Regional correlation of the Sonsela Member (Upper Triassic Chinle Formation) and detrital U-Pb zircon data from the Sonsela Sandstone bed near the Sonsela Buttes, northeastern Arizona, USA, support the presence of a distributive fluvial system

Prograding facies in terrestrial environments are likely to be associated with large distributive fluvial systems (“megafans”) that form most of the sediment volume of modern continental basins, and which draw their sediments directly from upland sources (Weissmann et al., 2010, 2015). Calibrating the progradation of these distributive systems may provide important clues as to how depositional rates are linked to the tectonic history of basin subsidence and source uplift (Kuhlemann and Kempf, 2002).

The Upper Triassic (Norian–Rhaetian) Chinle Formation consists of siliciclastic sediments deposited primarily by fluvial, lacustrine, and paludal systems (Blakey and Gubitosa, 1983; Dubiel and Hasiotis, 2011) on the landward side of the Cordilleran volcanic arc (Howell and Blakey, 2013; Ingersoll, 2012; Riggs et al., 2016). The Cordilleran arc was an important source of sediment transported east and northeast across Arizona, USA, to feed the northwesterly flowing Chinle stem fluvial system (Riggs et al., 2012, 2013, 2016). Compared with the Chinle stem river, the drastic differences in drainage direction for Late Triassic fluvial systems across Arizona, combined with their relative proximity to the arc sediment sources, raises the possibility that they were distributive fluvial systems (sensu Weissmann et al., 2010, 2015).

Trendell et al. (2012) advocated such an interpretation for the Sonsela Member (sensu Lucas, 1993; Heckert and Lucas, 2002; Woody, 2006; Martz and Parker, 2010), a coarse-grained unit of the Chinle Formation that was deposited across northern Arizona by low- to high-sinuosity fluvial systems (Howell and Blakey, 2013). Trendell et al. (2012) argued that the sediments of the uppermost Blue Mesa Member and overlying Sonsela Member were deposited by a large fluvial fan that prograded northeast across Arizona away from the Cordilleran arc, resulting in a coarsening upwards sequence being deposited in Petrified Forest National Park (PEFO), after 220 Ma (Ramezani et al., 2014). This interpretation is also supported by the radiating dispersal pattern of Late Triassic detrital zircon in the Sonsela Member at PEFO and the Vampire Formation and Waterman Formation in southern Arizona from a source in what is now the Mojave Desert (Riggs et al., 2013). In this paper, we present new U-Pb data for the type section of the Sonsela Sandstone bed, a unit that may be correlative with part of the Sonsela Member of PEFO, and discuss the implications for tracing the progradation of the Sonsela distributive fluvial system.

The Sonsela Sandstone bed was informally described as a course-grained unit within the Chinle Formation (Kiersch, 1955) but later was designated a type locality near the Defiance Uplift in the Navajo Nation of northern Arizona (Akers et al., 1958; Fig. 1) where it stands out prominently as two ledge-forming sandstones in the middle of otherwise slope-forming, fine-grained siltstones (Fig. 2). The Sonsela Sandstone bed was elevated to member-status within the Chinle Formation from work on stratigraphic sections at PEFO (Lucas, 1993; Heckert and Lucas, 2002; Woody, 2006), but the correlation between the “Sonsela Sandstone bed” at its type locality and the fossiliferous subunits of the “Sonsela Member” at PEFO has been exclusively based on the lithology and stratigraphic relationship of these units.

The Sonsela Member at PEFO contains numerous sandstone units, the most prominent of which is the Jasper Forest bed. The Jasper Forest bed (sensu Martz and Parker, 2010) contains many of the enormous silicified Triassic petrified logs after which PEFO is named. This bed occurs in a thick sandstone sequence that spans potentially as much as 9 m.y. (Ramezani et al., 2011; Atchley et al., 2013; Kent et al., 2018; Olsen et al., 2018) and its correlation to the type Sonsela Sandstone bed is not agreed upon by all workers (e.g., Lucas, 1993; Heckert and Lucas, 2002; Woody, 2006). The age of the Sonsela Sandstone bed at its type section could provide circumstantial support for the correlation with a given sandstone unit within the Sonsela Member of PEFO, although it is important to emphasize that lithostratigraphic units can be diachronous. Here we report the first detrital U-Pb zircon data from the Sonsela Sandstone bed from its type locality that are consistent with the correlation of the Sonsela Sandstone bed with the Jasper Forest bed of the Sonsela Member at PEFO.

We refer to the siliciclastic unit at the Sonsela Buttes between the underlying Bluewater Creek Member and overlying Petrified Forest Member of the Chinle Formation as the “Sonsela Sandstone bed” in order to retain its historical use. The Sonsela Sandstone bed was originally named as an informal lithostratigraphic unit prior to the adoption of standardized stratigraphic nomenclature (NACSN, 2005), but it was determined to be a scientifically and economically meaningful unit (Kiersch, 1955), and it was described and assigned a type section (Akers et al., 1958). The Sonsela Sandstone bed at its type locality is composed of three lithological facies, and it was correlated with other sandstone beds (described below) throughout the region. We refer to the formal “Sonsela Member” at PEFO as named by Lucas (1993) and defined by Woody (2003, 2006) and Martz and Parker (2010).

The Sonsela Member at PEFO includes five major units (Martz and Parker, 2010): the Camp Butte beds, the Lot’s Wife beds, the Jasper Forest bed, the Jim Camp Wash beds, and the Martha’s Butte beds. The Jasper Forest bed is considered the stratigraphic equivalent of the Kellogg Butte bed in the Devil’s Playground and the Rainbow Forest bed near the Rainbow Forest Museum, PEFO, Arizona (Martz and Parker, 2010; Parker and Martz, 2011; Martz et al., 2012).

The “Sonsela” bed (Kiersch, 1955, p. 5) was informally designated as a sandier division of the Petrified Forest Member of the Chinle Formation as it was recognized at the time. This was later named the “Sonsela sandstone bed” (Akers et al., 1958, p. 89) and was described and assigned a type section soon after to define a sandstone interval that divided the lower part of Herbert Gregory’s “Division C” of the Chinle Formation (Gregory, 1917, p. 43) into upper and lower parts of what was at that time referred to the Petrified Forest Member (Gregory, 1950; Cooley, 1957; Harshbarger et al., 1957).

At its type section a few kilometers east of Canyon DeChelly National Monument, Arizona at the Sonsela Buttes (Figs. 1 and 2), the Sonsela Sandstone bed was thought to overly the lower part of the Petrified Forest Member (Akers et al., 1958; Repenning et al., 1969; Stewart et al., 1972), but is now understood to overly the silty mudstone of the Bluewater Creek Member (informally the “lower red member”; Martz and Parker, 2010; Irmis et al., 2011) and is composed of three major divisions: a lower light gray sandstone, a middle blue-gray siltstone, and an upper, thicker light gray sandstone. These sandstones, especially the upper ledge-forming unit, are trough cross-bedded; the coarse bedloads contain volcanic clasts and chert cobbles that preserve Permian macrofossils similar to those found in the Kaibab Formation (Fig. 2B; McKee, 1937; Akers et al., 1958; Stewart et al., 1972). Where it crops out over 62,000 km² of northeastern Arizona and northwestern New Mexico, the Sonsela Sandstone bed is between 15 and 60 m thick and includes a variable number of individual sandstone lenses (Akers et al., 1958; Repenning et al., 1969; Stewart et al., 1972).

The Sonsela Sandstone bed at PEFO was raised to member-status (Lucas, 1993), but the Sonsela Member was restricted only to the conglomeratic sandstone capping Blue Mesa, between the Painted Desert Member and Blue Mesa Member within the Petrified Forest Formation of the Chinle Group. Although the wholesale status elevation of these units has not been widely accepted, most authors agree that the Sonsela Member is lithologically distinct at PEFO and is sufficiently widespread in the region to warrant member-status within the Chinle Formation (Raucci et al., 2006; Woody, 2006; Martz and Parker, 2010).

In an attempt to correlate the Sonsela Member as it occurs at PEFO to the three-part Sonsela Sandstone bed at its type locality, Heckert and Lucas (2002) divided the Sonsela Member at the park into an upper sandstone (the Agate Bridge bed), a middle silty unit (the Jim Camp Wash beds), and a lower sandstone (the Rainbow Forest bed). This tripartite division of the Sonsela Member was also hypothesized independently by Woody (2006) with a slightly different nomenclature provided (i.e., the Rainbow Forest beds, Jim Camp Wash beds, and Flattops One bed). Unfortunately, many of the correlations for the major sandstones in previous studies (e.g., Heckert and Lucas, 2002) were not accurately determined in the field or by walking out laterally extensive contacts between outcrop areas throughout the park, or by detailed mapping (Martz and Parker, 2010; Parker and Martz, 2017). Revisions to these correlations are summarized elsewhere (Raucci et al., 2006; Woody, 2006; Martz and Parker, 2010; Parker and Martz, 2011; Martz et al., 2012), but it is worth noting here that beds correlated across PEFO by Heckert and Lucas (2002) to their type sections of the Rainbow Forest bed and Agate Bridge bed are correlative to the Jasper Forest bed as used here (Martz and Parker, 2010).

The most current consensus divides the Sonsela Member at Petrified Forest National Park into five units, the Camp Butte beds, the Lot’s Wife beds, the Jasper Forest bed, the Jim Camp Wash beds, and the Martha’s Butte beds (Fig. 3; Martz and Parker, 2010; Martz et al., 2012). Using this framework, the Sonsela Sandstone bed at the Sonsela Buttes is correlated with the Jasper Forest bed at PEFO (Raucci et al., 2006, p. 158; Martz and Parker, 2010; Martz et al., 2012; Atchley et al., 2013). These correlations were primarily based on lithological similarity of the Jasper Forest bed/Rainbow Forest bed/Kellogg Butte bed with the upper sandstone at the type location of the Sonsela Sandstone bed (i.e., fining-upward, quartzose, fluvial trough cross-bedded sandstones containing limestone and chert clasts and silicified petrified logs), as well as the park-wide superpositional relationships of those named sandstone units to the Lot’s Wife beds and Jim Camp Wash beds below and above them, respectively, and historical use (e.g., Cooley, 1957; Lucas, 1993).

Correlations of the Sonsela Sandstone bed at its type locality with any of the variable sandstone-dominated units of the middle of the Chinle Formation in the vicinity of PEFO have been ambiguous. Previous authors (e.g., Cooley, 1957; Billingsley, 1985; Heckert and Lucas, 2002) agreed that the sandstone capping Blue Mesa at PEFO and the Sonsela Sandstone bed are lithologically identical and stratigraphically equivalent, but disagreed on the correlation with the Rainbow Forest bed, which crops out in the southern portion of PEFO. The Rainbow Forest bed was correlated with either the lower sandstone found at the type locality of the Sonsela Sandstone bed (Cooley, 1957; Heckert and Lucas, 2002) or was thought to occur ∼6 m above the sandstone capping Blue Mesa (Roadifer, 1966). It was also thought to be a distinct sandstone unit that is stratigraphically lower than the equivalent of the Sonsela Sandstone bed at PEFO (e.g., Billingsley, 1985; Ash, 1987; Murry, 1990).

The Sonsela Sandstone bed was previously sampled 13.3 km north of its type section (sample CP20 in Fig. 1) and detrital zircon U-Pb data were collected via laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) (Dickinson and Gehrels, 2008). The data were characterized by a significant component of Triassic Cordilleran arc-derived zircon between 227 Ma and 212 Ma, making up 29% of the sample, with three distinct modes in the spectra of youngest zircon grains at 215 Ma, 218 Ma, and 221 Ma (n = 90; Dickinson and Gehrels, 2008, 2010). Age-probability plots from LA-ICP-MS data for two samples in the lower part of the Sonsela Member at PEFO were determined from an aetosaur quarry in the Jim Camp Wash beds (sample 050508-1) and the Long Logs bed (Riggs et al., 2013). Cordilleran arc-derived grains make up 86% and 77% of each sample, respectively, with very minor age modes throughout the Proterozoic.

The first high-resolution chemical abrasion–thermal ionization mass spectrometry (CA-TIMS) zircon U-Pb dates from the Sonsela Member at PEFO were determined by Ramezani et al. (2011) from throughout the member (samples GPL, GPU, KWI, and SBJ in Figs. 1 and 3), yielding ²⁰⁶Pb/²³⁸U maximum ages for the maximum depositional age of the top and bottom of the Sonsela Member at PEFO of 213.124 ± 0.069 Ma and 219.317 ± 0.080 Ma, respectively. Sample GPL of that study was taken from the Jasper Forest bed and is dated at 218.017 ± 0.088 Ma.

More recent CA-TIMS zircon U-Pb dates were acquired from a geologic core drilled at Chinde Point (Fig. 1) by the Colorado Plateau Coring Project (Olsen et al., 2010, 2018; Kent et al., 2018). The published dates from the core sampled similar stratigraphic horizons as the outcrop sampled by Ramezani et al. (2011) within the upper part of the Sonsela Member at PEFO but without bed-level correlations in the core, stratigraphic location was only constrained by core depth. The three relevant CA-TIMS U-Pb dates from the core (Kent et al., 2018; Olsen et al., 2018) were 214.08 ± 0.20 Ma, 212.81 ± 1.25 Ma, and 213.55 ± 0.28 Ma.

A sample from the base of what was called the Blue Mesa Member was dated from Six Mile Canyon near Fort Wingate, New Mexico (Heckert et al., 2009, 2012; Irmis et al., 2011), but it is questionable that the Blue Mesa Member is present in New Mexico, and it is more likely that the sampled bed is situated within the lowest part of the Sonsela Member (see discussion below). Maximum depositional ages obtained via CA-TIMS for the Six Mile Canyon (SMC) bed at Six Mile Canyon (Fig. 3) were determined to be 219.3 ± 3.1 Ma (LA-ICP-MS, Heckert et al., 2009) and 218.1 ± 0.7 Ma (CA-TIMS, Irmis et al., 2011). An additional date was recovered from the same bed (220.9 ± 0.6 Ma, isotope dilution–thermal ionization mass spectrometry (ID-TIMS); Heckert et al., 2009, 2012) but the CA-TIMS date (218.1 ± 0.7 Ma; Irmis et al., 2011) is preferred because that method provides a more effective treatment for Pb-loss on the outer surface of the zircon and only those data are readily available for interpretation.

The thickness of each lithological unit at the type section of the Sonsela Sandstone bed (Akers et al., 1958; Museum of Arizona [MNA] locality 1809) was measured using a precision Jacob’s staff (ASC Scientific) with compass adaptor and a pocket transit compass (Brunton, Inc.). The color of unweathered rock samples was determined in spot holes using a Munsell rock color chart. GPS coordinates and photographs of the outcrop were taken using a Garmin Oregon 550 and Canon EOS Rebel XS digital camera. We collected 5 kg rock samples from units 4, 7, and 8 for geochronology from the measured section at the type locality of the Sonsela Sandstone bed (Fig. 3). This fieldwork was conducted under a geological reconnaissance permit issued by the Navajo Nation Minerals Department and the samples are reposited in trust of the Navajo Nation at the Museum of Northern Arizona in Flagstaff, Arizona.

Sample preparation and analysis was performed at the UTChron Geo-Thermochronometry Lab at the University of Texas at Austin. The entire sample was disaggregated in a jaw crusher and disc mill before hydrodynamic concentration using a Gemini water shaking table. A Franz Isodynamic Separator collected the non-magnetic fraction between two heavy mineral separations using methyl iodide (2.28 g/mL) and tribromomethane (2.89 g/mL). The zircon crystals were then mounted via double-sided adhesive tape on an acrylic disc and loaded into the large-format two-volume Helex sample cell for LA-ICP-MS analysis following the analytical procedures described by Marsh and Stockli (2015).

We randomly selected 122 zircon crystals from the sample taken in the upper sandstone (unit 8 of the measured section, MNA M.2576). The crystals were ablated with a 30 µm spot size using a PhotonMachines Analyte G.2 193 nm Excimer Laser for 30 s at 10 Hz. An Element2 High Resolution ICP-MS analyzed ²⁰⁶Pb, ²⁰⁷Pb, ²⁰⁸Pb, ²³²Th, ²³⁵U, and ²³⁸U. Analyses were depth-profiled (time-resolved) in order to monitor for Pb-loss within altered zones of a given zircon. Data reduction was performed using Iolite software with the IgorPro package (IgorPro, 2015; Paton et al., 2011) and VisualAge data reduction scheme (Petrus and Kamber, 2012). We excluded data from depth-variable intragrain age domains and domains characterized by high common Pb, Pb-loss, or U enrichment (Marsh and Stockli, 2015). The ICP was tuned using the NIST 612 glass (Jochum et al., 2011). A GJ1 zircon was used as an internal zircon standard (²⁰⁶Pb/²³⁸U 601.7 ± 1.3 Ma; Jackson et al., 2004) and was interspersed with unknown zircon at a 4:1 ratio. An in-house Pak1 zircon was used as a secondary standard (²⁰⁶Pb/²³⁸U 43.03 ± 0.01 Ma). Data were visualized using IsoPlot (Isoplot, 2015). Best ages were selected for each zircon based on careful evaluation of long-term University of Texas laboratory and this study’s data, carefully evaluating data precision, discordance, and data mode stability. ²⁰⁶Pb/²³⁸U ages are used for grains <850 Ma and ²⁰⁷Pb/²⁰⁶Pb ages are used for grains >850 Ma, with two-sigma internal error (Gehrels et al., 2008; Horstwood et al., 2016) and discordance assessed between ²⁰⁶Pb/²³⁸U and ²⁰⁵Pb/²³⁷U ages. All analytical and methodological details and U-Pb data are in Table S1¹. All U-Pb Data are also available from http://www.geochron.org/.

In order to assess the maximum depositional age (MDA) of the upper sandstone (unit 8) at the type locality of the Sonsela Sandstone bed, we calculated a weighted mean from the youngest coherent sub-population of detrital zircon from the Sonsela sandstone sample. LA-ICP-MS analyses can be more prone to the effects of Pb-loss than ID-TIMS or CA-TIMS, despite pre-analysis ablation (<1 micron) to clean the outer surface of each zircon prior to analysis (Gehrels et al., 2008; Mundil et al., 2008). In order to obtain a robust maximum depositional age, all U-Pb data (<240 Ma) were progressively filtered for discordance, ranging from 1%–5% and evaluated for data coherence in terms of mean age and mode. Hence, we applied a 3% discordance filter to select a subset of the zircon U-Pb data, based on topology of ranked single grain ages, preserving >50% of the data, and minimizing data characterized by older inheritance and lead loss, to calculate the most robust maximum depositional age for the Sonsela Sandstone bed.

The type section of the Sonsela Sandstone bed (Akers et al., 1958) crops out 5.63 km north of the western Sonsela Butte in Apache County, Arizona. (Figs. 1 and 2; Lower Wheatfields, Arizona 7.5 min quadrangle; 36.147298°N, 109.135812°W, datum WGS 84). There, the Bluewater Creek Member of the Chinle Formation (units 1–3 of the section measured for this study) underlies the Sonsela Member (units 4–8; Martz and Parker, 2010; Irmis et al., 2011), which contains two prominent sandstones separated by a thin siltstone interval (Figs. 2 and 3; Akers et al., 1958). The lower sandstone (unit 4) is 6.5 m thick and is lithologically similar to the upper sandstone. The middle siltstone interval (units 5 and 6) includes layers of greenish shale, greenish gray sandy siltstone, and grayish purple silty claystone. The upper sandstone (units 7 and 8) is a well-cemented, well-sorted, quartzose, medium-grained sandstone with coarse fractions contained in the bottom of fining-upwards sequences. It contains planar and trough cross-bedding, chert pebble stringers, and petrified wood. The top of the upper sandstone is truncated by erosion, but the unit itself is at least 18 m thick. The complete measured section and lithological descriptions of each unit are found below in Table 1 and can be seen in Figure 4.

The detrital age spectrum for the sample from the Sonsela Sandstone bed at its type locality (MNA M.2576) is characterized by a dominant Triassic age mode (41%) <250 Ma, which is likely derived from the early Mesozoic Cordilleran magmatic arc (Figs. 5A and 6C; Dickinson and Gehrels, 2008, 2010; Riggs et al., 2013, 2016). Of these Mesozoic zircon crystals, 43 yielded a Norian age (227–208.5 Ma; Cohen et al., 2018; Kent et al., 2017). The next largest detrital age components of the sample are zircons derived from the Yavapai-Mazatzal orogeny and the Archaean craton (each at 13.2%), Grenville orogeny (11.6%), and Mesoproterozoic plutons (10.7%). Detrital zircon associated with the Appalachian orogeny, peri-Gondwanan accreted terranes, Paleoproterozoic suture belts, and the Wopmay orogeny each contribute <5% to the sample.

The weighted mean of the 69 Triassic-aged zircon grains with <3% discordance is 216.6 ± 0.3 Ma (mean square weighted deviation [MSWD] = 4.9; Figs. 3, 5A, and 6). The distribution of discordance filtered ages shows a dominant age mode with younger ages and older ages due to lead loss and inheritance, respectively, resulting in an elevated MSWD value. The dominant age group exhibits a stable mean age and age mode for different discordance filters (1%–5%; Table S1 [^{footnote 1}]) and suggests a robust maximum depositional age of 216.6 ± 0.3 Ma.

Alternative approaches for determining MDA from a detrital zircon sample (Dickinson and Gehrels, 2009) include using the youngest age peak (ca. 218 Ma, Fig. 6B), the youngest single grain (205.8 ± 5.9 Ma), and the weighted mean of the youngest 2σ cluster of grains varying between 206 Ma and 210 Ma, depending on how many grains are used. The youngest age mode is fully congruent with the calculated weighted mean of 216.6 ± 0.3 Ma. The youngest single zircon appears to have experienced minor lead loss (Fig. 6) and is incongruent with previous CA-TIMS geochronology established from the Chinle Formation of northern Arizona, where the youngest MDA is ca. 208 Ma from the Owl Rock Member (Ramezani et al., 2011). The weighted mean age of ca. 217 Ma is also consistent with the lithostratigraphy at the Sonsela Buttes and previous CA-TIMS geochronology of the upper part of the Petrified Forest Member at PEFO, which is 212–210 Ma (Riggs et al., 2003; Ramezani et al., 2011; Kent et al., 2018).

The U-Pb zircon maximum depositional age provided here by LA-ICP-MS (216.6 ± 0.3 Ma) for the Sonsela Sandstone bed at the Sonsela Buttes is consistent with the lithostratigraphic correlation of this unit with the Jasper Forest bed, or more conservatively, with the lower part of the Sonsela Member at PEFO (the upper Lot’s Wife beds, Jasper Forest bed, and lowermost Jim Camp Wash beds; Raucci et al., 2006; Martz and Parker, 2010). Three previous U-Pb CA-TIMS dates constrain the age of the Jasper Forest bed at the park, 218.017 ± 0.088 Ma from the Jasper Forest bed and 213.870 ± 0.088 Ma from the Jim Camp Wash beds (Fig. 3; samples GPL and KWI, respectively; Ramezani et al., 2011) and 214.08 ± 0.20 Ma from the Jim Camp Wash beds (Fig. 3; sample 182Q1; Kent et al., 2018; Olsen et al., 2018) overlap with the date and error of the Sonsela Sandstone bed type section provided here (however, sample GPL may be too old for its core depth; Kent et al., 2018). Furthermore, our LA-ICP-MS data places the type Sonsela Sandstone bed squarely within the Norian Stage (Kent et al., 2017; Cohen et al., 2018).

At PEFO, the Adamanian–Revueltian holochronozone boundary is characterized by a turnover in the terrestrial biota as observed in phytosaurs, aetosaurs, and palynomorphs (Parker and Martz, 2011; Reichgelt et al., 2013; Baranyi et al., 2017; Martz and Parker, 2017; Beightol et al., 2018). If vertebrate fossils are found at the type Sonsela Sandstone bed in the future, we hypothesize that they would include taxa indicative of the Adamanian land vertebrate estimated holochronozone, such as the non-pseudopalatine leptosuchomorph phytosaur Smilosuchus and the desmatosuchin aetosaurs Scutarx deltatylus and Desmatosuchus smalli (Parker and Martz, 2011; Desojo et al., 2013; Stocker and Butler, 2013; Parker, 2016; Martz and Parker, 2017).

The Adamanian–Revueltian boundary is hypothesized to coincide stratigraphically with a persistent red “silcrete” bed in the lower part of the Jim Camp Wash beds at PEFO that also approximately coincides with evidence of increasing aridity in the region (Martz and Parker, 2010; Parker and Martz, 2011; Martz et al., 2012; Atchley et al., 2013). A marked difference in deposition of the Sonsela Member was reported at this horizon (Howell and Blakey, 2013); the lower part of the Sonsela Member (the Camp Butte beds, the Lot’s Wife beds, and the Jasper Forest bed) is made up of coarse-grained sheet sands that indicate relatively low basin subsidence, and upper part of the Sonsela Member (most of the Jim Camp Wash beds and the Martha’s Butte beds) is finer-grained with thin sand lenses representing a relatively high rate of basin subsidence. That change in sedimentation is also reflected in the detrital zircon provenance data for the Sonsela Member (Dickinson and Gehrels, 2008, 2010; Riggs et al., 2012, 2013).

The spectral age distribution of MNA M.2576 (Fig. 5B) is nearly identical to that of sample CP20 from several kilometers north of the type Sonsela Sandstone bed (Dickinson and Gehrels, 2008, 2010) and samples from the lower part of the Sonsela Member at PEFO (Riggs et al., 2013). MNA M.2576, CP20, and the Long Logs bed sample included age modes representing grains from the Grenville orogeny, Mesoproterozoic plutons, Yavapai-Mazatzal orogeny, and Paleoproterozoic suture belts (Fig. 5B; Dickinson and Gehrels, 2008, 2010). The sample from the Long Logs bed at PEFO includes age modes from Mesoproterozoic, Yavapai-Mazatzal, and Paleoproterozoic grains, but contains no zircon from the Grenville orogeny (Riggs et al., 2013). The similarities of the spectra from MNA M.2576, CP20, and the two samples from the lower part of the Sonsela Member at PEFO is consistent with the hypothesis that the Sonsela Sandstone bed at its type section correlates to the Jasper Forest bed, or at least the lower part of the Sonsela Member at PEFO (below the persistent red “silcrete”), which is dominated by Cordilleran arc-derived zircon (Howell and Blakey, 2013), rather than the upper part of the Sonsela Member at PEFO.

The recent CA-TIMS dates for the Chinle Formation (Ramezani et al., 2011, 2014; Irmis et al., 2011; Atchley et al., 2013; Kent et al., 2018; Olsen et al., 2018) also allow us to begin developing a more sophisticated model of Norian sediment deposition across the Colorado Plateau. The Monitor Butte Member of southern Utah, Cameron Member of northern Arizona, and the Bluewater Creek Member of northeastern Arizona and northwestern New Mexico are at least partially correlative with each other (Lucas, 1993; Lucas et al., 1997; Irmis et al., 2011, Kirkland et al., 2014; Martz et al., 2015), and with the lower part of the Blue Mesa Member type section at PEFO, with complex interfingering facies relationships between members occurring across the region (see more detailed discussions by Irmis et al., 2011, their supplemental data; Kirkland et al., 2014; Martz et al., 2015). In some areas, these units are capped with a drab mudstone-dominated unit similar to the upper part of the type section of the Blue Mesa Member at PEFO (Fig. 7; e.g., Lucas et al., 1997; Kirkland et al., 2014; Martz et al., 2015). The upper drab mudstones have even been assigned to the Blue Mesa Member (Lucas, 1993; Lucas et al., 1997). Although previous workers favored lithostratigraphic correlation of these drab mudstones, none addressed the possibility that the stratigraphic units might be regionally diachronous, a possibility warranting serious consideration given the striking difference in thickness and complexity between the overlying Sonsela Member at PEFO, Sonsela Sandstone bed at the Sonsela Buttes, and potentially equivalent Moss Back Member and related sandstones in southern Utah (e.g., Lucas et al., 1997; Kirkland et al., 2014).

The SMC sample at Six Mile Canyon, New Mexico (ca. 218 Ma), was derived from a sandstone just below a drab mudstone-dominated unit interpreted by previous authors to be correlative to the Blue Mesa Member, immediately overlying the Bluewater Creek Member (Irmis et al., 2011; also see Heckert and Lucas, 2002; Heckert et al., 2009, 2012). Subsequent U-Pb ages older than 219 Ma were reported from the type locality of the Bluewater Creek Member in New Mexico, only a few kilometers from Six Mile Canyon (Ramezani et al., 2014). Moreover, the Placerias Quarry at St. Johns south of PEFO (Fig. 1) occurs in a drab-colored mudstone-dominated facies resembling the upper part of the Blue Mesa Member at PEFO (Parker, 2018) with an age of 219.39 Ma (Fig. 7; Ramezani et al., 2014). These ages suggest that the drab, mudstone-dominated “upper Blue Mesa Member” in western New Mexico and in northern Arizona near St. Johns correlates chronostratigraphically with the lowest part of the Sonsela Member in PEFO (the Lot’s Wife beds; Parker, 2018), rather than with the upper part of the Blue Mesa Member mudstones at PEFO, which ceased deposition there at ca. 220 Ma (Ramezani et al., 2011; Atchley et al., 2013). This indicates that the “upper Blue Mesa Member” facies are regionally diachronous (Irmis et al., 2011).

It has been hypothesized that the deposition of the Sonsela Member deposition at PEFO was initiated by a massive alluvial fan that prograded to the northeast from the Cordilleran magmatic arc during the Norian (Trendell et al., 2012). A consequence of this model is that the base of the Sonsela Member should be diachronous, with distal parts of the fan being finer-grained compared to coarser and more proximal fan deposits that progressively overlaid them. Here we provide support for diachronous deposition of both fine-grained upper Blue Mesa Member facies and overlying coarse-grained Sonsela Member facies. Coarse-grained lower Sonsela Member facies at PEFO (the Lot’s Wife beds) were being deposited at ca. 218–219 Ma, contemporaneous with finer-grained upper Blue Mesa Member facies to the southeast (St. Johns), west (Zuni Mountains), and north (Navajo Nation); in all three regions, coarse-grained Sonsela Member facies were deposited later. At least in the Defiance Uplift, the Sonsela Sandstone bed may have been deposited around 214–218 Ma (Figs. 3 and 7), about the same time that the lower units within the Sonsela Member were being deposited in PEFO.

The interpretation presented here supports the idea that the upper part of the Blue Mesa Member at PEFO is a diachronous siliciclastic transition zone between the micaceous equivalent to the Bluewater Creek Member (the Newspaper Rock bed and lower part of the Blue Mesa Member; Martz and Parker, 2010; Ramezani et al., 2014) and the lowest part of the Sonsela Member (Fig. 7). Future analyses should target the Sonsela Member–Blue Mesa Member transition zone and the rest of the Sonsela Member and its partial equivalents in Utah (such as the Moss Back Member) using high-resolution CA-TIMS to improve age precision and resolution of the youngest mode of detrital zircon U-Pb data to refine cross-state correlations of the members of the Chinle Formation that preserve the rock record of the Adamanian–Revueltian boundary. Our LA-ICP-MS U-Pb data support the hypothesis that the uppermost Blue Mesa Member and the Sonsela Member (and the Sonsela Sandstone bed at its type locality) were deposited in a continental backarc basin (Howell and Blakey, 2013) or retroarc foreland basin (Riggs et al., 2016) by a large “megafan” prograding from the southwest into the northwesterly flowing Chinle stem fluvial system (Riggs et al., 2012, 2013). Finally, our data provide another example of how U-Pb geochronology can clarify stratigraphic relationships and the diachroneity of fluvial depositional systems, as well as their influences on regional biostratigraphy.

We thank Jake Tapaha, an intern at PEFO in 2012, for help measuring the section described in this paper. Fieldwork on the Navajo Nation was conducted under permits from the Navajo Nation Minerals Department issued by Mr. Ahktar Zaman and Mr. Brad Nesemeier. Any persons wishing to conduct geologic investigations on the Navajo Nation must first apply for and receive a permit from, P.O. Box 1910, Window Rock, Arizona 86515 and telephone number (928) 871–6587. We also thank Lisa Stockli and the graduate students at the UTChron Laboratory at the University of Texas at Austin analytical assistance and constructive discussions. We thank Timothy Rowe, Christopher Bell, Julia Clarke, and Hans-Dieter Sues for reviewing an early draft of this manuscript. Comments and suggestions from the associate editor and two anonymous reviewers greatly improved the manuscript. This is Petrified Forest National Park Paleontological Contribution No. 59. The conclusions presented here are those of the authors and do not represent the views of the United States Government.