Abstract

The Late Campanian (Late Cretaceous), upper part of the El Disecado Member, El Gallo Formation, Baja California, México, preserves a rich fossil assemblage of microvertebrates and macrovertebrates, silicified logs, macroscopic plant remains, and pollen that was likely deposited as the distal part of a subaerial fan. The unit was episodic and high energy, with its salient features deriving from active river channels and sheet, debris-flow deposits. Landscape stability is indicated by the presence of compound paleosol horizons, containing Fe2O3 mottling in B horizons, cutans, and calcium carbonate concretions. All of these features indicate wet/dry cyclicity in subsurface horizons, likely attributable to such cyclicity in the climate. Drainage was largely to the north and to a lesser extent, the west; however, some current flow to the south and east is preserved which, in conjunction with the proximal location of marginal marine deposits, suggest the influence of tides in this setting.

The fossil vertebrates preserved in this part of the El Disecado Member are almost exclusively allochthonous, preserved as disarticulated isolated clasts in hydraulic equivalence in the braided fluvial system. A relatively diverse microvertebrate assemblage is preserved, the largest components of which are first, dinosaurs, and second, turtles. Non-tetrapod fossils are relatively uncommon, perhaps reflecting an absence of permanent standing water in this depositional setting.

Here we report a high-precision U-Pb date of 74.706 + 0.028 Ma (2σ internal uncertainty), obtained from zircons in an airfall tuff. The tuff is located low within the sequence studied; therefore, most of the sedimentology and fossils reported here are slightly younger. This date, which improves upon previously published 40Ar/39Ar geochronology, ultimately allows for comparison of these El Gallo faunas and environments with coeval ones globally.

Primary stable isotopic nodules associated with roots in the paleosols of the terrestrial portion of the El Disecado Member are compared with ratios from similar sources from coeval northern and eastern localities in North America. Distinctive latitudinal gradients are observed in both δ13C and δ18O, reflecting the unique southern and western, coastal geographic position of this locality. These differences are best explained by differences in the floras that populated the northern and eastern localities, relative to the southern and western floras reported here.

RESUMEN

La parte superior del Miembro El Disecado de la formación El Gallo, Baja California, México, Campaniano Tardío (Cretácico Tardío), conserva una asociación fósil rica en microvertebrados y macrovertebrados, troncos silicificados, restos macroscópicos de plantas, y polen, la cual fue depositada posiblemente en la parte distal de un abanico subaéreo. La unidad fue depositada de forma episódica y bajo una alta energía, siendo sus rasgos más característicos canales de ríos trenzados y depósitos de flujos de escombros. La estabilidad del paisaje está indicada por la presencia de horizontes compuestos de paleosuelos que contienen horizontes B con un moteado de Fe2O3, cutans, y concreciones de carbonato de calcio. Todos estos rasgos indican una ciclicidad de humedad/aridez en horizontes debajo de la superficie atribuibles, probablemente, a la ciclicidad en el clima. El drenaje estaba orientado principalmente hacia el norte y en menor grado al oeste; sin embargo, se conservan algunas corrientes de flujo dirigidas hacia el sur y este, las cuáles junto con la ubicación proximal de los depósitos marinos marginales, sugieren la influencia de mareas en este ambiente.

Los vertebrados fósiles preservados en esta parte del Miembro El Disecado son casi exclusivamente alóctonos, se conservan como clastos aislados desarticulados en equivalente hidraúlico “hydraulic equivalence” en el sistema fluvial trenzado. Una asociación de microvertebrados relativamente diversa se conserva; los componentes más grandes son, en primer lugar, los dinosaurios y en segundo, las tortugas. Fósiles no tetrápodos son relativamente raros, reflejando probablemente la ausencia de agua permanente en este ambiente de depósito.

Aquí reportamos una edad U-Pb de alta precisión de 74.706 + 0.028 Ma (2σ de incertidumbre interna), obtenida de zircones de una toba de caída “airfall.” La toba está localizada en la parte inferior de la secuencia estudiada; por lo tanto, casi toda la sedimentología y fósiles registrados aquí son ligeramente más jóvenes. Esta fecha perfecciona la geocronología 40Ar/39Ar publicada previamente, y por último permite la comparación de la fauna y el ambiente de El Gallo con otros contemporáneos.

Proporciones primarias de isótopos estables (δ13C; δ18O) obtenidas de dientes aislados de hadrosaurio (dinosaurio) y de nódulos de carbonato asociados con raíces en los paleosuelos de la parte continental del Miembro El Disecado fueron comparadas con proporciones de fuentes similares de localidades norteñas y orientales contemporáneas en América del Norte. Gradientes latitudinales distintivos fueron observados en δ13C y δ18O, reflejando la singular posición geográfica costera al sur y occidente de esta localidad. Estas diferencias se explican mejor por las diferencias en las floras que habitaron las localidades norteñas y orientales relativas a las sureñas y occidentales registradas aquí.

INTRODUCTION

The Upper Cretaceous El Disecado Member of the El Gallo Formation (Kilmer, 1963), Baja California, México, has proven to be an abundant source of late Campanian fossil material. Fieldwork from the mid-1960s (Morris, 1973) and regular (yearly) collection, since 2004, of microvertebrates by researchers from the Instituto de Geología Universidad Nacional Autónoma de México (UNAM), the Los Angeles Country Museum, the Universities of Washington and Rhode Island (USA), and some affiliates, have produced a significant assemblage of macro- and micro-terrestrial vertebrates, including dinosaur and mammalian material (e.g., Romo de Vivar Martínez, 2011; López-Conde et al., 2018). A floral assemblage, dominated by fossil fruits, is also preserved (Hayes et al., 2018), along with an extensive undescribed palynoflora.

These Late Cretaceous fossil localities from the west coast of North America contain the southernmost and westernmost fossil vertebrates known from the Campanian of North America, and thus provide a biogeographically unique datum for comparison with better known, contemporaneous North American Campanian terrestrial biotas to the north and east. Here, we situate this fossil assemblage in its paleoenvironmental, taphonomic, and stratigraphic contexts. We present evidence that this portion of the El Disecado Member, here termed the Fossiliferous Terrestrial Sedimentary Sequence (FTS), represents a physically active and biologically productive distal terrestrial fan deposit, built by a sandy low-sinuosity fluvial system forming in a cyclical wet/dry climatic setting. Our new high-precision U-Pb zircon age from this member improves upon its legacy 40Ar/39Ar geochronology and helps place its rich fossil fauna in its regional biostratigraphic context.

To enrich our understanding of the context of the vertebrates, we analyzed stable isotopic ratios (18O and 13C) retrieved from both the teeth of hadrosaurs (dinosaurs) and from paleosol carbonates. These ratios allowed for comparison with approximately coeval isotopic data from much better known Campanian localities in the United States and Canada, and the isotopic ratios recovered from the El Disecado Member fit well into latitudinal isotopic trends observed in those regions.

Setting, Geological Context, and Stratigraphic Relationships of the El Disecado Member of the El Gallo Formation

The El Gallo Formation outcrops in the modern Peninsular Ranges, a low suite of mountains that lines the western part of Baja California. These mountains formed in a forearc basin, the result of arc magmatism driven by Late Cretaceous–early Paleocene subduction in northwestern Baja California (Busby et al., 1998, 2002; Busby, 2004). Busby et al. (1998) described strike-slip basin formation, tilting, and associated uplift to produce a thick sequence of outcrops including those described here.

Kilmer (1963) first described the El Gallo Formation as a ∼1300-m-thick sequence of terrestrial sedimentary rocks composed of three members (in ascending order): La Escarpa, El Disecado, and El Castillo, the last of which is encompassed within the El Disecado Member and restricted to the eastern side of the basin (Figs. 1 and 2).

Fulford and Busby (1993, p. 302) characterized the El Gallo Formation as a “retrogradational sequence of alluvial plain through floodplain deposits capped by tidal deposits” within a forearc basin. They recognized four “successions,” of which Successions 2–4 constitute the El Disecado Member of Kilmer (1963). The FTS is correlative with the uppermost part of Succession 2 and most of Succession 3. It constitutes ∼130 m, running from local sea level upsection to the stratigraphic level in which the El Disecado Member begins to produce marine fossils. Previous 40Ar/39Ar geochronology (Renne et al., 1991) determined a Campanian (Late Cretaceous) age for the El Gallo Formation and its vertebrate fauna (Fig. 2).

METHODS

U-Pb Geochronology

Tuff layers preserved in the FTS were sampled for U-Pb geochronology by chemical abrasion–isotope dilution–thermal ionization mass spectrometry (CA-ID-TIMS), following the general procedures described in Ramezani et al. (2011). The sample was disaggregated by standard crushing methods and heavy-mineral concentrates were obtained using a Frantz® magnetic separator and high-density liquids. Zircons were hand selected under a binocular microscope and pretreated by a chemical abrasion technique modified after Mattinson (2005); this technique involved thermal annealing in a furnace at 900 °C for 60 h, followed by partial dissolution in 29M HF at 210 °C in high-pressure vessels for 12 h. The chemically abraded grains were fluxed in and rinsed with several hundred microliters of dilute HNO3, 6M HCl and Millipore water, successively on the hot plate and in an ultrasonic bath, to remove the leachates.

Thoroughly rinsed zircon grains were spiked with the EARTHTIME ET535 mixed 205Pb-233U-235U isotopic tracer (Condon et al., 2015; McLean et al., 2015) prior to complete dissolution in 29M HF at 210 °C for 48 h and subsequent Pb and U purification via an HCl-based anion-exchange column chemistry (Krogh, 1973). Pb and U were loaded together onto single outgassed Re filaments along with a silica-gel emitter solution, and their isotopic ratios were measured on an Isotopx X62 multi-collector thermal ionization mass spectrometer equipped with a Daly photomultiplier ion-counting system at Massachusetts Institute of Technology. Pb isotopes were measured as mono-atomic ions in a peak-hopping mode on the ion counter, whereas U isotopes were measured as dioxide ions in a static mode using three Faraday collectors. A predicted sample 238U/235U ratio of 137.818 ± 0.045 (Hiess et al., 2012) was used in age calculation, along with an oxide correction based on an independently determined 18O/16O ratio of 0.00205 ± 0.00005.

Data reduction, calculation of dates, and propagation of uncertainties used the Tripoli and ET_Redux applications and algorithms (Bowring et al., 2011; McLean et al., 2011). The individual 206Pb/238U dates were corrected for initial 230Th disequilibrium based on a magma Th/U ratio of 2.8 ± 1.0 (2σ). The sample age was calculated based on the weighted-mean 206Pb/238U date of all of the analyses without any exclusions and reported with 95% confidence level uncertainties including both internal (analytical) uncertainty and total propagated uncertainty (in brackets), which included the decay constant uncertainties of Jaffey et al. (1971). The total uncertainty must be taken into account when dates from different chronometers (e.g., U-Pb of this study versus 40Ar/39Ar of Renne et al., 1991) are compared.

Sedimentary Geology

Paleoenvironments of the FTS were reconstructed via a conventional sedimentological approach (Graham, 1988), using a combination of 21 stratigraphic sections measured with a Jacob’s staff at key vertebrate-producing localities, observation of sandbody geometries, lithologies and sedimentary structures preserved in outcrop, measurement of cross-stratification orientations, and hand samples collected for laboratory analysis. From these data, we reconstructed minimum bedform sizes, from which we then deduced the intensities and directions of ancient flows.

GPS coordinates were recorded for all localities. These coordinates were mapped in Google Earth Pro (v. 7.3.2.5776–64 bit) and were then transferred to a 1:50,000 topographic map, “El Rosario de Arriba,” published by the Mexican governmental cartographic authority INEGI (www.inegi.gob.mx; 2003). This map served as the base map for all of the work reported here.

For grain-size determination, samples were soaked overnight in a 5% Calgon deionized water solution to deflocculate clays, and then were analyzed on a Malvern “Mastersizer 3000” laser-diffraction grain-size analyzer. Petrographic analysis was carried out using 30μ thin sections studied under an Olympus BH-2 petrographic microscope with an Olympus DP-73 camera attached.

Apparent paleocurrent flow directions were measured with a leveled Brunton compass, oriented to obtain the direction of the maximum dips of cross stratification as exposed in three dimensions. Ancient flow directions were ultimately reconstructed by (1) correction of initial trend readings for declination (10.5° east); (2) removal of structural dip (e.g., the trend measurements were rotated to flatness using Stereonet v. 9.9.6 for Mac), and (3) 8.6° counterclockwise rotation of modified dip orientations, undoing the clockwise rotation they underwent in the past 12 m.y. associated with the active period of the opening of the Gulf of California (DeMets and Merkouriev, 2016). Rose diagrams and statistics were generated using MatLab; statistical calculations are after Graham (1988).

Vertebrate Fossil Collection

Microvertebrates, the focus of the paleontological work reported here, were obtained by close surface prospecting and, when bone (generally, tiny fragments) was found, collecting sedimentary rock at the site in large bags for screen washing. Each of these collections of rock constitutes a locality (see “Vertebrate assemblages” below), and we returned annually to the most productive of them for more rock matrix (and microfossil) collection.

By far the greatest preponderance of microfossil retrieval was obtained by screen washing. Screen washing was performed by soaking the rock matrix overnight in water, followed by screening. The concentrate was then handpicked under a magnifying lens. On average, a single vertebrate microfossil is produced from every 23 kg of bulk-collected FTS matrix.

Stable Isotopes

Hadrosaur teeth were collected from a single FTS microfossil locality, ROS-50. Paleosol carbonate nodules were collected from two localities, “Silvinia” and “Granito de Mostaza” (all locality data available at the Instituto de Geología, UNAM). Tooth enamel, tooth dentine, and pedogenic carbonate samples weighing ∼1 mg were extracted using a diamond-tipped Dremel drill. Samples were then soaked in 0.1 M acetic acid for 24 h, rinsed three times with deionized water and then dried for 24 h. This removed any secondary authigenic carbonate.

Carbon- and oxygen-isotope ratios were obtained using an automated carbonate preparation device (KIEL-III) and a Finnegan MAT 252 isotope ratio mass spectrometer at the University of Arizona in Tucson. The isotope ratio measurement is calibrated based on repeated measurements of National Bureau of Standards (NBS)-19 (δ18O = –2.20‰, δ13C = +1.95‰), NBS-18 (δ18O = –23.20‰, δ13C = – 5.01‰), and CAR-1, an in-house carbonate standard (δ18O = –1.41‰, δ13C = +2.03‰). Analytical precision is ±0.1‰ for both carbon and oxygen. Isotope ratios are reported using the “δ” notation with δ13C and δ18O referring to (Rsample/Rstandard − 1) × 1000‰, where R is the ratio of the heavy isotope to the light isotope for carbon and oxygen, respectively. The standard for carbon is Vienna Pee Dee Belemnite (VPDB), and the standard for oxygen is Vienna Standard Mean Ocean Water (VSMOW).

RESULTS AND INTERPRETATIONS

Depositional Age

A suite of four closely spaced (∼50 cm) indurated ash tuffs, recognizable by phenoclasts of plagioclase, quartz, biotite, and other minerals in a clay-rich matrix, was present in an outcrop near the base of the FTS (Fig. 2). Each layer was 3–10 cm thick and preserved within gray siltstones with few distinguishing characteristics (Facies 3). The surficial ash layers we sampled for this study presented, when weathered at the surface, a dried white (7.5PB 9/2; Munsell Soil Color Charts, 1992) appearance with fine sand-sized mafic (biotite?) specks, and when freshly exhumed, a moist greenish gray (5G 5/1 gley; Munsell Soil Color Charts, 1992) aspect. An absence of sedimentary structures and the abundance of clays suggest that these layers represent (primary) airfall tuffs, likely to have been originally composed of a glass-rich matrix.

Complete Pb and U isotopic data are given in Table 1, and the age results are illustrated on the age plots of Figure 3. The stratigraphically highest layer of the tuff suite was sampled for U-Pb zircon geochronology by the cathodoluminescence–isotope dilution–thermal ionization mass spectrometry (CA-ID-TIMS) method (see Methods). Four single zircons analyzed from this sample all overlap in age within uncertainty and yield a weighted-mean 206Pb/238U date of 74.706 ± 0.028 (0.091) Ma with a mean square of weighted deviates (MSWD) of 1.4 (Fig. 3). The latter date represents the best estimate for the Campanian deposition of the corresponding strata.

Renne et al. (1991) reported 40Ar/39Ar geochronology based on laser-fusion analyses of single sanidine grains from four widely spaced tuff horizons throughout the El Gallo Formation, producing dates that ranged from 74.87 Ma to 73.59 Ma with 2σ analytical uncertainties of ± 0.10 to ± 0.18 m.y. Although this level of age precision (∼0.2%) was groundbreaking for its time, later improvements to the K decay constants (e.g., Min et al., 2000) and the age of the Fish Canyon sanidine fluence monitor (e.g., Kuiper et al., 2008) necessitate revisions to the above dates that exceed the reported uncertainties. Samples EG-48 and EG-42 with recalculated 40Ar/39Ar ages (2σ) of 75.17 ± 0.30 (2.96) Ma and 74.55 ± 0.18 (2.93) Ma, respectively, fall within the FTS part of the El Disecado Member (Figs. 1 and 2). Based on previously published map illustrations, we infer that our U-Pb tuff sample and sample EG-48 of Renne et al. (1991 = T3 tuff of Fulford and Busby, 1993) belong to the same stratigraphic level, notwithstanding a lack of precise sample location (e.g., GPS coordinates) for these legacy ages. Considering the large total propagated uncertainties of the 40Ar/39Ar dates due to large errors associated with the K decay constants, all three dates tend to overlap within total uncertainty.

Sedimentary Geology

Outcrops associated with the FTS are repetitively yellow- and gray-banded, corresponding to distinct facies (Fig. 4). These are all tilted uniformly throughout the full exposure of the FTS, with strikes of ∼120° and dips of ∼14° NE, presumably reflecting the active tectonic history of this unit. Directly adjacent to the Pacific Ocean, there also occur flat-lying, plane-bedded deposits superimposed upon the tilted deposits; we interpret the former to represent Quaternary or Holocene deposition, and thus these deposits do not form a part of this study.

The thicknesses of units within the FTS vary but form a semi-regular interbedded pattern. The contacts between the different colored beds are sharp. In these exposures, we recognize three important facies: (1) a medium- to coarse-grained, iron-stained (yellow), cross-stratified sandstone (“La Amarilla”), which unsurprisingly corresponds to yellow bands in outcrop; (2) a very fine- to fine-grained, gray mud-rich siltstone with variably sized intraclasts (the “GF facies“); and (3) a silt- to very fine-grained gray sandstone with weakly to moderately well-developed pedogenic features including carbonate nodules (the “paleosol facies”). Figure 5 shows some columnar sections measured at several localities; these sections exemplify sedimentary deposition in the FTS.

Facies 1 (La Amarilla)

Description.

La Amarilla is an iron-stained, yellow, texturally immature sandstone facies whose medium sand-sized (∼0.20–0.28 mm) framework constituents, though moderately well sorted (Fig. 6), are angular (Fig. 7). Compositionally this facies also is immature, including angular clasts of mono- and crypto-crystalline (chert) quartz, plagioclase, K-feldspar, biotite mica, a variety of clay-rich lithic fragments comprising the framework constituents. These are generally concentrated in laminae best measured on decimeter scales or occur as thin linings of concave-upward scour surfaces. The rock is weakly, but pervasively, cemented by clay and calcite cements. Petrographically, the rock suggests a mixed sedimentary and igneous provenance (Fig. 7). The textural and compositional immaturity, the presence of fresh feldspars and micas, in concert with the presence of sedimentary lithic fragments indicate a probable volcanic source with a strong sedimentary influence (such as perhaps a volcanic origin and transport through an eroding southern, sedimentary source). In the scheme of Dickinson and Suczek (1979), this lithology would be classified as having a “magmatic arc” provenance.

Facies 1 exhibits variable thicknesses on a multi-meter scale, ranging from ∼1 m to as much as 10 m; however, many of these deposits are multi-storeyed. Individual storeys can be identified by the superposition of channeliform sandstone bodies (Fig. 8A) within the beds (Bridge and Tye, 2000). Although these channels can be stacked and upper channels erode into lower ones (e.g., they are “amalgamated-truncated,” sensu Hajek and Heller, 2012), we estimate that individual storeys range from ∼1–2 m. Contacts with subjacent units are abrupt and are marked by convex-down (channeliform) shapes (Fig. 9).

The arcuate bases of the cross strata are commonly lined with coarser materials, including chert pebbles, mudstone intraclasts (some as large as 25 cm along the longest axis), petrified tree trunks (up to 6 m), and disarticulated dinosaur (hadrosaur) long bones. All of these features are associated (and commonly aligned) with the features of the sedimentary structures, indicating that they were being carried as intraclasts in the current (Fig. 8B). The width of the channeliform bodies varies; however, in general, these extend 20–40 m.

Internally, La Amarilla is characterized by small-, medium- and large-scale, trough cross stratification throughout (Fig. 8C). In general, the largest scale features in each bed are located basally, and the proportion of smaller-scale cross-stratification increases higher in each storey. We measured the trends of the cross-stratification dip directions to reconstruct the ancient paleoflow direction. There is significant variation, but overall they show a dominantly NNW flow (Fig. 10; see below).

Interpretation.

We interpret the cross stratification to represent a variety of large-scale sandy bars and smaller, sandy, subaqueous dunes, all migrating under strong unidirectional flows (Leclair and Bridge, 2001; Bridge, 2003). Based upon a rough estimate of the average size of the dune trough cross stratification (∼12 cm), dune heights may be inferred to range between 30 and 60 cm, using the method of Leclair and Bridge (2001). From these calculations, water depths of ∼3–5 m can be inferred (Bridge and Tye, 2000). These rough relationships allow us to calculate flow intensities in the 100–200 cm/sec range, obtained from grain size and the inferred sizes of bedforms and flow depths (Harms et al., 1982; see also Rubin and McCulloch, 1980 and Bhattacharya et al., 2016). Finally, using the formulation of Lynds et al. (2014), we obtained a slope of 0.00008 for the FTS, an extremely gentle slope of 0.8 m/km. This calculation, like the others, is a rough approximation; however, it gives a general sense of prevailing sedimentary conditions when the FTS part of the El Disecado Member was active.

The shapes of the sedimentary bodies indicate aqueous transport in erosive channels; large mudstone intraclasts and sharp contacts cutting subjacent units reveal the erosion. These channels would have avulsed and/or migrated episodically within a broad floodplain swath, each depositional event beginning with the highest flow intensities, and waning, as reflected by the decrease in the size of the grain sizes and cross stratification upwards in each bed. Distinct channeliform bodies within this facies (Fig. 8A) suggest that channel migration occurred as much by avulsion as by lateral erosion (see Collinson, 1996; Mohrig et al., 2000).

Sandy bars evidently developed, as indicated by bar-scale cross-stratification (Fig. 8C). However, stabilized (vegetated) bar sedimentary features reported in many braided systems are not preserved (Boothroyd and Ashley, 1975; Miall, 1977, 1978, 2006). This is not unexpected; full channel bodies (surrounded by mudstones) are rarely preserved; instead, most channels and bars have been partially eroded by younger channels and, following the characterization of Hajek and Heller (2012), would be classified as amalgamated to truncated. Taken in total, the evidence indicates that this was an active, episodic, sandy, low-sinuosity channelized fluvial system (Bridge and Tye, 2000; Bridge, 2003).

Facies 2 (GF)

Description.

Facies GF is neither volumetrically nor visually distinctive in the FTS; however, it produces microvertebrates (fossils generally but not exclusively <∼3 mm on the long axis) and is the most fossiliferous facies in the entire unit; for this reason, and because it bears upon the depositional setting, it commands our attention (Fig. 11A). Facies 2 is near black when moist (N4/ “dark gray”; Munsell Soil Color Charts, 1992); a clay-cemented, silt-to very fine sandstone (Fig. 6). Units are generally 1–2 m in thickness and locally lenticular. Its framework constituents are dominated by silt-sized angular clasts; however, chaotically distributed among these are subrounded to subangular lithic intraclasts ranging from very coarse sand- to granule-sized. These occupy <1% of the sample. The silt- and very fine sand-sized framework fraction of this siltstone consists of quartz, plagioclase, K-feldspar, biotite, and minor muscovite (Fig. 11B). Sedimentary structures are generally not visible in this facies; moreover, the large chert intraclasts are chaotically distributed in the siltstone matrix, in striking contrast to the well-sorted, well-developed, large-scale sedimentary structures visible in Facies 1.

Clays constitute some 20%–30% of this facies (Fig. 6). X-ray diffraction analysis in the laboratory of Dra. T. Pi (Instituto de Geología, UNAM) suggests that the dominant clay species in Facies GF is illite (see Supplemental Material1). These clays can appear aligned around the framework clasts, suggesting diagenetic (including pedogenic) alteration and/or recrystallization; ultimately, the contributions of diagenetic and syndepositional clays are impossible to tease apart. However, it is likely that a considerable portion of the clay-sized matrix, only part of which is actually composed of clay minerals, was deposited syndepositionally.

The fossils in Facies 2 (GF) are generally in the 1–3 mm range (although identifiable ones average ∼15 mm), and are commonly (but not exclusively) found nearly or completely disarticulated. Microvertebrate remains dominate the collections (Romo de Vivar Martínez, 2011; García-Alcántara, 2016; see below); however, tiny (∼3 mm, aperture notch to apex) gastropod shells are also present.

Interpretation.

The absence of sedimentary structures or other evidence of internal organization in Facies GF suggests non-Newtonian processes in the form of sediment gravity flows (e.g., Lowe, 1979; Miall, 2006), an interpretation concordant with the high percentages of silt- and clay-sized material in this facies (Fig. 6). The most likely type of flow, suggested by the grain sizes involved, would be hyperconcentrated sheet flows (Pierson and Costa, 1987), in which flow intensities reaching 10 m/sec have been recorded. Sheet flows are concordant with the relatively thin outcrops of this facies. High flow intensities are suggested by the disarticulated, broadly distributed microfossil content of Facies GF. This facies therefore represents a non-channelized, dense, clay- and silt-rich sediment slurry deposited in dramatic density contrast with the air, as a shallow sediment gravity flow at relatively high flow velocities (sensu Fraser and Suttner, 1986; Miall, 2006). Facies 2 (GF) may correspond to Fulford and Busby’s (1993) “laminated sand sheet floods,” although we did not observe laminations in this facies.

Facies 3 (Paleosols)

Description.

Facies 3, which occupies the preponderant portion of the FTS, is composed of a fine-grained suite of deposits showing pedogenic color banding (Figs. 4 and 12A). These units are dominated by clay- and silt-sized fractions. The remainder of the sedimentary rock consists of very fine sand-sized material, including quartz, plagioclase, K-feldspar, and sedimentary lithic fragments. These grains have the same maturity—both textual and compositional—as those in Facies 2.

Horizonation within this facies is commonly distinguished by color or textural changes—and ranges from 0.5–3 m; however, color banding within the bedding, weakly visible on the exterior weathered surface of the outcrop, generally occurs on a smaller scale, ranging from 0.5–1.5 m. The colors include drab grays, reds, and yellows (N5/; 4GY 4/1; 5GY 5/1; 10R 4/6; 2.5Y 6/6; Munsell Soil Color Charts, 1992). The color bands are produced by distinctive ferric oxyhydroxide mottling, best seen in fresh rock samples (Fig. 12B).

Also visible in some—but not all—outcrops are distinctive aggregated blocks of material best interpreted as soil structure (peds). Ped types present include blocky, subangular blocky, and some weakly planar features, likely platy peds (Retallack, 2001). In thin section, the peds reveal themselves to be aggregates of material with well-developed clayskins (cutans), suggesting illuviation in a B-horizon setting (Fig. 12C). Here, as in Facies 2, the dominant clay species is illite.

Most distinctive in Facies 3—but not necessarily characterizing all of it—are carbonate nodules. These range in diameter from several cm to ∼1 mm. They are generally aligned vertically in a weakly linear way, with some coarsest sizes associated with the upper contacts of bedding, and successively smaller sizes descending (Fig. 12D). They commonly preserve a central core, likely to have been the root around which they developed. They are distinct from carbonate concretions, also preserved in the FTS, which tend to be larger, less spherical, are horizontally oriented (commonly, in association with textural discontinuities between beds), and lack the vertical trending quality clearly visible in the descending associations of nodules. With calcite evidently mobilized diagenetically as cements (see above), these concretions are interpreted here as postdepositional, and thus unconnected with the active pedogenesis in this paleoenvironment.

Facies 3 exhibits evidence of significant bioturbation in some localities, such that all primary sedimentary features have been obliterated (Fig. 13A). The burrows are not large, generally between 2 and 5 mm in diameter, and traceable as far as 6 cm. Trace fossil burrows are common; they are not uniformly vertically oriented; many contain at least a portion that is subhorizontal. These burrows are superficially similar to those produced by insects in alluvial environments (Hasiotis and Bown, 1992); however, truly diagnostic features are lacking. Larger, vertically oriented conically shaped burrows are preserved as well; some of these are as much as 20 cm long (deep) and 6 cm in diameter (Fig. 13B). These, however, are preserved exclusively in Facies 1, and compare favorably burrows of the crayfish ichnogenus Camborygina sp. (Hasiotis and Mitchell, 1993).

Interpretation.

Facies 3 is a carbonate nodule-bearing paleosol facies. The color banding is best interpreted as the remnants of pedogenic subsurface horizons, where the mottling (Fig. 12B) is associated with a fluctuating water table, producing changing Eh and pH conditions that ultimately control the stability of dissolved ferric oxyhydroxides (Buol et al., 1980). Using the classification from the Soil Survey Staff (1975), the mottles would be classified as “many,” of medium size, and “distinct” in contrast; however, FTS diagenesis has clearly modified these features from their original condition in the active soil.

The presence of carbonate (calcite) nodules accords with a fluctuating water table perhaps due to limited rainfall; in the vertically oriented suites of nodules, translocated solubilized carbonates were likely associated with the movement of materials through the soil in and out of the roots, but during drier times, rainfall would have been insufficient to solubilize and remove the carbonate ions, which then would precipitate and accumulate as calcite (see Retallack, 2001).

A distinctive feature of the FTS paleosols is that they preserve the clear imprint of compound pedogenesis (e.g., Duchafour, 1991) such that superficial horizons, such as O and A horizons, are almost never recognizable. In most cases, all definitive traces of them have been removed, either through erosion, oxidization, or through re-purposing as subsurface horizons. What remains are thick (multi-meter) sequences of rooted, ped-bearing, mottled Bt horizons. Ultimately, compound pedogenesis and the absence of preserved diagnostic criteria hinder classification of these paleosols (Mora et al., 1993). Here, therefore, the features that we observe allow us to recognize some key processes these soils, which in turn allow us to infer aspects of the paleoenvironments in which they formed (see Retallack, 2001):

  • (1) Extensive mottling suggests the presence of fluctuating moisture (and, likely, Eh and pH) conditions in the soil subsurface.

  • (2) Peds, horizonation, cutans, and carbonate nodules, dominate the paleosols; these suggest active soil processes (in particular, illuviation) occurring in well-developed soils.

  • (3) Therefore, by contrast to Facies 1 and 2, Facies 3 represents zones of relative landscape stability in an otherwise episodically depositional setting.

Other Sedimentary Features

Although Faces 1 and 3 dominate the sedimentation of the FTS, several other sedimentary features are noteworthy as well. Plant material is found in several settings. In La Amarilla, large pieces of silicified wood, syndepositionally unpetrified, were carried as clasts within the active channels (Fig. 8D; see description of Facies 1 above). Unsilicified plant material is also present in fine-grained settings preserving small, floodplain ponds, identifiable by distinctive organic-rich (dark-colored), fine-grained, thin-bedded (<0.5 m), laminated silty mudstones associated with paleosol deposits. In such settings, delicate roots, leaves, and stems are preserved (Fig. 5A at the 12.03 mark; Fig. 14); we interpret these to be highly localized, low-energy floodplain ephemeral(?) pond deposits.

Pollen is abundant throughout the FTS. It is commonly preserved in darker, finer-grained deposits such as those associated with Facies 3; however, it is not generally a component of Facies 2. A diverse palynomorph assemblage is preserved in the FTS; however, describing it is beyond the scope of this study.

Finally, several plant localities exist, dominated largely by ferric oxyhydroxide-replaced angiosperm fruits (O. Cornales: Operculifructus; Hayes et al., 2018). Some of these localities consist of lenticular deposits of fine-grained laminated silt or fine sandstones postdepositionally ferric oxyhydroxide-cemented; others are clearly developed in Facies 3 (paleosols) as indicated by iron mottling of the siltstones. Iron-replaced roots, carbonate nodules, and iron-cemented horizons are commonly associated with these localities. The presence of well-developed soil features associated with these concentrations of plant material suggests that at these localities, at least, the plant materials are autochthonous.

Basal Lithology—“Ambar”

Description.

As noted above, the most basal exposures in this study here begin in, and just landward of the Pacific Ocean. These consist of a sequence we dubbed “Ambar,” which lies conformably beneath the FTS. Ambar is composed of interbedded cross-stratified sandstones and mudstones, with extensive burrowing (Fig. 15). These are tubular in shape, ∼1 cm in diameter, and up to 15 cm in depth. Backfilling spreite are clearly visible. The backfilled burrows are relatively densely packed, with each burrow separated by a few cm. These burrows are comparable to Skolithos, a well-known Phanerozoic marginal-marine trace fossil morphotype attributed to vermiform animals such as polychaetes (see Dodd and Stanton, 1990; Desjardins et al., 2010; Herringshaw et al., 2010; Vinn and Wilson, 2013). Other types of burrows are also present; these are narrower and include a horizontal component as well as a vertical one.

Interpretation.

“Ambar” bears a suite of facies—largely, bioturbated, organic-rich sandstones and mudstones—suggesting biotic activity, but it does not present any of the three major facies described above, nor does it contain any of the indicators of terrestriality (e.g., paleosols and terrestrial vertebrate fossils; see below). It does, however, contain palynomorphs, charcoal, and some small amber deposits, indicating proximity to terrestrial environments.

Limited outcrop exposure constrains the certitude with which the sedimentary environments represented at this locality can be identified here; however, an interpretation that is concordant with the features it presents is that of a marginal marine setting, such as a tidal flat. Among the observations in agreement with this interpretation are (a) thick sequences of fine-grained, organic-rich mudstones; (b) the extensive burrowing of a limited number (only two types) of trace fossil, reflective of high-abundance, low-diversity paleoecology characteristic of similar stressed environments; (c) channelized fine-grained sandstones, which in this setting would represent tidal channels; (d) extensive bioturbation, of typically marginal marine aspect (Rhoads, 1975).

Vertebrate Assemblages

The immediate impetus for understanding the paleoenvironmental setting of these deposits was the vertebrate fossil faunas that have been recovered from them. The most productive facies are La Amarilla and GF. La Amarilla produces relatively large dinosaur bones; most commonly, disarticulated limb bones and isolated vertebrae of adult hadrosaurs, although other substantial bones, such as a variety of turtle, have been found. Collections of large vertebrates were also made during the 1960s by paleontologists from the Los Angeles County Museum (Morris, 1973, 1981). Vertebrates notwithstanding, the most common fossil in La Amarilla is petrified wood. We did not collect La Amarilla systematically for fossil content, and thus our observations here about La Amarilla are anecdotal.

The most productive facies for microvertebrates is GF. To date, 87 fossil localities have been recorded in the FTS (Fig. 1). Since 2004, GF has each year been systematically screen washed. The microvertebrates GF produces are generally very small (1–2 mm; taxonomically identifiable fragments are less common, but average ∼15 mm), disarticulated, and commonly fragmentary. There are rare exceptions to this, including the semi-articulated remains of the squamate Dicothodon (Montellano-Ballesteros et al., 2005; Chavarría-Arellano et al., 2018a, 2018b), the tooth-bearing jaws of several vertebrates, and a near-complete multituberculate (mammal) skull (Wilson et al., 2011). However, although GF cannot be said to be rich in microvertebrates, after 14 years of collecting (as of this writing), a substantial microvertebrate assemblage has emerged, comprising some 1600 individual specimens (see Table 2; Fig. 16). This includes rays, at least two families of teleosts, amphibians including frogs and salamanders (Romo de Vivar Martínez et al., 2016), at least five different taxa of turtles (López-Conde et al., 2018) including the characteristic Naomichelys speciosa, plus Basilemys sp., Compsemys victa, Trionychidae indet., and cf. Chelydridae, at least four families of lizards (Chavarría-Arellano et al., 2018a, 2018b), crocodylomorphs, several families of theropod and ornithischian dinosaurs, and multituberculate and eutherian mammals (Romo de Vivar Martínez, 2011; García-Alcántara, 2016).

Finally, the FTS contains a few potentially in situ vertebrate fossils along with the preponderance of allochthonous specimens. Several interesting, and semi-articulated fossil specimens—notably a partial and fragmented baby hadrosaur and eggshell fragments—have been found at contacts between GF and the paleosol facies (Cabrera-Hernández, 2018; Cabrera-Hernández et al., 2018). These fossils breathe promise of nesting in the ancient FTS distal fan, a possibility that requires further exploration.

Stable Isotopes

Carbon- and oxygen-isotope data for hadrosaurid tooth enamel and dentine and from paleosol carbonate nodules from the FTS are reported in Table 3.

Two methods demonstrate that our results are primary: (1) an isotopic comparison of dentine and enamel from the same element; (2) an isotopic comparison of enamel carbonate and authigenic carbonate (i.e., paleosol nodules).

These comparisons rest on the assumption that the isotopic characteristics of dentine and paleosol nodules represent the burial and/or diagenetic environment, and if the isotopic characteristics of tooth materials are different, then the tooth enamel is likely preserving primary biological information. Dentine is more susceptible to diagenesis when compared to tooth enamel because of its higher porosity and larger apatite surface area; pedogenic nodules are by definition authigenic. A clear difference in both δ13C and δ18O between tooth enamel and paleosol carbonate (Table 2), as well as a difference in δ13C between tooth enamel and tooth dentine (Table 2), permit the conclusion that primary isotopic information is preserved in tooth enamel carbonate.

With confidence, then, that our isotopic results are primary, we integrated the FTS results with those previously published from the more northern and eastern sites, which allowed us to investigate latitudinal patterns in stable isotopic ratios. Figure 18 shows that there is a negative correlation between latitude and δ18O and a positive correlation between latitude and δ13C for both tooth enamel and nodules, with the highest δ18O and lowest δ13C being associated with the Baja locality. δ13C values from co-occurring tooth enamel and nodules broadly overlap at every locality where they are found together. Similar overlap is also observed in δ18O for the Two Medicine and Kaiparowits Formations, but not in the FTS, where δ18O of nodules is significantly lower than that of tooth enamel.

DISCUSSION

Environment of Deposition

The FTS represents a fully terrestrial setting. Supporting this are (1) the presence of a diverse terrestrial vertebrate and plant assemblage with no evidence of marine influence (see below); (2) the presence of well-developed paleosols, which in life were unambiguously subaerial features of the landscape; and (3) the presence of a suite of sedimentary features commonly associated with fluvial environments.

Both Facies 1 and 2 are the products of high-energy, unstable landscapes. However, as we have seen, considerable evidence for landscape stability exists in the form of the well-developed sequences of paleosols forming with compound pedogenesis (Facies 3). Time of formation of soils is notoriously difficult to estimate; however, the presence of well-developed sequences of peds, horizons, cutans, and root-originated carbonate nodules suggests significant time intervals of landscape stability. We interpret Facies 3 as floodplain deposition, representing land surfaces that likely functioned on multi-century timescales.

Ancient Flow Directions in the FTS

Some of the paleocurrent data (Fig. 10) show a direction opposite to the dominant trends in the data (e.g., south- and east-directed). We interpret this as the presence of tidal influence in this setting. Directionally bimodal cross-stratification is most commonly associated with tidal successions (e.g., Boggs, 2001); here, however, we propose that strong tidal bores likely made their way up some of the channels, producing sigmoidal cross stratification with slipface mud-couplets, features associated with tidal deposits (Prothero and Schwab, 2004) and recorded in La Amarilla (Fig. 17). The existence of tidal bores in the braided channels may have contributed to the unusually broad range of paleocurrent trends recorded and reinforces the interpretation of the FTS as a marginal marine deposit.

Paleocurrent data collected for this study show a strong pattern of northward flow, with a modest westward component. Unsurprisingly, sedimentary structures representing the largest bedforms show the least variation. These data imply, however, that at least the proximal source for the FTS was to the south, but results do not precisely accord with those obtained from Fulford and Busby’s (1993) limited (n = 11) data set for this member, in which a NW trend was obtained, or with their larger-scale results for the entire El Gallo Formation, in which they recognized a broad (nearly 180°) spread of paleocurrents (from a much larger database; n = 264), with a west-directed mean and considerable circular dispersion. They interpreted their results as concordant with a broad braid plain draining into a Cretaceous proto–Pacific Ocean.

Several possible scenarios could explain the more northern (and less western) trends of the paleocurrents we obtained. Scenario 1 is that the FTS was not in its present orientation, and that it has rotated as much as 90° from its Cretaceous position. Nothing in the known regional deformation would suggest that this occurred. In Scenario 2, the fossiliferous outcrops studied here may reflect local geography, and whereas overall depositional trends were likely westward, here, a local, small-scale northward depositional feature is being recorded.

Geomorphic Setting: Distal Subaerial Fan?

Petrographically as well as geographically, the active arc magmatic setting of the FTS suggests that it represents the geomorphic and depositional interface between the magmatic arc and the proto–Pacific Ocean. Because this likely involved the emergence of a channel draining mountains in combination with evidence for episodic and sheet flooding (Bull, 1972), a subaerial fan setting is suggested. The sandstones of the FTS are petrographically and compositionally immature, suggesting relative proximity to the source, and the fine-grained and sedimentary aspect of the system is also potentially concordant with the relatively distal component of a subaerial fan system.

While Miall (2006) regards fans as geomorphically unique rather than embodying unique sedimentary facies, Blair and McPherson (1994) note that in process, bedforms, sedimentary deposits as well as geomorphology, fans differ dramatically from other kinds of terrestrial/fluvial deposits. Their work details a broad list of fan features, including piedmont setting, the emergence of piedmont channels to unconfined flow, transport along the fan by mass-wasting and fluvial processes, rare and episodic deposition, the presence of poorly sorted, angular gravels, steep flow gradients, a “plano-convex” cross section (perpendicular to fan growth), a decrease in slope with distance from the source, and an absence of well-developed floodplains. These features, they aver, make fans unique among terrestrial depositional environments, and they reject what they perceive as a tendency of certain sedimentologists to merge braided river deposits with alluvial fans.

Indeed, the FTS contains retrievable features not associated by Blair and McPherson (1994) with fan deposition, most notably the absence of gravels, the presence of fine-grained sedimentation organized in well-developed floodplain deposits, and extremely low slopes. These features, however, are in agreement with the outer-fan setting of the “Low Sinuosity/Meandering Fluvial Fan” model of Stanistreet and McCarthy (1993). In this model, distal depositional slopes are predicted to be low (∼0.0003), comparable to that reported here for the FTS (0.00008). The disarticulated microvertebrate fauna, almost exclusively confined to the sheet-flow deposits, and macrovertebrate faunal elements (largely, but not exclusively hadrosaur bones), carried in the Facies 1 channels along with large tree trunks, reinforce the suggestion of episodic, high-energy flooding in a relatively stabilized coastal floodplain setting.

In summary, distinctive features that unambiguously characterize fans primarily rooted in geomorphology and process (Blair and McPherson, 1994) are either unretrievable in the FTS deposits, or are distinctly muted in this inferred distal setting, although episodicity and unconfined (sheet flood) deposition are preserved. Thus while acknowledging that the fit of the FTS to ideal arid fan deposits (Blair and McPherson, 1994) is imperfect, we tentatively propose a distal fan environment, such as that suggested by Stanistreet and McCarthy (1993) as a potential paleoenvironmental setting for the FTS. A marginal marine relationship is suggested by the stratigraphic proximity of marginal marine deposits to the FTS sedimentary package.

Climate

Carbonate nodules and soil-matrix mottling seen in paleosol Bt horizons send a clear signal of wet/dry cyclicity in the soil subsurface; cyclicity that most likely derived from the climate in which these soils developed. We infer generally dry conditions from the stable isotopic results (see below); however, as noted above, there is evidence for laterally restricted ponding on the floodplains (above; see Figs. 5A [at the 12.03 mark] and 14).

Vertebrate Taphonomy

The fragmentary, disarticulated quality of the vertebrates, and their marked correspondence to facies—large fossils in La Amarilla and microvertebrates within GF—strongly suggests that these fossils were carried into this system as allochthonous clasts in approximate hydraulic equivalence with flow. In La Amarilla, fossils—both vertebrate and plant—are organized in clear relationship to the cross-stratification and are associated with the scour fills that form at the toes of migrating three-dimensional subaqueous dunes. In GF, the fossils approximately match the size of the largest clasts in the system, and like them are chaotically distributed within it. They therefore must represent an upper limit of the competence of these debris flows; however, because they are potentially less dense than the other clasts in GF, we cannot determine whether they are associated with the larger clasts in matrix support, or whether they are themselves part of the matrix, otherwise composed of smaller, denser clasts.

Although the vertebrate fossils are certainly not in situ (they are clasts carried in the sedimentary system), their relatively fresh (not water-worn) condition suggests minimal travel; that is, they may have lived in association on the distal fan environment described here; however, smaller clasts (including allochthonous fossils), generally, show less rounding than larger clasts in siliciclastic depositional systems. Microvertebrate remains, including parts of a semi-articulated postcranial elements and a semi-articulated skull of two perinatal hadrosaurs (Fig. 5A;Cabrera-Hernández, 2018 and Cabrera-Hernández et al., 2018), are known from paleosol (Facies 3) deposits in the FTS, suggesting that, by contrast, facies GF most likely concentrated the fossils, rather than carried them from a long distance away. If this is true, their disarticulated quality potentially results from surficial gathering of isolated bones on the fan braid plains during storm events.

Stable Isotopes

Fricke et al. (2008, 2009) compared stable isotopic data from North American, late Campanian, terrestrial, dinosaur-bearing formations, including the Two Medicine, Judith River, Dinosaur Park, Kaiparowits, and Fruitland Formations with the goal of studying dinosaur migration; from these, they reconstructed a latitudinal gradient in isotope ratios. Given the late Campanian age of the FTS, as well as the fact that oceanic carbon isotope ratios varied by only ∼0.25‰ and temperature varied by only ∼1–2 °C during the late Campanian (Barrera and Savin, 1999), integrating FTS data into those from the Western Interior of the United States is a natural and logical step.

Spatial Patterns in δ13C and δ18O

The lack of isotopic overlap between samples from the FTS and other Campanian localities (Table 3; Fig. 18) is consistent with the FTS hadrosaurids being ecologically distinct from the other records reported in Fricke et al. (2009, 2015), an expected result given the location of the FTS on the Pacific side of the Cordillera. These differences stem from the δ13C of vegetation they were eating and δ18O of water they were drinking.

δ13C. One of the most remarkable aspects of the carbon isotope data presented in Table 3 and Figure 18 is the wide range in δ13C values of hadrosaur tooth enamel, and by extension, the wide range in δ13C values of Late Cretaceous vegetation. Assuming a δ13C offset between diet and tooth enamel of 19‰ for hadrosaurs (Fricke, 2007), average values for vegetation range from ∼22‰ in coastal settings in Montana and Alberta to ∼25‰ in more upland environments of Montana and Utah, to ∼29‰ in Baja. Taken together, this range in values is similar to that observed for modern C3 plants (e.g., O’Leary et al., 1992) and suggests that both now and in the past, the same factors operated to produce a range in δ13C values of modern C3 plants. Such factors include: (1) environmental variables that influence water availability to a plant, such as rainfall, humidity, salinity, and location in the canopy (e.g., sun versus shade); (2) the δ13C of the atmosphere, and (3) taxon-specific differences in fractionation during the process of photosynthesis (e.g., O’Leary et al., 1992).

Environmental variables such as the amount of precipitation are generally considered to play a dominant role in influencing plant δ13C, and several relations have been developed between precipitation and δ13C, with higher δ13C corresponding to drier conditions (e.g., Diefendorf et al., 2010). Such relations would suggest that El Gallo landscapes were wetter than those farther north; however, this interpretation is inconsistent with sedimentological data. The Judith River, Dinosaur Park, and Fruitland Formations contain coal and lignite beds, whereas the Kaiparowits and Two Medicine Formations include lake deposits; by contrast, the FTS shows evidence of wet/dry cyclicity and significant drainage in its paleosols. All of these observations suggest that the more northern localities were wetter than the El Gallo.

It is possible, therefore, that the δ13C of the atmosphere under some forest canopies differed from that of the open atmosphere due to inputs of soil-respired CO2. This canopy effect is associated with dense, closed canopies, and has been observed to produce leaves with unusually low δ13C (e.g., van der Merwe and Medina, 1991). For this factor to play a dominant role in explaining the spatial patterns in Figure 18, FTS forests would have to have had the most recycling of soil-respired CO2, and thus be the densest. Again, it is hard to reconcile this interpretation with the sedimentological evidence, which suggests that FTS landscapes were dominated by braided fluvial systems that were prone to episodic sedimentation associated with high flow intensities. In addition, organic matter is less common in the FTS than in its northern and eastern counterparts, suggesting a lack of vegetation that contrasts with the organic-rich deposits to the north.

A third, and perhaps the most likely, explanation for the wide range in δ13C is that different plant types dominated different landscapes. In particular, it is observed today that angiosperm plants typically have δ13C values that are 1‰–2‰ lower than that of gymnosperm plants (e.g., Heaton, 1999), thus a larger proportion of angiosperms to gymnosperms in Baja relative to other localities could explain the lower δ13C values observed there (Fig. 18). It is tantalizing, but hardly conclusive, that—aside from indeterminate petrified wood—the most common plants preserved in the FTS are angiosperm fruits (Hayes et al., 2018). Moreover, because angiosperms are thought to have evolved and undergone extensive radiation from low-to-high latitudes during the mid- to Late Cretaceous (Wing and Boucher, 1998), and it has been posited that they were able to do so because of their ability to occupy disturbed habitats (Wing and Boucher, 1998). FTS habitats may have been populated by a greater proportion of angiosperm taxa than were their counterparts to the north and east. FTS landscapes, located to the south of the others and dominated by avulsing river systems and episodic debris flows, could have been occupied by a flora assemblage much different than those localities to the north.

δ18O. Considering all the hadrosaur δ18O data together, a gradual decrease in δ18O with latitude is observed (Fig. 18). Such patterns in δ18O of precipitation are observed today, and at global scales are attributed to the gradual rainout of moisture from airmasses as they move from low to high latitudes (e.g., Rozanski et al., 1993). Precipitation patterns over western North America during the Campanian were, however, complicated by the existence of the Sevier highlands and of the Western Interior Seaway (e.g., Sewall and Fricke, 2013). Therefore regional-scale patterns in atmospheric movement and rainout may also account for some of the difference in δ18O observed in Figure 18. Indeed, Baja California del Sur was located on the western, windward side of the Cordillera, whereas all other sites were located leeward of the highlands and on the coast of the Western Interior Seaway. As a result, Baja California was dominated by Pacific-sourced airmasses that would have experienced little rainout prior to landfall such that δ18O of precipitation remained high, whereas localities to the north and east were more likely to have precipitation sourced from both Pacific and Western Interior Sea–sourced air masses (Sewall and Fricke, 2013). In reaching the eastern side of the Cordillera, Pacific-sourced air masses would have undergone extensive distillation, thus producing much lower δ18O in Alberta and Montana.

What is more unusual about the δ18O data is the large offset in δ18O between paleosol carbonate nodule and hadrosaur tooth enamel, not observed at the other northern and eastern localities. Although the δ18O of tooth enamel from a homeothermic hadrosaur is presumably insensitive to temperature, the δ18O of authigenic soil nodules is not. Thus, one possible explanation of a larger difference in δ18O is that nodules formed at higher temperatures in FTS soils relative to other localities with nodules. Because nodules today are observed to form at the onset of the dry season when evaporation-induced precipitation of calcite takes place (e.g., Breecker et al., 2009), the implication is that dry seasons in Baja were significantly warmer than those at other locations. Perhaps these warmer temperatures are indicative of a more seasonally extreme climate in the region, an inference which is concordant with its observed sedimentological characteristics (e.g., episodic sedimentation, and wet/dry cyclicity) relative to the other late Campanian localities studied.

CONCLUSIONS

The FTS, ∼130 m of terrestrial, fossil-bearing sedimentation in the upper part of the El Disecado Member of the El Gallo Formation, Baja California, México, is an important, heretofore unknown southern and western locality producing faunas and floras of late Campanian age in North America. The FTS represents a fully terrestrial environment, tentatively, a marginal-marine distal arid fan characterized by high-energy, episodic sedimentation. It likely developed in association with the then-nascent Peninsular Range of central and northern Baja California, and prograded out against the proto–Pacific Ocean of the late Campanian, an inference we derive from the close stratigraphic proximity of a marine unit (“Ambar”) conformably at its base. The FTS is stratigraphically dissected by cross-stratified, well-sorted channelized medium sand-sized sand bodies, suggesting high flow intensities and low-sinuosity fluvial channels.

Petrographically, FTS sediments are characterized by compositional and textural immaturity, reflecting relative proximity to the source. In the scheme of Dickinson and Suczek (1979), clast compositions indicate a magmatic arc, an inference that is concordant with the paleogeographic setting. Grain sizes are uniformly fine, suggesting some geographic separation from the source. A strong sedimentary lithic component pervades all samples, reinforcing the idea of some transport and associated reworking.

Along with the high-energy, episodic deposition, well-developed, thick, sequences of compound paleosols preserving subsurface horizons contain evidence of a fluctuating water table, implying a climatic regime that was characterized by wet/dry cyclicity, and likely tied to the episodicity of the sedimentation. Actively flowing water deposited FTS sedimentary rocks; however, the presence of extensive carbonate nodule development in the subsurface horizons suggests that episodes of aridity characterized this environment.

Deposition direction inferred from the dip orientations of the cross stratification reveals some circular dispersion, but was generally to the north; somewhat unexpected in this coastal setting, in which broadly west-directed flow might be anticipated. A local peculiarity of geomorphologic setting, however, could be responsible for the strong northern component of the deposition. We also found significant small-scale eastward flow, likely reflecting tidal influence.

The taphonomy and sedimentology of the FTS suggest that most of the individuals there were not preserved in situ, in that the faunas are generally disarticulated, and their bones found under conditions in which they would not likely have lived (active flowing fluvial channels and sheet-flow debris deposits associated with storm events). Two distinct size classes of vertebrate fossils are preserved: large vertebrate elements (e.g., hadrosaur femora; isolated vertebrae, etc.) and microvertebrate remains, generally 1–3 mm in size but, among diagnosable fossils, averaging ∼15 mm. The large fossil elements are associated with fluvial channel deposits, whereas the microvertebrates are strikingly correlated with fine-grained sheet-flow deposits. Floras in the FTS are relatively common, and at several localities, small (∼1 cm in diameter) multi-seed, spherical infructescences referred to the angiosperm genus Operculifructus (O. Cornales) are commonly preserved (Hayes et al., 2018). These are found tightly associated with other plant remains in paleosol deposits, suggesting that they, in contrast to the vertebrates, may be autochthonous in origin. Rare plant macrofossils, other than the ubiquitous, allochthonous silicified woody trunks found in the braided channel deposits, are preserved in laminated, lenticular clay-rich, mudstone deposits, suggesting preservation in local floodplain ponds.

Analyses of stable isotopic ratios of hadrosaur teeth and primary calcitic nodules demonstrate that isotopic paleolatitude trends established in other late Campanian localities farther to the north and west extend into the FTS, the most westerly and southward locality of the suite. Although a number of factors might account for the paleolatitude-correlated shifts, the simplest and most likely explanation is that the floras eaten by the hadrosaurs varied latitudinally, and this is reflected in both their isotopic content, and that of the dinosaurs that ate them. That the FTS, at least, contains a floral assemblage different from that of its coeval northern and eastern counterparts is hinted by the presence of Operculifructus infructescences preserved there (Hayes et al., 2018), a cornalean genus unknown from the localities well to the north and east of Baja California in which originally the latitudinal stable isotopic trends were originally identified (Fricke et al., 2009).

ACKNOWLEDGMENTS

We gratefully acknowledge generous support by the Programa de Apoyo a Proyectos de Investigación e InnovaciónTecnológica, for IN 100913 “Fauna de microvertebrados de la formación El Gallo (Cretácico Tardío, Campaniano) Baja California, México” and IN103616 “Microvertebrados de la formación El Gallo y algunos aspectos paleoambientales con base en isótopos estables de C y O” to Dr. Montellano, as well as by the Departamento de Paleontología, Instituto de Geología, Universidad Nacional Autónoma de México. We also acknowledge generous support of Franklin Research grants from the American Philosophical Society and a University of Washington College of Arts and Sciences grant to G. Wilson. Special thanks are owed to Dra. T. Pi, Instituto de Geología, Universidad Nacional Autónoma de México, for her detailed analysis of the clays in Facies 2 (GF), and to Dr. Simon Engelhart, University of Durham, UK, for his help with the grain-size analysis. We thank the faculty who participated in the field work along the way: Jesús Alvarado, Enrique Martínez, and Jeffrey A. Wilson. We also thank all the students who participated in the field seasons and helped collect the fossil material (K. Ángeles-Ortíz, D.J. Cabrera-Hernández, M.L. Chavarría-Arellano, M. Chen, M. de Sosa, D. García-Alcántara, A. García-Gil, M.T. González-Cruz, O. López-Conde, M.L. Martín-Aguilar, J. Barrientos, and V. Zavaleta-Villarreal), as well as the many students who picked fossils from matrix in the Instituto de Geología paleontological laboratory. We particularly wish to acknowledge the contributions of las familias Rojas and Acevedo, without whose support in El Rosario, this work could not have been carried out. University of Rhode Island (URI) undergraduate C. Tiley carried out the analyses in Figure 15 and Table 1 and constructed the figure. The stable isotopic laboratory work in the United States was supported, in part, by the Department of Geosciences, URI, and by the Geological Society of America. U-Pb geochronology was supported through National Science Foundation grant EAR-1424892 to JR (Sam Bowring). Special thanks are due to J.P. Battacharya for his thoughtful, careful review; this work benefitted greatly from his efforts.

1Supplemental Material. Consists of petrographic data, calculations of sedimentological parameters (bedform height; water depth; slope), and paleocurrent data. Please visit https://doi.org/10.1130/GEOS.S.12331337 to access the supplemental material, and contact editing@geosociety.org with any questions.
Science Editor: Andrea Hampel
Associate Editor: Cathy Busby
Gold Open Access: This paper is published under the terms of the CC-BY-NC license.