynomorph assemblage also from these facies. Record and Lehman (1989) interpreted the variation in carbonate and organic carbon isotopic values and clay mineralogy of this facies as recording multiple cycles in the expansion and submergence of brackish and saltwater marshes with poorly drained, organic-rich soils (histosols).

Interrupting the typical components of facies association D are thin intervals (1–2 m) of inclined heterolithic strata interpreted as tidal creek deposits (Table 1; Fig. 14). These deposits comprise facies D4, thin beds of ostreid shell coquina overlain by gently inclined thinly bedded siltstone and very fine-grained sandstone regularly alternating with carbonaceous shale and allochthonous lignite. Also laterally gradational with typical facies association D are much thicker (5–20 m) lenticular bodies of sandstone as much as ~500 m in width interpreted as estuarine channel deposits (Table 1; Fig. 14, sections 7 and 13). These deposits comprise facies D5, an aggradational fining-upward succession resting on a scoured erosional surface with sandy conglomerate composed of mud galls and abraded mollusc shells, overlain by tabular cross-stratified fine-grained sandstone rhythmically alternating with current ripple cross-laminated siltstone or very fine-grained sandstone, and gradationally overlain by thinly interbedded carbonaceous mudstone and ripple cross-laminated siltstone with soft-sediment deformation. The rhythmic internal stratification within these sandstone intervals is laterally extensive, uniform, and regularly alternating between the thick cross-stratified or current ripple cross-laminated sandstone beds and thin siltstone beds. These deposits are interpreted as estuarine in nature based on the regular alternation in sediment grain size and structure (Fig. 14, sections 7 and 13) that indicates periodic fluctuation in current strength, the presence of both oscillation ripples and unidirectional current ripples, flaser bedding, and mixed fresh- and brackish-water fauna. The cross-stratification in this facies exhibits highly varied orientation (Fig. 14). The tidal creek and estuarine channel deposits preserve a diverse mixture of marine and terrestrial vertebrates; otherwise, facies association D is poorly fossiliferous, with small ostreid bivalves (e.g., Crassostrea sp., Ostrea sp.) one of the only common elements.

Facies Association E

Facies association E is interpreted as an aggradational coastal fluvial floodplain succession (Table 1; Fig. 21). This facies association predominates in the lower part (unit 1) of the Alto Shale. Local examples have been described by Lehman (1982, 1985a), Sankey (1998), and Lewis (2011). Similar facies associations are recognized elsewhere in the Western Interior (e.g., Chan and Pfaff, 1991; Corbett et al., 2011; Roberts, 2013). These facies occur in coarsening-upward successions, beginning at the base with: facies E1, olive-gray or yellowish gray organic-rich, crudely bedded, bioturbated mudstone with thick ferromagnesian carbonate concretion horizons (interpreted as poorly drained coastal flood basin deposits); grading upward to facies E2, thinly bedded current-rippled siltstone or very fine-grained sandstone, overlain by thick (decimeter-scale) beds of ripple cross-laminated or trough cross-bedded fine-grained sandstone with lenses of granule-pebble conglomerate composed of mud galls, reworked concretions, wood fragments, abraded bones, and teeth (interpreted as crevasse splay deposits). These successions culminate in immature poorly drained soils (inceptisols and entisols) with subsoil pedogenic accumulation of ferromagnesian carbonates.

Vertebrate fossils are relatively common, and most of the vertebrate fauna known from the Aguja Formation has been obtained from this facies association. Abundant petrified woods as well as leaves and palynomorphs found in this facies are indicative of forested coastal environments with varied coniferous and dicotyledonous trees and shrubs (Baghai, 1998; Wheeler and Lehman, 2000, 2005) as well as monocotyledonous trees and shrubs (Manchester et al., 2010).

Facies Association F

Facies association F is interpreted as an aggradational inland fluvial channel and floodplain succession (Table 1; Fig. 21). This facies association predominates in the middle and upper parts (units 2 and 4) of the Alto Shale. Local examples have been described by Lehman (1985a, 1989a) and Atchley et al. (2004). Similar fluvial facies associations are recognized elsewhere in the Western Interior (e.g., Lorenz, 1981; Rogers, 1998). The fluvial channel and overbank floodplain deposits typically occur in fining-upward successions beginning at the base with: facies F1, crudely bedded pebble conglomerate composed of reworked calcareous concretions, wood, and abraded bones, grading upward to thickly bedded cross-stratified fine- to medium-grained sandstone (interpreted as fluvial channel deposits), and gradationally overlain by facies F2, ripple cross-laminated siltstone or very fine-grained sandstone and crudely bedded multicolored mudstone with horizons of calcite concretions (interpreted as fluvial overbank deposits). In most areas, the overbank mudstone beds compose the bulk of the deposits and are light gray, yellowish gray, purple, or red with little interstitial organic matter, suggesting deposition took place under better-drained floodplain environments and soils formed under more highly oxidizing conditions than in facies associations D or E. The most mature of these soils, with pedogenic calcium carbonate concretion horizons, are interpreted as calcic vertisols and alfisols (Lehman, 1989a; Atchley et al., 2004). These facies are in general poorly fossiliferous; however, isolated vertebrate fossils (typically in facies F1) and petrified woods are locally abundant. The woods are from a wide variety of coniferous and dicotyledonous trees and shrubs and indicate that these floodplain environments hosted diverse evergreen woodlands and forests (Wheeler and Lehman, 2000, 2005, 2009).

Facies Association G

Facies association G is interpreted as an aggradational lacustrine and fluvio-deltaic succession that accumulated within and around volcanic maar craters formed during deposition of the upper Alto Shale (unit 3, referred to above as the “pyroclastic interval”). Examples of this facies association have been described by Kovschak (1973), Lehman (1985a), Breyer et al. (2007), and Befus et al. (2008). Similar Upper Cretaceous volcaniclastic and crater-lake facies have been recognized within and surrounding volcanic centers elsewhere in the Western Interior (e.g., LaBranche, 1999; Grande et al., 2022). Facies association G is interstratified and laterally gradational with basaltic pyroclastic deposits that consist of rhythmically bedded coarse- and fine-grained lapilli tuff with isolated large basaltic bombs and lithic blocks derived from underlying Aguja Formation strata (Table 1; Fig. 22).

Facies association G is composed of: Facies G1, thinly laminated limestone and calcareous siltstone with leaf impressions, charophyte algae, small bivalves, and gastropods, rhythmically interbedded with bioturbated thickly laminated calcareous siltstone or very fine-grained sandstone (interpreted as lacustrine deposits; Fig 22). This facies is locally underlain, overlain, or laterally gradational with facies G2, thickly bedded climbing ripple cross-laminated and parallel-laminated siltstone or very fine-grained sandstone, gradational with trough cross-bedded conglomeratic fine-grained sandstone (interpreted as lacustrine deltaic deposits).

Also included in this facies association is facies G3, thick lenticular boulder-cobble breccia beds comprising intra-formational clasts (Fig. 22). These are interpreted as debris flow and rock fall deposits accumulated around crater margins (e.g., similar to those described by Smith, 1986; Gernon et al., 2009). Local air-fall tuff beds deposited more distant from the maar craters are also interstratified with these facies (e.g., Wick, 2023). Facies association G is much more limited in distribution and more highly varied in character than other sedimentary facies in the Aguja Formation. Facies association G is poorly fossiliferous; thus far, only a few vertebrate fossils and petrified woods have been reported (primarily from facies G1).

La Basa Sandstone Member

In most areas, the La Basa Sandstone exhibits a single thickening- and coarsening-upward facies succession ~10 m thick interpreted as comprising prodelta, delta front, mouth bar, and distributary channel deposits in a progradational deltaic succession (facies B1–B4; Fig. 12). Local variations on this succession include multiple progradational prodelta–delta front cycles (e.g., Fig. 12, section 21) or multiple distributary channel aggradational cycles (e.g., Fig. 5, section 17), resulting in a total thickness of as much as 20 m. In other areas, distributary channel deposits are absent, the section is typically thinner (~5 m), and paleocurrent evidence for dominantly offshore (eastwardly) sediment transport is lacking. In these areas, one or more deepening-upward facies intervals capped by ostreid or inoceramid bivalve coquina beds are typically included, and the section is instead interpreted as a retrogradational shoreface succession (Fig. 12, facies C1–C2; e.g., section 4). There is substantial variation in thickness and continuity of the La Basa Sandstone along strike. This suggests that deposition took place along an embayed low-energy coastline rather than along a high-energy coastline where longshore redistribution of sand would have generated more laterally contiguous sand bodies. Paleocurrent measurements on cross-stratification in the deltaic distributary deposits (facies B4; Fig. 12) and paralic shoreface deposits (facies C1; Fig. 12) indicate that sediment transport was to the east or southeast (offshore). Oscillation ripples (Fig. 12) indicate that the shoreline trend was north-south throughout deposition of the La Basa Sandstone.

Abajo Shale Member

The Abajo Shale exhibits multiple aggradational coastal marsh successions capped by lignite or coal beds (facies D1–D3; Fig. 14). This facies succession predominates in units 1 and 3 of this member and is interpreted as recording episodic expansion and submergence of coastal marshes. The sandy heterolithic deposits that predominate instead in unit 2 are interpreted as due to lateral migration of tidal creeks (facies D4, Fig. 14). Thick sandstone lenses interbedded with unit 1 near the base of the Abajo Shale are interpreted as estuarine channel deposits (facies D5; Fig. 14). Cross-stratification, primarily in the estuarine channel sandstones (Fig. 14), indicate that sediment transport continued toward the east or southeast during deposition of the Abajo Shale. Oscillation ripples (Fig. 14) indicate that the shoreline continued to trend generally north-south. The stratigraphic sequence from the La Basa Sandstone through unit 2 of the Abajo Shale represents a conformable progradational facies succession that records the seaward advance of paralic environments followed by emergent terrestrial environments. Unit 3 facies in the Abajo Shale are, however, similar to those found in unit 1, but in this case locally intertonguing with retrogradational shoreface deposits in the overlying Rattlesnake Mountain Sandstone, and instead record the initial landward retreat and subsequent submergence of coastal environments.

Rattlesnake Mountain Sandstone Member

Lehman (1985a) and Macon (1994) described the sedimentary facies of the Rattlesnake Mountain Sandstone and interpreted these deposits as comprising a retrogradational shoreface succession (facies association C; Fig. 17). In many areas, there is a single deepening-upward succession from upper to lower shoreface deposits, while in other areas, multiple such successions are superimposed. Where the Rattlesnake Mountain Sandstone is thickest, it also includes thick, poorly fossiliferous, cross-stratified beds interpreted as tidal inlet deposits (facies C4; Fig. 17). Paleocurrent data primarily from this facies (Fig. 17) indicate more complex sediment transport paths than for the underlying La Basa Sandstone and Abajo Shale. Oscillation ripples indicate that the shoreline continued to trend generally north-south, but cross-stratification (Fig. 17) indicates prevailing sediment transport was primarily oblique (offshore) to parallel (longshore). Low-angle accretionary bedding and channel margins preserved in the tidal inlet deposits (Fig. 17) are also oriented generally to the southwest, oblique to the paleo-shoreline. Scours and gutter casts instead suggest erosion resulting primarily from offshore-directed currents.

In areas where the Rattlesnake Mountain Sandstone is relatively thin (~5 m or less; Fig. 17, section 16), the retrogradational shoreface succession rests on a scoured erosional surface mantled by mollusc-shell coquina. This interface is interpreted as a wave-ravinement surface resulting from barrier retreat during transgression (e.g., reviewed by Zecchin et al., 2019). The presence of multiple stacked retrogradational shoreface successions in many areas indicates that transgression was punctuated (sensu Cattaneo and Steel, 2003; Fig. 17, section 7). Where the Rattlesnake Mountain Sandstone is thick (as thick as ~35 m; Fig. 17, section 14), it includes deposits interpreted as tidal inlet deposits (facies C4) below and intertonguing with the deepening-upward retrogradational shoreface succession. The tidal inlet deposits are typically interbedded with underlying mudstone (unit 3) of the Abajo Shale, and so, at least in part, the uppermost Abajo Shale likely includes back-barrier sediments deposited during transgression. Bar accretion surfaces in the tidal inlet deposits dip primarily to the south-southwest, thus inlet migration and longshore sediment transport was likely to the south. Cross-stratification in tidal inlet sandstones suggests that sediment transport during ebb-tide flow and by longshore currents predominated over landward-directed flood-tide flow (Fig. 17).

McKinney Springs Tongue

The McKinney Springs Tongue accumulated in muddy marine shelf environments (facies association A) and in prodeltaic environments (facies association B) below fair-weather wave base (Fig. 25). The lower part of this member includes the most distal (offshore) deposits represented in the entire succession (facies A1); it is composed of fossiliferous calcareous shale and marlstone (unit 1) and records initial deepening to middle shelf environments and subsequent deposition of shell beds and phosphorite (unit 2) under sediment-starved conditions (facies A2). Mosley (1992) described a low-diversity benthic marine foraminiferal fauna obtained mostly from units 1 and 2 of this member at two sites (McKinney Hills, Fig. 5, section 21, and La Clocha, Fig. 6, section 30). The foraminiferal fauna indicates that initial deposition took place in hyposaline, normally oxygenated, marginal marine waters. The upper part of this member instead consists of poorly fossiliferous shale with thin beds of siltstone (unit 3; facies A3) grading upward to silty shale interbedded with fine-grained sandstone and allochthonous lignite (unit 4; facies B1–B2). These facies record deposition in nearshore (inner shelf) environments on a texturally graded shelf with relatively high sedimentation rates (e.g., see Swift, 1970) and represent a shallowing-upward progradational succession (Fig. 25).

Terlingua Creek Sandstone Member

The Terlingua Creek Sandstone accumulated predominantly in delta front, distributary mouth bar, and distributary channel environments (facies B2–B4; Fig. 19). In most areas, there is a single progradational deltaic succession, 5–10 m in thickness (Fig. 19). Where this succession is relatively thin, distributary channel deposits (facies B4) are poorly developed and mouth bar deposits (facies B3) show more bioturbation, presumably recording interdistributary or channel margin environments. In contrast, other areas show as many as three superimposed progradational successions separated by slight angular discordance, with an aggregate thickness of as much as 20 m. These areas with unusually thick deposits extend from Chisos Pen (Fig. 5, section 17) northeast to Grapevine Hills (Fig. 5, section 19) and Persimmon Gap (Fig. 5, section 25). Presumably, these areas, where the total thickness of the Terlingua Creek Sandstone is at its greatest, expose the axes of one or more major deltaic distributary systems.

At one such exposure northwest of Grapevine Hills (Fig. 27), outcrops are sufficient to reveal the partial geometry of a distributary channel system. At this outcrop (Fig. 27), the depositional dip of delta front foreset bedding is evident at the northeastern end of the exposure and records at least three episodes of deposition with syndepositional deformation, each separated by slight angular discordance. The dip (3°–14°, mean = 7°) of the delta front bedding here exceeds that typical for marine deltas, even without considering its reduction by compaction; for example, Asquith (1974) showed a dip from 1° to 3° in Cretaceous delta front deposits in Wyoming (western USA). The excessive dips observed at the Grapevine Hills exposure likely reflect significant syndepositional deformation (e.g., due to slumping or growth faulting) of the delta front. Delta front deposits (facies B2) here are overlain by and interbedded to the southwest with distributary mouth bar deposits (facies B3), and both facies include numerous beds of allochthonous lignite, indicating that distributary outflow contained abundant particulate organic matter. At the southwestern end of the exposure, distributary channel deposits (facies B4) predominate, distinguished on the basis of pervasive large-scale trough cross-stratification. Multiple episodes of channel aggradation are evident, with varied orientation of bar accretion surfaces within both mouth bar and channel deposits indicative of episodic distributary channel switching. The entire succession is overlain by aggradational bay fill and crevasse deposits (facies associations D and E) in the Alto Shale.

Paleocurrent data from cross-stratification, primarily in facies B4 (distributary channel deposits), are unimodal and indicate sediment transport continued to be toward the east-northeast during deposition of the Terlingua Creek Sandstone (Fig. 19). Oscillation ripples indicate that the shoreline continued to be oriented predominantly north-south, as it was during deposition of the underlying La Basa and Rattlesnake Mountain Sandstones.

Interrupting the typical progradational deltaic succession (facies association B) in a few places are highly fossiliferous sandstone lenses with shell beds comprising inoceramid, cardid, and ostreid bivalves. These deposits are interpreted to represent retrogradational shoreface deposits accumulated during relatively short-lived episodes of shoreline retreat (facies association C). One such deposit is particularly well exposed along the eastern end of River Road (Fig. 6, section 31; see Lehman, 2008, “River Road sandstone”).

At the River Road outcrop (Fig. 28), retrogradational shoreface deposits (facies C1) comprise a thin veneer (2–3 m thick) resting on a scoured ravinement surface that rises ~2° to the southwest (landward) relative to stratification in underlying deltaic deposits (facies B2–B3). Resting on this erosional surface are thick beds of sandy abraded mollusc-shell coquina alternating with beds of whole articulated inoceramid bivalves in sets gently dipping to the south-southwest (obliquely landward); these are interpreted as storm-generated overwash deposits. These beds are truncated above by an erosional surface overlain by a thin layer of winnowed lag gravel composed of abraded shark’s teeth, bones, phosphate granules, and oyster shells. Resting on this deflation lag are beds of ripple cross-laminated sandstone in form sets with large, straight-crested, compound dunes or sand waves (Fig. 28). These beds are attributed to fair-weather reworking of the underlying overwash deposits. Oscillation ripples here, like elsewhere in the Terlingua Creek Sandstone, indicate that the shoreline was oriented predominantly north-south (Fig. 28). Dip of bedding in the overwash deposits suggests that storm waves approached shore from the east-northeast, while cross-stratification and bedform crest orientation in the sand wave deposits indicate sediment transport under presumed fair-weather conditions was to the north (longshore).

Alto Shale Member

The Alto Shale accumulated in a succession of environments, from poorly drained coastal marsh and floodplain environments (unit 1) to inland well-drained fluvial floodplain environments (units 2 and 4) and local lacustrine and deltaic environments that developed in areas around the maar craters in the pyroclastic interval (unit 3). The typical facies association in unit 1 reflects deposition immediately landward of the shoreline in coastal marshes and swamps (facies association D) and in deltaic interdistributary wetlands that record a succession from bay fill to crevasse sub-delta and coastal floodplain environments (facies association E; Fig. 21). Multiple aggradational facies successions are recorded in unit 1, and this facies association is highly fossiliferous; most of the terrestrial vertebrate fauna and flora known from the Aguja Formation has been collected from these facies.

In contrast, the facies associations in units 2 and 4 reflect deposition in well-drained fluvial channel and floodplain environments situated some greater distance inland from the shore (Fig. 21; facies association F). These facies record multiple fluvial aggradational successions comprising alternating fluvial channel (facies F1) and overbank floodplain deposits (facies F2) with well-developed calcic paleosols. Lacustrine and fluvio-deltaic deposits (facies G1–G2) are interstratified with these inland fluvial floodplain sediments in the pyroclastic interval and spatially associated with the maar craters but extending well beyond their limits in some places (Fig. 22). This suggests that lacustrine deposition may have been initiated due to drainage impoundment by the pyroclastic eruptions.

Cross-stratification in unit 1 (Fig. 21; facies associations D and E) indicates roughly bimodal sediment transport paths, one directed north-northeast, the other south-southeast, with the resultant vector mean directed nearly due east (as with the underlying Terlingua Creek Sandstone). This may indicate that during deposition of the lower part of the Alto Shale, significant sediment deposition was a result of crevassing in interdistributary environments or, alternatively, recording an incipient change in flow direction more evident in overlying unit 2. Cross-stratification in units 2 and 4 (Fig. 21; facies association F) shows pronounced unimodal sediment transport to the southeast. This likely records a significant change in paleo-slope during deposition of the upper Alto Shale preceding and following the pyroclastic eruptions (see discussion in Depositional History section).

Paleocurrent data summarized in the preceding Depositional Environments section for each member of the Aguja Formation indicate that the sediment source area for these strata lay primarily to the west-southwest of the present outcrop belt during deposition of the La Basa Sandstone up through the lower part of the Alto Shale. Sediment transport shifted direction during deposition of the upper Alto Shale and may have included source areas to the northwest. Aguja sandstones are composed primarily of monocrystalline quartz, volcanic lithic grains, and plagioclase and range from feldspathic litharenite or lithic arkose to sublitharenite and subarkose (Fig. 29; Lehman, 1985a; Bohanan, 1987; Schroeder, 1988; Macon, 1994). Together with the generally eastward transport direction, the abundance of plagioclase and volcanic lithic grains suggests that a primary source terrain for Aguja sediment was in the Mesozoic Cordilleran magmatic arc systems of western Mexico (Lehman, 1986, 1991; McDowell et al., 2001; Kortyna et al., 2023). This interpretation is consistent with Aguja detrital zircon assemblages documented by Colliver (2017) and Kortyna et al. (2023) with multiple dominant age modes between ca. 83 and 99 Ma that generally correspond with magmatism in the Middle Cretaceous Cordilleran arc and younger volcanic centers in the western Sierra Madre of Mexico, southern Arizona, and New Mexico. A prominent age mode also at ca. 164 Ma in Aguja detrital zircon samples could reflect direct derivation from the Jurassic Cordilleran (Nazas) arc in the same region or reworking of younger sediments originally derived from that source (Colliver, 2017; Kortyna et al., 2023). Lesser detrital zircon age modes at ca. 255, 447, 1068, 1421, and 1677 Ma almost certainly reflect reworking of zircons from younger sedimentary rock sources, inasmuch as Paleozoic and Proterozoic basement terranes with those ages are unlikely to have been exposed during Campanian time.

The average Aguja sandstone composition is Qm58F14Lt28 (with the numerals indicating percentages of monocrystalline quartz [Qm], feldspar [F], and total lithic grains [Lt]; chert grains are included at the Lt pole; Fig. 29). The relatively broad compositional variation in part reflects variation in grain size and sedimentary facies; Aguja paralic and deltaic sandstones seldom exceed very fine to fine grain sizes (as large as 0.25 mm), whereas the fluvial sandstones are typically fine to medium grained (as large as 0.5 mm). However, compositional variation also generally follows the stratigraphic succession; the La Basa Sandstone is the most quartzose, whereas sandstones from the upper part of the Alto Shale have the highest lithic grain content (Fig. 29). This change in sand composition coincides with the shift to southeastwardly paleocurrent direction. Although volcanic rock fragments are the dominant lithic grains in nearly all Aguja sandstones, the La Basa Sandstone also has appreciable sedimentary lithic grains (primarily chert), while sandstones in the Alto Shale also have noticeable amounts of metamorphic rock fragments (primarily polycrystalline quartz grains with muscovite microlites) and microcline. These grains could reflect direct contributions from lesser source areas in older sedimentary or metamorphic terrains, for example in the Chihuahua tectonic belt of Texas and Chihuahua or the Van Horn uplift of Texas (e.g., Lehman, 1991), or reworking of younger sediments derived from those sources as may be the case for Aguja detrital zircons with Paleozoic and Precambrian age modes documented by Colliver (2017) and Kortyna et al. (2023).

This stratigraphic trend in sand composition could simply be a result of the generally progradational nature of the Aguja Formation, with lower Aguja sediments reflecting a greater transport distance from the sediment source area, and thus are finer grained and more quartzose, while upper Aguja sediments were deposited closer to the source area, and so are coarser grained and lithic rich. Such a change may have been due, for example, to eastward encroachment of fold-and-thrust sheets in the Chihuahua tectonic belt (Muehlberger, 1980). Alternatively, the compositional gradient could instead record a gradual transition from predominantly recycled orogen to magmatic arc provenance (sensu Dickinson et al., 1983; Fig. 29) during deposition of the Aguja Formation and a greater influence of volcanic source areas in the northern Sierra Madre and/or southern Arizona and New Mexico compared to earlier sedimentary source areas within the Chihuahua tectonic belt (e.g., Lehman, 1986, 1991). The overall similarity in grain types throughout the Aguja, with variation primarily in just their relative abundances, suggests that the latter of these two interpretations is more likely. Moreover, the great similarity in detrital zircon age modes among samples taken from the La Basa Sandstone, Terlingua Creek Sandstone, and upper Alto Shale indicates that the primary sediment source areas remained constant throughout deposition of the Aguja Formation (Colliver, 2017; Kortyna et al., 2023).

Marine Biostratigraphy

The paralic sandstone units in the Aguja Formation and the intertonguing Pen Formation have yielded several species of ammonites useful for biostratigraphic correlation (Fig. 30; Young, 1963; Waggoner, 2006; Cobban et al., 2008). The most common ammonites, however, represent relatively long-ranging species. For example, Baculites haresi is found in all of the Aguja marine sandstone units (Waggoner, 2006) and occurs elsewhere in upper Santonian (Desmoscaphites bassleri Zone) through middle Campanian (Baculites obtusus Zone) strata in the Western Interior Basin, as well as Gulf and Atlantic Coastal Plains (e.g., Waggoner, 2006). Species of Placenticeras also have long range zones, and are additionally a subject of taxonomic uncertainty (e.g., Waggoner, 2006). Placenticeras syrtale appears to extend from Santonian into lower Campanian strata, possibly as high as the Baculites sp. (weak flank ribs) Zone (see discussions by Kennedy et al., 2001; Cobban, 2016). The range of Placenticeras intercalare extends well into the upper Campanian (Baculites compressus Zone), as does Placenticeras meeki (into the Baculites reesidei Zone; e.g., Cobban, 2016).

Several ammonites with restricted ranges, however, allow for more precise biostratigraphic correlation (Fig. 30). The uppermost part of the main body of the Pen Formation immediately below the contact with the La Basa Sandstone is within the range zone of the Tethyan ammonite Menabites (Delawarella) d elawarensis and the Western Interior ammonite Scaphites hippocrepis III (Waggoner, 2006). The range zones for these two species overlap in a restricted interval that places the base of the Aguja Formation in the upper part of the lower Campanian (ca. 82–81 Ma; Gale et al., 2020).

In addition, Young (1963, p. 113) reported a single specimen of the rare ammonite Menabites (Delawarella) s abinalensis from “the concretion horizon in the Terlingua formation [sic] on Tornillo Creek.” Later, Maxwell et al. (1967, p. 77) described this concretion horizon near McKinney Hills as within the Pen Formation “immediately beneath the Aguja Formation.” In this area, the La Basa Sandstone Member is close to its pinch-out (Fig. 24), and although the precise level of the collection site is unknown and no additional specimens have been recovered, this occurrence indicates that near its easternmost limit, the La Basa Sandstone is within or above the M. (D.) sabinalensis Zone. Cobban et al. (2008) indicated that the M. (D.) sabinalensis Zone is correlative with the Baculites sp. (smooth) Zone in the Western Interior.

Ammonites recovered from the Rattlesnake Mountain Sandstone include Pachydiscus (Pachydiscus) paulsoni and Hoplitoplacenticeras (Ho plitoplacenticeras) minor. The occurrence of H. (H.) minor is significant as it is elsewhere known only from the Wolfe City Sand in northeastern Texas, where it occurs together with the Western Interior ammonite Baculites maclearni (Cobban and Kennedy, 1993). This suggests that the Rattlesnake Mountain Sandstone could extend into the lowermost part of the middle Campanian (ca. 81–80 Ma; Gale et al., 2020). P. (P.) paulsoni is elsewhere known to occur exclusively within the lower Campanian range zone of Menabites (Delawarella) delaware nsis (Cobban and Kennedy, 1992a, 1992b, 1993).

Ammonites found in the Terlingua Creek Sandstone also include Pachydiscus (Pachydiscus) p aulsoni, but in addition, Trachyscaphites spiniger (Cobban et al., 2008) and Hoplitoplacenticeras aff. H. plasticum (H. n. sp. of Waggoner, 2006; W. Cobban, 1990, personal commun.). Species of Hoplitoplacenticeras in this morphological group extend into the lower middle Campanian (e.g., Kennedy et al., 1973) and both T. spiniger and Hoplitoplacenticeras spp. occur within the Baculites maclearni Zone in the Wolfe City Sand in northeastern Texas (Cobban and Kennedy, 1993). Collectively, ammonites from the youngest marine strata in the Aguja Formation indicate correlation with strata no younger than the Baculites maclearni Zone (ca. 81–80 Ma; Gale et al., 2020).

Radioisotope Age Determinations

Colliver (2017) reported U-Pb age distributions for two detrital zircon samples obtained from the Aguja Formation (Fig. 30). The locality data indicate that one of these is from the La Basa Sandstone near Chisos Pen (sample 031910K R-01 of Colliver, 2017; section 17 of the present study) and the other is from the lower part of the Alto Shale near Croton Spring (sample 031710K R-06 of Colliver, 2017; section 18 of the present study). Colliver (2017) calculated the maximum depositional age (MDA) for the La Basa Sandstone sample as ca. 83 Ma (mean of youngest three or more ages [n = 11] that overlap within 2σ uncertainty = 82.8 ± 4.7 Ma, based on data provided by Colliver, 2017; algorithm of Dickinson and Gehrels, 2009). The youngest single zircon grain in this sample is given as 75.3 ± 6.4 Ma (Colliver, 2017, p. 205). The MDA for the sample from the lower Alto Shale is ca. 80 Ma (mean of youngest three or more ages [n = 6] that overlap within 2σ uncertainty = 80.6 ± 2.8 Ma; based on data provided by Colliver, 2017). The youngest single zircon grain in this sample is listed as 75.9 ± 3.5 Ma (Colliver, 2017, p. 210).

Kortyna et al. (2023) also documented several detrital zircon samples from the Aguja Formation. Two of these are from the Terlingua Creek Sandstone on Dawson Creek (samples 16TO-17 and 16TO-18; Fig. 30). These samples yielded youngest statistical population MDA determinations (shown with alpha and beta uncertainties; Coutts et al., 2019; Herriott et al., 2019) of 78.6 ± 0.9/2.0 Ma (n = 3) and 79.6 ± 0.6/1.9 Ma (n = 9). Another zircon sample (16TO-19; Fig. 30) from the uppermost Alto Shale Member on Dawson Creek, 20 m below the contact with the Javelina Formation, provided a youngest statistical population MDA determination of 72.0 ± 1.1/2.0 Ma (n = 3).

All of the MDA determinations, those obtained from the mean of the youngest ages overlapping with 2σ uncertainty and by the youngest single grain, broadly overlap within uncertainty the ages determined by ammonite biostratigraphy. These MDA values are also comparable to ages determined for pyroclastic deposits within the upper part of the Alto Shale (see following paragraphs).

Leslie et al. (2018) reported an Ar-Ar minimum age population of 76.28 ± 0.06 Ma for a detrital sanidine sample from the same site in the uppermost Alto Shale on Dawson Creek where the detrital zircon sample (16TO-19) analyzed by Kortyna et al. (2023) was obtained (Fig. 30; section 16 of this report; note that the base of the Javelina Formation determined in the present study is at the 42 m level in the section shown by Leslie et al., 2018). The detrital sanidine minimum age population from this sample is notably older than the detrital zircon youngest statistical population MDA determined by Kortyna et al. (2023) but comparable to the detrital zircon MDA determined for the lower Alto Shale by Colliver (2017) and in particular, with the youngest single zircon grain ages. The pyroclastic interval in the upper part of Alto Shale is either not present or not exposed on Dawson Creek, and so its stratigraphic position relative to the sanidine sample of Leslie et al. (2018) is uncertain.

Basaltic pyroclastic deposits in the upper part of the Alto Shale have been documented at two sites, one on the eastern side of the Rosillos Mountains (Breyer et al., 2007; section 22 of the present study) and the other at Peña Mountain (Befus et al., 2008; section 8 of the present study). Deposits at both sites are interpreted as the fill of collapsed maar craters that resulted from phreatomagmatic eruptions contemporaneous with deposition of the upper Alto Shale. U-Pb ages were determined for zircon xenocrysts obtained from basaltic bombs in the coarse lapilli tuff at both sites, and the ages of the youngest zircon xenocrysts are thought to record the ages of the eruptions (Fig. 30). In this way, the Peña Mountain deposit is dated as 76.9 ± 1.2 Ma (Befus et al., 2008) and the Rosillos Mountains deposit at 72.6 ± 1.5 Ma (Breyer et al., 2007). If not recording the actual age of the eruptions, these ages could be viewed instead as analogous to MDA assessments, and if so, are comparable to the detrital zircon and sanidine MDA determinations reviewed in the preceding paragraphs.

Magnetostratigraphy

Two paleomagnetic polarity sequences have been described that include parts of the Alto Shale Member of the Aguja Formation. Sankey and Gose (2001) documented several intervals of normal and reversed polarity in the lower 50 m of the Alto Shale near Talley Mountain (Fig. 30). They correlated this section to the lower part of chron C32, corresponding to the boundary between C32r and C32n. This polarity reversal occurs at 73.6 Ma (Gale et al., 2020); however, the MDA for the lower part of the Alto Shale based on detrital zircons and ammonite biozones in the underlying Terlingua Creek Sandstone suggest that the lower Alto Shale is significantly older. A closer correspondence is obtained if the polarity reversal recorded in this section is instead correlated with the transition from C33r to the lower part of C33n (79.9 Ma; Gale et al., 2020; the interpretation shown in Fig. 30). This interpretation would require, however, that a short interval of normal polarity occur within C33r. C33r is typically shown as an interval of reversed polarity over its entire length (e.g., Gale et al., 2020); however, several high-resolution studies have found a short interval of normal polarity in the latter part of C33r (see discussion by Ward et al., 2012). The interpretation favored here is that the prominent polarity reversal detected by Sankey and Gose (2001) in the lower Alto Shale is from C33r to C33n, and that the short interval of normal polarity is within C33r (Fig. 30).

Leslie et al. (2018) documented a magnetic polarity reversal in a section of the upper 40 m of the Alto Shale on Dawson Creek (Fig. 30). They correlated this polarity reversal with the transition from subchron C32n to C31r, although they interpreted the contact between the two polarity zones as unconformable. This correlation is compatible with the U-Pb age for the youngest pyroclastic deposits in the Alto Shale (72.6 ± 1.5 Ma), although given the uncertainty in this age determination, a correlation with the polarity reversal from C33n to C32r is possible as well. The former interpretation is shown in Figure 30; however, evidence for an unconformity at that level in the stratigraphic section is dubious, and it is instead more likely that, if present, an unconformity lies ~35 m above that, at the base of the Javelina Formation (Lehman et al., 2018).

Terrestrial Vertebrate Biostratigraphy

There are as many as four land vertebrate biozones represented in the Aguja Formation (Fig. 30; Table S1 [see footnote 1]). Paleobotanical collections are also generally consistent with recognition of these four biozones (Table S2). The two older biozones (zones I and II) are relatively fossiliferous intervals and so are fairly well documented, but the younger two (zones III and IV) are only sparsely fossiliferous and as a result are thus far poorly known. Even the better-documented intervals are difficult to correlate with better-known faunas elsewhere in North America because some taxa appear to be endemic to the Big Bend region and other more cosmopolitan taxa are not particularly diagnostic as to age.

Zone I is within the middle part (unit 2) of the Abajo Shale. Vertebrate fossils rarely occur in the lower and upper parts of this member, and so it is unclear whether these may represent faunas distinct from that found in the middle of the unit. The “Lowerverse local fauna” (University of Texas Vertebrate Paleontology Laboratory TxVP 45947) occurs within this zone; it consists of a diverse assemblage of small terrestrial and aquatic vertebrates, including a variety of fishes (Wick, 2021b; Wick and Brink, 2022), amphibians (Wick, 2020, 2021a), lizards (Wick and Shiller, 2020), small theropods and birds (Wick et al., 2015), and mammals (Brink, 2016). Lehman et al. (2019) reviewed the larger vertebrate fauna including turtles, crocodylians, and dinosaurs known from all localities other than the Lowerverse site. Prieto-Márquez et al. (2019) described the hadrosaurian dinosaur Aquilarhinus palimentus from this zone. Although the zone I fauna is very diverse, most of the taxa known from sufficiently diagnostic material either are endemic to the Abajo Shale or exhibit relatively long stratigraphic ranges through the Campanian. The zone I fauna includes, however, several mammals (e.g., Symmetrodontoides, Aquiladelp his, Albertatherium) known primarily from Aquilan North American Land Mammal Age (NALMA) faunas but also some (e.g., Turgidodon) known otherwise from Judithian faunas (Brink, 2016). The zone I fauna may therefore be intermediate in age between the older typical Aquilan (ca. 85–80 Ma) and younger Judithian (ca. 79–75 Ma; e.g., Brink, 2016) NALMAs.

Zone II occurs in the lower part (unit 1) of the Alto Shale. Vertebrate fossils occur most abundantly within this interval; consequently, much of the paleontological literature has focused on this zone. Most faunal lists presented previously for the Aguja Formation are exclusively for zone II (Langston et al., 1989; Rowe et al., 1992; Anglen and Lehman, 2000; Lehman and Busbey, 2007; Wick and Corrick, 2015). Lehman and Busbey (2007) referred to this interval as the “Deinosuchus zone” in recognition of one of its most distinctive taxa; however, a few specimens from both the Abajo Shale (zone I) and upper Alto Shale (zone III) indicate that the stratigraphic range of Deinosuchus may extend below and above this zone (Lehman et al., 2018). Nevertheless, all of the significant specimens of the giant alligatoroid crocodylian Deinosuchus riograndens is were obtained from this interval (American Museum of Natural History AMNH 3073: Colbert and Bird, 1954; TxVP 43620; see recent review by Cossette and Brochu, 2020). The “Terlingua local fauna” (TxVP 43057, Oklahoma Museum of Natural History OMNH V58; Rowe et al., 1992) is within the zone II interval; it includes a diverse assemblage of small aquatic and terrestrial vertebrates, including fishes, lizards (Nydam et al., 2013), and mammals (Weil, 1992; Cifelli, 1994). The Talley Mountain micro-vertebrate sites (Sankey, 2001; Sankey and Gose, 2001) and the “Gaddis site” (Montgomery and Clark, 2016) are also within zone II. The bothremydid turtle Chupacabrachelys complexus (Lehman and Wick, 2010) occurs in this zone, and most of the other turtles that Tomlinson (1997) described for the Aguja Formation are from this interval as well. The WPA (Works Progress Administration) quarry sites described by Lehman (1982) and Davies (1983) are typical of the bone-bed accumulations in zone II. The ceratopsian Agujaceratops mariscalensis (Lehman, 1989b, 2007) and hadrosaurian dinosaurs described by Davies (1983), Davies and Lehman (1989), and Wagner (2001) are typical dinosaurs from zone II. Carpenter (1990; see also Bakker, 1988) and West (2020) described the ankylosaurian dinosaurs from this zone. The pachycephalosaur Texacephale langstoni (Longrich et al., 2010; see also Lehman, 2010; Jasinski and Sullivan, 2011) as well as small theropods, including the caenagnathid Leptorhynchos gaddisi and several ornithomimids (Longrich et al., 2013), are from this zone. Tyrannosaurs from zone II were described by Lehman and Wick (2013). Some taxa represented in zone II are held in common with typical faunas of the Judithian NALMA (e.g., the multituberculate Cimolomys clarki; Rowe et al., 1992); however, others (Paleomolops langstoni; Cifelli, 1994) in the Terlingua local fauna suggest that zone II may be somewhat older than typical Judithian faunas.

Zone III is within the middle part of the Alto Shale, below the pyroclastic interval. Although vertebrate fossils are relatively common and well preserved in this interval, only a few specimens have thus far been described. The ceratopsian Agujaceratops mavericus (Lehman et al., 2017) and the hadrosaurs Angulomastacator daviesi (Wagner and Lehman, 2009) and Malefica deckerti (Prieto-Márquez and Wagner, 2023) are all known only from zone III. Sankey (2006, 2008, 2010; see also Welsh and Sankey, 2008) described several micro-vertebrate sites and eggshell collections from this zone. Although additional prolific micro-vertebrate sites are also within this interval (e.g., TxVP 42880), their faunas have yet to be described. At present, other than endemic taxa, too little is known of the zone III fauna to assess its likely correlation with the NALMA zonation on biostratigraphic grounds.

Zone IV is within the uppermost Alto Shale, within and above the pyroclastic interval. Here also, vertebrate fossils are relatively common, but very few sites or specimens have thus far been adequately documented. A caenagnathid theropod (Wick et al., 2024) as well as a variety of turtles, crocodylians, and hadrosaurian and ceratopsian dinosaurs have been collected but are yet to be described. The “Running Lizard” site (Louisiana State University Museum of Natural History locality LSU VL-113) described by Standhardt (1986; see also Leslie et al., 2018) is within this interval. Standhardt (1986) suggested a Maastrichtian (Lancian NALMA) age for the Running Lizard fauna, based on a fragmentary tooth referred to Alphadon marshi; however, Leslie et al. (2018) determined that the specimen was not diagnostic. No other taxa of biostratigraphic utility have yet been described from the zone IV fauna. As a result, correlation of this fauna with the NALMA zonation is uncertain. The age of the youngest pyroclastic deposits in this interval (ca. 72 Ma) suggests, however, that the zone IV fauna may be much younger than those found in the underlying Alto Shale.

Age constraints for the ammonite biozones and for the magnetic polarity chrons (Gale et al., 2020; see also Scott, 2014) as well as U-Pb ages for the Aguja Formation pyroclastic deposits provide seven age calibration points, six of which can be linked to specific stratigraphic levels in the Aguja Formation. The composite section measured from Rattlesnake Mountain (Fig. 3, section 7) to Peña Mountain (Fig. 3, section 8) includes six of these tie points and provides the basis for a rudimentary age model for the Aguja Formation (Fig. 31).

The entire lower part of the Aguja, from the La Basa Sandstone up through and including the Terlingua Creek Sandstone, was deposited during a relatively brief time interval between the first occurrence datum for Scaphites hippocrepis III (81.5 Ma) and the last occurrence datum for Baculites maclearni (80.2 Ma), a time span of 1.3 m.y. Of course, deposition of the La Basa Sandstone may have begun sometime within or near the end of the range zone of S. hippocrepis III rather than at its first occurrence, and likewise deposition of the Terlingua Creek Sandstone may have concluded sometime before the last occurrence of B. maclearni, so this time span is a maximum estimate. Regardless, this part of the Aguja section is 115 m in thickness, for which the long-term (compacted) sediment accumulation rate is therefore at least 89 m/m.y. Aguja vertebrate biozone I occurs within this interval, and if the Aquilan-Judithian NALMA boundary is placed at ca. 79 Ma (e.g., Ramezani et al., 2022), then zone I is within the latter part of the Aquilan.

The end of polarity subchron C33r (78.6 Ma) occurs 30 m above the top of the Terlingua Creek Sandstone at Talley Mountain. Assuming the same is true at Rattlesnake Mountain, then the lower 30 m of the Alto Shale was deposited in a time span of ~1.6 m.y. (80.2–78.6 Ma). The long-term (compacted) sediment accumulation rate for the lower Alto Shale was therefore only ~19 m/m.y., significantly lower than for the lower wedge of the Aguja. Vertebrate biozone II occurs within this interval, and if the base of the Judithian NALMA is placed at ca. 79 Ma (e.g., Ramezani et al., 2022), then zone II straddles the Aquilan-Judithian boundary.

Most of the upper Alto Shale (130 m) lies below the Peña Mountain pyroclastic deposits (ca. 76.9 Ma; Fig. 31). Assuming that the U-Pb age determination for these deposits records the actual age of eruption and is not simply a maximum age, then the pre-pyroclastic interval of the upper Alto Shale was deposited from 78.6 to 76.9 Ma, a time span of 1.7 m.y. Aguja vertebrate biozone III occurs in this interval and therefore within the Judithian NALMA (e.g., Ramezani et al., 2022). The long-term (compacted) sediment accumulation rate for the pre-pyroclastic interval of the upper Alto Shale was ~77 m/m.y.; however, due to great uncertainty in the U-Pb age for the pyroclastic deposit at Peña Mountain, this accumulation rate may be highly inaccurate. Regardless, this interval records a shift in paleocurrent orientation from east-northeast to southeast, coarsening of sediment particle size, and a transition to cyclic fluvial aggradational deposition. The upper Alto Shale may be separated from the lower Alto Shale by a disconformity (sequence boundary), but if so, the precise position of the disconformity is obscure and the duration of any hiatus represented there is unknown.

The Rosillos Mountains pyroclastic deposits (ca. 72.6 Ma; Fig. 31) are the youngest of the pyroclastic deposits thus far known in the Aguja Formation, although given the confidence intervals, they nearly overlap in age with those found at Peña Mountain. It is not possible to determine the precise stratigraphic level of the Rosillos Mountains pyroclastics within the Aguja Formation. About 6 km to the southwest of the Rosillos Mountains exposure, however, at Grapevine Hills (Fig. 5, section 20), there is an interval of rhythmically bedded lacustrine deposits in the uppermost part of Alto Shale that is likely related to and correlative with the nearby pyroclastic deposits. Elsewhere in the few places (e.g., Peña Mountain: Fig. 3, section 8; Rooneys Place: Fig. 6, section 28) where pyroclastic deposits are clearly intercalated within the Aguja section, they also lie in the uppermost Alto Shale within 2–45 m of the contact with the overlying Javelina Formation. Moreover, if the peculiar blue and green chlorite-rich mudstone interval within the uppermost Alto Shale comprises floodplain sediment correlative with the pyroclastic deposits (see description of Alto Shale in Lithostratigraphy of the Aguja Formation section), then this interval also occurs in all areas near the top of the Alto Shale. Thus, it seems likely that everywhere in the Aguja Formation, the pyroclastic interval comprises a zone within 45 m of the top of the Alto Shale.

Again assuming that the U-Pb age determination for the Rosillos Mountains pyroclastic deposits (ca. 72.6 Ma) records the age of the eruption and not a maximum age, and if this age approximates the end of this regional episode of magmatic activity, then the entire pyroclastic interval within the Alto Shale may correspond with a significant hiatus in sedimentation or a stratigraphically condensed interval exceeding ~4 m.y. (between the oldest and youngest of the known eruptions). This period of magmatic activity may have led to local damming of stream drainages that resulted in lacustrine sedimentation associated with the pyroclastic deposits and diversion or reorganization of drainage such that the primary locus of sedimentation was routed elsewhere outside of the Big Bend area during this time.

The post-pyroclastic interval in the uppermost Alto Shale post-dates the youngest of the pyroclastic deposits (ca. 72.6 Ma) and records a period of renewed sedimentation at the end of Aguja deposition. This interval is highly variable in thickness, reaching a maximum of 35 m (Cow Heaven Mountain; Fig. 6, section 26) to 45 m (Peña Mountain; Fig. 3, section 8), but appears to be entirely absent is some areas (e.g., Sierra Aguja, section 3; Desert Mountain Overlook, section 1; Fig. 3). This variability in thickness may reflect varied depth of post-Aguja erosion if the base of the overlying Javelina Formation is unconformable with the Alto Shale (e.g., as suggested by Lehman et al., 2018).

The ca. 72 Ma MDA determination based on detrital zircons from 20 m below the top of the Alto Shale on Dawson Creek (Kortyna et al., 2023), correlation of the magnetic polarity reversal 35 m from the top of the Alto Shale with the end of polarity subchron C32n (71.6 Ma), and the estimated age of the base of the Javelina Formation at ca. 69.5 Ma (Leslie et al., 2018) collectively indicate that the post-pyroclastic interval of the Aguja likely extends into early Maastrichtian time (Campanian-Maastrichtian boundary at 72.1 Ma; Gale et al., 2020). If the same circumstances apply to the section nearby at Peña Mountain (Fig. 3), where the post-pyroclastic interval is 45 m thick, then as much as the uppermost 35 m of the Alto Shale is much younger than the underlying pyroclastic deposits—extending in age from ca. 72.6 to 69.5 Ma. Aguja vertebrate biozone IV occurs within the post-pyroclastic interval, and if the Judithian-Edmontonian NALMA boundary is placed at ca. 74 Ma (e.g., Ramezani et al., 2022), then zone IV is of Edmontonian age.

Initial deposition of the Aguja Formation resulted in a progradational deltaic facies succession that includes the upper part of Pen Formation through the La Basa Sandstone Member and lower part of the Abajo Shale Member (Fig. 32A). The top of this depositional interval is placed at the top of the La Basa Sandstone or, where the La Basa Sandstone is overlain by the Abajo Shale, at the top of unit 2 within the Abajo (e.g., Fig. 12). This succession of deposits likely comprises the highstand systems tract of the Zuni 3.5 sequence (Haq et al., 1988; or “KCa2” of Haq, 2014), and the sequence boundary lies at the top of this interval. The sequence boundary is placed at the top of unit 2 in the Abajo Shale because this unit has attributes indicative of a terrestrial stratigraphic condensed section (e.g., closely spaced paleosols, abundant vertebrate bones) compatible with the late highstand systems tract, and the overlying Abajo Shale (unit 3) locally intertongues with retrogradational shoreface deposits in the Rattlesnake Mountain Sandstone, suggesting that it was instead, at least in part, deposited in back-barrier environments during the subsequent transgression.

During this initial Aguja depositional interval, sediment transport was to the east-northeast and the shoreline trended north-south (e.g., Figs. 12 and 14). At its furthest extent, the sandy shoreface (represented by the pinch-out of the La Basa Sandstone) extended irregularly to or near the eastern edge of Big Bend National Park and likely into nearby neighboring areas of Coahuila, Mexico (vicinity of Pico Etero and at La Unión; Fig. 2); however, the actual coastline at that time was likely closer to the easternmost limit of the Abajo Shale, which extends through the central part of the park (Fig. 15). This initial progradational event was relatively short lived and took place over a period of ~400,000 years (from ca. 81.5 to 81.1 Ma; first occurrence [FO] Scaphites hippocrepis III to last occurrence [LO] Baculites sp. [smooth]).

The second phase of deposition resulted in a retrogradational shoreface and marine shelf succession that includes the uppermost part of the Abajo Shale, the overlying Rattlesnake Mountain Sandstone, and the lowermost part of the McKinney Springs Tongue (Pen Formation) up to the phosphate granule conglomerate bed interpreted to represent the maximum flooding surface (Fig. 32B). These deposits likely comprise the transgressive systems tract of the Zuni 4.1 sequence (Haq et al., 1988; or “KCa3” of Haq, 2014; or “Cam3” megacycle of Gale et al., 2020). Local intertonguing of the uppermost Abajo Shale and lower Rattlesnake Mountain Sandstone indicates that there were brief interludes of shoreline stasis during this punctuated transgression. The shoreline continued to trend north-south, but sediment transport paths were more varied with landward, longshore, as well as offshore-directed current systems responsible for deposition. During peak transgression, the maximum landward incursion of the shoreline extended to at or just beyond the present western limit of the Aguja outcrop belt (Fig. 26). Although the McKinney Springs Tongue is very thin (10 m or less) along the entire western edge of the outcrop, its pinch-out is preserved only at the far northwesternmost edge of the exposure (e.g., Steep Draw; Fig. 4, section 10). The transgressive event recorded in this interval was relatively short lived and took place over the course of ~500,000 years (from ca. 81.1 to 80.6 Ma; LO B. sp. [smooth] to FO Baculites maclearni).

The third phase of Aguja deposition produced a progradational deltaic succession including the upper part of the McKinney Springs Tongue, the Terlingua Creek Sandstone, and the lower part (unit 1) of the Alto Shale (Fig. 32C). This succession thickens to the east-southeast and is interpreted as the highstand systems tract of the Zuni 4.1 sequence (Haq et al., 1988). During this phase of deposition, the shoreline continued to trend north-south and sediment transport was dominantly directed offshore to the east (Fig. 19). Intertonguing of shelf deposits in the upper McKinney Springs Tongue with sandy shoreface and coquina beds in the Terlingua Creek Sandstone indicate that there were short-lived episodes of shoreline retreat that interrupted this otherwise progradational episode (e.g., Fig. 28, section 31). Voluminous prodelta, delta front, and deltaic deposits of the upper McKinney Springs Tongue and Terlingua Creek Sandstone were deposited relatively rapidly as the shoreline swept across the Big Bend region in ~400,000 years (from ca. 80.6 to 80.2 Ma; range zone of B. maclearni).

The lower part of the Alto Shale (unit 1) comprises aggradational bay fill, crevasse sub-delta, and coastal floodplain deposits that constructed the delta platform landward of the Terlingua Creek Sandstone shoreline during this depositional interval. These deposits generally thicken to the southeast (Fig. 32C). Sediment transport directions were varied but directed primarily to either the northeast or southeast. The bulk of these deposits accumulated slowly, over a period of ~1.6 million years (80.2–78.6 Ma; LO of B. maclearni to end of magnetic polarity subchron C33r). These strata likely comprise the stratigraphically condensed top-lapping interval of the highstand systems tract for the Upper Zuni A-4.1 sequence (Haq et al., 1988), and the upper sequence boundary presumably lies at the top of this succession. The fourth phase of deposition of the Aguja Formation resulted in a series of aggradational fluvial channel and floodplain successions that formed the upper part of the Alto Shale (units 2–4; Fig. 32D). These deposits consist primarily of well-drained fluvial floodplain sediments. The fluvial channel sandstone beds are coarser grained than in the underlying part of the Alto Shale, and sediment transport direction shifted to the southeast during deposition. These deposits are much thicker in the central part of the park area and thin to the southwest and northeast. This suggests that the central part of the Big Bend region (what would later become the Tornillo Basin; Lehman, 1991) had begun to subside at a greater rate than the surrounding basin margins. There are none of the marked facies changes, however, that might be expected if these marginal regions had experienced lower subsidence rates (e.g., greater proportion of fluvial channel deposits or more mature alluvial paleosols). Alternatively, it is possible that thinning of the upper Alto Shale along the basin margins could reflect instead primarily post-depositional erosion and that a significant disconformity exists at the base of the overlying Javelina Formation. For example, southwestern sections of the Alto Shale (Desert Mountain Overlook, section 1; Sierra Aguja, section 3; Fig. 3) lack units 3 and 4 of that member and appear to be truncated within the lower part of unit 2. In contrast, along the northeastern margin of the basin, even the uppermost parts of the Alto Shale (units 3 and 4, the pyroclastic and post-pyroclastic intervals) are present, and so thinning here must instead reflect original depositional thinning.

Although in many areas the upper Alto Shale can be separated into deposits accumulated prior to, during, and following the episode of pyroclastic eruptions, the distinction between these intervals is not everywhere clear or mappable. It is likely that parts of at least two depositional sequences are represented in this succession—one deposited over a period of ~1.7 million years that preceded the pyroclastic eruptions (unit 2; from 78.6 to 76.9 Ma), and one following the pyroclastic eruptions also deposited over a period of ~1.7 million years (unit 4; ca. 72.6–69.5 Ma). The two sequences appear to be separated by a significant hiatus in sedimentation or a condensed stratigraphic section recorded by the pyroclastic interval itself (unit 3).

The time interval recorded in deposits of the Aguja Formation (ca. 82–72 Ma) coincides with a dramatic change in the southern aperture of the Cretaceous Western Interior Seaway in North America. For example, in the most recent and widely known paleogeographic reconstructions of the Western Interior created by Blakey (2014; see also Blakey and Ranney, 2018; Deep Time Maps, 2021), the southern entrance to the seaway through the Gulf of Mexico is depicted as completely open at 82 Ma but essentially closed off and disconnected from the northern part of the seaway by 72 Ma (Fig. 33). The depositional history of the Aguja Formation provides a basis for interpreting likely processes responsible for this transformation.

The constriction and closure of the southern aperture of the seaway could have been due at least in part to a global drop in sea level (resulting in a forced regression) that began during deposition of the Terlingua Creek Sandstone (ca. 80 Ma). The eustatic sea level curve of Haq et al. (1988) indicates that the sea level lowstand at the end of their Upper Zuni A-3.5 sequence (or the “KCa3” cycle boundary of Haq, 2014) is dated at ca. 80 Ma and is interpreted to have been one of high amplitude, resulting in a major type 1 sequence boundary accompanied by lowstand fan deposits. However, sea level subsequently rose and remained at nearly the same level as it had been prior to that regression, and later Campanian sea level falls are believed to have been less significant than those that ensued during Maastrichtian time (Haq, 2014). Thus, it seems unlikely that a global fall in sea level alone could have been responsible for closing the southern entrance to the Western Interior Seaway.

The timing of closure of the southern entrance to the seaway also overlaps, however, with a significant but poorly understood regional episode of volcanism that resulted in the Balcones igneous province—a belt of magmatic activity that extended from the Big Bend region eastward into southern and central Texas and contemporaneous in part with the Arkansas igneous province in Louisiana, Arkansas, and Mississippi, southeastern USA (e.g., reviewed by Ewing, 2009). The Balcones igneous province comprises some 200 small volcanic centers that erupted in two pulses at ca. 80 Ma and 72 Ma. These pulses coincided with uplift and erosion focused locally in the vicinity of the most concentrated volcanic activity (the Uvalde uplift) and with broad regional uplift of the San Marcos arch through south-central Texas (Ewing, 2016).

The pyroclastic interval in the Aguja Formation is thought to be the westernmost expression of Balcones igneous province volcanism (Befus et al., 2008) and coincides here also with uplift and a marked change in sedimentation. The shift to southeastward sediment transport recorded in the upper Alto Shale Member of the Aguja at ca. 79 Ma appears to have preceded the outbreak of local volcanism in Big Bend and likely records incipient emergence of the Laramide-age Marathon uplift, which extends southward into Mexico where it is variously referred to as the Marathon-Burro uplift or Burro-Peyotes arch (e.g., Ewing, 2016). This uplift appears to have blocked the earlier eastward stream flow through the Big Bend region. The redirected southeastward stream flow evident in the upper Alto Shale persisted through Maastrichtian time during deposition of the Javelina Formation and continued well into Eocene time as the Marathon-Burro uplift was progressively unroofed (Lehman et al., 2018). Thus, rather than eustatic sea level fall, the closure of the southern aperture of the Western Interior Seaway has likely more to do with regional uplift associated with the onset of Laramide deformation and with the outbreak of volcanism in the Balcones igneous province, which may together reflect the same tectonic mechanism.

The youngest episode of Balcones province volcanism (ca. 72 Ma) was contemporaneous with deposition of the paralic San Miguel Formation in the Maverick Basin of southern Texas and adjacent Coahuila (Weise, 1979); these deposits record the initial encroachment of the strandline into the Rio Grande embayment of the Gulf of Mexico. The San Miguel sandstones are dominated by volcanic lithic grains, indicating that these sediments were sourced in part, or in large part, from the nearby volcanic edifices, and sand distribution patterns suggest sediment was supplied to the San Miguel shoreline from the northwest and north (Weise, 1979).

Hence, the western shoreline of the Western Interior Seaway must have shifted from the Big Bend region some 200 km eastward into south Texas and eastern Mexico sometime during the 8 m.y. time span following deposition of the Terlingua Creek Sandstone (ca. 80 Ma) to initial deposition of the San Miguel Formation (ca. 72 Ma, although some contend that the San Miguel may be younger; e.g., Mahmoud, 2004). Paleogeographic reconstructions produced by Blakey (2014) indicate that rivers in the Big Bend region throughout this time interval continued to flow northeastwardly to the Western Interior Seaway, with their mouths in the Texas Panhandle. Such a paleo-flow direction is contradicted by the southeastward paleocurrent data reported herein. In contrast, Ewing (2016) hypothesized that the deltaic and strandline deposits of the San Miguel Formation and the overlying Olmos Formation in southern Texas were actually fed by Aguja trunk streams (the “Bigfoot River” and “Olmos River”) flowing eastward through the Big Bend region into the Rio Grande embayment.

The marked eastward thinning of the Aguja Formation and paleocurrent orientations indicate, however, that the hypothesized paleo-river systems were unlikely to have been part of the same river system that deposited the upper Aguja or Javelina Formations. Instead, the southeastward-flowing river system initiated in the Big Bend area during deposition of the upper Alto Shale probably continued on its course to the southeast and emptied into the Sabinas Basin in eastern Coahuila 100 km further to the south and was separated from the Rio Grande embayment by the Burro-Peyotes arch (e.g., Ewing, 2016). The hypothesized paleo-rivers may instead have had their headwaters in the Marathon-Burro uplift and/or more locally in the Uvalde uplift and San Marcos arch. Regardless, the outbreak of volcanism in the Balcones igneous province between 80 and 72 Ma coincided with widespread uplift across central and southern Texas and northeastern Mexico, rerouting stream flow and shifting the strandline from the Big Bend region to the Rio Grande embayment. It seems likely that this tectonic and magmatic event was largely responsible for closing the southern entrance to the Western Interior Seaway.

The constriction and closure of the southern aperture of the Western Interior Seaway during Campanian time may explain why the Aguja Formation marine vertebrate fauna exhibits much greater affinity with marine vertebrates of the Gulf and Atlantic Coasts than with those of the Western Interior (Lehman and Tomlinson, 2004; Schubert et al., 2016). Likewise, the emergence of a southern land connection between western and eastern North America may explain why the Normapolles palynomorph flora of the Aguja Formation has more in common with the flora of Appalachia than with Laramidia (Baghai, 1994, 1996). This land bridge may also have allowed for terrestrial faunal interchange between the two North American biotic provinces. For example, ornithomimosaurian and ceratopsid dinosaurs might have followed such a migration route from Laramidia (Farke and Phillips, 2017; Tsogtbaatar et al., 2022). Several hadrosaurian dinosaurs found in the Aguja Formation are more closely related phylogenetically to those known from Appalachia (Prieto-Márquez and Wagner, 2023). Thus, the purported biotic provincialism between northern and southern parts of Laramidia may reflect, at least in part, the influence of a southern connection with Appalachia (e.g., Carpenter, 1982; Lehman, 1997; Gates et al., 2010, 2012).

The Aguja Formation preserves a remnant of the deposits that accumulated along the western shoreline of the Cretaceous Western Interior Seaway in the Big Bend region of Texas during Campanian time. The formation is herein subdivided into a series of formal members that provides a framework for detailed correlation, and a lectostratotype section is designated on Dawson Creek. Mappability of the member subdivisions is demonstrated for key outcrops, and the basic sedimentary facies are described and illustrated for each unit. Biostratigraphic correlation and available geochronologic data are used to provide an age model for the formation. Four terrestrial vertebrate biozones (zones I to IV) are proposed.

An initial progradational deltaic succession is recorded by deposition of the La Basa Sandstone Member and lower part of the Abajo Shale Member (which includes zone I). This was followed by a retrogradational succession that includes the upper part of the Abajo Shale, the overlying Rattlesnake Mountain Sandstone Member, and the lower part of the McKinney Springs Tongue of the Pen Formation. Deposition of this entire lower wedge of the Aguja took place relatively rapidly, from ca. 81 to 80 Ma. A third phase of deposition comprises a progradational deltaic succession that includes the upper part of the McKinney Springs Tongue, the Terlingua Creek Sandstone Member, and the lower part of the Alto Shale Member (which includes zone II). This succession records eastward migration of the strandline, the final withdrawal of the Western Interior Seaway from the Big Bend region, and took place from ca. 80 to 78 Ma.

The fourth phase of deposition comprises a series of fluvial aggradational successions that formed the upper part of the Alto Shale Member from ca. 78 to 77 Ma. Aguja vertebrate biozone III occurs within this interval. This phase of deposition coincides with a redirection of stream flow to the southeast, and the resulting deposits are much thicker in the central part of the Big Bend region, thin both to the southwest and northeast, and likely record initial subsidence in the Laramide Tornillo Basin. The latter part of this depositional phase was contemporaneous with an outbreak of basaltic pyroclastic eruptions that took place from ca. 77 to 72 Ma and was associated with a period of lacustrine deposition. Following this episode of magmatism, a final series of fluvial deposits accumulated (which include zone IV) from ca. 72 to 70 Ma.

A dramatic constriction in the southern entrance to the Western Interior Seaway through the Gulf of Mexico occurred during the latter phase of deposition of the Aguja Formation, corresponding with the shift to southeastward stream flow and the episode of pyroclastic eruptions. The Aguja Formation pyroclastic deposits are thought to be a westernmost expression of the Balcones igneous province in central and southern Texas, and so regional uplift associated with this episode of magmatism may have been responsible for closing the southern aperture of the Western Interior Seaway.

1Supplemental Material. Item S1: Locality information and descriptions for 34 measured stratigraphic sections of the Aguja Formation. Table S1: Aguja Formation vertebrate fauna thus far known from biozones I to IV. Table S2: Aguja Formation wood types, megaflora, and palynomorphs thus far known from vertebrate biozones I to IV. Please visit https://doi.org/10.1130/GEOS.S.25216061 to access the supplemental material, and contact editing@geosociety.org with any questions.
Science Editor: David E. Fastovsky
Associate Editor: Cathy Busby

This paper reflects the current state of our learning process that began nearly 40 years ago. Although the initial work was part of the lead author’s dissertation research, many students, advisors, colleagues, and fellow devotees of the Big Bend area have been involved in many ways, helping to create the current progress report. The authors of this report represent those that participated in the seemingly most significant ways, but in particular we also gratefully acknowledge W. Langston, Jr., G. Kocurek, W. Muehlberger, and E. Wheeler for their advice and support during the formative stages of this research; E. Lehman, N. LaFon, M. Sander, J. Browning, D. Evans, R. Record, J. Bohanan, J. Anglen, R. Schroeder, J. Over, J. Mosley, W. Straight, R. Kissel, A. Coulson, and S. Tomlinson for participating in fieldwork over the years and their own insightful research on the Aguja Formation that we have certainly benefitted from; the Science and Resource Management office in Big Bend National Park; and in particular, D. Corrick, P. Koepp, V. Davila, and M. Fleming, who provided assistance with the research planning and permitting process in the park over many years. D. Fastovsky, C. Kortyna, and an anonymous reviewer provided comments that improved the content and presentation of the manuscript. Most importantly, we thank E. Lehman, J. Wick, and all of our “significant others” for their patience over the years we have been engaged in this research. The Department of Geosciences at Texas Tech University provided support for preparation and publication of the manuscript.

Gold Open Access: This paper is published under the terms of the CC-BY-NC license.