Abstract

Thick late Miocene nonmarine evaporite (mainly halite and gypsum) and related lacustrine limestone deposits compose the upper basin fill in half grabens within the Lake Mead region of the Basin and Range Province directly west of the Colorado Plateau in southern Nevada and northwestern Arizona. Regional relations and geochronologic data indicate that these deposits are late synextensional to postextensional (ca. 12–5 Ma), with major extension bracketed between ca. 16 and 9 Ma and the abrupt western margin of the Colorado Plateau established by ca. 9 Ma. Significant accommodation space in the half grabens allowed for deposition of late Miocene lacustrine and evaporite sediments. Concurrently, waning extension promoted integration of initially isolated basins, progressive enlargement of drainage nets, and development of broad, low gradient plains and shallow water bodies with extensive clastic, carbonate, and/or evaporite sedimentation. The continued subsidence of basins under restricted conditions also allowed for the preservation of particularly thick, localized evaporite sequences prior to development of the through-going Colorado River.

The spatial and temporal patterns of deposition indicate increasing amounts of freshwater input during the late Miocene (ca. 12–6 Ma) immediately preceding arrival of the Colorado River between ca. 5.6 and 4.9 Ma. In axial basins along and proximal to the present course of the Colorado River, evaporite deposition (mainly gypsum) transitioned to lacustrine limestone progressively from east to west, beginning ca. 12–11 Ma in the Grand Wash Trough in the east and shortly after ca. 5.6 Ma in the western Lake Mead region. In several satellite basins to both the north and south of the axial basins, evaporite deposition was more extensive, with thick halite (>200 m to 2.5 km thick) accumulating in the Hualapai, Overton Arm, and northern Detrital basins. Gravity and magnetic lows suggest that thick halite may also lie within the northern Grand Wash, Mesquite, southern Detrital, and northeastern Las Vegas basins. New tephrochronologic data indicate that the upper part of the halite in the Hualapai basin is ca. 5.6 Ma, with rates of deposition of ∼190–450 m/m.y., assuming that deposition ceased approximately coincidental with the arrival of the Colorado River. A 2.5-km-thick halite sequence in the Hualapai basin may have accumulated in ∼5–7 m.y. or ca. 12–5 Ma, which coincides with lacustrine limestone deposition near the present course of the Colorado River in the region.

The distribution and similar age of the limestone and evaporite deposits in the region suggest a system of late Miocene axial lakes and extensive continental playas and salt pans. The playas and salt pans were probably fed by both groundwater discharge and evaporation from shallow lakes, as evidenced by sedimentary textures. The elevated terrain of the Colorado Plateau was likely a major source of water that fed the lakes and playas. The physical relationships in the Lake Mead region suggest that thick nonmarine evaporites are more likely to be late synextensional and accumulate in basins with relatively large catchments proximal to developing river systems or broad elevated terranes. Other basins adjacent to the lower Colorado River downstream of Lake Mead, such as the Dutch Flat, Blythe-McCoy, and Yuma basins, may also contain thick halite deposits.

INTRODUCTION

Thick and widespread evaporite deposits can develop in internally drained basins within continental rift settings in arid environments, as extension commonly fragments continental crust into numerous fault blocks and attendant half grabens. During the early phases of rifting, drainage networks may be relatively small and limit the extent and thickness of evaporite deposition. However, broader drainage networks integrating multiple basins typically evolve over time and may be coupled with regional subsidence, which collectively promote accumulation of more extensive evaporite sequences (e.g., McKenzie, 1978; Royden et al., 1980; Sclater and Christie, 1980; Bally and Oldow, 1984). Marine incursions into restricted embayments may further accelerate this process. Extensive syn- to postextensional evaporite deposits are therefore common in continental rifts, including many passive continental margins (Holwerda and Hutchinson, 1968; Lowell and Genik, 1972; Jackson and Seni, 1983; Tankard and Balkwill, 1989). For example, large provinces of extensive evaporite deposition in extensional settings are well documented especially in relatively arid regions, including the Gulf of Suez (Evans, 1988; Steckler et al., 1988), East Africa (Catuneanu et al., 2005; Chorowicz, 2005; Ebinger, 2005), Angola (Hudec and Jackson, 2002, 2004; Dickson et al., 2003), Gulf of Mexico (Dickinson, 2009; Pindell and Kennan, 2009; Stern et al., 2010), and Brazil (Meisling et al., 2001; Japsen et al., 2012).

Most of these evaporite provinces were largely derived from marine-fed water bodies, but some are related to extensive arid-region lake and playa deposits (Fig. 1), as exemplified in East Africa (Chorowicz, 2005; Ebinger, 2005) and the Dead Sea (Niemi et al., 1997). Depending on water composition and evaporation rates, evaporites may fill developing accommodation space very rapidly and can become particularly thick if the basins continue to subside. In particular, halite can collect at rates in excess of 10 cm/yr (Schreiber and Hsü, 1980). Evaporites deposited in lakes are commonly quite similar in appearance to many marine-fed deposits except for differences in trace-element content and the lack of tides affecting deposition. Generally, lake sedimentation is the product of water with entrained sediment inflow, but in arid regions evaporation rates become more important and water influx may become restricted to groundwater and springs, bringing little or no clastic load. In this case, the ionic content, pH, and rate of evaporation of the water become paramount controls (Benison et al., 2007), as described in models for basin filling presented in Lowenstein and Hardie (1985) and further developed in Renaut and Gierlowski-Kordesch (2010).

Although not as well documented, the thickness and extent of Miocene to Quaternary evaporite deposits in the Basin and Range Province within the southwestern United States (Fig. 2) may rival those of some world-renowned regions of evaporite deposition (e.g., Peirce, 1976, 1981; Johnson and Gonzales, 1978; Smoot and Lowenstein, 1991; Faulds et al., 1997; Rauzi, 2002a, 2002b). Many basins in this region contain substantial volumes of carbonate and Ca-sulfate deposits (anhydrite and gypsum), but some basins contain substantial amounts of halite. Particularly thick or widespread halite deposits occupy the Overton Arm, Detrital, and Hualapai basins in the Lake Mead region of southern Nevada and northwestern Arizona (Fig. 3; Mannion, 1963; Faulds et al., 1997), Luke and Picacho basins of central Arizona (Peirce, 1976), and Bristol Lake basin in southeastern California (Fig. 2; Rosen, 2000). Some of the halite deposits have little or no carbonate and limited amounts of sulfate, environments much like those found today in evaporative lakes in southern Western Australia (Benison et al., 2007; Jagniecki and Benison, 2010).

The regional setting and abundant textural and geochemical evidence indicate that most of the Miocene to Quaternary evaporite deposits within the Basin and Range Province were deposited in continental playas and shallow lakes in enclosed basins as opposed to a marine origin (Peirce, 1976; Rosen, 1991; Faulds et al., 1997). Within Arizona and neighboring parts of Nevada and California, these evaporites are inferred to be middle Miocene to early Pliocene, but definitive age constraints are sparse. Detailed sedimentological studies of the Quaternary Searles Lake (Smith, 1979, 2000) and Bristol Dry Lake evaporites (Rosen, 1991, 2000) provide important contemporary analogues for interpreting older and buried evaporites in the region. Lenticular wedge-shaped geometries of some of the evaporite deposits, such as the Quaternary Glauber salt bed of the Great Salt Lake (Eardley, 1962; Mikulich, 1971) and Red Lake halite in the Hualapai basin of northwestern Arizona (Faulds et al., 1997), suggest a synextensional origin for at least some of these evaporite deposits.

The halite deposits and related evaporites have important paleogeographic implications for the southwestern United States. They record a period of internal drainage that predates integration of most of the region by the Colorado River and its tributaries. As such, they provide a paleogeographic marker that can provide important temporal and spatial constraints on overall drainage evolution. In the Lake Mead region, for example, thick and widespread, late Miocene evaporite and lacustrine carbonate deposits appear to immediately predate arrival of the Colorado River. However, the temporal and spatial relationships between the evaporite and lacustrine deposits and the relevance of widespread late Miocene evaporite deposits in the Lake Mead region (e.g., halite, gypsum, and anhydrite) to development of the Colorado River have not been studied systematically. Further, the overall size and extent of the deposits in individual basins or groups of interconnected basins may reflect the relative length of time of the ponding episode and/or overall size of the catchment area. Such relations can be greatly impacted, however, by paleoclimate variations, which significantly affect the freshwater influx from adjacent ranges and potential for postburial halite dissolution.

Neogene halite deposits of the southwestern United States also have important economic implications, having been exploited for mineral salt production, evaluated for low-temperature geothermal resources, and assessed for storage of natural gas, natural gas liquids, compressed air energy, and nuclear wastes (Netherland and Sewell, 1977; Johnson and Gonzales, 1978; FERC, 1982; Rauzi, 2002b; Kostick, 2013). In the Lake Mead region, halite deposits in the Overton Arm area (Fig. 3) were exploited in several mines prior to the filling of Lake Mead and were analyzed for their potential effects on the water quality of the reservoir (U.S. Bureau of Reclamation, 1950). To the south of Lake Mead in northwestern Arizona, studies of the Red Lake halite in the Hualapai basin for gas storage date back to efforts by Southwest Gas Corporation in the mid-1950s (Neal and Rauzi, 1996; Rauzi, 2002b). Additionally, various developers have proposed and, in some cases, initiated permitting of natural gas storage projects at Red Lake.

The purpose of this paper is therefore to (1) provide an overview of the distribution and origin of late Miocene lacustrine and evaporative deposits (mainly halite) in major basins in the Lake Mead region; (2) review age constraints on these deposits, including existing data from previous efforts and several new 40Ar/39Ar dates and tephrochronologic correlations presented herein; (3) discuss the paleogeographic implications of these deposits, particularly in terms of elucidating the evolution of the Colorado River and the depositional environment characterizing the Lake Mead region immediately prior to integration by the river; and (4) consider whether other basins in the region may contain previously unrecognized thick halite deposits through a review of gravity and magnetic data. Our study area includes the broader Lake Mead region, which for the purposes of this study includes the area extending from Las Vegas Valley on the west to the Grand Wash Cliffs along the western margin of the Colorado Plateau on the east, as well as basins extending ∼75–100 km north and south of the present course of the Colorado River (Fig. 3). The regional geologic setting, geologic maps, well data in some areas, detailed gravity data (Fig. 4) with derivative depth-to-basement maps (Fig. 3; Langenheim et al., 2001a, 2001b, 2010), and aeromagnetic data were incorporated into our analysis of late Miocene basin-fill sedimentary sequences in the Lake Mead region. For the purposes of this paper, a basin is defined as a discrete structural entity bounded on at least one side by a major fault and containing relatively thick (>300 m) sequences of basin-fill sediments.

REGIONAL GEOLOGIC SETTING: CENOZOIC PALEOGEOGRAPHY

The Cenozoic paleogeographic evolution of the region is marked by a 180° drainage reversal that reflects an evolving tectonic setting. In the early Tertiary (Paleocene to Oligocene), streams flowed northeastward from broad highlands to the west of the Colorado Plateau into vast lowlands in the Colorado Plateau region (e.g., Potochnik and Faulds, 1998; Wernicke, 2011), as evidenced by widespread, southwesterly derived Paleocene–Eocene gravels along the western margin of the Colorado Plateau (Young, 1982). These highlands were built by crustal shortening and arc magmatism associated with the Sevier and Laramide orogenies. Extensive early Tertiary erosion stripped much of the sedimentary and volcanic cover from these highlands.

The highlands foundered in middle Tertiary time (Oligocene to Miocene) as the convergent plate margin transitioned into a transform boundary (e.g., Atwater and Stock, 1998), and extensional to transtensional tectonism took hold within the Basin and Range Province (e.g., Wernicke, 1992). In the Lake Mead region, extension generally began in early to middle Miocene time, as evidenced by tilt fanning (i.e., progressive upward decrease in tilt) in half grabens and exhumation of the footwalls of major normal faults (e.g., Anderson et al., 1972; Faulds et al., 1992, 2001a, 2001b; Beard, 1996; Duebendorfer and Sharp, 1998; Fitzgerald et al., 2009). As large-magnitude east-west extension swept through the region in the early to middle Miocene, major detachment faults and arrays of upper-plate normal faults dissected the former highlands, fragmenting them into panels of northerly trending fault blocks. In addition, major strike-slip faulting (Anderson, 1973; Longwell, 1974; Bohannon, 1984; Duebendorfer and Simpson, 1994; Duebendorfer et al., 1998; Fryxell and Duebendorfer, 2005; Umhoefer et al., 2010b) and a component of north-south shortening (Anderson et al., 1994; Duebendorfer and Simpson, 1994) further deformed the Lake Mead region north of the Colorado River in late Miocene time. As the region foundered and subsided during the early to late Miocene, the early Tertiary northeast-flowing drainage system was disrupted, and a long period of internal drainage ensued. Enclosed basins (generally half grabens) began filling with locally derived sediment, as evidenced by abundant early to middle Miocene fanglomerate, fluvial, and lacustrine deposits (e.g., Bohannon, 1984; Beard, 1996; Hickson et al., 2010; Lamb et al., 2010). Throughout much of the region (particularly areas south of Lake Mead), voluminous intermediate volcanism accompanied the onset of extension, transitioned to widespread bimodal magmatism during peak extension, and was followed by more isolated alkalic volcanism during the latter stages of rifting (Faulds et al., 1995, 2001a; Gans and Bohrson, 1998). Thus, some half grabens filled with thick sequences (>1 km) of volcanic rock.

By late Miocene to Pliocene time, the southern reaches of the transform boundary jumped inland to the eastern side of the Peninsula Ranges, the Gulf of California developed (Stock and Hodges, 1989; Oskin and Stock, 2003; Fletcher et al., 2007), and extension and strike-slip faulting waned in the Lake Mead region. The reduction in strain rates promoted widespread aggradation within composite basins, and a rugged middle Miocene topography of tilt blocks was buried by sediments in multiple basins, as evidenced by gravity studies (Langenheim et al., 2010). Although most of the region has been integrated into the Colorado River, facies patterns in late Tertiary basin fill are generally congruent with modern topography, as alluvial fans derived from flanking ranges interfinger with and give way to floodplain, lacustrine, and/or continental playa environments toward the basin centers.

Reduced strain rates and regional aggradation in late Tertiary time facilitated evolution of regional drainage systems that ultimately integrated large networks of basins. Because much of the region, especially large composite basins, had subsided beneath the level of the Colorado Plateau and base level lowered to sea level as the Gulf of California opened to the south, the developing late Miocene to Pliocene drainage network ultimately had an outlet to the south and southwest, opposite to the early Tertiary northeast-flowing drainage.

The Colorado River ultimately became the preeminent drainage system in the southwestern United States, as it captured much of the Rocky Mountains, most of the Colorado Plateau, and large portions of the Basin and Range Province (Fig. 2). Although controversy still surrounds its long-term evolution and the origin of the Grand Canyon (e.g., Flowers and Farley, 2012, versus Karlstrom et al., 2014), a consensus is developing that the lower Colorado River (i.e., from Lake Mead southward) developed through sequential south-directed filling and spilling of a chain of lakes through formerly closed basins (Spencer and Patchett, 1997; House et al., 2008) between ca. 6 and 5 Ma (Faulds et al., 2001c; Dorsey et al., 2007; House et al., 2008). As each lake spilled over to the south, the Colorado River lengthened in its wake. Limestone and siltstone of the Bouse Formation record the trail of lakes in multiple basins extending south from the Lake Mead region to the Blythe area (Spencer and Patchett, 1997; House et al., 2008; Spencer et al., 2008, 2013).

Substantial previous work has focused on the deposits left by the precursor lakes and Colorado River along the axial part of the drainage, including the Bouse Formation (e.g., Spencer et al., 2008, 2013) and fluvial deposits of the Colorado River (House et al., 2005, 2008), but relatively little research has addressed whether the evolving regional drainage system and ultimately the Colorado River also left behind a series of markers in the form of evaporite deposits in both axial and satellite basins (e.g., Faulds et al., 1997). We therefore evaluate the overall distribution and apparent age of known evaporite (e.g., halite) and related lacustrine deposits in the late Miocene basin fill, as well as evidence for possible undiscovered thick sequences of halite in the Lake Mead region. Although early to middle Miocene sedimentary deposits are widespread and also record key elements in the Cenozoic structural and paleogeographic evolution of the region (e.g., Bohannon, 1984; Beard, 1996; Umhoefer et al., 2010a; Hickson et al., 2010; Lamb et al., 2010, 2015), this paper focuses on the late Miocene sections due to their potential relevance to development of the Colorado River.

Lake Mead Region

In northwestern Arizona and southern Nevada, the Colorado River crosses an abrupt boundary between the Colorado Plateau and Basin and Range Province (Figs. 2, 3, and 5). Essentially flat, relatively unextended strata on the high-standing Colorado Plateau give way to moderately to steeply tilted fault blocks in the Basin and Range Province across a system of west-dipping normal faults, including the Grand Wash fault zone (Lucchitta, 1966, 1979) and South Virgin–White Hills detachment fault (Duebendorfer and Sharp, 1998; Brady et al., 2000; Fig. 3). The conspicuous, west-facing fault-line escarpment of the Grand Wash Cliffs, consisting of subhorizontal Paleozoic strata, marks the western margin of the Colorado Plateau within the footwall of the Grand Wash fault zone and rises ∼1.3 km above several east-tilted half grabens in the Basin and Range Province, including the Grand Wash Trough and Hualapai basin. Major extension progressed northward across this region in the Miocene, with the main episode occurring ca. 16–13 Ma to the south of Lake Mead and in the eastern Lake Mead region (Anderson et al., 1972; Faulds et al., 1992, 2008, 2010; Beard, 1996; Duebendorfer and Sharp, 1998) and ca. 13–9 Ma in the western Lake Mead area (Duebendorfer and Simpson, 1994; Harlan et al., 1998; Castor et al., 2000). Extension had generally ceased by ca. 6 Ma, especially to the south of Lake Mead. However, widely spaced Quaternary faults cut the northern part of the Lake Mead region (U.S. Geological Survey, 2010) and include segments of the Grand Wash, Wheeler Ridge, and Piedmont faults, which bound the Grand Wash Trough, Gregg basin, and Mesquite basin, respectively (Fig. 3).

The Colorado River emanates from the Grand Canyon within the western part of the Colorado Plateau and traverses the Lake Mead region from east to west orthogonal to the northerly trending structural grain (Figs. 2, 3, and 5). The origin of the Grand Canyon and Colorado River in this region has been the subject of significant study and debate (e.g., Powell, 1875, 1895; Blackwelder, 1934; Longwell, 1946; Hunt, 1956; Lucchitta. 1966, 1972, 1979, 1989; Lovejoy, 1980; Young and Spamer, 2001; Hill et al., 2008; Wernicke, 2011; Flowers and Farley, 2012; Karlstrom et al., 2014). Recent work has refined the timing of inception of the Grand Canyon and Colorado River (Spencer et al., 2001, 2013; Faulds et al., 2001c; House et al., 2005; Dorsey et al., 2007) and models for drainage development (Spencer et al., 2001, 2013; House et al., 2005, 2008; Crossey et al., 2015).

Several basins in southern Nevada and northwestern Arizona record a transition from dominantly lacustrine to through-flowing fluvial deposition between ca. 6 and 4.5 Ma that marks the arrival of the Colorado River (Faulds et al., 2001c; House et al., 2005, 2008). Late Miocene lacustrine carbonate deposits, in particular, chronicle a westward progression of lakes from ca. 12 to 5.5 Ma across the Lake Mead region. These lakes appear to predate full development of the Colorado River. From east to west, these lakes have been referred to as Lake Grand Wash, Lake Hualapai, and Lake Las Vegas (e.g., Spencer et al., 2013; Pearthree and House, 2014; Fig. 3B). The maximum levels of these lakes have been based on the uppermost elevations of late Miocene limestone deposits in the various basins, as discussed in subsequent sections.

The pre–Colorado River, late Miocene lacustrine deposits commonly overlie and interfinger with playa deposits, with sections composed of intercalated limestone, gypsum, anhydrite, halite, siltstone, and claystone. The finer-grained deposits grade laterally into generally locally derived fluvial sandstones and coarser-grained alluvial fan facies toward basin margins. This entire package of sediments, ranging from alluvial fans proximal to mountain fronts to siltstones and evaporites in the central parts of basins, has commonly been referred to as the Muddy Creek Formation (Bohannon, 1984). Numerous 40Ar/39Ar dates and tephra correlations constrain the Muddy Creek Formation and its equivalents throughout the region to the late Miocene–earliest Pliocene (Feuerbach et al., 1991; Spencer et al., 2001; Beard et al., 2007; Muntean, 2012). Sedimentologic and detrital zircon data indicate that the Muddy Creek Formation in the Virgin River area accumulated in interior drained basins and was not associated with a proto–Colorado River (e.g., Pederson, 2008; Forrester, 2009; Dickinson et al., 2014).

Halite Distribution and Origin

A group of poorly exposed, but thick (>1 km) late Tertiary evaporite deposits, consisting primarily of halite, stretches from southern Arizona to the Lake Mead region of southern Nevada (e.g., Holser 1970; Peirce 1976; Faulds et al., 1997; Rauzi, 2002a, 2002b, 2002c) (Figs. 2 and 3). Unusually thick evaporites have been documented within this group, including 1800 m of anhydrite in the Picacho basin (Peirce, 1976), >1200 m of halite in the Luke basin (Eaton et al. 1972), >500 m of halite in the Overton Arm basin (Mannion, 1963), and ∼2500 m of halite in the Red Lake area of the Hualapai basin (Faulds et al., 1997) (Figs. 2 and 3). Most of the thicker known evaporite deposits occur near the margin of the Basin and Range Province, proximal to the topographically elevated Colorado Plateau, as best exemplified by the thick halite in the Hualapai basin along the western margin of the Colorado Plateau (Fig. 3).

On the basis of the terrestrial Cenozoic setting for the bulk of the region and geochemical evidence, most of these halite and related evaporite deposits have been interpreted as nonmarine in origin (Mannion 1963; Holser, 1970; Eaton et al., 1972; Peirce 1976; Bohannon 1984; Faulds et al., 1997). For example, low bromine contents (2–6 ppm) in halite from the Luke (Eaton et al., 1972), Overton Arm (Holser, 1970), and Hualapai basins (Faulds et al., 1997), as well as S and O isotopic values from capping anhydrite in the Hualapai basin (Faulds et al., 1997), all indicate a nonmarine origin. Lovejoy (1980) related the unusually thick halite in the Hualapai basin to evaporation of Colorado River water, hypothesizing that a proto–Colorado River flowed through the Grand Canyon, debauched into the Grand Wash Trough, and terminated in the Hualapai basin. Interpretations of a late Miocene–early Pliocene marine incursion of a proto–Gulf of California into much of the lower Colorado River region (Lucchitta, 1979; Buising, 1990) have been largely refuted, at least in northern areas, on the basis of Sr isotopic values (0.7102–0.7114) and lack of marine fossils or marine geochemical signatures in late Miocene lacustrine limestone (e.g., Spencer and Patchett, 1997; Spencer et al., 2008, 2013). However, paleontologic evidence suggests that late Miocene–early Pliocene marine incursions may have extended as far north as the Blythe basin (McDougall, 2008; McDougall and Martinez, 2014) (Fig. 2), well south of the Lake Mead area, although marine organisms may have been transported by birds into saline lakes in the Blythe area (Spencer et al., 2013).

Abundant textural evidence (Fig. 6) suggests that the evaporites accumulated in continental playas. For example, poorly developed bedding, coarse interlocking halite crystals, the common occurrence of intercrystalline shale within the halite, and isolated displacive halite crystals and desiccation cracks in the associated shale (Faulds et al., 1997) indicate that the halite in the Hualapai basin originated primarily through intrasedimentary displacive growth in desiccated pans (playa mudflats), together with intervals of deposition in flooded saline pans (e.g., Gornitz and Schreiber, 1981; Rosen, 1994; Curial and Moretto, 1996; Schubel and Lowenstein, 1997). Relatively pure beds of finer-grained halite (cumulates) probably originated from subaqueous growth in a stratified hypersaline water body (e.g., Smoot and Lowenstein, 1991).

Possible sources of sodium chloride in the region include Permian redbeds (Supai Formation) in the Colorado Plateau, which are rich in chloride salts and interbeds of halite (Lucchitta, 1966; Eaton et al., 1972), thick Miocene calc-alkaline volcanic piles in the Basin and Range, and hydrothermal waters (Faulds et al., 1997). The Permian redbeds are a likely source of both Na+ and Cl, because the nearby Colorado Plateau provides a large, topographically elevated region from which to derive groundwater and surface runoff (Faulds et al., 1997). However, similar to the central Andes, where unusually thick nonmarine halite deposits were derived from volcanic source areas (Ericksen and Salas, 1989; Alonso et al., 1991), 1–4-km-thick Oligocene–Miocene volcanic piles in the Basin and Range and associated hydrothermal deposits may have also provided a significant source of Na+ and Cl. Modern hypersaline and soda lakes are commonly associated with active volcanic regions in rift settings (Kempe and Degens, 1985).

It is important to note that multiple processes modify the original textures of halite and other evaporites in continental playa settings and hypersaline lakes. These processes include desiccation, dissolution, recrystallization, and growth of displacive salt in muds. Such processes can obliterate or significantly modify original bedding, crystalline form, and fluid inclusions (e.g., Shearman, 1970). This, in turn, can create difficulties in accurately dating such deposits utilizing such techniques as tephrochronology. Tephras are commonly well preserved in lacustrine environments and form widespread uniform beds, which can be geochemically correlated over broad regions (e.g., Perkins and Nash, 2002). In contrast, dissolution, desiccation, and recrystallization can significantly disrupt layers of tephra in evaporite sequences, making it much less likely to recognize such layers in drill core. The thick sequences of halite in the region and dominance of halite in some stratigraphic columns (e.g., Hualapai basin; Faulds et al., 1997) compounds the dating dilemma. Nonetheless, without interbedded lavas, tephras offer the most viable means by which to date the halite deposits, and thus we dedicated significant effort to identifying tephras in available core from the region.

It should also be noted that that evaporites can accumulate rapidly. High rates of evaporation in arid regions permit optimum accumulation of sulfates at rates of 1–40 m/1000 yr and halite at 10–100 m/1000 yr, as observed in modern marine-sourced basins (Schreiber and Hsü, 1980). For example, deposition of the 2-km-thick upper Miocene evaporite section on the floor of the Mediterranean occurred within ∼0.63 m.y. (CIESM, 2008; Roveri et al., 2008; Manzi et al., 2009). In continental basins with sufficient accommodation space, evaporite deposition may be as rapid as in marine-fed basins. However, assuming the ionic input into a continental-sourced basin is only one-tenth that of marine sources, the time interval available for deposition of 2.5 km of halite, as in the Hualapai basin, may have been as long as several million years.

Halite is preserved when it is buried beneath the level of the stagnant point in the phreatic zone. The upper phreatic zone is typically fresh water recharged by topographically driven flow from the adjacent, basin-bounding ranges. Below the stagnant point, lateral groundwater movement is slow, and the groundwater becomes saturated with respect to halite by dissolution of halite near the basin margins. Buried halite is therefore preserved in the basin depocenter. Preservation of halite occurs when subsidence results in burial of the halite below the stagnant point and/or where the groundwater becomes saturated with respect to halite. Preservation conditions are met during periods of a prolonged arid climate, rapid subsidence, and/or where the bounding ranges lack sufficient elevation to provide the hydraulic head to flush the central basin by topographically driven flow.

METHODS

Geophysical

Although no new geophysical data were obtained in this study, available detailed gravity and aeromagnetic surveys (Fig. 4) and depth-to-basement estimates derived from the gravity data (Fig. 3A; Langenheim et al., 2001a, 2001b, 2010) were utilized to interpret the extent of halite in several basins in the Lake Mead region. Halite has specific density and magnetic properties that make gravity and aeromagnetic data well suited to identify basins likely to contain salt deposits and to delineate the extent of halite within basins with known accumulations. Both methods offer significant advantages, versus drilling, resistivity, or seismic methods, in that there are substantial historic databases, both in the public domain and commercially available.

Data acquisition and inversion methods are described in Langenheim et al. (2001a, 2001b, 2010). Gravity measurements are precise to ∼0.1 mGal, and the resulting anomalies are estimated to be ±0.5 mGal or better. Aeromagnetic anomalies are estimated to be ±1–2 nT or better. To estimate thickness of basin fill, the gravity field is separated into a component caused by density variations within the basement and a component caused by low-density basin fill. The basin-fill anomaly is then converted to thickness using density-depth relationships, in the case of this region, based on sonic velocity information (logs and interval velocities used to migrate seismic-reflection data) as a proxy for density. An important assumption is that the density of the basin-fill deposits can be related to velocity using standard relationships (e.g., Gardner et al., 1974). For the depth-to-basement map shown in Figure 3, the density of the upper 500 m is assumed to be ∼2100 kg/m3, but it increases with depth to ∼2400 kg/m3 by 1.2–1.3 km. The effect of the assumed density-depth relationship means that the shape of the basin is fairly well constrained, but that the absolute depth of the basin may be overestimated if the real density contrast is higher than that assumed and underestimated if the actual density contrast is lower than that used in the inversion.

Halite deposits have an approximate mean density of 2100 kg/m3. In contrast, quartzose and feldspathic sandstones and conglomerates have a density of 2300–2600 kg/m3, increasing in density with burial compaction. This contrast in density produces a negative gravity response, or gravity “low.” An example of a modeled gravity response (3 mGal) for a half graben containing a 300-m-thick halite deposit and a half graben without any halite is illustrated in Figure 7. This anomaly is certainly resolvable and can be used to delineate known halite accumulations or identify potential halite deposits. However, thick, porous, sedimentary fill will also yield a gravity low relative to crystalline and metamorphic basement. In our analysis of individual basins, we compared the maximum extent of subsurface halite by assuming a constant density-depth relationship of the same density of halite with those previously published estimates of basin fill assuming no halite shown in Figure 3. Additional methods, however, are needed to test whether a gravity low is a halite deposit or a thick accumulation of clastic sediments and to estimate the proportion of halite to clastic sediments within the basin fill.

Halite also has a low magnetic susceptibility (–10.1 k × 10–6 SI) in contrast to iron-bearing quartzose sandstones (440 k × 10–6 SI) and shales (641 k × 10–6 SI) (Prieto, 1993). This weak contrast in magnetic susceptibility has been used to map halite bodies using aeromagnetic data in the Gulf of Mexico (Vixo et al., 1987; Prieto, 1993), because such bodies have a characteristically low magnetic response on the order of a few nT. In the Lake Mead region, this susceptibility contrast is enhanced by clasts of magnetic volcanic and basement rocks and by iron oxide and manganese coating of clasts (as observed in buried deposits) resulting from the prolonged arid climate. The magnetic response of halite produces anomalies of 4–20 nT, depending on the depth of burial, with a shallow halite body producing the largest amplitude anomaly and the characteristic lower and to higher values over the edges of a tabular body. This signal, although measurable with the existing data, may be difficult to isolate, particularly if the basement rocks are also strongly magnetic.

Seismic-reflection data are available in some areas and have been applied to interpreting basin geometries and evolution, most notably in the Hualapai, Detrital, and Las Vegas basins, as well as the Virgin River depression (Bohannon et al., 1993; Faulds et al., 1997; Langenheim et al., 2001a, 2001b). These data provide important constraints for the various depth-to-basement estimates using gravity. The sonic velocity of halite (4500–5000 km/s) is considerably higher than predicted by its low density using standard velocity-density relationships (4500 km/s predicts a density of 2540 kg/m3 using the Gardner et al. (1974) relationship), so comparison of basin thickness from seismic-reflection data with that derived from gravity can help constrain whether halite is a significant component of the basin fill.

Electrical methods can also be useful for identifying subsurface halite bodies. When groundwater wells are geophysically analyzed, they are typically surveyed with conductivity, or resistivity wire line logs. Freshwater has relatively poor conductivity (e.g., high resistivity). As chloride content increases, however, conductivity increases and resistivity decreases due to dissolution of halite by drilling fluids or increasingly brackish groundwater proximal to evaporite deposits (Dewan, 1983). Therefore, although evaporites may not be penetrated by a groundwater well, trends in conductivity/resistivity can indicate proximity to buried evaporites. Electrical surveys, both airborne and ground based, can also be used to map conductivity anomalies associated with brackish or saline water. These data are available in a few areas and have been applied to groundwater issues, in particular interpreting grain size, groundwater quality, and possible presence of evaporites in the upper 400–500 m in parts of the Hualapai, Detrital, Sacramento, and Virgin River basins (Zohdy et al., 1994; Truini et al., 2012).

40Ar/39Ar Geochronology

Seven 40Ar/39Ar dates were obtained to constrain the ages of basin deposits and timing of major extension, particularly within the Grand Wash Trough (Table 1). Six samples were dated at the New Mexico Geochronology Research Laboratory (NMGRL), and one sample was dated at the U.S. Geological Survey in Denver. Interpretations of the ages are summarized briefly herein, and analytical data are presented in Appendix A (Tables A1, A2, and A3). A more rigorous data discussion can be obtained at www.ees.nmt.edu/Geol/labs/Argon_Lab/NMGRL_homepage.html (click on data) for the NMGRL data. Basalt samples were subjected to step-wise heating and yielded simple plateau age spectra that correspond to eruption ages. The sanidine ages are determined from laser fusion of many splits (15–30) of fine-grained separates and are therefore susceptible to xenocrystic contamination. Eruption ages are based on either combining the youngest population of apparent ages (samples JF-97-144, JF-98-308, and MW-98-36) or calculating maximum eruption ages from the youngest population of results that yield indistinguishable 40Ar/39Ar ratios (JF-98-155). A maximum eruption age is assigned to these, because the total 40Ar measured in these samples is assumed to be equal to the radiogenic 40Ar.

Tephrochronology

Tephrochronology is the characterization, correlation, and age calibration of tephra layers. It has contributed substantially to knowledge of the stratigraphy and chronology of late Neogene sedimentary strata in the western United States (e.g., Sarna-Wojcicki and Davis, 1991; Christiansen and Yeats, 1992; Perkins et al., 1998; Perkins and Nash, 2002). Of particular importance in this area are large-magnitude explosive eruptions generated along the Yellowstone hotspot track and the more focused eruptions from the southwest Nevada volcanic province. Extensive tephra layers from these eruptions, ranging in age from ca. 16 to 5 Ma, provide both relative and quantitative ages for many extensional basins in the Basin and Range (Perkins and Nash, 2002; Perkins et al., 2014). Relative dating is provided by the chemical correlation of tephra from basin to basin using electron microprobe analyses of major elements in glass shards, the dominant material in the tephra layers. Isotopic ages provide a chronologic calibration of select tephra in these basins, while remaining ages are estimated by interpolation between dated horizons.

Glass shards from three tephra beds from surface exposures in the Grand Wash Trough and Nellis basin, as well as one sample from core in the Hualapai basin, were analyzed at the University of Utah following methods described in Perkins et al. (1998). Glass shard compositional data from electron microprobe analyses of individual shards are presented for the four samples in Table 2. The core sample (RL1-2053) correlates with the 5.62 ± 0.11 Ma Conant Creek Tuff. One sample from the Nellis basin correlates with the 5.62 ± 0.11 Ma tuff of Wolverine Creek. Two samples from the Grand Wash Trough correlated with the 11.01 ± 0.03 Ma Cougar Point Tuff. The context of these correlations is discussed in the individual descriptions of each basin.

RESULTS

In the sections below, we describe the late Miocene evaporite and lacustrine deposits in several individual basins in the Lake Mead region (Fig. 3) in order to glean the spatial and temporal relations of such deposits to one another and the Colorado River. For each basin, we present (1) geological overviews, (2) descriptions of the known distribution of late Miocene lacustrine and evaporite deposits, (3) geochronological constraints on the timing of basin development and late Miocene sedimentary facies, (4) potential for undiscovered thick sections of halite in the basin-fill sediments, and (5) a summary of late Miocene depositional environments. More detailed descriptions of the Grand Wash Trough and Hualapai basin are provided due to superb exposures and availability of core for analysis, respectively. These attributes, in turn, furnished a sedimentologic, geochronologic, and geophysical framework with which to evaluate and catalogue lacustrine and evaporite deposits elsewhere in the Lake Mead region. For all analyzed basins in the Lake Mead region in this study, Table 3 summarizes the age, thickness, and elevation of late Miocene lacustrine limestone, gypsum-anhydrite, and halite deposits, overall basin-fill thicknesses, and potential for thick halite deposits.

Grand Wash Trough

The Grand Wash Trough is a 75-km-long basin bounded on the east by the Grand Wash Cliffs along the western margin of the Colorado Plateau and on the west by the White Hills and South Virgin Mountains (Fig. 3A). The southern part of the trough includes two east-tilted half grabens, which are separated by Wheeler Ridge in the north and the Lost Basin Range in the south. The southeastern half graben, here termed the southern Grand Wash Trough, developed in the hanging wall of the west-dipping northern Grand Wash fault, is centered in the Grapevine Wash area, and contains less than ∼1 km of Tertiary basin-fill sedimentary rocks (Langenheim et al., 2010). In the western part of the Grand Wash Trough, the Gregg basin is a relatively narrow east-tilted half graben that lies in the hanging wall of the west-dipping Wheeler Ridge and Lost Basin Range faults. As the Wheeler Ridge fault dies out northward to the north of Lake Mead, Gregg basin and southern Grand Wash Trough coalesce to form a large composite basin more than 4 km deep (Langenheim et al., 2010), here referred to as the northern Grand Wash Trough (Fig. 3).

Dissection by the Colorado River and its tributaries has produced excellent exposures of the upper part of the Tertiary section in both the southern Grand Wash Trough and Gregg basin (Figs. 3 and 5). The middle to late Miocene section in these areas has been referred to as the rocks of the Grand Wash Trough after Bohannon (1984). The upper part of this section temporally correlates with the Muddy Creek Formation. In ascending order, the Tertiary section in the southern Grand Wash Trough consists of at least 250 m of middle to late Miocene fanglomerate, more than 120 m of a sandstone-siltstone facies with locally interbedded gypsum, and as much as 210 m of late Miocene limestone (Figs. 8 and 9; Longwell, 1936; Lucchitta, 1966; Bohannon, 1984; Wallace, 1999; Blythe, 2005; Wallace et al., 2005; Crossey et al., 2015). The Miocene units interfinger and thicken eastward toward the deeper parts of the half graben. The limestone in the Grand Wash Trough is known as the Hualapai Limestone and has been correlated with similar limestone elsewhere in the central to eastern Lake Mead region (Longwell, 1928, 1936; Lucchitta, 1966). In the Grand Wash Trough, exposures of the Hualapai Limestone range from ∼500 to 912 m above sea level (asl). Gypsum crops out as low as ∼360 m asl (Fig. 8 and Table 3; Wallace et al., 2005). Detrital zircon data indicate a local provenance for the sandstone-siltstone facies in the southern Grand Wash Trough (Crossey et al., 2015).

The rocks of the Grand Wash Trough are bracketed between ca. 15.5 and 7.5 Ma, as constrained by 40Ar/39Ar dates originally reported by Faulds et al. (2001b) but with the analytical data and further discussion presented herein (Fig. 9 and Table 1). The older age is based on a 15.51 ± 0.04 Ma 40Ar/39Ar date on sanidine from a rhyolite tuff tilted ∼30°–35° near the base of the Tertiary section on the west flank of Grapevine Mesa (Figs. 5 and 8; Table 1; Wallace et al., 2005). A gently tilted (∼5°–10°) nonwelded tuff in the lower exposed part of the sandstone-siltstone facies in the Pearce Ferry area (Figs. 8 and 10D) has yielded a maximum 40Ar/39Ar age on sanidine of 13.26 ± 0.06 Ma, whereas tephras in the lowermost part of the Hualapai Limestone in Grapevine Wash (Figs. 8 and 10C) and directly beneath the Hualapai Limestone near Airport Point both geochemically correlate with an 11.01 ± 0.03 Ma tuff derived from the Bruneau-Jarbidge volcanic field in southernmost Idaho (Table 2; Wallace et al., 2005). Fine-grained sanidine from this tephra also yielded an 40Ar/39Ar age of 11.20 ± 0.13 Ma (Fig. 9; Table 1), which is compatible with the geochemical correlation. In the same part of the section, Crossey et al. (2015) geochemically correlated a tephra with 12.07–11.31 Ma ash-flow tuffs derived from calderas in the Snake River Plain region. Furthermore, as reported herein, K-feldspar from a tephra intercalated in the upper part of the nearly flat-lying Hualapai Limestone at Grapevine Mesa yielded an 40Ar/39Ar maximum eruptive age of 7.52 ± 0.11 Ma (Figs. 8 and 9; Table 1).

These data bracket both the timing of tilting and age of the Hualapai Limestone in the Grand Wash Trough. Significant tilting and presumably extension occurred between ca. 15.5 and 7.5 Ma, as indicated by the progressive upward decrease in tilt (from 35° to subhorizontal) within the middle to late Miocene section. However, the base of the Tertiary section is tilted much less than underlying Paleozoic strata (35° versus ∼60°+), suggesting an earlier onset to extension. On the basis of regional relations (e.g., Beard, 1996; Faulds et al., 2010), we infer that major extension began ca. 16 Ma in the Grand Wash Trough and that the main episode of extension occurred between ca. 16 and 13 Ma. Additionally, our dates and those from Crossey et al. (2015) bracket the age of the Hualapai Limestone in the Grand Wash Trough between ca. 12 and 7.5 Ma. It is important to note that the middle to upper parts of the Hualapai Limestone onlap the Grand Wash fault in the southern Grand Wash Trough (Fig. 8), indicating that this part of the fault zone became inactive prior to ca. 8–9 Ma.

40Ar/39Ar dates from basalt flows along the Grand Wash Cliffs to the north of Lake Mead further constrain the structural and paleogeographic evolution of the Grand Wash Trough. Billingsley and Wellmeyer (2003) suggested that a basalt flow capping Snap Point on the elevated western margin of the Colorado Plateau directly east of the Grand Wash Trough was erupted from a dike at Snap Point and flowed down the steep erosional escarpment of the Grand Wash Cliffs through Snap Canyon to the Nevershine Mesa area in the Grand Wash Trough, where a presumed correlative basalt is intercalated in alluvial fan sediments (Fig. 11). The basalt flow at Snap Point had yielded a K-Ar age of 9.07 ± 0.80 Ma (Reynolds et al., 1986), but the Nevershine Mesa basalt had not been previously dated. Our 40Ar/39Ar dates and geochemical analyses of the basalt flows at Snap Point and Nevershine Mesa support the Billingsley and Wellmeyer (2003) hypothesis. The two basalts yielded essentially identical geochemistry and 40Ar/39Ar dates of 8.93 ± 0.03 Ma and 8.94 ± 0.05 Ma, respectively (Fig. 12). The basalt at Nevershine Mesa overlies probable correlatives of the Hualapai Limestone, and both the limestone and basalt flow onlap the Grand Wash fault (Figs. 11A and 11D). These relations suggest that the Hualapai Limestone in this area is older than 8.9 Ma and that the present physiography of the Grand Wash Trough and abrupt western margin of the Colorado Plateau were largely established by 8.9 Ma.

Another basalt flow at Sandy Point in the Grand Wash Trough provides an additional, critical paleogeographic marker, as it overlies Colorado River gravels ∼100 m above the present grade of the river (Figs. 5, 8, and 10A). This basalt flow yielded an 40Ar/39Ar age of 4.49 ± 0.23 Ma (Fig. 9; Tables 1 and A2). It was originally dated at 3.79 ± 0.46 Ma via the K/Ar method (Damon et al., 1978). The underlying river gravels accumulated within a canyon cut significantly below the level of the Hualapai Limestone. Thus, the first appearance of the Colorado River in the Grand Wash Trough is bracketed between ca. 7.5 Ma (age of the upper Hualapai Limestone) and 4.5 Ma (age of Sandy Point basalt).

The Wheeler Ridge and Lost Basin Range faults accommodated 275–500 m of offset of the Hualapai Limestone (Lucchitta, 1966; Wallace et al., 2005; Seixas et al., 2015) and continued to accommodate displacement as the Colorado River incised from ca. 5 Ma to present (Seixas et al., 2015), as evidenced by gentle local tilting (∼5°) of both the early Pliocene basalts and Colorado River sediments (Howard and Bohannon, 2001; Howard et al., 2010), as well as offset early Pleistocene alluvial fan deposits (Wallace et al., 2005). Thus, mild extension continued to affect the area in Pliocene–Pleistocene time.

Tertiary strata are generally not well exposed in the deeper (>4 km) northern part of the Grand Wash Trough (Fig. 3) nor has this area been penetrated by deep wells. The Grand Wash fault in this area has accommodated Quaternary displacement in contrast to areas to the south, where late Miocene strata onlap the fault. In the transitional region between the northern and southern parts of the Grand Wash Trough, the Hualapai Limestone overlies extensive late Miocene gypsum, which may approach 500 m in thickness (Figs. 3 and 11; Billingsley and Wellmeyer, 2003). Similar strata may lie in the subsurface farther north, where gravity data indicate the thickest (>4 km) sedimentary fill within the Grand Wash Trough. Here, a 30–40 mGal gravity low (Fig. 4A; Langenheim et al., 2010) may indicate that halite makes up some of the basin fill. Based on inversion of the anomaly and assuming a constant density contrast consistent with halite, the maximum thickness of halite in this area is ∼1.5 km. Such a thickness would produce reasonable basement gravity values and is also consistent with low magnetic values in the area (Fig. 4B).

The late Miocene depositional environment of the Grand Wash Trough ranged from alluvial fans proximal to the margins to continental playa and lacustrine generally in the interior portions. Deposition of the sandstone-siltstone facies probably occurred in an evaporative interior continental playa, as evidenced by the intercalated gypsum, thin bedding, and mudcracked bedding planes (Wallace et al., 2005). The lack of fluvial textures, such as crossbeds or ripple marks, local abundance of gypsum rinds, interfingering and bordering fanglomerate facies, and interbedded gypsum appear to rule out a through-going drainage. However, the relatively thick deposits of gypsum (∼500 m) and possible thick halite (as much as 1.5 km) in the central to northern reaches of the Grand Wash Trough suggest significant input of groundwater and/or surface water to the basin in late Miocene time.

The basin transitioned to a wetter setting ca. 12–11 Ma as reflected by the Hualapai Limestone. Although some early interpretations suggested a marine origin for the Hualapai Limestone (Blair, 1978; Blair and Armstrong, 1979), 87Sr/86Sr isotopic ratios (0.7114–0.7195; Spencer and Patchett, 1997; Roskowski et al., 2010; Lopez Pearce et al., 2011), fossil assemblages, petrography, and δ13C-δ18O isotopic geochemistry (Faulds et al., 1997, 2001c; Wallace, 1999; Wallace et al., 2005; Crossey et al., 2015) indicate deposition in a lacustrine to nonmarine wetland setting. This lake and/or wetlands has been referred to as Lake Grand Wash (e.g., Spencer et al., 2013), with a maximum level of ∼912 m (Fig. 3B) corresponding to the current elevations of the uppermost exposures of the Hualapai Limestone. Because the lake was fresh, it probably had an outlet either through surface runoff or groundwater sapping into nearby basins. 87Sr/86Sr and δ13C-δ18O trends through the section of limestone suggest that Lake Grand Wash was initially fed by springs sourced from the Colorado Plateau ca. 12–8 Ma and that meteoric groundwater was progressively introduced after ca. 8 Ma just prior to development and integration by the Colorado River (Crossey et al., 2015). It is important to note that the absence of exotic rounded river gravels and fluvial textures in the sandstone-siltstone facies combined with the presence of the capping lacustrine Hualapai Limestone suggest that the Colorado River was not responsible for deposition of the late Miocene sediments in the Grand Wash Trough (e.g., Longwell, 1946; Lucchitta, 1966, 1972).

Hualapai Basin

The Hualapai basin is an ∼60-km-long, north-northwest–trending depression situated ∼40 km south of the Colorado River and sandwiched between the western margin of the Colorado Plateau on the east and Cerbat Mountains on the west (Fig. 3). The basin is a gently to moderately east-tilted half graben developed in the hanging wall of the southern Grand Wash fault (Fig. 13B). The half graben probably formed primarily during the main episode of middle Miocene extension (ca. 16–13 Ma), but minor amounts of normal faulting have deformed the area since ca. 8.7 Ma (Faulds et al., 2010). Quaternary scarps have not been observed along the southern Grand Wash fault or elsewhere in the basin. Despite its proximity to the Colorado River and Grand Canyon, the Hualapai basin remains an internally drained, closed depression and contains the Red Lake playa. Due to a lack of dissection by tributaries of the Colorado River, synextensional middle to late Miocene strata within the basin are obscured by more recent flat-lying sediments in contrast to the highly dissected basins along the course of the Colorado River to the north, such as the Grand Wash Trough. Nonetheless, gravity (Fig. 4A; ∼50–60 mGal low in the northern part of the basin; Langenheim et al., 2010), drill-hole (Fig. 13A; Table 4), and seismic-reflection data (Faulds et al., 1997) indicate that the Hualapai basin contains a thick (∼3.9 km) sequence of Miocene to Quaternary basin-fill strata, with tilts decreasing up-section from ∼25° to 0°.

As inferred from exposures in surrounding mountain ranges and along the western margin of the Colorado Plateau, analysis of drill core, and seismic-reflection profiles, the stratigraphy of the Hualapai basin consists of a thick section of primarily Miocene sedimentary and volcanic rocks. In ascending order, this section includes: (1) ∼750 m of lower to middle Miocene volcanic and sedimentary rock possibly resting on Cambrian strata and/or Proterozoic gneiss and granite; (2) ∼335 m of middle Miocene volcanic and sedimentary rock; (3) fanglomerates along the margins that interfinger with evaporites in the central part of the basin; 4) up to 2500 m of halite referred to as the Red Lake halite and intercalated with minor shale (5%–10%) and anhydrite; and (5) ∼600 m of late Miocene–Quaternary shale and lesser amounts of gypsum, anhydrite, and conglomerate (Fig. 9; Faulds et al., 1997). It is important to note that the thick halite does not crop out anywhere, either within or along the margins of the basin.

Drill-hole and geophysical data constrain the overall thickness and extent of the Red Lake halite deposit. In four drill holes within the basin, the top of the halite varies from 209 to 418 m in elevation (Fig. 13A; Tables 3 and 4), whereas the top of the gypsum-anhydrite sequence that caps the section of evaporites ranges from 286 to 440 m asl. The EP-1 well (drilled by El Paso Natural Gas; Tables 3 and 4) penetrated a thick succession of halite from depths of 597 m (1958 ft depth) to 1827 m (5995 ft) near the center of the Hualapai basin (Fig. 13B), yielding a minimum thickness of halite of ∼1230 m. Documented halite in the EP-1 well ranges from 258 m above sea level to ∼972 m below sea level (bsl). On the basis of the reflective character of the halite in a seismic-reflection profile, the total thickness of halite has been inferred at ∼2500 m (Faulds et al., 1997), suggesting that the halite may extend downward to elevations in excess of 2 km below present sea level. These elevations are consistently lower than the exposures of the Hualapai Limestone in the Grand Wash Trough, although minor tectonism between the two basins since the late Miocene cannot be ruled out. Gravity data are compatible with the interpretation that the Hualapai basin is dominated by a 2.5-km-thick sequence of halite (Figs. 3, 4, and 13B). Inversion of the gravity low using a relationship of increasing density with depth produces a basin thickness of more than 10 km, which does not agree with the seismic-reflection data. Modifying the density-depth relationship to include a significant thickness of low-density fill, such as halite, produces basin depths more consistent with the seismic-reflection data (Langenheim et al., 2010).

On the basis of regional relations, the age of the Red Lake halite was previously bracketed between ca. 13 and 8 Ma (Faulds et al., 1997), but our new tephrochronologic data (Table 2) and recent work indicate a somewhat younger age of ca. 12–5.6 Ma. Because the halite occupies the gently tilted upper part of the basin fill and extension in the region is constrained in nearby areas to ca. 16–8 Ma (Faulds et al., 2010), with the main pulse occurring from 16 to 13 Ma, the gently tilted halite is probably younger than ca. 13 Ma. Our analysis of core from the EP-1 well in the Hualapai basin (Fig. 13) revealed an intercalated tephra at a depth of 626 m (2053 ft) or 229 m in elevation. The tephra lies ∼79 m beneath the top of the halite and 136 m beneath the top of the capping anhydrite. This tephra geochemically correlates (Table 2) with the Conant Creek Tuff, which is a Yellowstone hotspot tuff sourced from the eastern Snake River Plain. The Conant Creek Tuff has yielded an 40Ar/39Ar age of 5.62 ± 0.11 Ma from the near source ash-flow tuff (relative to an age of 28.201 ± 0.46 Ma for sanidine from the Fish Canyon Tuff; Kuiper et al., 2008). Thus, the halite in the Hualapai basin is now bracketed between ca. 12 to slightly younger than 5.6 Ma and is largely contemporaneous with the Hualapai Limestone to the north.

The textures (Fig. 6) and bromine content of the halite and S and O isotopic values of intercalated and capping anhydrite indicate that halite deposition resulted from groundwater discharge and ponding in an intracontinental playa (Faulds et al., 1997). Our recent, more thorough analysis of core from the EP-1 well revealed a compositionally, texturally, and sedimentologically monotonous sequence of halite. This sequence consists of halite (>90%), interbedded with lesser fine clastics and anhydrite (Figs. 6B, 6C, and 6E). Generally, the halite layers are massive, pure, and poorly bedded. In addition, salt crystals are clear, lack primary sedimentary features, generally exceed 2 cm in length (Fig. 6A), and range up to 10 cm long. Some of the halite fills dissolution cavities within the halite masses. Fine clastics or anhydrite stringers outline some of the halite crystals, with relatively common displacive halite crystals in siltstone (Fig. 6B). The observed fabrics are typical of postdepositional textures, which resulted from rock and groundwater interaction. Primary sedimentary halite is typically small in crystal size and rich in milky fluid inclusions (Benison and Goldstein, 1999; Goldstein, 2001). Entrapment of fluid inclusions generally takes place at high rates of crystal growth during primary deposition of the halite. The observed lack of fluid inclusions and large blocky crystals indicate that much of the Red Lake halite formed from slow growth and/or recrystallization processes probably due to periods of exposure, localized dissolution, and reprecipitation in shallow playas.

The extreme thickness of the halite, apparent lack of limestone, common displacive growth textures, and relative lack of interlayered clastics suggest that the Hualapai basin had no outlet and served as a regional sink for saline groundwater and surface water that evaporated in extensive salt pans within a continental playa (e.g., Fig. 1; Neev and Emery, 1967; Smoot and Lowenstein, 1991). A similar depositional environment presently characterizes the Dead Sea area (Neev and Emery, 1967; Abed and Yaghan, 2000). Two coeval sedimentary processes probably contributed to halite deposition in the Hualapai basin: (1) evaporation from a standing body of brine, and (2) intra-sedimentary displacive growth of halite in a desiccated salt pan. The common displacive growth textures (Fig. 6B) and relative lack of interlayered fine clastics indicate that much of the halite was derived from groundwater brines under arid conditions, which accounts for lack of clastic influx into the basin center area. The basin playa most likely served as a prolonged, regional discharge zone for large volumes of generally saline groundwater (Faulds et al., 1997). The displacive growth of halite probably took place by both seeping, hypersaline groundwater and from incoming fresh water at the surface of the salt pan and subsequent evaporation (Fig. 1). Poorly defined bedding throughout much of the sequence resulted from a sporadic influx of undersaturated waters that dissolved and then reprecipitated the underlying halite both as irregular crystal overgrowths and as interstitial precipitates upon re-concentration due to rapid evaporation. During the early stages of halite deposition, the groundwater in the basin may have been relatively fresh but would become saline through discharge and subsequent evaporation. However, as the halite sequence grew in thickness, any groundwater within the central part of the basin would have likely been saline due to dissolution of some of the previously deposited, subsurface halite.

Virgin-Detrital Trough

The Virgin-Detrital trough is a continuous ∼140-km-long, 15–40-km-wide, northerly trending topographic depression that straddles Lake Mead and the Colorado River in southern Nevada and northwestern Arizona (Fig. 3). It consists of three distinct parts, including a central portion referred to as the Temple Bar basin, a northern segment called the Overton Arm basin, and a southern leg referred to as the Detrital basin.

Temple Bar Basin

The Temple Bar basin occupies the central Lake Mead region proximal to the Colorado River and central part of the Virgin-Detrital trough (Fig. 3). It contains the confluence of the Virgin and Colorado Rivers. It is as much as ∼25 km wide and is sandwiched between the northern White Hills and South Virgin Mountains to the east and Black Mountains to the west. The Temple Bar basin is essentially an east-tilted half graben in the hanging wall of the South Virgin–White Hills detachment fault and largely consists of east-tilted middle to late Miocene fanglomerate, megabreccia, and volcanic rocks bracketed between ca. 15.2 and 6 Ma. Upward-decreasing tilts suggest synextensional deposition of the Miocene section, with the main episode of extension occurring from ca. 16.5 to 11 Ma (Duebendorfer and Sharp, 1998; Blythe et al., 2010). Quaternary fault scarps have only been observed along the northeastern margin of the Temple Bar basin to the north of Lake Mead (Fig. 3; U.S. Geological Survey, 2010) and possibly along the far western edge along the northern part of the Detrital fault. However, late Miocene strata in the area are locally tilted gently (<15°) and cut by minor faults (Beard et al., 2007).

The upper part of the basin fill consists of little-deformed late Miocene fanglomerate that interfingers with and gives way to the Hualapai Limestone and gypsum toward the basin center (Howard et al., 2010). The upper Miocene section onlaps the bounding detachment fault and fills paleovalleys cut into Proterozoic rocks in the footwall of the detachment. Some of these paleochannels connect eastward to the Gregg basin in the Grand Wash Trough (Howard et al., 2010). Scattered exposures of Hualapai Limestone rim the central part of the basin (Beard et al., 2007), crop out at elevations ranging from ∼360 m to as high as 720 m (Spencer et al., 2013), and range in thickness from ∼25–100 m (Table 3; Howard et al., 2010). Interbedded tephras and an 8.4 Ma basalt flow bracket the age of the limestone and related sediments (including gypsum) between ca. 10 and 6 Ma (Howard et al., 2010). A tephra interbedded within the middle part of the Hualapai Limestone directly south of Temple Bar yielded an 40Ar/39Ar age of 5.97 ± 0.04 Ma (Spencer et al., 2001), which suggests that the limestone in the Temple Bar area is largely younger than that in the Grand Wash Trough (Spencer et al., 2013). This date has been widely used as a younger age constraint for the Hualapai Limestone in the Lake Mead area.

Gypsum and mudstone indicate a depocenter in the central part of the basin directly west of Temple Bar (Fig. 3; Longwell, 1936; Beard et al., 2007). Here, extensive gypsum crops out along the shores of Lake Mead in an area known as the “Big Gypsum Ledges.” The gypsum is exposed at elevations ranging from 360 to 439 m asl (Table 3; Peirce, 1974; Beard et al., 2007) and contains tephras correlated between 7.5 and 6.3 Ma (Howard et al., 2010). Gypsum in this area is at least ∼200 m thick.

Gravity data and surface exposures of gently tilted strata indicate that the sediments in the Temple Bar basin are as much as 1.5 km thick (Langenheim et al., 2010). If composed solely of halite, the basin fill is no more than 1 km thick. Smooth, generally low magnetic values are consistent with the presence of some halite, although the survey flight lines are widely spaced in most of this area. Halite observed in drill holes in the nearby northern Detrital basin (Fig. 3) is also suggestive of some halite in the Temple Bar area.

The late Miocene depositional environment in the Temple Bar basin was probably similar to that of the Grand Wash Trough, as it ranged from alluvial fans proximal to the margins (e.g., Blythe et al., 2010) to continental playa and lacustrine in interior portions. Deposition of the gypsum and mudstone occurred in an evaporative playa through groundwater and/or surface water input. The lack of widespread fluvial sediments suggests that a through-going drainage did not exist. The basin transitioned to a wetter setting ca. 10–6 Ma as reflected by the Hualapai Limestone, which accumulated in a lacustrine to nonmarine wetland environment. This lake has been referred to as Lake Hualapai (Fig. 3B; Spencer et al., 2013). On the basis of current elevations of the uppermost exposures of the Hualapai Limestone between the Detrital and Temple Bar basins (e.g., Beard et al., 2007; Felger and Beard, 2010), maximum levels of the lake have been inferred at ∼720 m. Sr isotopes from the limestone (0.7137–0.7145; Roskowski et al., 2010) suggest that the lake had a similar origin to Lake Grand Wash to the east, with springs and possibly some runoff serving as the primary source of water and a probable outlet that prevented significant evaporation.

Overton Arm Basin

The Overton Arm basin is an ∼50-km-long, ∼10–15-km-wide, northerly trending topographic depression that extends northward along the Virgin River from Lake Mead between the South Virgin Mountains to the east and Muddy Mountains to the west (Fig. 3). Gravity data indicate that basin-fill sediments in the southern part of the basin are relatively thin (generally <1 km thick). However, a 25 mGal gravity low occupies the northern part of the basin (Fig. 4A) and lies between two left-lateral faults, the Bitter Ridge–Hamblin Bay and Hen Spring–Rogers Spring faults. This part of the basin contains as much as 3–4 km of basin-fill sediments (Langenheim et al., 2010) and has been interpreted as a pull-apart within a left step in the sinistral fault system (Campagna and Aydin, 1994). Significant deformation has affected the Overton Arm basin since the late Miocene, as evidenced by faulting and folding of the Muddy Creek Formation (Figs. 14B and 14D). Quaternary fault scarps are limited but include segments of the northeast-striking sinistral Rogers Spring fault (Fig. 14B) on the west side of the basin and northerly trending normal faults on the eastern margin (Fig. 3).

Significant halite occupies the Overton Arm basin 20–30 km north of the Colorado River, as evidenced by drill holes directly west of Lake Mead and sparse outcrops (Fig. 14; Longwell, 1928, 1936; Mannion, 1963). Exposures of halite in this area are the only observed outcrops of appreciable halite in the Lake Mead region. The exposed halite crops out as discrete sedimentary layers as thick as ∼4 m (Fig. 14C); these layers are not diapirs. Halite in the Overton Arm area was used as a source of salt by Native Americans and early settlers in the region and was mined prior to the filling of Lake Mead (Longwell, 1936; Netherland and Sewell, 1977). However, nearly all halite exposures are now submerged beneath the reservoir. The halite is generally pure and coarsely crystalline (Netherland and Sewell, 1977), with textures resembling those in the halite of the Hualapai basin. Siltstone interbeds are common within the halite, and anhydrite is typically interbedded in the upper and lower parts of the halite sequence. Glauberite is also relatively common and makes up more than 3% of the evaporite sequence. The halite interfingers with and is overlain by gypsum, siltstone, sandstone, and clay of the Muddy Creek Formation (Mannion, 1963; Muntean, 2013). The Muddy Creek Formation in the area is roughly bracketed between ca. 11 and 6 Ma (Howard et al., 2010). A 6 Ma basalt flow (Feuerbach et al., 1991) is intercalated in the Muddy Creek Formation ∼40–400 m above the top of the halite (Mannion, 1963) and places a younger age constraint on the halite in the Overton Arm basin (Figs. 14B–14D).

Five holes drilled by Stauffer Chemical Company in the early 1960s penetrated halite in the Overton Arm basin with thicknesses ranging from ∼305 m to more than 530 m (Fig. 14A; Table 4). Three of the holes bottomed in halite (Netherland and Sewell, 1977). The top of the halite ranges from ∼400 m (surface exposures above sea level) down to 260 m bsl (Table 3). The halite extends to at least 706 m bsl, as recorded in well SSC2. Documented gypsum-anhydrite in the area ranges up to ∼100 m in thickness, as observed in outcrops and drill holes. The late Miocene section is, however, deformed into folds and cut by faults in the area. The halite thins and becomes more impure to the south toward Echo Bay. Overall, the known halite extends over an area exceeding ∼90 km2.

The thickest known part of the halite (>530 m in the SSC2 drill hole) corresponds to the margin of the deepest part of the basin, as marked by the negative gravity anomaly (Figs. 3, 4, and 14C) and would account for about half of the observed gravity low. However, the thickest halite probably lies ∼5 km to the northeast of the drill holes near the center of the gravity low and presumably in the deepest part of the basin. Based on a gravity inversion with a constant density contrast of –600 kg/m3, the maximum thickness of halite in this basin is probably ∼1.5 km. Such a thickness is consistent with basement gravity values and low magnetic values in the area (Fig. 4). Gypsum-anhydrite is also probably significantly thicker in the deeper part of the basin as compared with that observed in outcrops and drill holes to the southwest.

Although core samples were not available for analysis, we envision a late Miocene depositional environment in the Overton Arm basin similar to that of the Hualapai basin. Proximal alluvial-fan facies graded basinward to a broad continental playa, where groundwater and/or surface water evaporated in salt pans yielding relatively thick sections of halite and gypsum. In contrast to the Temple Bar basin and southern Grand Wash Trough, however, lacustrine limestone did not accumulate in the Overton Arm basin, suggesting that a long-standing fresh water lake with a relatively stable level did not exist in the area in the late Miocene. The broad extent and appreciable thickness of the sandstone-siltstone facies in the Muddy Creek Formation in the Overton Arm basin (Fig. 14B) suggest that clastic input into this basin was significantly greater than that in the Hualapai basin, which in turn implies a greater fluvial contribution.

Detrital Basin

The Detrital basin is a 10–15-km-wide, northerly trending elongated basin that extends southward from Lake Mead for more than 90 km, with the Black Mountains on the west and White Hills and Cerbat Mountains on the east (Fig. 3). It connects northward with the Temple Bar and Overton Arm basins. Two major lobes comprise the Detrital basin and are separated by an interbasinal high ∼35–50 km south of Lake Mead, where basin deposits are less than 500 m thick. The northern lobe is a broad relatively shallow graben with sediments as much as 1.2 km thick, as evidenced by gravity data (Langenheim et al., 2010). It is bound by the steeply east-dipping Detrital fault on the west and west-dipping Blind Goddess fault on the east. Quaternary scarps and a major north-trending gravity and magnetic gradient mark the Detrital fault. The Detrital fault appears to have accommodated modest west-tilting of the Detrital basin, as evidenced by gently west-dipping reflectors in a seismic-reflection profile imaging the basin (Faulds, 1999).

Several holes drilled by Goldfield Consolidated Mining Company and Phelps Dodge Corporation from 1957 to 1959 in the northern lobe penetrated evaporites, including gypsum ∼16 km south of Lake Mead and halite farther south (Fig. 15). Evaporite tops in these wells range in elevation from 360 to 488 m asl (Peirce, 1974) with the uppermost halite at 338–414 m asl (Tables 3 and 4), similar to elevations of the top of the halite in the Hualapai basin to the east. Maximum observed thickness of halite in wells within the Detrital basin is 229 m, but this is a minimum as several wells bottom in halite. Transient electromagnetic profiles flown across this basin indicate low resistivity values to depths of 400 m, suggesting that evaporite deposits (including halite) extend to those depths (Truini et al., 2012). A gravity inversion with a constant density contrast of –600 kg/m3 suggests that the maximum thickness of halite in this basin is ∼800 m. The presence of substantial halite in the northern Detrital basin is particularly notable, as this area is not marked by a large negative gravity anomaly in contrast to the thick halite deposits in the Hualapai and Overton Arm basins.

The halite in the northern Detrital basin appears to be similar in composition and age to other halite deposits in the region. Although no samples were available, well logs indicate that the Detrital basin halite largely consists of massive interlocking crystals with minor amounts of intercalated siltstone (logs on file with Arizona Geological Survey), similar to that within the Hualapai basin. The Detrital basin halite appears to directly underlie and partially interfinger with gypsum deposits that correlate with the ca. 7.5–6.3 Ma gypsum in the western part of the Temple Bar basin (Virgin basin of Howard et al., 2010). We therefore infer that this halite is similar in age to the >6 Ma halite in the Overton Arm basin and also temporally correlative with at least the upper part of the halite in the Hualapai basin. It is also probable that the evaporite sequence in the Detrital basin is similar in age to the ca. 10–6 Ma evaporite-lacustrine section in the Temple Bar basin, because the two basins essentially merge in the Lake Mead area.

The southern lobe of Detrital Valley is underlain by a narrow (∼15-km-wide), north- to northwest-trending composite basin. The southern and eastern parts of this composite basin (Fig. 3) merge southward with the northern depocenter of Sacramento Valley and probably comprise an east-tilted half graben developed in the hanging wall of the west-dipping Cerbat Mountains fault (Langenheim et al., 2010). The northern part of this southern lobe is probably a west-tilted half graben developed in the hanging wall of the Mockingbird Mine fault and/or southern continuation of the Detrital fault zone. Although basin-fill sediments are generally less than 2 km thick, the southern lobe of Detrital basin may locally contain as much as ∼3 km of sediment (Fig. 3), with a major depocenter (2–3-km-thick basin fill) to the west of Dolan Springs, as evidenced by Bouguer and isostatic gravity data (Oppenheimer and Sumner, 1981; Langenheim et al., 2010).

Because drill-hole data indicate appreciable halite to the north in the shallower part of the basin, we speculate that the deeper southern lobe may also contain a relatively thick sequence of halite, and thus the depths to basement shown in Figure 3 may be overestimated. A gravity inversion, assuming that halite comprises all basin fill, suggests that the maximum depth of the basin is ∼1.5 km and thus the maximum thickness of the halite is also ∼1.5 km. Airborne transient electromagnetic profiles indicate low-resistivity values in the central and eastern parts of the gravity low at depths of ∼200–400 m, which were interpreted as saturated fine-grained basin fill (Truini et al., 2012). An alternate interpretation for these deposits is evaporites, such as halite. However, thick halite has yet to be documented by drill holes in this area.

The late Miocene depositional environment in Detrital basin was similar to that in the Overton Arm and Hualapai basins. It appears to have been characterized by proximal alluvial-fan facies grading basinward to a continental playa, where groundwater and/or surface water evaporated in salt pans yielding halite and gypsum. Available well data suggest relatively minor clastic input, which implies a minimal fluvial contribution in the central part of the basin. The apparent lack of carbonate deposition suggests that most of the Detrital basin did not contain a long-standing fresh-water lake during the late Miocene but was instead dominated by ephemeral saline lakes and playas.

Boulder, Nellis, and Las Vegas Basins

The Boulder basin in the western Lake Mead region essentially merges northwestward with the narrow (5–10-km-wide), west-northwest–trending Nellis basin, which in turn wraps around the north end of Frenchman Mountain and joins the Las Vegas basin to the west (Fig. 3). The northern part of the north-northeast–striking Frenchman Mountain fault zone truncates the Nellis basin on the west and forms an abrupt eastern margin to a 2–4-km-deep subbasin of Las Vegas Valley. The right-lateral Las Vegas Valley shear zone, which has accommodated ∼65 km of slip since the middle Miocene (Longwell, 1974; Wernicke et al., 1988), essentially bounds all three basins on the north. The main episode of extension in this area occurred ca. 13–9 Ma, as evidenced by decreasing tilts up-section within major half grabens (e.g., Duebendorfer and Wallin, 1991; Castor et al., 2000).

On the basis of gravity data, the Boulder basin appears to be a broad shallow (generally <1-km-thick section of basin-fill sediments) basin that lies directly east of Frenchman Mountain and consists primarily of middle to late Miocene sedimentary strata. Miocene strata in the basin are generally tilted gently to moderately eastward, with the magnitude of tilting progressively decreasing up-section from ∼50° in middle Miocene strata to subhorizontal in latest Miocene (ca. 6 Ma) units (Castor et al., 2000). Thus, the basin may correspond, at least in part, to an east-tilted half graben developed in the hanging wall of the Saddle Island detachment (Duebendorfer and Wallin, 1991). However, the northern extent of the Saddle Island detachment is poorly defined. The northern part of Boulder basin is complicated by an east- to northeast-trending fold belt. If once a half graben or series of half grabens, the Boulder basin has since been significantly modified by north-south shortening, possibly associated with the intersection of the left-lateral Lake Mead fault system and right-lateral Las Vegas Valley shear zone (see Anderson et al., 1994; Duebendorfer and Simpson, 1994). Although Boulder basin contains as much as 3 km of middle to late Miocene sedimentary rocks, it has a relatively subdued gravity signature indicative of a relatively shallow (generally <1 km) basin. This may result from the gentle dips of some of the bounding faults, such as the Saddle Island detachment. Another possible contributing factor arises from higher densities of these older, generally more consolidated sedimentary rocks as compared to those assumed in the gravity inversion. Quaternary faults are sparse within the Boulder basin, with one Holocene scarp (Meadview Slope fault) along the southeastern margin (Fig. 3; U.S. Geological Survey, 2010). However, late Miocene strata are locally deformed by minor faults and folds.

Relatively thick sections of middle to late Miocene sedimentary rocks and isostatic gravity data (Langenheim et al., 2001b) indicate that a small late Tertiary basin lies directly north of the Frenchman Mountain block and contains as much as 2 km of basin-fill sediment. This basin has been referred to as the Nellis basin, because it contains the eastern part of Nellis Air Force Base (Castor et al., 2000). The west-northwest elongated basin may have developed in a right step or pull-apart near the eastern end of the Las Vegas Valley shear zone (Fig. 3; Langenheim et al., 2001b). Isostatic gravity data define a 7-km-long basin that trends ∼N70°W, parallel to the eastern part of the shear zone. If the basin is filled with sediments of the same density of halite, the basin would be no more than 800 m deep. A pronounced magnetic low is also present here, although most, if not all, of the amplitude of this low results from the juxtaposition of magnetic basement beneath Frenchman Mountain to the south against nonmagnetic basement to the north across the Las Vegas Valley shear zone.

The Las Vegas Valley is a broad basin bound by Frenchman Mountain and the River Mountains on the east, Spring Mountains on the west, McCullough Mountains to the south, and a series of northerly trending ranges along the Las Vegas Valley shear zone to the north (Fig. 3A). The basin contains a complex system of subbasins associated with major normal faults and strands of the right-lateral Las Vegas Valley shear zone, as evidenced by isostatic residual and aeromagnetic data (Langenheim et al., 2001b). A system of west-dipping normal faults and right-lateral faults (e.g., the Frenchman Mountain fault) bounds Las Vegas basin on the east (Matti et al., 1993; Castor et al., 2000). Quaternary scarps mark several of these faults. Some of the deeper subbasins (2–4 km) occupy the northern and northeastern parts of Las Vegas Valley and are probably associated with right steps in the Las Vegas Valley shear zone (Langenheim et al., 2001b). Quaternary sediments blanket most of Las Vegas basin and obscure older basin fill. Seismic-reflection data suggest >2-km-thick sections of basin-fill sediments in subbasins along the northern margin of the valley. Taylor et al. (2008) noted that well logs show gypsum units in the Miocene section within Las Vegas basin, but little information is available on the overall composition of sediments and facies distribution within the basin. If the basin fill is composed entirely of halite, the inversion using the seismic constraints produces basement gravity values that are 10–20 mGal higher than gravity measurements on nearby basement outcrops. Given the uncertainties of the geophysical data, halite as thick as 300 m could be present in the basin.

Widespread late Miocene gypsum and a capping limestone unit mark the uppermost part of the Miocene section in the northern part of the Boulder basin and adjacent Nellis basin. There appears to be little break in deposition between gypsum and limestone, as evidenced by concordant dips and an apparent lack of major unconformities. Surface exposures and exploratory drilling indicate that the late Miocene gypsum covers at least 13 km2 and locally exceeds 40 m in thickness (Castor and Faulds, 2001). The limestone unit is 2–50 m thick and both overlies and interfingers with the upper part of the gypsum. The limestone crops out at elevations ranging from ∼595 m to 720 m asl, whereas the gypsum is exposed at elevations ranging from ∼450 m to 660 m asl. Fossil assemblages, textures, and 87Sr/86Sr isotopic values indicate a nonmarine origin for the limestone (Castor and Faulds, 2001; Roskowski et al., 2010). On the west, limestone exposures are truncated by northern strands of the Frenchman Mountain fault (Fig. 3A). A downthrown correlative of the limestone and gypsum units may be present beneath Quaternary sediments in the northeastern part of the Las Vegas basin but have yet to be documented by drill holes.

Geochronologic data indicate that the limestone and gypsum in the Nellis and Boulder basins are late Miocene (ca. 8.5–5.5 Ma) and temporally correlate with the regionally extensive Muddy Creek Formation. For example, 40Ar/39Ar and K/Ar data from previous studies in a variety of units indicate that strata beneath the gypsum range from ca. 11.6 to 8.5 Ma (e.g., Bohannon, 1984; Feuerbach et al., 1991; Harlan et al., 1998; Castor et al., 2000). Moreover, a tephra intercalated in redbeds ∼20 m beneath the limestone geochemically correlates with the 5.62 ± 0.11 Ma tuff of Wolverine Creek (e.g., Perkins et al., 1998; Pederson et al., 2001), as initially described by Castor and Faulds (2001) but with analytical data reported herein (Table 2). This datum provides a maximum age for the overlying limestone.

87Sr/86Sr values from the late Miocene limestone average 0.7108, similar to those of the Bouse Formation and substantially lower than any values obtained from the Hualapai Limestone to the east. Roskowski et al. (2010) interpreted the limestone in the Nellis basin as accumulating in a paleolake fed by Colorado River water and representing the upstream end of the Bouse lake system. This paleolake was reported to have an elevation up to ∼675 m and is referred to as Lake Las Vegas after Spencer et al. (2013), but we note that the Nellis limestone and associated marl crop out at elevations as high as 720–747 m asl (Fig. 3B).

The late Miocene depositional environment of the presumably connected Boulder, Nellis, and Las Vegas basins resembled that of other basins in the Lake Mead region but with some key differences. Similarities include marginal alluvial fan complexes and a continental playa setting, where gypsum accumulated. However, significant gypsum deposition probably did not begin until ca. 8.5 Ma, and limestone deposition initiated shortly after ca. 5.6 Ma, both somewhat later than in basins to the east. Details are lacking from the Las Vegas basin, but late Miocene deposits in the Nellis basin are truncated by faults bounding Las Vegas basin (Fig. 3), suggesting that at least some of the sedimentary fill within the northeastern Las Vegas basin correlates with that exposed in the Nellis basin. The lacustrine limestone within the Nellis basin suggests that the corresponding lake had an outlet, possibly to the west or south.

Virgin River Depression

The Virgin River depression is a large northeast-trending basin that covers >1500 km2 in southeastern Nevada, northwestern Arizona, and the southwest corner of Utah (Fig. 3A). The tectonic development and overall architecture of this basin have been assessed through geologic and geophysical investigations, including seismic-reflection and gravity surveys (Carpenter, 1989; Bohannon et al., 1993; Langenheim et al., 2001a). The basin is bound on the west by the Mormon Mountains and northern Muddy Mountains and on the east by the Virgin and Beaver Dam Mountains. The Virgin River, a major tributary of the Colorado River, flows through the central part of the depression (Fig. 3).

Gravity and seismic-reflection data indicate that the depression consists of two distinct sedimentary basins that correspond to east-tilted half grabens. The western half graben has been referred to as the Mormon basin and the eastern graben as the Mesquite basin (Bohannon et al., 1993). West-dipping listric normal fault systems bound the half grabens on the east. Late Miocene–early Pliocene sediments locally cover these faults, but the Piedmont fault cuts Quaternary alluvium on the east side of the Mesquite basin (Bohannon et al., 1993), and Quaternary scarps have been observed along much of the southeastern margin of the basin (Fig. 3; U.S. Geological Survey, 2010). A pronounced negative gravity anomaly (∼60–70 mGal) suggests that basin-fill sediment in the northern part of the Mesquite basin reaches thicknesses of 8–10 km (Langenheim et al., 2001a), nearly double that of other deep basins in the region. The Mormon basin is as much as 4.5–5 km deep, consistent with gravity data and seismic-reflection interpretations (Bohannon et al., 1993).

The upper part of the basin fill in the Virgin River depression is dominated by the late Miocene to early Pliocene Muddy Creek Formation composed primarily of fine- to medium-grained fluvial sandstone with subordinate intercalated conglomerate and siltstone, some of which is lacustrine (e.g., Williams et al., 1997; Forrester, 2009; Dickinson et al., 2014). The age of the Muddy Creek Formation in this area is poorly constrained but probably ranges from ca. 10 to 4 Ma (see summary in Dickinson et al., 2014). Seismic-reflection and well data indicate 1–2 km of late Miocene to early Pliocene sediments of the Muddy Creek Formation, including gypsiferous units, in the Virgin River depression (Bohannon et al., 1993). Only the upper ∼250 m of the Muddy Creek Formation are exposed in the area (Forrester, 2009). However, the Mobil Virgin River 1A well (Fig. 3), the only deep drill hole in the depression, penetrated 884 m of the Muddy Creek Formation in the Mormon basin. The upper 671 m in this well consist primarily of siltstone, whereas abundant gypsum was observed from 671 to 884 m depth (90–303 m below sea level in elevation; Bohannon et al., 1993). The lacustrine sediments (gypsum and siltstone) are more abundant to the south in the Mormon basin, but these beds are subordinate in surface outcrops (Muntean, 2012). Detrital zircon data preclude paleo–Colorado River sediment in at least the upper part of the Muddy Creek Formation in the Virgin River depression (Dickinson et al., 2014).

Thick halite and other evaporites have not been documented in the Virgin River depression, but well control and resistivity information are lacking in deeper parts of the basins. If the Mesquite basin is filled with sediments of the same density as halite, the thickest parts of the sedimentary section would be ∼5 km, more compatible with other basins in the region. Such a great thickness of halite, however, would result in basement gravity values under the basin that are 20–30 mGal higher than any measured on basement outcrops surrounding the depression. Although this is possible, the correlation of unusually high basement gravity values with the basin is suspicious. Halite as thick as 1 km, however, is possible, given the effects on the basement gravity and constraints from seismic-reflection data, as well as considering the areal extent and size of the Mesquite basin.

Although deep well data are lacking, the late Miocene depositional environment of the Virgin River depression was probably characterized by a continental playa, significant fluvial input in some areas, and surrounding alluvial fans. On the basis of detrital zircon data from the upper parts of the Muddy Creek Formation, fluvial input probably emanated from the northeast along an incipient Virgin River (Dickinson et al., 2014). Siltstone, gypsum, and possibly some halite accumulated in the playa. The upward progression from gypsum to siltstone noted in the Mobil Virgin River 1A well may record a trend toward an increase in water input in the late Miocene similar to many other basins in the region. However, it is noteworthy that late Miocene carbonate has not been documented in the Virgin River depression, suggesting lack of a long-standing, relatively stable fresh-water lake.

DISCUSSION

Our analysis of the distribution, age, and origin of late Miocene lacustrine and evaporative deposits in major basins of the Lake Mead region provides insights into the paleogeographic evolution of the region, particularly with respect to the relationships of such deposits to major episodes of extension and development of the Colorado River. Large-magnitude middle to late Miocene extension generated a vast system of half grabens and grabens within which multiple playas and lakes developed immediately prior to arrival of the Colorado River. Unusually thick nonmarine halite sequences accumulated in some of these basins. In this section, we review (1) the impacts of Miocene extension on the late Miocene paleogeography, (2) depositional model immediately prior to development of the Colorado River, (3) potential economic implications of the unusually thick halite deposits, and (4) possible regional and global analogues to the late Miocene setting in the Lake Mead region. Representative middle Miocene to Pliocene stratigraphic columns, as well as the approximate elevations of the major lacustrine and evaporite units in major basins in the Lake Mead region, are shown in Figure 16.

Paleogeographic Implications of Middle to Late Miocene Extension

As demonstrated by tilt fanning in the dated deposits in southern Grand Wash Trough and similar relationships in nearby areas (e.g., Faulds et al., 1992, 2010; Duebendorfer and Simpson, 1994; Beard, 1996; Duebendorfer and Sharp, 1998), major extension generated basins in the region from ca. 16 to 9 Ma (Fig. 9). Late Miocene to early Pliocene strata (ca. 12–5 Ma) are locally faulted and tilted (<5°) in the eastern Lake Mead region (Fig. 8; Wallace et al., 2005; Howard et al., 2010), but the primary episode of extension and basin development occurred from ca. 16–13 Ma in the Grand Wash Trough and Hualapai basin. The general physiography of the eastern Lake Mead region, including the abrupt western margin of the Colorado Plateau, series of deep basins in the hanging wall of the Grand Wash fault zone, and Grand Wash Cliffs, were established by ca. 9 Ma, as evidenced by the age of basin-fill deposits in the southern Grand Wash Trough (Figs. 8 and 9) and White Hills (Duebendorfer and Sharp, 1998; Faulds et al., 2010), as well as the correlative basalt flows at Snap Point and Nevershine Mesa adorning the Grand Wash Cliffs (Figs. 3, 11, and 12). In contrast, tilt fanning brackets major extension in the western Lake Mead region to between ca. 13 and 9 Ma (Duebendorfer and Simpson, 1994; Castor et al., 2000). Latest Miocene to early Pliocene strata in the central to western Lake Mead region are also locally tilted and faulted, as exemplified in both the Overton Arm and Nellis basins (Figs. 3 and 14B; Beard et al., 2007).

These relations indicate that the extensive late Miocene lacustrine and evaporite deposits in the Lake Mead region are primarily late synextensional and largely postdate the major episode of extension. For example, the Hualapai Limestone in the eastern to central Lake Mead region is constrained to ca. 12–7.5 Ma in the Grand Wash Trough and ca. 10 to <6 Ma in the Temple Bar area (Figs. 3 and 16; Table 3). The thick halite sequence in the Hualapai basin is likely ca. 12 to <5.6 Ma, similar in age to the Hualapai Limestone. Thus, these deposits postdate the ca. 16–13 Ma episode of major extension. Similarly, generally subhorizontal late Miocene gypsum and limestone deposits in the western Lake Mead region (Boulder and Nellis basins) are bracketed between ca. 8.5 and 5 Ma, which immediately postdates the 13–9 Ma period of major extension. The general younging trend to the west of the late Miocene lacustrine and evaporite deposits in the Lake Mead region is consistent with both the apparent westward progression of extension and proposed downstream filling and spilling of lakes inferred for the evolution of the lower Colorado River (e.g., House et al., 2008; Spencer et al., 2008).

The late synextensional timing for extensive lacustrine and evaporite deposition is also consistent with the structural and paleogeographic evolution of an actively extending region. Significant accommodation space must be generated in half grabens to host such deposits (Fig. 1A), and thus it would be unlikely to form thick sequences during the early stages of extension. Furthermore, significant drainage evolution is needed to develop broad enough catchments to supply sufficient groundwater and/or surface water over prolonged time periods to accumulate appreciable sequences of lacustrine and evaporite deposits. High rates of extension would promote continued isolation of individual basins and hinder regional drainage and groundwater flow-path development, especially in arid regions. Once extension wanes, basins can more easily fill and drainage networks can progressively integrate originally isolated basins into broader regional drainage systems. In addition, lower overall topographic relief during the waning stages of extension will reduce steep-gradient fluvial environments and thus the dominance of alluvial fan facies while inducing development of broad, low-gradient plains, shallow water bodies, and associated fine-grained deposits. This, in turn, would promote a high surface-to-volume ratio for late synextensional water bodies and thus greater evaporative surfaces and resulting evaporite deposits. Broader water bodies would also increase the likelihood of groundwater connections through interbasinal sills and possibly some mountain blocks, aided in both cases by preexisting faults. Additionally, continued mild extension and associated relative subsidence of basins would be critical in preserving thick halite by allowing for its subsidence below the stagnant point in the phreatic zone.

In addition to rates of extension, climatic conditions are crucial in the drainage development of any extended region. Our inferences on the relations between tectonism and drainage development in the Lake Mead region may best apply to a region that experienced (1) a major pulse of extension with relatively high strain rates followed by significantly reduced rates in the waning stages and (2) a relatively persistent arid climate. The latter appears to have characterized the late Miocene in the southwestern United States (e.g., Chapin, 2008). In fact, the mean annual temperature in the late Miocene may have been 3–8 °C warmer than the contemporary climate, as evidenced by carbonate clumped-isotope thermometry from the Bouse Formation and Hualapai Limestone (Huntington et al., 2010; Wernicke, 2011).

Late Miocene–Early Pliocene Depositional Model

The spatial distribution of late Miocene lacustrine limestone and evaporite deposits in the Lake Mead region reflects a broad depositional system of interconnected lakes and continental playas (Figs. 1, 3, and 17). This system initiated at ca. 12–11 Ma in the Grand Wash Trough and continued to ca. 5.6–5.3 Ma in much of the Lake Mead region, including the Hualapai basin in the east, Virgin-Detrital trough in the central part of the region, and Boulder and Nellis basins in the west. This system of lakes and playas immediately predates arrival of the Colorado River (Fig. 16).The late Miocene–early Pliocene lacustrine deposits (e.g., Hualapai Limestone and Bouse Formation) and associated lakes have long been recognized in the region (e.g., Longwell, 1936; Lucchitta, 1966; Wallace et al., 2005; House et al., 2008; Spencer et al., 2008, 2013), but our review of the regional extent and temporal relationships of evaporite deposits (e.g., halite and gypsum) documents that extensive playas were developing in concert with the lakes.

The most extensive late Miocene limestone and gypsum deposits in the Lake Mead region crop out within ∼25 km of the subsequently developed Colorado River. For example, most of the Hualapai Limestone in the Grand Wash Trough and Temple Bar basin lies within ∼10–25 km of the present course of the river (Fig. 3). The most extensive gypsum deposits crop out in the northern part of the southern Grand Wash Trough and along or directly south of the present course of the Colorado River in the central Temple Bar and northern Detrital basins. In the western Lake Mead region, exposures of gypsum and limestone in the Boulder and Nellis basins are concentrated ∼10–25 km northwest of the current path of the Colorado River, with the limestone extending westward to the faulted eastern margin of the deep northeastern lobe of the Las Vegas basin. Notably, late Miocene limestone consistently overlies gypsum throughout the Lake Mead region, albeit the lower part of the limestone locally interfingers with the upper section of gypsum, and interbeds of both occur in the other.

In contrast to the limestone and gypsum, thick sequences of halite have been documented in basins ∼20–50 km to both the south and north of the Colorado River, including the Hualapai, northern Detrital, and Overton Arm basins (Figs. 3 and 16A) but not proximal to the Colorado River. It is possible that some halite may also occupy the axial basins (i.e., proximal to the Colorado River), but significant gravity lows do not coincide with these areas (Fig. 4), suggesting that thick halite is unlikely. Moreover, large-magnitude and areally extensive gravity lows imply that thick halite may lie in the subsurface of several other satellite basins to the Colorado River in the region, including the northern Grand Wash, southern Detrital, northeastern Las Vegas, Mormon Mesa, and Mesquite basins (Fig. 3), in all cases, tens of kilometers away from the axial drainage and associated late Miocene limestone deposits.

Although some faulting has occurred in the region since the late Miocene, broadly similar elevations (Fig. 16B) of limestone and evaporite (gypsum and halite) deposits and similar Sr isotopic values in the Hualapai Limestone (Roskowski et al., 2010; Spencer et al., 2013) in various basins suggest an interconnected hydrologic system in the eastern to central Lake Mead region during at least part of the late Miocene. Limestone elevations range from ∼500–912 m asl in the Grand Wash Trough and 367–720 m asl in the Temple Bar area (Table 3). Gypsum elevations range from ∼370–750 m asl in the Grand Wash Trough and ∼390–440 m asl in the Temple Bar and Detrital basins. We note that down-to-the-west displacement along the Wheeler Ridge and Lost Basin Range faults since the late Miocene is at least partly responsible for the higher elevations in the Grand Wash Trough. Halite ranges from less than ∼2000 m bsl to 418 m asl in the Hualapai basin, less than ∼170–414 m asl in the Detrital basin, and less than 706 m bsl to 163 m asl in the Overton Arm basin. These elevations demonstrate the progression from gypsum to limestone up-section in several of the basins, as well as the lower elevations of halite in some of the satellite basins, compared to limestone in axial basins. Coeval deposition of limestone in axial basins and halite in satellite basins suggests some groundwater and/or surface water input from standing lakes to adjacent playas. The lower elevations of halite in the Hualapai and Overton Arm basins further imply that some of the satellite basins were subsiding more rapidly and essentially deeper than axial basins, as supported by depth-to-basement calculations (Fig. 3). However, continued tectonism may have induced continued subsidence since the late Miocene in some areas, such as the Overton Arm basin.

Elevations of limestone (∼595–675 m) and gypsum (∼450–595 m) farther west in the Nellis basin are also broadly similar with those of limestone and gypsum to the east (Fig. 16B; Table 3), but Sr isotopic values of the Nellis limestone differ significantly from those from the Hualapai Limestone (Roskowski et al., 2010), suggesting that the hydrologic system in the western Lake Mead region differed from that to the east. Nonetheless, the late Miocene stratigraphic succession from gypsum to limestone in the Boulder and Nellis basins is similar to that in basins to the east, although thick halite has not been documented in the Boulder, Nellis, and Las Vegas basins.

These spatial and temporal patterns of lacustrine and evaporite deposition suggest that the late Miocene to earliest Pliocene paleogeography of the Lake Mead region was initially characterized by a system of axial playas, within which silts and gypsum were deposited, with halite accumulating in broad salt pans within satellite basins to the north and south, at least in the eastern to central Lake Mead region (Fig. 17). These playas were probably largely fed by groundwater discharge, as evidenced by textures within the halite, interbedded muds, and apparent scarcity of fluvial sediments. To the north, however, greater fluvial input appears to have characterized the Virgin River depression. Nonetheless, the influx of groundwater and/or surface water appears to have increased region-wide during the late Miocene, because the initial axial playas gave way to lakes, marshes, and expanding fluvial systems, as marked by the Hualapai Limestone in the eastern to central Lake Mead region, fluvial-dominated sediments in the upper part of the Muddy Creek Formation in the Virgin River depression, and Nellis limestone in the western part of the region.

On the basis of the lateral distribution, elevation, age, and geochemistry of the lacustrine carbonate deposits, three sequential late Miocene paleolakes have previously been defined in the Lake Mead area: (1) Lake Grand Wash, (2) Lake Hualapai, and (3) Lake Las Vegas (Spencer et al., 2013; Pearthree and House, 2014; Figs. 16 and 17). Lake Grand Wash (ca. 12–7.5 Ma, 912 m) inundated the Grand Wash Trough and Hualapai basin and was contained by inferred paleodams between Gregg Basin and Temple Bar on the west and between the Grand Wash Trough and Virgin River depression on the northwest. Deposits of the Hualapai Limestone reach a maximum elevation of 912 m in the Grand Wash Trough (Wallace et al., 2005; Beard et al., 2007). Lake Hualapai (ca. 10–5.6 Ma, 720 m) covered the Temple Bar area and backed up into northern Detrital Valley and Grand Wash Trough but probably did not occupy the Hualapai basin. Deposits of the Hualapai Limestone reach a maximum elevation of 720 m between Detrital Valley and the Temple Bar area and provide the basis for this lake level elevation (e.g., Beard et al., 2007; Felger and Beard, 2010). Based on the western and northwestern extent of the Hualapai Limestone, Lake Hualapai may have been dammed ∼10–15 km west and north of Temple Bar (Fig. 17). Notably, the more than 275 m of displacement along the Wheeler Ridge fault since the late Miocene can account for the difference in elevation between Lake Hualapai and Lake Grand Wash. Identical in upper elevation to Lake Hualapai, Lake Las Vegas (ca. 5.6–5 Ma, 720 m) covered the entire Lake Mead area up to at least ∼720 m, as defined by the upper elevation of the Nellis Limestone. Our estimate for the upper elevation of Lake Las Vegas at 720 m is significantly higher than the 650 m level surmised in previous analyses (e.g., Spencer et al., 2013; Pearthree and House, 2014). Lake Las Vegas was probably dammed in the Black Canyon area (Fig. 17), where it eventually spilled over to form paleo–Lake Mohave and initiate deposition of the Bouse formation (e.g., Pearthree and House, 2014).

The isotopic geochemistry of the Hualapai and Nellis limestone deposits suggests hydrochemically distinct basins persisted in the western versus the eastern-central Lake Mead region up to ca. 6 Ma. The Grand Wash and Hualapai lakes were probably initially fed by springs with an increasing influx of meteoric water from ca. 8–6 Ma (Crossey et al., 2002, 2006, 2015), which may have been a direct precursor to the Colorado River (Faulds et al., 2001c). As the axial basins transitioned to more permanent lakes and wetlands, halite deposition appears to have continued in at least some of the satellite basins (Hualapai and Overton Arm basins), persisting to ca. 6 to <5.6 Ma. Thus, we envision a system of axial playas progressively replaced by lakes and becoming more integrated by surface drainage through the late Miocene. These axial basins were consistently or at least episodically connected, either through surface spillways or groundwater, to broad salt pans in continental playas in satellite basins (Figs. 17 and 18). The interconnectedness of the axial basins in the eastern to central Lake Mead region is supported by relatively similar Sr isotopic values in the scattered exposures of the Hualapai Limestone (Roskowski et al., 2010; Spencer et al., 2013). Some of the satellite basins with extensive salt pans (e.g., northern Detrital basin) appear to have been directly connected to axial basins (e.g., Temple Bar) and essentially represent the more distal parts of large composite basins. In other cases, however, relatively low topographic sills likely separated the satellite basins from the axial basins. For example, a low topographic sill composed primarily of alluvial fan deposits probably separated the Hualapai basin from axial basins and lakes to the north. Permeable alluvial fan sediments may have permitted significant groundwater influx from both Lake Grand Wash and Lake Hualapai into the Hualapai basin, facilitating deposition of the unusually thick sequence of halite. Our hypothesized interconnectedness between the axial and satellite basins (e.g., Hualapai, Overton Arm, and others) could be tested through isotopic analysis of late Miocene gypsum and anhydrite deposits exposed in the associated basins and available from core in the Hualapai basin.

The Lake Mead region was integrated into the Colorado River between ca. 5.3 and 4.9 Ma, as evidenced by the ca. 6 Ma upper Hualapai Limestone in the Temple Bar basin, ca. 6 Ma uppermost halite in the Overton Arm basin, <5.6 Ma limestone in the Nellis basin, <5.6 Ma uppermost halite in the Hualapai basin, and 4.5 Ma Sandy Point basalt intercalated with Colorado River gravels in the Grand Wash Trough (Fig. 16). The Colorado River debouched into the Blythe area by ca. 4.9 Ma (Spencer et al., 2013) and possibly into the Gulf of California as early as ca. 5.3 Ma (Dorsey et al., 2007). Because there is no evidence that any of the lacustrine and evaporite deposits in the Lake Mead region are any younger than ca. 5.3–4.9 Ma, it follows that the Lake Mead region was integrated into the Colorado River between ca. 5.3 and 4.9 Ma.

The approximate coincidence in the timing of cessation of halite deposition in satellite basins (shortly after 5.6 Ma), termination of lacustrine deposition in axial basins, and regional development of a through-going Colorado River suggests that deposition of both the late Miocene lacustrine deposits and evaporites in the Lake Mead region was directly linked to immediate precursors of the Colorado River. This begs the question as to whether a proto–Colorado River fed the lakes and playas, and if so, from where did it originate. However, a lack of fluvial sediments either intercalated with or underlying the limestone and evaporite deposits (e.g., Lucchitta, 1989), as well as the isotopic geochemistry of the Hualapai Limestone (Roskowski et al., 2010; Crossey et al., 2015), appear to preclude a major river system emptying into the Lake Mead region prior to ca. 5.6 Ma. The interpretation of Lovejoy (1980) that an early Colorado River entered into the Grand Wash Trough, turned sharply south, and flowed into the Hualapai basin, thus producing the thick halite, is not supported by the Sr isotopic values in the Hualapai Limestone in the Grand Wash Trough and lack of extensive fluvial and deltaic deposits in the Grand Wash Trough and Hualapai basin.

The late Miocene lakes and playas in the Lake Mead region were probably fed instead by a complex system of springs and lesser amounts of surface water (e.g., Crossey et al., 2015) derived primarily from the Colorado Plateau. The Colorado Plateau is implicated as the main source by the spatial association of the thickest and apparently oldest late Miocene limestone and halite in basins (Grand Wash Trough and Hualapai basin) directly adjacent to its western margin. The springs were probably associated with extensive and complex carbonate aquifers in the southwestern Colorado Plateau (e.g., Huntoon, 1996, 2000; Crossey et al., 2006, 2011; Hill and Polyak, 2014). This further implies that Permian redbeds on the Colorado Plateau were a primary source of Na and Cl for the late Miocene halite deposits. The proposed link between evaporite deposits in the Lake Mead region with groundwater derived from the Colorado Plateau could also be tested through isotopic analysis (e.g., Sr) of late Miocene gypsum and anhydrite deposits.

The transition from evaporitic to normal lacustrine deposition in the axial basins of the Lake Mead region indicates an increasing input of fresh water from this groundwater system and possibly some surface water during the late Miocene. This may reflect a combination of a progressive increase in the size of the catchment, resulting in part from waning tectonism, and possibly evolving climatic effects. As axial lakes filled with fresh water, the regional groundwater table rose, and a prolonged period of groundwater discharge and evaporation ensued in satellite basins (Figs. 1, 17, and 18), ultimately producing some of the thickest known nonmarine halite deposits.

Lake Grand Wash was the first in the series of lakes to develop in the Grand Wash Trough and the neighboring Gregg and Hualapai basins from ca. 12 to 7.5 Ma (Fig. 17A). The drainage may have emanated from a series of springs in the southern Grand Wash Trough, where water ponded in lakes and marshes leading to accumulation of the Hualapai Limestone. The lake in the southern Grand Wash Trough probably drained to both the north and south into continental playas in the deeper northern Grand Wash Trough and Hualapai basin, respectively, within which thick halite and/or gypsum accumulated. Some of the water in Lake Grand Wash may have also drained westward into Lake Hualapai in the Temple Bar and Detrital basins.

Lake Hualapai ponded in at least the Temple Bar and northern Detrital basins (Fig. 17B) beginning as early as ca. 10 Ma and continuing to at least 5.6 Ma, thus partially overlapping in time with Lake Grand Wash. Lake Hualapai probably drained into deeper basins to both the north and south (e.g., Overton Arm, Mormon, and Mesquite to the north and Hualapai and southern Detrital to the south) through a combination of groundwater sapping and possibly surface runoff and spillovers. Sr isotopic data from the Hualapai Limestone deposited in both Lake Grand Wash and Lake Hualapai indicate no connection with the Colorado River (Spencer and Patchett, 1997; Roskowski et al., 2010; Spencer et al., 2013).

Lake Las Vegas was the last in the series of late Miocene lakes to develop in the Lake Mead region (Fig. 17C). An initial playa setting formed by ca. 8 Ma and was followed by lacustrine carbonate deposition (Nellis limestone) shortly after ca. 5.6 Ma, which may have persisted to ca. 5.0 Ma. Elevations of the Nellis limestone suggest that this lake had maximum levels similar to that of Lake Hualapai, although late Miocene to recent tectonism, especially in the Frenchman Mountain-Nellis basin area, has probably altered these elevations. Nonetheless, it is possible that Lake Las Vegas backed up into the same region as Lake Hualapai, leading to continued groundwater sapping and/or spillover into several satellite basins, such as the Overton Arm and Hualapai basins.

It seems likely that the deep satellite basins distributed tens of kilometers to the north and south of the present course of the Colorado River, such as the Hualapai, northern Grand Wash, Mesquite, and Overton Arm basins (Fig. 17), had to fill to a certain threshold before the drainage could expand significantly and spillover into the next set of downstream catchments. For example, the relatively deep (∼4 km) Hualapai basin and northern Grand Wash Trough probably served as initial terminal sinks for an incipient drainage system that fed Lake Grand Wash in the eastern Lake Mead region. Significant filling of these basins probably occurred prior to appreciable spillover into the Temple Bar basin. Similarly, it is likely that the deeper basins in the Virgin River depression and Virgin-Detrital trough served as sinks farther west and north. Both Lake Hualapai and Lake Las Vegas may have initially drained to the north into the Overton Arm and Mesquite basins, as well as southward, possibly through groundwater sapping, into the Hualapai and southern Detrital basins. However, such interpretations should be considered tentative, with available data, due to the vagaries of erosion and variable local uplift and subsidence resulting from subsequent tectonism and sediment loading.

It is noteworthy that the Nellis limestone and underlying gypsum deposits in the western Lake Mead region lie well to the northwest of the present course of the Colorado River, with no intervening bedrock barriers to the deep northeastern lobe of Las Vegas basin (Fig. 3). It is therefore probable that Lake Las Vegas connected and possibly emptied westward into the Las Vegas basin (Fig. 17). Interestingly, the Nellis limestone has similar Sr isotopic ratios to that of the Bouse Formation, suggesting that it was associated with an incipient Colorado River rather than with the Hualapai Limestone to the east (Roskowski et al., 2010; Spencer et al., 2013). We therefore suggest that the incipient Colorado River in the Lake Mead region briefly emptied into Las Vegas basin prior to spilling to the south through Black Canyon and becoming integrated with basins to the south of Lake Mead. Such a model is compatible with evaporites partially accounting for the significant gravity low in the northeastern part of Las Vegas Valley. Well logs indicate that the late Miocene section within the Las Vegas basin contains relatively thick gypsum (Taylor et al., 2008). As a proto–Colorado River in the Lake Mead region initially spilled over into a system of downstream lakes and eventually emptied into the Gulf of California as a through-going fluvial system by ca. 5.3–4.9 Ma (e.g., Dorsey et al., 2007; Spencer et al., 2013), the corresponding loss of axial lakes in the Lake Mead region combined with the likely channeling of groundwater systems in the southwestern part of the Colorado Plateau to the Grand Canyon would have collectively induced a significant decrease in groundwater discharge in satellite basins and ultimately terminated significant evaporite deposition even in areas, such as the Hualapai basin, that remained internally drained.

Available geochronologic data place some constraints on rates of deposition of halite in the Lake Mead region. For example, 79 m of halite and 136 m of halite and anhydrite accumulated in the Hualapai basin above the 5.6 Ma tephra prior to integration of the region into the Colorado River by 5.3–4.9 Ma. This suggests depositional rates of evaporites of ∼0.19–0.45 m/1000 yr (or ∼190–450 m/m.y.), assuming that evaporite deposition ceased between ca. 5.3 and 4.9 Ma. If the Lake Mead region was integrated into downstream lakes earlier than 5.3 Ma, evaporite depositional rates would be higher. Such rates are compatible with documented rates of halite deposition (e.g., Schreiber and Hsü, 1980; Roveri et al., 2008; Manzi et al., 2009). These rates are also compatible with the entire 2.5-km-thick sequence of halite and capping anhydrite in the Hualapai basin accumulating in ∼5–7 m.y. (from ca. 12 Ma to 5 Ma), or roughly coincidental with deposition of the ca. 12–6 Ma Hualapai Limestone in the region.

Economic Implications

Evaporite minerals, such as halite and gypsum, are important industrial minerals that have many societal applications. For example, late Miocene gypsum is currently mined in the western Lake Mead area and processed into wallboard by PABCO Gypsum. Miocene–Pliocene halite in the Luke basin west of Phoenix (Fig. 2) has also been mined for its mineral content. In addition, man-made caverns within the Luke salt body are used for storage of liquid petroleum gas (Spencer, 2005). Clearly, the thick and widespread halite deposits in the Lake Mead region represent an important economic resource with many potential industrial applications (e.g., mineral resources and natural gas storage). It should be noted, however, that water use and brine disposal are key challenges for developing salt cavern storage in the Neogene basins of this arid region.

Nonetheless, our synthesis indicates that the Lake Mead region may contain as much as ∼700 km3 of halite. Because most halite in the region accumulated in half grabens and is partly synextensional, lenticular wedge-shaped bodies probably characterize the geometry of the halite deposits, similar to that observed in the Hualapai basin (e.g., Faulds et al., 1997). These geometries can be modeled as quarter ellipsoids for purposes of estimating volumes (Fig. 19). The largest halite deposits in the Lake Mead region appear to be hosted by the Hualapai and Mesquite basins, where halite volumes probably approach and possibly exceed 200 km3.

Regional and Global Analogues

Although the southwestern United States contains dozens of Cenozoic basins (Figs. 2 and 3), thick halite has been documented in relatively few (e.g., Mannion, 1963; Peirce, 1976; Faulds et al., 1997). However, most Cenozoic halite deposits in this region, as well as many elsewhere, are buried and do not crop out (e.g., Peirce, 1976; Faulds et al., 1997). In such cases, the regional setting and geophysical techniques, particularly gravity and magnetics, can be utilized to determine the most likely locations for subsurface deposits and to design other geophysical surveys, such as electrical, seismic reflection, and even heat flow, to test whether the gravity and magnetic lows are caused by halite, rather than other low-density fill, and to ultimately determine the depth and thickness of potential halite deposits.

The spatial and temporal relations in the Lake Mead region between lacustrine and evaporative sequences and development of the Colorado River provide a conceptual model for examining the potential for thick halite and related evaporites in downstream reaches of the Colorado River south of Lake Mead and in major tributaries of the Colorado River in the arid southwestern United States, especially the Gila River drainage in southern Arizona (Fig. 2). For example, lacustrine carbonate facies are more common in the Lake Mead region within the axial basins proximal to the subsequent Colorado River, with thick evaporite sequences (particularly halite) characterizing nearby satellite basins that presumably served as terminal sinks to segments of the developing drainage system. Some of these basins served as regional sinks for prolonged periods (>5 Ma) and thus accumulated thick (1–2.5 km) evaporite sequences. These relationships indicate that position within a drainage system should be factored into estimating the potential for thick halite and related evaporites, as a general correlation may exist between this position, time interval of ponding, and thickness of evaporite sequences.

Such a model has relevance to understanding both the evolution of these drainage systems as well as for guiding exploration models for identifying thick, economically viable halite deposits. For example, assuming that the Gila River (Fig. 2) evolved similar to the Colorado River through a chain of progressive downstream sapping and spilling of lakes, the thickest halite would be expected in the more upstream reaches, still within the Basin and Range Province but proximal to the highlands of the Colorado Plateau. In such areas, large quantities of fresh water could be derived from the Colorado Plateau, either through surface drainage or groundwater systems, large deep basins would facilitate prolonged ponding of waters and contain sufficient accommodation space for accumulation of thick evaporative sequences, and a warm arid climate would afford rapid evaporation rates (Figs. 1 and 2). Notably, thick halite has been documented within this general setting in the Gila River drainage network, including the Luke and Picacho basins in south-central Arizona (Fig. 2; Peirce, 1976, 1981; Rauzi, 2002c; Spencer, 2005).

All basins proximal to the lower Colorado River to the south of Lake Mead also have significant potential for containing thick halite deposits. However, the lower reaches of the Colorado River were integrated relatively rapidly through sequential, downstream spillovers (e.g., House et al., 2005, 2008; Spencer et al., 2008, 2013). Thus, compared to the Lake Mead region, where some lakes and associated playas persisted for as long as ca. 5 Ma (e.g., Lake Grand Wash with apparent terminal sinks in Hualapai basin and northern Grand Wash Trough; Figs. 3B and 17), lakes downstream from Lake Mead, associated with development of the Colorado River, were probably more transient features, possibly lasting for only a few hundred thousand years or less. Shorter increments of ponding of the developing Colorado River would presumably result in thinner evaporative sequences. However, once the Colorado River had evolved to the point of progressively cascading southward in a series of lakes toward the Gulf of California, it probably had a much larger catchment area compared to earlier drainage and groundwater systems emptying into the Lake Mead region. Considering the potential high rates of evaporite deposition (e.g., Schreiber and Hsü, 1980) and large volumes of water emptying into regional, albeit short-lived sinks, it is possible that some of the basins proximal to the Colorado River south of Lake Mead contain relatively thick sequences of halite. These include the Dutch Flat, Sacramento, Chemehuevi, Blythe-McCoy, Yuma, and Mohawk basins (Fig. 2). Thick halite may be more likely in the Blythe-McCoy and Yuma basins due to possible late Miocene–early Pliocene marine incursions in this area (McDougall, 2008; McDougall and Martinez, 2014). Nonetheless, documentation of appreciable halite in these basins awaits comprehensive analysis of well and geophysical data.

The overall extent of thick Neogene evaporite deposits in the southwestern United States, stretching from central Arizona to the Lake Mead region, marks a belt of evaporites that rivals those in other parts of the world, including the central Andes (Alonso et al., 1991), southern Spain (Rodríguez-Fernández and Sanz de Galdeano, 2006; García-Veigas et al., 2015), and many passive continental margins (Tankard and Balkwill, 1989). Some of the basins in central and southern Spain may be broadly analogous to those in the southwestern United States. For example, although the Granada Basin began as part of a much larger and more extensive marine basin from ca. 8.5 Ma to 7.2 Ma, it gradually became cut off from the ocean and continued to widen while remaining deep but nonmarine (Rodríguez-Fernández and Sanz de Galdeano, 2006; García-Veigas et al., 2013, 2015; Navarro-Hervas et al., 2014). During a final period of cut-off from marine water sources (end Tortonian), marine halite deposition transitioned into nonmarine halite with added CaCl2 (hydrothermal). The top of the section accumulated as a shallow nonmarine gypsum-anhydrite deposit. Although the initial phase of marine deposition in the Granada basin clearly differs from the sequence of events in the Lake Mead region, the overall extent of nonmarine evaporites in southern Spain is comparable to those in the Lake Mead region and southwestern United States. Moreover, the relationships between the thick evaporite sequences and chronology of both regional extension and drainage development in the southwestern United States may elucidate the evolution of other nonmarine evaporative belts in extensional settings throughout the world and help to guide exploration efforts for thick subsurface halite deposits.

CONCLUSIONS

The late Miocene landscape in the Lake Mead region contained a series of lakes, wetlands, and playas, which stretched from the mouth of the Grand Canyon in the Grand Wash Trough westward to the Las Vegas basin. Thick late Miocene nonmarine evaporite (primarily halite and gypsum) and related lacustrine limestone deposits accumulated in lakes and playas throughout the region. 40Ar/39Ar and tephrochronologic data, regional relationships, and a progressive upward decrease in tilt in some basins indicate that these deposits are late synextensional, ranging from ca. 12 to 5 Ma. The uppermost deposits immediately predate the arrival of the Colorado River. The late synextensional setting allowed for development of sufficient accommodation space in half grabens and deposition of relatively thick sequences of lacustrine and evaporite sediments. Continued subsidence of basins also allowed for the preservation of thick sequences of halite below the stagnant level in the phreatic zone. Furthermore, waning rates of extension facilitated development of drainage networks large enough to supply appreciable groundwater and some surface water to lakes and playas. Lower overall topographic relief in the late synextensional setting also induced development of broad, low-gradient plains and shallow water bodies, which, in turn, promoted high surface-to-volume ratios for late synextensional water bodies and thus greater evaporative surfaces and resulting evaporite deposits.

The distribution, age, and composition of late Miocene deposits in the Lake Mead region suggest a broad depositional system involving axial playas and subsequent lakes (along and proximal to the eventual course of the Colorado River), with extensive playas and salt pans in satellite basins within ∼50 km of the axial basins. In the axial basins, gypsum deposition transitioned to limestone accumulation as fresh-water input increased during the late Miocene. Evaporite deposition dominated many of the satellite basins, with thick halite (∼200–2500 m) accumulating in the Hualapai, Detrital, and Overton Arm basins. Large-magnitude negative gravity anomalies indicate that thick undiscovered halite may comprise a significant part of the fill within several other satellite basins, including the northern Grand Wash, Mesquite, southern Detrital, and northeastern Las Vegas basins (Fig. 3). Total halite volume in the region probably exceeds 700 km3. On the basis of the age and distribution of lacustrine limestone, it would appear that late Miocene fresh-water lakes first formed ca. 12 Ma in the Grand Wash Trough adjacent to the Colorado Plateau (Lake Grand Wash) and then progressed westward to the central Lake Mead area (Lake Hualapai) ca. 10 Ma, and finally to the Las Vegas area by ca. 5.6 Ma (Lake Las Vegas). Sr isotopic values from the <5.6 Ma limestone deposited in Lake Las Vegas are similar to the Bouse Formation and may reflect early arrival of the Colorado River in the region (Roskowski et al., 2010; Spencer et al., 2013). All of the late Miocene lakes in the region overlapped spatially and/or temporally (Figs. 17 and 18). The thickest and most long-lived late Miocene lacustrine and evaporite deposits thus far documented in the region reside in half grabens in the eastern Lake Mead region, suggesting that the Colorado Plateau was a major source of groundwater and possibly some surface water to the basins. Isotopic analyses of late Miocene gypsum and anhydrite throughout the region are needed, however, to test this hypothesis, as well as our models for a series of interconnected basins and evolving hydrologic regimes in the late Miocene (Figs. 17 and 18).

The relations between the thick evaporite sequences and chronology of both regional extension and drainage development in the Lake Mead region may elucidate the evolution of other nonmarine evaporite sequences in the southwestern United States and elsewhere, helping to guide exploration efforts for thick subsurface halite deposits. Our conceptual model suggests that the most extensive nonmarine evaporite sequences will be late synextensional and occur in basins with large catchments proximal to developing river systems and/or broad elevated terranes (e.g., Colorado Plateau).

This work was funded by a variety of sources over several years, including grants awarded to Faulds from the National Science Foundation (EAR99-10977 and EAR04-09913) and EDMAP program of the U.S. Geological Survey (Cooperative agreement #1434-HQ-97-AG-07146). In addition, Unocal and LK Energy partially funded some of the research on the halite deposits. The U.S. Geological Survey in Las Vegas kindly provided a field vehicle for substantial amounts of this work, for which we thank Gary Dixon and Peter Rowley. We also thank the National Park Service at the Lake Mead National Recreation Area for extensive logistical support over many field seasons, including boat access into remote areas. Kent Turner and Darlene Carnes with the Lake Mead National Recreation Area were especially helpful. We also greatly appreciate Mark Odegard, Grizzly Geosciences, Inc., and Bill Cathey, Earthfield Technology, for drawing our attention to and demonstrating the utility of gravity and magnetic data in delineating salt bodies in these basins. This research has also benefited from fruitful discussions with Jon Spencer, Sue Beard, Kyle House, Keith Howard, Gary Dixon, and Mark Wallace. We also thank David Davis at the Nevada Bureau of Mines and Geology for discovering obscure reports describing drill holes in the Lake Mead area and Holly McLachlan for assistance with preparing figures portraying 3D perspectives of the wells. Reviews by Melissa Lamb, Dave Miller, Karl Karlstrom, and an anonymous individual greatly improved this manuscript.

APPENDIX A: 40Ar/39Ar DATA