Earth and Planetary Science Letters

Transcription

1 Earth and Planetary Science Letters 270 (2008) Contents lists available at ScienceDirect Earth and Planetary Science Letters journal homepage: www. elsevier. com/ locate/ epsl δ 44 Ca evolution in a carbonate aquifer and its bearing on the equilibrium isotope fractionation factor for calcite Andrew D. Jacobson a,, Chris Holmden b a Department of Earth and Planetary Sciences, Northwestern University, 1850 Campus Dr, Evanston, IL, 60208, USA b Saskatchewan Isotope Laboratory, Department of Geological Sciences, 114 Science Place, University of Saskatchewan, Saskatoon, Canada SK S7N 5E2 A R T I C L E I N F O A B S T R A C T Article history: Received 29 November 2007 Received in revised form 17 March 2008 Accepted 27 March 2008 Available online 9 April 2008 Editor: R.W. Carlson Keywords: calcium isotopes groundwater calcite precipitation fractionation To improve understanding of Ca isotope transport during water-rock interaction on the continents, we measured dissolved δ 44 Ca values along a 236 km flow path in the Madison aquifer, South Dakota, where fluids have chemically evolved according to dolomite and anhydrite dissolution, calcite precipitation, and Cafor-Na ion-exchange over a timescale spanning ~15 kyr. We used a reactive transport model employing rate data constrained from major ion mass-balances to evaluate the extent to which calcite precipitation and ionexchange fractionate Ca isotopes. Elevated δ 44 Ca values during the initial and final stages of water transport possibly result from calcite precipitation under supersaturated conditions and Ca-for-Na ion-exchange, respectively. However, for the bulk of the flow path, δ 44 Ca values evolve by mixing between anhydrite and dolomite dissolution, with no fractionation during calcite precipitation under saturated conditions. We attribute the absence of Ca isotope fractionation to the long timescale of water-rock interaction and slow rate of calcite precipitation, which have enabled fluids to chemically and isotopically equilibrate with calcite. We therefore conclude that the equilibrium Ca isotope fractionation factor between calcite and water (Δ cal w ) is very close to zero. To the extent that the Madison aquifer typifies other groundwater systems where calcite slowly precipitates from solutions at or near chemical equilibrium, this study suggests that groundwater contributions to δ 44 Ca variability on the continents can be modeled according to simple mixing theory without invoking isotope discrimination Elsevier B.V. All rights reserved. 1. Introduction Calcium isotope geochemistry (δ 44 Ca) has recently emerged as a promising tool for examining the chemical evolution of Earth's atmosphere and oceans. To this end, most work has focused on the marine realm with the aim of reconstructing how climatic and tectonic processes have influenced the oceanic Ca cycle (Zhu and Macdougall, 1998; De La Rocha and DePaolo 2000; Schmitt et al., 2003b; Fantle and DePaolo, 2005; Farkas et al., 2007a; Farkas et al., 2007b; Sime et al., 2007). Interpreting secular variations in the marine δ 44 Ca record requires information about processes controlling the δ 44 Ca composition of continental weathering inputs. Pioneering studies have revealed that calcite precipitation, ion-exchange, and plant uptake of Ca can preferentially remove light isotopes from solution, thus suggesting that development of a heavy isotope reservoir on the continents could impact the isotope balance of Ca in seawater (Russell and Papanastassiou, 1978a; Schmitt et al., 2003a; Corresponding author. Tel.: ; fax: address: adj@earth.northwestern.edu (A.D. Jacobson). DePaolo, 2004; Lemarchand et al., 2004; Marriott et al., 2004; Wiegand et al., 2005; Tipper et al., 2006). More recently, analyses of pore fluids within slowly recrystallizing marine carbonate sediments have demonstrated that no Ca isotope fractionation occurs during calcite precipitation under conditions of chemical equilibrium (Fantle and DePaolo, 2007). This implies that the equilibrium Ca isotope fractionation factor between calcite and water (Δ cal w ) equals 0 (α cal w =1.0000) and that other instances of fractionation during calcite precipitation are kinetically controlled. On the continents, aquifers represent a comparable environment where relatively long water-rock interaction times and slow reaction kinetics may facilitate chemical and isotopic equilibration between water, calcite, and other phases involved in the distribution and transport of dissolved Ca. Hence, aquifers offer a powerful opportunity to test fundamental assumptions concerning Ca isotope fractionation effects while simultaneously providing direct insight into processes controlling δ 44 Ca variability in terrestrial waters. In this study, we examined the downgradient evolution of dissolved δ 44 Ca values along a 236 km flow path in the Madison aquifer, South Dakota a confined carbonate aquifer recharging in the igneous Black Hills. The Madison aquifer offers a unique opportunity to explore Ca isotope effects during water-rock interaction because X/$ see front matter 2008 Elsevier B.V. All rights reserved. doi: /j.epsl

2 350 A.D. Jacobson, C. Holmden / Earth and Planetary Science Letters 270 (2008) Table 1 Ca isotope geochemistry of water samples from the madison aquifer Sample Distance (km) (1) fluids and rock have chemically equilibrated over a much longer timescale (up to ~15 kyr) than can be simulated in laboratory experiments and (2) previous studies have quantified the rates and mass-balances of geochemical reactions controlling bulk Ca concentrations along the aquifer flow path (Plummer et al., 1990; Jacobson and Wasserburg, 2005). Here, we use new δ 44 Ca measurements and existing model rates for the addition and removal of Ca to evaluate the extent to which Ca isotope fractionation may be associated with calcite precipitation and Ca-for-Na ion-exchange. This work represents the first comprehensive analysis of Ca isotope transport in groundwater systems. We find that throughout most of the aquifer, δ 44 Ca traces mixing of Ca from dolomite and anhydrite dissolution, with no apparent isotope discrimination during calcite precipitation under conditions of chemical equilibrium. 2. Study site and methods Ca (μmol/l) δ 44 Ca±2σ int S ± 0.03 S ±0.04 W ±0.03 W ± 0.04 W ± 0.03 W ± 0.05 W ±0.04 W ±0.04 W ±0.03 W ±0.04 Data taken from Jacobson and Wasserburg (2005). 2σ int=internal precision. Detailed descriptions of the study site as well as the samples and their collection method are given in Jacobson and Wasserburg (2005). Briefly, the Madison aquifer is the middle of three Paleozoic aquifers in parts of MT, ND, SD, WY, and Canada. The aquifer is composed of Mississippian age Madison Limestone, which is a grey to buff colored shallow-water marine carbonate containing dolomite (70%), calcite (25%), and anhydrite (4%) with accessory goethite, halite, hematite, quartz, kaolinite, illite, and mixed layer illite-smectite (Thayer, 1983; Busby et al., 1991). Shale and siltstone belonging to the Englewood Formation confine the bottom of the aquifer, while shale and sandstone belonging to the Minnelusa Formation confine the top. The Madison Limestone was locally uplifted and tilted during the Laramide orogeny, which exposed Precambrian basement rocks and Tertiary syenite porphory intrusions comprising the core of the Black Hills. Recharge to the aquifer primarily occurs by hydrologic loss as streams originating in gneissic and igneous watersheds cross karstified beds of Madison Limestone draping the flanks of the intrusion (Downey, 1984; Long and Putnam, 2002). Following recharge, groundwater flows in a west to east direction with velocities ranging from 2 to 25 m/yr (Downey, 1984; Plummer et al., 1990; Long and Putnam, 2002). Vertical water mass mixing is negligible (Downey, 1984). The suite of samples analyzed in this study consists of two surface water samples (S1 and S2) assumed to represent recharge waters to the aquifer and eight groundwater samples (W1 W8) assumed to lie along a 236 km hydrologic flow path between the Black Hills and central SD [see Fig. 1a,b in Jacobson and Wasserburg (2005)]. S1 was collected from a tributary stream draining the core of the Black Hills, whereas S2 was collected where the main stem of the tributary intersects outcrop of Madison Limestone. W1 W8 were collected from wells completed in the Madison aquifer. Groundwater ages range from ~0 30 yrs (W1 W4), ~70 yrs (W5), and ~15,000 yrs (W6 W8) (Plummer et al., 1990). All samples were filtered through acid-cleaned 0.45 μm filters, acidified to ph=2 with ultrapure HNO 3, and stored in acid-cleaned LDPE bottles prior to analysis. Two dolomite samples (DOL1 and DOL2) were collected near the location of S2, and two anhydrite samples (ANH1 and ANH2) were taken from Madison Limestone drill core currently stored at the USGS Core Research Center in Denver, CO (Blankennagel et al., 1979). All rock samples were reduced to a fine powder in a shatterbox, and 1-g sub-samples were completely digested in 6N ultrapure HCl. Following the procedure outlined in Holmden (2005), δ 44 Ca values were measured using a Triton MC-TIMS at the Saskatchewan Isotope Laboratory and a 43 Ca 42 Ca double-spike to correct for instrumental mass discrimination. Care was taken to ensure complete spike-sample equilibration before purification of Ca from matrix elements using Teflon columns packed with AG MP-50 cation exchange resin. Total procedural blanks were negligible (~100 ng). Sample:blank ratios were 500:1. Approximately 5 to 8 μg of purified Ca was loaded onto out-gassed Ta filaments, using ultrapure H 3 PO 4 acid, and measurements were made using a single filament assembly. Observation of the 41 K beam demonstrated absence of an isobaric interference of 40 K on 40 Ca. All 44 Ca/ 40 Ca ratios are reported in delta notation (δ 44 Ca)=[( 44 Ca/ 40 Ca) sample /( 44 Ca/ 40 Ca) sw 1] The internal precision (2σ int ) of the analyses ranged between ±0.030 and Repeated analyses of samples and internal standards during the period of study yielded an external precision (2σ ext ) of ±0.08. The value obtained for NIST 915a was Results Table 1 reports data for the water samples, and Fig. 1 shows δ 44 Ca values as a function of distance along the flow path. The surface water collected from the core of the Black Hills (S1) has δ 44 Ca= 0.93, whereas the surface water collected at the hydrologic loss zone (S2) has δ 44 Ca= Following recharge, dissolved δ 44 Ca values decrease from 0.72 to 1.31 with increasing downgradient distance between W1 and W7. The one exception to this trend is W8 (the oldest water sample collected), which has δ 44 Ca= Table 2 shows δ 44 Ca values for the dolomite and anhydrite samples. Within the uncertainty of the measurements, the dolomite samples are distinct from each other ( 1.10 and 0.88 ). The anhydrite samples yielded lower and more uniform δ 44 Ca values ( 1.29 Fig. 1. δ 44 Ca vs. Distance (km) for Madison aquifer surface- and ground-water samples (open and filled circles, respectively). Error bars show 2σ ext =±0.08. Solid horizontal lines display average δ 44 Ca of dolomite (DOL) and anhydrite (ANH). Shaded regions surrounding solid lines denote 2σ ext. Solid curve displays best fit" model output. Dashed curves A and B display model output for flow path segment S2 W7 with Δ cal w= 2.1 and 1.5, respectively. Dashed curves C and D display model output for flow path segment S2 W7 considering temperature dependency of Δ cal w [0.02 / C; Lemarchand et al. (2004)] with initial Δ cal w = 2.1 and 1.5, respectively. Dashed curve E displays model output for flow path segment S2 W7 with initial Ca concentration set to 400 μmol/l; see text.

3 A.D. Jacobson, C. Holmden / Earth and Planetary Science Letters 270 (2008) Table 2 Ca isotope geochemistry of dolomite (DOL) and anhydrite (ANH) samples from the madison aquifer SampIe and 1.21 ). The lighter value for DOL1 relative to DOL2 cannot be explained by the presence of trace anhydrite with low δ 44 Ca, because the anhydrite concentration in DOL1 is small (~ g/g) (Jacobson and Wasserburg, 2005). At present, Ca isotope data for dolomite and anhydrite are scarce. Relative to our values for seawater and NIST 915a, Carboniferous dolomite has δ 44 Ca= 0.76 (n=2 Steuber and Buhl, 2006), and Neoproterozoic cap carbonate dolomicrite has δ 44 Ca= 0.87 (Kasemann et al., 2005). Similarly, Russell and Papanastassiou (1978b) reported values of 1.5 and 2.1 for anhydrite, whereas in a more recent survey, Hensley (2006) reported values for both anhydrite and gypsum ranging from 0.20 to According to batch precipitation experiments, the δ 44 Ca value of anhydrite is lower by ~1 relative to the fluid from which it precipitates (Hensley, 2006). 4. Discussion Ca (μmol/g) δ 44 Ca±2σ int DOL ±0.03 DOL ±0.04 ANH ±0.05 ANH ±0.04 2σint =internal precision. Previous researchers have shown that groundwater in the Madison aquifer chemically evolves by dedolomitization (Plummer and Back, 1980; Back et al., 1983; Plummer et al., 1990; Naus et al., 2001; Jacobson and Wasserburg, 2005). In this reaction, the irreversible dissolution of anhydrite in waters in equilibrium with respect to dolomite and calcite leads to dolomite dissolution and calcite precipitation according to the following generalized reaction: CaSO 4 þ CaMgðCO 3 Þ 2 2CaCO 3 þ Mg 2þ þ SO 2 4 ð1þ Typically, the amount of Ca added to solution, mostly by anhydrite dissolution and to a lesser extent dolomite dissolution, exceeds the amount of Ca removed by calcite precipitation such that dissolved Ca accumulates during downgradient water transport (Plummer and Back, 1980; Back et al., 1983; Plummer et al., 1990; Naus et al., 2001; Jacobson and Wasserburg, 2005). Calcite precipitation occurs during small fluctuations about the equilibrium state. Indeed, intrinsic carbonate mineral reaction rates in the Madison aquifer are extremely slow (~ yr 1 ; Jacobson and Wasserburg, 2005) and comparable to those reported for recrystallizing marine carbonate sediments (Fantle and DePaolo, 2007). Halite dissolution and Ca-for-Na ion-exchange also occur, but for the bulk of the flow path (S1 W7), these reactions are negligible. The one exception is the flow path segment between W7 and W8, where ion-exchange and calcite precipitation remove Ca at nearly equivalent rates. The net effect is that Ca concentrations linearly increase between S1 and W7 but then moderately decline between W7 and W8 (Jacobson and Wasserburg, 2005). As waters approach equilibrium with respect to anhydrite with increasing downgradient distance, their δ 44 Ca composition approaches that of anhydrite (Fig. 2a). We therefore interpret the bulk of the data to reflect the progressive addition of anhydritederived Ca during groundwater flow. However, notable exceptions to this overall pattern exist, especially at the extremes of the flow path. Waters in the immediate vicinity of the recharge zone appear 44 Ca enriched ( 0.71 ) relative to dolomite ( 0.88 and 1.10 ) and S1 ( 0.93 ). The δ 44 Ca value measured for S1 might reflect the weathering of Black Hills igneous rocks, which would be expected to have low δ 44 Ca values relative to dolomite. With only two dolomite samples, it is conceivable that we incompletely characterized the composition of dolomite and that the average δ 44 Ca value is closer to Hence, dolomite dissolution could explain the δ 44 Ca increase between S1 and S2. Alternatively, calcite precipitation (Tipper et al., 2006) could fractionate Ca isotopes during surface water transport and groundwater infiltration. The two surface waters (S1 and S2) are supersaturated with respect to calcite and dolomite, whereas all groundwater samples (W1 W8) are in equilibrium (Jacobson and Wasserburg, 2005). As shown in Fig. 2b and c, however, an immediate correlation between δ 44 Ca and the calcite and dolomite saturation state of the waters is not evident. Removal of 40 Ca by plants could also explain the relatively high δ 44 Ca values of infiltrating waters (Schmitt et al., 2003a; Bullen et al., 2004; Wiegand et al., 2005; Holmden and Belanger, 2006). In the case of W8, which is 44 Ca enriched relative to the samples that precede it, we reason by inference to previous laboratory experiments that Ca-for-Na ion-exchange could fractionate Ca isotopes (Russell and Papanastassiou, 1978a). As discussed above, the flow path segment between W7 and W8 experiences a large effect from ion-exchange relative to the preceding segment. To better understand the δ 44 Ca evolution of the aquifer waters, we adapt the steady-state 1D reactive-transport model presented in Jacobson and Wasserburg (2005). Consistent with previous studies of the Madison aquifer, the model implicitly assumes that the composition of recharge waters has not changed appreciably over the time span represented by the samples (Plummer et al., 1990). The equation describing the rate of change in the amount of dissolved Ca in water (w) over distance (dx) is: v dca w dx ¼ F dol þ F anh F ixc ; ð2þ where ν is the water velocity, F dol and F anh are the rates of Ca addition by dolomite and anhydrite dissolution, respectively, and F cal and F ixc are the rates of Ca removal by calcite precipitation and Ca-for-Na ion- Fig. 2. δ 44 Ca vs. the saturation state of waters (logω) with respect to anhydrite (a), calcite (b), and dolomite (c). Surface- and ground-water samples shown with open and filled circles, respectively. Error bars show 2σ ext = ±0.08. Dashed lines show ±0.25 uncertainty of saturation state calculations. Saturation state data taken from Jacobson and Wasserburg (2005).

4 352 A.D. Jacobson, C. Holmden / Earth and Planetary Science Letters 270 (2008) exchange, respectively. Similarly, the equation describing the downgradient evolution of dissolved δ 44 Ca values is: vca w ðxþ dd44 Ca w dx ¼ F dol d 44 Ca dol d 44 Ca w þ F anh d 44 Ca anh d 44 Ca w ; F cal D cal w F ixc D ixc w where Ca w (x) is the Ca concentration in water at position x determined from Eq. (1), δ 44 Ca dol and δ 44 Ca anh are the Ca isotope compositions of dolomite and anhydrite, respectively; Δ cal w and Δ ixc w are the approximate isotope fractionation factors associated with calcite precipitation and Ca-for-Na ion-exchange, respectively; and the remainder of the parameters are described above. As discussed here, Δ j w =δ 44 Ca j δ 44 Ca w 10 3 lnα j w, where j refers to either calcite precipitation or Ca-for-Na ion-exchange, and α j w is the isotope fractionation factor. The model only considers fractionation during removal of dissolved Ca because at present, no evidence exists for fractionation during weathering. We use Eqs. (1) and (2) to estimate Δ j values that could best explain the observed water data. We set ν=2 m/yr (Jacobson and Wasserburg, 2005); δ 44 Ca dol = 0.99 (the average of DOL1 and DOL2); and δ 44 Ca anh = 1.25 (the average of ANH1 and ANH2). The rate parameters (F values) are given in Table 3. These rates, which were determined via inverse mass-balances for Cl, Na, Mg, SO 4, and total inorganic carbon (C T ), are effective, in that they describe the rate at which dissolution, precipitation, and ion-exchange either add or remove Ca to produce the observed water chemistry for a specified value of ν. As discussed in Jacobson and Wasserburg (2005), two sets of F values exist, one for flow path segment S1 W7 and one for W7 W8. To evaluate the apparent 44 Ca enrichment of the recharge waters relative to the bedrock δ 44 Ca values, we divide segment S1 W7 into two smaller segments, S1 S2 and S2 W7. This is permissible because the reaction rates are linear. The F values that apply to segment S1 W7 also apply to any smaller segment between S1 and W7. The high δ 44 Ca value for S2 superimposed on the otherwise decreasing trend between S1 and W7 (Fig. 1) dictates that we divide S1 W7 into a minimum of two smaller segments. We could further subdivide S1 S2 and S2 W7, but because the reaction rates are linear, this would not change the model Δ j w values obtained. For S1 S2, the best fit is obtained by setting Δ cal w = 2.1 (Fig. 1; solid curve). This suggests that calcite precipitation could enrich infiltrating water in 44 Ca by preferentially removing 40 Ca. The dissolved δ 44 Ca values cannot be attributed to ion-exchange because the model shows that the relatively modest rate of Ca removal by ion-exchange (a factor of ~40 lower than that attributed to calcite precipitation) would require an unreasonably high fractionation factor. We note that the Δ cal w value of 2.1 is larger than that determined in prior studies of inorganic calcite precipitation in both field and laboratory settings (e.g., 0.92, Tipper et al., 2006; 1.5±0.25, Lemarchand et al., 2004). This discrepancy could arise because the model attributes all fractionation to calcite precipitation, whereas 40 Ca uptake by plants might also be important. Moreover, the model fractionation factor greatly depends on the Ca concentration and isotope composition of the initial input water (S1). Sample S1 was Table 3 Reaction rates controlling the input and output of dissolved Ca to Madison aquifer groundwater, Segment F dol F anh F cal F ixc ( 10 2 μmol Ca/l/yr) S1 W W7 W Calculated according to data presented in Jacobson and Wasserburg (2005). Negative sign indicates removed from solution. S1 S2 and S2 W7 experience the same reaction rates as S1 W7; see text. ð3þ collected during the summer, when stream flow is lowest and evaporative concentration is highest. However, most recharge occurs during peak discharge in the spring, when water from melting snow is expected to be more dilute. As such, a lower initial Ca concentration would yield a lower model Δ cal w value. If we set Δ cal w = 1.5, then the model predicts an input water Ca concentration of ~400 μmol/l, which is a factor of ~2.5 lower than the concentration for S1. As described above, it is also conceivable that we incompletely characterized δ 44 Ca dol and that the actual average value is closer to S2 and W1. This scenario decreases the importance of calcite precipitation and points to dolomite dissolution as the cause for the isotopic increase between S1 and S2. In either case, while apportioning isotopic effects among calcite precipitation versus biological processes, and determining seasonal and spatial variability in surface water δ 44 Ca values presently hampers reliable estimation of Δ cal w in the Black Hills recharge zone, our data qualitatively suggest that some form of Ca isotope redistribution could occur during the initial stages of water transport and infiltration. Given that ion-exchange is unimportant for Ca isotope transport in segment S1 S2, we set Δ ixc w =0 for segment S2 W7. However, the same logic does not hold for Δ cal w. Setting Δ cal w equal to 2.1, or even as low as 1.5, produces a large offset between the measured and modeled data (Fig. 1; dashed curves A and B). Although groundwater temperatures increase downgradient from a minimum of 10 C to a maximum of 70 C at a rate of ~0.25 C/km, the offset persists even if we account for the effect of temperature on Δ cal w (0.02 / C) (Lemarchand et al., 2004) (Fig. 1; dashed curves B and C). In general, Ca isotope fractionation during inorganic calcite precipitation is only weakly dependent on temperature (Marriott et al., 2004; Gussone et al., 2005; Sime et al., 2005; Steuber and Buhl, 2006). The closest match is obtained by setting Δ cal w =0, which indicates that calcite precipitation under conditions of chemical equilibrium does not fractionate Ca isotopes in this region of the aquifer (solid curve in Fig. 1). This is consistent with the absence of a correlation between the calcite saturation state of the groundwater samples and their δ 44 Ca composition (filled circles in Fig. 2b). When Δ cal w is set to 0, the resulting theoretical δ 44 Ca values are largely insensitive to assumed changes in the dolomite weathering contribution. For example, the difference between the best fit model output (solid curve in Fig. 1) and model output obtained by setting δ 44 Ca dol equal to one of the two values measured (either 1.10 or 0.88 ) is much smaller than the ±0.08 uncertainty of the measurements. The same holds by eliminating the dolomite weathering contribution entirely (F dol =0). These sensitivity analyses confirm that anhydrite dissolution dominates the δ 44 Ca evolution of groundwater. Dolomite dissolution imparts only a small effect in the best fit case because the rate of Ca addition by dolomite dissolution is a factor ~5 lower than the corresponding rate for anhydrite dissolution (Table 3). A slightly better fit is obtained by setting the initial input Ca concentration to the model predicted value of 400 μmol/l (Fig. 1; dashed curve E). Regardless, the majority of data analyzed in this study clearly demonstrate that δ 44 Ca conservatively traces the progressive effect of anhydrite dissolution in a manner exactly identical to 87 Sr/ 86 Sr (Jacobson and Wasserburg, 2005). Lastly, for segment S7 S8, we set Δ cal w =0, because calcite precipitation under conditions of chemical equilibrium does not fractionate Ca isotopes in the preceding segment. This leaves Ca-for-Na ion-exchange as the most plausible mechanism to explain the large isotopic enrichment in W8 relative to W7, even at the relatively high temperature of ~70 C. The model yields Δ ixc w = 1.0, which is consistent with previous laboratory experiments showing that ion-exchange preferentially selects 40 Ca over 44 Ca (Russell and Papanastassiou, 1978a). Although we cannot entirely rule out the delivery of 44 Ca enriched fluids by cross-formational mixing, we view this scenario as unlikely. Firstly, previous work has demonstrated that cross-formational mixing is minimal (Downey, 1984). Secondly, both underlying and overlying formations contain anhydrite,

5 A.D. Jacobson, C. Holmden / Earth and Planetary Science Letters 270 (2008) which presumably has low δ 44 Ca values. Highly evolved waters in these aquifers should therefore be isotopically indistinguishable from those in the Madison aquifer. This is the first evidence that Ca isotope fractionation by ion-exchange may be occurring in nature with an apparent net isotope effect similar to many reported instances of biotic and abiotic calcite precipitation. Because ion-exchange is a ubiquitous phenomenon, it could be important for generating Ca isotope variability in other settings. However, we caution that further investigations are required to confirm this hypothesis. 5. Conclusions and implications Throughout most of the Madison aquifer, δ 44 Ca dynamically traces mixing between anhydrite- and dolomite-derived Ca, with no apparent isotope fractionation when calcite precipitates during small fluctuations about the state of chemical equilibrium. Modeling suggests that calcite precipitation and Ca-for-Na ion-exchange could fractionate Ca isotopes during the initial and final stages of water transport, respectively. However, given the limited number of samples characterizing these regions of the aquifer, these findings are less robust than the main conclusion of the study demonstrating that the equilibrium fractionation factor between calcite and water is very close to zero. We attribute the absence of Ca isotope fractionation to the long transit time of water (~15 kyr) and relatively slow rate of calcite precipitation, which have permitted chemical and isotopic equilibration between calcite and water. Our study supports the contention that the equilibrium Ca isotope fractionation factor between calcite and water (Δ cal w ) is very close to zero (Fantle and DePaolo, 2007). We therefore reason that instances of 44 Ca enrichment on the continents attributed to calcite precipitation (e.g., Tipper et al., 2006) must originate from kinetic effects, most likely controlled by the magnitude of the intrinsic precipitation rate and/or the degree of supersaturation of the fluid with respect to calcite (Gussone et al., 2003; Steuber and Buhl, 2006; Fantle and DePaolo, 2007). Although additional work is required to understand the possible role of ion-exchange in creating δ 44 Ca variability, our findings imply that in aquifers where Δ cal =0 groundwater δ 44 Ca values can be predicted by simple mixing, in a manner identical to other widely utilized radiogenic isotope tracers (e.g, 87 Sr/ 86 Sr). We therefore suggest that groundwater can be discounted as a significant source of 44 Ca enriched water in carbonate-rich settings, and propose instead that simple dilution and mixing during surface-groundwater interactions might serve to reduce the range of δ 44 Ca variability imparted to surface waters by rapid calcite precipitation and plant uptake. Acknowledgements This work was supported by NSF grant EAR awarded to A. Jacobson. T. D. Bullen and A. Galy provided comments on an earlier version of this manuscript. Two anonymous reviewers helped improve the version in print. The authors thank editor R. W. Carlson for handling the manuscript. References Back, W., Hanshaw, B.B., Plummer, L.N., Rahn, P.H., Rightmire, C.T., Rubin, M., Process and rate of dedolomitization: mass transfer and 14 C dating in a regional carbonate aquifer. Geol. Soc. Amer. Bull. 94, Blankennagel, R.K., Howells, L.W., Miller, W.R., Hansen, C.V., Preliminary data for Madison Limestone test well 3, NW 1/4, SE 1/4, Sec. 35, T. 2N., R. 27E, Yellowstone Country, Montana. USGS Open File Report, pp. Bullen, T.D., Fitzpatrck, J.A., White, A.F., Schulz, M.S., Vivit, D.V., Calcium stable isotope evidence for three soil calcium pools at a granitoid chronosequence. In: Wanty, R.B., Seal, R.R. (Eds.), Water-Rock Interaction, Proceedings of the Elventh International Symposium on Water-Rock Interaction. London, Taylor, & Francis, Saratoga, Springs, New York, pp Busby, J.F., Plummer, L.N., Lee, R.W., Hanshaw, B.B., Geochemical evolution of water in the Madison aquifer in parts of Montana, South Dakota and Wyoming. USGS Professional Paper, 1273-F: 89. De La Rocha, C.L., DePaolo, D.J., Isotopic evidence for variations in the marine calcium cycle over the Cenozoic. Science 289, DePaolo, D.J., Calcium isotope variations produced by biological, kinetic, radiogenic, and nucleosynthetic processes. In: Johnson, C.M., Beard, B.L., Albarede, F. (Eds.), Geochemistry of Non-traditional Stable Isotopes. Reviews in Mineralogy. Mineralogical Society of America, Washington DC, pp Downey, J.S., Geohydrology of the Madison and associated aquifers in parts of Montana, North Dakota, South Dakota, and Wyoming. USGS Professional Paper, 1273-G, p. 47. Fantle, M.S., DePaolo, D.J., Variations in the marine Ca cycle over the past 20 million years. Earth Planet. Sci. Lett. 237, Fantle, M.S., DePaolo, D.J., Ca isotopes in carbonate sediment and pore fluid from ODP Site 807A: The Ca 2+ (aq)-calcite equilibrium fractionation factor and calcite recrystallization in Pleistocene sediments. Geochimica et Cosmochimica Acta 71, Farkas, J., Bohm, F., Wallmann, K., Blenkinsop, J., Eisenhauer, A., Geldren, R.v., Munnecke, A., Voigt, S., Veizer, J., 2007a. Calcium isotope record of Phanerozoic oceans: implications for chemical evolution of seawater and its causative mechanisms. Geochim. Cosmochim. Acta 71, Farkas, J., Buhl, D., Blenkinsop, J., Veizer, J., 2007b. Evolution of the oceanic calcium cycle during the late Mesozoic: evidence from δ 44/40 Ca of marine skeletal carbonates. Earth Planet. Sci. Lett. 253, Gussone, N., Eisenhauer, A., Heuser, A., Dietzel, M., Bock, B., Bohm, F., Spero, H.J., Lea, D.W., Bijma, J., Nagler, T.F., Model for kinetic effects on calcium isotope fractionation (δ 44 Ca) in inorganic aragonite and cultured planktonic foraminifera. Geochim. Cosmochim. Acta 67, Gussone, N., Bohm, F., Eisenhauer, A., Dietzel, M., Heuser, A., Teichert, B.M.A., Reitner, J., Worheide, G., Dullo, W.-C., Calcium isotope fractionation in calcite and aragonite. Geochim. Cosmochim. Acta 69, Hensley, T.M., Calcium isotopic variation in marine evaporites and carbonates: Applications to Late Miocene Mediterranean brine chemistry and Late Cenozoic calcium cycling in the oceans, University of California, San Diego, Ph.D. Thesis, 124 pp. Holmden, C., Measurement of δ 44 Ca using a 43 Ca 42 Ca double-spike TIMS technique, Summary of Investigations Industry and Resources, vol. 1. Saskatchewan Geological Survey, Sask, p. 7. Holmden, C., Belanger, N., Calcium isotope fractionation in a boreal forst ecosystem. Geochim. Cosmochim. Acta 70, A261 suppl. S. Jacobson, A.D., Wasserburg, G.J., Anhydrite and the Sr isotope evolution of groundwater in a carbonate aquifer. Chem. Geol. 214, Kasemann, S.A., Hawkesworth, C.J., Prave, A.R., Fallick, A.E., Pearson, P.N., Boron and calcium isotope composition of Neoproterozoic carboante rocks from Namibia: evidence for extreme environmental change. Earth Planet. Sci. Lett. 231, Lemarchand, D., Wasserburg, G.J., Papanastassiou, D.A., Rate-controlled calcium isotope fractionation in synthetic calcite. Geochim. Cosmochim. Acta 68, Long, A.J., Putnam, L.D., Flow-system analysis of the Madison and Minnelusa aquifers in the Rapid City area, South Dakota Conceptual Model. USGS Water Resources Investigations Report, pp. Marriott, C.S., Henderson, G.M., Belshaw, N.S., Tudhope, A.W., Temperature dependence of δ 7 Li, δ 44 Ca, and Li/Ca during growth of calcium carbonate. Earth Planet. Sci. Lett. 222, Naus, C.A., Driscoll, D.G., Carter, J.M., Geochemistry of the Madison and Minnelusa aquifers in the Black Hills area, South Dakota. USGS Water Resources Investigations Report, pp. Plummer, L.N., Back, W., The mass balance approach: application to interpreting the chemical evolution of hydrologic systems. Am. J. Sci. 280, Plummer, L.N., Busby, J.F., Lee, R.W., Hanshaw, B.B., Geochemical modeling of the Madison Aquifer in parts of Montana, Wyoming, and South Dakota. Water Resour. Res. 26, Russell, W.A., Papanastassiou, D.A.,1978a. Calcium isotope fractionation in ion-exchange chromatography. Anal. Chem. 50, Russell, W.A., Papanastassiou, D.A., 1978b. Ca isotope fractionation on the Earth and other solar system materials. Geochim. Cosmochim. Acta 42, Schmitt, A.-D., Chabaux, F., Stille, P.J., 2003a. The calcium riverine and hydrothermal isotopicfluxesand theoceanic calcium mass balance. EarthPlanet. Sci. Lett. 6731,1 16. Schmitt, A.-D., Stille, P.J., Vennemann, T., 2003b. Variations of the 44 Ca/ 40 Ca ratio in seawater during the past 24 million years: evidence from δ 44 Ca and δ 18 O values of Miocene phosphates. Geochim. Cosmochim. Acta 67, Sime, N.G., De La Rocha, C.L., Galy, A., Negligible temperature dependence of calcium isotope fractionation in 12 species of planktonic foraminfera. Earth Planet. Sci. Lett. 232, Sime, N.G., De La Rocha, C.L., Tipper, E.T., Tripati, A., Galy, A., Bickle, M.J., Interpreting the Ca isotope record of marine biogenic carbonates. Geochim. Cosmochim. Acta 71, Steuber, T., Buhl, D., Calcium-isotope fractionation in selected modern and ancient marine carbonates. Geochim. Cosmochim. Acta 70, Thayer, P.A.,1983. Relationship of porosity and permeability to petrology of the Madison Limestone in rock cores from three test wells in Montana and Wyoming. USGS Professional Paper, 1273-C. 29 pp. Tipper, E.T., Galy, A., Bickle, M.J., Riverine evidence for a fractionated reservoir of Ca and Mg on the continents: implications for the oceanic Ca cycle. Earth Planet. Sci. Lett. 247, Wiegand, B.A., Chadwick, O.A., Vitousek, P.M., Wooden, J.L., Ca cycling and isotopic fluxes in forested ecosystems in Hawaii. Geophys. Res. Lett. 32, doi: / 2005GL Zhu, P., Macdougall, D.J.,1998. Calcium isotopesinthe marineenvironmentandtheoceanic calcium cycle. Geochim. Cosmochim. Acta 62,