Accepted Manuscript

A new estimate of detrital redox-sensitive metal concentrations and variability

in fluxes to marine sediments

Devon B. Cole, Shuang Zhang, Noah J. Planavsky

PII: S0016-7037(17)30471-4
DOI: http://dx.doi.org/10.1016/j.gca.2017.08.004
Reference: GCA 10409

To appear in: Geochimica et Cosmochimica Acta

Please cite this article as: Cole, D.B., Zhang, S., Planavsky, N.J., A new estimate of detrital redox-sensitive metal
concentrations and variability in fluxes to marine sediments, Geochimica et Cosmochimica Acta (2017), doi: http://
dx.doi.org/10.1016/j.gca.2017.08.004

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
 A new estimate of detrital redox-sensitive metal concentrations and variability in

fluxes to marine sediments

Devon B. Cole1*, Shuang Zhang1, Noah J. Planavsky1

1
Department of Geology & Geophysics, Yale University, New Haven, CT 06511, USA
*
Corresponding Author: Devon Cole, Devon.Cole@yale.edu (617) 823-3950

Abstract

The enrichment and depletion of redox sensitive trace metals in marine sediments have

been used extensively as paleoredox proxies. The trace metals in shale are comprised of

both detrital (transported or particulate) and authigenic (precipitated, redox-driven)

constituents, potentially complicating the use of this suite of proxies. Untangling the

influence of these components is vital for the interpretation of enrichments, depletions,

and isotopic signals of iron (Fe), chromium (Cr), uranium (U), and vanadium (V)

observed in the rock record. Traditionally, a single crustal average is used as a cutoff for

detrital input, and concentrations above or below this value are interpreted as redox

derived authigenic enrichment or depletion, while authigenic isotopic signals are

frequently corrected for an assumed detrital contribution. Building from an extensive

study of soils across the continental United States – which upon transport will become

marine sediments – and their elemental concentrations, we find large deviations from

accepted crustal averages in redox-sensitive metals (Fe, Cr, U, V) compared to typical

detrital tracers (Al, Ti, Sc, Th) and provide new estimates for detrital contributions to the

ocean. The variability in these elemental ratios is present over large areas, comparable to

the catchment-size of major rivers around the globe. This heterogeneity in detrital flux

highlights the need for a reevaluation of how the detrital contribution is assessed in trace
metal studies, and the use of confidence intervals rather than single average values,

especially in local studies or in the case of small authigenic enrichments.

1. Introduction

1.1 Redox Proxies

Enrichments of transition metals in marine sediments, such as Cr, V, U, Mo, and Fe,

are some of the most widely used redox proxies in modern and ancient sedimentary

systems (e.g., Tribovillard et al., 2006). The basic idea behind using these trace elements

as redox proxies is straightforward—their solubility is strongly controlled by local redox

conditions. With Cr, V, U, and Mo (herein referred to redox sensitive trace metals),

increased solubility in oxic conditions results in large metal enrichments under oxygen-

poor conditions. Iron, though the system is slightly more complex, is also typically

enriched in anoxic sediments as the result of a benthic shuttle from reducing sediments

underlying oxic waters and an efficient Fe trap in anoxic settings (e.g., Lyons and

Severmann, 2006). Mobilization of redox sensitive trace metals occurs during oxidative

chemical weathering at Earth’s surface, and subsequent riverine transport to the ocean.

Upon entering reducing conditions in the marine system, in either the water column or

the sediment pile, these metals can be reduced and or converted to more insoluble particle

reactive phases (e.g., Algeo and Maynard, 2004), resulting in precipitation and

sequestration within the sediment. Because solubility is linked to redox state, metal

enrichments in sediments have the potential to act as a very powerful tracer of past redox

conditions.

Redox sensitive trace metals enrichments have also been used extensively to explore

global marine redox conditions (e.g., Algeo, 2004; Partin et al., 2013; Reinhard et al.,
2013; Scott et al., 2008). The dissolved marine reservoir size of redox sensitive trace

metals is directly linked to the global marine redox landscape (e.g., Algeo, 2004;

Emerson and Huested, 1991; Hastings et al., 1996). Assuming a constant input term, in a

well oxygenated ocean the size of the dissolved marine redox sensitive metal reservoir is

larger than in an anoxic ocean due to inhibited marine burial under oxic conditions. This

results in large metal enrichments in the few anoxic regions of a well-oxygenated ocean

and small enrichments in anoxic settings in a largely anoxic ocean (see Reinhard et al.,

2013; Scott et al., 2008). One of the major appeals of this approach is that these

enrichments can be directly translated into a quantitative mass balance models (e.g.,

Partin et al., 2013; Reinhard et al., 2013).

More recently, this approach has been taken a step further with the use of redox
53 98 238
sensitive metal isotope systems (e.g., δ Cr, δ Mo, δ U). Beyond enrichments, isotopic

data can provide much more detailed information about the surface cycling of these

elements and potentially the nature of the reductive sinks (i.e., euxinic, ferruginous,

suboxic). These systems are also appealing as they can be translated into quantitative

isotope mass balances, which have led to substantial insights into the evolution of global

redox conditions (e.g., Arnold et al., 2004; Dahl et al., 2010; Kendall et al., 2011; Tissot

and Dauphas, 2015; Wang et al., 2016a; Wang et al., 2016c; Weyer et al., 2008). It

should be recognized however, that a detrital correction has to be applied for many

potential seawater archives, thus first requiring recognition of the potential for variation

within this term.

Chromium, vanadium, and uranium are emerging isotope proxies of particular interest

as they have the potential to transform our understanding of variable redox systems.

Acknowledgment of the heterogeneity in the detrital input of these metals is critical in for
the quantification of authigenic isotope signals, and for the recognition of small but

significant redox-driven authigenic enrichments of Cr, V, and U. Further, Cr and V can

potentially provide a unique signal for weakly reducing (suboxic) or oscillating redox

conditions, as these two elements are the most easily reduced of the redox-sensitive suite

of trace metals (Algeo and Maynard, 2004; Calvert and Pedersen, 1993; Li et al., 2015b).

Chromium can be reduced via nitrate reduction, making it the most easily reduced of the

traditional suite of trace metals (Gueguen et al., 2016; Rue et al., 1997). Vanadium also

has unique redox behavior as it can be reduced twice, and can be sourced from suboxic

sediments (Calvert and Pedersen, 1993; Hastings et al., 1996). In the case of Cr, small

enrichments indistinguishable from the detrital component might be expected, however

these could likely be revealed with isotopic analyses, which would distinguish an

unfractionated detrital flux from a fractionated authigenic component (assuming

deposition under an oxidizing earth atmosphere). The V isotope system is also likely to

provide a unique signature of suboxic conditions, making both of these systems critically

important to improving the resolution of our understanding of the oxygenation of the

Earth’s oceans.

1.2 Detrital contribution

Although there have been, without question, large steps forward in our understanding

of marine redox evolution from trace metal geochemistry, uncertainties about the

magnitude of background particulate flux – or detrital signal – unrelated to the redox

conditions at the time of deposition have been problematic and may have led, in some

cases, to spurious conclusions from redox sensitive trace metal studies. The detrital

particulate component of a specific metal will eventually be delivered to the sediment

regardless of the redox state of the depositional environment. Therefore, a baseline

detrital value must be established for paleoredox proxies of interest, in order to resolve

enrichments beyond this background value (e.g., Tribovillard et al., 2006). It is

assessment of this authigenic component – the target of both enrichment and isotopic

studies – that is critical to making paleoenvironmental interpretations.

All redox-sensitive trace elements are delivered to the ocean in both dissolved form

and as detrital particulate material. For some metals, such as molybdenum (Mo), there is

a very small detrital component, in most cases making redox interpretations robust with

respect to, and regardless of, changing background detrital input (Crusius et al., 1996).

Molybdenum however, is likely reduced via the thiomolybdate pathway whereby Mo

becomes particle-reactive, and increasingly so with each successive sulfidation reaction

(i.e., dithio-, trithio-, tetrathiomolybdate), though each successive step is roughly one

order of magnitude slower (Erickson and Helz, 2000; Helz et al., 1996). Thus,

enrichments are strongly tied to sulfide levels and the presence of persistent euxinia

(Erickson and Helz, 2000; Scott and Lyons, 2012). Other redox-sensitive metals, such as

Cr and V, have the potential to be much more easily reduced, and thus function as more

sensitive indicators of reducing environments (Algeo and Maynard, 2004; Rue et al.,

1997). Similarly, U also responds to all reducing conditions, though reduction occurs

primarily in the sediment column (Algeo and Maynard, 2004). Unfortunately, Cr and V,

in particular, have a very large detrital component, adding significant complexity to the

disentanglement of environmental signals (authigenic enrichments) and provenance

(background detrital input) (e.g., Perkins et al., 2008). Indeed, the majority of transition

metals used in redox studies (i.e., Fe, Cr, V, U) have a substantial detrital component
(e.g., Algeo, 2004; Reinhard et al., 2013; Sahoo et al., 2012), requiring careful

assessment in order to interpret enrichments truly driven by redox conditions.

Understanding the average composition of the crust has been the focus of numerous

studies since as early as the late 19th century, eventually expanding to estimates of

average crustal trace element values (e.g., Clarke, 1889; Condie, 1993; Eade and Fahrig,

1971; Gao et al., 1998; Goldschmidt, 1937; Kamber et al., 2005; Rudnick and Gao, 2014;

Shaw et al., 1967; Taylor and McLennan, 1985; Turekian and Wedepohl, 1961;

Wedepohl, 1995). Broadly, there are two basic approaches used to make these estimates –

either calculating weighted average values for rocks exposed on Earth’s surface (e.g.,

Clarke, 1889; Condie, 1993), or compiling average composition from siliciclastic

sedimentary rocks or glacial deposits which sample the crust (e.g., Goldschmidt 1937;

Taylor and McLennan, 1985). The strengths and weaknesses of each of these studies and

the level of consistency between them is discussed at length by Rudnick and Gao (2014),

who, present their own best estimate of average crustal composition which we refer to

throughout this study. It should be noted however, that non-negligible variation remains

in the compilation of these estimates, particularly for trace metals.

While likely broadly representative of global means, especially considering the

overall consistency of these numerous studies, none of the work outlined above provides

constraints on regional-scale variability — a critical missing link in understanding redox-

sensitive trace metal enrichments in sediments. This is especially relevant to studies

working on a basinal or local scale, or in the case of small but non-negligible authigenic

enrichments. The latter is particularly pertinent in both the Precambrian and in restricted

basins, where the marine reservoir of metals such as Cr, U, and V is already depleted due

to widespread anoxic waters (Algeo, 2004; Algeo and Rowe, 2012; Partin et al., 2013;
Reinhard et al., 2013). Further, at very low oxygen levels (i.e., much of the

Precambrian), there is an effect on the detrital flux of the redox-sensitive metals as a

result of the availability of oxidants. For Fe, in particular, oxygen levels of at least 0.01 %

but possibly as high as 1% present atmospheric levels (Crowe et al., 2013) are needed to

limit runoff of Fe(II). Below these levels, it is possible the detrital flux could appear

‘depleted’.

Traditionally, researchers have relied on accepted crustal average values to elucidate

redox-induced enrichments, and occasionally to demonstrate depletions (e.g., Reinhard et

al., 2014; Zhang et al., 2016). In the place of empirical constraints on detrital flux to a

basin, which are often limited to recent deep sea cores, or even the relationship of a basin

with the open ocean (frequently the case in deep time), crustal averages are used as

cutoffs for authigenic enrichment with no consideration for potential variability both

laterally (between basins) or through time (within a single basin experiencing changes in

detrital input source, rate of erosion, or rate of deposition). Unfortunately, this can result

in the failure to clearly distinguish a redox driven signal, especially in the case of

depletions or small enrichments. Some researchers (e.g., Algeo and Maynard, 2004) have

used trace metal concentrations in grey shales within the same or adjacent sedimentary

units as a baseline for detrital input (assuming oxic deposition of these sediments),

however this still fails to account for variation of the detrital input through time, and

lithology (e.g., low organic content) is not a practically robust redox indicator (e.g.,

Myrow, 1990). Further, such an approach is not always possible in many settings. Others

such as Perkins et al. (2008) have used redox-sensitive/detrital tracer ratios to identify a

slope associated with detrital contribution and identify points above this lower limit to be

authigenically enriched. Unfortunately this approach requires a shift in the depositional

redox environment within the measured section, and if that shift does not reflect a change

from a system where authigenic metal enrichment is not possible, then this slope will

represent a mixing line between authigenic and detrital components, rather than a solely

detrital contribution. More recently, variability has been demonstrated for Cr/Ti in

modern oxic marine sediments (Gueguen et al., 2016). However, there is potential for

processes resulting in the loss or gain of metals in oxic sediment, adding complexity to

the problem, and suggesting that a clearer picture of the detrital contribution may come

from investigating continental source material.

To investigate the potential variability of detrital input to the ocean we have

examined the source of this material – modern soils – from the United States Geologic

Survey (USGS) North American Soil Geochemical Landscapes Project covering the

continental United States. The results of the sampling campaign are available from the

USGS archives and were released in USGS open file report 2014-1082 (Smith et al.,

2014). We suggest that topsoils provide a reasonable analogue for marine detrital

contribution as these sediments will be eroded and directly transferred to the marine

system. These findings indicate that a more holistic approach, beyond the use of a mean

crustal cutoff value to quantify the detrital contribution, is required to make a significant

and unique paleoenvironmental interpretation of depositional redox conditions. This

database has allowed a new estimate of surficial variability in trace metal composition of

sediments just prior to their transport to the ocean, allowing a first approximation of

confidence intervals (‘error bars’) associated with enrichment-based redox

interpretations.

2. Methods
In order to provide new estimates of the detrital contribution of redox-sensitive metals to

marine sediments, and emphasize both the complexity and variability of these systems we

have investigated a suite of redox-sensitive metal/detrital tracer ratios from across the

continental United States. Our analysis builds from a soil database (see Smith et al.,

2014), which includes elemental concentrations of 45 major and trace elements from

more than 4,850 localities. All of the samples in this database were collected between

2007 and 2010 based on the sampling procedure outlined in Smith et al. (2014). All

samples were sieved to < 2mm grain size, specifically targeting the fine fraction and to

avoid the anomalous analysis of individual large grains. All elemental analyses were

made on inductively-coupled plasma atomic emission spectrometry (ICP-AES) or

inductively-coupled plasma mass spectrometry (ICP-MS) following a multi-acid total

digestion. For further details on collection and analyses, see Smith et al. (2014).

Using this database, we have compiled ratios of widely used redox-sensitive proxies

to their most commonly employed detrital tracers. That is, we have examined the ratios of

often studied elements with non-conservative behavior during redox-dependent surface

processes relative to tracers that are considered conservative (i.e., they will remain in

particulate form). These ratios, rather than just concentrations, are employed in order to

ideally circumvent dilution effects, changes in sedimentation rate, and changes in flux

related to continental weathering. The following ratios were compiled, and the data was

then interpolated over maps of the continental US; V/Al, V/Ti, V/Sc, U/Al, U/Th, Cr/Ti,

Cr/Al, Fe/Al, and the detrital tracer Al/Ti. We have then examined the distribution of

values for each elemental ratio in comparison to widely accepted continental crustal

average estimates of Rudnick and Gao (2014).

 In addition, we have examined large-scale variation that can present on catchment-

size scales. To quantify the lateral persistence of deviation from crustal means in the

database, a 1000 iteration Monte Carlo resampling procedure was executed for Cr/Ti,

V/Al, and Fe/Al. More specifically: An area was randomly selected from a database of

the world’s 405 largest modern drainage basins from the Global Runoff Data Center

(GRDC, 2016) and a height-width ratio (ranging from 0.5 to 2) was randomly assigned to

a rectangle of that area. This rectangle was then overlain on the geographic dataset for

each elemental ratio and the mean and standard deviation of the data points within that

area were calculated. This process was then repeated 1000 times for each elemental ratio.

Herein, the means generated from this resampling procedure are referred to as the

catchment resampled means.

Finally, to investigate variability with soil depth for cases where the primary physical

erosion is the result of catastrophic erosion events (i.e., mass wasting, e.g., Osterkamp

and Friedman (2000)), we have examined the difference and variability between the

topsoil (0 – 5 cm depth), A horizon (composite of soil A horizon), and C horizon (soil C

horizon – partially weathered parent material) or a sample from 80-100 cm depth if the C

horizon is deeper than 1 m (Smith et al., 2014). Ratios of Cr/Ti and V/Al from each of

these three horizons were examined to investigate shifts in mean values and variability,

both laterally and vertically. This vertical variation through the soil profile also provides

information about preferential loss during weathering, and the variable nature of these

processes, in addition to evidence for inherited variability from the regolith.

3. Results
For all elemental ratios considered, we find significant variance away from crustal

average values of Rudnick and Gao (2014) in topsoils across the continental US. Cr, V,

and Fe show very large distributions, while U remains closer to the crustal average. These

observations of significant heterogeneity are consistent regardless of the detrital tracer

chosen, though the relationship between soil means and crustal means can vary

depending on the detrital tracer (e.g., V/Ti and V/Al). Further, we find that variations

from the crustal averages are consistently maintained on large, laterally continuous

scales, demonstrated by the catchment resampled means of topsoil ratios for Cr/Ti, V/Al,

and Fe/Al. Means and standard deviations of the U.S. soil data and crustal averages for

each elemental ratio considered are summarized in Table 1.

3.1 Chromium

The mean topsoil value for Cr/Ti is 0.015 ± 0.094 (2SD), which is substantially lower

than the continental crustal average of 0.02 (Rudnick and Gao, 2014), while the mean

topsoil value for Cr/Al is 9.015 ± 46.101 (2SD) compared to the crustal average of

11.288 (Rudnick and Gao, 2014). Rudnick and Gao (2014) report 1 sigma errors,

however these are based on variation between other published crustal estimates, and are

rarely considered in redox-focused geochemical studies, thus, herein we refer only to the

crustal average. Chromium shows the largest disparity between the topsoil mean and the

continental crustal averages of all elemental ratios considered in this study. The Cr/Ti

data is also highly variable and this variability is maintained in the catchment resampled

data, where the catchment resampled mean is 0.015 ± 0.015 (2SD) (Fig.1). Further, when

the standard deviations for each of the 1000 resampled areas are compiled, the mean of

these standard deviations is 0.016, demonstrating sustained variation on all sampling

scales. The similarity of the resampled mean to that of the raw data demonstrates that our

resampling procedure accurately reflects the true distribution of the data. In addition, we

have examined the soil data from the A and C horizons, in addition to the topsoil (Fig. 2).

The A horizon mean is 0.013 ± 0.014, while the C horizon mean is 0.015 ± 0.128. With

consideration of the large standard deviations for all three soil horizons, these means

show no significant difference related to soil depth, using a T-test (p-value = 0.94 for C

horizon compared to topsoil, and p-value = 0.88 for C horizon compared to A horizon).

3.2 Vanadium

Three ratios were examined for V (V/Al, V/Ti, V/Sc), along with catchment resampling

for V/Al, and the percentage change between the C horizon and the topsoil ratios of V/Al

(Figs. 3, 4). The mean topsoil value for V/Al is 13.279 ± 12.758 (2SD) compared to the

crustal average of 11.9 (Rudnick and Gao, 2014). The mean topsoil value for V/Ti is

229.283 ± 246.471 (2SD) compared to the crustal average of 252.8 (Rudnick and Gao,

2014). Finally, the mean topsoil value for V/Sc is 9.135 ± 5.598 (2SD) compared to the

crustal average of 6.9 (Rudnick and Gao, 2014). The large-scale lateral variability of V

was tested with catchment resampling for V/Al, where the catchment resampled mean is

13.127 ± 6.182 (2SD), while the mean 2SD of the resampled areas is 9.155 (Fig. 3).

Significantly, this mean is similar to the topsoil mean, indicating that the resampling

procedure accurately reflects the distribution of the original data. We also examined

variation within the soil profile for V/Al, and find that the mean for the C horizon is

13.470 ± 13.582 (2SD). Using a t-test, we find that the C horizon and topsoil do not show

a statistically significant difference (p-value = 0.23). The percentage change of V/Al

between the C horizon and topsoil is also highly variable with an average shift of -3.47%

± 70.36 (2SD) (Fig. 4).

3.3 Uranium

Uranium was compared to the two detrital tracers Al and Th. The topsoil mean for U/Al

is 0.578 ± 1.134 (2SD) in comparison to the crustal mean of 0.33 (Rudnick and Gao,

2014), while the topsoil mean of U/Th is 0.285 ± 0.369 (2SD) compared to the crustal

mean of 0.26 (Rudnick and Gao, 2014) (Fig. 5). The catchment resampled mean for U/Th

is 0.282 ± 0.102 (2SD), while the mean 2SD of the resampled areas is 0.179.

3.4 Iron

Iron was investigated in comparison to Al. The topsoil mean of Fe/Al is 0.476 ± 1.223

(2SD) in comparison to the crustal average of 0.48 (Rudnick and Gao, 2014). The

catchment resampled mean is 0.472 ± 0.302 (2SD), while the mean 2SD of the resampled

areas is 0.406 (Fig. 6). The mean of the C horizon is 0.475 ± 0.427 (2SD), consistent

with, on average, quantitative iron retention during weathering. Using a T-test we find

that the C horizon and topsoil do not show a statistically significant difference (p-value =

0.90).

3.5 Aluminum and Titanium

Finally, we examined Al/Ti ratios, which are utilized as a detrital or provenance tracer.

The topsoil mean of Al/Ti is 19.026 ± 17.640 (2SD) in comparison to the crustal mean of

21.24 (Rudnick and Gao, 2014) (Fig. 7). We have also considered the relationship

between Cr/Ti and Al/Ti. Because Cr contribution is primarily derived from mafic
igneous rocks, an inverse relationship might be expected between Cr/Ti and Al/Ti, given

that these ratios are lower and higher, respectively, in felsic versus mafic igneous rocks.

If this is the case, the use of Ti as a normalizing detrital tracer should also help to

normalize for the influence of mafic vs. felsic contributions of Cr. However we find no

correlation between these two ratios (r-squared = 0.002) (Fig. 8).

We have also examined the relationship between Al contents and each of the four

primary ratios (Cr/Ti, V/Al, U/Th, Fe/Al), and compared this data to the global average

values for suspended sediment in rivers from Viers et al. (2009). We find that our

distributions show no relationship with Al contents, and reflect the published riverine

values and their ranges, despite some slight offsets, which are discussed below (Fig. 9).

4. Discussion

Accepting the significant heterogeneity of trace metal concentrations and ratios in surface

sediments observed in this study, we consider the use of crustal cutoff values for

distinguishing authigenic enrichments an inaccurate oversimplification. The pitfalls of

normalization of trace elements have been discussed previously (Van der Weijden, 2002),

however we seek to emphasize an awareness of the surface heterogeneity that can plague

normalization schemes, based on empirical data from surface sediments. The potential for

deviations from crustal means on catchment-size scales is very high, especially for V and

Cr, as discussed below. This type of variation can present on both lateral, basin-wide

scales, as well as vertically, in the rock record (Walther, 1894), leading to flawed

paleoredox reconstructions. Though this study does not build from a global dataset, the

continental US is composed of a wide variety of bedrock compositions, climates, and

drainage basins that collectively can at least provide an approximation of the magnitude

of variability that may be expected from a larger, global dataset. It is likely that pollution
has affected some samples of this study, however it is unlikely that these samples drive

the broad spatial variability observed in the elemental ratios examined. Minimally,

recognition of the capacity for background variability is especially critical in the analyses

of individual stratigraphic sections, or even multiple sections within a single basin. The

magnitude of the variation observed in plausible detrital contributions of Cr and V

indicates that the resolvability of truly redox-driven signals is, in many cases, too poor to

draw any conclusions without further analytical work.

4.1 Chromium

Chromium variability was assessed in comparison to Ti and Al – the most commonly

used detrital tracers for Cr (e.g., Cole et al., 2016b; Crowe et al., 2013; Gueguen et al.,

2016; Konhauser et al., 2011; Reinhard et al., 2013; Reinhard et al., 2014). Traditionally,

the continental crust average Cr/Ti value of 0.02 (Rudnick and Gao, 2014) is used as a

cutoff value to characterize the detrital contribution to marine sediments, and values

above 0.02 are considered evidence of authigenic enrichment resulting from reduction

under an anoxic water column. However, significant heterogeneity observed in both

Cr/Ti and Cr/Al of topsoils across the US (Fig. 1) indicates a confidence interval rather

than a single crustal average cutoff value is a more useful means of gauging Cr

enrichments in marine sediments. Importantly, this interval is so large that its practical

applications will be notably limited, requiring either the analyses of large and statistically

substantial data sets, or else use of alternative geochemical techniques such as isotopic

analyses (see 5.1 Recommendations).

Critically, we find not only major variability away from this crustal average value in

individual samples, but also persistent and widespread areas where topsoil Cr/Ti values
fall well below the canonical crustal average. This is most apparent in the catchment

resampled distributions, which show that sub-crustal means and substantial variation

persist on catchment-size scales (Fig. 1C). Most crucially, it is this potential to sustain

deviations away from the mean crustal value over a large and continuous areal extent that

will most significantly impact the resolvability of the background detrital contribution in

the sedimentary record. Additionally, considering the observed distributions, it is likely

that a detrital flux well below the crustal average used in many Cr isotope studies could

be sustained, leading to the concealment of significant authigenic enrichments critical to

the interpretation of isotopic analyses (e.g., Cole et al., 2016b; Gueguen et al., 2016).

Further, within the soil profile, the mean Cr/Ti value of the topsoil, the A horizon,

and the C horizon all fall below the crustal average (Fig. 2). Most notably, we find that

overall variability is essentially maintained vertically through the soil profiles on a large

scale, indicating that varying degrees and rates of erosion – whether a result of mass

wasting or continuous removal – will have little effect on the variation potential of the

detrital flux to the ocean. Further, considering the significant range of percentage change

in Cr/Ti between the C horizon and topsoil, it is likely a large variety of surface processes

enable the mobilization of Cr, making the detrital flux from a particular regolith difficult

to classify based only on rock type. These findings suggest that resolving true authigenic

enrichments based solely on concentrations will be very difficult, and generally limited to

scenarios with either a detailed understanding of source material and sedimentation rate,

or else exceptionally high enrichments that are confined to the Phanerozoic.

This lack of resolvability is particularly troublesome when redox interpretations based

on small to moderate enrichments are applied on local or basinal scales, or through

individual stratigraphic sections. At this scale, variation in the detrital flux has the
potential to exert a major control on the Cr/Ti signal, greatly skewing redox

interpretations if this flux is dissimilar from the crustal mean. This includes scenarios

with a Cr rich detrital flux that could appear to have an authigenic enrichment or, in

contrast, the concealment of significant enrichments when coupled with a Cr-poor detrital

flux. Both cases are particularly relevant in Precambrian studies: there is potential for the

former in the Archean while weathering an ultramafic crust and the likelihood of the

latter case through the Mesoproterozoic where the marine reservoir of Cr has been drawn

down by widespread anoxia, making only muted authigenic enrichments possible relative

to the Phanerozoic (e.g., Reinhard et al., 2013). Further there is significant variation in

the mapped distributions between Cr/Al and Cr/Ti (Fig 1), suggesting that the choice of

detrital tracer could also have an outsized impact on the interpretation of redox-driven

enrichments.

While the findings of this study elucidate the presence of a large error bar on the

detrital contribution of Cr to shale, challenging the framework of several studies (e.g., Li

et al., 2015a; Reinhard et al., 2013), we suggest that as the geographic and temporal scale

of a study increases, redox trends tend to become more robust. While some error in

interpretation is likely for individual samples, when interpreted in concert with multiple

formations globally or as broad-scale trends through time, such interpretations are likely

to remain robust.

Finally, as an emerging isotope system (e.g., Cole et al., 2016b; Frei et al., 2009; Frei

et al., 2013; Planavsky et al., 2014; Wang et al., 2016c; Wille et al., 2013), it is vital that

researchers recognize the issues of resolvability of the authigenic Cr component, which is

the target of these investigations. In some cases, large datasets and a statistical approach

can remedy the ambiguity of small authigenic signals (e.g., Cole et al., 2016). In other
cases, as our understanding of the system develops, Cr isotopes have the potential to

provide much higher resolution insights regarding redox of past depositional

environments (see 5.2 Applications of isotope systems).

4.2 Vanadium

Vanadium was compared to three different commonly used detrital tracers (Al, Ti, Sc) in

order to demonstrate not only variability in V, but also the impact of the chosen detrital

tracer and the heterogeneity within these systems that are generally considered stable,

constant and frequently interchangeable. Notably, all three V ratios not only show large

standard deviations, but also the means of the soil data show no consistent offset

direction from crustal averages – the V/Ti soil mean falls below the crustal mean, while

both the V/Sc and V/Al soil means fall above the crustal means (Fig. 3). The significance

of these incongruous relationships between soil and crustal means is twofold; first, the

lack of consistency between tracers reflects further under-recognized source rock

variability related to bedrock type and crustal heterogeneity, and secondly, this further

emphasizes that on the scale of small enrichments, the choice of a particular detrital

tracer compared solely to the crustal mean could significantly impact the interpretation of

the redox signal.

We find, based on catchment resampling of V/Al, that variability in V flux is

maintained over large drainage-basin areas, demonstrating the potential for sustained

detrital inputs substantially different from the crustal average in the V system (Fig. 3).

This widespread lateral variability that exists in the present will result in similar

variability through time within the sediment column in the depositional basin (Middleton,

1973; Walther, 1894). While the expression of variability in background detrital V/Al
will be the result of a number of factors including weathering and sedimentation rates, we

suggest without a local understanding of both the processes at work and V/Al ratio of the

source material, a single V/Al cutoff for detrital input cannot be implemented without a

confidence interval that accounts for this heterogeneity.

We also find, based on the percentage change in V/Al between the C horizon and the

topsoil samples from individual sites, that there is potential for both increases and

decreases in the ratio as a result of local processes (Fig. 4). While preferential loss of V

from the soil will primarily be the result of chemical weathering – resulting in dissolution

and removal of that V from the detrital flux entirely – the potential for variation vertically

through the soil profile can influence the nature of the detrital flux, especially if the

primary physical weathering in a region is the result of high energy catastrophic erosion

(i.e storms) as opposed to the low energy more sustained removal of only the topsoil

(e.g., Osterkamp and Friedman, 2000; Wolman and Miller, 1960).

Further, the percentage change in V/Al from C horizon to topsoil provides a measure

of the lateral variation in the potential preferential mobilization of V. We find both

relative losses and gains in V between the C horizon and the topsoil, indicative of the

susceptibility of V to mobilization through a variety of pathways including a ligand-

bound fraction and oxidation and dissolution (Emerson and Huested, 1991; Shiller and

Mao, 2000; Wehrli and Stumm, 1989). Such added complexities in the surface cycling of

V only make unraveling the detrital and authigenic contributions to marine sediment

more difficult. This highlights the possibility of varying detrital contribution from a

specific area based not just on bedrock composition, but relative activity of surface

processes including chemical weathering, oxidation and reduction within the soil profile,

and mobilization via organic and oxide sorption. Without additional data that can provide
more highly resolved information on the redox setting, it is very difficult to make

interpretations based on either depletions or enrichments in V relative to a crustal

average.

4.3 Uranium

Uranium was compared to Al and Th, the most commonly used U detrital tracers (e.g.,

Adams and Weaver, 1958; Anbar et al., 2007; Andersen et al., 2014; Calvert and

Pedersen, 1993; Morford et al., 2001; Partin et al., 2013; Wignall and Myers, 1988).

Similar to other proxies considered, variation from crustal means does persist, however

the standard deviations, especially for U/Th, remain much smaller than those for V or Cr.

This suggests that the U detrital flux is much more consistent, and less likely to result in

major stratigraphic deviations in the depositional system than V or Cr. Notably, the

topsoil mean for U/Th is very similar to the crustal average, however the U/Al mean

displays considerably more offset (Fig. 5). This is likely at least in part, the result of

similarities in atomic size and elemental behavior of U and Th, and provides a

compelling case for choosing a detrital tracer that is most chemically similar to the

element of interest.

The smaller scale variability in the U/Th system may make the use of a confidence

interval a practical alternative in some cases for distinguishing significant authigenic

enrichment (see 5.1 Recommendations). Despite these more muted effects compared to

other redox-sensitive systems, it still remains important to recognize the potential for

small variations. This is particularly the case when considering small-scale changes in U

enrichments in the rock record, especially in individual stratigraphic sections.

Additionally, in isotopic studies, the isotopic value of the U detrital flux is constant (e.g.,
De Corte, 2016), however it is likely important to consider the potential for variation in

the magnitude of this term in the overall U isotope mass balance. In such studies, the use

of a confidence interval rather than a crustal cutoff remains the most conservative and

accurate approach, and it is likely conclusions regarding paleoredox conditions will

remain ambiguous without the use of additional techniques.

4.4 Iron

Iron, and in particular iron speciation, has been the focus of decades of work seeking to

improve our understanding of depositional paleoredox conditions. The use of iron as

local redox proxy, although more complicated, is also based on Fe scavenging under

anoxic conditions (e.g., Poulton and Raiswell, 2002). Iron is, of course, more soluble

under anoxic conditions. However, Fe is typically enriched in anoxic sediments because

of a benthic shuttle of iron from reducing marine sediments (anoxic sediments underlying

an oxic water column) and a strong Fe trap in anoxic settings (foremost in euxinic

settings) (Lyons and Severmann, 2006; Raiswell and Canfield, 1998; Severmann et al.,

2008). Yet, similar to other redox-sensitive metals discussed, studies utilizing Fe

enrichments require an understanding of the portion of reactive iron species relative to

the total iron content of the sediment (e.g., Berner, 1970; Poulton and Canfield, 2005;

Raiswell and Canfield, 1998). Significantly, there is the potential to convert more highly

reactive species to the unreactive silicate bound fraction via metamorphism, thus

disguising redox driven enrichments. It has been suggested (Raiswell et al., 2011) that

Fe/Al ratios could help resolve evidence of this type of alteration, by exposing

enrichments relative to crustal averages (i.e., beyond the silicate fraction).

 Here, Fe/Al was examined in the topsoil dataset, as this ratio has been employed by

researchers, often in conjunction with C and S systematics relating to pyrite formation, to

constrain the detrital Fe contribution and further elucidate redox settings (e.g., Cruse and

Lyons, 2004; Lyons and Severmann, 2006; Lyons et al., 2003; Partin et al., 2013) (Fig 6).

Enriched Fe/Al ratios are relied upon as an indicator of reducing environments, however

we find significant heterogeneity in the crustal ratio of these elements, suggesting –

similar to the other redox tracers discussed herein – that, minimally, a confidence interval

must be used in assessing sedimentary Fe/Al values, rather than a single average cutoff.

As with other proxies discussed, the potential for detrital variability is too great to draw

realistic conclusions about redox based on Fe/Al alone, especially on a local scale. Using

catchment resampling, we again demonstrate the potential of capturing detrital signals

that deviate from crustal averages over large areas, resulting in sustained and non-

negligible variation from crustal mean values in the detrital flux to the adjacent

depositional basin.

4.5 Aluminum and Titanium

In addition to a suite of redox-sensitive elements, we have also examined heterogeneity in

Al/Ti ratios, which are a provenance tracer and a ratio generally considered to not be

susceptible to alteration during surface processes. This ratio (Al/Ti) has been used to

distinguish mafic and felsic source rocks (e.g., Hayashi et al., 1997; Huang et al., 2013).

The variability of Al/Ti observed in this study suggests that the detrital tracers

themselves, which are relied upon for remaining relatively constant through the time

interval of many sedimentary redox studies (at least on the scale of individual

stratigraphic units), have the potential to vary greatly themselves (Huang et al., 2013)
(Fig. 7). Given that Al/Ti ratios are likely to reflect the felsic to mafic composition of the

parent rock and that the composition of the parent can greatly impact the supply of redox-

sensitive trace metals (e.g., Cr is concentrated in mafic rocks), it might be possible for

provenance tracers to ‘correct’ detrital redox-sensitive trace element variability.

However, we find no relationship between Cr/Ti and Al/Ti (Fig. 8, r-squared value =

0.002). This not only suggests that the choice of a particular detrital tracer can impact the

result of a study, but also suggests that redox-sensitive trace element variability cannot be

easily accounted for with a provenance tracer.

Considering the lack of a strong relationship between Cr/Ti and Al/Ti, and given the

lack of clear redox-sensitive metal depletion in the topsoil, we suggest that surficial

variability is primarily driven by inherited variation, rather than the result of variable

surface weathering within the soil. The majority of the Earth’s weatherable surface area

(i.e., exposed continental area) is comprised of sedimentary rocks (Bluth and Kump,

1991), driving this inherited variability and adding complexity beyond what is observed

in igneous rocks. While, for example, Cr is primarily sourced from mafic rocks and it

may appear possible correct for this with a detrital tracer such as Ti, we find instead that

the variation within the weatherable sources is too great for this to be an effective method

of quantifying the detrital contribution.

4.6 Aluminum content

Lastly, we have compared each of our redox sensitive metal ratios for which we have

suggested new confidence intervals (Cr/Ti, V/Al, U/Th, Fe/Al) to aluminum content, and

then examined this data in relation to average global river suspended sediment values

from Viers et al., (2009) (Fig. 9). We find that for all ratios, increasing Al content results
in sustained variability within that ratio, and that this variability remains too high to

further constrain detrital flux ratios based on Al content. While our database only

investigates the fine sediment fraction (< 2mm), it is this fraction that is likely to be the

most consistent (e.g. larger grain sizes tend to be even more variable; e.g., Bouchez et al.,

2011). Further, the global averages fall within our suggested confidence intervals for all

four ratios (Fig. 9), providing a second line of evidence for our dataset being generally

representative of the detrital material being transported to the marine system. However, it

is apparent that the river averages tend toward higher ratios and Al concentrations than

the majority of our soil data. This is likely a result of sorption effects of dissolved

elements present in the riverine system, as well as the exclusion of material being

transported in the bed load. As such, we suggest our examination of soil data as opposed

to an extensive survey of riverine data provides a simpler window into establishing a

baseline range for detrital fluxes that is not complicated by additional factors involved in

riverine transport.

5. Future Directions

5.1 Recommendations

We suggest use of confidence intervals for Cr/Ti, V/Al, U/Th, and Fe/Al based on the

catchment resampled mean values and two times the standard deviation of these means,

which are summarized in Table 2. For U and V, the least variable ratios (U/Th and V/Al)

were chosen and are recommended for future studies. In some cases, these confidence

intervals (especially for U/Th) may provide a narrow enough constraint to elucidate

significant enrichments in some (likely Phanerozoic) studies, where very large

enrichments are likely to develop. In other cases, especially for Cr and V, the observed
variation in soils is so extensive that the confidence intervals define a very broad range

that will generally fail to distinguish authigenic enrichments, except for in the most

extreme cases of enrichments (e.g., Bakken Shale, Karma and Parslow, 1989). In these

cases, further analyses such as isotopic work will be required to definitively classify the

paleoredox conditions.

As shown in this study, the use of single crustal averages to distinguish authigenic

enrichments of most redox-sensitive metals does not account for the considerable crustal

heterogeneity in these systems. This leads to an oversimplification of these redox proxies,

most of which frequently lack the resolvability to independently distinguish redox driven

enrichments without additional analyses. These issues are most problematic in analyses

based on individual stratigraphic sections or basins (e.g., Li et al., 2015b; Zhang et al.,

2016). As data sets increase in size and global distribution, there is potential to highlight

redox-driven enrichments using a statistical approach. For example, to deal with detrital

Cr/Ti element variability, Cole et al. (2016) compared large datasets from several

formations from different time intervals with similar Cr/Ti ranges. The basic idea behind

this approach was that although a single formation may have an anomalous Cr/Ti ratio,

large datasets from many units capture the full variability of different crustal inputs, that

is, samples with high and low detrital Cr/Ti ratios.

While this study is based on a dataset restricted to the continental United States, we

suggest that this assessment of detrital variability represents a major step forward in

understanding the potential for lateral, large-scale variation. The U.S. also encompasses a

wide variety of exposed lithologies, weathering regimes, and climates, acting as a

reasonable microcosm of the global weatherable shell. As larger databases emerge, this

could be assessed with a similar study using a more expansive global dataset, however
considering the catchment scale variability observed here, we suggest that our

recommended confidence intervals provide at least a first approximation of expected

variation, and would certainly be a step forward from the use of a single cutoff value.

5.2 Applications of Isotope Systems

Iron can provide a success story for the future of redox-sensitive transition metals, as a

more thoroughly explored system than other transition metals discussed here. The use of

both Fe speciation and the Fe isotope system have proven very powerful tools to improve

the resolvability of the system beyond the more ambiguous enrichments (e.g., Poulton

and Canfield, 2005; Severmann et al., 2008; Sperling et al., 2015). Using a similar

framework of coupled techniques from other systems (Cr, U, V), it is conceivable that the

resolution of redox studies can be greatly enhanced. To add such resolvability to trace

metal redox proxies, we highlight the utility of emerging isotopic systems such as Cr and

V to distinguish authigenic enrichments, providing higher resolution constraints on local

marine redox. Further, isotopic approaches can prove particularly useful in identifying

potential shifts in the efficiency of chemical weathering (i.e. supply of oxidants), but if

taken a step further, this can also provide information about the subsequent impact on the

magnitude of the detrital flux. While our understanding of these systems is still

developing, the established basic framework can already function as a valuable new tool

to distinguish moderately reducing systems, and provide improved resolvability in redox

studies.

The Cr isotope system has the potential to provide new insights about weakly

reducing environments or those with vacillating redox conditions, as Cr is easily and

rapidly removed through a variety of reduction pathways, including nitrate reduction, and
sequestered in the sediment (Graham and Bouwer, 2010; Pettine et al., 1994; Rue et al.,

1997). The detrital Cr fraction appears in most cases to retain the unfractionated bulk

silicate earth value of δ53Cr = – 0.124 ± 0.101‰ (Gueguen et al., 2016; Schoenberg et al.,

2008; Wang et al., 2016b), however, in the case of a well-oxidized earth atmosphere,

authigenic Cr will have undergone oxidation during terrestrial weathering, and

subsequent reduction in the marine system, imparting a fractionated signal (Ellis et al.,

2002; Kotaś and Stasicka, 2000; Reinhard et al., 2014; Schauble et al., 2004), resulting in

a recognizable authigenic fraction. For example, in a hypothetical sample with a Cr/Ti

ratio of .025, let us consider two cases; 1) 100% of the Cr is detrital or 2) 50% of the Cr

is detrital, while the other 50% is derived from seawater. If we assume a seawater δ53Cr

value of 1.0 ‰, the sample from the first scenario would retain an unfractionated crustal

value, while the second would likely have a distinctly fractionated value of ~ 0.5 ‰,

representing the relative isotopic mass balance of the two components. Significantly, both

of these scenarios are likely based on our predicted range of detrital Cr/Ti ratios.

Acknowledging the very limited extent of Cr isotope variability in oxic sediments,

despite variation in Cr/Ti (Gueguen et al., 2016), a coupled isotope – enrichment

approach is likely to yield highly resolved information regarding redox conditions of

individual depositional environments that more poorly constrained enrichments alone are

unable to provide (Cole et al., 2016a).

Vanadium, similar to Cr, also has the potential to elucidate suboxic conditions (sensu

Scott et al., 2008), filling in the gradient of potential redox environments. It has been

suggested that V can be sourced from suboxic sediment (Hastings et al., 1996; Morford

and Emerson, 1999; Sahoo et al., 2012; Thomson et al., 2015), released along with Mn2+

(Morford et al., 2005) – thus a decrease in V could potentially signal a shift to a more
oxidized system. Additionally, V can be reduced twice – first from its oxidized state

V(V) to V(IV) under mildly reducing conditions and then removed from the sediment via

formation of ligands or adsorption (Emerson and Huested, 1991; Morford and Emerson,

1999; van der Sloot et al., 1985), and secondly to V(III) in the presence of H2S which can

be removed as oxides (V2O3), hydroxide V(OH)3 or incorporated into geoporphyrins

(Breit and Wanty, 1991; Wanty and Goldhaber, 1992). This two-step reduction of V

lends this system to tracking varying states of anoxia in the marine system resulting from

different solubilities of V phases. Unfortunately, because of the large potential range of

detrital V contribution, it is unlikely that such shifts can be recognized based on changes

in concentration alone. However, with the emergence of the V isotope system, though

still in incipient stages, there is potential that isotopic fractionations paired with elemental

concentrations can enhance our understanding of these intermediate redox environments

(Prytulak et al., 2013). Further it is also conceivable that with a large enough dataset, a

statistical analysis of V concentration changes through large time intervals, and the

relationship of these V concentrations with other redox-sensitive trace metals, that some

broad-scale shifts in suboxic area of the ocean can be recognized.

In sum, despite the current limitations of these nascent systems, major insights can

still be gained from simplistic applications of the Cr and V isotope systems. The coupled

approach described above has the capacity to provide high-resolution redox

interpretations regardless of fluctuations in detrital input or more moderately reducing

conditions and in spite of a poor understanding of the V and Cr global isotope mass

balances.

Uranium isotopes, though also still in the early stages, have been explored more

widely than the Cr and V systems (e.g., Andersen et al., 2014; Andersen et al., 2016;
Asael et al., 2013; Kendall et al., 2013; Kendall et al., 2015; Tissot and Dauphas, 2015;

Weyer et al., 2008). Uranium is primarily reduced within the sediment column (Algeo

and Maynard, 2004; McManus et al., 2005), and reduction under anoxic conditions

imparts large fractionations from seawater values. Thus, the areal extent of this sink can

impact the U isotopic value of seawater compared to the riverine input, the latter of which

is considered constant (De Corte, 2016). Using a quantitative mass balance framework,

these relationships can be used to reconstruct the areal extent of reducing marine

environments (e.g., Kendall et al., 2015; Tissot and Dauphas, 2015). However, detrital

corrections must be made in order to elucidate the isotopic value of the authigenic

component (e.g., Asael et al., 2013; Lu, 2016), yet the potential for variation of this term

is rarely considered. Use of the confidence intervals suggested in this study can thus

provide a means of assessing the potential variation in this term, the subsequent ‘error

bars’ on estimations of reducing ocean area.

6. Conclusions

In examining some of the most commonly used redox-sensitive transition metals and

a variety of detrital tracers in the soils of the continental US, we find pervasive and

laterally extensive heterogeneity in all systems that suggests the detrital contribution of

these elements in the rock record must be assessed with extreme caution. The prevalence

of detrital material with trace metal/detrital tracer ratios significantly different from those

of the average crust suggests that using crustal averages to estimate detrital metal fluxes

is a serious oversimplification. Recognition of surficial variability is necessary in

sedimentary redox studies, and must be considered especially carefully in local and

basinal contexts. We suggest new confidence intervals, based on catchment resampled

means for Cr/Ti, V/Al, U/Th, and Fe/Al to quantify the potential for detrital contribution

in these systems, which can help to elucidate enrichments clearly driven by redox

conditions even in the face of major surficial variability. In cases where enrichments

remain ambiguous, the application of emerging isotope systems can provide a means to

improve the resolvability of trace element enrichment redox proxies, and the coupled use

of these tools will be a major step forward in our ability to understand a spectrum of

redox environments.

Acknowledgements

The authors acknowledge funding from the Alternative Earths Team of the NASA

Astrobiology Institute. N.J.P. acknowledges funding from the Alfred P. Sloan

Foundation. The authors also gratefully acknowledge the detailed and thoughtful reviews

from Julien Bouchez and Tom Algeo, which improved the manuscript considerably.
References
Adams, J.A. and Weaver, C.E. (1958) Thorium-to-uranium ratios as indicators of
sedimentary processes: example of concept of geochemical facies. AAPG Bulletin
42, 387-430.

Algeo, T.J. (2004) Can marine anoxic events draw down the trace element inventory of
seawater? Geology 32, 1057-1060.

Algeo, T.J. and Maynard, J.B. (2004) Trace-element behavior and redox facies in core
shales of Upper Pennsylvanian Kansas-type cyclothems. Chemical Geology 206,
289-318.

Algeo, T.J. and Rowe, H. (2012) Paleoceanographic applications of trace-metal

concentration data. Chemical Geology 324–325, 6-18.

Anbar, A.D., Duan, Y., Lyons, T.W., Arnold, G.L., Kendall, B., Creaser, R.A., Kaufman,
A.J., Gordon, G.W., Scott, C., Garvin, J. and Buick, R. (2007) A Whiff of Oxygen
Before the Great Oxidation Event? Science 317, 1903-1906.

Andersen, M.B., Romaniello, S., Vance, D., Little, S.H., Herdman, R. and Lyons, T.W.
(2014) A modern framework for the interpretation of 238U/235U in studies of
ancient ocean redox. Earth and Planetary Science Letters 400, 184-194.

Andersen, M.B., Vance, D., Morford, J.L., Bura-Nakić, E., Breitenbach, S.F.M. and Och,
L. (2016) Closing in on the marine 238U/235U budget. Chemical Geology 420,
11-22.

Arnold, G.L., Anbar, A.D., Barling, J. and Lyons, T.W. (2004) Molybdenum Isotope
Evidence for Widespread Anoxia in Mid-Proterozoic Oceans. Science 304, 87-90.

Asael, D., Tissot, F.L.H., Reinhard, C.T., Rouxel, O., Dauphas, N., Lyons, T.W.,
Ponzevera, E., Liorzou, C. and Chéron, S. (2013) Coupled molybdenum, iron and
uranium stable isotopes as oceanic paleoredox proxies during the
Paleoproterozoic Shunga Event. Chemical Geology 362, 193-210.

Berner, R.A. (1970) Sedimentary pyrite formation. American Journal of Science 268, 1-
23.

Bluth, G.J. and Kump, L.R. (1991) Phanerozoic paleogeology. Am. J. Sci 291, 284-308.

Bouchez, J., Gaillardet, J., France-Lanord, C., Maurice, L., Dutra-Maia, P. (2011) Grain
size control of river suspended sediment geochemistry: Clues from Amazon River
depth profiles. Geochemistry, Geophysics, Geosystems 12.

Breit, G.N. and Wanty, R.B. (1991) Vanadium accumulation in carbonaceous rocks: A
review of geochemical controls during deposition and diagenesis. Chemical
Geology 91, 83-97.
Calvert, S.E. and Pedersen, T.F. (1993) Geochemistry of Recent oxic and anoxic marine
sediments: Implications for the geological record. Marine Geology 113, 67-88.

Clarke, F.W. (1889) The relative abundance of the chemical elements. Philosophical
Society of Washington Bulletin X 1, 131-142.

Cole, D.B., Planavsky, N.J., Gaines, R.R. and Li, C. (2016a) A Chromium isotope
perspective on Burgess Shale-type preservation, Geological Society of America
Annual meeting, Denver, CO.

Cole, D.B., Reinhard, C.T., Wang, X., Gueguen, B., Halverson, G.P., Gibson, T.,
Hodgskiss, M.S.W., McKenzie, N.R., Lyons, T.W. and Planavsky, N.J. (2016b) A
shale-hosted Cr isotope record of low atmospheric oxygen during the Proterozoic.
Geology.

Condie, K.C. (1993) Chemical composition and evolution of the upper continental crust:
Contrasting results from surface samples and shales. Chemical Geology 104, 1-
37.

Cruse, A.M. and Lyons, T.W. (2004) Trace metal records of regional paleoenvironmental
variability in Pennsylvanian (Upper Carboniferous) black shales. Chemical
Geology 206, 319-345.

Crusius, J., Calvert, S., Pedersen, T. and Sage, D. (1996) Rhenium and molybdenum
enrichments in sediments as indicators of oxic, suboxic and sulfidic conditions of
deposition. Earth and Planetary Science Letters 145, 65-78.

Dahl, T.W., Hammarlund, E.U., Anbar, A.D., Bond, D.P.G., Gill, B.C., Gordon, G.W.,
Knoll, A.H., Nielsen, A.T., Schovsbo, N.H. and Canfield, D.E. (2010) Devonian
rise in atmospheric oxygen correlated to the radiations of terrestrial plants and
large predatory fish. Proceedings of the National Academy of Sciences 107,
17911-17915.

De Corte, B.P. (2016) Uranium isotope ratios in modern and Precambrian soils,
Lawrence University.

Eade, K.E. and Fahrig, W.F. (1971) Geochemical evolutionary trends of continental
plates: A preliminary study of the Canadian Shield. Department of Energy, Mines
and Resources.

Ellis, A.S., Johnson, T.M. and Bullen, T.D. (2002) Chromium Isotopes and the Fate of
Hexavalent Chromium in the Environment. Science 295, 2060-2062.

Emerson, S.R. and Huested, S.S. (1991) Ocean anoxia and the concentrations of
molybdenum and vanadium in seawater. Marine Chemistry 34, 177-196.
Erickson, B.E. and Helz, G.R. (2000) Molybdenum(VI) speciation in sulfidic waters::
Stability and lability of thiomolybdates. Geochimica et Cosmochimica Acta 64,
1149-1158.

Frei, R., Gaucher, C., Poulton, S.W. and Canfield, D.E. (2009) Fluctuations in
Precambrian atmospheric oxygenation recorded by chromium isotopes. Nature
461, 250-253.

Frei, R., Gaucher, C., Stolper, D. and Canfield, D.E. (2013) Fluctuations in late
Neoproterozoic atmospheric oxidation — Cr isotope chemostratigraphy and iron
speciation of the late Ediacaran lower Arroyo del Soldado Group (Uruguay).
Gondwana Research 23, 797-811.

Gao, S., Luo, T., Zhang, B., Zhang, H., Han, Y., Zhao, Z., Hu, Y., (1998) Chemical
composition of the continental crust as revealed by studies in East China.
Geochemica et Cosomchimica Acta 62, 1959-1975.

Goldschmidt, V.M. (1937) The principles of distribution of chemical elements in

minerals and rocks. The seventh Hugo Müller Lecture, delivered before the
Chemical Society on March 17th, 1937. Journal of the Chemical Society
(Resumed), 655-673.

Graham, A.M. and Bouwer, E.J. (2010) Rates of Hexavalent Chromium Reduction in
Anoxic Estuarine Sediments: pH Effects and the Role of Acid Volatile Sulfides.
Environmental Science & Technology 44, 136-142.

GRDC (2016) The Global Data Runofff Center, 56068 Koblenz Germany.

Gueguen, B., Reinhard, C.T., Algeo, T.J., Peterson, L.C., Nielsen, S.G., Wang, X., Rowe,
H. and Planavsky, N.J. (2016) The chromium isotope composition of reducing
and oxic marine sediments. Geochimica et Cosmochimica Acta 184, 1-19.

Hastings, D.W., Emerson, S.R., Erez, J. and Nelson, B.K. (1996) Vanadium in
foraminiferal calcite: Evaluation of a method to determine paleo-seawater
vanadium concentrations. Geochimica et Cosmochimica Acta 60, 3701-3715.

Hayashi, K.-I., Fujisawa, H., Holland, H.D. and Ohmoto, H. (1997) Geochemistry of
∼1.9 Ga sedimentary rocks from northeastern Labrador, Canada. Geochimica et
Cosmochimica Acta 61, 4115-4137.

Helz, G.R., Miller, C.V., Charnock, J.M., Mosselmans, J.F.W., Pattrick, R.A.D., Garner,
C.D. and Vaughan, D.J. (1996) Mechanism of molybdenum removal from the sea
and its concentration in black shales: EXAFS evidence. Geochimica et
Cosmochimica Acta 60, 3631-3642.

Huang, J., Chu, X., Lyons, T.W., Planavsky, N.J. and Wen, H. (2013) A new look at
saponite formation and its implications for early animal records in the Ediacaran
of South China. Geobiology 11, 3-14.
Jones, B. and Manning, D.A.C. (1994) Comparison of geochemical indices used for the
interpretation of palaeoredox conditions in ancient mudstones. Chemical Geology
111, 111-129.

Kamber, B.S., Greig, A. and Collerson, K.D. (2005) A new estimate for the composition
of weathered young upper continental crust from alluvial sediments, Queensland,
Australia. Geochimica et Cosmochimica Acta 69, 1041-1058.

Karma, R. and Parslow, G. (1989) Sedimentology and geochemistry of the Bakken

Formation (Devonian-Mississippian) in southern Saskatchewan. Summary of
Investigations, 89-84.

Kendall, B., Brennecka, G.A., Weyer, S. and Anbar, A.D. (2013) Uranium isotope
fractionation suggests oxidative uranium mobilization at 2.50 Ga. Chemical
Geology 362, 105-114.

Kendall, B., Gordon, G.W., Poulton, S.W. and Anbar, A.D. (2011) Molybdenum isotope
constraints on the extent of late Paleoproterozoic ocean euxinia. Earth and
Planetary Science Letters 307, 450-460.

Kendall, B., Komiya, T., Lyons, T.W., Bates, S.M., Gordon, G.W., Romaniello, S.J.,
Jiang, G., Creaser, R.A., Xiao, S., McFadden, K., Sawaki, Y., Tahata, M., Shu,
D., Han, J., Li, Y., Chu, X. and Anbar, A.D. (2015) Uranium and molybdenum
isotope evidence for an episode of widespread ocean oxygenation during the late
Ediacaran Period. Geochimica et Cosmochimica Acta 156, 173-193.

Konhauser, K.O., Lalonde, S.V., Planavsky, N.J., Pecoits, E., Lyons, T.W., Mojzsis, S.J.,
Rouxel, O.J., Barley, M.E., Rosiere, C., Fralick, P.W., Kump, L.R. and Bekker,
A. (2011) Aerobic bacterial pyrite oxidation and acid rock drainage during the
Great Oxidation Event. Nature 478, 369-373.

Kotaś, J. and Stasicka, Z. (2000) Chromium occurrence in the environment and methods
of its speciation. Environmental Pollution 107, 263-283.

Li, C., Planavsky, N.J., Love, G.D., Reinhard, C.T., Hardisty, D., Feng, L., Bates, S.M.,
Huang, J., Zhang, Q., Chu, X. and Lyons, T.W. (2015a) Marine redox conditions
in the middle Proterozoic ocean and isotopic constraints on authigenic carbonate
formation: Insights from the Chuanlinggou Formation, Yanshan Basin, North
China. Geochimica et Cosmochimica Acta 150, 90-105.

Li, C., Planavsky, N.J., Shi, W., Zhang, Z., Zhou, C., Cheng, M., Tarhan, L.G., Luo, G.
and Xie, S. (2015b) Ediacaran Marine Redox Heterogeneity and Early Animal
Ecosystems. Scientific Reports 5, 17097.

Lu, X. (2016) Reconstruction of Local and Global Marine Redox Conditions During
Deposition of Late Ordovician and Early Silurian Organic-Rich Mudrocks in the
Siljan Ring District, Central Sweden.
Lyons, T.W. and Severmann, S. (2006) A critical look at iron paleoredox proxies: New
insights from modern euxinic marine basins. Geochimica et Cosmochimica Acta
70, 5698-5722.

Lyons, T.W., Werne, J.P., Hollander, D.J. and Murray, R.W. (2003) Contrasting sulfur
geochemistry and Fe/Al and Mo/Al ratios across the last oxic-to-anoxic transition
in the Cariaco Basin, Venezuela. Chemical Geology 195, 131-157.

McManus, J., Berelson, W.M., Klinkhammer, G.P., Hammond, D.E. and Holm, C.
(2005) Authigenic uranium: Relationship to oxygen penetration depth and organic
carbon rain. Geochimica et Cosmochimica Acta 69, 95-108.

Middleton, G.V. (1973) Johannes Walther's Law of the Correlation of Facies. Geological
Society of America Bulletin 84, 979-988.

Morford, J.L. and Emerson, S. (1999) The geochemistry of redox sensitive trace metals in
sediments. Geochimica et Cosmochimica Acta 63, 1735-1750.

Morford, J.L., Emerson, S.R., Breckel, E.J. and Kim, S.H. (2005) Diagenesis of
oxyanions (V, U, Re, and Mo) in pore waters and sediments from a continental
margin. Geochimica et Cosmochimica Acta 69, 5021-5032.

Morford, J.L., Russell, A.D. and Emerson, S. (2001) Trace metal evidence for changes in
the redox environment associated with the transition from terrigenous clay to
diatomaceous sediment, Saanich Inlet, BC. Marine Geology 174, 355-369.

Myrow, P.M. (1990) A New Graph for Understanding. Journal of Geological Education
38, 16.

Osterkamp, W. and Friedman, J. (2000) The disparity between extreme rainfall events
and rare floods— with emphasis on the semi-arid American West. Hydrological
Processes 14, 2817-2829.

Partin, C.A., Bekker, A., Planavsky, N.J., Scott, C.T., Gill, B.C., Li, C., Podkovyrov, V.,
Maslov, A., Konhauser, K.O., Lalonde, S.V., Love, G.D., Poulton, S.W. and
Lyons, T.W. (2013) Large-scale fluctuations in Precambrian atmospheric and
oceanic oxygen levels from the record of U in shales. Earth and Planetary Science
Letters 369–370, 284-293.

Pettine, M., Millero, F.J. and Passino, R. (1994) Reduction of chromium (VI) with
hydrogen sulfide in NaCl media. Marine Chemistry 46, 335-344.

Perkins, R.B., Piper, D.Z., and Mason, C.E. (2008) Trace-element budgets in the
Ohio/Sunbury shales of Kentucky: Constraints on ocean cirulation and primary
productivity in the Devonian-Mississippian Apalachian Basin. Palaeogeography,
Palaeoclimatology, Palaeoecology 265, 14-29.
Planavsky, N.J., Reinhard, C.T., Wang, X., Thomson, D., McGoldrick, P., Rainbird,
R.H., Johnson, T., Fischer, W.W. and Lyons, T.W. (2014) Low Mid-Proterozoic
atmospheric oxygen levels and the delayed rise of animals. Science 346, 635-638.

Poulton, S.W. and Canfield, D.E. (2005) Development of a sequential extraction

procedure for iron: implications for iron partitioning in continentally derived
particulates. Chemical Geology 214, 209-221.

Poulton, S.W. and Raiswell, R. (2002) The low-temperature geochemical cycle of iron:
From continental fluxes to marine sediment deposition. American Journal of
Science 302, 774-805.

Prytulak, J., Nielsen, S.G., Ionov, D.A., Halliday, A.N., Harvey, J., Kelley, K.A., Niu,
Y.L., Peate, D.W., Shimizu, K. and Sims, K.W.W. (2013) The stable vanadium
isotope composition of the mantle and mafic lavas. Earth and Planetary Science
Letters 365, 177-189.

Raiswell, R. and Canfield, D.E. (1998) Sources of iron for pyrite formation in marine
sediments. American Journal of Science 298, 219-245.

Raiswell, R., Reinhard, C.T., Derkowski, A., Owens, J., Bottrell, S.H., Anbar, A.D. and
Lyons, T.W. (2011) Formation of syngenetic and early diagenetic iron minerals in
the late Archean Mt. McRae Shale, Hamersley Basin, Australia: New insights on
the patterns, controls and paleoenvironmental implications of authigenic mineral
formation. Geochimica et Cosmochimica Acta 75, 1072-1087.

Reinhard, C.T., Planavsky, N.J., Robbins, L.J., Partin, C.A., Gill, B.C., Lalonde, S.V.,
Bekker, A., Konhauser, K.O. and Lyons, T.W. (2013) Proterozoic ocean redox
and biogeochemical stasis. Proceedings of the National Academy of Sciences 110,
5357-5362.

Reinhard, C.T., Planavsky, N.J., Wang, X., Fischer, W.W., Johnson, T.M. and Lyons,
T.W. (2014) The isotopic composition of authigenic chromium in anoxic marine
sediments: A case study from the Cariaco Basin. Earth and Planetary Science
Letters 407, 9-18.

Rudnick, R. and Gao, S. (2014) Composition of the Continental Crust, in: Holland, H.D.,
Turekian, K.K. (Eds.), Treatise on Geochemistry, Second ed. Elsevier, Oxford,
pp. 1-51.

Rue, E.L., Smith, G.J., Cutter, G.A. and Bruland, K.W. (1997) The response of trace
element redox couples to suboxic conditions in the water column. Deep Sea
Research Part I: Oceanographic Research Papers 44, 113-134.

Sahoo, S.K., Planavsky, N.J., Kendall, B., Wang, X., Shi, X., Scott, C., Anbar, A.D.,
Lyons, T.W. and Jiang, G. (2012) Ocean oxygenation in the wake of the
Marinoan glaciation. Nature 489, 546-549.
Schauble, E., Rossman, G.R. and Taylor Jr, H.P. (2004) Theoretical estimates of
equilibrium chromium-isotope fractionations. Chemical Geology 205, 99-114.

Schoenberg, R., Zink, S., Staubwasser, M. and von Blanckenburg, F. (2008) The stable
Cr isotope inventory of solid Earth reservoirs determined by double spike MC-
ICP-MS. Chemical Geology 249, 294-306.

Scott, C. and Lyons, T.W. (2012) Contrasting molybdenum cycling and isotopic
properties in euxinic versus non-euxinic sediments and sedimentary rocks:
Refining the paleoproxies. Chemical Geology 324–325, 19-27.

Scott, C., Lyons, T.W., Bekker, A., Shen, Y., Poulton, S.W., Chu, X. and Anbar, A.D.
(2008) Tracing the stepwise oxygenation of the Proterozoic ocean. Nature 452,
456-459.

Severmann, S., Lyons, T.W., Anbar, A., McManus, J. and Gordon, G. (2008) Modern
iron isotope perspective on the benthic iron shuttle and the redox evolution of
ancient oceans. Geology 36, 487-490.

Shaw, D., Reilly, G., Muysson, J., Pattenden, G. and Campbell, F. (1967) An estimate of
the chemical composition of the Canadian Precambrian Shield. Canadian Journal
of Earth Sciences 4, 829-853.

Shiller, A.M. and Mao, L. (2000) Dissolved vanadium in rivers: effects of silicate
weathering. Chemical Geology 165, 13-22.

Smith, D.B., Cannon, W.F., Woodruff, L.G., Solano, Federico and Ellefsen, K.J. (2014)
Geochemical and Mineralogical Maps for Soils of the Conterminous US: Open
File Report 2014-1082, in: USGS (Ed.).

Sperling, E.A., Wolock, C.J., Morgan, A.S., Gill, B.C., Kunzmann, M., Halverson, G.P.,
Macdonald, F.A., Knoll, A.H. and Johnston, D.T. (2015) Statistical analysis of
iron geochemical data suggests limited late Proterozoic oxygenation. Nature 523,
451-454.

Taylor, S.R. and McLennan, S.M. (1985) The continental crust: its composition and
evolution.

Thomson, D., Rainbird, R.H., Planavsky, N., Lyons, T.W. and Bekker, A. (2015)
Chemostratigraphy of the Shaler Supergroup, Victoria Island, NW Canada: A
record of ocean composition prior to the Cryogenian glaciations. Precambrian
Research 263, 232-245.

Tissot, F.L.H. and Dauphas, N. (2015) Uranium isotopic compositions of the crust and
ocean: Age corrections, U budget and global extent of modern anoxia.
Geochimica et Cosmochimica Acta 167, 113-143.
Tribovillard, N., Algeo, T.J., Lyons, T. and Riboulleau, A. (2006) Trace metals as
paleoredox and paleoproductivity proxies: An update. Chemical Geology 232, 12-
32.

Turekian, K.K. and Wedepohl, K.H. (1961) Distribution of the Elements in Some Major
Units of the Earth's Crust. Geological Society of America Bulletin 72, 175-192.

van der Sloot, H.A., Hoede, D., Wijkstra, J., Duinker, J.C. and Nolting, R.F. (1985)
Anionic species of V, As, Se, Mo, Sb, Te and W in the Scheldt and Rhine
estuaries and the Southern Bight (North Sea). Estuarine, Coastal and Shelf
Science 21, 633-651.

Van der Weijden, C.H. (2002) Pitfalls of normalization of marine geochemical data using
a common divisor. Marine Geology 184, 167-187.

Viers, J., Dupre, B., Gaillardet, J., (2009) Chemical composition of suspended sediments
in World Rivers: New insights from a new database. Science of the Total
Environment 407, 853-868.

Walther, J. (1894) Einleitung in die Geologie als Historische Wissenschaft.

Wang, X., Planavsky, N.J., Reinhard, C.T., Hein, J.R. and Johnson, T.M. (2016a) A
Cenozoic seawater redox record derived from 238U/235U in ferromanganese
crusts. American Journal of Science 316, 64-83.

Wang, X., Planavsky, N.J., Reinhard, C.T., Zou, H., Ague, J.J., Wu, Y., Gill, B.C.,
Schwarzenbach, E.M. and Peucker-Ehrenbrink, B. (2016b) Chromium isotope
fractionation during subduction-related metamorphism, black shale weathering,
and hydrothermal alteration. Chemical Geology 423, 19-33.

Wang, X., Reinhard, C.T., Planavsky, N.J., Owens, J.D., Lyons, T.W. and Johnson, T.M.
(2016c) Sedimentary chromium isotopic compositions across the Cretaceous
OAE2 at Demerara Rise Site 1258. Chemical Geology 429, 85-92.

Wanty, R.B. and Goldhaber, M.B. (1992) Thermodynamics and kinetics of reactions
involving vanadium in natural systems: Accumulation of vanadium in
sedimentary rocks. Geochimica et Cosmochimica Acta 56, 1471-1483.

Wedepohl, K.H. (1995) The composition of the continental crust. Geochimica et

Cosmochimica Acta 59, 1217-1232.

Wehrli, B. and Stumm, W. (1989) Vanadyl in natural waters: Adsorption and hydrolysis
promote oxygenation. Geochimica et Cosmochimica Acta 53, 69-77.

Weyer, S., Anbar, A.D., Gerdes, A., Gordon, G.W., Algeo, T.J. and Boyle, E.A. (2008)
Natural fractionation of 238U/235U. Geochimica et Cosmochimica Acta 72, 345-
359.
Wignall, P.B. and Myers, K.J. (1988) Interpreting benthic oxygen levels in mudrocks: A
new approach. Geology 16, 452-455.

Wille, M., Nebel, O., Van Kranendonk, M.J., Schoenberg, R., Kleinhanns, I.C. and
Ellwood, M.J. (2013) Mo–Cr isotope evidence for a reducing Archean
atmosphere in 3.46–2.76 Ga black shales from the Pilbara, Western Australia.
Chemical Geology 340, 68-76.

Wolman, M.G. and Miller, J.P. (1960) Magnitude and frequency of forces in geomorphic
processes. The Journal of Geology, 54-74.

Zhang, S., Wang, X., Wang, H., Bjerrum, C.J., Hammarlund, E.U., Costa, M.M.,
Connelly, J.N., Zhang, B., Su, J. and Canfield, D.E. (2016) Sufficient oxygen for
animal respiration 1,400 million years ago. Proceedings of the National Academy
of Sciences.

Figure Captions

Figure 1. Distribution of Cr/Al and Cr/Ti in topsoils from the continental US. A)
Distribution of Cr/Al raw data. B) Distribution of Cr/Ti raw data and C) the catchment
resampled distribution. Red dashed lines denote the mean of the raw or resampled
distribution respectively, while the black dashed line denotes the crustal mean value of
Rudnick and Gao, 2014. D and E) Raw data interpolated across the continental US for
Cr/Al and Cr/Ti respectively.

Figure 2. Distribution of Cr/Ti in topsoil (A), A horizon (B), and C horizon (C). Red
dashed lines denote the mean of the raw data, while the black dashed line denotes the
crustal mean value of Rudnick and Gao (2014). D) Distribution of the percentage change
between C horizon Cr/Ti and topsoil. E) Interpolated map of the data in (D).

Figure 3. Topsoil distributions (left) and interpolated maps (right) of V/Ti (A), V/Sc (B),
and V/Al (C) which shows the raw data (dark blue) and catchment resampled distribution
(light blue). Red dashed lines denote the mean of the raw data, while the black dashed
line denotes the crustal mean value of Rudnick and Gao (2014).

Figure 4. Distribution of V/Al in topsoil (A), A horizon (B), and C horizon (C). Red
dashed lines denote the mean of the raw data, while the black dashed line denotes the
crustal mean value of Rudnick and Gao (2014). D) Distribution of the percentage change
between C horizon V/Al and topsoil. E) Interpolated map of the data in (D).
Figure 5. Topsoil distributions (left) and interpolated maps (right) of U/Al (A) and U/Th
(B). Red dashed lines denote the mean of the raw data, while the black dashed line
denotes the crustal mean value of Rudnick and Gao (2014).

Figure 6. Distribution of Fe/Al in topsoils from the continental US. A) Distribution of

raw data and B) the catchment resampled distribution. Red dashed lines denote the mean
of the raw or resampled distribution respectively, while the black dashed line denotes the
crustal mean value of Rudnick and Gao (2014). C) Raw data interpolated across the
continental US.

Figure 7. Distribution of Al/Ti in topsoils from the continental US (left) and interpolated
data (right). Red dashed lines denote the mean of the raw or resampled distribution
respectively, while the black dashed line denotes the crustal mean value of Rudnick and
Gao ( 2014).

Figure 8. Al/Ti compared to Cr/Ti from topsoils across the continental US.

Figure 9. A) V/Al , B) Cr/Ti, C) U/Th, D) Fe/Al all compared to Al content. Grey circles
and error bars represent average global river suspended sediment element concentrations
and 2SD from Viers et al. (2009). Black dashed lines show the upper and lower limit of
our recommended confidence intervals while red dashed lines show the crustal means of
Rudnick and Gao (2014).
Table 1. Soil and crustal means

2 2 2 Upper
Elemental Topsoil A horizon C horizon
StdDev StdDev StdDev Continental Units
mean1 1 1
Ratio mean mean
(±) (±) (±) Crust2
V/Ti 229.28 246.47 252.80 ppm/wt%
V/Al 13.28 12.76 13.34 12.92 13.47 13.58 11.90 ppm/wt%
V/Sc 9.13 5.60 6.90 ppm/ppm
Cr/Ti 0.015 0.094 0.013 0.014 0.015 0.128 0.02 ppm/ppm
Cr/Al 9.015 23.050 11.29 ppm/wt%
U/Al 0.578 1.134 0.33 ppm/wt%
U/Th 0.285 0.369 0.26 ppm/ppm
Fe/Al 0.476 1.223 0.475 0.427 0.48 wt%/wt%
Al/Ti 19.03 17.64 21.24 wt%/wt%
1
Smith et al., (2014)
2
Rudnick and Gao, (2014)
Table 2. Recommended Confidence Intervals

Catchment Upper
Elemental 2 StdDev Confidence
resampled Continental
Ratio (±) Interval
mean Crust1
6.95 –
V/Al 13.13 6.18 19.31 11.90
Cr/Ti 0.015 0.015 0 – 0.030 0.02
0.180 –
U/Th 0.282 0.102 0.384 0.26
0.170 –
Fe/Al 0.472 0.302 0.773 0.48
1
Rudnick and Gao, (2014)
 A. D.
0.06
f [normalized]

0.04

0.02 >30

0.0
0 10 Cr/Al 20 30

0.08
B.
0.06
E.
0.04
f [normalized]

0.02
>0.06

0.08 C.

0.06

0.04

0.02

0.0 0.02 Cr/Ti 0.04 0.06

 A. Topsoil 0.08 D.
0.06
0.04
0.04
0.02
0
B. A Horizon -300 -200 -100 0 100
0.06 Cr/Ti (Topsoil relative to C horizon) (%)
f [normalized]

E.
0.04
0.02

C. C Horizon
0.06
0.04
0.02

0 0.01 0.02 0.03 0.04 0.05

Cr/Ti
 A. 0.06
f [normalized]

0.04

0.02

0 200 400 600

V/Ti

B. 0.15

0.1
f [normalized]

0.05

0 2 4 6 8 10 12 14 16 18
V/Sc

C. 0.15
0.1

0.05
f [normalized]

catchment
0.1 resamp.

0.05

0 5 10 15 20 25 30 35
V/Al
 A. Topsoil D.
0.1 0.1

0.05 0.05

0
B. A Horizon -80 -40 0 40 80
E. V/Al (C horizon relative to topsoil) (%)
f [normalized]

0.1

0.05

C. C Horizon
0.1

0.05

0 5 10 15 20 25 30 35 40
V/Al
 A.

0.10

0.06

0.02
f [normalized]

0 0.5 1.0 1.5 2.0

U/Al
B.
0.12

0.08

0.04

0 0.2 0.4 0.6 0.8

U/Th
 0.08 A. C.

0.06

0.04
f [normalized]

0.02

0.12 B.

0.08

0.04

0 0.5 1.0 1.5

Fe/Al
 0.06
0.05

0.04
f [normalized]

0.03
0.02

0.01

0 20 40 60 80
Al/Ti
 R2 = 0.002
80

60

Al/Ti
40

20

0 0.02 0.04 0.06 0.08 0.1

Cr/Ti
 0.15
A. B.

50

0.1
V/Al *104

Cr/T i
30

0.05

10

0
0 4 8 12 16 0 4 8 12 16
Al (wt %) Al (wt %)
1.5 C. D.
1.8

1.4
1

Fe/A l
U/Th

0.5
0.6

0.2

0
0 4 8 12 16 0 4 8 12 16
Al (wt %) Al (wt %)