Synoptic reconstruction of a major ancient lake system: Eocene Green River Formation, western United States M. Elliot Smith* Alan R. Carroll Brad S. Singer Department of Geology and Geophysics, University of Wisconsin, 1215 West Dayton Street, Madison, Wisconsin 53706, USA ABSTRACT Members. Sediment accumulation patterns than being confi ned to a single episode of arid thus refl ect basin-center–focused accumula- climate. Evaporative terminal sinks were Numerous 40Ar/39Ar experiments on sani- tion rates when the basin was underfi lled, initially located in the Greater Green River dine and biotite from 22 ash beds and 3 and supply-limited accumulation when the and Piceance Creek Basins (51.3–48.9 Ma), volcaniclastic sand beds from the Greater basin was balanced fi lled to overfi lled. Sedi- then gradually migrated southward to the Green River, Piceance Creek, and Uinta ment accumulation in the Uinta Basin, at Uinta Basin (47.1–45.2 Ma). This history is Basins of Wyoming, Colorado, and Utah Indian Canyon, Utah, was relatively con- likely related to progressive southward con- constrain ~8 m.y. of the Eocene Epoch. Mul- stant at ~150 mm/k.y. during deposition of struction of the Absaroka Volcanic Prov- tiple analyses were conducted per sample over 5 m.y. of both evaporative and fl uctuat- ince, which constituted a major topographic using laser fusion and incremental heating ing profundal facies, which likely refl ects the and thermal anomaly that contributed to a techniques to differentiate inheritance, 40Ar basin-margin position of the measured sec- regional north to south hydrologic gradient. loss, and 39Ar recoil. When considered in tion. The most rapid sediment accumulation The Greater Green River and Piceance Creek conjunction with existing radioisotopic ages for the entire system (>1 m/k.y.) occurred Basins were eventually fi lled from north and lithostratigraphy, biostratigraphy, and between 49.0 and 47.5 Ma, when volcani- to south with Absaroka-derived detritus at magnetostratigraphy, these new age deter- clastic materials from the Absaroka and/or sedimentation rates 1–2 orders of magnitude minations facilitate temporal correlation of Challis volcanic fi elds entered the Green greater than the underlying lake deposits. linked Eocene lake basins in the Laramide River Formation lakes from the north. Rocky Mountain region at a signifi cantly Our new ages combined with existing Keywords: Ar-Ar, Absaroka, Uinta Basin, increased level of precision. To compare our paleomagnetic and biostratigraphic control Piceance Creek Basin, land-mammal ages, lake results to the geomagnetic polarity time scale permit the fi rst detailed synoptic comparison type, Laramide, Eocene, Green River Formation and the regional volcanic record, the ages of lacustrine depositional environments in of Eocene magnetic anomalies C24 through all the Green River Formation basins. Cou- INTRODUCTION C20 were recalibrated using seven 40Ar/39Ar pled with previously published paleocurrent ages. Overall, the ages obtained for this study observations, our detailed correlations show Large lakes are widely recognized for their are consistent with the isochroneity of North that relatively freshwater lakes commonly importance as economic resources and as American land-mammal ages throughout the drained into more saline downstream lakes. archives of faunal, fl oral, and climatic evolution study area, and provide precise radioisotopic The overall character of Eocene lake depos- (e.g., Bradley, 1929; Franczyk et al., 1989; Wilf, constraints on several important biostrati- its was therefore governed in part by the 2000) but are less well understood with regard to graphic boundaries. geomorphic evolution of drainage patterns the geomorphic evolution of the landscapes sur- Applying these new ages, average sedi- in the surrounding Laramide landscape. rounding them (cf. Surdam and Stanley, 1980; ment accumulation rates in the Greater Freshwater (overfi lled) lakes were initially Pietras et al., 2003a; Carroll et al., 2006). The Green River Basin, Wyoming, were approx- dominant (53.5–52.0 Ma), possibly related Eocene Green River Formation (Hayden, 1869) imately three times faster at the center of the to high erosion rates of remnant Cretaceous of Wyoming, Colorado, and Utah represents one basin versus its ramp-like northern margin strata on adjacent uplifts. Expansion of bal- of the best-documented ancient lake systems and during deposition of the underfi lled Wilkins anced-fi ll lakes fi rst occurred in all Green has long been a type example for understanding Peak Member. In contrast, sediment accu- River Formation basins at 52.0–51.3 Ma and lacustrine depositional systems (Bradley, 1929; mulation occurred faster at the edge of again between 49.6 and 48.5 Ma. Evapora- Eugster and Surdam, 1973; Carroll and Bohacs, the basin during deposition of the bal- tive (underfi lled) lakes occurred in various 1999). Since Marsh (1871) speculated that the anced fi lled to overfi lled Tipton and Laney basins between 51.3 and 45.1 Ma, coincident Green River Formation lakes were hydrologi- with the end of the early Eocene climatic cally connected, numerous authors have pro- optima and subsequent onset of global cool- posed temporal correlations of its strata across *Present address: Department of Geology, So- noma State University, 1801 East Cotati Avenue, ing defi ned from marine record. However, the Uinta uplift (e.g., Bradley, 1931; Roehler, Rohnert Park, California 94928, USA, e-mail: evaporite intervals in the different depocen- 1974; Surdam and Stanley, 1980). However, [email protected] ters were deposited at different times rather due to the absence of intervening strata between GSA Bulletin; January/February 2008; v. 120; no. 1/2; p. 54–84; doi: 10.1130/B26073.1; 14 fi gures; 5 tables; Data Repository Item 2007211. 54 For permission to copy, contact [email protected] © 2007 Geological Society of America Synoptic reconstruction of the Eocene Green River Formation Figure 1. Map showing the locations of Eocene basins and basin-bounding uplifts. Compiled from Ross et al. (1955), Grose (1972), Witkind and Grose (1972), Bond and Wood (1978), Stewart and Carlson (1978), Tweto (1979), Love and Christiansen (1985), Constenius (1996), and Mitchell (1998). Eocene stratal thicknesses are from Robinson (1972). Geological Society of America Bulletin, January/February 2008 55 Smith et al. depocenters (Fig. 1), lateral facies changes, and which tend to be more susceptible to chemi- series of basement uplifts of central Colorado limited radioisotopic age control, temporal cor- cal weathering than sanidine (Renne, 2000; and southeastern Wyoming (Fig. 1). These relation has been insuffi ciently precise to address Smith et al., 2006). Age determinations were basins typically contain 1–2 km packages of more specifi c questions concerning basin evo- further limited in accuracy by the unavoid- coarse-grained alluvial strata (Dickinson et al., lution. This paper provides an 40Ar/39Ar–based able inclusion of inherited or altered biotite 1988). Extensional basins are strike-elongate age framework that allows for the fi rst detailed grains in the large samples analyzed (cf. Smith grabens and half grabens that overlie the former delineation on upstream-downstream relation- et al., 2006), and limited in their precision by Cordilleran fold and thrust belt, are commonly ships between the sequences of lakes that occu- lower resolution mass spectrometry. Recently, bounded by normal faults that reactivate Cordil- pied the Green River Formation basins and pro- 40Ar/39Ar geochronologic studies of tuff beds leran thrust faults, and often contain thick (2– vides a fundamental measurement of lacustrine using smaller samples and more sensitive mass 5 km) but areally restricted packages of alluvial sediment accumulation rates over ~8 m.y. of the spectrometry have begun to signifi cantly refi ne and lacustrine strata (Constenius, 1996). Eocene Epoch. the timing of the Green River Formation and Volcanism occurred over broad areas of the Until recently, mammalian biostratigraphy related alluvial strata (Wing et al., 1991; Smith northwestern United States during the Eocene was the only available method for determin- et al., 2003, 2004, 2006). This study integrates and provided both fallout tuffs and volcani- ing the relative age of strata in the terrestrial 25 new age determinations with detailed facies clastic sediment to the Green River Formation basins that contain the Green River Formation and geochemical analyses to construct the lake basins (Fig. 1; Surdam and Stanley, 1980; (Wood et al., 1941; Lillegraven, 1993; Robinson most comprehensive and highly resolved chro- Fritz and Harrison, 1985; Armstrong and Ward, et al., 2004). However, mammalian fossils are nostratigraphic model available for any major 1991). Major volcanic centers include the Absa- typically preserved in adjacent alluvial depos- pre-Quaternary lake system. roka Volcanic Province, Challis volcanic fi eld, its that can be diffi cult to correlate with lake and Lowland Creek Volcanics; minor fi elds are deposits using the physical characteristics of GEOLOGIC SETTING scattered throughout the region (Fig. 1). The the strata (cf. Clyde et al., 2004; Smith et al., stratigraphic and time-stratigraphic constraints 2004). The temporal resolution of mammalian The Green River Formation of Wyoming, for Eocene volcanic fi elds in Wyoming, Mon- biostratigraphy is also fundamentally limited Colorado, and Utah was deposited in a series of tana, and Idaho are summarized in Figure 5. by