USGS - science for a changing world

Geosciences and Environmental Change Science Center

Paleoclimate Variability of the American Southwest

Paleoclimate from Lake Deposits

We study lake deposits using sediment cores and outcrop samples to obtain detailed climate records for Pleistocene and Holocene conditions in the Mojave Desert, an area south of most previously established records (see figure below). These records can be combined with studies of flood histories, speleothems, and packrat middens to improve understanding of seasonal and longer-term shifts in winds and air-mass sources that govern temperature and precipitation delivery in this area. The Mojave Desert occupies a transition zone between western California, which dominantly receives westerly winter Pacific moisture, and areas to the east that dominantly receive southerly summer monsoon moisture; climate records from the Mojave may potentially contain a sensitive history of the changing influence of these two moisture sources.

Map of Great Basin showing drainage area and present and Pleistocene lakes
Pluvial lakes in the Great Basin at their maximum late Pleistocene extent. Darker blue areas are late Pleistocene pluvial lakes: Bonneville, Lahontan, Russell (Mono), Searles, Manix, Manly, Kawich, and Mojave. (from Morrison, 1991)

Previous studies of Lake Mojave (Wells and others, 2003) suggest that it reached high levels before and after more northerly lakes. Previous studies of exposed sediments of Lake Manix (see figure below), the terminus of the Mojave River before its integration downstream to form Lake Mojave, showed that a long middle to late Pleistocene paleolake sediment record is preserved (Jefferson, 2003). Recent studies of cores in the Gulf of California (Barron and others, 2004) suggest well-defined time-series changes in seasonal wind directions over the last 16,000 years that should have a direct impact on Mojave Basin air temperature and moisture regime, including seasonality of precipitation. We have improved and extended records like these into past glacial and interglacial periods using detailed studies of sediment cores and outcrops in the Lake Manix basin.

Map of Lake Manix basin showing different lake levels
Map of shorelines, Quaternary faults, and geography of the Lake Manix basin. MF is Manix fault, DLF is Dolores Lake fault, CF is Calico Fault, and "es" shows areas of active dunes and eolian sheet sands.


We describe, sample, and analyze sediments from cores and outcrops to determine:

  1. Environment of deposition
  2. Source of sediment and water
  3. Paleohydrologic status (low vs. high lake levels, fresh vs. saline water, cool vs. warm)
  4. Local and regional vegetation change
  5. Age

To accomplish this, we analyze particle size, organic and inorganic carbon, magnetic properties, microfossil content (especially ostracode species indicative of hydrologic conditions), pollen content, stable isotopes of carbon and oxygen from ostracodes and tufa, and ages determined by 14C, amino-acid racemization, and uranium-series techniques.

Stratigraphy of Manix Core

The Manix core was obtained from a site northwest of the confluence of Manix Wash and the Mojave River. Outcrops in this area formed the basis for Jefferson's (2003) reconstruction of the history of fluctuations of Lake Manix. New mapping combined with detailed study of the core sediment (see figure below) revised Jefferson's (2003) interpretations to include shorter lake cycles and somewhat different timing for parts of the section. From paleomagnetic analyses, Lake Manix endured as the terminus of the Mojave River from at least 480,000 to 25,000 years ago.

Lake Manix core Stratigraphic sketch of Manix core based on sedimentology, with positions of chronologic age control. Black bars show thicknesses of intervals with anomalous declination and inclination interpreted as magnetic excursions. PI, paleointensity maxima (magnetic data courtesy of Steve Lund, University Southern California). Intervals denoted by dashed lines exhibit slight disturbance, including small fractures and deformation, that may record nearby earthquake events. Gaps indicate missing core. (Larger view of figure)

Buried soil profiles or soil horizons within the lacustrine section can be identified in the core and outcrops, and suggest multiple periods of non-deposition on the low-gradient floor of the Cady subbasin during the Pleistocene. Radiocarbon ages of Anodonta shells sampled from outcrop in the uppermost stratigraphic unit range from 30.4 to 40.1 cal ka. The Manix tephra bed (approximately 185 ka) serves as a valuable time marker in the Cady subbasin. We have evidence for seven paleomagnetic excursions in the core sediments and estimate the bottom of the Manix core to be about 500,000 years old.

Particle size, bedding characteristics, and magnetic and ostracode studies (see figure below) reveal many fluctuations in lake level through time. One of the most important results from the core is that lacustrine episodes at Manix do not correlate solely with glacial climates as demonstrated in basins farther north and as originally hypothesized. Lakes have inundated Cady and Afton sub-basins repeatedly and consistently from oxygen isotope stage (OIS) 12 through OIS 2, including during interstadial OIS 3 and interglacials OIS 5, 7, and 9.

Chart of Lake Manix core Particle size, carbonate content, and interpreted lake level from the Manix core. Plot on left is the ratio of medium and fine sand to fine silt and clay; smaller values indicate finer sediment and deeper water. Second plot shows percent sand, clay, and carbonate. Third plot shows interpreted lake level (L numbers are lake phases). Pink bands are periods of soil formation; dashed gray lines are brief dry periods; gray band is Manix tephra. Plot on right shows independent 18O curve for marine cores and boundaries of major glacial (even numbers) and interglacial (odd) periods. (Larger view of figure)

An important event in the Lake Manix basin occurred at about the time the tephra was deposited. Before this time, the lake was confined to the area west of Buwalda Ridge. During this event, the natural dam of fan gravel holding back the lake failed abruptly and a large flood entered the Afton subbasin to the east. The ~7 m of sediment above the tephra in the core had been interpreted as perennial-lake deposits, but our analysis indicates that these sediments were deposited in a fluctuating environment of shallow lakes and mudflats, probably because most of the water-holding capacity of the basin after the flood lay to the east of the core site. Sediment patterns in the core suggest that millennial- to centennial-scale cycles of wet and dry periods may be recorded by alternating soils and fluvial, shallow-lake, and mudflat beds.

Ostracode species distribution in the Manix core

Ostracodes are small crustaceans that inhabit nearly all aquatic environments. Ostracodes secrete calcium carbonate bivalved shells that are frequently preserved in lake sediments. Ostracode species distribution is dictated by several environmental parameters including water chemistry, water permanence, and water temperature.

The Manix core contains three distinct ostracode faunal zones. Lake sediments at the bottom of the core (40-43 m) are dominated by Limnocythere bradburyi with few L. ceriotuberosa. Lacustrine sediments between approximately 30 and 38 m contain a variable mix of L. platyforma, L. ceriotuberosa, and L. robusta. Above 30 m, only L. ceriotuberosa is present.


(Not to scale; actual shells are barely visible without magnification.)

a) male Limnocythere robusta, right shell, b) male Limnocythere bradburyi, left shell, c) female Limnocythere platyforma right shell, d) male Limnocythere ceriotuberosa, left shell.

Limnocythere bradburyi is an extant ostracode species currently found in several shallow, alkaline lakes west of Mexico City, Mexico (Forester, 1985). Other than one isolated occurrence in southwestern New Mexico, near the Mexico-USA border, this ostracode is not currently known in the United States (Forester, 1985). Monthly air temperatures at Mexico City fluctuate within a few degrees of 15°C and never fall below freezing, and precipitation is greatest during the months of July through September. In contrast, Limnocythere ceriotuberosa, also an extant ostracode species, is common in shallow alkaline lakes like those in the Great Plains region of the United States. In this region evaporation exceeds precipitation, monthly air temperatures range from sub-freezing to approximately 25°C, and precipitation is greatest during the months of April through August. Limnocythere ceriotuberosa-bearing lakes and ponds receive a large portion of their water in the spring as snowmelt runoff. Both L. bradburyi and L. ceriotuberosa inhabit water bodies with the same chemical composition (e.g., Forester, 1985). Their current geographical separation is seemingly controlled by the frost line (currently approximately congruent with the Mexican-USA border) and differences in the timing of surface runoff and seasonal precipitation maxima.

Limnocythere platyforma has not been collected live, so nothing direct is known about its hydrologic implications or habitat preferences. Steinmetz (1988) concluded that the valves of L. platyforma and L. ceriotuberosa can be considered the same. The extra calcification that typifies L. platyforma valves may be related to lower total dissolved solids and low calcium concentrations (e.g., Keyser, 2005) at the time of valve calcification. The L. platyforma valve morphology is currently considered to be a freshwater variation of L. ceriotuberosa.

Limnocythere robusta is an extinct species and nothing directly is known about its hydrologic implications. Forester (1985) proposes that L. robusta may be related to, or possibly a variant of, L. bradburyi.

Some of the more intriguing aspects of the Manix ostracode stratigraphy are the low species variability and the relative lack of species fluctuations. The growth and shrinking of pluvial lakes usually results in a predictable transition through numerous ostracode species assemblages (e.g., Owens Lake; Carter, 1997; Death Valley; Lowenstein and others, 1999; Lake Bonneville; Oviatt and others, 1999; Balch and others, 2005). Large, cold, dilute lake phases that might typify a glacial climate are commonly represented in a paleontological sense by ostracodes such as Cytherissa lacustris. Smaller, more brackish lake phases that might typify a warm and dry interglacial are commonly represented by ostracodes such as Limnocythere ceriotuberosa, Limnocythere staplini, and Candona rawsoni. Some lakes may eventually be reduced to valley-bottom, groundwater-supported wetlands, which are commonly characterized by very diverse ostracode species assemblages (e.g., Oviatt and others, 1999; Balch and others, 2005). These wetlands may include ostracode species such as Candona acuminata, Cyprideis beaconensis, Limnocythere staplini, Limnocythere sappaensis, Cypridopsis vidua, Physocypria spp. and a variety of other species. The oscillating pattern of past glacial-interglacial intervals generates cyclical fluctuations in lake size and chemical composition, which then creates cyclical fluctuations in a lake's ostracode species assemblages.

In the Manix deposits there are very few fluctuations in ostracode species assemblages, even though the lake sediments were deposited over several glacial-interglacial cycles. Reasonable explanations for the lack of ostracode variability are related to the location of the Manix core and the temporal resolution of the core sediments. The core contains sediments that were deposited along the margin of the Manix basin rather than at the depocenter of the basin. Paleolakes that occupied the basin may have only inundated the core site during the medium-sized to the largest lake phases. If true, then the preserved ostracode assemblages would be consistently similar. Additionally, the ostracode samples were taken from the core at approximately 20 cm intervals, which would equate to millennial scale (1000-yr) resolution. High-resolution variability in the ostracode assemblages may have been overlooked at such a scale.

Chart showing calcite and ostracode concentrations from the Manix core plotted by age Sample-averaged δ18O and δ13C values in Limnocythere spp. calcite (panels c and d) and ostracode concentrations (panels e to j; note change in scales) from the Manix core plotted by age. Panels a and b are the SPECMAP marine and Devils Hole δ18O records and glacial-interglacial stages. Orange bars, soils in the Manix core; blue bar, Manix tephra; blue envelope in panels c and d, ± 1δ for the individual sample means; solid horizontal lines in panels c and d, average stable isotope value for the entire Manix core; dashed horizontal lines in panels c and d, ± 1δ of the core average. (Larger view of figure)

Stable isotopes from ostracodes in the Manix core

Each oxygen atom has 8 protons and between 8 and 10 neutrons in its nucleus. Oxygen atoms, therefore, have atomic masses of 16, 17, or 18, depending on the number of neutrons present. Atoms with the same number of protons but different atomic masses are called isotopes. Oxygen atoms with a mass of 16 (16O) are by far the most common isotope of oxygen, comprising about 99.8% of all oxygen atoms. Oxygen atoms with a mass of 18 (18O) are the second most common isotope of oxygen, but account for only about 0.2% of all oxygen atoms. The ratio of 16O and 18O can be measured in ostracode valves, which are made of calcite (CaCO3, where any of the three oxygen atoms can be 16O or 18O). Oxygen-isotope ratios from solids, like ostracode calcite, are compared to a standard (the PeeDee Belemnite; PDB). A calcite sample with an 18O/16O ratio that is the same as the standard is reported as "0 PDB". A calcite sample with an 18O/16O ratio that is higher or lower than the standard is reported as a positive or negative PDB value, respectively. A similar approach is used for water samples, except the standard is Vienna Standard Mean Ocean Water (VSMOW).

The ratio of oxygen atoms in ostracode calcite is derived from the water in which ostracodes make their valves. The ratio of oxygen atoms in Lake Manix was dependent on several variables, but especially on the amount of Mojave River water entering the Manix basin (rivers typically have much more negative 18O/16O ratios than lakes) versus the amount of evaporation that occurred once the Mojave River was impounded in the Manix basin. Evaporation preferentially removes the lighter 16O atoms from lakes and over time Lake Manix became more enriched in the heavier 18O atoms, leading to progressively more positive 18O/16O values. The 18O/16O ratio in ostracode calcite from Lake Manix sediments recorded these changes. Ostracode calcite samples with more positive values probably represent periods of increased evaporation, a decrease in the influx of river water to the lake, or a combination of these factors. A reduction in the amount of Mojave River water entering Manix basin might have been accomplished by a reduction in the amount of snowmelt draining from the San Bernardino Mountains, or by diversions of the Mojave River into any of the basins that are upstream of, or adjacent to, the central Manix basin (Harper and Coyote basins; e.g., Meek, 2004, 1999). Conversely, ostracode calcite samples with more negative values likely represent periods of decreased evaporation, an increase in the influx of Mojave River water to Lake Manix, or a combination of these variables.

The oxygen-isotope values from the Manix core are overall much higher (more positive) than expected (compared to values from other pluvial lakes), with only a few prominent excursions. The high values suggest that evaporation played a large part in the hydrologic balance of the various Lake Manix lake cycles. Lakes that filled the Manix basin were relatively shallow, but with large surface areas. In these kinds of settings, evaporation would have been a dominant factor in the oxygen-isotope budget of the lake(s). Complicating the oxygen- isotope interpretations is the fact that the Mojave River may have flowed into other basins upstream of, or adjacent to, the core site (Harper basin and Coyote subbasin; e.g., Meek, 1999), where the waters would have ponded and evaporated prior to reaching the Manix basin. Additionally, the highest and largest stands of Lake Manix spilled over into Coyote subbasin (e.g., Meek, 2004). The incorporation of Coyote subbasin would have substantially increased Lake Manix's surface area and increased its evaporation potential. As a result, the largest lakes in the history of Lake Manix may also be represented in the core by unexpectedly high oxygen-isotope values.

The average Lake Manix ostracode oxygen-isotope value is about 0 ± 1 PDB, which, when adjusted for vital effects (-1 PDB) and calcification temperatures (6 to 15°C), indicate that average lake-water oxygen isotope values would have been about -2 ± 2 parts per mil VSMOW. The most negative ostracode oxygen-isotope values (e.g., 29 m depth; -4.6 parts per mil PDB) equate to a lake-water value of about -6 VSMOW. These values seem high considering that the Mojave River, which would have been fed primarily by snowmelt from the San Bernardino Mountains, likely would have had much lower oxygen-isotope values (< -15 VSMOW). The largest negative deviations from ~ 0 PDB likely reflect intervals were the influx of the Mojave River was large, more persistent, or more directly flowing to near the core site (e.g., Lake Cycle 5). The higher oxygen-isotope values (0 to +2 PDB) tend to occur at the beginning or the end of lake cycles and likely represent stable lake conditions where evaporation exceeded river inflow, or perhaps intervals where the Mojave River was no longer flowing directly into the Manix basin and was flowing into one of the upstream basins and evaporating prior to reaching Lake Manix. The highest oxygen-isotope values (17-18 m; ~ +5 PDB) occur after deposition of the Manix tephra and the catastrophic failure at Buwalda Ridge (Lake Cycle 6). Once this former lake threshold failed, the depocenter of Lake Manix shifted northeastward, and the core site would no longer have been submerged except by lakes that were large enough to back-fill into the Cady subbasin from the new threshold in Afton basin (e.g., possibly indicated by the increase in L. ceriotuberosa at ~ 11 mbs). The core site was probably not much more than a broad mudflat, or perhaps submerged under very shallow lakes, after the failure of the Buwalda Ridge threshold.


Balch, D.P., Cohen, A.S., Schnurrenberger, D.W., Haskell, B.J., Valero Garces, B.L., Beck, W.J., Cheng, H., and Edwards, L., 2005, Ecosystem and paleohydrological response to Quaternary climate change in the Bonneville Basin, Utah: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 221, p. 99-122.

Barron, J.A., Bukry, David, and Bischoff, J.L., 2004, High resolution paleoceanography of the Guaymas Basin, Gulf of California, during the past 15,000 years: Marine Micropaleontology, v. 50, p. 185-207.

Carter, Claire, 1997, Ostracodes in Owens Lake core OL-92: alternation of saline and freshwater forms through time, in Smith, G.I., and Bischoff, J.L., eds., An 800,000-year paleoclimatic record from core OL-92, Owens Lake, Southeast California: Geological Society of America Special Paper 317, p. 113-119.

Forester, R.M., 1985, Limnocythere bradburyi n. sp.: A modern ostracode from central Mexico and a possible Quaternary paleoclimatic indicator: Journal of Paleontology, v. 59, p. 8-20.

Jefferson, G.T., 2003, Stratigraphy and paleontology of the middle to late Pleistocene Manix Formation, and paleoenvironments of the central Mojave River, southern California, in Enzel, Y., Wells, S.G., and Lancaster, N., eds., Paleoenvironments and paleohydrology of the Mojave and southern Great Basin deserts: Geological Society of America Special Paper 368, p. 43-60.

Keyser, Dietmar, 2005, Histological peculiarities of the noding process in Cyprideis torosa (Jones) (Crustacea, Ostracoda): Hydrobiologia, v. . 538, p. 95-106.

Lowenstein, T.K., Brown, Christopher, Roberts, S.M., Ku, Teh-Lung, Luo, Shangde, Yang, Wenbo, 1999, 200 k.y. paleoclimate from Death Valley salt core: Geology, v. 27, p. 3-6.

Meek, Norman, 2004, Mojave River history from an upstream perspective, in Reynolds, R.E., ed., Breaking up—the 2004 Desert Symposium field trip with abstracts from the 2004 Desert Symposium [PDF]: California State University, Desert Studies Consortium, and LSA Associates, Inc., p. 68.

Meek, Norman, 1999, New Discoveries about the Late Wisconsinan history of the Mojave River System: San Bernardino County Museum Association Quarterly, v. 46, p. 113-117.

Morrison, R.B., 1991, Quaternary stratigraphic, hydrologic, and climatic history of the Great Basin, with emphasis on Lakes Lahontan, Bonneville, and Tecopa, in Morrison, R.B., ed., Quaternary Nonglacial Geology: Conterminous U.S: Geological Society of America, p. 283-320.

Oviatt, C.G., Thompson, R.S., Kaufman, D.S., Bright, Jordan, and Forester, R.M., 1999, Reinterpretation of the Burmester core, Bonneville Basin, Utah: Quaternary Research, v. 52, p. 180-184.

Steinmetz, J.J., 1988, Biostratigraphy and paleoecology of limnic ostracodes from the Late Pleistocene Manix formation [MS thesis]: Pomona, California State Polytechnic University, 33 p.

Wells, S.G., Brown, W.J., Enzel, Yehouda, Anderson, R.Y., and McFadden, L.D., 2003, Late Quaternary geology and paleohydrology of pluvial Lake Mojave, southern California, in Enzel, Yehouda, Wells, S.G., and Lancaster, Nicholas, eds., Paleoenvironments and paleohydrology of the Mojave and southern Great Basin deserts: Geological Society of America Special Paper 368, p. 79-114.

Accessibility FOIA Privacy Policies and Notices

Take Pride in America logo logo U.S. Department of the Interior | U.S. Geological Survey
Page Contact Information: Website Manager
Page Last Modified: Thursday March 26 2015