Geologic CO2 input into groundwater and the atmosphere, Soda Springs, ID, USA

Chemical Geology 339 (2013) 61–70

Contents lists available at SciVerse ScienceDirect

Chemical Geology journal homepage: www.elsevier.com/locate/chemgeo

Geologic CO2 input into groundwater and the atmosphere, Soda Springs, ID, USA J.L. Lewicki a,⁎, G.E. Hilley b, L. Dobeck c, T.L. McLing d, B.M. Kennedy f, M. Bill f, B.D.V. Marino e a

U.S. Geological Survey, Menlo Park, CA 94025, USA Department of Geological and Environmental Sciences, Stanford University, Stanford, CA 94305, USA Department of Chemistry and Biochemistry, Montana State University, Bozeman, MT 59717, USA d Idaho National Laboratory, Idaho Falls, ID 83415, USA e Planetary Emissions Management, Inc., Cambridge, MA 02142, USA f Earth Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA b c

a r t i c l e

i n f o

Article history: Accepted 20 June 2012 Available online 30 June 2012 Keywords: Springs CO2 degassing Eddy covariance Mantle Metamorphic decarbonation Accumulation chamber

a b s t r a c t A set of CO2 flux, geochemical, and hydrologic measurement techniques was used to characterize the source of and quantify gaseous and dissolved CO2 discharges from the area of Soda Springs, southeastern Idaho. An eddy covariance system was deployed for ~ one month near a bubbling spring and measured net CO2 fluxes from −74 to 1147 g m −2 d −1. An inversion of measured eddy covariance CO2 fluxes and corresponding modeled source weight functions mapped the surface CO2 flux distribution within and quantified CO2 emission rate (24.9 t d −1) from a 0.05 km 2 area surrounding the spring. Soil CO2 fluxes (b1 to 52,178 g m −2 d −1) were measured within a 0.05 km2 area of diffuse degassing using the accumulation chamber method. The estimated CO2 emission rate from this area was 49 t d −1. A carbon mass balance approach was used to estimate dissolved CO2 discharges from contributing sources at nine springs and the Soda Springs geyser. Total dissolved inorganic carbon (as CO2) discharge for all sampled groundwater features was 57.1 t d −1. Of this quantity, approximately 3% was derived from biogenic carbon dissolved in infiltrating groundwater, 35% was derived from carbonate mineral dissolution within the aquifer(s), and 62% was derived from deep source(s). Isotopic compositions of helium (1.74–2.37 Ra) and deeply derived carbon (δ13C ≈ 3‰) suggested contribution of volatiles from mantle and carbonate sources. Assuming that the deeply derived CO2 discharge estimated for sampled groundwater features (~35 t d −1) is representative of springs throughout the study area, the total rate of deeply derived CO2 input into the groundwater system within this area could be ~ 350 t d−1, similar to CO2 emission rates from a number of quiescent volcanoes. Published by Elsevier B.V.

1. Introduction Quantification of CO2 emissions from volcanic, geothermal, and metamorphic (VGM) systems over a range of temporal and spatial scales is required to monitor volcanic activity, processes associated with geothermal energy production, and health and environmental hazards associated with elevated CO2 concentrations, as well as to explore for geothermal energy potential. In addition, although present-day emissions of CO2 derived from non-volcanic (not from the craters and flanks of volcanoes) sources such as metamorphism and subsurface magmatism are likely globally significant (e.g., Kerrick et al., 1995; Kerrick, 2001), a relatively limited number of estimates of this degassing are available (e.g., Kerrick et al., 1995; Seward and Kerrick, 1996; Chiodini et al., 1999, 2000; Rogie et al., 2000; Becker et al., 2008; Evans et al., 2008; Frondini et al., 2008). Further studies are required to quantify the contribution of CO2 degassing from non-volcanic sources to the global carbon cycle. A number of methods have been applied to characterize the discharge of CO2 from soil, groundwater, and surface water sources in ⁎ Corresponding author. Fax: +1 650 329 4538. E-mail address: [email protected] (J.L. Lewicki). 0009-2541/$ – see front matter. Published by Elsevier B.V. doi:10.1016/j.chemgeo.2012.06.013

VGM systems. For example, the accumulation chamber technique (e.g., Chiodini et al., 1998) has been a reliable technique to measure the spatial distribution of soil CO2 fluxes and estimate total CO2 emission rates in VGM regions. Limitations of the accumulation chamber method, however, include the measurement's small spatial scale (b1 m 2) and inability to continuously measure variations in CO2 fluxes at a large number of point locations across a study area. Spring flow rate measurements have been coupled with aqueous chemical and isotopic data to estimate dissolved discharges of geologic CO2 (e.g., Chiodini et al., 1999, 2000; Evans et al., 2002, 2006; Frondini et al., 2008), whereas direct measurements of gaseous CO2 emissions from bubbling springs and pools have been carried out by trapping the gas in an inverted container and measuring the CO2 concentration of the gas and the gas outflow rate from the container (e.g., Rogie et al., 2000; Werner et al., 2000a). Eddy covariance (EC) is a micrometeorological technique (e.g., Baldocchi, 2003) proposed as a method to measure CO2 emissions from VGM systems (Werner et al., 2000b; Anderson and Farrar, 2001; Werner et al., 2003; Lewicki et al., 2008, 2012). EC has the advantage of providing an automated and time and space-averaged flux measurement, with a spatial scale significantly larger (m 2–km 2) than that of aforementioned alternative

62

J.L. Lewicki et al. / Chemical Geology 339 (2013) 61–70

approaches. Lewicki et al. (2012) showed that EC could be used to map the spatial distribution of and quantify volcanic CO2 emissions from the soil. However, this ability has yet to be tested where geologic CO2 is emitted from ground and surface water sources. Numerous carbonated springs and extensive travertine deposits are distributed throughout southeastern Idaho. The geochemistry of a number of these springs has been studied as part of regional hydrologic and geothermal investigations (e.g., Young and Mitchell, 1973; Mitchell, 1976; Muller and Mayo, 1983; Hutsinpiller and Parry, 1985; Mayo et al., 1985; Semenza, 2011). While metamorphic decarbonation of marine carbonates was suggested to be a significant source of the CO2 in springs (Mayo et al., 1985), no attempt has yet been made to quantify rates of CO2 degassing. We investigate non-volcanic CO2 degassing near the town of Soda Springs, Idaho. Here we apply a set of CO2 flux, geochemical, and

hydrologic measurement techniques to quantify gaseous and dissolved CO2 discharges. Specifically, we test for the first time the ability of EC to quantify gaseous CO2 emissions from a spring source. Soil CO2 emissions from an area of diffuse degassing are measured using the accumulation chamber method. Furthermore, dissolved CO2 discharges from contributing sources are estimated at selected springs using a carbon mass balance approach (Chiodini et al., 1999, 2000). Finally, we make a preliminary estimate of total dissolved CO2 discharge derived from deep geologic source(s) for the entire study area. 2. Geologic and hydrogeologic background The study area is located in southeastern Idaho, within the Sevier fold-and-thrust belt and the northeasternmost Basin and Range province (Fig. 1a). The area is also located adjacent to the southern border of the eastern Snake River Plain volcanic province. Structures

Fig. 1. (a) Map of Soda Springs study area with groundwater sampling locations (modified from Armstrong, 1969; Mabey and Oriel, 1970; Semenza, 2011). Town of Soda Springs and cross-section A–A′ are also shown. (b) Geologic cross section A–A′ with springs mapped by Semenza (2011). Spring waters are derived from Paleozoic carbonate aquifers and have been estimated to circulate to depths of about 1000 to 2000 m (Mansfield, 1927; Mayo et al., 1985). Source locations and migration pathways of deeply derived CO2 are unknown.

J.L. Lewicki et al. / Chemical Geology 339 (2013) 61–70

63

related to Cretaceous to early Cenozoic contraction in the Sevier fold-and-thrust belt are overprinted by normal faults associated with Eocene to Holocene east–west extension (e.g., Armstrong and Oriel, 1965; Armstrong, 1969). Mountain ranges and basins are now hosted in the footwalls and hanging walls, respectively, of these extensional structures. Locally, the Soda Springs Hills on the west side of the study area are composed primarily of early to middle Paleozoic marine quartzites, limestones, and dolomites. On the east side of the study area, the Aspen Range consists mainly of late Paleozoic and Mesozoic marine limestones, dolomites, siltstones, and shales. Quaternary basalts and intercalated sediments associated with the Blackfoot Lava Field located north of the study area fill valleys (Fig. 1). In addition, several ~100 ka rhyolite domes (Armstrong et al., 1975) are situated within the Blackfoot lava field ~20 km north of Soda Springs. Numerous (close to 100) springs, both thermal and non-thermal, have been mapped within the study area (Semenza, 2011). The proximity of many springs and travertine deposits associated with active and extinct springs to extensional faults suggests that these faults act as conduits for ascending groundwater (e.g., Mansfield, 1927; Mayo et al., 1985). Spring waters in the study area are derived from Paleozoic carbonate aquifers (Fig. 1b) and have been estimated to circulate to depths of about 1000 to 2000 m (Mansfield, 1927; Mayo et al., 1985). Recharge of these aquifers occurs in areas where Paleozoic formations outcrop in the northwest-trending mountain ranges (Mayo et al., 1985). While metamorphic decarbonation of marine carbonates was suggested to be a significant source of carbon in high pCO2 springs in the region (Mayo et al., 1985), the source locations and migration pathways of deeply derived CO2 are unconstrained. In addition to the springs present in the area, the Soda Springs geyser (Fig. 1a) is a man-made CO2-charged geyser that erupts from a well drilled in 1937.

was measured by a LI-COR LI-190SA quantum sensor at 2 m height every 5 s and averaged over 30 min. Net turbulent flux of CO2 in the vertical direction (Fc) was calculated as the temporal covariance of CO2 density (c) and vertical wind speed (w),

3. Methods

F c ¼ w ′ c′ ;

3.1. Measurement and modeling of soil CO2 fluxes

where the overbar indicates a temporal average (30 min) and primes denote the instantaneous turbulent deviation of the measurement ′  Coordinate rotation, the from its temporal mean (e.g., w ¼ w−w). Webb–Pearman–Leuning correction for effects of heat and water vapor fluctuation on air density, raw signal de-spiking, and filtering of Fc data according to stationarity criteria were applied as described by Lewicki et al. (2008) and Lewicki and Hilley (2009). Fc data were filtered according to a test for development of turbulent conditions as described by Lewicki et al. (2012). Precision of Fc measurements is generally ±7% during the daytime and ±12% during the nighttime (Baldocchi, 2003 and references therein). Each Fc measurement is sourced from a unique surface area upwind of the sensors that depends on factors such as sensor height (zm), atmospheric stability, and surface roughness height (z0). A source weight function, f(xm − x′, ym − y′, zm − z0, t), that describes the relative contribution of each upwind element of surface CO2 flux (Qc) at point (x', y', z' = z0) to Fc measured at point (xm, ym, zm) can be modeled such that:

Soil CO2 flux was measured with a WEST Systems Fluxmeter based on the accumulation chamber method, with precision of ±10% (Chiodini et al., 1998). Soil CO2 flux was measured at 178 points within a 0.05 km 2 area adjacent to Sulfur spring (Figs. 1a and 2) on 2–3 October 2011. A stochastic simulation procedure based on a sequential Gaussian simulation algorithm from Geostatistical Software Library (Deutsch and Journel, 1998) was used to model the spatial distribution of soil CO2 flux and estimate CO2 emission rate from the study area as described by Lewicki et al. (2007). The mean and 95% lower and upper bounds of the simulated CO2 emission rates are assumed to be the characteristic CO2 emission rate for the study area and its uncertainty, respectively. A LI-COR LI-8100A Automated Soil CO2 Flux System was installed near Mammoth spring (Fig. 1a) from 30 September to 4 October 2011. This system was configured with four chambers that measured soil CO2 flux approximately once an hour. To characterize ecosystem respiration, one chamber was installed at a background site within pastureland with no visual evidence of geologic CO2 emissions. 3.2. Measurement and modeling of net CO2 fluxes We deployed an EC station 55 m northeast of the primary discharge point of Mammoth spring from 8 September to 9 October 2011 (Figs. 2a and 3a). A Gill-Solent WindMaster Pro sonic threedimensional anemometer/thermometer measured wind speeds in three orthogonal directions and sonic temperature at 10 Hz. A LI-COR LI-7500 open-path CO2-H2O infrared gas analyzer measured CO2 and water vapor densities at 10 Hz. Both sensors were mounted on a tripod tower at 3.8 m height. Photosynthetically active radiation

Fig. 2. Simulated map of log soil CO2 flux based on accumulation chamber measurements made at black dots within 0.05 km2 area around Sulfur spring. Lower and upper 95% bounds on CO2 emission rate (DCO2) are in parentheses. Mapped bubbling pools are shown as light blue areas.

ð1Þ

h

 i ∞ ∞ ′ ′ ′ ′ ′ ′ ′ F c ðxm ; ym ; zm Þ ¼ ∫−∞ ∫−∞ Q c x ; y ; z ¼ z0 ⋅f xm −x ; ym −y ; zm −z0 dx dy

ð2Þ (e.g., Horst and Weil, 1992; Schmid, 1997). If the spatial distribution of Qc(x ', y ', z ' = z0) is relatively constant over a particular observation interval, this distribution can be modeled using an inversion of Fc(xm, ym, zm) and f(xm − x′, ym − y′, zm − z0) that maintains smoothness of Qc(x ', y ', z ' = z0) (Lewicki and Hilley, 2009; Lewicki et al., 2009, 2012). We modeled a source weight function corresponding to each of 290 Fc measurements according to Horst and Weil (1992) using the inputs: (1) zm = 3.8 m; (2) z0 = 0.05 m; (3) measured mean horizontal wind

64

J.L. Lewicki et al. / Chemical Geology 339 (2013) 61–70

of the LI-8100A (18 g m −2 d −1) was subtracted from each Fc. Because (i) a clear relationship was not found between measured photosynthetically active radiation and Fc and (ii) Fc measurements were made after plant senescence, removal of a photosynthetic uptake component of Fc (Lewicki and Hilley, 2009; Lewicki et al., 2009) was deemed unnecessary. The model domain for the inversion was 400 × 600 m (Fig. 3b). As described in Lewicki et al. (2009) and Lewicki and Hilley (2009), there are typically more elements of surface CO2 flux to be modeled than available EC observations. Also, the best-fit solutions for surface CO2 flux distributions often vary abruptly and unrealistically in space. Therefore, two constraints were applied to the inversion that → required (i) Q c values to be non-negative by using the constrained optimization functionality of the CVX convex optimization code (e.g., Boyde and Vandenberghe, 2004) and (ii) a → finite-difference approximation of curvature between adjacent Q c values to be minimized, along with the misfit between observed and modeled Fc (e.g., Harris and Segall, 1987). A smoothing weight applied to the curvature finite difference approximation was systematically varied to find the value resulting in a compromise between spatial continuity across the model solution space and misfit between measured and modeled Fc (Lewicki and Hilley, 2009; Lewicki et al., 2009, 2012).

3.3. Groundwater sampling and analysis

Fig. 3. (a) Aerial photograph of Mammoth spring area with mapped bubbling pools (blue areas), primary spring outlet (yellow dot), eddy covariance (EC) station location (red dot), and EC average 90% flux source area (isopleth shown as white line). Photo of → primary spring outlet is shown in inset. (b) Map of modeled surface CO2 flux (Q c) and 2 CO2 emission rate (DCO2) within the EC average 90% flux source area (0.05 km ) for a smoothing weight (wsm) of 20. Black box shows extent of model domain.

direction; (4) cross-wind turbulence near the surface (σv/u*, where σv and u* are the standard deviation of wind speed in the cross-wind direction and friction velocity, respectively); (5) calculated Monin– Obukhov length (L), if it corresponded to unstable atmospheric conditions (i.e., only measurements corresponding to L b 0 were considered). The area, on average, from which 90% of the source weights were contained during the September–October 2011 observation period (the EC average 90% flux source area; Fig. 2) was calculated as described by Lewicki et al. (2012). We carried out an inversion of 290 Fc measurements and corresponding modeled source weight functions to model the spatial → distribution of surface CO2 fluxes (Q c). To remove the ecosystem component of CO2 flux from the total measured Fc used in the inversion (i.e., isolate the geologic CO2 emission signal), the average soil CO2 flux measured over a four-day period by the background chamber

We sampled nine springs (Mammoth, Sulfur, Hooper, Octagon, Outhouse, Ninety Percent, Barrel, Lover's Delight, and Mormon) and the Soda Springs geyser (Fig. 1a) in September–October 2011. Water temperature and pH were measured in the field at the time of sample collection. Filtered (0.45 μm) water samples were collected in 40 ml glass vials with septa caps for analysis of sulfate and total dissolved inorganic carbon concentrations ([SO4] and [DIC]tot, respectively) and stable carbon isotopes of DIC. Duplicate samples were acidified for analysis of calcium and magnesium concentrations ([Ca] and [Mg], respectively). All water samples were refrigerated until analysis. Free gases from the bubbling Hooper, Octagon, Mammoth, and Sulfur springs were trapped in an inverted funnel and collected in pre-evacuated Pyrex bottles for stable carbon isotopic analysis of CO2. The CO2 concentration of these gases ([CO2]gas) was measured (±0.1 vol.%) in the field by pumping gases from the inverted funnel to a Landtech GEM2000 landfill gas analyzer. Gases from Hooper, Octagon, and Mammoth springs were trapped in an inverted funnel and collected in Cu tubes that were then sealed with refrigeration clamps for noble gas analysis. Spring discharge rates were either measured in September–October 2011 using a pygmy meter following U.S. Geological Survey methods (Buchanan and Somers, 1969) or visually estimated. We estimated a sum of flow rates from numerous small discharge points at Sulfur and Outhouse springs. The uncertainty (±1σ) of water discharge measurements was estimated as 20% (Sauer and Meyer, 1992). Aqueous geochemical measurements were made for the purpose of carrying out carbon mass balance calculations (Section 3.4). The [Ca] and [Mg] of water samples were determined by a Perkin Elmer SCIEX Elan DRC II inductively coupled plasma mass spectrometer. The [SO4] was determined by a Dionex ICS-2100 integrated reagent-free ion chromatography system. The [DIC]tot was determined by a Shimadzu TOC-VCSH inorganic/organic carbon analyzer. Uncertainties (±1σ) of [Ca], [Mg], [SO4], and [DIC]tot analyses are given in Table 1. Carbon isotope compositions of DIC (δ 13Ctot; ± 0.19‰) and CO2 gas (δ 13CCO2; ± 0.37‰) samples were analyzed using a Micromass JA Series Isoprime isotope ratio mass spectrometer connected to a Tracegas™ preconcentrator interface associated with an autosampler. Noble gas abundances and isotopic ratios were measured as described by Kennedy et al. (1985) and Hiyagon and Kennedy (1992). Analytical uncertainties (± 1σ) are given in Table 2.

J.L. Lewicki et al. / Chemical Geology 339 (2013) 61–70

65

Table 1 Chemical and isotopic data. Site

Latitude

Longitude

T (°C)

pH

[CO2]gas (vol. %)

[DIC]tot (mmol l−1)t

±

[Ca] (mg l−1)

±

[Mg] (mg l−1)

±

[SO4] (mg l−1)

±

δ13Ctot (‰)

δ13CCO2 (‰)

Hooper Octagon Mammoth Sulfur Geyser Outhouse Ninety Percent Barrel Lover's Delight Mormon

42° 42° 42° 42° 42° 42° 42°

111° 111° 111° 111° 111° 111° 111°

10.2 14.2 15.8 14.9 25.8 29.7 10.5

6.0 6.0 5.7 5.9 6.3 6.0 6.5

70.0 99.9 99.0 99.6 nd nd nd

35.5 61.0 44.0 19.8 52.4 54.1 19.3

0.3 0.6 1.4 0.2 6.4 0.9 0.2

96.80 463.80 101.50 197.90 595.00 629.30 99.60

1.51 1.90 1.10 0.67 1.50 1.80 0.71

127.70 181.40 144.70 48.50 177.60 186.30 105.60

0.51 0.45 0.45 0.21 1.20 1.80 0.51

56.07 449.71 51.03 290.83 800.70 889.26 23.84

0.17 1.03 0.24 0.95 1.21 1.51 0.01

1.9 2.9 2.3 4.6 4.5 4.2 2.7

−1.4 −1.2 −0.3 2.3 nd nd nd

40′ 39′ 42′ 38′ 39′ 39′ 41′

44.304″N 35.657″N 48.574″N 43.519″N 26.207″N 20.285″N 1.838″N

36′ 36′ 37′ 30′ 36′ 38′ 37′

13.317″W 9.798″W 25.082″W 22.846″W 18.887″W 39.645″W 56.400″W

42° 40′ 37.037″N 42° 39′ 53.131″N

111° 36′ 8.194″W 111° 36′ 3.773″W

8.6 8.9

6.1 6.2

nd nd

35.9 42.2

0.4 1.3

99.60 144.50

1.20 0.98

140.10 178.10

0.70 0.88

55.75 65.05

0.08 0.02

1.3 1.5

nd nd

42° 40′ 29.607″N

111° 35′ 38.114″W

13.9

7

nd

11.6

0.1

111.00

1.10

46.90

0.30

28.60

0.02

1.6

nd

of [DIC]ext, respectively. Eq. (5) was rearranged to estimate δ 13Cext for each sample:

3.4. Groundwater carbon balance We applied a carbon mass balance approach following the methods of Chiodini et al. (1999, 2000) to estimate the discharges of DIC derived from contributing sources from the sampled groundwater features. The measured [DIC] tot is the sum of DIC derived from dissolution of carbonate minerals within the aquifer ([DIC]carb) and DIC derived from sources external to the aquifer ([DIC]ext): ½DICtot ¼ ½DICcarb þ½DICext :

ð3Þ

The [DIC]ext is derived from both biogenic carbon dissolved in infiltrating groundwater ([DIC]inf) and deep source(s) ([DIC]deep) such as metamorphic decarbonation and/or the mantle. Assuming that (1) the measured [Ca], [Mg], and [SO4] of each groundwater sample were derived entirely from dissolution of Ca-Mg carbonates and sulfates and (2) no CO2 degassing, and consequently, carbonate mineral precipitation occurred prior to groundwater discharge at the surface, [DIC]carb was calculated according to: ½DICcarb ¼ ½Ca þ ½Mg−½SO4 :

ð4Þ

The aforementioned assumptions are tested and discussed in following sections. We then estimated [DIC] ext for each sample using Eq. (3), measured [DIC]tot, and calculated [DIC]carb. The carbon isotope mass balance of the system is described by the equation: 13

13

13

½DICtot ⋅δ Ctot ¼ ½DICcarb ⋅δ Ccarb þ ½DICext ⋅δ Cext ;

ð5Þ

where δ 13Ctot, δ 13Ccarb, and δ 13Cext are the isotopic values measured for [DIC]tot, representative of the carbonate aquifer(s), and

h  i  13 13 13 δ Cext ¼ δ Ctot −δ Ccarb ½DICcarb =½DICtot ½DICtot =½DICext

ð6Þ

assuming δ 13Ccarb was equal to the average value measured for local upper Paleozoic marine carbonate rocks (3‰; Muller and Mayo, 1983). A plot of δ 13Cext versus [DIC]ext was then used to estimate [DIC]inf, [DIC]deep, and δ 13Cdeep. The total and deeply derived DIC discharges were calculated for each spring by multiplying [DIC] tot and [DIC]deep, respectively, by the groundwater discharge rate. Further details are found in Chiodini et al. (1999, 2000). 4. Results 4.1. Physical and chemical characteristics of sampled features Physical parameters and chemical and isotopic compositions measured for the ten groundwater samples in this study are shown in Tables 1–3. Mammoth is a cold (15.8 °C) bubbling ([CO2]gas = 99.0 vol.%) spring that discharges in a grassy wetland area (Figs. 1 and 3a). A series of bubbling pools were located near the primary spring outlet, the extent of which we mapped using a hand-held GPS (Fig. 3a). Measured water discharge from Mammoth spring was 190 l s −1. Sulfur is a series of small, bubbling ([CO2]gas = 99.6 vol.%), cold (14.9 °C) springs and bubbling pools located at the mouth of Sulfur canyon (Fig. 1a). We mapped the extent of the bubbling pools with a hand-held GPS (Fig. 2). Free sulfur was evident in suspension in the spring waters. In addition, diffuse, “cold” CO2 and H2S degassing was observed within ~ 0.05 km 2 area around the springs and pools

Table 2 Noble gas data. Site

(R/Ra)mb

Hooper 2.096 Octagon 1.744 Mammoth 2.367 10 °C ASWd a b c d

±

(R/Ra)cc

0.054 2.119 0.196 1.750 0.089 2.389

±

[36Ar] cm3 cm−3 total gas

0.086 8.2398E −07 0.218 8.9164E −08 0.114 9.2754E −07 1.3376E −06

±

F(4He)a

±

F(22Ne) a

1.7788E−09 30.8248 1.8959 0.6270 3.6686E−10 31.8266 1.9614 0.2394 1.4344E−09 13.3425 0.8206 0.2127 0.2166 0.2722

±

F(84Kr) a

0.0125 1.2530 0.0160 2.4438 0.0065 2.1217 1.9412

±

F(132Xe)a

0.0116 1.8520 0.0210 7.4810 0.0158 4.3474 3.6767

F(i) = (i/36Ar)sample/(i/36Ar)air. (R/Ra)m = [(3He/4He)sample/(3He/4He)air]. (R/Rm)c = (R/Ra)m corrected for air/air saturated water using the measured F(4He) and F(22Ne) values. The composition of 10 °C air‐saturated water (ASW) was calculated from empirical parameters given in Smith and Kennedy (1983).

±

40

Ar/36Ar

0.0554 299.29 0.2074 300.55 0.1399 296.06

± 1.42 1.62 1.06

66

J.L. Lewicki et al. / Chemical Geology 339 (2013) 61–70

(Fig. 2). Total water discharge from Sulfur spring was estimated as 5 l s −1. Hooper is a bubbling ([CO2]gas = 70.0 vol.%) cold (10.2 °C) spring with a measured discharge rate of 19 l s −1. Octagon is a bubbling ([CO2]gas = 99.9 vol.%) cold (14.2 °C) spring that discharges from a travertine mound at an estimated rate of 2 l s −1. Lover's Delight is a bubbling cold (8.9 °C) spring that discharges from a travertine mound at an estimated rate of 2 l s −1. Barrel is a bubbling cold (8.6 °C) spring with a measured discharge rate of 17 l s −1. Outhouse is a series of numerous small, bubbling, warm (29.7 °C) springs discharging on the banks of Alexander Reservoir at total estimated rate of 8 l s −1. Ninety Percent and Mormon were the only sampled springs that were not bubbling. Ninety Percent is cold (10.5 °C), with a measured discharge rate of 91 l s −1. Mormon is a cold (13.9 °C) spring with a measured discharge rate of 20 l s −1. The Soda Springs geyser is located in downtown Soda Springs (Fig. 1a). Hourly geyser eruptions from the wellbore are controlled on a timer and warm (25.8 °C), bubbling water flows from the well bore between hourly eruptions at a measured rate of 51 l s −1. Outflow from the wellbore actively deposits travertine. The δ 13C values of DIC (δ 13Ctot) in all samples ranged from 1.3 to 4.6‰ (Table 1). The δ 13C values of CO2 gas (δ 13CCO2) collected from Hooper, Octagon, Mammoth, and Sulfur springs (− 0.3 to 2.3‰; Table 1) were lighter than the corresponding δ 13C of DIC, as expected based on isotopic fractionation between gaseous and aqueous carbon species over the measured temperature range (Friedman and O'Neil, 1977), and relative concentrations of aqueous carbon species over the measured pH range of the springs. Because the actual δ 13C composition of carbon in each of the sampled bubbling springs is somewhere between that of the CO2 gas and DIC, the measured δ 13C values of DIC are regarded as overestimates. Table 2 shows the noble gas abundances and isotopic compositions of Hooper, Mammoth, and Octagon spring gas samples. Helium isotopic compositions ((R/Ra.)c) ranged from 1.75 to 2.39. A plot of F( 132Xe) versus F( 22Ne) is shown in Fig. 4, where F(i) is equal to i/ 36 Ar for the sample divided by that ratio for air. The Hooper spring gas sample lies along a mixing line between the compositions of air and 10 °C air-saturated water (ASW). The relatively low [CO2]gas (70.0 vol. %) was measured in-situ at Hooper spring and was in agreement with a previous study that reported [CO2]gas of 72.79 vol.% and [N2]gas of 25.82 vol.% here (Jeandel et al., 2010). Contribution of air to Hooper spring gas may therefore have been due to natural spring dynamics, rather than contamination during sampling. Both Mammoth and Octagon spring gas samples have lower F( 22Ne) and higher F( 132Xe) values than 10 °C ASW (Fig. 4). The composition of Mammoth spring gas lies along a compositional trajectory of residual water after batch (single stage) gas loss at 10 °C from an initial composition like 10 °C ASW. The composition of the Octagon sample is consistent with groundwater degassing and subsequent mixing with

air. Addition of air may have occurred within the travertine mound from which Octagon spring issues.

4.2. Gaseous CO2 emissions at Mammoth and Sulfur springs Fig. 5a shows the net turbulent CO2 flux (Fc) time series measured by the EC station at Mammoth spring. Data gaps resulted from filtering according to quality control criteria. Fc ranged from − 74 to 1147 g m −2 d −1, with a mean and standard deviation of 350 and 281 g m −2 d −1, respectively. Variability of measured Fc was due to natural spatial and temporal variability of geologic and ecosystem CO2 fluxes within the study area, changes in the source weight function with time, and measurement uncertainty. On average during the observation period, 90% of measured Fc was sourced from a 0.05 km 2 area, within ~ 200 m of the EC station (Fig. 3). Relatively high Fc was measured when the EC station was located downwind of the primary spring outlet and bubbling pools (mean horizontal wind direction ~ 150–225 o; Figs. 3a and 5b). → A wsm of 20 was used in the inversion of F c and modeled source weight functions because it provided a compromise between spatial continuity across the model solution space and misfit between mea→ → sured and modeled F c(Fig. 6). Fig. 3b shows modeled Q c and CO2 emission rate within the EC average 90% flux source area. We modeled an → → area of relatively high Q c southwest of the EC station. Integrating Q c over the EC average 90% flux source area yielded a CO2 emission ; Fig. 3b). For reference, the influence rate of 24.9 tonnes d −1 (t d −1→ in of different wsm on modeled Q c and CO2 emission rate is shown → Fig. 7. With increasing wsm, the area(s) of relatively high Q c southwest of the EC station became smoother and spread out. The CO2 emission rate showed no systematic relationship with wsm, and varied from 23.4 to 30.2 t d −1 over the broad wsm range investigated. Soil CO2 fluxes measured in the Sulfur spring area ranged from b 1 to 52,178 g m −2 d −1 (Fig. 2). Observed variability of these measurements was likely due primarily to natural spatial variability of CO2 fluxes within the study area and to a lesser extent to natural temporal variability and measurement uncertainty. The highest CO2 fluxes were measured in areas with native sulfur deposition and strong H2S scent. Estimated total CO2 discharge from the 0.05 km 2 area was 49 t d −1.

Table 3 Discharge data. Estimated DIC discharges are in t CO2 d−1. Site

Hooper Octagon Mammoth Sulfur Geyser Outhouse Ninety Percent Barrel Lover's Delight Mormon a

Water discharge

Total DIC discharge

(l s−1)

(t d−1)

19 2a 190 5a 51 8a 91 17 2a 20

2.6 0.5 31.7 0.4 10.2 1.5 6.7 2.3 0.3 0.9

Discharge visually estimated.

±

DICdeep discharge

±

(t d−1) 0.5 0.1 6.4 0.1 2.4 0.3 1.3 0.4 0.1 0.2

1.5 0.3 19.3 0.4 10.0 1.5 1.1 1.3 0.2 0.0

0.3 0.1 4.0 0.1 2.4 0.3 0.4 0.3 b0.1 0.1

Fig. 4. Xenon-neon F-value plot with compositional trajectory of residual water after batch (single stage) gas separation at 10 °C from an initial composition like 10 °C air-saturated water (ASW; solid line). Dots show separation steps at different moles gas/moles water ratios. Dashed lines show mixing between air and 10 °C ASW and air and residual water after gas loss. Hooper, Octagon, and Mammoth spring compositions are plotted as gray dots.

J.L. Lewicki et al. / Chemical Geology 339 (2013) 61–70

Fig. 5. (a) Eddy covariance net CO2 flux (Fc) time series measured at Mammoth spring. (b) Fc versus mean horizontal wind direction.

4.3. Dissolved carbon discharges The calculated dissolved inorganic carbon concentrations derived from dissolution of carbonate minerals within the aquifer(s) ([DIC]carb; Eq. (4)) was positive for all samples except Sulfur, Outhouse and Soda Springs geyser. For these latter three samples, calculated [DIC]carb was less than zero, leading to [DIC]ext > [DIC]tot (Eq. (3)). Our assumption that the [Ca], [Mg], and [SO4] of groundwaters were derived entirely from dissolution of Ca-Mg carbonates and sulfates and neither CO2 degassing nor carbonate mineral precipitation occurred prior to groundwater discharge at the surface was likely violated for these samples (see Section 5 for further discussion). Therefore, for the purpose of carbon discharge calculations, we assumed that [DIC]ext = [DIC]tot for Sulfur spring, Outhouse spring, and Soda Springs geyser samples. Fig. 8 shows a plot of δ13Cext versus [DIC]ext for groundwater samples. Theoretical curves on Fig. 8 represent the compositional evolution of groundwaters resulting from addition of external carbon derived from a mantle source (δ13C = −6‰), local upper Paleozoic carbonate rocks (δ13C = 3‰), and a deep carbon source with δ 13C = 8‰ to an

67

infiltrating water with [DIC] = 1 mmol l −1 and δ13C = −17‰. The carbon isotopic composition of CO2 dissolved in infiltrating water (δ13Cinf) is poorly constrained for the study area and likely highly variable, due to a wide range of vegetation/ecosystem types (e.g., sagebrush steppe, mixed conifer forests, pastureland) within potential groundwater recharge areas. We therefore set δ 13Cinf approximately equal to the δ13Cext of Mormon spring (−17‰), which had a negligible [DIC]deep component. Accounting for 4–5‰ isotopic fractionation between the plant biomass and soil CO2 derived from respiration processes (Cerling et al., 1991) yields corresponding δ13C values of plant biomass within recharge areas from −21 to −22‰, which fall within the high range of values typical of C3 plants (O'Leary, 1988). Groundwater data in Fig. 8 can be explained by addition of deeply derived carbon with a δ13C value lying between the mantle (−6‰) and 8‰ to infiltrating water with [DIC]= 1 mmol l −1 and δ13C =−17‰. Most of these data indicate a deep carbon source with an isotopic composition close to that of local upper Paleozoic carbonate rocks (average value= 3‰). Supplement 1 shows a similar plot as that presented in Fig. 8, but uses a δ13Cinf of −23‰ to calculate theoretical curves. While the corresponding δ13C values of plant biomass (−27 to −28‰) are more typical of C3 plants (O'Leary, 1988), the [DIC]inf and δ 13C value(s) of the deep carbon end member(s) defined by the plot are similar to those defined in Fig. 8. We estimated [DIC]deep for each water sample by subtracting [DIC]inf (1 mmol l −1; Fig. 8) from [DIC]ext. Calculated total and deeply derived DIC (as CO2) discharges are given in Table 3. Errors of total and deeply derived DIC discharge estimates were propagated based on the uncertainties of chemical analyses and spring discharge measurements and assuming an uncertainty of ±1 mmol l −1 for the estimated [DIC]inf (Table 3). Potential CO2 degassing and associated carbonate mineral precipitation prior to sampling and/or the presence of SO4 in excess of that derived from dissolution of Ca-Mg sulfates would increase the uncertainty of our discharge estimates. However, because we have no quantitative information on these processes, they were not accounted for in error propagation. Total DIC discharges ranged from 0.3 (Lover's Delight) to 31.7 (Mammoth) t d −1, while deeply derived DIC discharges ranged from near zero (Mormon) to 19.3 (Mammoth) t d −1. Total DIC discharge for all sampled groundwater features was 57.1 t d −1. Of this quantity, approximately 3% was derived from biogenic carbon dissolved in infiltrating groundwater, 35% was derived from carbonate mineral dissolution within the aquifer(s), and 62% (35.5 t d −1) was derived from deep source(s). 5. Discussion and conclusions We demonstrated for the first time the ability to roughly map the location of and quantify geologic CO2 emissions from a focused groundwater source using an inversion of EC measured CO2 fluxes and modeled source weight functions (Fig. 3). The CO2 emission rate from the Mammoth spring area estimated based on the inversion (24.9 t d −1) was 79% of the total dissolved CO2 discharge estimated as the product of [DIC]tot and spring outflow rate (Table 3). High pCO2 springs like Mammoth rapidly re-equilibrate to atmospheric conditions upon discharge at the surface, degassing CO2 and precipitating carbonate minerals, for example: þ2

Ca

→

Fig. 6. Plot of log surface CO2 flux (Q )c roughness versus misfit between measured and modeled net CO2 flux, Fc (as weighted residual sum of squares; WRSS) for different smoothing weights (wsm).

þ 2HCO3

−

¼ CaCO3 þ H2 O þ CO2 :

ð7Þ

Chiodini et al. (1999) demonstrated that a high pCO2 spring in Italy lost more than 53% of its DIC to CO2 degassing almost immediately upon emergence at the surface. Thus, the observed difference between the CO2 emission rate estimated based on the EC inversion and the total dissolved CO2 discharge at Mammoth spring is likely reasonable and highly encouraging. These results suggest EC as a valuable tool for the mapping and quantification of CO2 emissions

68

J.L. Lewicki et al. / Chemical Geology 339 (2013) 61–70

→

Fig. 7. Modeled surface CO2 flux (Q )c and CO2 emission rate (DCO2) within the eddy covariance (EC) average 90% flux source area for different smoothing weights (wsm). White dots show EC station location. White boxes show extent of model domain.

from ground and surface water features such as springs and lakes. In addition, the spatial resolution of modeled surface CO2 fluxes (Fig. 3b) could be improved by deploying multiple EC stations in a network configured to maximize ability to map the surface CO2 flux distribution and quantify CO2 emission rate, while minimizing the number of stations required (Lewicki and Hilley, 2012). At Sulfur spring, the diffuse CO2 emission rate of 49 t d −1 from the 0.05 km 2 area estimated based on accumulation chamber measurements overwhelmed the total dissolved CO2 discharge of 0.4 t d −1.

Fig. 8. Diagram of δ13C of external DIC (δ13Cext) versus external DIC concentration ([DIC]ext) for groundwater samples (after Chiodini et al., 1999; 2000). Theoretical curves represent compositional evolution resulting from addition of carbon derived from a mantle source (δ13C = −6‰; dotted-dashed curve), local upper Paleozoic carbonate rocks (δ13C = 3‰; solid curve), and a deep carbon source with δ13C = 8‰ (dashed curve) to infiltrating water with [DIC] = 1 mmol l−1 and δ13C = −17‰ (gray box).

Furthermore, because CO2 emissions from bubbling pools (Fig. 2) were unaccounted for, this gaseous emission rate was an underestimate. Eddy covariance could be a useful alternative method to measure total CO2 emissions from areas such as Sulfur spring where degassing to the atmosphere occurs from both the soil and ground/ surface water features. The DIC in Ninety Percent and Mormon springs was primarily sourced from carbonate mineral dissolution within the aquifer(s), whereas carbon derived from deep source(s) external to the aquifer(s) contributed significantly to [DIC]tot in remaining samples (Table 3). In all groundwaters, biogenic carbon dissolved in infiltrating groundwater accounted for only 2–8% of [DIC]tot. Helium isotopic compositions of gases collected from Mammoth, Octagon, and Hooper springs (Table 2) and the Soda Springs geyser (Jeandel et al., 2010) indicated a mantle helium component. In addition, while CO2 produced by metamorphic decarbonation is enriched in 13C relative to the source rock (Chacko et al., 1991), the δ13Cext versus [DIC]ext diagram indicated addition of deeply derived carbon with a value similar to that of local carbonate rocks (Fig. 8). It is likely, therefore, that the deeply derived carbon is a mixture of carbon from mantle and metamorphic sources. Basin and Range extension and intrusion and/or eastern Snake River Plain intrusion are likely sources of heat for metamorphic reactions and mantle volatiles in groundwaters. The use of [Ca], [Mg], and [SO4] to estimate [DIC]carb (Eq. (3)) relied on the assumptions that these concentrations were derived entirely from dissolution of Ca-Mg carbonates and sulfates within the aquifer(s) and neither CO2 degassing nor carbonate mineral precipitation occurred prior to groundwater sampling (i.e., “no-sinks” assumption). Violations of the “no-sinks” assumption can cause [DIC]deep to be underestimated and δ 13C of the deep carbon source to be overestimated (e.g., Chiodini et al., 2000). On the other hand, if SO4 in groundwaters is in excess of that derived from dissolution of Ca-Mg sulfates, then [DIC]deep and δ 13C of the deep carbon source could be overestimated and underestimated, respectively. It is likely that some degree of degassing (± carbonate mineral deposition)

J.L. Lewicki et al. / Chemical Geology 339 (2013) 61–70

occurred at all of the bubbling springs prior to sampling, leading to underestimation of [DIC]deep. In particular, Octagon and Lover's Delight springs and the Soda Springs geyser actively deposit travertine, noble gas compositions showed evidence for degassing at Mammoth and Octagon springs, and Sulfur spring issues from within an area of high diffuse CO2 emissions. Several sources of the relatively high [SO4] observed at certain springs within the region (and H2S at Sulfur spring) have been proposed, including reactions involving organic-rich sedimentary rocks, gypsum, anhydrite, and/or native sulfur deposits (Hutsinpiller and Parry, 1985; Mayo et al., 1985). It is therefore possible that SO4 in excess of that derived from gypsum/anhydrite dissolution is present in the Outhouse, Soda Springs geyser, Sulfur, and Octagon samples, which could counteract to some extent the impact of CO2 degassing and carbonate mineral precipitation on the calculation of [DIC]deep. Sulfur isotopic analysis, however, would be required to better constrain SO4 origin in these high SO4 groundwaters. The sampled groundwater features in this investigation represented about 10% of groundwater discharge from the study area (Semenza, 2011). If we can roughly assume that the chemical and isotopic compositions of sampled features, and thus the deeply derived CO2 discharge estimated for these features (~ 35 t d −1), are representative of springs throughout the study area, then the total rate of deeply derived CO2 input into the groundwater system within this area could be ~ 350 t d −1. This value is similar to CO2 emission rates estimated for quiescent volcanoes such as Mammoth Mountain, USA (250–520 t d −1; Gerlach et al., 1999; Sorey et al., 1998), Vulcano, Italy (259–500 t d −1; Chiodini et al., 1996; Inguaggiato et al., 2012), Miyakejima, Japan (100–150 t d −1; Hernández et al., 2001), and Teide, Spain (380 t d −1; Hernández et al., 1998). This work provides a preliminary estimate of discharge of deeply sourced CO2 for the Soda Springs area. A comprehensive geochemical study, including radiocarbon as a tracer for deeply derived CO2 (Evans et al., 2002, 2006), of groundwaters throughout southeastern Idaho should be carried out to quantify the rate of non-volcanic CO2 degassing in this region. Supplementary related to this article can be found online at http:// dx.doi.org/10.1016/j.chemgeo.2012.06.013. Acknowledgments We thank two anonymous reviewers for their helpful comments on this manuscript and LI-COR, Inc. for the use of the LI-COR LI-8100A Automated Soil CO2 Flux System. We are grateful to the Hilda Thompson Ranch and town of Soda Springs, ID for site access and field logistical support. This work was funded by Assistant Secretary for Fossil Energy, Office of Sequestration, Hydrogen, and Clean Coal Fuels, NETL, of the U.S. Dept. of Energy under Contract No. DE-AC02-05CH11231. References Anderson, D.E., Farrar, C.D., 2001. Eddy covariance measurement of CO2 flux to the atmosphere from an area of high volcanogenic emissions, Mammoth Mountain, California. Chemical Geology 177, 31–42. Armstrong, F.C., 1969. Geologic map of the Soda Springs quadrangle, southeastern Idaho. U.S. Geol. Surv. Misc. Geol. Inv. Map I-557, scale 1:48,000. Armstrong, F.C., Oriel, S.S., 1965. Tectonic development of Idaho–Wyoming thrust belt. Bulletin of the American Association of Petroleum Geologists 49, 1847–1865. Armstrong, R.L., Leeman, W.L., Malde, H.E., 1975. K-Ar dating Quaternary and Neogene rocks of the Snake River Plain, Idaho. American Journal of Science 275, 225–251. Baldocchi, D.D., 2003. Assessing the eddy covariance technique for evaluating carbon dioxide exchange rates of ecosystems: past, present, and future. Global Change Biology 9, 479–492. Becker, J.A., Bickle, M.J., Galy, A., Holland, T.J.B., 2008. Himalayan metamorphic CO2 fluxes: quantitative constraints from hydrothermal springs. Earth and Planetary Science Letters 265, 616–629. Boyde, S., Vandenberghe, L., 2004. Convex Optimization. Cambridge Univ. Press, New York. 730 pp.

69

Buchanan, T.J., Somers, W.P., 1969. Discharge Measurements at Gaging Stations. U.S. Geol. Surv. Tech. Water Resour. Invest. Bk., 3. ch. A8, 65 pp. Cerling, T.E., Solomon, D.K., Quade, J., Bowman, J.R., 1991. On the isotopic composition of carbon in soil carbon dioxide. Geochimica et Cosmochimica Acta 55, 3403–3405. Chacko, T., Mayeda, T.K., Clayton, R.N., Goldsmith, J.R., 1991. Oxygen and carbon isotopic fractionation between CO2 and calcite. Geochimica et Cosmochimica Acta 55, 2867–2882. Chiodini, G., Frondini, F., Raco, B., 1996. Diffuse emission of CO2 from the Fossa crater, Vulcano Island (Italy). Bulletin of Volcanology 58, 41–50. Chiodini, G., Cioni, G.R., Guidi, M., Raco, B., Marini, L., 1998. Soil CO2 flux measurements in volcanic and geothermal areas. Applied Geochemistry 13, 543–552. Chiodini, G., Frondini, F., Kerrick, D.M., Rogie, J., Parello, F., Peruzzi, L., Zanzari, A.R., 1999. Quantification of deep CO2 fluxes from central Italy. Examples of carbon balance for regional aquifers and of soil diffuse degassing. Chemical Geology 159, 205–222. Chiodini, G., Frondini, F., Cardellini, C., Parello, F., Peruzzi, L., 2000. Rate of diffuse carbon dioxide Earth degassing estimated from carbon balance of regional aquifers: the case of central Apennine, Italy. Journal of Geophysical Research 105, 8423–8434. Deutsch, C.V., Journel, A.G., 1998. GSLIB: Geostatistical Software Library and User's Guide. Oxford Univ. Press, New York. 369 pp. Evans, W.C., Sorey, M.L., Cook, A.C., Kennedy, B.M., Shuster, D.L., Colvard, E.M., White, L.D., Huebner, M.A., 2002. Tracing and quantifying magmatic carbon discharge in cold groundwaters: lessons learned from Mammoth Mountain, USA. Journal of Volcanology and Geothermal Research 114, 291–312. Evans, W.C., Bergfeld, D., van Soest, M.C., Huebner, M.A., Fitzpatrick, J., Revesz, K.M., 2006. Geochemistry of low-temperature springs northwest of Yellowstone caldera: seeking the link between seismicity, deformation, and fluid flow. Journal of Volcanology and Geothermal Research 154, 169–180. Evans, M.J., Derry, L.A., France-Lanord, C., 2008. Degassing of metamorphic carbon dioxide from the Nepal Himalaya. Geochemistry, Geophysics, Geosystems 9, Q04021, http://dx.doi.org/10.1029/2007GC001796. Friedman, I., O'Neil, J.R., 1977. Compilation of Stable Isotope Fractionation Factors of Geochemical Interest. U.S. Geol. Surv. Prof. Pap. 440-KK, 12 pp. Frondini, F., Caliro, S., Cardellini, C., Chiodini, G., Morgantini, N., Parello, F., 2008. Carbon Dioxide Degassing from Tuscany and Northern Latium (Italy). Global and Planetary Change 61, 89–102. Gerlach, T.M., Doukas, M.P., McGee, K.A., Kessler, R., 1999. Airborne detection of diffuse carbon dioxide emissions at Mammoth Mountain. Geophysical Research Letters 26, 3661–3664. Harris, R., Segall, P., 1987. Detection of a locked zone at depth on the Parkfield, California segment of the San Andreas fault. Journal of Geophysical Research 92, 27945–27962. Hernández, P.A., Pérez, N.M., Salazar, J.M., Nakai, S., Notsu, K., Wakita, K., 1998. Diffuse emissions of carbon dioxide, methane, and helium-3 from Teide volcano, Tenerife, Canary Islands. Geophysical Research Letters 25, 3311–3314. Hernández, P.A., Salazar, J.M., Shimoike, Y., Mori, T., Notsu, K., Pérez, N., 2001. Diffuse emission of CO2 from Miyakejima volcano, Japan. Chemical Geology 177, 175–185. Hiyagon, H., Kennedy, B.M., 1992. Noble gases in CH4-rich gas fields, Alberta, Canada. Geochimica et Cosmochimica Acta 56, 1569–1589. Horst, T.W., Weil, J.C., 1992. Footprint estimation for scalar flux measurements in the atmospheric surface layer. Boundary-Layer Meteorology 2, 279–296. Hutsinpiller, A., Parry, W.T., 1985. Geochemistry and geothermometry of spring water from the Blackfoot reservoir region, southeastern Idaho. Journal of Volcanology and Geothermal Research 26, 275–296. Inguaggiato, S., Mazot, A., Diliberto, I.S., Inguaggiato, C., Madonia, P., Rouwet, D., Vita, F., 2012. Total CO2 output from Vulcano island (Aeolian Islands, Italy). Geochemistry, Geophysics, Geosystems 13, Q02012, http://dx.doi.org/10.1029/2011GC003920. Jeandel, E., Battani, A., Sarda, P., 2010. Lessons learned from natural and industrial analogues for storage of carbon dioxide. International Journal of Greenhouse Gas Control 4, 890–909, http://dx.doi.org/10.1016/j.ijggc.2010.06.005. Kennedy, B.M., Lynch, M.A., Reynolds, J.H., Smith, S.P., 1985. Intensive sampling of noble gases in fluids at Yellowstone: early overview of the data, regional patterns. Geochimica et Cosmochimica Acta 49, 1251–1261. Kerrick, D.M., 2001. Present and past nonanthropogenic CO2 degassing from the solid earth. Reviews of Geophysics 39, 565–585. Kerrick, D.M., McKibben, M.A., Seward, T.M., Caldeira, K., 1995. Convective hydrothermal CO2 emission from high heat flow regions. Chemical Geology 121, 285–293. Lewicki, J.L., Hilley, G.E., 2009. Eddy covariance mapping and quantification of surface CO2 leakage fluxes. Geophysical Research Letters 36, L21802, http://dx.doi.org/ 10.1029/2009GL040775. Lewicki, J.L., Hilley, G.E., 2012. Eddy covariance network design for mapping and quantification of surface CO2 leakage fluxes. International Journal of Greenhouse Gas Control 7, 137–144, http://dx.doi.org/10.1016/j.ijggc.2012.01.010. Lewicki, J.L., Hilley, G.E., Tosha, T., Aoyagi, R., Yamamoto, K., Benson, S.M., 2007. Dynamic coupling of volcanic CO2 flow and wind at the Horseshoe Lake tree kill, Mammoth Mountain, California. Geophysical Research Letters 34, L03401, http:// dx.doi.org/10.1029/2006GL028848. Lewicki, J.L., Fischer, M.L., Hilley, G.E., 2008. Six-week time series of eddy covariance CO2 flux at Mammoth Mountain, California: performance evaluation and role of meteorological forcing. Journal of Volcanology and Geothermal Research 171, 178–190, http://dx.doi.org/10.1016/j.jvolgeores.2007.11.029. Lewicki, J.L., Hilley, G.E., Fischer, M.L., Pan, L., Oldenburg, C.M., Dobeck, L., Spangler, L., 2009. Eddy covariance observations of surface leakage during shallow subsurface CO2 releases. Journal of Geophysical Research 114, D12302, http://dx.doi.org/ 10.1029/2008JD011297.

70

J.L. Lewicki et al. / Chemical Geology 339 (2013) 61–70

Lewicki, J.L., Hilley, G.E., Dobeck, L., Marino, B.D.V., 2012. Eddy covariance imaging of diffuse volcanic CO2 emissions at Mammoth Mountain, CA, USA. Bulletin of Volcanology 74, 135–141, http://dx.doi.org/10.1007/s00445-011-0503-y. Mabey, D.R., Oriel, S.S., 1970. Geologic Map of the Blackfoot Reservoir Area, Caribou County, Idaho. U.S. Geol. Surv. Prof. Paper 646-E. Mansfield, G.E., 1927. Geography, geology, and mineral resources of part of southeastern Idaho, with description of Carboniferous and Triassic fossils, by G.H. Girty. U.S. Geol. Surv. Prof. Paper, 152. 453 pp. Mayo, A.L., Muller, A.B., Ralston, D.R., 1985. Hydrogeochemistry of the Meade thrust allochthon, southeaster Idaho, U.S.A., and its relevance to stratigraphic and structural groundwater flow control. Journal of Hydrology 76, 27–61. Mitchell, J.C., 1976. Geothermal Investigations in Idaho, Part 6, Geochemistry and Geologic Setting of the Thermal and Mineral Waters of the Blackfoot Reservoir Area, Caribou County, Idaho. Idaho Dep. Water Rosour., Water Inf. Bull., No. 30. 47 pp. Muller, A.B., Mayo, A.L., 1983. Ground-water circulation in the Mead thrust allochthon evaluated by radiocarbon techniques. Radiocarbon 25, 357–372. O'Leary, M.L., 1988. Carbon isotopes in photosynthesis. Bioscience 38, 328–336. Rogie, J.D., Kerrick, D.M., Chiodini, C., Frondini, F., 2000. Flux measurements of nonvolcanic CO2 emissions from some vents in central Italy. Journal of Geophysical Research 105, 8435–8445. Sauer, V.B., Meyer, R.W., 1992. Determination of error in individual discharge measurements. U.S. Geol. Surv. Open File Rep, pp. 92–144. 21 pp.

Schmid, H.P., 1997. Experimental design for flux measurements: matching scales of observations and fluxes. Agricultural and Forest Meteorology 87, 179–200. Semenza, N.J., 2011. Map of the natural springs in the Soda Springs area, Caribou County, Idaho, M.S. Thesis, Idaho State University. Seward, T.M., Kerrick, D.M., 1996. Hydrothermal CO2 emission from the Taupo Volcanic Zone, New Zealand. Earth and Planetary Science Letters 39, 105–113. Smith, S.P., Kennedy, B.M., 1983. The solubility of noble gases in water and NaCl brine. Geochimica et Cosmochimica Acta 47, 503–515. Sorey, M.L., Evans, W.C., Kennedy, B.M., Farrar, C.D., Hainsworth, L.J., Hausback, B., 1998. Carbon dioxide and helium emissions from a reservoir of magmatic gas beneath Mammoth Mountain, California. Journal of Geophysical Research 103, 15,303–15,323. Werner, C., Brantley, S.L., Boomer, K., 2000a. CO2 emissions related to the Yellowstone volcanic system 2. Statistical sampling, total degassing, and transport mechanisms. Journal of Geophysical Research 105, 10,831–10,846. Werner, C., Wyngaard, J.C., Brantley, S.L., 2000b. Eddy-correlation measurement of hydrothermal gases. Geophysical Research Letters 27, 2925–2928. Werner, C., Chiodini, G., Voigt, D., Caliro, S., Avino, R., Russo, M., Brombach, T., Wyngaard, J., Brantley, S., 2003. Monitoring volcanic hazard using eddy covariance at Solfatara volcano, Naples, Italy. Earth and Planetary Science Letters 210, 561–577. Young, H.W., Mitchell, J.C., 1973. Geothermal investigations in Idaho, part 1, geochemistry and geologic setting of selected thermal waters. Idaho Dept. Water Resour. Water Inf. Bull., No. 30. 43 pp.