Earth and Planetary Science Letters 459 (2017) 381–393

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

www.elsevier.com/locate/epsl

88
Enrichment of Sr in continental waters due to calcium carbonate
precipitation
Netta Shalev a,b,∗ , Ittai Gavrieli b , Ludwik Halicz b,c , Amir Sandler b , Mordechai Stein b ,
Boaz Lazar a
a
Institute of Earth Sciences, the Hebrew University of Jerusalem, Edmond J. Safra campus, 91904, Jerusalem, Israel
b
Geological Survey of Israel, 30 Malkhei Israel Street, 95501, Jerusalem, Israel
c
Faculty of Chemistry, Biological and Chemical Research Centre, University of Warsaw, Żwirki i Wigury 101, 02-089, Warsaw, Poland

a r t i c l e i n f o a b s t r a c t

Article history: δ 88/86 Sr data published over the last few years suggest that continental waters are enriched with 88
Sr
Received 27 April 2016 as compared to the rocks in their drainage basins. In an attempt to understand this phenomenon, this
Received in revised form 17 November 2016 study established the fractionation in the 88 Sr/86 Sr ratio during precipitation of continental carbonates
Accepted 23 November 2016
(i.e., carbonates precipitated on land from surface, pedogenic, or ground waters), and evaluated the
Available online 12 December 2016
Editor: M. Bickle
contribution of this process to the 88 Sr-enrichment in rivers. For this, stable and radiogenic Sr isotopes
(88 Sr, 87 Sr and 86 Sr) were measured in calcite samples and their precipitating waters collected in various
Keywords: continental environments, such as soil, cave, streams and groundwater.
strontium isotopes The results indicate that continental carbonates are 88 Sr-depleted relative to their precipitating waters,
δ 88/86 Sr placing them as one of the most 88 Sr-depleted reservoirs on earth. The average difference in δ 88/86 Sr
88
Sr enrichment values between waters and solid CaCO3 (tufas or speleothems) that they precipitate is carb-water =
isotope fractionation −0.218 ± 0.014h (1SD). An even larger fractionation (εcarb-water = −0.285 ± 0.02h) was measured
continental waters
in groundwater with particularly high carbonate-alkalinity and high carbonate precipitation rate that
continental carbonates
depleted ∼65% of the Sr in the groundwater, resulting in substantial 88 Sr-enrichment in the residual
dissolved Sr (δ 88/86 Sr = 0.656h). Results also suggest that pedogenic carbonate precipitation in soil
profile removes 50–85% of the Sr from the recharging soil–water, thereby increasing the δ 88/86 Sr value of
the soil-water from ∼ 0.18h to 0.3h–0.6h. Similar 88 Sr-enrichment was observed in drip water from
a karst cave.
A maximum removal flux of Sr into continental carbonates of about 20 Gmol(Sr) ·y−1 is required to yield
the reported 88 Sr-enrichment in global rivers (δ 88/86 Sr = 0.32h) relative to their rock sources when
using the fractionation factor derived in this study, carb-water = −0.218h, and the published δ 88/86 Sr
composition of marine carbonates of 0.16h. This surprisingly large flux requires that ∼40% of the
originally weathered Sr should co-precipitate with continental carbonates. This calculation may suggest
that CaCO3 precipitation is a significant mechanism in the continental Sr cycle. It is possible however
that other mechanisms such as Sr uptake by plants or incongruent weathering of silicates, could also
contribute to the riverine 88 Sr-enrichment. Alternatively, the average δ 88/86 Sr value of marine carbonate
rock sections undergoing weathering is higher than suggested in current literature. It is concluded that
additional δ 88/86 Sr data on marine carbonate rock sections should be collected to resolve the enigma of
the high value of riverine δ 88/86 Sr.
© 2016 Elsevier B.V. All rights reserved.

1. Introduction

The continental strontium cycle and its effect on the strontium

*Corresponding author at: Institute of Earth Sciences, the Hebrew University, Ed-
mond J. Safra campus, Jerusalem, Israel, 91904. Fax: +972 2 5662581.
E-mail addresses: netta.shalev@mail.huji.ac.il (N. Shalev), ittai.gavrieli@gsi.gov.il
(I. Gavrieli), ludwik@gsi.gov.il (L. Halicz), sandler@gsi.gov.il (A. Sandler),
motis@mail.huji.ac.il (M. Stein), boaz.lazar@mail.huji.ac.il (B. Lazar).
content and isotope composition of the oceans have traditionally
been evaluated through the 87 Sr/86 Sr isotope ratio of the major
rivers that drain the exposed continental terrains (e.g. Faure, 1986;
Veizer, 1989). Recently, with the advances in mass spectromet-

http://dx.doi.org/10.1016/j.epsl.2016.11.042
0012-821X/© 2016 Elsevier B.V. All rights reserved.
382 N. Shalev et al. / Earth and Planetary Science Letters 459 (2017) 381–393
Fig. 1. Major Sr fluxes and their published mean δ 88/86 Sr values (After: Krabbenhöft
et al., 2010; Moynier et al., 2010; Charlier et al., 2012; Vollstaedt et al., 2014; Pearce
et al., 2015).
ric analysis of the stable isotopes of strontium (δ 88/86 Sr), the
Sr continental–ocean budget was re-evaluated (e.g., Halicz et al.,
2008; Krabbenhöft et al., 2010; Wei et al., 2013; Chao et al., 2015;
Pearce et al., 2015). It was suggested that processes occurring on
land are accompanied by Sr isotope fractionation that affect the
δ 88/86 Sr of rivers and, in turn, that of the ocean (e.g., Chao et
al., 2015; Pearce et al., 2015). The modern ocean has a δ 88/86 Sr
value of 0.39h (e.g., Fietzke and Eisenhauer, 2006; Ohno and Hi-
rata, 2007; Halicz et al., 2008), which is higher than the modern
riverine discharge (0.32h; Fig. 1; Krabbenhöft et al., 2010; Pearce
et al., 2015). However, the rivers are 88 Sr-enriched relative to both
carbonate and silicate rock sources. While the δ 88/86 Sr value of ter-
restrial silicate rocks is estimated at 0.30h (Moynier et al., 2010;
Charlier et al., 2012), the mean δ 88/86 Sr value of carbonate rocks
exposed on the continents is largely uncertain. The flux-weighted
average δ 88/86 Sr value of modern biogenic marine carbonates and
the Phanerozoic mean marine carbonates, based on brachiopods
and belemnites record, are 0.21h and 0.16h, respectively, signif-
icantly lower than that of modern seawater (Halicz et al., 2008;
Krabbenhöft et al., 2010; Vollstaedt et al., 2014).
Based on riverine 87 Sr/86 Sr, the proportion of Sr derived from
weathering of carbonate and silicate rocks is estimated to be about
60:40, respectively (Pearce et al., 2015 and references therein).
Given this ratio and the δ 88/86 Sr compositions of carbonate and sil-
icate rocks quoted above (16h and 0.30h, respectively; Moynier
et al., 2010; Charlier et al., 2012; Vollstaedt et al., 2014), the
δ 88/86 Sr of the global rivers should have been ∼ 0.22h. How-
ever, this value is considerably lower than the measured river-
ine δ 88/86 Sr composition of 0.32h (Krabbenhöft et al., 2010;
Pearce et al., 2015), reinforcing the suggestion that weathering-
related processes fractionate Sr isotopes on the continents (Fig. 1).
Several mechanisms were proposed for the 88 Sr-enrichment of
continental waters (summarized in Chao et al., 2015): 1) Prefer-
ential adsorption of the lighter isotopes onto particles; 2) Bio-
logical fractionation during uptake of Sr (De Souza et al., 2010;
Andrews et al., 2016; Bullen and Chadwick, 2016); 3) Incongru-
ent weathering of silicate minerals, during which the heavy 88 Sr
isotope is preferentially released into solution (Halicz et al., 2008;
Wei et al., 2013; Chao et al., 2015; Stevenson et al., 2016); and
4) Precipitation of 88 Sr-depleted continental carbonates, which en-
riches the residual solution with 88 Sr (Chao et al., 2013, 2015).
Continental carbonates precipitate on land from surface, pedo-
genic, or ground waters (Deocampo, 2010) in a range of environ-
ments, including rivers (tufa), lakes, soil (pedogenic carbonates),
veins, caves and through local re-crystallization of marine car-
bonates by freshwater. The total amount of these carbonates is
not known, but elevated Sr/Ca ratios in Himalayan water were
explained by the precipitation of an unexpectedly large amount
of continental carbonate in the catchment area: 50–80% of the
originally weathered Ca (Galy et al., 1999; Jacobson et al., 2002;
Bickle et al., 2015). If this estimate is representative of other en-
vironments in which continental carbonates precipitate, then car-
bonate sink is huge, and may also be a sink for Sr. However,
measurements of δ 88/86 Sr values of such carbonates and their
associated waters are rare and inconclusive (Chao et al., 2013;
Shalev et al., 2013; Chao et al., 2015; Pearce et al., 2015).
The present study aims at providing new data on the impact
of precipitation of continental carbonates on the δ 88/86 Sr value of
continental waters. The stable and radiogenic Sr isotopic compo-
sitions of different types of continental carbonates and associated
waters in several environments, including streams, groundwaters
and soils, were investigated. The dual isotopic system of δ 88/86 Sr
and 87 Sr/86 Sr was used to determine the sources and evolution of
continental Sr for each environment, and to quantify the Sr iso-
tope fractionation. Finally, the implications of the new results for
the global continental Sr budget are addressed.
2. Materials and methods
2.1. Materials
Several continental carbonates and their precipitating waters
from Israel, and one such pair from France, were selected to
represent a range of continental carbonate types and environ-
ments, including groundwater-tufa, streams-tufa, desert springs-
tufa, speleothems and pedogenic carbonates. Sampling details are
presented in Table A.1 (Appendix A).
2.1.1. Stream waters and tufas
Several streams, draining mainly Mesozoic carbonate rocks,
were sampled (Fig. 2): the Banias and Dan streams are fed by
the Jurassic limestone aquifer of Mt. Hermon in northern Israel,
which is characterized by its highly developed karst system. Tufa
was sampled within the Banias stream in the form of a thin coat-
ing on bedload boulders, and within the Dan stream in the form of
a coating on a submerged metal pipe. The base flow at the David
stream, located in the Judean Desert, eastern Israel, is derived from
springs issuing from an Upper Cretaceous carbonate aquifer. Tufa,
up to few tens of centimeters thick, covers the rocks wherever
the water forms small waterfalls. A sample was collected from the
uppermost ∼2 cm tufa closest to the water stream. In addition,
the Huveaune River, Provence, France, draining the Jurassic karstic
limestones of the Sainte Baume Massif was sampled. CaCO3 precip-
itation, enhanced by the presence of Cyanobacteria, forms terraces
and local ponds along the river. Soft carbonate deposits that cover
the river bed were sampled, along with river water.
2.1.2. Drip water and speleothems
Drip water and its associated carbonates (stalagmites) were
sampled in the well-studied Soreq cave, central Israel (e.g. Bar-
Matthews et al., 1991; Fig. 2). The stalagmites were sampled
from their surface (∼2 mm), representing recent precipitation. The
10–40 m thick caprock above the cave consists of Cenomanian
dolostone, locally covered by ∼30 cm thick terra rossa soil, which
evolved mostly from settled dust (Yaalon, 1997). Soreq caprock and
Jerusalem dust were sampled as well.
 N. Shalev et al. / Earth and Planetary Science Letters 459 (2017) 381–393
Fig. 2. Location of sampling sites in Israel and France. Colors of circles denote
different types of environments: stream waters and tufas (blue), drip water and
speleothems (green), Hula Valley groundwater and carbonates (purple) and pedo-
genic carbonates (brown). (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article).
2.1.3. Hula Valley groundwater and tufa
A massive tufa cap precipitates on the outlet pipe of a leak-
ing abandoned gas exploration well, Hula Valley-1 (HV-1), in the
Hula Valley, located within the Dead Sea rift valley in northern
Israel (Fig. 2). Groundwater and tufa from this well and ground-
water from an additional sealed well, Gonen-6 (G-6), were sam-
pled. These wells are drilled into an artesian aquifer at a depth of
a few hundred meters, where water is associated with methane
(Nissenbaum and Magaritz, 1991). The origins of methane are
∼2500 m thick organic-rich sediments, which were deposited in
freshwater lakes and swamps from Mid-Miocene to Recent (Bein,
1986).
Water from G-6 well was sampled in 2011 during an acci-
dental sealing failure. Two samples were collected: the first was
immediately acidified and therefore preserves the original water
(sample G-6), whereas the second was stored and cooled with lit-
tle head space, allowing degassing of CO2 . This latter bottle was
opened in 2014 and the water and CaCO3 that precipitated in it
were analyzed (samples G-6-bottle and G-6-bottle-carb., respec-
tively).
2.1.4. Pedogenic carbonates
Two pedogenic calcic nodules were sampled from the lower
and upper parts of a soil profile near Noam ephemeral stream, cen-
tral Israel. The soil is a light brown clayey loam that developed on
Eocene chalk. The chalk bedrock was also sampled.
383
2.2. Samples preparation and analyses
Carbonate samples were homogenized and 50–500 mg aliquots
were dissolved in 0.5M acetic acid at 50 ◦ C. The solutions were
then evaporated to dryness, dissolved in concentrated HCl, evapo-
rated to dryness, dissolved in concentrated HNO3 and evaporated
again. Finally, the samples were dissolved in 8N HNO3 .
Water samples were filtered (0.45 μm) in the field and stored at
4 ◦ C until treated. Treatment included evaporation to dryness and
re-dissolution in 8N HNO3 .
Sr isotopic abundances were determined by MC-ICP-MS (Nu
Plasma-I, Nu Instruments, Wrexham, UK, fitted with the Aridus de-
solvating nebulizer) at the Geological Survey of Israel. The double-
spike technique was used for instrumental mass bias correction
(Shalev et al., 2013). Since this method requires measuring both
spiked and unspiked portions of a sample, two aliquots were
taken from each sample. The spiked portion was mixed with the
84
Sr–87 Sr spike to obtain a total Sr sample/spike ratio of 2.3.
The 84 Sr–87 Sr spike solution (84 Sr-0.4953, 86 Sr-0.0049, 87 Sr-0.4586
and 88 Sr-0.0412, total Sr concentration = 6.3 mg/l) was prepared
and calibrated in IFM-GEOMAR, Kiel, Germany (Krabbenhöft et al.,
2009).
A column containing Eichrom-Sr-Resin (50–100 mesh) was used
to separate Sr from the sample matrix. Following the loading of
the sample to the column, the matrix was rinsed with 8N HNO3 ,
after which Sr was eluted with 3.5 ml of 0.05N HNO3 . Overall Sr
recovery was >95%.
Major elements (Ca, Mg, Sr, K and Na) were determined using
ICP-OES (Optima 3300, Perkin–Elmer), with precision of ±5%. Five
mg/l Sc were added to all samples as an internal standard and the
results were quantified on a calibration curve produced for each
session. Accuracy was checked by running an international stan-
dard (IAPSO or USGS M-184-198) with each session.
2.3. Data reduction
The double-spike inversion was calculated using the Double-
Spike MATLAB ToolBox software (Rudge et al., 2009). The inputs
for this algorithm are the measured (instrumental biased) isotopic
compositions (84 Sr, 86 Sr, 87 Sr and 88 Sr) of the spiked and unspiked
portions of the sample, after correction for 86 Kr, 84 Kr and 87 Rb
isobaric interferences (Shalev et al., 2013). In addition, the exact
isotopic composition of the spike is needed. The algorithm yields
the isotopic composition of the unspiked portions of the sample
(84 Sr, 86 Sr, 87 Sr and 88 Sr), corrected for instrumental mass bias.
The 88 Sr/86 Sr ratio obtained from the MATLAB algorithm is re-
ported in the common permil notation relative to SRM987 stan-
dard (88 Sr/86 Sr = 8.375209; Nier, 1938):
 
(88 Sr/86 Sr)sample
δ 88/86 Sr[h] = − 1 × 1000 (1)
(88 Sr/86 Sr)SRM987
The δ 88/86 Sr long-term average of IAPSO seawater standard is
0.382(10)h (2SEM, n = 58), identical to the 0.386(5)h reported
earlier (Krabbenhöft et al., 2009). Each IAPSO replicate was mea-
sured several times during one measurement session and the
2SEM errors on the average δ 88/86 Sr of each session are presented
in Fig. 3. Note that all values from individual sessions are similar,
within error, to the long-term average (Fig. 3).
87
Sr/86 Sr ratios were determined following two different proce-
dures, the “traditional” procedure (87 Sr/86 Sr), and the double-spike
procedure that obtains the “true ratio” (87 Sr/86 Sr∗ , after Krabben-
höft et al., 2009, 2010; Table 1). The “traditional” procedure, which
is determined from the unspiked measurement, assumes that the
88
Sr/86 Sr ratio is constant and equal to 8.3752 (i.e., δ 88/86 Sr = 0h).
The instrumental mass discrimination for 87 Sr/86 Sr is corrected
384 N. Shalev et al. / Earth and Planetary Science Letters 459 (2017) 381–393
Fig. 3. Long term δ 88/86 Sr measurements of IAPSO seawater standard. Average of
each measurement-session is shown together with the long-term average value. Er-
ror bars are 2SEM. The number of measurements curried out in each session is
indicated. All results are identical within the long-term average ±2SEM range (yel-
low rectangle). (For interpretation of the references to color in this figure legend,
the reader is referred to the web version of this article).
based on this assumption, thereby effectively erasing the relatively
minor variation in the radiogenic 87 Sr/86 Sr ratio, which may arise
from natural fractionation processes. In contrast, the double spike
algorithm of Rudge et al. (2009) corrects all three isotopic ra-
tios (88 Sr/86 Sr, 87 Sr/86 Sr and 84 Sr/86 Sr), yielding the true ratios.
The long-term average of 87 Sr/86 Sr∗ of the IAPSO seawater stan-
dard obtained using this technique is 0.709296(8) (2SEM, n = 58),
within error of the 0.709312(9) reported by Krabbenhöft et al.
(2009). A validation procedure for the analytical process, which
considers the correspondence between the traditional 87 Sr/86 Sr
and 87 Sr/86 Sr∗ ratios, while taking into account the natural frac-
tionation in 87 Sr/86 Sr system, is presented in Appendix B. However,
the results presented hereafter are the traditional 87 Sr/86 Sr ratios,
thereby allowing comparison with previous data.
3. Results
The 87 Sr/86 Sr ratios of most samples, including the marine car-
bonate rocks and the carbonate fraction of settled dust samples,
lie within the range of 0.70727–0.70830 (Table 1; Fig. 4A). This
range is consistent with 87 Sr/86 Sr ratios of Mesozoic–Cenozoic ma-
rine carbonate rocks and the carbonate fractions of terra rossa soil
and loess in Israel (Haliva-Cohen et al., 2012). The silicate fraction
of the settled dust has a much higher 87 Sr/86 Sr ratio (0.71071; Ta-
ble 1), similar to ratios previously reported for silicate fractions
of dust in Israel (e.g., Haliva-Cohen et al., 2012). The Hula valley
samples with lower 87 Sr/86 Sr ratios (∼0.7062) are probably re-
lated to waters originating from the Neogene–Quaternary basalts,
(e.g., ∼0.7032; Stein and Hofmann, 1992) and basaltic soils at the
Golan Height and the Galilee Mountains bordering the Hula valley.
Springs that discharge the basaltic aquifers are characterized by
low 87 Sr/86 Sr ratios, contributing non-radiogenic Sr to the associ-
ated groundwater (e.g. 87 Sr/86 Sr ratios ∼0.7047, Spiro et al., 2011).
The carbonates and waters sampled during the current study
display a wide range of δ 88/86 Sr values (0.004h–0.656h; Table 1;
Fig. 4B). Within this range, the carbonates are relatively 88 Sr-
depleted (δ 88/86 Sr = 0.004h–0.118h), with an average of 0.073 ±
0.037h (1SD, N = 10). Four of the water samples (Soreq cave,
Dan stream, Huveaune river and David stream), which derive their
Sr from marine carbonate terrains of different ages, have similar
low δ 88/86 Sr values, averaged to 0.250 ± 0.026h (1SD). More 88 Sr-
enriched values were measured in the Banias stream (0.324h)
and in the Hula Valley groundwater (up to 0.656h). The δ 88/86 Sr
values of the four marine carbonate rocks – two Cenomanian dolo-
stone, Turonian limestone and Eocene chalk (0.204 − 0.376h) –
Fig. 4. Sr isotopic compositions of the studied materials. A) 87 Sr/86 Sr ratios. Errors
are within symbols (2SEM); B) δ 88/86 Sr values. Error bars are 2SEM. Continental car-
bonate (yellow) and their precipitating waters (blue) are presented wherever these
waters exist. Country carbonate rocks (brown) and dust carbonates (pale brown) are
also shown. (For interpretation of the references to color in this figure legend, the
reader is referred to the web version of this article).
are heavier than the Phanerozoic mean marine carbonates (0.16h;
Vollstaedt et al., 2014).
4. Discussion
4.1. Sr isotope fractionation during precipitation of continental
carbonates
Sr isotope fractionation during carbonate precipitation was
evaluated using the isotopic compositional differences between
pairs of precipitate and associated water (carb-water ):
carb-water = δcarbonate − δwater (2)
During such a fractionation process, the δ 88/86
Sr value varies but,
by definition, the 87 Sr/86 Sr ratio remains unchanged (see Ap-
pendix B for discussion of the actual fractionation that takes
place in the 87 Sr/86 Sr system). Hence, isotope fractionation in
the 88 Sr/86 Sr system is described by horizontal lines on a dual
87
Sr/86 Sr versus δ 88/86 Sr plot. To ensure the co-genetic relations
between the precipitate and water, only pairs that fell on the same
horizontal line (fractionation line) were considered. Materials that
derive their Sr from different sources are not likely to fall on the
same horizontal line because of the different radiogenic contribu-
tion of 87 Sr.
4.1.1. Fractionation between freshwaters and carbonates
Four out of five pairs of cave/stream waters and associated
carbonates fall on four different horizontal lines (Fig. 5A), in-
dicating the co-genetic relation of each distinct pair and their
different Sr sources. Still, all four pairs yield similar carb-water
values (average = −0.218 ± 0.014h; 1SD; Table 2), suggesting
that −0.218h may be considered as the 88 Sr-fractionation fac-
tor, ε (≈carb-water ), in the process of precipitation of carbon-
ates from continental waters. This further implies that the light
Table 1
Sample location, chemical composition and Sr isotopic compositions. 87
Sr/86 Sr∗ is the true ratio and 87
Sr/86 Sr is the “traditional” ratio.

Sample Description Location Mg Ca Sr Na K Sr/Ca δ 88/86 Sr 2se n 87

Sr/86 Sr∗ 2se n 87
Sr/86 Sr 2se n
(mM/M) (h) ·10−5 ·10−5

Continental carbonates (mg/g)

11-2-sg-151 Stalagmite Soreq cave, central Israel 4.7 394 0.1 n.d. n.d. 0.13 0.060 0.029 4 0.70820 1.9 4 0.70818 1.1 4

N. Shalev et al. / Earth and Planetary Science Letters 459 (2017) 381–393
5-3-sg-151 Stalagmite Soreq cave, central Israel 5.3 385 0.1 n.d. n.d. 0.13 0.068 0.031 4 0.70831 1.3 4 0.70829 0.8 4
Banias stream tufa-1 Tufa Banias stream, northern Israel 1.6 316 0.2 bdl 0.2 0.31 0.100 0.037 4 0.70730 3.2 4 0.70727 0.8 3
Dan stream tufa-1 Tufa Dan stream, northern Israel 1.1 76 0.03 bdl 0.3 0.20 0.062 0.050 4 0.70768 2.4 4 0.70766 0.8 3
Huveaune river tufa∗ Tufa Huveaune river, Provence, France 0.3 240 0.1 bdl bdl 0.14 0.004 0.068 4 0.70787 5.6 4 0.70786 3.8 3
David stream tufa Tufa David stream, eastern Israel 10.8 345 0.64 0.17 0.05 0.85 0.024 0.010 4 0.70796 3.7 4 0.70796 2.8 4
HV-1 tufa Tufa Hula Valley, northern Israel 12.7 309 0.8 0.4 bdl 1.24 0.118 0.043 4 0.70625 3.2 4 0.70623 1.4 3
G-6-bottle-carb.∗ Carbonate Hula Valley, northern Israel n.d. n.d. n.d. n.d. n.d. 0.74 0.095 0.056 7 0.70623 4.9 7 0.70619 3.0 4
Noam-3 carb. Pedogenic carbonate Noam eph. stream, central Israel n.d. n.d. n.d. n.d. n.d. 0.91 0.096 0.030 10 0.70830 3.1 10 0.70824 3.8 3
Noam-8 calcrete Pedogenic carbonate Noam eph. stream, central Israel 5.2 378 0.34 0.1 0.3 0.41 0.105 0.045 8 0.70819 2.8 8 0.70814 4.3 3

Continental freshwaters and seawater standard (mg/l)

11-2-151 Drip water Soreq cave, central Israel 30.4 73 0.1 n.d. n.d. 0.69 0.259 0.022 4 0.70825 2.0 4 0.70816 1.7 4
Banias stream water-1 Stream water Banias stream, northern Israel 12.5 72 0.3 6.5 2.4 1.81 0.324 0.026 5 0.70743 4.7 5 0.70730 2.1 3
Dan stream water-1 Stream water Dan stream, northern Israel 4.7 63 0.1 2.8 1.1 0.55 0.263 0.019 4 0.70748 4.5 4 0.70736 0.8 2
Huveaune river water River water Huveaune river, Provence, France 6.2 76 0.1 3.0 0.2 0.83 0.236 0.046 5 0.70794 2.9 5 0.70784 1.3 3
David stream water Stream water David stream, eastern Israel 29 50 0.6 46.5 2.7 5.24 0.242 0.039 4 0.70803 3.1 4 0.70795 2.8 4
HV-1 groundwater Groundwater Hula Valley, northern Israel 290 43 0.5 137 15 5.57 0.656 0.042 5 0.70648 3.5 5 0.70624 6.3 2
G-6 Groundwater Hula Valley, northern Israel 187 458 1.3 311 16 1.30 0.390 0.020 7 0.70635 8.0 7 0.70621 0.9 4
G-6-bottle Groundwater, bottled for 3 years Hula Valley, northern Israel 180 275 0.9 344 17 1.56 0.475 0.018 7 0.70638 1.2 7 0.70619 1.4 4
IAPSO Seawater standard 1310 466 7.7 7.58 0.382 0.010 58 0.70930 7.9 56 0.70915 0.9 26

Marine carbonate rocks and dust-carbonate (mg/g)

Soreq caprock-2-dol Cenomanian dolostone Soreq cave, caprock, central Israel 112 242 0.16 n.d. n.d. 0.07 0.346 0.010 4 0.70761 1.5 4 0.70749 0.9 4
Chalk NOAM-9 Eocene chalk Noam stream, central Israel 4.7 258 0.84 12.4 1.0 0.33 0.204 0.021 8 0.70790 1.9 8 0.70784 0.6 4
Limestone SA-568B-flo. Turonian limestone Jerusalem, central Israel 1.1 337 0.08 14.7 0.8 0.02 0.340 0.025 8 0.70768 2.0 8 0.70757 2.8 4
Dolostone SA-574A-flo. Cenomanian dolostone Meron Mt., northern Israel 89.2 182 0.08 8.1 2.3 0.04 0.376 0.026 8 0.70790 3.0 8 0.70776 1.9 4
Dust JER-14-carb.∗∗ Carbonate fraction of settling dust, 2014 Jerusalem, central Israel 24.2 318 0.8 4.9 4.0 0.24 0.182 0.034 8 0.70836 2.5 7 0.70830 2.3 8
Dust JER-14-sil.∗∗ Silicate fraction of settling dust, 2014 Jerusalem, central Israel 14.8 14.1 0.1 4.8 12.9 0.183 0.033 0.71076 2.3 7 0.71071 3.2 8

Concentrations in the solids are in mg in the carbonate fraction per gram bulk dry solid, except where concentrations were normalized to gram wet bulk solids (∗) or to gram of the carbonate/silicate fraction (∗∗). Isotopic
measurements were done on the carbonate fraction. n.d.: not determined. bdl: below detection limit.

385
386 N. Shalev et al. / Earth and Planetary Science Letters 459 (2017) 381–393

Table 2
Stable Sr isotopic differences (carb-water ) and Sr partition coefficients (K d ) calculated from the current data.

Solid sample Water sample Description  Error,  Kd Error, K d

Freshwater
11-2-sg-151 11-2-151 Stalagmite-Drip water −0.199 0.036 0.182 0.012
Banias-tufa-1 Banias-water-1 Tufa-Stream water −0.224 0.045 0.173 0.007
Huveaune-tufa Huveaune-water Tufa-River water −0.232 0.082 0.168 0.012
David-stream-tufa David-stream-water Tufa-Stream-water −0.218 0.040 0.155 0.006
Average stream waters and tufas (1SD) −0.218 0.014 0.169 0.011

Reduced groundwater
HV-1-tufa HV-1-water Tufa-Groundwater −0.538 0.060 0.222 0.011
G-6-bottle-carb G-6 Tufa-Groundwater −0.295 0.059 0.570
G-6-bottle-carb G-6-bottle Tufa-Groundwater −0.380 0.045 0.474 0.014
Average groundwater and tufas (1SD) −0.405 0.123 0.422 0.180
Rayleigh distillation calibrated values −0.285 0.020 0.45 0.02

Pedogenic carbonate
Noam-3-carb. Not available Pedogenic carbonate-soil water <−0.085
Noam-8-calc. Not available Pedogenic carbonate-soil water <−0.085

isotope, 86 Sr, is preferentially incorporated into the solid phase.

This behavior is similar to Ca and Mg isotope incorporation in
travertine and speleothems (Tipper et al., 2006; Galy et al., 2002;
Immenhauser et al., 2010).
The carb-water of Sr isotopes was experimentally shown by
Böhm et al. (2012) to reflect the kinetics of the reaction and, there-
fore, to vary as a function of the partition coefficient of strontium
between solution and calcite (K d ):

(Sr/Ca)carbonate
Kd = (3)
(Sr/Ca)solution
The calculated partition coefficients of the four carbonate-water
pairs range between 0.155 and 0.182 (Table 2). Following Böhm
et al. (2012), these K d values should have been accompanied by a
slightly more negative carb-water (between −0.29h and −0.25h)
than those measured in the present study. This difference may sug-
gest that the discrimination against the heavy isotopes is slightly
smaller in continental carbonates than in seawater under similar
K d or be due to additional kinetic factors that are beyond the
scope of the present study.

4.1.2. Pedogenic carbonates

Strontium in the Noam soil profile has two potential sources:
settled dust and Eocene chalk bedrock (Yaalon, 1997). 87 Sr/86 Sr ra-
tios in the pedogenic carbonates of this profile are similar to the
ratio measured in the dust, whereas the δ 88/86 Sr values are sig-
nificantly lower than in the dust and the bedrock (Fig. 5B). While
this isotope difference between the pedogenic carbonates and their
lithogenic sources indicates isotope fractionation, the composition
of the precipitating waters (for the carb-water ) is not available and
must be estimated. The silicate fraction of the dust is less soluble
than the carbonate fraction and is characterized by low Sr concen- Fig. 5. Sr isotope compositions of continental carbonates and their Sr sources on a
87
tration, high 87 Sr/86 Sr ratio and a similar δ 88/86 Sr value (Table 1). Sr/86 Sr versus δ 88/86 Sr plot. Dashed red lines: horizontal fractionation lines. Error
bars are 2SEM. A) Sr isotope fractionation between freshwaters (blue) and carbon-
Hence, its strontium contribution to the pedogenic carbonates is
ates (yellow); B) Sr isotope compositions of pedogenic carbonates and potential
likely small and will not affect the δ 88/86 Sr value of the pedogenic carbonate sources. Yellow: pedogenic carbonates, orange: carbonate dust, brown:
carbonates (although it could affect its 87 Sr/86 Sr). Accordingly, only bedrock, dashed green line: mixing line of Sr sources. (For interpretation of the ref-
the carbonate sources are considered in the following calculation erences to color in this figure legend, the reader is referred to the web version of
(Fig. 5B). this article).

The mixing line between the carbonate fraction of the dust and
the bedrock is the locus of all possible initial Sr isotope composi-
tions of water infiltrating through the Noam soil profile, before the
onset of secondary carbonate precipitation. The initial composition
of the precipitating water, before the onset of the pedogenic car-
bonate precipitation, is thus given by the intersection between the
horizontal line and the mixing line joining the potential sources
(Fig. 5B). The isotope fractionation occurring during the precipi-
tation of pedogenic carbonate in this soil must be accompanied
by a change in the isotopic composition of the precipitating wa-
ter along the horizontal fractionation line, whereby the remaining
Sr becomes more 88 Sr-enriched. Accordingly, the isotopic composi-
tion of the water from which the carbonate precipitated must have
been heavier or identical to this initial composition, depending on
the amount of precipitated carbonate. Thus, the resulting isotopic
difference between the initial δ 88/86 Sr value of the water and the
sampled carbonates, carb-initial ≈ −0.085h, is a minimum value
 N. Shalev et al. / Earth and Planetary Science Letters 459 (2017) 381–393
for the isotopic fractionation. The difference between the carbonate
and its instantaneous precipitating water may be more negative if
the water has previously precipitated carbonate, resulting in more
isotopically enriched residual Sr in the water.
Assuming that the Sr fractionation factor in the pedogenic
carbonate is similar to that of the tufa/speleothem-water pairs
(carb-water = −0.218h), the 88 Sr-enrichement in the water and
the fraction of Sr precipitated in pedogenic carbonates (1 − f Sr )
prior to the formation of the sampled carbonate can be estimated
by applying a Rayleigh distillation equation:
(ε /1000)
δwater = (δ0 + 1000) · f Sr − 1000
where f Sr is the remaining fraction of Sr in the water, ε is the
enrichment factor (ε ≈ ), δ0 is initial water composition as de-
duced from the intersection between the MDF line and the mix-
ing line (∼0.18h; Fig. 5B), and δwater , the δ 88/86 Sr values of the
precipitating water, is calculated from the δ 88/86 Sr of the product
sampled carbonates. These carbonates may be the instantaneous
product, the cumulative product, or a mixture of both. Therefore
the δ 88/86 Sr values of the products were calculated for both ex-
treme scenarios, respectively:
δinst-prod = δwater + ε
ε ln f Sr
δcum-prod = δ0 + ε ln f Sr −
1 − f Sr
where δinst-prod and δcum-prod are the δ 88/86 Sr values of an instan-
taneous carbonate precipitate and a cumulative precipitate, respec-
tively. Taking the measured δ 88/86 Sr value of the pedogenic carbon-
ates (∼0.10h) as either δinst-prod or δcum-prod , the δ 88/86 Sr value of
the precipitating water, as calculated from equations (4)–(6), is be-
tween 0.3h and 0.6h, while the residual Sr ( f Sr ) is 15% to 50%.
Hence, soil waters leave 50–85% of their original Sr (1 − f Sr ) in
the soil profile. This substantial removal of Sr is consistent with
the elevated Sr/Ca ratios in the pedogenic calcites relative to their
Sr sources (0.41–0.91 and 0.24–0.33 mM/M, respectively, Table 1).
Since K d < 1 (Table 2), such high fraction of Sr removal requires
also that nearly all the Ca (>90%) is precipitated as pedogenic cal-
cite. This condition may be met in the semi-arid climate of Noam
site during the dry season, when soil water is subjected to massive
evaporation. This further implies that the remaining soil water at
the end of the dry season has a significantly elevated δ 88/86 Sr value
and Sr/Ca ratio. During the wet season this water is flushed into
the groundwater reservoir. Therefore, soils are a potential signifi-
cant 88 Sr-enrichment medium for infiltrating water that returns to
the continental reservoirs (e.g. groundwater). Below, we show that
in other environments with extreme conditions of carbonate pre-
cipitation, the enrichment of groundwater in 88 Sr may also be very
high.
4.1.3. Fractionation between groundwater with high alkalinity and
carbonates
The Hula groundwater is characterized by extremely high alka-
linity (>100 meq/l in G-6-bottle). Massive Ca-carbonate precipita-
tion from the Hula groundwater (Fig. 6A), enhanced by degassing
of CO2 , is reflected in the chemical compositions of the three wa-
ter samples collected from this unique system (Fig. 6B). Ca and
Sr concentrations are highest in the sample from the sealed well
(G-6), intermediate in the bottled sample with small headspace
(G-6-bottle), and lowest in the leaking water from well HV-1.
Due to the substantial depletion in Sr concentration (leaking wa-
ter from well HV-1 contains only ∼35% of the Sr relative to G-6)
and the consequent increase in δ 88/86 Sr values of the water (up
to 0.656h), the carbonates and waters cannot be considered as
instantaneous carbonate-water pairs. However, the similarities in
387
Mg, K and Na (Fig. 6B), and the fact that all samples fall on the
same horizontal fractionation line (Fig. 6C), suggest that the wa-
ters have the same origin and differ only by the amounts of CO2
lost and Ca-carbonate precipitated.
Thus, the three water samples can be treated as representing
the progressive impact of carbonate precipitation on a Rayleigh-
distillation plot (Fig. 6D). In the following calculation, sample G-6,
which has the highest Sr concentration, is taken to represent the
parent groundwater composition, even though it could have al-
ready precipitated some Ca-carbonate in the subsurface.
The 88 Sr-isotope enrichment factor, ε (≈), was determined by
(4) fitting a Rayleigh distillation curve (equation (4)) to the measured
isotopic compositions of the water samples. The best-fit curve
(blue curve in Fig. 6D) yields ε = −0.285 ± 0.02h (error denotes
the range of ε values which yield curves that fall within the er-
ror bars of the measured water samples). This fractionation factor
is more negative than those found in tufas and speleothems (Ta-
ble 2), as well as in most marine carbonates (e.g., Krabbenhöft et
al., 2010; Vollstaedt et al., 2014). Similar large fractionations were
previously measured in some coccolithophores (biological marine
calcite), which were cultured at relatively high temperature that
increased their growth rate and K d (Stevenson et al., 2014), and
in a sample of inorganic calcite synthesized at high growth rate
(5)
(Böhm et al., 2012). Thus, our results corroborate previous sug-
gestions that precipitation kinetics impact the isotope enrichment
(6)
factor, and that enhanced Ca-carbonate precipitation is accompa-
nied by increased Sr isotopic fractionation.
Using ε = −0.285h and equations (5) and (6), the δ 88/86 Sr
curves of the instantaneous product (δinst-prod ) and the cumulative
product (δcum-prod ) were plotted (Fig. 6D) together with the two
samples of carbonate precipitates. The results suggest that the car-
bonate from bottle of G-6 (G-6-bottle-carb) is a cumulative product
that precipitated from the water after sampling and following de-
gassing. In contrast, the tufa sampled from the massive carbonate
coating on HV-1 well pipe (Fig. 6A) is not a direct product of the
sampled water but contains several layers that, over the years, pre-
cipitated from the leaking groundwater. Its isotopic composition
suggests that it precipitated from water with higher f Sr than the
present leaking water (Fig. 6D).
The Sr partition coefficient (K d , equation (3)) between the
groundwater and its carbonate precipitate was determined using
the Rayleigh distillation equation for the Sr/Ca ratio:
( K −1 )
(Sr/Ca)solution = (Sr/Ca)solution-initial · f Ca d (7)
The best fit of equation (7) for the measured data was found to be
with K d = 0.45 ± 0.02. This value is significantly higher than that
found in other natural calcite and is another indicator of the high
precipitation rate from the Hula groundwater system.
4.2. Impact of precipitation of continental carbonate on δ 88/86 Sr of
continental waters
The δ 88/86 Sr values of the continental waters (>0.236h) that
drain marine carbonate terrains are heavier than the estimated
mean δ 88/86 Sr of marine carbonates (0.16h, Vollstaedt et al., 2014)
and the “unfractionated” δ 88/86 Sr of global rivers (0.22h; dis-
cussed above, section 1). The contribution of silicate-derived Sr is
likely minimal in such terrains, particularly given the much lower
concentration of Sr in silicates. The 88 Sr may be enriched dur-
ing precipitation of continental carbonates such as those discussed
above. Some examples of such enrichment processes are discussed
below.
The sources of strontium at the Soreq cave are well constrained
and include the dolostone caprock and the settled atmospheric
dust which is a major component of the terra rossa soil above the
388 N. Shalev et al. / Earth and Planetary Science Letters 459 (2017) 381–393

Fig. 6. Hula Valley tufa and groundwater. A) Photo of the massive carbonate coating on HV-1 well pipe; B) Major cations concentrations in the Hula groundwater; C) Sr isotope
compositions on a 87 Sr/86 Sr versus δ 88/86 Sr plot. Yellow: carbonates, blue: groundwater, dashed red line: horizontal fractionation line. Error bars are 2SEM; D) δ 88/86 Sr versus
the remaining fraction of Sr in the precipitating water. The blue line is a Rayleigh distillation curve fitted to the Hula groundwater samples (blue diamonds). The theoretical
instantaneous and cumulative carbonate precipitates are denoted by the yellow dashed and solid lines, respectively. The carbonate samples (yellow triangles) are located
based on the f Sr of their associated water, although they are not true instantaneous water–precipitate pairs. Arrow on carbonate of HV-1 points towards more 88 Sr-depleted
water from which it must have precipitated. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article).

cave (e.g. Sandler et al., 2015). The 87 Sr/86 Sr ratios indicate that
only a small amount of Sr is derived from the silicate fraction of
the dust, as is the case for the pedogenic carbonates (section 4.1.2
above). Therefore, only the carbonate sources for Sr are discussed.
The original composition of the drip water (δ0 ) is represented by
the intersection of the mixing line between these sources (Fig. 7)
with the horizontal line. The slightly higher δ 88/86 Sr of the sampled
drip water suggests an earlier phase of carbonate precipitation.
Applying equation (4) with the following parameters: 1) ε ≈
carb-water = −0.218h (Table 2); 2) δ0 – the initial δ 88/86 Sr value
as deduced from Fig. 7; and 3) δwater – the δ 88/86 Sr value of the
sampled drip water, indicates that the fraction of Sr that precipi-
tated as calcite during the infiltration path prior to reaching the
cave (1 − f Sr ) is ∼0.2. i.e., about 20% of the original Sr in the
water precipitated before reaching the dripping site. Such a signif-
icant Sr removal requires removal of ∼70% of the Ca during calcite
precipitation, which is consistent with the elevated Sr/Ca ratio of
the dripping water (0.7 mM/M) relative to its Sr sources (0.24 and
0.07 mM/M in dust and dolostone, respectively, Table 1) and is
lower than the Sr removal calculated in the soil profile (50–85%;
section 4.1.2 above).
The Sr-depleted-88 Sr-enriched groundwater that leaks from the
HV-1 well (δ 88/86 Sr = 0.656 ± 0.042h) is among the heaviest wa-
ters that have been reported so far in the literature. This excep-
tionally high value is attributed to two factors: 1) the high frac-
tion of Ca-carbonate precipitated (1 − f Ca ), accompanied by high
Sr removal (1 − f Sr ); and 2) a large isotopic fractionation dic-
tated by the high precipitation rate. The highest published δ 88/86 Sr
values (up to 0.82h) were measured in mud volcano fluids in Tai-
wan (Chao et al., 2013), which are similarly depleted in Sr and
Ca due to massive Ca-carbonate precipitation associated with the
degassing of CO2 from exceptionally highly alkaline groundwater.
While the Hula groundwater derives its high alkalinity from the
decomposition of organic matter, the alkalinity in the mud vol-
 N. Shalev et al. / Earth and Planetary Science Letters 459 (2017) 381–393
Fig. 7. Sr isotope compositions of the Soreq cave system on a 87 Sr/86 Sr versus
δ 88/86 Sr plot. Yellow: speleothems, orange: dust, brown: caprock, dashed red line:
fractionation line, dashed green line: mixing line of Sr sources. Error bars are 2SEM.
(For interpretation of the references to color in this figure legend, the reader is re-
ferred to the web version of this article).
cano fluids is derived by inorganic water–rock interactions (Chao
et al., 2013). These two cases emphasize the significant role of
carbonate-alkalinity, whether organic or inorganic, in determining
aquatic δ 88/86 Sr value.
The above examples, and the calculated δ 88/86 Sr of the infil-
trating soil water (0.3h to 0.6h, see section 4.1.2), emphasize
the important role that precipitation of continental carbonates has
in determining the δ 88/86 Sr value of continental water on a local
scale. The possible impact of continental carbonates precipitation
on the δ 88/86 Sr value of the global riverine flux to the ocean is dis-
cussed below.
4.3. Impact of continental calcium carbonate precipitation on δ 88/86 Sr
of the global riverine flux
The average δ 88/86 Sr value of 31 samples of continental carbon-
ates measured in the current and previous studies (Shalev et al.,
2013; Chao et al., 2015; Fruchter et al., 2016) is 0.076 ± 0.053h.
This places continental carbonates as one of the most 88 Sr-depleted
reservoirs on Earth (Fig. 8A). The current study demonstrates that
precipitation of continental carbonates enriches continental water
with 88 Sr on local scale, but the question is whether this mech-
anism can explain the global 88 Sr-enrichment in rivers flowing to
the oceans relative to their rock sources (Krabbenhöft et al., 2010;
Pearce et al., 2015).
If the rivers’ 88 Sr-enrichment is assumed to be due only to
the preferential sequestration of 86 Sr in the continental carbon-
ate reservoir, the Sr removal to this reservoir can be globally
estimated. In the calculation presented below, the following as-
sumptions are made: 1) riverine Sr discharging to the oceans
has δ 88/86 Srrivers = 0.32h (Krabbenhöft et al., 2010; Pearce et
al., 2015); 2) the 88 Sr/86 Sr isotope enrichment factor during
continental carbonate precipitation is ε = −0.218h, as mea-
sured in this research; and 3) the average value of silicate rocks
is δ 88/86 Srsil-rocks = 0.30h, that of marine carbonate rocks is
δ 88/86 Srcarb-rocks = 0.16h, and their Sr contributions to global
rivers is 40:60, respectively (Moynier et al., 2010; Charlier et
al., 2012; Vollstaedt et al., 2014; Pearce et al., 2015). Thus, the
initial “unfractionated” isotopic composition of continental water
(δ 88/86 Srunfract ) is 0.22h.
In addition to the existing natural variability of all of these pa-
rameters (see discussion in Appendix C), there is a significant un-
certainty in the choice of the initial stable-isotopic value of Sr de-
rived from weathering of marine carbonates (δ 88/86 Srcarb-rocks ). The
389
mass balance calculation presented by others and adopted here as-
sumes δ 88/86 Srcarb-rocks = 0.16h for the marine carbonate source
rocks (after Vollstaedt et al., 2014). However, as discussed below,
the more representative δ 88/86 Srcarb-rocks value is probably more
88
Sr-enriched. This assumed value for δ 88/86 Srcarb-rocks (0.16h)
was calculated by Vollstaedt et al. (2014) based on the Phanero-
zoic mean of well-selected brachiopods and belemnites (chosen
for reconstructing seawater values), whereas the mean δ 88/86 Sr of
bulk marine carbonate sediments is not well constrained. Based
on modern biogenic marine carbonates, Krabbenhöft et al. (2010)
estimated the modern marine carbonate flux-weighted average
value to be 0.21h. However, in the present study we measured
higher δ 88/86 Sr values in four exposed marine carbonate rocks:
two dolostones (Upper Cretaceous), one limestone (Upper Creta-
ceous) and one chalk (Eocene), with an average of 0.32 ± 0.08h
(1SD; Table 1 and Fig. 8A). In addition, three of the four car-
bonate matrix samples from the geological record measured by
Vollstaedt et al. (2014) are also significantly heavier than bra-
chiopods or belemnites from the same sample (with averages of
0.30h versus 0.12h, respectively). Given this implied uncertainty
in δ 88/86 Srcarb-rocks , its impact on the riverine 88 Sr-enrichment was
tested.
Inserting the assumed values in equation (4) implies that only
62% of the initial weathered Sr remains in the rivers that reach
the oceans (i.e., 38% of the Sr is removed). However, if as sug-
gested above, δ 88/86 Srcarb-rocks is > 0.16h, the fraction of Sr re-
moved, 1 − f Sr , is smaller accordingly (Fig. 8B). The uncertainty
in the fraction of precipitated Sr, 1 − f Sr , which stems from nat-
ural variability of the parameters, is about ±0.08 (Appendix C).
Palmer and Edmond (1989) estimated that the riverine Sr flux
to the ocean is ∼33.3 Gmol(Sr) ·y−1 . Based on this value and
δ 88/86 Srcarb-rocks = 0.16h, the maximum total Sr weathering flux
would be ∼53 Gmol(Sr) ·y−1 , of which ∼20 Gmol(Sr) ·y−1 is de-
posited as continental carbonates (sink). If correct, this continental
Sr sink is huge, and is equivalent to about 10% of the marine Sr
sink (174 Gmol(Sr) ·y−1 ; Krabbenhöft et al., 2010).
Estimates of the total annual continental carbonate deposition
are rare due to the diversity of carbonate types (e.g., pedogenic car-
bonates, cave deposits, rivers tufa, lakes carbonates and secondary
carbonates in rocks). Based on water chemical compositions and
radiogenic Sr ratios, Bickle et al. (2015) explained the elevated
Sr/Ca in downstream Himalayan water by the precipitation of sur-
prisingly large amount of continental carbonates in the catchment
area (50–80% of the originally weathered Ca, 0.2 ≤ f Ca ≤ 0.5). This
conclusion is in accordance with field observations and Ca and
Mg isotopes compositions of the Himalayan water (English et al.,
2000; Tipper et al., 2006, 2008). Based on the Sr partition co-
efficient found in the present study for carbonates (K d = 0.169;
section 4.1.1; Table 2), this implies that 11–24% of the original
weathered Sr is removed to these continental carbonates (1 − f Sr ):
f Sr = ( f Ca ) K d (8)
Even if such extensive Sr removal (1 − f Sr = 11–24%) into continen-
tal carbonates reservoirs occurs on a global scale, a process that
cannot be ruled out, it will elevate the δ 88/86 Sr of global rivers by
only up to 0.06h. For this process to account for the entire 88 Sr-
enrichment in the riverine Sr, the original unfractionated waters
(δ 88/86 Srunfract ) should be 0.26–0.29h (and not the 0.22h as-
sumed earlier). To attain this value, δ 88/86 Srcarb-rocks should be be-
tween 0.23h and 0.29h (Fig. 8B). Considering the measured val-
ues of the marine carbonate rocks, the higher δ 88/86 Srcarb-rocks val-
ues suggested here seem reasonable. It is therefore suggested that
the average δ 88/86 Sr value of exposed marine carbonate rocks un-
dergoing weathering is higher than previously suggested (0.16h).
However, other fractionation processes that operate to enrich the
riverine stable Sr isotope composition, such as biological uptake of
390 N. Shalev et al. / Earth and Planetary Science Letters 459 (2017) 381–393

our data indicates that the continental carbonates comprise one

Fig. 8. Global continental Sr budget: reservoirs and fluxes. A) Measured δ 88/86 Sr
values of continental Sr reservoirs. Data from current study denoted by black x
symbols; Colored symbols are published data from: Ohno and Hirata, 2007; Halicz
et al., 2008; Krabbenhöft et al., 2010; Moynier et al., 2010; De Souza et al., 2010;
Charlier et al., 2012; Raddatz et al., 2013; Chao et al., 2013; Shalev et al., 2013;
Wei et al., 2013; Neymark et al., 2014; Stevenson et al., 2014; Vollstaedt et al., 2014;
Chao et al., 2015; Pearce et al., 2015; Voigt et al., 2015; Bullen and Chadwick, 2016.
Black lines are estimated average of the reservoirs: silicates 0.30h (Moynier et al.,
2010; Charlier et al., 2012); Marine carbonate Phanerozoic mean 0.16h (Vollstaedt
et al., 2014) and weighted average of modern carbonates 0.21h (Krabbenhöft
et al., 2010); Riverine flux weighted average 0.32h (Krabbenhöft et al., 2010;
Pearce et al., 2015); Continental carbonates average 0.076h (current study; Shalev
et al., 2013; Chao et al., 2015; Fruchter et al., 2016); B) Sensitivity tests for the
fraction of Sr removed to the fractionated reservoir on the continents. The tested
initial compositions (δ 88/86 Srunfract ) are shown on the lower horizontal axis and
were calculated from δ 88/86 Srsil-rocks = 0.30h for silicate rocks and different val-
ues for carbonate rocks (δ 88/86 Srcarb-rocks , presented in the upper horizontal axis).
The black line shows the results obtained with ε = −0.218h. The blue dashed
lines indicate different estimations for δ 88/86 Srcarb-rocks (0.16h from Vollstaedt et
al., 2014; 0.21h from Krabbenhöft et al., 2010; and 0.32h from this study).
Sr (De Souza et al., 2010; Andrews et al., 2016; Bullen and Chad-
wick, 2016), incongruent weathering of silicate minerals (Wei et
al., 2013; Chao et al., 2015; Stevenson et al., 2016) or adsorption
of the lighter isotopes onto particles (Chao et al., 2015), cannot
be ruled out. It is therefore concluded that the limiting factor for
studying the continental stable Sr isotope budget is the small num-
ber of measurements of bulk exposed marine carbonate rocks.
5. Summary
of the most 88 Sr-depleted reservoirs on earth, with an average
value of δ 88/86 Sr = 0.076 ± 0.053h. The isotopic differences be-
tween waters and tufas or speleothems have a relatively narrow
range (carb-water = −0.218 ± 0.014h ≈ ε ). Based on this isotopic
fractionation and the δ 88/86 Sr values of pedogenic carbonates and
their Sr sources, we suggest that the associated soil water that
precipitated these carbonates has lost 50–85% of its Sr to the pe-
dogenic carbonate, leaving the remaining soil water 88 Sr-enriched
with δ 88/86 Sr values between 0.3h and 0.6h. An even larger iso-
tope fractionation was measured in tufa precipitated from a high
carbonate-alkalinity groundwater (ε = −0.285 ± 0.02h, Hula Val-
ley), possibly reflecting an enhanced kinetic effect related to the
rapid precipitation of the tufa. Thus, our study clearly shows that
precipitation of carbonates from waters on the continents may
cause a significant increase in the stable Sr isotopic ratio of the
continental waters.
Given the estimate of the stable Sr isotopic fractionation de-
termined here, and the assumed composition of the silicate and
carbonate source rocks and their relative Sr contribution to the
global rivers, the maximum global Sr sink in continental carbon-
ates is calculated to be ∼20 Gmol(Sr) ·y−1 (38% of the originally
weathered Sr). Thus, extensive continental carbonates precipita-
tion (11–24% of the weathered Sr), such as inferred for the Hi-
malayan rivers (Bickle et al., 2015), cannot be ruled out, and
may elevate the δ 88/86 Sr of global rivers by up to 0.06h. How-
ever, even this high fraction of Sr sequestration in continental
carbonates cannot explain the 88 Sr-enrichment of global rivers
(δ 88/86 Sr = 0.32h) relative to the assumed weathered composi-
tion (δ 88/86 Sr = 0.22h). It is therefore suggested that the δ 88/86 Sr
value of exposed marine carbonate rocks undergoing weathering is
higher than previously suggested (δ 88/86 Sr = 0.16h) implying that
the global weathered stable Sr isotopic composition is >0.22h.
The large uncertainty in this value is thus the limiting factor in
the quantification of the continental Sr reservoir and isotopic bud-
get.
Acknowledgements
We thank N. Teplyakov, Y. Zakun and the GSI geochemical
division staff for their help in the laboratory work. We thank
Y. Burstyn for sharing his knowledge of the Soreq cave and for
help in the field. A. Babad and F. Gelman provided samples from
the Hula Valley. We also thank the staff of the Israel Nature and
Parks Authority for helping with stream sampling. We gratefully
acknowledge C. Pearce, B. Peucker-Ehrenbrink and M.J. Bickle (ed-
itor) for their constructive comments that considerably improved
earlier version of this manuscript. This study was funded by: DFG
German–Israeli–Palestinian trilateral project TRION (Ei272/30-1);
the Israeli Ministry of Science, Technology and Space (an Eshkol
scholarship to NS); and Dalia and Dan Maydan Fellowship (to
NS).
Appendix A. Sampling details
Table A.1
Sampling details.

Water/solid sample Sampling date Latitude Longitude Temp. pH

This study evaluates the role of precipitation of calcium carbon- (◦ C)

ates in continental environments in determining the δ 88/86 Sr val- David stream water 12/05/2015 31.470 35.390
HV-1 groundwater 21/07/2013 33.107 35.612
ues of continental waters, which are generally more 88 Sr-enriched
G-6 24/11/2011 33.112 35.622
than their source rocks. Banias stream water-1 30/09/2013 33.236 35.678 17.2 8.7
For this, we analyzed the δ 88/86 Sr values and 87 Sr/86 Sr ratios Dan stream water-1 30/09/2013 33.243 35.646 16.2 8.0
of calcium carbonates and their precipitating waters from various 11-2-151 06/03/2013 31.756 35.024 19.6 8.2
continental environments (e.g., springs deposits, tufas, pedogenic Huveaune river water 12/06/2013 43.345 5.762
Noam 04/10/2011 31.562 34.804
carbonates and cave-speleothems). Combined with other studies,
 N. Shalev et al. / Earth and Planetary Science Letters 459 (2017) 381–393 391

Appendix B. 87 Sr/86 Sr∗ vs. traditional 87 Sr/86 Srtrad :

a measurement verification procedure

The radiogenic 87 Sr/86 Sr ratio was determined following two

different procedures, the “traditional” procedure, to obtain the
“traditional ratio” (87 Sr/86 Srtrad ), and the double-spike procedure,
to obtain the “true ratio” (87 Sr/86 Sr∗ , after Krabbenhöft et al., 2009,
2010). Both procedures assume that the instrumental fractionation
is proportional to the relative mass differences between the iso-
topes and hence will follow the slope of mass-dependent fraction-
ation (MDF) (Fig. B.1A). When using the δ notation (equation (1)
in the main text) for both 87 Sr/86 Sr and 88 Sr/86 Sr ratios (relative to
SRM987 standard: 87 Sr/86 Sr = 0.710240 and 88 Sr/86 Sr = 8.375209;
Nier, 1938), the MDF line will follow a linear function with a slope
close to 0.5 (0.504–0.507; after Young et al., 2002). This line in-
tersects the δ 87/86 Sr axis at b and not at the origin because of the
radiogenic contribution of 87 Sr:

δ 87/86 Sr = 0.5 · δ 88/86 Sr + b (B.1)

The common notation for the ratio that includes the radiogenic
87
Sr is 87/86 R (87/86 R = 87 Sr/86 Sr ratio). Converting the δ 87/86 Sr no-
tation to the R notation, using equation (1), yields a new linear
function:
 87/86 
87/86 R SRM987 · 0.5 b · 87/86 R SRM987
R= · δ 88/86 Sr +
1000 1000
87/86
+ R SRM987 (B.2)
87 86
Given that Sr/ Sr of SRM987 is 0.710240 (Nier, 1938), equa- Fig. B.1. Comparison of 87 Sr/86 Sr∗ and 87 Sr/86 Srtrad . A) Schematic illustration of
tion (B.2) can be rewritten to yield: the true, the measured and the “traditional” 87 Sr/86 Sr ratios on 87 Sr/86 Sr versus
δ 88/86 Sr diagram. The mass-dependent fractionation (MDF) line connecting these
87/86
R = 0.00036 · δ 88/86 Sr + b(R) (B.3) three points is indicated by red solid and dashed lines; B) Comparison between the
“traditional” 87 Sr/86 Sr ratios obtained by the “traditional” technique and the same
87
87 Sr/86 Sr ratios (b(R) values) obtained by applying the MDF linear equation (equa-
where b(R) is the Y intercept on a Sr/86 Sr vs. δ 88/86 Sr plot
tion (B.3)) with the 87 Sr/86 Sr∗ and δ 88/86 Sr values obtained from the double-spike
(Fig. B.1A) and is equal to: algorithm for each sample. The blue line indicates the 1:1 proportion between the
values obtained in both procedures. (For interpretation of the references to color in
b · 87/86 R SRM987
b(R) = + 87/86 R SRM987 (B.4) this figure legend, the reader is referred to the web version of this article).
1000
calculate b(R) for each sample. The b(R) values of all samples cal-
(b is the Y intercept on a δ 87/86 Sr vs. δ 88/86 Sr plot, equation (B.1)). culated by this method were found to be identical, within error,
The “traditional” ratio (87 Sr/86 Srtrad ) assumes that the 88 Sr/86 Sr to the 87 Sr/86 Srtrad ratios obtained by the “traditional” procedure
ratio in nature is constant and equal to the 88 Sr/86 Sr ratio in (Fig. B.1B), thereby reinforcing the results. Note that, theoretically,
SRM987 (i.e., δ 88/86 Sr = 0h) and the instrumental (as well as nat- 87
Sr/86 Srtrad and b(R) are identical, but, in practice, the former
ural) mass discrimination for 87 Sr/86 Sr is corrected based on this was determined by a single measurement, whereas in the present
assumption. Therefore, by definition, the intersection point, b(R) , of study the latter was calculated from an additional measurement of
all MDF lines (equation (B.3)) should be equal to the “traditional” the spiked sample, and using the calculation approach presented
value (87 Sr/86 Srtrad ; Fig. B.1A). Accordingly, the MDF equation can above.
be rewritten as:
87/86 Appendix C. Sensitivity tests of global estimated f Sr
R = 0.00036 · δ 88/86 Sr + 87/86 R trad (B.5)
Using equation (B.3) and the δ 88/86 Sr and 87 Sr/86 Sr∗ values ob- A calculation of the global fraction of Sr removed to the conti-
tained with the double spike technique, it is now possible to nental carbonate reservoir out of the initial weathered Sr (1 − f Sr )

Table C.1
Sensitivity tests on the estimated 1 − f Sr (0.38) based on the uncertainty in the input parameters. Each uncertainty indicates the published natural and analytical variability
of the chosen parameter, unless otherwise mentioned.

Parameter Value Uncertainty Range of f Sr due to uncertainty Sr removal % (1 − f Sr ) · 100

Global riverine δ 88/86 a
Sr 0.32h 0.005h 0.61–0.64 36–39
Isotope enrichment factor during −0.218h 0.014h 0.60–0.64 36–40
continental carbonate precipitation, ε
Riverine Sr derived from silicate rocksb 40% 20% 0.55–0.71 29–45
δ 88/86 Sr of silicate rocksc 0.30h 0.07h 0.55–0.71 29–45
δ 88/86 Sr of carbonate rocksd 0.16h 0.02h 0.59–0.66 34–41

a b
From: Krabbenhöft et al. (2010) and Pearce et al. (2015); Uncertainty is based on Gaillardet et al. (1999) and Pearce et al. (2015), but in order to account for all
possible values, we considered a larger range of values; c From: Moynier et al. (2010) and Charlier et al. (2012); d Secular variability of selected carbonates from:
Vollstaedt et al. (2014) and Pearce et al. (2015). The uncertainty in the δ 88/86 Sr of carbonate rocks, which stems from the type of carbonate measured, is further discussed in
the text.
392 N. Shalev et al. / Earth and Planetary Science Letters 459 (2017) 381–393
is presented in section 4.3, yielding 1 − f Sr = 38%. However, there
are natural variabilities on the parameters used in the calculation,
and the sensitivity tests conducted on these suggest that 1 − f Sr
ranges between 29–45% (Table C.1) implying an uncertainty of
about ±8% on this value.
Appendix D. Supplementary material
Supplementary material related to this article can be found on-
line at http://dx.doi.org/10.1016/j.epsl.2016.11.042.
References
Andrews, M.G., Jacobson, A.D., Lehn, G.O., Horton, T.W., Craw, D., 2016. Radio-
genic and stable Sr isotope ratios (87 Sr/86 Sr, δ 88/86 Sr) as tracers of riverine
cation sources and biogeochemical cycling in the Milford Sound region of Fiord-
land, New Zealand. Geochim. Cosmochim. Acta 173, 284–303. http://dx.doi.org/
10.1016/j.gca.2015.10.005.
Bar-Matthews, M., Matthews, A., Ayalon, A., 1991. Environmental controls of
speleothem mineralogy in a karstic dolomitic terrain (Soreq Cave, Israel).
J. Geol. 99 (2), 189–207. http://dx.doi.org/10.1086/629484.
Bein, A., 1986. Early evolution and transformation of organic matter in the active
continental Jordan Rift Valley. Org. Geochem. 10 (4–6), 751–757. http://dx.doi.
org/10.1016/S0146-6380(86)80012-2.
Bickle, M.J., Tipper, E., Galy, A., Chapman, H., Harris, N., 2015. On discrimination
between carbonate and silicate inputs to Himalayan rivers. Am. J. Sci. 315 (2),
120–166. http://dx.doi.org/10.2475/02.2015.02.
Böhm, F., Eisenhauer, A., Tang, J., Dietzel, M., Krabbenhöft, A., Kisakürek, B., Horn,
C., 2012. Strontium isotope fractionation of planktic foraminifera and inorganic
calcite. Geochim. Cosmochim. Acta. http://dx.doi.org/10.1016/j.gca.2012.04.038.
Bullen, T., Chadwick, O., 2016. Ca, Sr and Ba stable isotopes reveal the fate of soil
nutrients along a tropical climosequence in Hawaii. Chem. Geol. 422, 25–45.
http://dx.doi.org/10.1016/j.chemgeo.2015.12.008.
Chao, H.C., You, C.F., Liu, H.C., Chung, C.H., 2015. Evidence for stable Sr isotope frac-
tionation by silicate weathering in a small sedimentary watershed in southwest-
ern Taiwan. Geochim. Cosmochim. Acta 165, 324–341. http://dx.doi.org/10.1016/
j.gca.2015.06.006.
Chao, H.C., You, C.F., Liu, H.C., Chung, C.H., 2013. The origin and migration of mud
volcano fluids in Taiwan: evidence from hydrogen, oxygen, and strontium iso-
topic compositions. Geochim. Cosmochim. Acta 114, 29–51. http://dx.doi.org/
10.1016/j.gca.2013.03.035.
Charlier, B.L.A., Nowell, G.M., Parkinson, I.J., Kelley, S.P., Pearson, D.G., Burton,
K.W., 2012. High temperature strontium stable isotope behaviour in the early
solar system and planetary bodies. Earth Planet. Sci. Lett. 329–330, 31–40.
http://dx.doi.org/10.1016/j.epsl.2012.02.008.
Deocampo, D.M., 2010. The geochemistry of continental carbonates. In: Alonso-
Zarza, A.M., Tanner, L.H. (Eds.), Developments in Sedimentology, vol. 62. Elsevier,
pp. 1–59. Chapter 1.
English, N.B., Quade, J., DeCelles, P.G., Garzione, C.N., 2000. Geologic control of Sr
and major element chemistry in Himalayan Rivers, Nepal. Geochim. Cosmochim.
Acta 64 (15), 2549–2566. http://dx.doi.org/10.1016/S0016-7037(00)00379-3.
Faure, G., 1986. Principles of Isotope Geology. Wiley/The University of Michigan.
Fietzke, J., Eisenhauer, A., 2006. Determination of temperature-dependent stable
strontium isotope (88 Sr/86 Sr) fractionation via bracketing standard MC-ICP-MS.
Geochem. Geophys. Geosyst. 7 (8). http://dx.doi.org/10.1029/2006GC001243.
Fruchter, N., Lazar, B., Nishri, A., Almogi-Labin, A., Eisenhauer, A., Stein, M., 2016.
88
Sr/86 Sr fractionation and calcite accumulation rate in the Sea of Galilee.
Geochim. Cosmochim. Acta. Submitted.
Gaillardet, J., Dupré, B., Louvat, P., Allègre, C.J., 1999. Global silicate weathering
and CO2 consumption rates deduced from the chemistry of large rivers. Chem.
Geol. 159, 3–30.
Galy, A., France-Lanord, C., Derry, L.A., 1999. The strontium isotopic budget of
Himalayan Rivers in Nepal and Bangladesh. Geochim. Cosmochim. Acta 63
(13–14), 1905–1925. http://dx.doi.org/10.1016/S0016-7037(99)00081-2.
Galy, A., Bar-Matthews, M., Halicz, L., O’Nions, R.K., 2002. Mg isotopic composition
of carbonate: insight from speleothem formation. Earth Planet. Sci. Lett. 201 (1),
105–115. http://dx.doi.org/10.1016/S0012-821X(02)00675-1.
Halicz, L., Segal, I., Fruchter, N., Stein, M., Lazar, B., 2008. Strontium stable isotopes
fractionate in the soil environments? Earth Planet. Sci. Lett. 272 (1–2), 406–411.
http://dx.doi.org/10.1016/j.epsl.2008.05.005.
Haliva-Cohen, A., Stein, M., Goldstein, S.L., Sandler, A., Starinsky, A., 2012. Sources
and transport routes of fine detritus material to the Late Quaternary Dead Sea
basin. Quat. Sci. Rev. 50, 55–70. http://dx.doi.org/10.1016/j.quascirev.2012.06.
014.
Immenhauser, a., Buhl, D., Richter, D., Niedermayr, a., Riechelmann, D., Dietzel, M.,
Schulte, U., 2010. Magnesium-isotope fractionation during low-Mg calcite pre-
cipitation in a limestone cave – field study and experiments. Geochim. Cos-
mochim. Acta 74 (15), 4346–4364. http://dx.doi.org/10.1016/j.gca.2010.05.006.
Jacobson, A.D., Blum, J.D., Walter, L.M., 2002. Reconciling the elemental and Sr iso-
tope composition of Himalayan weathering fluxes: insights from the carbonate
geochemistry of stream waters. Geochim. Cosmochim. Acta 66 (19), 3417–3429.
http://dx.doi.org/10.1016/S0016-7037(02)00951-1.
Krabbenhöft, A., Eisenhauer, A., Böhm, F., Vollstaedt, H., Fietzke, J., Liebetrau, V., Au-
gustin, N., Peucker-Ehrenbrink, B., Müller, M.N., Horn, C., Hansen, B.T., Nolte,
N., Wallmann, K., 2010. Constraining the marine strontium budget with natural
strontium isotope fractionations (87 Sr/86 Sr∗ , δ 88/86 Sr) of carbonates, hydrother-
mal solutions and river waters. Geochim. Cosmochim. Acta 74 (14), 4097–4109.
http://dx.doi.org/10.1016/j.gca.2010.04.009.
Krabbenhöft, A., Fietzke, J., Eisenhauer, A., Liebetrau, V., Bohm, F., Vollstaedt,
H., 2009. Determination of radiogenic and stable strontium isotope ratios
(87 Sr/86 Sr; δ 88/86 Sr) by thermal ionization mass spectrometry applying an
87
Sr/84 Sr double spike. J. Anal. At. Spectrom. 24, 1267–1271.
Moynier, F., Agranier, A., Hezel, D.C., Bouvier, A., 2010. Sr stable isotope composi-
tion of Earth, the Moon, Mars, Vesta and meteorites. Earth Planet. Sci. Lett. 300
(3–4), 359–366. http://dx.doi.org/10.1016/j.epsl.2010.10.017.
Neymark, L.A., Premo, W.R., Mel’nikov, N.N., Emsbo, P., 2014. Precise determination
of δ 88 Sr in rocks, minerals, and waters by double-spike TIMS: a powerful tool in
the study of geological, hydrological and biological processes. J. Anal. At. Spec-
trom. 29 (1), 65. http://dx.doi.org/10.1039/c3ja50310k.
Nier, A.O., 1938. The isotopic constitution of Strontium, Barium, Bismuth, Thal-
lium and Mercury. Phys. Rev. 5 (4), 275–278. http://dx.doi.org/10.1103/PhysRev.
54.275.
Nissenbaum, A., Magaritz, M., 1991. Bromine-rich groundwater in the Hula Val-
ley, Israel. Naturwissenschaften 78 (1989), 217–218. http://dx.doi.org/10.1007/
BF01136083.
Ohno, T., Hirata, T., 2007. Simultaneous determination of mass-dependent iso-
topic fractionation and radiogenic isotope variation of strontium in geochemical
samples by multiple collector-ICP-mass spectrometry: analytical sciences. Anal.
Sci. 23 (11), 1275–1280.
Palmer, M.R., Edmond, J.M., 1989. The strontium isotope budget of the modern
ocean. Earth Planet. Sci. Lett. 92 (1), 11–26. http://dx.doi.org/10.1016/0012-
821X(89)90017-4.
Pearce, C.R., Parkinson, I.J., Gaillardet, J., Charlier, B.L.A., Mokadem, F., Burton, K.W.,
2015. Reassessing the stable (δ 88/86 Sr) and radiogenic (87 Sr/86 Sr) strontium iso-
topic composition of marine inputs. Geochim. Cosmochim. Acta 157, 125–146.
http://dx.doi.org/10.1016/j.gca.2015.02.029.
Raddatz, J., Liebetrau, V., Rüggeberg, A., Hathorne, E., Krabbenhöft, A., Eisen-
hauer, A., Böhm, F., Vollstaedt, H., Fietzke, J., López Correa, M., Freiwald, A.,
Dullo, W.C., 2013. Stable Sr-isotope, Sr/Ca, Mg/Ca, Li/Ca and Mg/Li ratios in
the scleractinian cold-water coral Lophelia pertusa. Chem. Geol. 352, 143–152.
http://dx.doi.org/10.1016/j.chemgeo.2013.06.013.
Rudge, J.F., Reynolds, B.C., Bourdon, B., 2009. The double spike toolbox. Chem.
Geol. 265 (3–4), 420–431. http://dx.doi.org/10.1016/j.chemgeo.2009.05.010.
Sandler, A., Meunier, A., Velde, B., 2015. Mineralogical and chemical variabil-
ity of mountain red/brown Mediterranean soils. Geoderma 239–240, 156–167.
http://dx.doi.org/10.1016/j.geoderma.2014.10.008.
Shalev, N., Segal, I., Lazar, B., Gavrieli, I., Fietzke, J., Eisenhauer, A., Halicz, L.,
2013. Precise determination of δ 88/86 Sr in natural samples by double-spike
MC-ICP-MS and its TIMS verification. J. Anal. At. Spectrom. 28 (6), 940–944.
http://dx.doi.org/10.1039/c3ja50039j.
De Souza, G.F., Reynolds, B.C., Kiczka, M., Bourdon, B., 2010. Evidence for mass-
dependent isotopic fractionation of strontium in a glaciated granitic water-
shed. Geochim. Cosmochim. Acta 74 (9), 2596–2614. http://dx.doi.org/10.1016/
j.gca.2010.02.012.
Spiro, B., Ashkenazi, S., Starinsky, A., Katz, A., 2011. Strontium isotopes in Melanop-
sis sp. as indicators of variation in hydrology and climate in the Upper Jordan
Valley during the Early–Middle Pleistocene, and wider implications. J. Hum.
Evol. 60, 407–416. http://dx.doi.org/10.1016/j.jhevol.2010.07.026.
Stein, M., Hofmann, A.W., 1992. Fossil plume head beneath the Arabian litho-
sphere? Earth Planet. Sci. Lett. 114 (1), 193–209. http://dx.doi.org/10.1016/
0012-821X(92)90161-N.
Stevenson, E.I., Aciego, S.M., Chutcharavan, P., Parkinson, I.J., Burton, K.W., Blakowski,
M.A., Arendt, C.A., 2016. Insights into combined radiogenic and stable strontium
isotopes as tracers for weathering processes in subglacial environments. Chem.
Geol. 429, 33–43. http://dx.doi.org/10.1016/j.chemgeo.2016.03.008.
Stevenson, E.I., Hermoso, M., Rickaby, R.E.M., Tyler, J.J., Minoletti, F., Parkinson, I.J.,
Mokadem, F., Burton, K.W., 2014. Controls on stable strontium isotope fraction-
ation in coccolithophores with implications for the marine Sr cycle. Geochim.
Cosmochim. Acta 128, 225–235. http://dx.doi.org/10.1016/j.gca.2013.11.043.
Tipper, E.T., Galy, A., Bickle, M.J., 2006. Riverine evidence for a fractionated reser-
voir of Ca and Mg on the continents: implications for the oceanic Ca cycle.
Earth Planet. Sci. Lett. 247 (3–4), 267–279. http://dx.doi.org/10.1016/j.epsl.2006.
04.033.
Tipper, E.T., Galy, A., Bickle, M.J., 2008. Calcium and magnesium isotope sys-
tematics in rivers draining the Himalaya–Tibetan–Plateau region: lithologi-
cal or fractionation control? Geochim. Cosmochim. Acta 72 (4), 1057–1075.
http://dx.doi.org/10.1016/j.gca.2007.11.029.
 N. Shalev et al. / Earth and Planetary Science Letters 459 (2017) 381–393
Veizer, J., 1989. Strontium isotopes in seawater through time. Annu. Rev. Earth
Planet. Sci. 17 (May 1989), 141–167. http://dx.doi.org/10.1146/annurev.ea.17.
050189.001041.
Voigt, J., Hathorne, E.C., Frank, M., Vollstaedt, H., Eisenhauer, A., 2015. Variabil-
ity of carbonate diagenesis in equatorial Pacific sediments deduced from ra-
diogenic and stable Sr isotopes. Geochim. Cosmochim. Acta 148, 360–377.
http://dx.doi.org/10.1016/j.gca.2014.10.001.
Vollstaedt, H., Eisenhauer, A., Wallmann, K., Böhm, F., Fietzke, J., Liebetrau, V.,
Krabbenhöft, A., Farkaš, J., Tomašových, A., Raddatz, J., Veizer, J., 2014. The
Phanerozoic δ 88/86 Sr record of seawater: new constraints on past changes in
oceanic carbonate fluxes. Geochim. Cosmochim. Acta 128, 249–265. http://dx.
doi.org/10.1016/j.gca.2013.10.006.
393
Wei, G., Ma, J., Liu, Y., Xie, L., Lu, W., Deng, W., Ren, Z., Zeng, T., Yang, Y., 2013.
Seasonal changes in the radiogenic and stable strontium isotopic composition
of Xijiang River water: implications for chemical weathering. Chem. Geol. 343,
67–75. http://dx.doi.org/10.1016/j.chemgeo.2013.02.004.
Yaalon, D.H., 1997. Soils in the Mediterranean region: what makes them different?
Catena 28 (3–4), 157–169. http://dx.doi.org/10.1016/S0341-8162(96)00035-5.
Young, E.D., Galy, A., Nagahara, H., 2002. Kinetic and equilibrium mass-
dependant isotope fractionation laws in nature and their geochemical and
cosmochemical significance. Geochim. Cosmochim. Acta 66 (6), 1095–1104.
http://dx.doi.org/10.1016/S0016-7037(01)00832-8.