Journal of Geochemical Exploration: Krishangi D. Groover, John A. Izbicki
Journal of Geochemical Exploration: Krishangi D. Groover, John A. Izbicki
U.S. Geological Survey, California Water Science Center, 4165 Spruance Road, Suite 200, San Diego, CA 92101, United States of America
⁎
Corresponding author.
E-mail address: kgroover@usgs.gov (K.D. Groover).
https://doi.org/10.1016/j.gexplo.2018.09.005
Received 4 February 2017; Received in revised form 25 August 2018; Accepted 3 September 2018
0375-6742/ Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).
Please cite this article as: Groover, K.D., Journal of Geochemical Exploration, https://doi.org/10.1016/j.gexplo.2018.09.005
K.D. Groover, J.A. Izbicki Journal of Geochemical Exploration xxx (xxxx) xxx–xxx
sample density of one sample per 1600 square kilometers (km2). Grid- 2016).
based sampling approaches used in continental-scale surveys are ef- The most productive aquifer within the Mojave River Groundwater
fective at determining the chemical composition of a large area, but Basin is the floodplain aquifer, which extends along the length of the
may neglect local-scale variations in chemistry that could be captured Mojave River (California Department of Water Resources, 2003; Stamos
by other, more detailed sampling approaches. et al., 2001). The floodplain aquifer is commonly < 1 to 2 km wide and
The study area in the western part of the Mojave Desert, California, typically < 70 m thick (Stamos et al., 2001), and is surrounded and
between 100 and 200 kilometers (km) northeast of Los Angeles (Fig. 1), underlain by a regional aquifer composed of older Mojave River de-
is composed of multiple closed basins containing alluvium eroded from posits (Pliocene to Pleistocene age) and alluvium eroded from local
surrounding mountains. Weathering of minerals within alluvium and mountains. Older Mojave River alluvium covers a large area within the
rock from these areas may release a range of elements into groundwater Mojave River Groundwater Basin, ranging from Deep Creek in the
under the proper geochemical conditions (Hem, 1992). Except for some southwest and extending nearly to Barstow in the central part of the
streams at high elevations on the edge of the study area, there are no study area (Fig. 1B; Stamos et al., 2001). For the purposes of this report,
perennial streams, and groundwater is the only dependable source of the older Mojave River deposits also include the Victorville fan (Fig. 1B;
water supply. Meisling and Weldon, 1989), which was deposited by streams that were
Groundwater quality in alluvial aquifers underlying the western tributary to the Mojave River before the fan was beheaded by erosion
Mojave Desert has been most recently described by Dawson and Belitz along the San Andreas Fault. The Morongo Groundwater Basin, and
(2012) and Metzger et al. (2015), and some recent work has linked parts of the study area not included in the Mojave River Groundwater
naturally occurring trace elements in groundwater, such as chromium Basin, are composed of numerous small alluvial basins separated by
and vanadium, to weathering of minerals eroded and transported from faults and bedrock outcrops (California Department of Water Resources,
rocks in upland drainages (Izbicki et al., 2008; Wright and Belitz, 2003).
2010). Some high concentrations of trace elements dissolved in Bedrock is dominated by felsic plutonic and volcanic rocks, with
groundwater in exceedance of California maximum contaminant levels lesser amounts of mafic plutonic and volcanic, metamorphic, and se-
(MCL) may be naturally occurring and attributable to local geologic dimentary rocks that are commonly overlain by alluvial and playa de-
sources (Dawson and Belitz, 2012; Izbicki et al., 2008; Metzger et al., posits (Fig. 1B; Dibblee, 1967; Jennings et al., 2010). Rocks have been
2015). Given the widespread occurrence of trace elements in ground- locally altered and mineralized by hydrothermal activity, and a number
water from alluvial aquifers in the area, geochemical-survey data tar- of active and historic mining operations have targeted gold, silver, and
geting alluvium within stream channels eroded from specific geologic boron resources (U.S. Geological Survey, 2005a, 2005b). Hydro-
source terrains may help identify natural geologic sources of selected thermally altered rocks have not been formally mapped at the regional
trace elements. scale, and are not shown on Fig. 1B.
The purpose of this paper is to evaluate concentrations of selected The San Bernardino Mountains, composed largely of granitic rock,
trace elements from natural geologic sources in alluvium and rocks in form the southern boundary of the study area and contain high con-
the western Mojave Desert, and assess the potential mobility of selected centrations of uranium (U.S. Geological Survey, 2004). The San Gabriel
trace elements such as arsenic, chromium, and uranium to ground- Mountains southwest of Victorville are composed of a wider variety of
water. The scope of this study included field collection and elemental rock types including the Pelona Schist (Ehlig, 1958, 1968), which is a
analyses of samples from alluvium in stream channels and from rock in gray mica-schist (Ehlig, 1958; Evans, 1982) containing a minor (5% of
upland source areas that compose specific geologic source terrains. The the total mapped formation) mafic greenschist facies (Evans, 1982).
study also included comparison of elemental analyses from various The greenschist facies contain actinolite and mica with high con-
sample matrixes, made using different techniques, to determine how centrations of chromium, iron, and nickel (Izbicki et al., 2008). Sedi-
available these elements may be to groundwater under changing pH, ment eroded from the Pelona Schist composes the Sheep Creek alluvial
redox, and ionic strength. fan that extends northward from the mountains towards El Mirage (dry)
This paper is part of a larger study of trace element concentrations Lake (Fig. 1) (Izbicki et al., 2008; Meisling and Weldon, 1989; Miller
in rock, alluvium, and groundwater in the western Mojave Desert. and Bedford, 2000; Stamos et al., 2001).
Ongoing work addresses the aqueous geochemistry of selected naturally Movement along the right-lateral San Andreas Fault, near the
occurring trace elements dissolved in groundwater in the area. southwestern margin of the study area (Fig. 1B), has transported the
mafic Pelona Schist to the northwest over the last one to five million
2. Description of study area years to its present position in the San Gabriel Mountains near the
headwaters of Sheep Creek (Cox et al., 2003; Meisling and Weldon,
The study area includes the Mojave River and Morongo 1989). This movement has changed geologic source areas contributing
Groundwater Basins (California Department of Water Resources, 2003), alluvium to the Mojave River and the Victorville fan over time (Cox
and smaller groundwater basins within Fort Irwin National Training et al., 2003; Izbicki et al., 2008). Consequently, older Mojave River
Center to the northeast (Fig. 1A). Elevations within the study area range deposits contain a greater fraction of material eroded from the Pelona
between 450 and 2600 m, and precipitation ranges from 110 milli- Schist than present-day Mojave River alluvium (Cox et al., 2003;
meters per year (mm/yr) near Barstow in the central part of the study Meisling and Weldon, 1989). Multiple smaller northwest striking faults
area, up to 1000 mm/yr, in the higher altitudes of the San Bernardino also are active within the study area (Dibblee, 1967; Glazner et al.,
and San Gabriel Mountains to the southwest (Fig. 1A; Izbicki et al., 2002; Miller and Yount, 2002), and movement along these faults may
2008). have resulted in similar, smaller-scale changes in local alluvial source
The area is internally drained, with several large ephemeral streams terrains.
that flow from south to north. The Mojave River, the largest of these Groundwater in alluvial aquifers within the study area is generally
streams, extends nearly 200 km from the San Bernardino and San oxic, with neutral to alkaline pH (Dawson and Belitz, 2012; Metzger
Gabriel mountains south of Victorville to its terminus east of Afton et al., 2015). Except for groundwater in the floodplain aquifer re-
Canyon (Fig. 1A; Thompson, 1929). The Mojave River is typically dry charged by occasional large flows along the Mojave River, most
along most of its reach, except where groundwater discharge maintains groundwater in the study area was recharged many thousands of years
perennial streamflow near the Upper and Lower Narrows at Victorville, ago (Izbicki and Michel, 2004). Groundwater ages in the regional
and near Afton Canyon. Large flows, extending the entire length of the aquifer increase to > 20,000 years away from the mountain front and
Mojave River from the mountain front to Afton Canyon, occur on away from the Mojave River (Ball and Izbicki, 2004; Izbicki and Michel,
average every 5 to 7 years (Lines, 1996; Stamos et al., 2001; Seymour, 2004). The chemistry and age of groundwater combine to create
2
K.D. Groover, J.A. Izbicki Journal of Geochemical Exploration xxx (xxxx) xxx–xxx
Fig. 1. Sample locations (A), and generalized surficial geology (B) within the western Mojave Desert, southern California.
3
K.D. Groover, J.A. Izbicki Journal of Geochemical Exploration xxx (xxxx) xxx–xxx
conditions that may allow certain naturally occurring trace elements analyzer (Olympus, Waltham, MA), equipped with a 4-watt Ta/Au X-
such as arsenic, chromium, uranium, and vanadium in alluvium to ray tube operated between 15 and 40 keV in “Soil” mode. Each mea-
enter groundwater at concentrations that may be a public health con- surement was done for 2 min across an oval area approximately 5 mm
cern (Ball and Izbicki, 2004; Dawson and Belitz, 2012; Hem, 1992; across its larger diameter, and was generally repeated 3 times per
Izbicki and Michel, 2004; Metzger et al., 2015). sample to assess variability in composition. Measurements were opti-
mized for high concentrations of chromium (based on instrument per-
formance with standard reference materials) by adjusting instrument
3. Methods settings when samples with high chromium concentrations were en-
countered.
A total of 161 samples of alluvium, 36 samples of rock, and 20 Two NIST (National Institute of Standards and Technology) soil
subsamples of alluvium from 8 wells (drilled between 2011 and 2013) standards (2710a and 2711a), a USGS rock standard (BHVO-2), and a
were collected between January 2014 and June 2016 (Supplementary silicon dioxide blank (99.99% purity) were measured daily using por-
Data A). Most alluvial samples were collected from small stream table X-ray fluorescence (pXRF) before data collection to monitor the
channels draining distinct geologic source terrains. Detailed, closely- accuracy of the measurements (Supplementary Data B). Data quality
spaced sampling along the length of the Mojave River and other large associated with measurements, chromium-optimized measurements,
streams was done to evaluate changes in alluvium chemistry with and non-optimized measurements are presented in Table 1. Most
changes in bedrock along the stream length. Most alluvium samples measurements were within ± 10% of the NIST- and USGS-certified
consisted of silt, sand, and granules to small pebbles, and generally did standard values. Blank data showed contamination at low levels for
not contain material that had been extensively altered by soil-forming titanium (12–20 mg/kg) and iron (9–10 mg/kg), representing 0.74 and
processes (dissolution and precipitation, hydrolysis, etc.). Rock samples 0.05% of the average concentrations for these elements, respectively.
from selected locations provide compositional end members for dif- Laboratory XRF (LXRF) measurements for major and trace elements
ferent geologic source terrains. Samples were taken to the U.S. were completed on splits from 33 samples of alluvium and core sam-
Geological Survey (USGS) San Diego office to allow for consistent ples. For major elements, samples were powdered, homogenized, and
sample preparation and handling prior to measurement. The average fused into a glass disk at 1000 °C. For trace elements, samples were
sample density in this study is one sample per 80 km2, which is much powdered, mixed with elvacite (a plastic binder), and compressed into
denser sampling than the one sample per approximately 1600 km2 disks using 23 tons of pressure (Supplementary Data C). Samples were
within the study area available from the most recent continental-scale analyzed using a wavelength dispersive LXRF (Philips MagiX PRO) at
survey by Smith et al. (2014). San Diego State University. The LXRF is equipped with a 4000-watt X-
A total of 831 measurements were completed on alluvium and core ray tube, approximately 1000 time more powerful than the X-ray tube
material, and 112 measurements were done on samples of rocks. in the pXRF. LXRF data were compared to pXRF data to determine
Sample preparation for alluvial material consisted of manually re- accuracy of pXRF data.
moving cobbles and twigs from the sample and either air drying or oven Sequential extractions were done on splits from 40 samples of al-
drying samples at 80 °C for 4 h prior to measurement. Sample pre- luvium and core material to determine the concentrations of selected
paration for rocks consisted of rinsing a freshly exposed surface with trace elements on operationally-defined sorption sites on the surfaces of
tap water, and allowing the sample to air dry prior to measurement. mineral grains. Extractions, done using progressively stronger extrac-
Samples were analyzed for 27 elements (Supplementary Data A) using a tion solutions, were prepared at the USGS San Diego Office using a
portable energy-dispersive DP-4000 Delta Premium X-ray fluorescence
Table 1
Portable X-ray fluorescence (pXRF) measurements compared to National Institute of Standards and Technology (NIST) 2710a, NIST 2711a, and U.S. Geological
Survey (USGS) BHVO-2 standard reference materials.
All measurements Chromium-optimized measurements Non-optimized measurements
(n = 239) (n = 78) (n = 161)
Element Average percent of reference Intercept Average percent of reference Intercept Average percent of reference Intercept
concentration concentration concentration
“N.D.” = no data available; “Optimized” = measurements were optimized by adjusting instrument beam settings to measure chromium accurately.
a
Data not available from chromium-optimized measurements.
b
Standard reference concentrations not available for all standards. Bold font indicates accuracy of ± 10% of standard reference value or better; italics indicate
statistically significant differences between optimized and non-optimized measurements on the basis of a Wilcoxon ranked sum test with a confidence criterion of
α = 0.05.
4
K.D. Groover, J.A. Izbicki Journal of Geochemical Exploration xxx (xxxx) xxx–xxx
5
K.D. Groover, J.A. Izbicki Journal of Geochemical Exploration xxx (xxxx) xxx–xxx
Fig. 2. Comparison between elemental concentrations in alluvium measured using portable (pXRF) and laboratory (LXRF) X-ray fluorescence, western Mojave
Desert, southern California.
350 mg/kg in rocks from Joshua Tree and Iron Mountain, respectively. average concentration in bulk continental crust.
Nickel, typically associated with chromium-bearing minerals in Uranium concentrations in alluvium ranged from < 2.8 mg/kg to
mafic rocks (Garver et al., 1996; Izbicki et al., 2008; Morrison et al., 13 mg/kg (Table 4), with only 12% of samples containing detectable
2015; Reimann and Caritat, 1998), had a similar distribution to chro- uranium (Fig. 5). The 2.8 mg/kg detection limit for uranium is greater
mium in alluvium and rock samples. The highest nickel concentrations than the average concentration in bulk continental crust of 0.91 mg/kg
(in exceedance of the average concentration in bulk continental crust of (Taylor and McLennan, 1995) and U.S. soils of 2.1 mg/kg (Smith et al.,
105 mg/kg; Taylor and McLennan, 1995) were in playa/lake deposits, 2014). The highest concentration of uranium in alluvium was measured
and in alluvium and rock eroded from mafic terrains and Pelona Schist in a fine-grained, clay-rich sample collected from older Mojave River
(Supplementary Data A). Only 4% of alluvium measurements and 23% deposits west of Iron Mountain (13 mg/kg), with other high con-
of rock measurements had concentrations of nickel in exceedance of the centrations measured in alluvium eroded from unaltered and
6
K.D. Groover, J.A. Izbicki Journal of Geochemical Exploration xxx (xxxx) xxx–xxx
Arsenic (As) N.D. N.D. N.D. N.D. 4.2.2. Principal component analyses
Calcium (Ca) 1.41 0.90 1.05 −7950 Principal component analysis (PCA) (Hotelling, 1933; Wold et al.,
Chromium (Cr) 1.04 0.84 5.10 1.4 1987) was used to understand the spatial variability and differences in
Copper (Cu) 0.79 0.94 0.85 2.4 elemental concentration data of alluvium eroded from different geo-
Iron (Fe) 1.71 0.90 1.43 380
logic source terrains. PCA is a multivariate statistical technique used to
Lead (Pb) 0.66 0.31 3.50 2.8
Manganese (Mn) 0.91 0.91 0.01 1.9 transform potentially correlated, and potentially non-linear data (in this
Molybdenum (Mo) 0.73 0.88 0.83 1.7 case, elemental concentrations of alluvium measured using pXRF) into
Nickel (Ni) 0.78 0.89 1.77 20 new variables called principal components that are uncorrelated, linear
Potassium (K) 0.94 0.52 0.56 11,600
combinations of the original data. The values of the principal compo-
Rubidium (Rb) 0.94 0.92 4.28 10
Strontium (Sr) 0.98 0.95 18.41 51
nents are known as scores. Eigenvectors calculated from the PCA that
Tin (Sn) N.D. N.D. N.D. N.D. describe the direction and magnitude of elemental contributions to the
Titanium (Ti) 0.95 0.73 0.14 −23 PCA scores were used to evaluate elemental co-occurrence in alluvium
Uranium (U) 1.00 0.95 0.52 2.1 eroded from different geologic terrains. PCA results are sensitive to the
Vanadium (V) 1.05 0.66 13.36 300
scaling and magnitude of the data. As a consequence, data from rocks,
Zinc (Zn) 0.79 0.98 2.02 0.85
Zirconium (Zr) 1.32 0.91 12.53 11 which commonly have higher elemental concentrations than alluvium,
were not used to calculate PCA scores. Instead, PCA scores for rocks
N.D. = no data, element not measured using laboratory X-ray fluorescence (Srn) were calculated from eigenvectors estimated from alluvial con-
(LXRF) or no detections were made during measurements; number of sam- centration data according to the following equation:
ples = 33; LXRF data were the independent variable and portable X-ray fluor-
escence (pXRF) data were the dependent variable in the calculations. Bold font ¯
indicates elements with slopes within ± 10% of 1 between the two methods.
Xr Xa
Srn = E n
a
(1)
hydrothermally altered felsic volcanic rocks east of Barstow (7 mg/kg)
and northwest of Barstow (6 mg/kg), respectively. Uranium was com- where a = alluvium, r = rock; γ = is the individual element As, Ca,
monly detected in alluvium within Pipes Wash that drains the uranium- Cr…; Xrγ = concentration of element γ for a given rock measurement,
rich granitic rock in the San Bernardino Mountains (U.S. Geological ¯
Xa = mean concentration of element γ in alluvium measurements, and
Survey, 2004), but was less commonly detected in recent Mojave River
σ = standard deviation from the mean for element γ in alluvium;
deposits despite the abundance of granitic fragments eroded from ur-
E = nth eigenvector for element γ where n = 1, 2, 3, …
anium-rich rocks in the San Bernardino Mountains (U.S. Geological
The first and second principal components captured 45% of the total
Survey, 2004). Uranium was detected more frequently and at higher
variability in the pXRF data; the addition of the third principal com-
concentrations in rocks. Concentrations in rocks were as high as
ponent explained 57% of the total variability (Table 5).
180 mg/kg in hydrothermally altered felsic volcanic rock northwest of
PCA results, grouped by geologic terrain, were plotted with the first
Table 4
Arsenic, chromium, and uranium concentrations measured using portable X-ray fluorescence (pXRF) in generalized geologic materials, western Mojave Desert,
southern California.
Geologic material Arsenic (mg/kg) Chromium (mg/kg) Uranium (mg/kg) Number of samples
Alluvium
Felsic 3.0 11 18 54 < 2.8 6.2 10% 22
Felsic (volcanic) 11 24 32 68 < 2.8 7.0 28% 8
Hydrothermally altered volcanics 30 125 22 43 < 2.8 6.0 24% 6
Mojave River alluvium 2.2 36 7.6 52 < 2.8 3.4 4% 16
Older Mojave River alluvium 3.2 49 27 87 < 2.8 13 14% 13
Mafic 4.6 31 54 300 < 2.8 3.5 1% 15
Mafic (volcanic) 6.4 13 60 135 < 2.8 3.8 7% 7
Mafic (Pelona Schist) 3.8 5.8 150 300 < 2.8 5.0 30% 6
Sedimentary 4.3 15 24 78 < 2.8 4.3 9% 50
Playa/lake deposits 8.9 16 58 490 < 2.8 < 2.8 0% 8
Carbonate 5.8 10 25 133 < 2.8 6.7 10% 5
Rock
Felsic 5.6 13 20 150 < 2.8 12 37% 7
Felsic (volcanic) 8.6 30 34 45 < 2.8 7.5 31% 4
Hydrothermally altered volcanics 45 1400 27 350 < 2.8 180 14% 3
Mafic 4.5 11 64 420 < 2.8 < 2.8 0% 6
Mafic (volcanic) 7.6 20 154 400 < 2.8 8.8 18% 10
Mafic (Pelona Schist) 5.1 55 5000 8600 < 2.8 18 12% 3
Carbonate 2.5 4.5 8.1 16 < 2.8 3.1 11% 3
Reporting level 1.4 3.3 2.8 Total: 192a
Median measured concentration of alluvium 4.3 26 < 2.8
Average bulk continental crust 1.0 185 0.91
Concentrations in milligrams per kilogram (mg/kg). Bulk continental crust concentrations cited from Taylor and McLennan, 1995.
a
Total sample count does not include 25 samples that were measured after being sieved or powdered. Raw data are available in Supplementary Data A.
7
K.D. Groover, J.A. Izbicki Journal of Geochemical Exploration xxx (xxxx) xxx–xxx
8
K.D. Groover, J.A. Izbicki Journal of Geochemical Exploration xxx (xxxx) xxx–xxx
9
K.D. Groover, J.A. Izbicki Journal of Geochemical Exploration xxx (xxxx) xxx–xxx
Table 5
Mean concentration, standard deviation, and the first, second, and third ei-
genvectors from principal component analysis (PCA) of selected portable X-ray
fluorescence (pXRF) measurements of alluvium, western Mojave Desert,
southern California.
Element Mean Standard First Second Third
deviation eigenvector eigenvector eigenvector
samples collected from Pipes Wash (Fig. 5; not shown in Fig. 7). The
frequent detection of uranium in Pipes Wash alluvium was likely due to
higher concentrations of uranium in source rocks from the San Ber-
nardino Mountains (U.S. Geological Survey, 2004).
Samples also were collected from several small stream channels that
drain felsic quartz monzonite bedrock with small outcrops of mafic rock
near Joshua Tree. Chromium concentrations within the channels in-
creased locally near mafic outcrops and then decreased farther down-
stream (not shown on Fig. 7).
Changes in the elemental composition of alluvium in Pipes Wash
and smaller washes near Joshua Tree is likely a consequence of the
relatively low volume of alluvium transported by these smaller systems,
and the greater sensitivity of those small systems to changes in bedrock
geology. This contrasts with the large volume of alluvium transported
by the Mojave River that dilutes additions of locally-derived alluvium
from mafic rock in Iron Mountain (Fig. 7A). However, regional changes
in geologic source terrain east of Barstow were large enough to affect
the elemental composition of alluvium in downstream reaches of the
Mojave River.
10
K.D. Groover, J.A. Izbicki Journal of Geochemical Exploration xxx (xxxx) xxx–xxx
Fig. 6. First and second principal component scores from elemental concentrations in rock (calculated) and alluvium grouped by rock and alluvium eroded from
felsic, mafic and hydrothermal terrains (A); rock and alluvium eroded from carbonate, sedimentary, and alluvium in playa/lake terrains (B); and for alluvium from
the Mojave River, older Mojave River, and Pelona Schist (C), western Mojave Desert, southern California.
decreasing contributions of alluvium eroded from mafic rocks in the they were associated with crystalline aluminum, iron, and manganese
San Gabriel Mountains and increasing contributions of alluvium from oxide coatings and other minor phases including secondary sulfides.
felsic rocks in the San Bernardino Mountains. This occurred over the Concentrations of extracted elements were lower in the 0.2 M ammo-
past one million years as a result of movement of geologic-source ter- nium oxalate and 0.2 M ammonium oxalate plus 0.1 M ascorbic acid
rains contributing alluvium to the Mojave River, including the mafic extractions modified from Wenzel et al. (2001). Comparison of data
Pelona Schist, along the San Andreas Fault. from the two extractions suggests that the extractions proposed by
A Wilcoxon ranked sum test (with a confidence criterion of Wenzel et al. (2001) were not sufficiently rigorous to extract all the
α ≤ 0.05) was used to test differences between the chemical composi- trace elements sorbed to aluminum, iron, and manganese oxides on the
tion of recent versus older Mojave River deposits to confirm the shift in surfaces of mineral grains.
source rock type as a consequence of the movement along the San Of the trace elements measured, only arsenic was present at sig-
Andreas fault. Median concentrations of elements in the two deposits nificant concentrations in the specifically sorbed fraction (Fig. 9; Ap-
were statistically different for all elements measured using pXRF, ex- pendix Table S3). This differs from the distribution of sorbed arsenic on
cept for nickel and uranium. The lack of a statistically significant dif- soils, where most arsenic is in the amorphous phase (Wenzel et al.,
ference in uranium concentrations in alluvium may result because ur- 2001). In general, results from this study are similar to results obtained
anium is high in both the San Bernardino and San Gabriel Mountains. for alluvium elsewhere in the Mojave Desert (Izbicki et al., 2008, 2015),
Nickel concentrations in both recent and older Mojave River deposits except that previous data showed uranium concentrations nearly an
were generally low, with median concentrations of 23 and 34 mg/kg, order of magnitude greater, with about half of the uranium contained
respectively. The decrease in the ratio of chromium to nickel from 0.79 within the non-specifically sorbed, 0.05 M KCl fraction and likely as-
in older Mojave River deposits to 0.31 in recent Mojave River deposits sociated with carbonate minerals. Uranium concentrations may be
is consistent with removal of the mafic Pelona Schist–which has a high higher in those data because many of the previous samples were allu-
chromium to nickel ratio of 4.3–from the geologic source areas con- vium eroded from the uranium-rich Pipes Wash drainage.
tributing to the river. Extracted elements were present at higher concentrations in fine-
grained deposits (on the basis of Wilcoxon ranked sum tests with a
4.3. Sequential extractions confidence criterion of α ≤ 0.05) (Fig. 9). This is consistent with
greater surface area and a greater abundance of iron, manganese, and
Elements sorbed onto coatings on mineral grains are potentially aluminum oxide surface sites available for exchange on clay and silt
soluble and may enter groundwater under the proper geochemical sized material. Kim et al. (2011) observed a similar correlation between
conditions (Morrison et al., 2015). In contrast, elements contained increasing trace-element concentrations, including arsenic and chro-
within primary mineral grains must weather from the mineral before mium, and decreasing grain size. Izbicki et al. (2008) reported higher
they can enter groundwater, and may be of lesser concern to ground- chromium concentrations in extractions from samples distal from the
water consumers. Selected elements sorbed on the surfaces of mineral San Gabriel Mountains in the Sheep Creek fan. This coincided with an
grains were extracted from samples of alluvium and core material increase in iron and manganese sorption sites that they attributed to
(Appendix Table S1) using progressively stronger extraction solutions finer grain-size.
to evaluate elements associated with operationally defined surface- Comparison of extraction data to total elemental abundance mea-
sorption sites (Table 2). Sequential extraction results (Appendix Tables sured by pXRF indicates that arsenic, uranium, and manganese are
S2–S6) were compared to total elemental concentrations measured by present at higher concentrations in the extracts from material on the
pXRF to evaluate the potential availability of selected trace elements to surfaces of mineral grains than within the grains (Fig. 9). In contrast,
groundwater. Uranium concentration data measured using LXRF chromium and vanadium were present in lower concentrations in the
(Supplementary Data C) were used in place of pXRF data that were less extracts than within mineral grains. This pattern is consistent across the
than the detection limit of 2.8 mg/kg. range of size fractions from clay to coarse sand.
Arsenic, chromium, uranium, and vanadium in extraction solutions Surface coatings in finer-grained, clayey deposits contained ur-
were present at the highest concentrations in the strong-acid (4 N anium at concentrations as high as 70% of the total concentration
HNO3) extraction modified from Chao and Sanzolone (1989), indicating measured by pXRF, with almost all the uranium present in the strong-
11
K.D. Groover, J.A. Izbicki
Fig. 7. Selected trace-element concentrations in alluvium along the Mojave River (A), Sheep Creek and Fremont Wash (B), Oro Grande Wash (C), and Pipes Wash (D),
western Mojave Desert, southern California. Drainage locations shown on Fig. 1.
acid extraction fraction and associated with crystalline oxides. Some the strong acid and the specifically sorbed phases, where sorption is pH
coarse-grained samples of alluvium immediately west of Barstow con- dependent. Arsenic sorption on the specifically sorbed phase increased
tained uranium at concentrations as high as 50% of the total pXRF or with decreasing particle size.
LXRF concentration in the strong-acid extraction (Figs. 5, 9). Arsenic in Sequential extraction and pXRF data suggest that some elements
fine-grained, clayey deposits was almost equally distributed between such as manganese, uranium, and especially arsenic in fine-grained
K.D. Groover, J.A. Izbicki Journal of Geochemical Exploration xxx (xxxx) xxx–xxx
Fig. 8. Chromium, nickel, potassium, and rubidium concentrations in older Mojave River deposits plotted against approximate age, western Mojave Desert, southern
California.
sediments may be more available for release to groundwater than continental crust of 1 mg/kg. Arsenic concentrations as high as 125 and
chromium or vanadium if redox conditions change. This is attributed to 1400 mg/kg were associated with alluvium and rock, respectively,
chromium and vanadium residing largely in primary minerals. Arsenic collected from hydrothermally altered felsic volcanic terrain.
mobilization in particular may be sensitive to changes in pH of Chromium concentrations in the Mojave Desert were generally low, and
groundwater (Wenzel et al., 2001), or increases in ionic strength that only 4% of alluvium samples had chromium concentrations greater
are accompanied by increasing concentrations of ions, such as phos- than average concentrations in bulk continental crust of 185 mg/kg.
phate, which compete with arsenic for sorption sites (Welch et al., However, chromium concentrations were as high as 490 and 8600 mg/
2000). kg in playa lake deposits eroded from the Pelona Schist and rocks
Manganese sorbed to the surfaces of mineral grains in clayey de- collected from the Pelona Schist, respectively. Median chromium con-
posits, and to a lesser extent coarse-grained felsic volcanic terrain (not centrations were as low as 7.6 mg/kg in recent Mojave River alluvium.
shown in Fig. 9), could potentially oxidize insoluble trivalent chro- Uranium concentrations were generally less than the pXRF reporting
mium, Cr(III), on surface exchange sites, to highly soluble hexavalent level of 2.8 mg/kg, which was greater than the average bulk continental
chromium, Cr(VI) (Morrison et al., 2015). Once oxidized, Cr(VI) may crust abundance of 0.91 mg/kg. However, uranium concentrations as
subsequently enter groundwater under redox and pH conditions present high as 13 and 180 mg/kg were measured in clayey older Mojave River
within the study area (Dawson and Belitz, 2012; Metzger et al., 2015). alluvium and hydrothermally altered volcanic rocks, respectively. In
However, sequential extraction and pXRF data show that oxidation of general, erosion of hydrothermally altered and mineralized felsic vol-
Cr(III) to Cr(VI) under natural conditions may be limited because canic rocks yields alluvium with high arsenic concentrations, erosion of
weathering of chromium from mineral grains prior to its release to chromium-rich mafic rocks (specifically the Pelona Schist) yields allu-
groundwater is a comparatively slow process. vium with high chromium concentrations, and erosion from felsic
volcanic and plutonic source rocks yield generally uranium-rich sedi-
ments.
5. Conclusions Principal component analysis (PCA) quantified the relation between
geologic source terrain and alluvium eroded from those terrains. PCA of
Systematic comparison between pXRF and LXRF data, and with pXRF data demonstrate higher concentrations of iron, copper, chro-
NIST-traceable standard reference material, showed pXRF data were mium, and to a lesser extent titanium, manganese, and nickel (in
suitable for evaluating the occurrence and spatial distribution of trace principal component one) characterize alluvium eroded from mafic
elements of potential public-health concern in groundwater including terrain, while higher concentrations of rubidium and potassium (in
arsenic, chromium, and uranium. principal component two) characterize alluvium eroded from felsic
Arsenic concentrations were generally high in the Mojave Desert, terrain. In contrast, alluvium eroded from hydrothermally altered vol-
and 96% of samples had measurable arsenic greater than the pXRF canics was characterized by high concentrations of mafic and felsic
reporting level of 1.4 mg/kg compared to the average concentration in
13
K.D. Groover, J.A. Izbicki Journal of Geochemical Exploration xxx (xxxx) xxx–xxx
Acknowledgments
This study was funded by the Mojave Water Agency, with additional
funding from Joshua Basin Water District, the Lahontan Regional Water
Quality Control Board, and the Fort Irwin National Training Center. The
authors thank Amanda Lopez and Tony Winkel (Mojave Water Agency),
Jill Judy-Densmore, Joseph Nawikas, and Andy Morita (U.S. Geological
Survey) for their assistance in the field collecting samples; Joan
Fig. 9. Concentrations of arsenic, chromium, and other selected trace and Kimbrough and Heather Webb of the Department of Geological
major elements in sequential extractions normalized to concentrations of ele-
Sciences at San Diego State University for assistance with preparation
ments in alluvium measured using X-ray fluorescence (XRF), western Mojave
and analysis of samples using the laboratory XRF; and Russell Johnson
Desert, southern California.
(U.S. Geological Survey, retired) for his assistance in preparing the
sequential extractions. The authors also thank Kathy Thorbjarnarson
elements, and also had high concentrations of zinc, vanadium, and (San Diego State University) and David Miller (U.S. Geological Survey)
arsenic. for their advice, thoughtful comments, and assistance during this study.
Changes in alluvium chemistry occur along stream channels because
of factors such as dilution, mineral grain size, and the accretion of al- References
luvium from different bedrock source terrains along the stream length.
If a geochemically distinctive rock outcrop (e.g., a mafic-rich outcrop Ball, J.W., Izbicki, J.A., 2004. Occurrence of hexavalent chromium in ground water in the
versus a felsic outcrop) is sufficiently large relative to the volume of western Mojave Desert, California. Appl. Geochem. 19, 1123–1135.
sediment in the stream derived from upstream sources, the elemental Brady, L.L., Jobson, H.E., 1973. An experimental study of heavy-mineral segregation
under alluvial - flow conditions. U.S. Geological Survey Professional Paper 562-K.
signature of alluvium eroded from this outcrop may be distinguishable. (38 pp.).
If the outcrop is small relative to the volume of sediment in the stream California Department of Water Resources, 2003. California's groundwater. In: California
derived from upstream sources, the elemental signature of alluvium Department of Water Resources Bulletin 118, . http://www.water.ca.gov/
groundwater/bulletin118/update_2003.cfm (August 2015, 246 pp.).
may be diluted. The volume of alluvium transported by the Mojave Chaffee, M.A., 1975. Geochemical exploration techniques applicable in the search for
River is large enough to dilute elemental signatures of alluvium copper deposits. In: U.S. Geological Survey Professional Paper 907-B, (26 pp.).
14
K.D. Groover, J.A. Izbicki Journal of Geochemical Exploration xxx (xxxx) xxx–xxx
Chao, T.T., Sanzolone, R.F., 1989. Fractionation of soil selenium by sequential partial Report 00-222, . http://wrgis.wr.usgs.gov/open-file/of00-222 (March 2016).
dissolution. Soil Sci. Soc. Am. J. 53, 385–392. Miller, D.M., Yount, J.C., 2002. Late Cenozoic tectonic evolution of the north-central
Cox, B.F., Hillhouse, J.W., Owen, L.A., Enzel, Y., Wells, S.G., Lancaster, N., 2003. Pliocene Mojave Desert inferred from fault history and physiographic evolution of the Fort
and Pleistocene evolution of the Mojave River, and associated tectonic development Irwin area, California. Geol. Soc. Am. Mem. 195, 173–197. https://doi.org/10.1130/
of the Transverse Ranges and Mojave Desert, based on borehole stratigraphy studies 0-8137-1195-9.173.
and mapping of landforms and sediments near Victorville, California. In: Miller, D.M., Schmidt, K.M., Mahan, S.A., McGeehin, J.P., Owen, L.A., Baron, J.A.,
Paleoenvironments and Paleohydrology of the Mojave and Southern Great Basin Lehmkuhl, F., Löhrer, R., 2010. Holocene landscape response to seasonality of storms
Deserts: Boulder, Colorado, Geological Society of America Special Paper 368, pp. in the Mojave Desert. Quat. Int. 215 (1–2), 45–61. https://doi.org/10.1016/j.quaint.
1–42. 2009.10.001.
Dawson, B.J., Belitz, K., 2012. Status of groundwater quality in the California Desert Morrison, J.M., Goldhaber, M.B., Mills, C.T., Breit, G.N., Hooper, R.L., Holloway, J.M.,
Region, 2006–2008: California GAMA Priority Basin Project. In: U.S. Geological Diehl, S.F., Ranville, J.F., 2015. Weathering and transport of chromium and nickel
Survey Scientific Investigations Report 2012–5040, (110 pp.). from serpentinite in the Coast Range ophiolite to the Sacramento Valley, California,
Deer, W.A., Howie, R.A., Zussman, J., 1992. An introduction to the rock-forming mi- USA. Appl. Geochem. 61, 72–86. https://doi.org/10.1016/j.apgeochem.2015.05.
nerals, 2nd ed. Pearson Education Limited, Essex, England (696 pp.). 018.
Dibblee, T.W., 1967. Areal geology of the western Mojave Desert, California. In: U.S. Morton, A.C., Hallsworth, C.R., 1999. Processes controlling the composition of heavy
Geological Survey Professional Paper 522, (153 pp.). mineral assemblages in sandstones. Sediment. Geol. 124, 3–29.
Ehlig, P.L., 1958. The Geology of the Mount Baldy Region of the San Gabriel Mountains, Neter, J., Wasserman, W., 1974. Applied Linear Statistical Models. Richard D. Irwin, Inc.,
California. University of California, Los Angeles (Ph.D. thesis: 195 pp.). Homewood, Illinois (842 pp.).
Ehlig, P.L., 1968. Causes of distribution of Pelona, Rand, and Orocopia Schist along the Painter, S., Cameron, E.M., Allan, R., Rouse, J., 1994. Reconnaissance geochemistry and
San Andreas and Garlock faults. In: Dickinson, W.K., Grants, A. (Eds.), Proceedings of its environmental relevance. J. Geochem. Explor. 51 (3), 213–246. https://doi.org/
Conference on Geologic Problems of San Andreas Fault System. 11. Stanford 10.1016/0375-6742(94)90008-6.
University Publications, Geological Sciences, pp. 294–306. Pettijohn, F.J., 1941. Persistence of heavy minerals and geologic age. J. Geol. 49 (6),
Evans, J.G., 1982. The Vincent Thrust, eastern San Gabriel Mountains, California. In: U.S. 610–625.
Geological Survey Bulletin 1507, (15 pp.). Pettijohn, F.J., Potter, P.E., Siever, R., 1987. Sand and Sandstone, 2nd ed. Springer-
Garbarino, J.R., Kanagy, L.K., Cree, M.E., 2005. Determination of elements in natural- Verlag, New York (553 pp.).
water, biota, sediment and soil samples using collision/reaction cell inductively Reimann, C., 2005. Sub-continental scale geochemical mapping: sampling, quality control
coupled plasma-mass spectrometry. In: U.S. Geological Survey Techniques and and data analysis issues. Geochem.: Explor., Environ., Anal. 5 (4), 311–323.
Methods, (book 5, sec. B, chap.1, 88 p. Method ID: I-2020-05 https://nwql.usgs.gov/ Reimann, C., Caritat, P., 1998. Chemical Elements in the Environment. Springer-Verlag,
Public/rpt.shtml?TM-5-B1, accessed March 21, 2018). Berlin, Germany (398 pp.).
Garver, J.I., Royce, P.R., Smick, T.A., 1996. Chromium and nickel in shale of the Taconic Sato, T., Murakami, T., Yanase, N., Isobe, H., Payne, T.E., Airey, P.L., 1997. Iron nodules
foreland: a case study for the provenance of fine-grained sediments with an ultra- scavenging uranium from groundwater. Environ. Sci. Technol. 31, 2854–2858.
mafic source. J. Sediment. Res. 66 (1), 100–106. Seymour, W.A., 2016. Hydrologic and Geologic Controls on Groundwater Recharge Along
Glazner, A.F., Walker, D.J., Bartley, J.M., Fletcher, J.M., 2002. Cenozoic evolution of the the Mojave River Floodplain Aquifer. San Diego State University, San Diego,
Mojave block of southern California. Geol. Soc. Am. Mem. 195, 19–41. California (M.S. thesis, 59 pp. http://sdsu-dspace.calstate.edu/bitstream/handle/
Gough, L.P., Severson, R.C., Shacklette, H.T., 1988. Element concentrations in soils and 10211.3/175391/Seymour_sdsu_0220N_11429.pdf?sequence=1 (December 2017)).
other surficial materials of Alaska. In: U.S. Geological Survey Professional Paper Shacklette, H.T., Boerngen, J.G., 1984. Element concentrations in soils and other surficial
1458, (53 pp.). materials of the conterminous United States. In: U.S. Geological Survey Professional
Gough, L.P., Wang, B., Smith, D.B., Gustavsson, N., 2005. Geochemical landscapes of Paper 1270, (63 pp.).
Alaska – new map presentations and interpretations for 23 elements in surficial Shacklette, H.T., Hamilton, J.C., Boerngen, J.G., Bowles, J.M., 1971. Element composi-
materials. In: U.S. Geological Survey Professional Paper 1716, (36 pp.). tion of surficial materials of the conterminous United States. In: U.S. Geological
Graham, G.E., Taylor, R.D., Lee, G., Tripp, R.B., 2015. Targeting Cu-Au and Mo resources Survey Professional Paper, 574-D.
using multi-media exploration geochemistry: an example from Tyonek Quadrangle, Smith, D.B., Cannon, W.F., Woodruff, L.G., Solano, F., Ellefson, K.J., 2014. Geochemical
Alaska Range, Alaska. J. Geochem. Explor. 52–65. https://doi.org/10.1016/j.gexplo. and mineralogical maps for soils of the conterminous United States. In: U.S.
2015.014. Geological Survey Open-File Report 2014-1082, https://doi.org/10.3133/
Guthrie, V.A., Kleeman, J.D., 1986. Changing uranium distributions during weathering of ofr20141082. (386 pp.).
granite. Chem. Geol. 54 (1–2), 113–126. https://doi.org/10.1016/0009-2541(86) Stamos, C.L., Martin, P., Nishikawa, T., Cox, B.F., 2001. Simulation of ground-water flow
90075-6. in the Mojave River Basin, California. In: U.S. Geological Survey Water-resources
Hawkes, H.E., 1976. The early days of exploration geochemistry. J. Geochem. Explor. 6 Investigations Report 01-4002, (113 pp.).
(1–2), 1–11. Taylor, S.R., McLennan, S.M., 1995. The geochemical evolution of the continental crust.
Hem, J.D., 1992. Study and interpretation of the chemical characteristics of natural Rev. Geophys. 33, 241–265.
water. In: U.S. Geological Survey Water-supply Paper 2254, (263 pp.). Thompson, D.G., 1929. The Mojave Desert region, California: a geographic, geologic, and
Hotelling, H., 1933. Analysis of a complex of statistical variables into principal compo- hydrologic reconnaissance. In: U.S. Geological Survey Water-supply Paper 578,
nents. J. Educ. Psychol. 24 (6), 417–441. (759 pp.).
Hughes, M.G., Keene, J.B., Joseph, R.G., 2000. Hydraulic sorting of heavy-mineral grains Tieh, T.T., Ledger, E.B., Rowe, M.W., 1980. Release of uranium from granitic rocks during
by swash on a medium-sand beach. J. Sediment. Res. 70 (5), 994–1004. in situ weathering and initial erosion (central Texas). Chem. Geol. 29 (1–4), 227–248.
Izbicki, J.A., Michel, R.L., 2004. Movement and age of ground water in the western part of https://doi.org/10.1016/0009-2541(80)90022-4.
the Mojave Desert, Southern California, USA. In: U.S. Geological Survey Water-re- U.S. Environmental Protection Agency, 2007. Method 7010: graphite furnace atomic
sources Investigations Report 03-4314, (35 pp.). absorption spectrometry, Revision 0. https://www.epa.gov/sites/production/files/
Izbicki, J.A., Ball, J.W., Bullen, T.D., Sutley, S.J., 2008. Chromium, chromium isotopes 2015-12/documents/7010.pdf, Accessed date: 21 March 2018.
and selected trace elements, western Mojave Desert, USA. Appl. Geochem. 23, U.S. Geological Survey, 2004. The National Geochemical Survey - database and doc-
1325–1352. https://doi.org/10.1016/j.apgeochem.2007.11.015. umentation. In: U.S. Geological Survey Open-file Report 04-1001, . http://mrdta.
Izbicki, J.A., Kulp, T., Suarez, D., Barnes, T., Ajwani, C., Kim, T.J., Tseng, I., 2015. In-situ usgs.gov/geochem/ (February 2016).
Arsenic Removal During Groundwater Recharge Through Unsaturated Alluvium. U.S. Geological Survey, 2005a. Active Mines and Mineral Processing Plants in the United
Water Research Foundation (Web Report #4299, 59 pp. http://www.waterrf.org/ States in 2003. U.S. Geological Survey, Reston, Virginia. http://mrdata.usgs.gov/
PublicReportLibrary/4299.pdf (March 2016)). mineplant/ (February 2016).
Jennings, C.W., with modifications by Gutierrez, C., Bryant, W., Saucedo, G., and Wills, U.S. Geological Survey, 2005b. Mineral Resources Data System. U.S. Geological Survey,
C., 2010, Geologic map of California: California Geological Survey, Geologic Data Reston, Virginia. http://mrdata.usgs.gov/mrds/ (February 2016).
Map No. 2, scale 1:750,000. Webb, J.S., Fortescue, J., Nichol, I., Tooms, J.S., 1964. Regional Geochemical
Kim, C.S., Wilson, K.M., Rytuba, J.J., 2011. Particle-size dependence on metal(loid) Reconnaissance in the Namwala Concession Area, Zambia: Technical Communication
distributions in mine wastes: implications for water contamination and human ex- 47. Geochemical Prospecting Research Centre.
posure. Appl. Geochem. 26, 484–495. https://doi.org/10.1016/j.apgeochem.2011. Webb, J.S., Thornton, I., Thompson, M., Howarth, R.J., Lowenstein, P.L., 1978. The
01.007. Wolfson Geochemical Atlas of England and Wales. Oxford University Press, Oxford
Lines, Gregory C., 1996. Ground-Water and Surface-Water relations along the Mojave (70 pp.).
River, southern California. In: U.S. Geological Survey Water-resources Investigations Welch, A.H., Westjohn, D.B., Helsel, D.R., Wanty, R.B., 2000. Arsenic in ground water of
Report 95-4189, (42 pp.). the United States: occurrence and geochemistry. Ground Water 38 (4), 589–604.
Meisling, K.E., Weldon, R.J., 1989. Late Cenozoic tectonics of the northwestern San Wenzel, W.W., Kirchbaumer, N., Prohaska, T., Stingeder, G., Enzo, L., Adriano, D.C.,
Bernardino Mountains, southern California. Geol. Soc. Am. Bull. 101, 106–128. 2001. Arsenic fractionation in soils using an improved sequential extraction proce-
Metzger, L.F., Landon, M.K., House, S.F., Olsen, L.D., 2015. Mapping selected trace ele- dure. Anal. Chim. Acta 436, 309–323.
ments and major ions, 2000–2012. In: Mojave River and Morongo Groundwater Wold, S., Esbensen, K., Geladi, P., 1987. Principal component analysis. Chemom. Intell.
Basins, Southwestern Mojave Desert, San Bernardino County, California: U.S. Lab. Syst. 2, 37–52.
Geological Survey Data Release, https://doi.org/10.5066/F7Q23X95. (http://ca. Wright, M.T., Belitz, K., 2010. Factors controlling the regional distribution of vanadium in
water.usgs.gov/mojave/water-quality.html). groundwater. Ground Water 48 (4), 515–525. https://doi.org/10.1111/j.1745-6584.
Miller, D.M., Bedford, D.F., 2000. Geologic map database of the El Mirage Lake area, San 2009.00666.x.
Bernardino and Los Angeles counties, California. In: U.S. Geological Survey Open-file
15