Fluvial siliciclastic rocks bracketing the Cretaceous- Paleogene (K-Pg) boundary in the San Juan Basin, New Mexico (USA), provide records of regional fluvial and tectonic evolution during the Laramide orogeny. Petrographic analyses of sandstones from the Upper Cretaceous Fruitland Formation and Kirtland Formation and the Paleocene Ojo Alamo Sandstone and Nacimiento Formation show that the rivers depositing these sediments were sourced in areas where unroofing of crystalline basement rocks took place, introducing an increasing proportion of immature detrital grains into the fluvial system through time. After the Cretaceous-Paleogene boundary, rivers deposited an increasing amount of microcline and orthoclase feldspar relative to plagioclase feldspar, suggesting a growing source in unique crystalline basement rocks. Geochemical analyses show significant differences between Al- and K-poor Upper Cretaceous sandstones and Al- and K-rich lower Paleocene sandstones in the San Juan Basin.
The high proportion of sand-sized material in the Ojo Alamo Sandstone suggests that it was deposited in a basin with a low ratio of sediment supply to accommodation. However, magnetostrati-graphic age constraints suggest it had a relatively high sedimentation and/or subsidence rate of as much as 0.38 m/k.y. The sediment supply must have been high in order to deposit a basin-wide coarse sand-dominated package, suggesting rapid creation of topographic relief in the San Juan uplift, the proposed source area of the Ojo Alamo fluvial system.
The observed sedimentary architecture and age constraints of the Ojo Alamo Sandstone, including kilometers-wide sand bodies and limited overbank mudstones throughout most of the outcrop area, are difficult to reconcile with accepted models of aggradation and avulsion in large fluvial systems, but available age and lithologic data make difficult a complete understanding of Paleocene San Juan Basin fluvial systems and basin evolution. Here, we present new lithologic, petrographic, and thickness data from San Juan Basin K-Pg fluvial siliciclastic units and interpretations of their origins.
Since the 1970s, Ojo Alamo Sandstone (New Mexico, USA) research has focused primarily on the interpretation of the formation's position at or near the Cretaceous-Paleogene (K-Pg) boundary (e.g., Fassett et al., 2002; Fassett, 2009; Lucas et al., 2009; Flynn et al., 2020), its potential economic role as a hydrocarbon reservoir or uranium source (e.g., Vizcaino and O’Neill, 1977), and its importance as a marker of regionally extensive Laramide tectonic activity (e.g., Baltz, 1967; Dickinson et al., 1988; Cather, 2004; Blum and Pecha, 2014; Cather et al., 2019). Few studies have offered basin-scale sedimentation, tectonic, and paleo-environmental interpretations for the Ojo Alamo Sandstone. Recent advances in understanding of fluvial sedimentary processes and preservation potential (Owen et al., 2015a, and references therein), Laramide tectonic deformation history of the southern Rocky Mountains (Heller et al., 2012, and references therein), and radiometric and magnetostratigraphic age controls for Cretaceous and Paleogene rocks of the San Juan Basin (SJB) (Donahue, 2016; Cather et al., 2019; Flynn et al., 2020) allow for updated interpretations of the depositional history, lithostratigraphic relationships, and tectonic significance of the Ojo Alamo Sandstone, which we present here.
The K-Pg boundary in the SJB of northwestern New Mexico is located between the Upper Cretaceous Kirtland Formation and the Paleocene Ojo Alamo Sandstone. The exact position of the boundary was subject to conflicting interpretations throughout the 20th century (e.g., Bauer, 1916; Ree-side, 1924; Dane, 1936; Baltz et al., 1966; Lindsay et al., 1981; Fassett, 1985). Though still debated (e.g., Lucas et al., 2009; Fassett et al., 2011), the K-Pg boundary is generally accepted to lie at the base of the Ojo Alamo Sandstone sensuBaltz (1967), i.e., at the base of the Kimbeto Member of the Ojo Alamo Sandstone sensuPowell (1973) and Cather et al. (2019). We use that position throughout this paper and note that the exact position of the boundary does not affect our interpretations of basin sedimentary architecture and tectonics.
The intricacies of spatial and temporal relationships between aggradation-progradation, degradation, system quiescence, accommodation, and extrabasinal controls lead to complex sedimentary records in fluvially dominated terrestrial basins. Recent work shows that many fluvial siliciclastic packages preserved in modern continental sedimentary basins are the deposits of prograding distributive fluvial systems (DFSs) or megafans (Hartley et al., 2010; Weissmann et al., 2010, 2013, 2015; Kukulski et al., 2013). Other workers highlight the abundance of both modern and ancient fluvial deposits that do not meet the criteria for a DFS, suggesting that tributary fluvial systems are at least as likely as DFSs to be preserved in the rock record (Fielding et al., 2012; Latrubesse, 2015). In either type of system, incision, reworking of sediment, and downcutting occur when the rate of sediment volume supply is greater than the rate at which sediment can be accommodated by subsidence in the basin (Shanley and McCabe, 1994; Holbrook et al., 2006). Conversely, when the creation of accommodation is greater than the rate of sediment volume supply, the fluvial system experiences aggradation. Aggrading and/or prograding fluvial systems cannot uniformly deposit sediment throughout a basin at any one point in time. Rather, deposition occurs in and around individual active channels occupying a small portion of the total basin area (Shukla et al., 2001; Nichols and Fisher, 2007). Over time, as the active channels change position, most or all of the basin area experiences deposition and sediment accumulation. Basin-wide lithologically uniform beds or deposits therefore are assumed to be diachronous, though the duration of their diachroneity can vary (Nichols and Fisher, 2007).
Sedimentation in Laramide basins in the North American Cordillera from the Late Cretaceous to the Eocene was simultaneous with uplift in the regions surrounding the basins (Dickinson et al., 1988; Clinkscales and Lawton, 2018; Lawton, 2019), and these sediments commonly provide the highest-resolution records of local to regional deformation and associated syntectonic responses. Laramide basin sediments have been investigated for insights into paleogeography, paleotectonics, and mantle processes (Dickinson et al., 1988; Heller et al., 2003; Yonkee and Weil, 2015; Heller and Liu, 2016). These investigations share the premise that regional flexural response was responsible for and contemporaneous with creation of uplifts and basins that led to (1) erosion of materials in the uplifted areas surrounding basins, (2) transport downgradient toward basins, and (3) contemporaneous and subsequent deposition and storage within basins as the basins subsided. As such, the fluvial deposits preserved in Laramide basins are directly related to the fluvial, magmatic, climatic, and tectonic conditions present during their erosion, transport, and deposition.
Here, we document lithologic and petrographic changes in SJB K-Pg fluvial sedimentary rocks. These changes record evolution of the fluvial systems that deposited sediment in the SJB during the Laramide orogeny and allow for increased understanding of regional paleogeography. In addition, we report on the fluvial sedimentology of the Ojo Alamo Sandstone, which comprises the first record of SJB sedimentation during and after the major Laramide reorganization of paleogeographic and paleofluvial features. This formation shows that basin- wide fluvial sedimentation can occur rapidly and does not always produce the sedimentary architecture expected in an avulsive fluvial system.
The SJB is located in the Four Corners region of New Mexico and Colorado (Fig. 1) in the Navajo physiographic section of the Colorado Plateau physiographic province (Fenneman and Johnson, 1946). The SJB is an asymmetrical broken-fore-land structural basin formed during the Laramide orogeny, which deformed Paleozoic through early Cenozoic strata (Dickinson et al., 1988; Cather, 2004). The axial trace of the SJB is arcuate, located near the northern and eastern basin margins with steeply dipping northern and eastern limbs and shallowly dipping southern and western limbs. Like most of the sedimentary cover of the Colorado Plateau, the central, southern, and western SJB is relatively undeformed, with regional dips of <5°. The eastern basin margin is highly deformed along the Nacimiento uplift, with Mesozoic and Cenozoic units steeply dipping to overturned (Woodward et al., 1972). The northern basin margin is moderately deformed along the San Juan uplift, Archuleta anti-clinorium, and Hogback monocline, with Paleozoic through Cenozoic units moderately to steeply dipping (Steven et al., 1974). The western and southern basin margins are slightly deformed along the Defiance upwarp and Zuni uplift, respectively, with Meso zoic and Cenozoic units gently dipping (Baltz, 1967). The structure of the SJB leads to a bullseye pattern in map view, with Eocene sedimentary units in the central basin surrounded by rings of outcrop-ping Paleocene and Cretaceous units.
The SJB was located on the western margin of the Western Interior Seaway during its maximum inundation from ca. 95 Ma (Cenomanian) until ca. 74 Ma (Campanian) and accumulated as much as 1900 m of marine sands and muds during three major transgressive-regressive episodes (Leipzig, 1982; Klute, 1986; Roberts and Kirschbaum, 1995). Following the retreat of the Western Interior Seaway at ca. 74 Ma, the SJB accumulated as much as 625 m of fluvioclastic sediment during the Campanian and Maastrichtian (Klute, 1986; Sikkink, 1987; Cather, 2004; Donahue, 2016). These deposits include the Fruitland Formation and the Kirtland Formation. During the onset of the major Laramide regional deformation in the Paleocene, the SJB accumulated as much as 700 m of fluvioclastic sediment (Sik-kink, 1987; Williamson and Lucas, 1992; Williamson, 1996; Russell, 2009). These deposits in the northern SJB include the McDermott Formation and Animas Formation, and in the central and southern SJB, the Ojo Alamo Sandstone and Nacimiento Formation. The contemporaneous relationships of these units are illustrated in Figure 2. Fluviatile deposition continued into the Eocene after major Laramide uplifts occurred adjacent to the SJB, resulting in the accumulation of as much as 650 m of fluvioclastic sediment of the San Jose Formation (Smith, 1992, 1988; Milner et al., 2005). Episodic deposition likely continued through the Oligocene (e.g., Cather et al., 2008), but SJB bedrock units younger than Eocene are not preserved within the basin. A generalized stratigraphic column showing units relevant to this paper is presented in Figure 2.
Throughout the 20th century, the issues of age and nomenclature surrounding the K-Pg units in the SJB were addressed by Sinclair and Grainger (1914), Reeside (1924), Dane (1936), Baltz (1953, 1967), Anderson (1960), Baltz et al. (1966), O’Sullivan et al. (1972), Clemens (1973), Fassett (1973, 1974), Lindsay et al. (1978, 1981), Klute (1986), and Sikkink (1987). Baltz et al. (1966) designated the lower conglomerate and dinosauriferous middle shale of Bauer (1916) to the Kirtland Formation and restricted the name Ojo Alamo Sandstone to the upper conglomerate and pebbly sandstone. Fassett (1966) extended this designation to include the medium- to thickly interbedded shales and channel sandstones over-lying the shales of the Kirtland Formation at Mesa Portales in the southeasternmost outcrop area near Cuba, New Mexico (these facies are absent in other portions of the SJB). The debate over these designations has continued for half a century and is fueled by contradictory absolute age estimates, geographically limited field studies, varying interpretations of conformable versus disconformable contacts, and nomenclature obfuscation.
Powell (1973) provided measurements of >7000 paleocurrent- direction indicators throughout the Ojo Alamo Sandstone outcrop area. His interpretation of these data suggests a mean flu-vial transport azimuth of 138°. Given the direction and consistency of paleocurrent indicators, Powell suggested a single sedimentary source area to the northwest of the SJB in the vicinity of the present western San Juan Mountains or La Plata Mountains (Colorado). Powell's interpreted depositional environment for the Ojo Alamo Sandstone is an alluvial plain that resulted from the downgradient transport of sands and gravels from alluvial fans located proximal to the source area.
An investigation of outcrops of Ojo Alamo Sandstone by Klute (1986) included sedimento-logic, petrographic, and paleocurrent analyses and interpretations. Klute interpreted the depositional environment of the Ojo Alamo Sandstone as South Saskatchewan–or Platte-type sandy braided rivers sensuMiall (1981). The sandy braided river models of Miall (1981, 1996) best fit most of the Ojo Alamo Sandstone sedimentary features described by Klute (1986). However, none of the modern river analogs in Miall (1981, 1996) or Klute (1986) match all of the features observed in the Ojo Alamo depositional system. Klute attributed the relative homogeneity of the Ojo Alamo Sandstone throughout most of the study area to the avulsion of active channels through the basin and the subsequent reworking of sediment during channel shifting and flood events. Klute's petrographic analyses show that the Ojo Alamo Sandstone is enriched in potassium feldspar and sedimentary lithic grains relative to the underlying Kirtland Formation fluvial sandstone.
Sikkink (1987) analyzed lithofacies in the Ojo Alamo Sandstone and adjacent units in order to provide interpretations of depositional history near the K-Pg boundary in the SJB. Sikkink interpreted the Ojo Alamo Sandstone to be synchronous with both the Animas Formation and the Nacimiento Formation; interfingering of the Ojo Alamo Sandstone and McDermott Formation is noted in the northwestern SJB near Farmington, New Mexico. The interpretations proposed by Sikkink (1987) include a source area in an early Laramide magmatic center near the present-day La Plata Mountains that shifted shortly thereafter to a source in a basement-cored uplift near the present-day central to eastern San Juan Mountains. Sikkink suggested that K-Pg McDermott Formation and Animas Formation conglomeratic fluvial and debris-flow deposits in the northernmost SJB were shed from these emerging highlands while the finer-grained fluvial deposits, including the Ojo Alamo Sandstone, represent the distal meandering stream counterparts.
In the early 21st century, several studies focused on the Ojo Alamo Sandstone because of isolated dinosaur fossils found within it (e.g., Fassett et al., 2002, 2011; Fassett, 2009; Sullivan et al., 2005; Lucas et al., 2009). These studies were focused in the geographic and stratigraphic vicinity of the dinosaur fossils in question and focused new attention on the age of what was variously interpreted as the Ojo Alamo Sandstone (sensuBaltz et al., 1966) or the Kimbeto Member of the Ojo Alamo Sandstone (sensuPowell, 1973). Of interest to this study is that the magnetostratigraphy utilized by the above studies suggests that the unconformity below the Ojo Alamo Sandstone represents a deposition hiatus of 2–4 m.y. (Lindsay et al., 1981; Butler and Lindsay, 1985; Lucas et al., 2006). However, paleomagnetic and geochronologic evidence from Flynn et al. (2020) suggests the duration of the sub–Ojo Alamo Sandstone unconformity is much shorter. Furthermore, radiogenic isotope age estimates derived from detrital zircons and sanidines in the Kirtland Formation, Ojo Alamo Sandstone, and Nacimiento Formation suggest that the hiatus between these formations was as short as 1.5 m.y. (zircon) and 0.4 m.y. (sanidine) (Mason et al., 2013; Donahue, 2016), further decreasing the length of the K-Pg hiatus in the SJB.
Stratigraphic section sites were selected on the basis of correlation of the Ojo Alamo Sandstone with other units, amount of exposure, completeness of section, and accessibility. Stratigraphic sections were measured with Jacob's staff and tape. In addition to our own measurements, we utilized petrographic data and lithologic descriptions from measured sections previously published by Baltz et al. (1966), Baltz (1967), Powell (1973), Leipzig (1982), Klute (1986), Sikkink (1987), and Wegert and Parker (2011). We analyzed publicly available electric well logs from the New Mexico Oil Conservation Division in order to estimate subsurface thickness of pertinent units. Petrographic data were collected with thin section point counts using the Folk method (Folk, 1957). Whole-rock bulk geochemical analyses of sandstones and mudstones were performed on a Rigaku ZSX Primus II X-ray fluorescence spectrometer by the Analytical Chemistry Laboratory in the Department of Earth and Planetary Sciences at the University of New Mexico (Albuquerque) in order to investigate geochemical trends among K-Pg formations preserved in the SJB.
FIELD OBSERVATIONS AND RESULTS
Ojo Alamo Sandstone
The 5–120-m-thick Ojo Alamo Sandstone contains pebbly arenite and wacke, mud-clast conglomerate, claystone, and siltstone. The unit's thickness varies both in outcrop and in the subsurface throughout the basin. Generalized stratigraphic columns with major lithologic designations and their locations are presented in Figure 3. In most areas, the formation is characterized by pebbly medium- to coarse-grained arenite with local lenses and horizons of claystone. Beds range in thickness from 0.3 to 10 m. Sandstones are typically light brown to yellow brown and in some places have developed reddish brown weathering surfaces or desert varnish. Claystones and siltstones are typically light brown, brown, or dark brown and are commonly covered in colluvium. Sandstone beds commonly overlie scour surfaces in claystones, while claystones conformably overlie gradational boundaries. Sand grains are angular to subrounded and consist of quartz, feldspar, sedimentary and volcanic lithic fragments, chert, and mafic minerals. Medium- and coarse-grained sandstones are predominately clay cemented with minor silica cements. Less common fine-grained sandstones are cemented with roughly equal parts clay and silica. Bedforms include both trough and tabular crossbeds, convolute bedding, horizontal plane bedding, lenticular bedding, wavy bedding, and fluid-escape structures. Silicified stumps, wood fragments, and logs as much as 34 m long are common in the sandstones of the Ojo Alamo Sandstone. Leaf fossils are present but uncommon in the formation's claystones (Flynn and Peppe, 2019).
Conglomerates and conglomeratic sandstones of the Ojo Alamo Sandstone are either pebble conglomerate or mud-clast conglomerate. Pebble conglomerates contain clasts from 2 to 100 mm (−1 to −7 ϕ) in diameter and include subrounded to well-rounded pebbles of chert, alkali feldspar granite, quartzite, silicified wood, trachyandesite, sandstone, limestone, and metapelite. Mud-clast conglomerates contain clasts from 30 to 500 mm (−5 to −9 ϕ) in diameter of subangular to very angular claystone. The mud-clast conglomerate lithofacies is restricted to cut-and-fill structures that represent scouring in both intraformational claystones as well as the underlying Cretaceous claystones at the base of the Ojo Alamo Sandstone.
Descriptions of Ojo Alamo Sandstone Facies
We categorize the Ojo Alamo Sandstone into three broad facies associations: (1) amalgamated channel belt deposits, (2) overbank floodplain deposits, and (3) isolated channel-fill deposits, described below and in Table 1.
Amalgamated Channel Belt Deposits
The amalgamated channel belt facies association contains poorly to moderately well-sorted medium- to coarse-grained pebbly crossbedded and plane-bedded sandstones. These lithological compositions and sedimentary structures, along with abundant petrified wood and lack of marine or estuarine fossils, suggest deposition in fluvial channel environments (Miall, 1978; Robinson and McCabe, 1998; Bridge, 2003; Gibling, 2006). Tabular and trough crossbeds represent downstream and oblique bar-face deposits (Best et al., 2003) and range from 0.3 to 1.5 m in thickness. The abundance of tangential crossbeds suggests high sediment transport rates (Bagnold, 1954; Sallenger, 1979; Bridge, 2003). Planar horizontal beds represent vertically accreted bar-top deposits and range from 0.2 to 1.5 m in thickness. These beds suggest high bed shear stress and sediment transport rates (Bridge and Best, 1988, 1997; Bridge, 2003). These facies commonly exhibit highly contorted bedding, especially near the tops of stories. Cut-and-fill structures are found both at the base of and within these deposits (Fig. 4). Where these facies overlie mud-stones, clay clasts from the underlying deposits are found within the sandstones and conglomerates.
The deposits of this facies association form simple and multi-story bodies as much as 11 km wide and as much as 19 m thick. Individual channel belts are impossible to recognize in the field. In the southern and western SJB, these facies contain multiple stories that represent erosion within former channel deposits, suggesting reworking of material and a low ratio of accommodation to sediment supply (Kjemperud et al., 2008; Owen et al., 2015b). Multiple stories, lateral continuity, and the lack of identifiable channel-belt margins collectively suggest that vertical and lateral amalgamation of channel deposits was a widespread process occurring throughout both the temporal and the spatial range of deposition of fluvial units (Friend et al., 1979; Wang et al., 2011). While there is no significant downcurrent thinning of these facies, they do exhibit a downgradient reduction in mean and maximum grain size from coarse sand and cobbles in the northwest to medium sand and pebbles at Mesa Portales in the southeast.
Overbank Floodplain Deposits
The overbank floodplain facies association contains sandy mudstones with interbedded fine-grained sandstones. This facies association's laterally discontinuous sand lenses, organic- rich fossil hash layers, and leaf fossils (Flynn et al., 2014; Flynn and Peppe, 2019) suggest deposition in flu-vial floodplain environments (Kraus and Gwinn, 1997; Alexander and Fielding, 2006). Paleosols are uncommon; where present, they lack horizonation and are identified based solely on root traces. Bedding is difficult to observe in most outcrops but is thin and tabular to laminar where observed. Mud-stones in these facies contain considerable sand and silt, suggesting deposition in close proximity to channels, high sediment supply, or high flood magnitude (Guccione, 1993; Pizzuto, 1987; Owen et al., 2015b). The sandy and poorly sorted nature of these deposits is similar to that described in the proximal overbank deposits of Slingerland and Smith (2004) and Hajek and Edmonds (2014), which they show to be associated with vertically accreting systems in aggradational settings. These facies conformably overlie sandy channel deposits and are overlain by amalgamated channel belt deposits and cut-andfill structures.
Deposits of this facies association form laterally extensive complexes with widths of >1 km and thicknesses of as much as 7 m. These deposits are commonly truncated by channel sandstones (Fig. 5A). In the southern SJB at Mesa Portales and the western SJB at Head Canyon, these facies separate as many as four thick, laterally extensive amalgamated channel belt complexes, leading to the bluff-and-slope topography common in the Ojo Alamo outcrop area at those locations (Fig. 5B).
Isolated Channel-Fill Deposits
The isolated channel-fill facies association contains poorly to moderately sorted medium- to coarse-grained sandstones and pebbly sandstones that fill simple channels with well-defined channel geometry in cross-section. Their geometry (Fig. 4A) suggests incision of a simple channel into overbank or channel deposits followed by deposition from a non-migrating channel. These deposits commonly contain clay rip-ups where overlying mudstones. Both symmetric and asymmetric channel forms are present. Wings associated with levee and/or crevasse-splay deposits are absent. Some isolated channel-fill deposits are found in close proximity to one another in the same stratigraphic interval (Fig. 4B), suggesting possible anastomosing fluvial form. Given the lithologic similarity to the sandstones of the amalgamated channel belt facies through which they incised, these isolated channel-fill deposits might represent possible episodes of low stage during which the depositional rivers assumed an anastomosing form that was not present during episodes of higher stage (Miall, 1981, 1996). Alternatively, these deposits could represent cross-bar channels (Sambrook Smith et al., 2009). This interpretation does not require a stage-dependent variation in fluvial form and is in accord with inferred reworking of sediment arising from the observed lithologic homogeneity between isolated channel-fill deposits and amalgamated channel belt deposits.
Deposits of this facies association form lens-shaped bodies of 5–90 m width and 1–9 m thickness that cut into underlying sandstones and mudstones. These deposits are overlain by sandstones along distinct boundaries. Most isolated channel-fill deposits in the Ojo Alamo Sandstone exhibit no appreciable upward-fining trends, suggesting that they are not abandoned main channels (Miall, 1996; Bridge, 2003). Flynn and Peppe (2019) reported 15–25-m-wide lenticular carbonaceous shales they interpreted as abandoned channel fills analogous to oxbow lakes in modern meandering fluvial systems, an interpretation with which we agree. These deposits are difficult to observe except where exposed in vertical cliff faces, leading to a probable underestimation of their abundance. Given their relative thinness and narrowness when compared to geometries of amalgamated channel belt deposits, we do not interpret the channel dimensions recorded in the isolated channel-fill facies as representative of the main channels of the Ojo Alamo Sandstone fluvial depositional system.
The sedimentary architecture of the Ojo Alamo Sandstone varies across the outcrop area. In some sections, such as the Shannon Bluffs south of Farmington (Fig. 1), the Ojo Alamo Sandstone forms a single 12–70-m-thick amalgamated complex channel belt sandstone cliff with no observed mudstones. These multi-story sand bodies are laterally continuous for at least 3 km. In other locations, such as Head Canyon and Mesa Portales, the Ojo Alamo Sandstone consists of laterally continuous sand bodies separated by mudstone interbeds. Where mudstones are present, they are commonly laterally truncated by overlying channel deposits, making mudstones less laterally extensive than sandstones. In the Bisti/De-Na-Zin Wilderness Area and Escavada Wash (Fig. 1), the Ojo Alamo Sandstone has a 6–10-m-thick single-story form. These thinner single-story forms contain abundant very coarse sand and pebbles, perhaps representing high-energy bar-head or confluence scouring and minimal deposition of relatively coarse material.
Petrographic data from SJB Upper Cretaceous and lower Paleogene sandstones are presented in Figure 6. Sandstones from the Fruitland Forma tion are well-sorted fine- to medium-grained subrounded to rounded quartz arenites. Kirtland Formation sandstones show significant variation in mineralogical composition; they are moderately to well-sorted fine- to medium-grained subrounded to subangular arkosic arenites and lithic arenites. Sandstones from the Ojo Alamo Sandstone are poorly to moderately sorted medium- to coarse-grained subrounded to very angular arkosic arenites; two of the 39 samples are lithic arenites. Sandstones from the lower Nacimiento Formation (Arroyo Chijuillita Member of Williamson and Lucas  and Kutz Member of Cather et al. ) are moderately to well-sorted fine- to medium-grained angular to well-rounded feldspathic arenites and arkosic arenites cemented with clays and amorphous silica.
Detrital grains in SJB K-Pg sandstones include quartz, plagioclase feldspar, potassium feldspar, chert, lithic fragments, hornblende, micas, amphibole, and opaque minerals assumed to be oxides. Both monocrystalline and polycrystalline quartz grains are present. The wide range of angularity in quartz and chert grains in the Kirtland Formation, Ojo Alamo Sandstone, and Nacimiento Formation suggests that rivers depositing all three of these formations were reworking sands from older Phanerozoic units upstream. Detrital feldspars in the Ojo Alamo Sandstone display a wide range of physical weathering and chemical alteration by sericitization and vacuolization (Fig. 7), with chemically altered feldspars becoming less common upsection.
Relative proportions of plagioclase feldspar and potassium feldspar for sandstones from the Kirtland Formation and Ojo Alamo Sandstone are presented in Figure 8. Mean values for 25 Kirtland Formation sandstones are 62% plagioclase feldspar and 38% potassium feldspar. Mean values for 39 Ojo Alamo Sandstone sandstones are 37% plagioclase feldspar and 63% potassium feldspar. These data show that while the total feldspar content in both formations is not significantly different, there is significant variation in the ratios of plagioclase feldspar to potassium feldspar in the SJB across the K-Pg boundary, with the Ojo Alamo Sandstone being richer in potassium feldspar. In both formations, potassium feldspars include orthoclase and microcline.
The bulk geochemical compositions of 57 sandstones from the Ojo Alamo Sandstone and immediately adjacent formations are presented in Table 2. SiO2 and Al2O3 compose 60%–89% of the sandstones by mass. The Ojo Alamo Sandstone is richer in K and Al than the sandstones of the Kirt-land Formation (Figs. 9 and 10). The geochemical composition of the Nacimiento Formation sandstones is more variable and overlaps with that of the Kirtland Formation and Ojo Alamo Sandstone.
Changing Sediment Sources across the K-Pg Boundary
As shown in Figure 6, significant petrographic variation exists among the sandstones of the Fruit-land Formation, Kirtland Formation, Ojo Alamo Sandstone, and Nacimiento Formation. The quartz arenites of the Fruitland Formation (Leipzig, 1982) suggest that the rivers that deposited the Fruitland Formation during middle to late Campanian time were sourcing mature sediments, likely reworked from underlying Cretaceous marine and shoreface sands. The scarcity of feldspars and lithic grains is indicative of either very distal deposits of a mature fluvial system or a quartz-rich source area, such as a broad coastal plain underlain by beach deposits of the ultimate Western Interior Seaway regression, now preserved in the SJB as the Pictured Cliffs Sandstone. The quartz-feldspar-lithic (QFL) composition of the Fruitland Formation plots in the craton interior–continental block provenance categories of Dickinson et al. (1983). The increase in feldspar and lithic grains in sandstones of the Kirtland Formation could be indicative of unroofing of basement rocks in the sediment source area, an increase in volcanic activity in the drainage basin, a significant reorganization of regional drainage patterns leading to a different sediment source area, or some combination of these. Given the conformable character of the Fruitland Formation–Kirtland Formation contact (Baltz, 1967; Klute, 1986), significant reorganization of regional drainage patterns between the deposition of these two formations is unlikely. Instead, we interpret the increase in feldspar and lithic grains in Kirtland Formation sandstone as the result of evolving upland source areas related to the early phase of Laramide localized uplift. The sand-dominated fluviatile Vermejo Formation in the Raton Basin ~250 km east of the SJB (Fig. 1) is coeval with the Kirtland Formation and contains detrital grains indicative of as much as 1 km of localized uplift in the headwaters of its depositional fluvial system (Cather, 2004). Similar-scale uplift and fluvial response likely occurred in the source areas of rivers in the SJB, leading to the increase in feldspathic and lithic detrital grains in the Kirtland Formation. In light of the paleocurrent synthesis of Cather et al. (2019), indicating southeastward direction of transport, and paleogeographic information from Baltz (1967), Cather (2004), Heller et al. (2012), and Lawton (2019), we suggest that this relief might have been created in the first stages of the Laramide San Juan or Defiance uplifts to the northwest of the SJB or the regional uplift and unroofing of the Mogollon Rim to the southwest of the SJB (Fig. 1).
Sandstones in the Ojo Alamo Sandstone show an increase in both the total proportion of lithic grains and in the ratio of potassium feldspar to plagioclase feldspar relative to the underlying Cretaceous sandstones. The increase in the ratio of potassium feldspar to plagioclase feldspar and the proportions of hornblende, mica, and heavy mineral grains in the Ojo Alamo Sandstone suggest development of a sediment source in unique crystalline igneous or metamorphic terrains, potentially the Needle Mountains (Colorado) in the western San Juan uplift ~125 km north of the SJB or the Mogollon Rim ~300 km southwest of the SJB. The paleo-current analyses of Powell (1973) and Sikkink (1987) both support a northerly source area for the Paleocene fluvial system that deposited the Ojo Alamo Sandstone. The Needle Mountains in the western San Juan Mountains contain the nearest and largest exposures of crystalline basement rocks and, in light of the above paleocurrent analyses, were upstream of the Ojo Alamo Sandstone depocenter, making them the likely source for basement-derived detritus in the Ojo Alamo Sandstone. Donahue (2016), Bush et al. (2016), and Pecha et al. (2018) presented data showing a significant population of detrital zircons from the Ojo Alamo Sandstone with ages as young as 68.0 ± 1.4 Ma, suggesting a young igneous source within the Ojo Alamo drainage area, most likely the Colorado Mineral Belt–associated La Plata magmatic center ~100 km north of the SJB. This interpretation agrees with the paleocurrent analyses and the presence of minerals in both the La Plata magmatic center and the more proximal McDermott Formation (O’Shea, 2009; Gonzales, 2010; Wegert and Parker, 2011).
Sandstones in the lower Nacimiento Formation have similar petrographic compositions to the Ojo Alamo Sandstone. Unaltered detrital grains of plagioclase feldspar, potassium feldspar, and hornblende suggest a source area in crystalline bedrock, likely shared with that of the Ojo Alamo Sandstone. Sericitized feldspars and well-rounded quartz and chert grains suggest reworking of sediment from older sedimentary units. The >5-km-thick exposed section of Paleozoic and Mesozoic sedimentary rocks between the San Juan uplift and the SJB depocenter provides abundant potential sources for the variety of grain compositions and textures in Nacimiento Formation sandstones. The petrographic composition, interfingering contact with the underlying Ojo Alamo Sandstone, and earliest Paleocene age of the Nacimiento Formation all suggest that it is the deposit of the same fluvial system(s) that deposited the Ojo Alamo Sandstone.
Regional Landscape Evolution Recorded in SJB K-Pg Deposits
The relative homogeneity, lateral surface and subsurface stratigraphic continuity, and improved age constraints from the Ojo Alamo Sandstone and adjacent units provide a unique opportunity for interpretation of the sedimentary, tectonic, and fluvial factors at work during deposition across the K-Pg boundary, shown in paleogeographic maps in Figures 11–13. Campanian and lower Maastrichtian fluvial deposits in this study area record rivers flowing from highlands in the southwest toward the coastal plain of the retreating Western Interior Seaway (Potochnik, 1989; Cather et al., 2012; Dickinson, 2013) (Fig. 11). Detrital zircons in these sediments are predominately Phanerozoic in age, suggesting a limited source of Proterozoic crystalline basement rocks during this depositional period (Dickinson et al., 2012; Bush et al., 2016; Donahue, 2016; Pecha et al., 2018). The orientation of these rivers was similar to the north- northeast- flowing Turonian rivers proposed by Blum and Pecha (2014), indicating that the lower Kirtland Formation provides the final record of sedimentation in the study area prior to significant regional drainage reorganization caused by the Laramide orogeny. The relatively large areal expanse of these Turonian through Maastrichtian deposits suggests limited localized basin subsidence. Farther west in the North American Cordillera, Campanian deposits record evidence of broad flexural responses to the beginning of Laramide-style deformation (Aschoff and Steel, 2011; Leary et al., 2015), but no such evidence is yet recognized in the SJB.
After deposition of the Fruitland Formation and the majority of the Kirtland Formation by north-northeast-flowing rivers, there was a period of non-deposition and erosion in the SJB (Fig. 12). This erosion removed more sediment from the southeastern SJB than from the northwestern SJB (Fassett, 1974, 1985; Newman, 1987), resulting in preservation of younger pre-erosion rocks in the northwest than in the southeast. This differential erosion is perhaps due to uplift in the southeast related to the incipient Nacimiento uplift, passage of a flexural high associated with the transition from broad-wavelength Sevier tectonism to shorter-wavelength Laramide tectonism beneath the SJB during this interval, or increased subsidence to the east of the SJB leading to lowering of local base level. More work is needed to increase understanding of the timing and effects of tectonism in the area. We propose that this period of non-deposition and erosion represents the interval of SJB drainage reorganization resulting from regional landscape changes associated with the onset of Laramide tectonism. The late Maastrichtian age of this depositional hiatus (Williamson and Weil, 2008) is in accord with proposed regional- (Mackey et al., 2012; Bush et al., 2016; Cather et al., 2019) to continental-scale (Galloway et al., 2011; Cather et al., 2012; Blum and Pecha, 2014) drainage reorganization episodes. Given the lack of sedimentary rocks in the SJB from this interval, it is unclear whether or not the La Plata magmatic center or the San Juan uplift contributed sediment to the SJB via south-southeast-flowing streams during this interval. As shown in Figure 12, we interpret the erosive Maastrichtian fluvial systems to have had the same general orientation as those that deposited the Kirtland Forma tion. The variation in the thickness of the Ojo Alamo Sandstone across the SJB (Fig. 14) might be explained by accommodation of earliest Paleocene (i.e., Ojo Alamo Sandstone) sediment in incised valleys left by north-northeast-flowing Maastrichtian erosive streams.
The Maastrichtian hiatus ended with the deposition of the first basin-filling sediments by south- southeast–flowing fluvial systems in the latest Maastrichtian and earliest Paleocene, beginning with the Naashoibito Member of the Kirtland Formation and the Ojo Alamo Sandstone. Detrital zircon (Donahue, 2016; Pecha et al., 2018), volcaniclastic grain petrography (O’Shea, 2009; Gonzales, 2010; Wegert and Parker, 2011), and paleocurrent (Powell, 1973; Sikkink, 1987) evidence suggests a source area in the La Plata magmatic center and San Juan uplift, requiring a south- southeast paleoflow direction for the Ojo Alamo depositional system (Fig. 13). The accommodation of Laramide uplift–derived sediment in the SJB represents two significant changes in regional landscape evolution: first, a >90° shift in regional drainage directions resulting from reorganization of regional-scale topography; and second, a preferential storage of sediment in isolated subsiding basins rather than on wide coastal plains. The first shift is recorded nearly simultaneously in other basins (Heller et al., 2012). The second shift is in accord with K-Pg sedimentary records on the western Gulf Coastal Plain in Texas (USA) that show a lack of earliest Paleocene sediments there (Galloway et al., 2011; Mackey et al., 2012). The south-southeast–flowing Paleocene depositional system persisted through the deposition of the Nacimiento Formation.
Avulsion and Preservation during Ojo Alamo Sandstone Deposition
where l is a roughness length scale in meters, and r is the basin-wide sedimentation rate in meters per thousand years (Wang et al., 2011). Sedimentation durations below TC will produce sediments that record both autogenic and allogenic variations. For roughness length scales, we use the depth of the thickest channel deposits we observe in the Ojo Alamo Sandstone of 10 m. Sedimentation rates are calculated from available age constraints and formation thickness data (Table 3). Our TC estimates range from 26 k.y. (using the highest sedimentation rate) to 43 k.y. (using the lowest sedimentation rate). These TC values essentially represent the duration required for the fluvial system to have filled its deepest channels with sediment. These estimates are an order of magnitude less than the duration of sedimentation for the entire Ojo Alamo Sandstone estimated above. Therefore, it is likely that the Ojo Alamo Sandstone's sedimentary architecture is the result of predominately allogenic processes. This conclusion is supported by the formation's wide geographic distribution, which would have been unlikely to have occurred in a system controlled by autogenic processes.
Estimated Ojo Alamo Sandstone sedimentation rates are 1.1–6.2 times greater than estimated sedimentation rates of the overlying Nacimiento Formation of Williamson (1996), suggesting a rapidly growing sediment source, rapidly subsiding basin, or combination of the two. The Nacimiento Formation estimated sedimentation rates of Williamson (1996) are minimum estimates because they do not take into account the numerous sedimentary hiatuses evidenced by well-developed paleosols throughout the formation (Hobbs and Fawcett, 2014). In fluvial-dominated continental basins, high sedimentation rates are usually associated with mudstones because a subsiding basin creates accommodation for storage of even fine-grained easily transportable sediment (Blair, 1987; Valero et al., 2017). Therefore, the deposition of as much as 120 m of coarse sands by the Ojo Alamo fluvial system in as little as ~328 k.y. is paradoxical. With the exception of minor muds preserved at Mesa Portales, a large proportion of the fine sands and muds must have bypassed through the SJB. The downgradient depocenter is not known and likely was removed by post-deposition uplift and erosion. By the late Paleocene, fluvial systems were routing sediment from the southern Rocky Mountains Laramide highlands to the western Gulf of Mexico coastal plain (Galloway et al., 2011; Mackey et al., 2012), but Paleocene sediments along the presumed paleoriver paths between the SJB and the Gulf of Mexico coastal plain are not preserved. In light of these considerations, the Ojo Alamo Sandstone in the SJB highlights the problems associated with assuming that the present structural basin in which fluvial sedimentary rocks are preserved can be directly equated to the fluvial basin in which those sediments were deposited. The lithologic composition (e.g., predominant coarse sandstones and scarcity of mudstones) and the proximity to major Laramide tectonic features (e.g., <1 km from the Nacimiento fault, a Laramide reverse fault with ~2 km of vertical offset; Woodward et al., 1992) suggest that the present outcrop and subsurface area of the Ojo Alamo Sandstone represents a small portion of the area of the entire fluvial system in which the sediments were deposited. Recent studies of well-preserved intact paleofluvial basins have shown that the deposits of a single fluvial system show significant lithologic and stratigraphic variation over spatial scales of hundreds of kilometers (Klausen et al., 2014, 2015; Owen et al., 2015b). Deposits of such systems are subject to post-deposition deformation (Klausen et al., 2014) and partial exhumation and removal (Owen et al., 2015b), making interpretation of the remnants of the system subject to inherent problems associated with incomplete stratigraphic records. While there is no evidence for major vertical downcutting or erosion after deposition of the Ojo Alamo Sandstone, some portion of it has been removed around the basin margins.
Several workers have addressed the topic of recognizing avulsion deposits in fluvial sedimentary packages (e.g., Smith et al., 1989; Kraus and Wells, 1999; Slingerland and Smith, 2004; Jones and Hajek, 2007). Though the specific type of avulsion affects the exact sedimentary signature of the event, avulsion deposits are generally recognized as having (1) a lower coal or paleosol overlain by a general coarsening-upward sequence of heterolithic fine sandy deposits and channel sandstones belts, and (2) an upper paleosol representing abandonment of the developed channel system (Kraus and Wells, 1999). The Ojo Alamo Sandstone lacks paleosols, abundant overbank mudstones, and heterolithic fine sandy deposits, making interpretation of the avulsion history of its depositional system difficult. The formation is spatially widespread, exhibits subparallel paleocurrent indicators throughout its outcrop area, and appears to have a common source area in the La Plata magmatic center–San Juan uplift area, which all suggest that it was deposited by a fluvial system characterized by migrating channels. Avulsion is a requisite process for such systems to form. However, the sedimentary architecture of the Ojo Alamo Sandstone does not provide the record of avulsive processes outlined above. Reconciling the seeming necessity of avulsion in the system with the observed lack of well-defined avulsion deposits requires interpretation of the processes and products of the Ojo Alamo fluvial system. One possible cause, illustrated in Figure 15, is that the Ojo Alamo fluvial system did not undergo random avulsion in the typical sense. Fassett (1985) suggested that the Ojo Alamo Sandstone depositional system shifted eastward through time due to development of accommodation in that direction during the development of the Laramide foreland basin and uplifts (Fig. 15). This could explain the fining-eastward lithologic changes noted by Sikkink (1987). If over-bank muds were stored preferentially at any one time on the eastern side of the fluvial system due to increased accommodation there, then perhaps they were subsequently removed when the channel belt shifted eastward. In this scenario, the only remnants of the overbank muds are the thin discontinuous mudstones observed among the channel belt deposits. The abundant channel scours into and through muds (Fig. 4A) are evidence of mud removal during Ojo Alamo Sandstone deposition; perhaps this process was more important than has been previously considered. Flynn et al. (2020) reported paleomagnetic evidence for the Ojo Alamo Sandstone–Nacimiento Formation contact being time transgressive in nature, with the contact younging to the south. Those authors interpreted that time transgression as the result of downstream progradation of a distributive fluvial system. Similar paleomagnetic and geochronologic investigations in the subsurface and along the southern and eastern margins of the Ojo Alamo Sandstone outcrop area are necessary to elucidate whether the formation records an eastward younging. Better constraints on west-to-east age relationships of the Ojo Alamo Sandstone are needed in order to develop this hypothesis more fully, but the Ojo Alamo Sandstone presents a case in which the prevailing accepted explanations for processes and products of avulsion in continental fluvial systems are problematic.
An alternative interpretation of the Ojo Alamo Sandstone is that it represents the deposits of roughly parallel coalescing alluvial systems draining the La Plata magmatic center, San Juan dome, and/or Colorado Mineral Belt uplifts during an episode of rapid uplift. This interpretation is supported by the subparallel paleocurrent indicators of Powell (1973) and Sikkink (1987), perpendicular-to-paleoflow lithologic changes of Sikkink (1987), and thickness variation (Fig. 14) of the formation. The rapid uplift of a northerly source area would have provided an abundance of coarse material for south-flowing streams that rapidly filled accommodation space and bypassed most fines through the SJB. By the time of deposition of the overlying Nacimiento Formation, the ratio of sediment supply to accommodation was lower, leading to accumulation of more fines.
Petrographic and paleocurrent evidence from fluvial siliciclastic rocks bracketing the K-Pg boundary in the SJB provides records of evolution of paleorivers’ source areas through the Late Cretaceous and Paleocene. The first rivers depositing sediment in the SJB after the retreat of the Western Interior Seaway flowed north-northeast and deposited the quartz sands of the Fruitland Formation. By the Maastrichtian Age, the same north-northeast–flowing rivers deposited the arkosic sands of the Kirtland Formation, suggesting unroofing of crystalline bedrock in their source areas southwest of the SJB. Reorganization of regional fluvial systems after deposition of most of the Kirtland Formation in the Maastrichtian resulted in removal of some early Maastrichtian and possibly upper Campanian fluvial sediments and a basin-wide erosional unconformity. This reorganization is likely related to the onset of Laramide tectonism at ca. 67 Ma. By the beginning of the Paleocene at 66 Ma, SJB paleorivers flowed south-southeast from the La Plata magmatic center–San Juan uplift and through the SJB, depositing the arkosic sandstones and conglomerates of the Ojo Alamo Sandstone. The Ojo Alamo Sandstone's relative enrichment in potassium feldspar relative to plagioclase feldspar suggests a source area rich in potassium feldspar, the nearest known and most likely such area being the Needle Mountains of the western San Juan uplift. Early to middle Paleocene sandstones of the Nacimiento Formation show similar paleocur-rent indicators and mineralogic composition to the Ojo Alamo Sandstone. This petrographic information, combined with the interfingering depositional contact between the Ojo Alamo Sandstone and the Nacimiento Formation, suggest that the latter records the continuation of the same fluvial system(s) that deposited the former in an incipient and evolving broken foreland basin.
In the northern and central SJB, the Ojo Alamo Sandstone occurs entirely within the Paleogene portion of chron C29r, which gives it a maximum deposition duration of as 328 ± 15 k.y. and a sedimentation rate of as high as 0.39 m/k.y., indicating significant subsidence during the early Paleocene. Because the wide sandstone bodies that compose the majority of the Ojo Alamo Sandstone are associated with low ratios of accommodation to sediment supply, the Ojo Alamo Sandstone must represent a period of high sediment supply during which predominately coarse sand and pebbles were deposited in the SJB. Given the presence in the Ojo Alamo Sandstone of Maastrichtian-aged detrital zircons from the La Plata magmatic center and the abundance of potassium feldspar presumably from crystalline basement rocks in the San Juan uplift, this high sediment supply is likely related to the rapid development of positive topographic relief in the source area of the Ojo Alamo fluvial system. Finer sands and muds were bypassed through the system, possibly deposited in a more distal reach of the paleoriver system that since has been uplifted and eroded.
Standard sedimentary models of avulsion stratigraphy are not compatible with the sedimentary architecture observed in the Ojo Alamo Sandstone. The formation's high estimated sedimentation rates, predominance of coarse sands, and lack of paleosols provide little evidence of long-term landscape stability in the fluvial system during the earliest Paleocene depositional interval. Because of this, evidence of avulsions in the Ojo Alamo fluvial system is difficult to recognize. However, the formation's conformable relationship with the overlying Nacimiento Formation, lateral continuity, and broad geographic distribution all suggest that it likely was deposited by a long-lived and spatially migrating fluvial system. A potential explanation that reconciles the absence of expected avulsion stratigraphy with the necessity of a spatially migrating fluvial system is that the system did not avulse with the typical compensation-driven processes associated with rivers in unconfined basins. Rather, the Ojo Alamo fluvial system potentially was driven unidirectionally eastward across the incipient SJB by relatively increased uplift in the west, relatively increased subsidence in the east, or some combination of the two during the onset of Laramide deformation in the earliest Paleocene. The observed lack of avulsion stratigraphy throughout most of the Ojo Alamo Sandstone alternatively could be explained by the interpretation that the formation represents not the deposits of one fluvial system but rather the coalesced deposits of multiple allu-vial systems draining a rapidly uplifted northerly source area. Further work is needed regarding the timing of uplifts adjacent to the SJB in order better to constrain SJB depositional history and to inform understanding of relationships between tectonics and sedimentation.
Thorough and constructive comments from Associate Editor Cathy Busby and reviewers Chris Clinkscales and Andrew Flynn allowed for considerable improvements to an earlier version of this manuscript. The authors sincerely thank them for their input.