Abstract

Limestone beds underlain and overlain by alluvial fan conglomerate near Amboy, California, are very similar in many respects to parts of the Bouse Formation, suggesting that an arm of the Pliocene Bouse water body extended across a wide part of the southern Mojave Desert. The deposits are north of the town of Amboy at and below an elevation of 290 m, along the northern piedmont of the Bristol “dry” Lake basin. The Amboy outcrops contain the Lawlor Tuff (4.83 Ma), which is also found in an outcrop of the Bouse Formation in the Blythe basin near Buzzards Peak in the Chocolate Mountains, 180 km southeast of Amboy. Bouse exposures near Amboy are ∼3.4 m thick, white, distinctly bedded, with limestone and calcareous sandstone as well as stromatolite mounds; we interpret these as nearshore deposits. The Bouse at Amboy contains ostracodes, diatoms, and mollusks that indicate saline lake or estuarine environments with an admixture of fresh-water forms. Along with wading bird tracks and a spine from a marine fish, these fossils suggest that the deposits formed in saline waters near a fresh-water source such as a perennial stream. Beds of the outcrop dip southward and are 113 m above the surface of Bristol Playa, where similar age sediments are buried 270+ m deep, indicating significant faulting and vertical tectonics in this part of the Eastern California Shear Zone during the past 5 m.y. Confirmation of the Bouse Formation at Amboy strengthens previous assignments to the Bouse Formation for mudstones in driller logs at Danby “dry” Lake, California, and suggests that areally extensive arms of the Bouse water body were west of the Blythe basin. The Bristol basin arm of the lower Bouse basin probably was restricted from the main water body by narrow passages, but Bouse sediment there is similar to that in the Blythe basin, suggesting generally similar water chemistry and environmental conditions. Examining the degree to which Bouse deposits in the western arms differed from Bouse deposits in the Blythe basin offers an approach to test whether the southernmost Bouse water body was deposited in an estuarine or lacustrine setting.

INTRODUCTION

The origin and evolution of the Colorado River are key to understanding many topics of current interest, including dynamics of river evolution, dispersal of aquatic species, and tectonics associated with the San Andreas fault plate boundary and the opening of the Gulf of California. Study of the Pliocene Bouse Formation nearest to the Gulf of California, the deposits of which formed in a terminal lake in the Colorado River system and/or an estuary of the Gulf of California, may yield information on several of these topics. Over the past five decades study of stratigraphy, fossils, and chemistry of the Bouse Formation has converged toward an understanding that the Blythe basin is Miocene to early Pliocene in age, and locally >400 m deep (Metzger, 1968; Buising, 1990). However, connections with marine waters seemingly required by fossil species (Metzger et al., 1973; McDougall, 2008) have been difficult to reconcile with isotopic composition of Bouse carbonate materials in this and basins farther north that are more similar to modern Colorado River water than seawater (Spencer and Patchett, 1997; Roskowski et al., 2010). Basins upstream of the Blythe basin have been compellingly demonstrated to have formed by sequential fill of lakes, cutting through their dams to spill to the next lower basin, followed by another fill cycle (House et al., 2008). This mode of origin requires that the upper basins were lacustrine; similar Sr isotopic values for carbonates in these northern basins and the Blythe basin (Roskowski et al., 2010) strongly argue for a lacustrine origin for the Blythe basin. However, isotopic models based on O, C, and Sr support a marine origin for the southern Bouse system by mixing of groundwater, river water, and marine water (Crossey et al., 2011).

We describe a key exposure near the town of Amboy, California (Fig. 1), far to the west of the main Blythe basin. We argue that the sediments, fossils, and age of the deposit indicate that it represents the Bouse Formation. The environment indicated by analysis of the fossils and sediment is one of brackish water with admixtures of nearby fresh water, which could represent either estuarine or brackish lacustrine settings. Future studies of the potential environmental differences of the highly evaporative western arms of the Bouse basin and the tectonics after Bouse deposition may improve understanding of its connections with marine waters.

PREVIOUS WORK

Twentieth century debates about the distribution mechanism of fresh-water fishes (Blackwelder, 1934; Hubbs and Miller, 1948) postulated drainages from the central Mojave Desert through the Bristol-Cadiz-Danby trough to the Colorado River (Fig. 1). Minimal evidence of either fresh-water or brackish-water organisms in Bristol Lake sediments from the surface to 270 m depth (ca. 3.65 Ma; Rosen, 1992) refutes this concept of a fresh-water drainage system (Bassett et al., 1959; Brown and Rosen, 1992; Rosen, 1992). However, in the Danby basin and perhaps also in the Cadiz Basin (Fig. 1), sediments containing brackish-water ostracodes and foraminifera (Smith, 1970; Brown and Rosen, 1992, 1995) have been interpreted as an extension of the Pliocene Bouse embayment. Recent studies of the Bouse Formation support Blackwelder’s (1934) idea that downstream integration of the Colorado River occurred by successive lake overflow events, creating fresh-water lakes in several basins, all north of a barrier at Topock (e.g., Spencer and Patchett, 1997; House et al., 2008). However, sediments in the Blythe basin south of Topock (Fig. 1), from Parker to Cibola, may in part be older than other Bouse beds and may reflect interplay of fresh and estuarine to marine conditions (Smith, 1970; Metzger et al., 1973; Todd, 1976; McDougall, 2008; Spencer et al., 2008). Many interesting questions remain regarding the geographic extent and paleoenvironment of the lower Blythe basin system.

In addition to the compelling stratigraphic and chronologic studies that indicate that the Bouse deposits in basins upstream of the Blythe basin are lacustrine (House et al., 2008), isotopic studies cast strong support for lacustrine origins for all of the Bouse deposits (Spencer and Patchett, 1997; Roskowski et al., 2010; Spencer et al., 2013); these authors have shown that fossils, carbonate crusts, and secondary carbonate in the Bouse Formation have 87Sr/86Sr values between 0.7102 and 0.7114, in contrast with seawater values of ∼0.709. Measured values for Bouse deposits in northern basins, where lacustrine origins are clear and the age of the carbonate is demonstrably Pliocene, are similar to values for the southern basin and to modern Colorado River values. However, reports of open-marine faunal assemblages in the lower Bouse Formation of the Blythe basin and indications that the lower Bouse is as old as or older than the Bouse in upper basins (e.g., McDougall, 2008) lend support to the interpretation that part or all of the lower Bouse Formation in the Blythe basin represents open-marine to estuarine deposition. It is difficult to develop a scenario for which an estuary is blocked off from the ocean and becomes a deep lake (Spencer et al., 2013). Avian introduction of fauna into a saline lake may resolve the paradox (Spencer et al., 2008), but more work is needed. However, a river-dominated estuary may be chemically similar to saline lakes, and mixtures of groundwater, meteoric, and marine waters can be consistent with measured isotopic values for the lower Bouse Formation (Crossey et al., 2011).

GEOLOGIC SETTING

In this paper we describe deposits near Amboy, in Bristol Lake basin, and discuss their relations with similar deposits of the Colorado River trough from Blythe to Parker (Fig. 1), across an area of several mountain ranges separated by generally broad valley floors. The Colorado River south of Blythe is 65 m above sea level, and is flanked by uplands, much well above 100 m elevation. As illustrated by the 300 m reference elevation (Fig. 1), topographically low areas of the upland extend west and northwest from the Colorado River trough to Palen Lake and Bristol Lake. Mountains of the Eastern Transverse Ranges separate the Palen-Ford lakes low from the Salton Trough farther west. Mountains of the Mojave Desert rise west, north, and east of the Bristol-Cadiz lake lows; these low areas currently connect with the Colorado River trough via passes distinguished in Figure 1; the passes range from ∼30 to ∼110 m above the nearby low areas.

Much of the area from Blythe northwest to Bristol Lake basin is in the Eastern California shear zone (ECSZ) of Dokka and Travis (1990), a region of overall dextral shear northeast of the San Andreas fault plate boundary. This shear zone has accommodated several tens of kilometers of dextral shear since the late Miocene (ca. 10 Ma) within a complex array of dextral and sinistral strike-slip faults (e.g., Richard, 1993; Langenheim and Powell, 2009). Thick sedimentary basins of uncertain age underlie some of the low areas, suggesting persistent releasing bends in strike-slip faults that bound the basins (Richard, 1993; Jachens et al., 2002; Langenheim and Powell, 2009). Separations on these faults of the ECSZ are incompletely known, but several have established separations of 6–8 km (Howard and Miller, 1992; Richard, 1993; Langenheim and Powell, 2009). Throw on the faults is generally poorly known, but several are along linear steep mountain fronts (Fig. 1), suggesting that throw on the faults may be several hundreds of meters. Faults shown in Figure 1 are mostly Quaternary in age, although some faults along the northeast margin of the ECSZ are older than widespread late Pleistocene deposits (Howard, 2002) and are thought in some cases to have last ruptured in the early or middle Pleistocene (Bedford et al., 2010; Phelps et al., 2012). Bassett et al. (1959), Rosen (1992), and Richard (1993) described evidence that Pliocene and early Pleistocene strata are deformed in basins of the ECSZ. The area we describe is near the northwest end of the Bristol-Cadiz-Danby trough, a topographic low that trends east-southeast to the Parker area on the Colorado River.

AMBOY AREA GEOLOGY

Quaternary piedmont gravels ring the discharging playa of Bristol Lake (Rosen, 1992; Bedford et al., 2010). The northern piedmont flanks the southern Bristol Mountains, which are underlain by Mesozoic granitoids and early Miocene volcanic and sedimentary rocks (Miller et al., 1982; Miller, 1994). Strike-slip faults of the ECSZ cut these rocks; the faults strike northwest and are dextral with as much as 6 km of horizontal separation (Howard and Miller, 1992). Quaternary (the past 2.6 m.y.) deformation of deposits under Bristol Lake is indicated by anomalously high depositional rates and faults encountered by boreholes (Rosen, 1992, 2000).

During regional geologic mapping of the southern Bristol Mountains in the 1980s, one of us (Miller) identified outcrops of limestone in gentle hills within the piedmont north of Amboy. One large outcrop 3 km north of the town of Amboy [Fig. 2; Universal Transverse Mercator (UTM) Zone 11, North American Datum (NAD) 83, 615026 m E, 3828006 m N) was excavated before that time. Other outcrops are scattered about the area (Fig. 2), but only three expose more than a few thin sandstone beds or stromatolitic mounds.

Limestone, sandstone, and mudstone crop out in a few hills and in gullies of the piedmont (Fig. 2). These rocks are above alluvial fan conglomerate, which in turn is on the 18.8 Ma Peach Spring Tuff (Hillhouse et al., 2010; Miller et al., 2010; Ferguson et al., 2013) 0.3 km to the north (Fig. 2). Limestone, sandstone, and mudstone are exposed discontinuously and exhibit variation in strike and dip, indicating structural complexities. The southward dip of limestone beds (278°, 26° south), as well as downward-steepening dips of underlying conglomerate and tuff, suggest that the section was tilted by folding or faulting. Although deformation may have changed the elevation of the outcrop, it is now at 290 m elevation, within the elevation range of Bouse Formation outcrops along the Colorado River from Parker to Cibola (maximum ∼330 m; Spencer et al., 2008). Furthermore, strontium isotopic ratios in the Amboy limestone overlap those of the so-called Bouse marl in the Blythe basin, supporting the correlation of the Amboy locality with the Bouse Formation (Spencer et al., 2013). Faults cut the Bouse beds and underlying gravels in two exposures, and one outcrop-scale anticline plunges gently to the northwest. Both faults are near the main outcrop; they strike northwest and are normal, down to the southwest. Mudstone is restricted to one exposure; the mudstone did not yield any fossils.

DESCRIPTION OF THE BOUSE FORMATION NEAR AMBOY

An excavated exposure of limestone (Fig. 2) was examined in detail and sampled to allow fossil and sediment description. This reference section is ∼3.4 m thick (Figs. 3 and 4), and contains a tephra bed 40–45 cm thick near or at its base. Much of the section is pale in color and consists of silty to sandy limestone containing abundant crystalline calcite matrix and grains of calcite. Beds near the base and top contain more local clastic sand and gravel, and these intervals also commonly contain stromatolitic algal mounds (tufa of Metzger, 1968). We divided the reference section into eight intervals based on outcrop descriptions of lithology, bedding characteristics, and color, and labeled these A to H, from base to top (Fig. 3). The following descriptions are based on field and microscope study. In addition to study of the reference limestone section, we studied a quarried mudstone section that appears to be below limestone at a lower elevation (Fig. 2).

Interval A

Stromatolitic mounds (Fig. 5) are on locally derived alluvial fan conglomerate at the base of the section. The mounds encrust boulders and cobbles and in many cases coalesce upward to form a single wide mound. The mounds are formed of calcite with prismatic structure in dense radial growths (Fig. 5B) alternating with more porous growth. The mounds incorporate very little terrigenous material. Adjacent to mounds at the reference section as much as 25 cm of thin-bedded sandstone is above the alluvial fan conglomerate. Color varies from pale brown to white, apparently a result of the amount of dark colored sand derived from local granitic sources. The lowest sandstone beds are arkosic with notable biotite, and cemented by white crystalline calcite. Upward some beds are nearly white in color, and contain less arkosic sand and more calcite in aggregates that suggest coarse to very coarse sand grains of calcite, now recrystallized. A few recrystallized calcite masses are shaped like fragments of mollusk shells. The mounds and sand are overlain directly by ash of interval B. Calcite veins cut sandstone of interval A. One out-of-place sandy limestone bed, apparently belonging to this interval, had a bird track, described in a later section.

Interval B

Gray fine-grained ash makes up interval B; the ash is ∼40–45 cm thick and drapes over algal mounds. The ash is pure and very well sorted in the basal 5 cm, and upward is faintly thin bedded with wisps of silt and flakes of biotite, suggesting reworking of the upper part. The tephra was identified as the 4.83 Ma Lawlor Tuff by tephrochronologic correlation (A.M. Sarna-Wojcicki, 2005, written commun.; Sarna-Wojcicki et al., 2011). Tuff at this location was studied by Harvey (2014) to strengthen correlations. Harvey found that ages of zircon in the tuff and the isotopic composition of that zircon both support the correlation with the Lawlor Tuff.

Interval C

Interval C is composed of parallel-bedded white to tan calcareous sandstone and silty limestone. Beds are 1–5 cm thick and laterally persistent (Fig. 4), and some show faint lamination. Thicker beds are coarser grained, and sand particles are composed of fine- to medium-grained arkosic sand in a crystalline calcite matrix. Thinner beds have greater calcite content, and a few having chalky appearance are diatom rich. A few beds carry sand-sized calcite fragments of indeterminate origin. Diatoms, ostracodes, and carbonate fragments are apparent.

Interval D

The very thinly bedded interval D is similar to interval C, but beds are laminated and cross-laminated, and most are very fine grained sand and silt. Beds are composed of calcareous sand with arkosic sand admixtures, and a few beds contain very fine gravel. The interval appears to have lenticular bedding but in detail the lenses are low-angle cross-stratified beds with variable replacement by silica. Some beds with ripple cross-laminations show sediment transport to the north. Coarse beds have local scoured bases, at most 3 cm deep. Silicified beds have been replaced by microcrystalline silica and cut by gypsum veins from an undetermined source. Many beds contain abundant diatoms, and a few contain abundant ostracodes.

Interval E

Interval E is distinguished by its contorted bedding and coarse sandstone. Its planar base is gently disconformable on the underlying rippled sand. The lower half is the most tightly folded, with fold axes that trend about eastward with vergence to the south. Beds are composed of either medium to coarse calcite and arkosic sand or low-density limestone containing abundant diatoms. The top of the interval is truncated across bedding.

Interval F

Above the contorted beds of interval E is an interval of sandstone, some with nodular-appearing blebs. The beds in interval F are thin, and a few have ripple laminations indicating north-directed sediment transport. The coarse beds low in the interval contain 15%–30% arkosic material, whereas the overlying nodular beds contain 20%–40%. Nodules appear to be lenses consisting of cemented calcite sand in beds of mixed calcite and arkose sand. The nodules probably originally were medium-grained sand, whereas the mixed sand enclosing the nodules is coarse grained. A few beds contain scattered fine gravel-sized lithics. Sand-sized calcite fragments range from lumps apparently derived from stromatolites, to cylinders and branching tubes, to a few whole and fragmented gastropods.

Interval G

Interval G is ∼0.8 m thick and poorly exposed, consisting of lithic-rich thin-bedded coarse sandstone and conglomerate. Beds are white to pale brown and appear to be planar. Arkosic and lithic material is moderately to well sorted and constitutes 40%–50% of the beds; the remainder is calcite chips and chunks of varying color, all in a fine crystalline calcite matrix.

Interval H

Interval H is composed mainly of poorly sorted alluvial conglomerate with clasts as large as boulder size. Scattered carbonate caps on clasts and a few thin beds of moderately sorted very coarse sandstone and fine conglomerate indicate that the deposits represent environments at the margin of a water body. As much as 1.5–2 m higher is a line of carbonate masses in the alluvial conglomerate that suggests stromatolites in a later standing-water incursion or a period of groundwater discharge.

Mudstone Unit

Mudstone in a quarry exposure at ∼275 m elevation consists of very thin bedded to laminated brown mudstone and sandstone. The structure appears to be complex, with beds of several orientations, but exposures are poor. Limestone is nearby to the west, but the mudstone is not interbedded with any rock types observed in the reference section, so its stratigraphic position is uncertain. The rock types range from clayey siltstone to well-sorted siltstone and very fine sandstone. Fine mica is a constituent in all mudstones, and rosettes of gypsum are common. Several ichnofossils, described in the following, were recovered at the quarry (Reynolds, 2012).

PALEONTOLOGY AND ENVIRONMENT

All samples shown in Figure 3 were studied for ostracodes, diatoms, and foraminifera, and examined under a microscope in situ. No foraminifera were recovered (K. McDougall, 2011, written commun.). In addition, large samples from intervals that appeared to have fragments of mollusks were processed for macrofossils.

Stromatolitic algal mounds (Fig. 5) in the Amboy section (Bristol basin) directly overlie alluvial fan conglomerate, and therefore record deposition at the margin of a standing body of water. The mounds are made up of thick authigenic calcite layers and formed in shallow water. They have no well-documented counterparts in the thicker Bouse section in the Blythe basin, but may be similar to the Bouse tufa, which occurs on nearshore bedrock in areas where water was shallow and clear (Metzger, 1968; Metzger et al., 1973; Buising, 1990; Reynolds et al., 2007).

Ostracodes

Ostracode fossils occurring throughout the limestone in the reference section include Cyprideis beaconensis and probable Cytheromorpha (revised from possible Limnocythere staplini reported in Reynolds et al., 2008). C. beaconensis is not reported from the Bouse Formation of the Blythe basin and is similar to C. castus reported in the Bouse Formation from the Parker area, so we carefully studied these ostracodes to determine species. The ostracode C. beaconensis has morphometric valve characteristics that distinguish it from C. castus, including size, sulcus and hinge characteristics, and dorsal bulges (Reynolds et al., 2008).

Ostracodes studied in 12 samples (Fig. 3) are summarized in Table 1. Ostracode valves in all samples have a finely crystalline texture, originally interpreted as calcite overgrowths, and are poorly preserved and abraded. Several valves were manually broken and found to be hollow. It appears that at least some presumed valves have been completely replaced by secondary calcite, and are simply molds of the original valves. The crystalline texture obscures much of the valve topography, making identification more difficult, but C. beaconensis and probable Cytheromorpha are present in all samples bearing ostracodes. A few samples, generally the samples in the center of the section, also contain an unidentified species of Candona. Candona likely represent a lower salinity setting than the other ostracodes species, and are not known from marine settings and, to our knowledge, are not known from estuarine settings.

Within the Cyprideis group, the common environmental theme (whether continental or nearshore marine) is a persistent environment with low alkalinity and low calcium solutions, such as the perennial springs and marshes surrounding the Great Salt Lake, the Salton Sea, and estuarine and coastal marine waters. C. castus, present in the Bouse Formation in the Blythe basin, is less able to withstand transport from marine settings to lakes, because of limited brood pouches, than is C. beaconensis. Cytheromorpha has hydrochemical requirements similar to those of the Cyprideis group. C. beaconensis has not been reported in the Bouse Formation from Parker to Cibola (R. Forester, 2008, written comm.), and has been collected live along the shore of the Salton Sea (Hurlbert et al., 2001; Detwiler et al., 2002). Fresh water from a source such as the Colorado River is apparently incompatible with the environmental requirements of C. beaconensis; to date it has not been collected from that river. Shallow, brackish waters under elevated rates of evaporation at the northwest end of a lacustrine or estuarine embayment would have been suitable habitat for C. beaconensis.

Diatoms

Study of 12 samples yielded 47 species and 29 genera (Table 2), despite the fact that 2 samples were barren and 2 had only fragments of diatoms. As with the ostracodes, poor preservation suggests transport before deposition. It may also create preservation bias by size sorting and by silicification. The most common genera are Actinocyclus, Cocconeis, Diploneis, Epithemia, Fragilaria, Mastigloia, Paralia, and Pseudostaurosira; of these, Diploneis, Fragilaria, and Paralia account for more than half of the total. Most of the species present in the samples are benthic, brackish-water taxa that occur in estuaries of the Gulf of California and California coast. Actinocyclus is generally marine. However, Fragilaria and Pseudostaurosira are common in all samples and are fresh-water species. Epiphytic species, such as Cocconeis spp., indicating the presence of aquatic macrophytes throughout the section, are most abundant in interval C. The mix of fresh-water and salt-water species may indicate a saline lake or an estuary, with nearby spring or stream input.

Macrofossils

Hints of carbonate-encrusted fragments of mollusk shells and ubiquitous calcareous tubes in intervals E and F led to processing of large samples of intervals A, E, and F for macrofossils. Processing included wet sieving followed by cleaning and study under microscope by P.I. LaFollette (2014, written commun.). Mollusks are rare at the outcrop, but washing recovered a few specimens from intervals E and F. Two or three genera are represented from interval E, but the only positive identification is Batillaria californica (Taylor, 1983). A few calcareous polychaetes encrusting worms and two pelecypods remain unidentified. The gastropods and polychaetes reinforce a brackish-water environment.

A small fish spine collected from interval C above the Lawlor Tuff was studied by G.R. Smith (2010, written commun.). It is identified as a medial spine, probably a dorsal spine of a mullet (Mugillidae) or a silversides (Atherinpsidae). The spine is broken on one side but is extremely slender and was probably long before it was broken. An atherinid, Colpichthys regis, is known from the Bouse Formation in the Blythe basin, and this spine may represent the same species. Another possible fish that has long, extremely slender dorsal spines is a baby striped mullet, Mugil cephalus. Both species are marine and/or estuarine, although they can be found in saline lakes. The spine matches the estuarine forms of atherinids much more closely than fresh-water forms.

Ichnofossils

Tracks and burrows were found in thin sandy limestone beds of interval A (Table 3) and in thin-bedded mudstone at the mudstone locality (Fig. 2). The mudstone yielded distinct wading bird tracks that are referable to ichnogenera Alaripeda and Avipeda (Sarjeant and Reynolds, 2001), as well as Aviadactyla (Reynolds, 2012). Modern shorebirds of these genera, such as killdeer, sanderling, and sandpiper, inhabit shores of lakes and seas. The killdeer nest in meadows and dry uplands. The sanderling mostly occurs along ocean beaches, nesting in moors and marshes. The sandpiper mostly occurs along beaches, tidal flats, and marshes, and nests close to water (Sibley, 2003). Horizontal and vertical burrows in these mudstone beds apparently were made by invertebrates such as shrimp, crabs, and isopods (Reynolds, 2012). Beach and tidal-flat environments are indicated.

Synthesis of Paleontology

Ostracodes and diatoms both indicate substantial transport and carbonate encrustation of the sediment. Stromatolitic mounds, ichnofossils, and the diatom taxa indicate shallow water. The observed mix of fresh-water and brackish-water species throughout the section might be explained by mixing of two environments in close proximity. Wading bird tracks reinforce the shallow-water interpretations for parts of the section and strongly support beach or tidal flats. The fish bone from interval C apparently requires a marine or estuarine environment. Mollusks and polychaetes indicate a brackish-water setting.

DISCUSSION

Physical Stratigraphy

The reference section is composed of limestone with admixtures of locally sourced arkosic sand and fine gravel. Limestone consists of fine crystalline matrix, and in places appears to be clastic, because sand-sized clots of calcite matrix have contrasting color. Fragments of shell and stromatolites occur in some beds. The reworked carbonate suggests a carbonate-productive shore zone, almost reef like, with material being broken and redistributed by waves to the foreshore and offshore. No actual reefs were present, because calcite fragments are present in the wave zone, not in a back-reef environment (Braithwaite, 2014). Carbonate production is consistent with highly evaporative conditions. All deposition took place in a shallow-water environment, based on sedimentary structures and grain size. Thin, laterally continuous beds suggest deposition on a gently dipping beach foreshore, consistent with fetch-limited beaches (Nordstrom and Jackson, 2012). Grain-size changes indicate a reduction of wave energy from the base of the section upward into interval C, and then sediment coarsening and increasing energy upward. The initial transgression of the water body may have been rapid, as no coarse beach deposits were reworked from underlying alluvial fan gravel, and stromatolites formed on that gravel require shallow water depths and clear water conditions for suitable transmission of sunlight. Increasing amount of arkosic sand near the base and top of the section reflect shallowing water, and deposits of interval C apparently represent deepest water, near wave base (Table 3). However, the trend toward shallower water conditions in the top of the section may be governed primarily by alluvial influx, rather than actual water-level decline. Increasing arkosic sand and gravel upward in interval G, followed upward by coarse, poorly sorted gravel of interval H, suggests progradation of alluvial fans. The repeated occurrence of better sorted sand beds and intervals of carbonate mounds in interval H suggests that water level remained high while fan gravels built across the previously deposited waterlain sediment.

The 4.83 Ma Lawlor Tuff was deposited on stromatolite mounds, indicating a tephra-fall component, but the upper part of the ash bed adjacent to mounds is faintly bedded and impure, suggesting redeposition, perhaps from a nearby fluvial source that brought ash from a wide area of the subaerial landscape. Detrital zircons in the ash reinforce this interpretation (Harvey, 2014).

Ripples in several parts of the section indicate deposition above wave base and sediment transport to the north; similarly, slump folds indicate mass failure directed offshore to the south. There is no direct evidence for repeated fluctuations in water level as might be expected for a closed lake basin with no threshold control, nor are there structures suggestive of tidal conditions in a marine setting. The lack of unconformities, other than the truncation of folds caused by slumping in interval E, and the lack of indications for subaerial plant encroachment and soil development suggest that prolonged water-level drawdowns did not occur.

The presence of bird tracks in the gypsiferous mudstone locality identifies these sediments as nearshore, possibly tidal mudflat or coarse-sediment–starved muds. Unfortunately, the mudstone cannot be correlated to the main section and its interpretation is uncertain in terms of age and position in the Bouse Formation. It is clear that mudstones are not restricted to basin-center deep-water environments, as was suggested by Metzger (1968).

Fossils and Environment

Poor preservation of microfossils owing to leaching, breakage, abrasion, and overgrowths indicate transport with sediment, probably along shore in the wave zone. The benthic diatom taxa support sediment inferences for a shallow standing-water environment. Tracks of wading birds in the mudstone indicate mud flats, consistent with the estuarine setting suggested by the fish spine and some diatoms or with a lake with fluctuating water levels. Other fossils indicate that both fresh- and saline-water habitats were present in the general area, as microfossils that support both environments were mixed and deposited throughout the section. One scenario would be an estuary or brackish lake with a perennial large stream nearby, which would provide local fresh-water habitat. (We consider an estuary to be a semienclosed coastal water body having a connection to the open sea and within which seawater is diluted with fresh water from land drainage; see Cameron and Pritchard, 1963.) In the case of a possible lower Bouse estuary, it may have been very strongly influenced by input of Colorado River waters.

Ostracodes present throughout indicate saline lake to marine environments, and although not compatible with fresh Colorado River water chemistry, would be consistent with highly evaporative Colorado River water in a partially restricted arm of a lake. A few ostracodes in the central part of the section represent fresher lake waters or transport from a proximal fresh-water environment. Diatoms provide similar environmental clues, i.e., fresh to marine conditions, mixed during transport of the fossils. It is interesting that epiphytic species indicate that macrophytes were present throughout the section, and thus the water body did not exceed salinity tolerance levels for aquatic plants.

Molluscan fauna is depauperate, unlike the diverse fauna typical of coastal estuaries and fresh-water settings. The restricted fauna may indicate brackish, hostile environments such as those associated with seasonal increases in salinity and temperature in a restricted arm of a lake. Polychaetes are versatile organisms, although ∼90% of species are known from marine waters. Stromatolites and microbialites (e.g., Noffke and Awramik, 2013) may be responsible for carbonate sand-sized particles making up much of the sediment, but other sources, such as detritus from macrofossils and inorganic precipitates, cannot be ruled out. Sand and pebble grains are not carbonate coated, suggesting that inorganic precipitation was unimportant.

Age

Chemical correlation with the 4.83 Ma Lawlor Tuff (Sarna-Wojcicki et al., 2011) indicates an early Pliocene age for the Amboy sediments. A similar tephra found in the lower part of the section of the Bouse Formation at ∼310 m elevation in the Chocolate Mountains (Fig. 1; Metzger, 1968) also correlates with the Lawlor Tuff. Dating by U-Pb of zircon and zircon isotopic tracers for these ash beds confirms the Lawlor correlation (Harvey, 2014). This early Pliocene time is very close to the Hemphillian-Blancan North American Land Mammal Age boundary, providing an opportunity to refine timing by studying vertebrate fossils.

Correlations of tephra recovered from deep cores in Bristol Lake indicate that the basin started filling before 3.7 ± 0.2 Ma (Rosen, 1991, 1992). Conservative estimates based on depositional rates suggest that the basin might have started filling earlier than 6 Ma (Rosen, 1992). However, deposits equivalent to those reported herein were not described, and the Lawlor Tuff was not recovered in borings. The borings that penetrated the oldest sediment penetrated a lower section of alluvial fan sediments that lacked datable materials. It is possible that unconformities within the alluvial fan sequence cut out the Bouse deposits. Marine or lacustrine deposits containing a brackish-water fauna have been documented from beneath Cadiz Lake (Fig. 1), suggesting that the Bouse Formation might be present in that basin (Brown and Rosen, 1992). Fauna and stratigraphy strongly suggest that that the Bouse Formation is present below 80 m in Danby Lake (Brown and Rosen, 1992; McDougall, 2008).

Basin Hypsometry

Most study of the lower Bouse Formation was conducted in the Blythe basin, the broad, deep basin with thick Miocene and Pliocene deposits that is now transected by the Colorado River (Fig. 1). West of that basin are two broad low areas (arms) that would be inundated by water levels indicated by the highest Bouse Formation exposures (∼330 m; Spencer et al., 2008, 2013) if Pliocene topography was similar to modern topography. Although the active ECSZ in most of the area west of the Blythe basin requires that Pliocene topography differed from the modern, the coincidence of several long-term depocenters with modern topographic lows (Richard, 1993; Jachens et al., 2002; Langenheim and Powell, 2009) suggests that some low areas have persisted for long periods of time. No information is available for the persistence of the topographic divides over the past 5 m.y. Here we explore potential sequences of flooding events that would be expected to occur during a hypothetical rapid rise in water level during Bouse Formation deposition.

One western arm coincides with the broad valleys marked by low points at Ford and Palen dry lakes (Fig. 1). This arm would be inundated as water levels reached ∼138 m, adding considerable evaporative loss to a deepening water body as it spread westward.

The other western arm has a threshold level of ∼271 m next to the West Riverside Mountains. Water filling west of that threshold would cross another shallow threshold near the location of Rice, California, inundating the Danby Lake area. Continuing rise of water level would then cross the threshold north of the Iron Mountains (∼295 m), and inundate the Cadiz Lake and Bristol Lake area, which we term the Bristol basin. A modern low divide at the northern Calumet Mountains separates the Cadiz Lake and Bristol Lake watersheds (Bishop, 1963; Howard, 2002). The presence of the Bouse Formation in the Bristol basin indicates that water inundations somewhat like that described here occurred during Pliocene time, although we do not know what elevations the thresholds between basins were.

The fact that the Bristol basin is located far west of either marine-water input or fresh-water input from the Colorado River may provide paths for better understanding the Bouse water body. If the water body was a brackish lake that became increasingly saline in a closed-basin configuration before finally overtopping a paleodam at the Chocolate Mountains (e.g., Spencer et al., 2013), the waters in the Bristol basin would have had limited exchange with waters of the Blythe basin, and that should have resulted in significantly different temperature, salinity, chemistry, and fauna in the two subbasins. Water level in the Bristol basin would have fluctuated seasonally and on longer time scales, and these fluctuations would have been magnified by water level dropping below internal thresholds, stranding it and enhancing drawdown. Evidence for strongly saline waters and for significant water-level fluctuations is lacking in the Bouse Formation near Amboy. The only evidence for hypersaline conditions comes from core logs from Cadiz and Danby Lakes (Brown and Rosen, 1992, Fig. 3 therein) where evaporite deposits overlie strata referred to as “Bouse sediments” (Brown and Rosen, 1992, 1995) that contain microfossils suitable for brackish-water conditions. The evaporite deposits apparently reflect the termination of the Bouse water body. Although evidence for strongly saline waters in the Bristol basin are absent, we have recorded different fauna at Amboy compared to the Blythe basin and the Sr isotopic values also differ somewhat in being more enriched than Bouse carbonates collected along the Colorado River corridor (Spencer et al., 2013). These lines of evidence support restricted circulation in the Bristol basin.

If the Bouse water body were estuarine, water levels would not be subject to seasonal climatic fluctuations, but would be driven by a combination of sea-level change, tidal currents, and tectonics. Tides would be expected, although they tend to be reduced in segmented water bodies (Nordstrom and Jackson, 2012). Tidal action would fairly effectively mix water from basin to basin, making the estuarine environment more uniform than that of a saline lake. Although the general similarity of deposits at Amboy with those of the Blythe basin (Reynolds et al., 2008; Spencer et al., 2013) supports the estuarine interpretation, more detailed study of environmental differences among the subbasins may be fruitful. For example, Sr isotopes of the Bristol basin deposits are more similar to those of the Blythe basin than modern seawater (Spencer and Patchett, 1997; Roskowski et al., 2010), but most values are enriched above those for the Blythe basin. Can a highly evaporative arm of a large estuary or more groundwater contributions attain these values? Carefully posed hypotheses that compare a restricted arm of a river-dominated estuary versus a saline lake may be testable by detailed isotopic study. In addition, duration of the water body should be quite distinctive; Spencer et al. (2013) estimated a 30 k.y. duration for a saline lake, whereas an estuary would probably have a duration several orders of magnitude greater.

Tectonics and Basin Evolution

Several deposits that potentially correlate with the Bouse Formation are in the subsurface of the southeastern Mojave Desert. In addition to the deposits north of Amboy described herein, deposits intersected in borings under Cadiz and Danby Lakes have been suggested to be the Bouse Formation (Bassett et al., 1959; Brown and Rosen, 1992; McDougall, 2008) based on sedimentology and fossils. Near Pinto Basin (Fig. 1; east of Eagle Mountain), borings conducted by GeoPentech in 2003 (State Water Resources Control Board, 2010) encountered mudstone >300 m thick at surface elevations below 300 m. Although no limestone was described from this area, only one borehole penetrated the mudstone to underlying bedrock. The mudstone deposits may correlate with the Bouse Formation. Near Ford Dry Lake (Fig. 1), borings have been interpreted as penetrating Bouse Formation greater than 600 m thick (Bureau of Land Management, 2010).

The deposits we studied are more than 600 m higher than Pliocene sediments that are buried under Bristol Lake playa 5 km southeast of Amboy. The exposed Bouse rocks are folded and faulted, and generally dip southward 20°–30°. Exposures are sparse, offering few constraints for tectonic interpretations, but the association with northwest-striking dextral faults that were active during the Quaternary (Howard and Miller, 1992; Bedford et al., 2010) suggests an origin as a releasing bend or transtensional stepover in the ECSZ. The most likely candidates are the south Bristol Mountains and Broadwell Lake faults (Fig. 1), which strike northwest along the southwest side of the Bristol Mountains and in the Lava Hills, respectively, and project southeastward into the Bristol Lake area. The south Bristol Mountains fault apparently forms a barrier between saline water in Bristol Lake and fresh groundwater to the east (Rosen, 1992, 2000). In addition, northwest-striking faults mapped in the Calumet Mountains (Bishop, 1963; Howard, 2002) may extend into the basin and form the southern counterparts for a dextral releasing bend. Although a broad cover of Holocene distal fan deposits (Bedford et al., 2010) adjacent to Bristol Lake masks details of buried faults and folds, it seems clear that post-Bouse tectonism is well represented in the Bristol basin. The Bristol-Granite fault farther east is less well understood because it was last active during the early Pleistocene (Bedford et al., 2010), but it projects to the basin divide north of the Iron Mountains and could control its elevation.

An improved understanding of the tectonics of the greater Blythe basin may provide insight into the aquatic environment of the Bouse Formation. If the area has been little deformed since the time of Bouse Formation deposition, the high elevation of the deposits (∼330 m; Spencer et al., 2008) strongly favors a lacustrine origin. In contrast, substantial uplift of the region over the past 5 m.y. would accommodate the estuarine interpretation. The middle Pliocene was a time of high sea levels, perhaps from ca. 4.6 to 3.5 Ma (Wardlaw and Quinn, 1991) and probably extending through the Pliocene optimum to ca. 2.9 Ma (Raymo et al., 2009). However, Pliocene sea levels were no higher than ∼40 m above present levels, more likely 22 ± 5 m (Miller et al., 2012), and therefore cannot account for more than a small fraction of the elevation differential observed for the Bouse Formation, even if high sea levels coincided with deposition.

The temporal and spatial patterns of uplift and its causes are beyond the scope of this paper, but we note that commonly cited causes for uplift include both mantle processes that change buoyancy of the lithosphere, and erosion accompanied by isostatic uplift.

In particular, uplifts flanking rift zones are common (Chéry et al., 1992) and might be expected adjacent to the Salton Trough, which initiated in late Miocene (Lonsdale, 1989; Stock and Hodges, 1989) to early Pliocene time (Shirvell et al., 2009; Dorsey et al., 2011). The Eastern Transverse Ranges may owe part of their high elevation to this origin. Uplift of the southeastern Mojave Desert driven by erosional denudation seems to be unlikely because climate became increasingly arid in the late Pliocene (Peryam et al., 2011) and because small Pliocene basins exist within the area we studied, indicating that most eroded materials were stored locally. Similar conclusions from locations farther northwest in the Mojave Desert (Cox et al., 2003; Miller et al., 2011) reinforce this interpretation.

Epirogenic regional uplift of the southern Mojave Desert may have occurred since the early Pliocene. This area was tectonically active as part of the ECSZ, including deformation of Bouse sediment in the Blythe basin (Richard, 1993). Studies of geomorphology (e.g., Karlstrom et al., 2011) and thermochronology (e.g., Hoffman et al., 2011; Lee et al., 2013) conclude that significant erosion and uplift have occurred on the western Colorado Plateau in the past ca. 6–10 m.y., but inferred causes for denudation differ, with both erosional and tectonic uplift origins invoked. As these studies are expanded to include areas west within the Basin and Range province, we will be able to better understand the Pliocene tectonic evolution of the Mojave Desert.

Fossils and isotopic composition of the Bouse Formation have inherent ambiguity and do not resolve the question of whether the lower Bouse Formation was in part estuarine. Evaluating offset along faults in the ECSZ and regional lithospheric changes associated with the shoulder of Salton rift may help resolve the tectonics of the last 5 m.y. to address the question of whether the required uplift of estuarine deposits could have occurred. Understanding of tectonics can be improved by more detailed investigations of faults. Hypsometry and basin dynamics can be studied by examining Bouse sediment near divides to test whether enhanced tidal currents (Nordstrom and Jackson, 2012) or catastrophic inflow deposits (House et al., 2008) are present. Shallow geophysical methods can help with both paths of inquiry. In addition, study of Colorado River gravels overlying the Bouse Formation may permit us to test whether uplift took place after Bouse deposition.

CONCLUSIONS

Exposures of sediments north of Amboy strongly resemble Bouse Formation outcrops in the Blythe basin along the Colorado River, including similar elevation, age (presence of the Lawlor Tuff, 4.83 Ma), and lithology. This lower Bouse basin is, in turn, similar in lithology to higher basins farther north along the Colorado River, basins that have been shown to be lacustrine. The similarity of sediment in all basins argues for a strong Colorado River signal, a finding that is supported by similar Sr isotopic signals throughout. Fossil diatoms, ostracodes, and mollusks at Amboy indicate deposition in brackish waters in either a saline lake or estuary, and also include a few fresh-water forms. Wading bird tracks indicate beach or tidal environments, and a broken fish spine belongs to one of two genera that favor marine habitat.

If the Pliocene hypsometry resembled modern topography, western arms of the Bouse water body would have had restricted circulation with the main body along the modern Colorado River, and would have been isolated by small drops in water level. Evidence for isolation and evaporite deposition is lacking in these exposures, but fossil and Sr isotopic evidence suggests that water chemistry in the Bristol basin differed from the main Blythe basin. Carefully contrasting the basins, employing indicators for aquatic environment and water chemistry, may help to discriminate between estuarine and lacustrine origins for the lower Bouse Formation. Study of the topographic divides, particularly their longevity, would also help to frame the discussion.

We thank Becky Dorsey, Rick Forester, Keith Howard, George Jefferson, Kris McDougall, Michael Rosen, and Jon Spencer for sharing their ideas on the Bouse Formation. We appreciate the tephra identification by Andrei Sarna-Wojcicki, bone identification by Gerry Smith, and Kris McDougall’s willingness to hunt for foraminifera. P. LaFollette, M. Roeder, T. Howe, P. Riseley, and C. Powell studied mollusks. J.R. Miller helped evaluate habitats for modern shorebirds. Reviews by Becky Dorsey, Kyle House, Phil Pearthree, an anonymous reviewer, and editor Karl Karlstrom significantly helped us improve this paper.