Chemical Geology 391 (2015) 123–137

Contents lists available at ScienceDirect

Chemical Geology

j ourna l homepage: www.e lsev ie r .com/ locate /chemgeo

Solute sources and water mixing in a flashy mountainous stream(Pahsimeroi River, U.S. Rocky Mountains): Implications on chemicalweathering rate and groundwater–surface water interaction

Benjamin Hagedorn a,⁎, Robert B. Whittier b,c

a Department of Geological Sciences, California State University, Long Beach, CA 90840, USAb Department of Geology and Geophysics, School of Ocean and Earth Sciences and Technology, University of Hawaii at Manoa, Honolulu, HI 96822, USAc State of Hawaii, Department of Health, Honolulu, HI 96813, USA

⁎ Corresponding author. Tel.: +1 562 985 4198.E-mail address: Klaus.Hagedorn@csulb.edu (B. Hagedo

http://dx.doi.org/10.1016/j.chemgeo.2014.10.0310009-2541/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 19 June 2014Received in revised form 23 October 2014Accepted 29 October 2014Available online 11 November 2014

Editor: David R. Hilton

Keywords:Chemical weatheringStrontium isotopesGroundwater–surface water interactionBaseflowRocky Mountains

Identifying solute sources and mixing processes between various water types is challenging in geologically di-verse, fractured rock settings, where various minerals contribute to solute loads andmixing between groundwa-ter and surface water can occur at difficult-to-delineate point locations. In such regions, chemical indicators thatallow constraining characteristic mineral fingerprints of drainage lithology or aquifer end-members are critical.This study assesses solute sources and water mixing in the Pahsimeroi River, a small streamwith highly variabledischarge draining twoheavily deformedmountain ranges of the U.S. RockyMountains. Solute inputs to themainstreamwere estimated using an end-membermixingmodelwith 87Sr/86Sr andNa/Sr ratios as tracers consideringthat these depict the compositional variability pertaining to the various carbonate and silicate lithologies of thebasin. Our results show that the mean solute input from carbonate-dominated terrains decreases from 95.9%in the headwaters to about 36.4% in the lowlandswhile the corresponding inputs fromvolcanic andmetamorphicsources increase from3.63% to 51.0% and 0.45% to 12.6%, respectively. Data further indicate highly variable chem-ical weathering rates with highest values observed in the steeper uplands. Calculated CO2 consumption rates(mean value: 0.14 Mmol km−2 a−1) are lower than the reported continental average (0.21 Mmol km−2 a−1)possibly due to lower than average streamflow at the time of sampling and significant input of carbonatesolute-enriched baseflow in the downstream sections where the basin shallows and decreases in width. Theherein delineated gaining stream segments are consistent with those deduced from downstream seepage runswhich suggests that groundwater sustains perennial flow in the agriculturally developed lowlands.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Assessing the long term reliability of water resources associatedwith Rocky Mountain streams is important because they comprisesensitive ecosystems and also supply water for hydroelectric power,irrigation and potable usage. Robust information about stream wateravailability and quality is critical for sustainable management of theseresources. In general, the availability of stream water depends uponcomponents the specific watershed's water budget; particularly precip-itation, evapotranspiration as well as the degree to which streams aredischarging to or are fed by local groundwater (baseflow). Streamwater quality and ability to provide potable supply or to supportwildlife, in turn, depend upon the watershed geology, climate, andland use regime(s) as well as the degree of mixing between surfaceand groundwater reservoirs. Climate change and increasing water

rn).

usage since the early to mid-20's century caused a significantdecline in streamflow in watersheds of the western U.S. and the RockyMountains in particular (Rood et al., 2005; Luce and Holden, 2009;Luce et al., 2013). Given that continuing decline in future decades islikely, assessments of sustainable limits for water extraction are critical.Such assessments require a detailed understanding of watermixing andsolute sources.

Documenting water mixing in streams is challenging given that itcan vary seasonally between wet and dry periods, and also spatially de-pending on local hydrogeologic conditions. For instance, in sedimentarysystems where surface runoff as well as diffuse recharge from localgroundwater can contribute to streamflow (Winter et al., 1998;Lambs, 2004; Langhoff et al., 2006; Anibas et al., 2011), numericalmodels relying on extensive climatic and hydrogeological parametersare commonly applied to quantify the contribution of various tributariesand baseflow to streamflow. These methods, however, are subject touncertainty particularly in crystalline (i.e., fractured) rock settingswhere mixing can occur at difficult-to-delineate point locations alongfractures or fault planes (Winter et al., 1998; Sophocleous, 2002). In

124 B. Hagedorn, R.B. Whittier / Chemical Geology 391 (2015) 123–137

125B. Hagedorn, R.B. Whittier / Chemical Geology 391 (2015) 123–137

such regimes, hydrochemical surveys can provide valuable informationon solute sources and water mixing if reliable source end-members canbe established.

Water mixing in streams has been analyzed using various tracers in-cluding major ions, trace compounds (e.g., CFCs), stable (O andH) isotopes and radioactive (e.g., 14C, 3H, 222Rn) isotopes (Genereux andPringle, 1997; Negrel et al., 2003; Negrel and Petelet-Giraurd, 2005;McCallum et al., 2010; Cartwright et al., 2011; Guida et al., 2013). Howev-er, selecting appropriate tracers for a specific setting is not straightfor-ward, considering that there can be site specific limitations. Forexample, the applicability of 18O and D tracers is compromisedwhen var-ious source waters (e.g., surface runoff, baseflow) and stream water dis-play overlapping values because they are derived from the samemeteoric water source. 222Rn values, which have been successfully usedto quantify baseflow contributions due to the 222Rn gradient betweengroundwater (high values due to prolonged water–rock interactions)and surface water (low values due to shortened water rock-interactions) may be unsuitable in turbulently flowing river segmentscharacterized by significant 222Rn degassing (Cartwright et al., 2014) orwhen infiltrating rainwater exhibits 222Rn concentration similar to riverwater values (Stellato et al., 2008). There may also be problems when222Rn values of local aquifer end-members cannot be reliably constrained,particularlywhenmultiple, mineralogically distinct aquifers contribute tobaseflow. In such cases, 87Sr/86Sr ratios can facilitate refined assessmentsas 87Sr/86Sr values specifically reflect chemical weathering input of Sr-bearing minerals (i.e., carbonates yielding high Sr contents but low87Sr/86Sr values, K-silicates yielding low Sr contents but high87Sr/86Sr ratios and Ca/Na plagioclase yielding intermediate Sr con-tents and low 87Sr/86Sr ratios) (Negrel et al., 1993; Galy et al.,1999; Grosbois et al., 2001; Harrington and Herzceg, 2003; Wuet al., 2009; Hagedorn et al., 2011). Additionally, major ion data – ifcorrected for atmospheric input and apportioned among appropriatechemical weathering reactions (Galy and France-Lanord, 1999;Spence and Telmer, 2005; Ryu and Jacobson, 2012) – may provideadditional constraints on chemical weathering and water solutesources.

This study applies 87Sr/86Sr ratios and major ion systematics tounderstand the contribution of water originating from mineralogi-cally distinct sub-basins to the Pahsimeroi River in the Idaho RockyMountains. The river is a major tributary of the Salmon River andhas become increasingly exploited for local agriculture irrigation. Itdrains two heavily deformed mountain ranges (Fig. 1) which areeach characterized by diverse mineralogies and numerous normalfaults that juxtapose impermeable bedrock with pervious alluvialsediments. As a result, groundwater discharges into the river viasprings or baseflow at multiple locations causing various high-flowand low-flow segments along the river reaches (Liberty et al.,2006). Because of increasing levels of groundwater extraction inthe watershed, there is a need for an understanding of the dominantwater source(s) along the entire reaches of the Pahsimeroi. Unlikemany water resources in the western U.S., there is the opportunityto use results to understand the hydrochemistry of the streamahead of major development. Such studies can thus be implementedin other areas where broad-scale watershed studies have identifiedthat there is a high risk of streamflow and water quality changedeveloping in the future.

Fig. 1. (a) Location of Pahsimeroi Watershed (ESRI, 2010), USGS and Idaho Power (IP) operateLane (2), Circle Pi (3), Dowton Lane (4), Burstedt Lane (5), Ellis (6), Morgan Creek Spring (7),Spring South Fork Grouse Creek (12), Grouse Creek (13), Double Springs (14), Merriam LakeFalls Creek (19), Big Creek Spring (20), Patterson Creek II (21), Patterson Creek I (22), Big CreeGoldburg Creek at Hatch Lane (27), Goldburg Creek I (28), Goldburg Creek II (29), Upper GoLower Spring (33), Circle Pi Upper Spring (34), Spring at the Flying Joseph (35), Baldwin SB. Hatch Well (40), Sulfur Creek Well (41), Sulfur Creek Pump (42), Church Well (43), MillerPrecipitation Station II (48), Precipitation Station IV (49), Upper Double Springs (50), Lower(54). (b) Simplified geology of study area (modified after Fisher et al., 1992; Link and Janecke,

2. Local physiographic setting

The Pahsimeroi Watershed is located in the central Idaho RockyMountains covering a total area of approximately 2150 km2. ThePahsimeroi River is about 95 km long and flows in a northwesterly direc-tion between the Lemhi Range to the northeast, and the Lost River Rangeto the southwest. Altitude changes from ~3180m above sea level (asl) inthe headwaters in the highlands of the Lost River Range to ~1490m asl atthe confluence with the Salmon River near the small community of Ellis(Fig. 1). Thewatershed is characterized by deeply incised, steepmountainterrain and narrow stream channels in the uplands and more gently dip-ping slopes in the central portions of the valley.

Precipitation is highest from November through February and lowestfrom June through September and ismainly received as snow in the high-lands (WRCC, 2014). Spatially, precipitation varies stronglywith elevationwith annual means as high as 1000 mm in the mountains and as low as250 mm/yr in the central valley (PRISM, 2010). Mean air temperaturesin Challis (located approximately 24 km southwest of Ellis) vary between−5.6 °C in January and 20.3 °C in July (WRCC, 2014) and generally de-crease at a mean lapse rate of about 0.4 °C per 100 m elevation increase(Blandford et al., 2008). Because the short annual growing season in thevalley (mid-June to the end of August) coincides with the annual dry pe-riod, approximately 0.15 km3 of water is diverted annually within thebasin to irrigate about 120 km2 (Williams et al., 2006).

The study area is only sparsely populated with most of the populace(~100 people) concentrated in May and Ellis. Land cover in the valleycomprises shrub/rangelands (55%), grass/pasture/hay lands (19%),natural forest (17%), water/wetlands (7%) and grain crops (2%) (NRCS,2008). The Pahsimeroi is an important anadromous fish spawningstream (Maret et al., 2005) and alike other rivers in the Rocky Moun-tains, it has experienced significant declines in resident fish (i.e., troutand salmon) populations; arguably caused by water diversions andclimate change-induced changes in water temperature (Rieman et al.,2007; Rieman and Isaak, 2010). This has led to concerned efforts tominimize out-of-stream diversions particularly in the upper reaches ofthe river, the main juvenile fish rearing space (SOI-DFG, 2014).

2.1. Local geology

The study area is located in the heavily deformed Basin and RangeProvince of the western U.S.; a region that experienced multiplelithospheric extension events and associated volcanism since the latePaleogene. The general stratigraphy in the region as mapped by Fisheret al. (1992), McIntyre and Hobbs (1987) and Janecke (1993) issummarized from oldest to youngest as: (1) Precambrian (Belt Super-group) metamorphic basement (quartzites and metagreywackes),(2) Palaeozoic limestone and dolomite containing varying amounts ofinterbedded siltstone, shale and lithic fragments, (4) Palaeogene Challisvolcanics (intermediate lava flows and associated sediments); and(5) recent basin fill deposits (alluvial and fluvial gravel, sand and silt)(Fig. 1). The basement rocks cover large proportions of the LemhiRange and comprise the Mid-Proterozoic Swauger Formation contain-ing mainly fluvio-lacustrine feldspathic arenites (N80% quartz, b12%K-feldspar; b1% lithic fragments) (Connor et al., 1991). Carbonaterocks occur mainly in the southwestern parts of the watershed in theLost River Range as thick limestone and dolomite banks (Fisher et al.,

d stream gages, and water sampling localities at: Headwater Pahsimeroi River (1), McCoyMorgan Creek (8) Lawson Creek (9), Sulfur Creek Spring (10), Carlson Lake Spring (11),(15), Leatherman Lake Spring II (16), Leatherman Lake Spring I (17), Morse Creek (18),k I (23), Iron Creek Spring (24), Big Springs Creek at Hooper Lane (25), Big Creek II (26),ldburg Creek I (30), Upper Goldburg Creek II (31), Cutler/Dowton Spring (32), Circle Piprings (36), Lone Pine Spring (37), Lone Pine Large Spring (38), IDF&G Well #2 (39),Well (44), Precipitation Station III (45), Precipitation Station I (46), Whittier Ranch (47),Double Springs (51), Carlson Lake (52), Merriam Lake Trail (53) and Leatherman Pass1999 and Liberty et al., 2006).

Fig. 2.Daily streamflow at Ellis, Burstedt Lane, Dowton Lane, P9 and Furey Lane (see Fig. 1 for stream gage locations). Streamflowbecomes perennial between the Furey Lane and P9 gages.

126 B. Hagedorn, R.B. Whittier / Chemical Geology 391 (2015) 123–137

1992; Link and Janecke, 1999). Volcanic rocks comprise intermediatelavas and associated pyroclastic rocks (McIntyre and Hobbs, 1987).Rhydacites occur in the northern section of the Lost River Range,where-as the western flanks mainly comprise K rich trachyandesites and to alesser extent latites (McIntyre et al., 1982; Moye et al., 1988; NormanandMertzman, 1991). Sulfides (mainly pyrite) occur asminor accessoryphases within most lithological units, either as a primary phase in theigneous rocks or as diagenetic phase in the (meta)sedimentary rocks.

Basin fill deposits comprise laterally discontinuous gravel, sand andsilt layers of eroded drainagematerials. Seismic surveys indicate a fairlysymmetrical drainage basin geometry with greatest depths (~2000 m)in the southern central portions at McCoy Lane (Fig. 1) and shallowestparts (~200 m) towards the northwest at Dowton Lane (Liberty et al.,2006). Numerous northwest–southeast trending normal faults alongthe valley boundaries and in its interior (e.g., Goldburg Fault) (Libertyet al., 2006) document the local extensional history. Springs occuralong these faults and form bordering wetlands (Fig. 1). These springsreportedly comprise the main source of streamflow in the watershed(Williams et al., 2006).

2.2. Local hydrology

Streamflow changes with season with highest values in late fall/early winter due to surface runoff pulses as well as irrigation returnflows. These characteristics are common in mountainous watershedsin the area (Tang et al., 2012); as is the typical El Niño-Southern Oscilla-tion pattern in discharge recordswith peak flows in Ellis (lowland gage;Fig. 1) occurring in the wet years 1984, 1998 and 2010 (Fig. 2).

Groundwatermovement in the area generally follows topography andoccurs down-valley to the northwest. However, a lack of concurrentwater level and well construction data for the Pahsimeroi Basin prohibitsa detailed evaluation of hydraulic gradients. Mean depth to groundwaterdata for 10 shallow wells (data from USGS National Water InformationSystem) varies between 5.56mand 32.1mover short (10s ofmeters) lat-eral distances indicating significant vertical gradients reflective for semi-confined conditions which are characteristic for laterally discontinuousfluvial sand/silt units. Meinzer (1924) reported that depth to the watertable along the Pahsimeroi River fluctuates N3 m over the course of ayear with the largest fluctuations occurring around irrigated tracts nearMcCoy Lane and Ellis. Seasonally, the water table rises rapidly in springdue to streamflow contributions and irrigation and is highest in June

and September and lowest in February and early May (Meinzer, 1924).Based on the generalized water budget by Williams et al. (2006; andreferences therein), groundwater underflow through the mouth of thebasin represents less than 1% of the annual precipitation for thePahsimeroi Basin indicating that groundwater flow through the basinoutlet is minimal with respect to the annual water budget.

In an effort to better understand the hydrology of the PahsimeroiWa-tershed, Williams et al. (2006) conducted seepage runs along the entirelength of the river in August (irrigation season) and November (post irri-gation season) of 2005 (Fig. 3). They characterized four primary reaches(in order of occurrence) as: (1) a losing upper reach from the headwatersand downstream for approximately 28.8 km, (2) a 9.17 km long gainingreach of extensive springs and wetlands in the vicinity of McCoy Lane,(3) a mid-valley 12.7 km long losing/sink reach, and (4) a lower29.8 km long gaining reach towards the outlet. The average gain/loss ofstreamflow estimated during the 2005 seepage runs is highly variableranging from ~−100 to +300 L/s/km (Williams et al., 2006; Fig. 3b).That there is an overall trend of increasing baseflow contributiontowards the lower reaches suggests that at least beyond Furey Lane,where the basin shallows and decreases in width, groundwater sustainsperennial flow.

The picture that emerges is that the Pahsimeroi Watershed is locatedin a complex geologic setting where water from 3 lithologic regimes con-tributes to streamflow: (1) the carbonate dominated portions of the LostRiver Range, (2) the volcanic dominated portions of the Lost River Range,and (3) the metamorphic (quartzite)-dominated sections of the LehmiRange. While sub-basin runoff and spring water sampled within the re-spective mountain ranges are likely characterized by watershedlithology-specific chemical fingerprints, water derived from within thebasin depocenter is expected to represent an integrated average of thechemical heterogeneity of bedrock formations. In the following sections,geochemical tracerswill be applied to characterize each of these potentialsources and assess their contribution to streamflow.

3. Materials and methods

Water from rivers, springs, precipitation and shallow domestic wellswas collected during eight sampling campaigns carried out betweenFebruary 2006 and October 2007 (Appendix A). Surfacewaterwas sam-pled at approximately 20 cm water depth during broad daylight. Allwells except the Sulfur Creek Well were equipped with submersible

Fig. 3. (a) Map of Pahsimeroi Basin showing the sinks and spring complex areas, and gaining and losing segments of the Pahsimeroi River as observed during seepage runs in August andNovember of 2005 (after Williams et al., 2006). (b) Distributed gains and losses observed along the Pahsimeroi River stretches during August and November 2005 seepage runs (afterWilliams et al., 2006).

127B. Hagedorn, R.B. Whittier / Chemical Geology 391 (2015) 123–137

128 B. Hagedorn, R.B. Whittier / Chemical Geology 391 (2015) 123–137

pumps and were sampled at hose bibs. The Sulfur CreekWell was sam-pled with an ABS pipe bailer. Rain and snow was collected from raingages from February to September 2007.

Collected samples were filtered through 0.45 μm membrane fil-ters, acidified (with 1% HNO3) for cation analysis and stored inHDPE bottles (for major ions) and glass VOA vials (for D and 18O sam-ples) free of headspace below 4 °C. Water pH, temperature and elec-tric conductivity were measured immediately after collection usingportable YSI probes that were calibrated before each sampling cam-paign. Alkalinity was determined in the field via Gran titration usingHach digital titrators and reagents. Major ions were analyzed at theUniversity of Hawaii by ion chromatography (Dionex) for anionsand by ICP-OES and AAS for cations. The precision of major ions in di-lute waters based on repeat analyses was generally better than±15%. The major ion charge balance error (CBE) is generally betterthan±10% (mean andmedian values of 3.7% and 0.4%, respectively).Some samples show an excess of cations which most likely reflectsinaccuracies in the titration of HCO3 particularly in dilute (EC b 50 μS/cm)waters.

Stable O and H isotopes were measured at the SIRFER Laboratoryof the University of Utah using a Finnigan MAT 252 mass spectrome-ter and an automated Finnigan MAT H/Device. δ18O and δD valueswere measured relative to internal PLRM and SLRM standards thatwere calibrated against accepted NIST and/or IAEA water referencematerials.

87Sr/86Sr values were analyzed on a VG Sector thermal ionizationmass spectrometer (TIMS) at the SOEST Isotope Laboratory of the Uni-versity of Hawaii at Manoa. All samples were dried down on a hotplateand re-dissolved in 3.5 M HNO3, and Sr was separated using standardion exchange techniques. Measured 87Sr/86Sr ratios were corrected forinstrumentalmass fractionation using 86Sr/88Sr of 0.1194. The total pro-cedural blanks for Sr varied between about 1 and 3 ng compared to sam-ple strontium loads of severalmicrograms. Reproducibility and accuracyof Sr isotope runs have been periodically checked by running the Stan-dard Reference Material NBS 987, with a mean 87Sr/86Sr value of0.7102438 ± 0.000014 (2σ standard deviation).

Fig. 4. δ18O vs. δD values of sampled waters along with the Global Meteoric Water Line (GMWprecipitation data.

4. Results and discussion

4.1. Basic hydrochemical patterns

Appendix A lists the geochemistry of the main stream PahsimeroiRiver (MSPR) waters subdivided according to sampling locations(Fig. 1) into headwaters (sampled about 60 km southeast of the conflu-ence), upper reaches (sampled at McCoy Lane, about 40 km southeast ofthe confluence), mid reaches (sampled between Hooper Lane andDowton Lane, between 11 and 20 km southeast of the confluence)and lowland (sampled between Burstedt Lane and Ellis, between 3.8and 0.2 km from the confluence). The dataset also incorporates thechemistries of tributaries/springs sampled within the metamorphicrock-dominated Lemhi Range (source waters — metamorphics) and thecarbonate/volcanic rock-dominated Lost River Range (source waters —carbonates, source waters — volcanics), respectively. Also included areprecipitation data as well as tributary, spring and groundwater samplescollected within the alluvial plains.

4.1.1. Stable isotopes (δD, δ18O)Values of δD and δ18O of sampled waters range from −128.5‰ to

−152.2‰ and from −17.2‰ to −19.6‰, respectively (Fig. 4). Allpoints align close to the global meteoric-water line (Craig, 1961) aswell as within the local precipitation (rain and snow) range indicatinga local meteoric water origin and a lack of significant evaporation oroxygen isotope exchanges between water and bedrock material. Sea-sonal changes are most pronounced in rainfall samples illustrated bythe Whittier Ranch samples that display a shift of δD and δ18O from−24.5‰ and −187.6‰ in February 2007 to −13.0‰ and −110.0‰in June 2007. Compared to precipitation, variability of δD and δ18Ovalues of surface water, spring water and groundwater samples is rela-tivelymuted and samples plot over a narrow range indicating amajorityof water input from summer precipitation (Fig. 4) and some degree ofmixing and homogenization affecting these waters along their flowpaths. There is no correlation between altitude and δ18O values of springwater and precipitation even if we subdivide the dataset into smaller

L; data from Craig, 1961) and Local Meteoric Water Line (LMWL) which is based on local

Fig. 5. Downstream trends in electric conductivity (a), δ18O (b) and 87Sr/86Sr (c) data.

129B. Hagedorn, R.B. Whittier / Chemical Geology 391 (2015) 123–137

sampling groups (e.g., summer vs. winter precipitation, Lost RiverRange vs. Lehmi Range mountain front precipitation). This observationis consistent with the complex climatic setting with variable airmass trajectories and local topography-driven weather phenomena(e.g., temperature inversions resulting from cold dry air pooling invalleys) that may cause high spatial temperature and relative humidityvariability (Whiteman et al., 2001; Blandford et al., 2008) and potential-ly high δ18O variability in precipitation.

Unlike other river systems in the world (Herzceg et al., 1993;Cartwright et al., 2011), the Pahsimeroi does not exhibit a steadydownstream increase in δD and δ18O (Fig. 5). This implies multiplewater inputs to the main stream along the entire reaches of the river.

4.1.2. Major ion systematicsThere are no significant seasonal electric conductivity variations

(Appendix A) in the MSPR; there is however a consistent downstreamtrend towards increasing electric conductivity values (Fig. 5a) possiblyreflecting the combined effects of progressive chemical weatheringand water mixing. Only two out of a total of 76 samples exhibit nitratelevels higher than EPA's maximum contaminant level (MCL) of10mg/L (EPA, 2014). These samples came from anunused groundwaterwell (Sulfur CreekWell) located in a pasture with an open borehole cov-ered by a sheet of plywood. The NO3 is likely introduced through theborehole and the samples were excluded from further analysis. Overall,however, agricultural (i.e., fertilizer) contamination does not appear tosignificantly affect waters in the basin and solute loads are mainly at-tributed to precipitation, chemical weathering and mixing.

Ionic ratios (Ca/Na vs. Mg/Na; Fig. 6) of MSPR headwater and car-bonate tributary samples align close to the theoretical carbonate lithol-ogy end-member indicating a dominant contribution of water derivedfrom the carbonate portions of the Lost River Range (e.g., Grouse Creekand/or Leatherman Lake Spring) to the headwaters. There is also an in-creasing input from silicate weathering in the MSPR mid reach to low-land samples implying that the contribution from the volcanic and/ormetamorphic terrains increases downstream.MSPR upper reachwaterssampled near McCoy Lane plot significantly closer to the silicate end-member than the alluvial spring samples (e.g., Lone Pine Spring, BaldwinSpring) that were collected in close proximity to these waters (Fig. 6).This would per se imply that the contribution of groundwater tostreamflow near McCoy Lane was only minor at the time of sampling.However, in this case, the discrepancy could also be ascribed to inflowfrom the Goldburg Creek tributary which originates in the metamorphicLemhi Range and is hence more influenced by a silicate major ion fin-gerprint. These competing major ion signatures (e.g., carbonate domi-nated spring vs. silicate dominated tributary waters) make aquantitative MSPR solute source assessment challenging. Given thataside from the carbonate tributary waters, all tributary, ground- andspring waters exhibit more or less overlapping major ion compositions,refinedmixing assessments and a clear water and solute source distinc-tion cannot be reliably conducted on the basis of cationic ratios alone.

4.1.3. Strontium isotopes87Sr/86Sr values in the Pahsimeroi Watershed range from 0.70797

(IDFG Well#2) to 0.75596 (Patterson Creek). There are no significantseasonal variations, as illustrated by the MSPR headwater and outletsamples which vary by only 0.00175 between August 2006 and Febru-ary/May 2007. In agreement with major ion trends discussed above,increasing 87Sr/86Sr values in downstreamMSPR samples imply increas-ing silicate weathering inputs (Fig. 5c). Spatially, 87Sr/86Sr are highest(N0.72) in waters of the metamorphic Lehmi Range and lowest(b0.72) in waters of the carbonate/volcanic rock dominated Lost RiverRange. There is a relative depletion of 87Sr/86Sr in MSPR and adjacentspring water and groundwater samples collected from the gainingstretches of the Pahsimeroi (Fig. 7a). This indicates that groundwaterand spring water derived from the Lost River Range recharge the riverand dilute the silicate weathering signature in the MSPR waters. A lack

of correlation between 87Sr/86Sr and 1/Sr (R2 b 0.2; relationship notshown), however, militates against simple binary (carbonate–silicate)mixing in the system.

Plots of 87Sr/86Sr vs. (Ca+Mg)/(Na+K), Ca/Sr andNa/Sr (Fig. 8a–c)provide more useful information on mixing as they are capable ofdepicting the compositional variability pertaining to the various carbon-ate and silicate lithologies (Negrel et al., 1993; Hagedorn et al., 2011).High 87Sr/86Sr values (N0.73) coupled with generally low (Ca + Mg)/(Na+K) ratios, such as those observed for themetamorphic sourcewa-ters, imply high contributions from K silicates; particularly orthoclase,biotite and other micas that contain elevated contents of incompatibleRb. Intermediate 87Sr/86Sr values (b0.73, N0.71) and highly variable(Ca + Mg)/(Na + K), Ca/Sr and Na/Sr ratios in the volcanic tributarysamples suggest input of less radiogenic Sr from plagioclase of the

Fig. 6. Ca/Na vs. Mg/Na values in Pahsimeroi waters (lithology end-member fields from Gaillardet et al., 1999).

130 B. Hagedorn, R.B. Whittier / Chemical Geology 391 (2015) 123–137

volcanic rocks which exhibit quite variable Ca and Na compositions inthe area (McIntyre et al., 1982; Norman and Mertzman, 1991). As Sr isexcluded from the carbonate lattice during water rock interaction, dis-solution of carbonate will lead to elevated Ca/Sr ratios. Thus, high Ca/Sr combined with low 87Sr/86Sr values (b0.71) and low Na/Sr, such asobserved for the carbonate source waters and MSPR headwater sam-ples, are consistent with primary (marine) limestone weathering.Even though solute input of accessory calcite with relatively high87Sr/86Sr values has been shown to be significant in metamorphic base-ment systems (Galy et al., 1999), like the one under consideration in thepresent study, the observation that metamorphic source waters displaythe lowest overall Ca/Sr valuesmilitates againstmore thanminor solutecontributions from these sources. However, accessory calcite derivedfrom igneous sources may very well exhibit the 87Sr/86Sr, Na/Sr, Ca/Srand (Ca + Mg)/(Na + K) patterns of Ca-plagioclase and may – assuch – bias the signal of the volcanic source waters towards a calciteweathering fingerprint. While the relative contribution of igneousrock-derived secondary calcite cannot be fully ruled out, a lack of mac-roscopic secondary calcite deposits in thefield suggests its solute contri-bution to be only of secondary importance in the volcanic sourcewaters.

MSPR and tributary samples exhibit distinct 87Sr/86Sr vs. (Ca+Mg)/(Na+K), Ca/Sr andNa/Sr trends reflective of the solute contributions ofthe dominant lithologies as well as atmospheric precipitation. In itsheadwaters, the MSPR retains a signature reflective for mainly carbon-ate rocks. This signature becomes increasingly mixed with silicate de-rived solutes as the waters drain the alluvial sediments of the basin'sdepocenter. Tributary inflow from the metamorphic Lehmi Range(e.g., Patterson Creek) and the volcanic sections of the Lost RiverRange (e.g., Morgan Creek) combined with baseflow input further ho-mogenizes the solute content ofMSPRwaters particularly in the gainingsegments near McCoy Lane and in the lowlands.

4.2. Chemical weathering rates

Chemical weathering of silicates by atmospheric H2CO3 and thesubsequent riverine transport of dissolved HCO3 to the ocean is animportant control on atmospheric CO2 levels over geologic time(Berner et al., 1983).Modeling studies using runoff and bedrock geologyinput data predict that rivers draining the U.S. Rocky Mountains, on a

continental and global scale, exhibit very high CO2 consumption ratesdue to the combined effects of runoff and lithology exposure(Amiotte-Suchet and Probst, 1995; Amiotte-Suchet et al., 2003;Moosdorf et al., 2011; Li et al., 2014). Given the abundance of silicate li-thologies in the Pahsimeroi Watershed, these high CO2 consumptionrate predictions should be reflected in the collected geochemistrydata. To estimate CO2 consumption in this study, solute contents wereapportioned among specific weathering reactions using major ion in-version techniques similar to those described in previous studies(Stallard and Edmond, 1981; Galy and France-Lanord, 1999; Spenceand Telmer, 2005; Ryu and Jacobson, 2012). Such techniques requirethe following assumptions: (1) dissolved cations are not fractionatedduring dissolution, (2) all Cl is derived solely from evapotranspirationof precipitation, (3) there is no major ion input from anthropogenicsources, and (4) the only acids that contribute to mineral weatheringare carbonic acid (derived from atmospheric or soil CO2) and, to a lesserdegree, sulfuric acid (derived from oxidation of sulfideminerals). Giventhese constraints and the local lithologic setting, the following fourweathering reactions are considered to contribute to solute yields inthe Pahsimeroi Watershed: (1) carbonic acid–carbonate weathering(CACW), (2) carbonic acid–silicate weathering (CASW), (3) sulfuricacid–carbonate weathering (SACW), and (4) sulfuric acid–silicateweathering (SASW). The herein presented major ion inversion tech-niques have generally only been applied to surface (stream)water sam-ples. Here, spring waters and groundwater are included in the analysisassuming they are derived from local mountain front recharge and arenot significantly affected by secondary (e.g., ion exchange) reactions.

As a first step, we filtered out the atmospheric solute input via:

i�W ¼ iW–ClW i=Clð ÞP ð1Þ

where i⁎ is the corrected concentration of solute i, and iW, ClW and ClPcorrespond to the measured concentration of i and Cl in the collectedwater (W) and precipitation (P) sample, respectively. The ID03 precipi-tation data used here comprises the geometric meanmajor ion value ofprecipitation that was collected at an elevation of 1807 m asl about140 km southeast of Ellis between 1980 and 2013 (NADP, 2014). Thisdata is considered a valid long-term proxy of mountain front precipita-tion chemistry in the study area.

Fig. 7. Natural neighbor interpolation of measured 87Sr/86Sr (a) and calculated mean CWRcarb/CWRsil (b) values in Pahsimeroi waters. Depicted gaining and losing stream segments aretaken from Williams et al. (2006).

131B. Hagedorn, R.B. Whittier / Chemical Geology 391 (2015) 123–137

132 B. Hagedorn, R.B. Whittier / Chemical Geology 391 (2015) 123–137

133B. Hagedorn, R.B. Whittier / Chemical Geology 391 (2015) 123–137

Next, we apportioned corrected solute concentrations among SACWand SASW, first calculating the input of SO4 as:

SO4SACW ¼ SO4�WR þ 1 ð2Þ

SO4SASW ¼ SO4�W–SO4SACW ð3Þ

where R refers to the ratio of carbonic acid consumed by silicateweathering to carbonic acid consumed by carbonate weathering (2).Because SACW generates a Ca, Mg vs. SO4 ratio of 2, cations derived bySACWwere calculated as:

CaSACW þMgSACWð Þ ¼ 2SO4SACW: ð4ÞSASW-derived cation yields were then calculated as:

NaSASW ¼ SO4SASW Ca=Nað Þsil þ Mg=Nað Þsil þ 0:5 K�=Na�

� �W þ 0:5

� �−1

ð5Þ

KSASW ¼ NaSASW K�=Na�

� �W ð6Þ

CaSASW ¼ NaSASW Ca=Nað Þsil ð7Þ

MgSASW ¼ NaSASW Mg=Nað Þsil: ð8Þ

Crucial input parameters in the calculation of weathering rates ininverse modeling studies are the silicate-bedrock Ca/Nasil and Mg/Nasilratios. Uncertainties in these ratios can impart a significant error inthe weathering yield estimates, particularly in basins that are not largeenough to average the chemical heterogeneity of bedrock forma-tions (Hren et al., 2007). Most past studies applied mean Ca/Nasiland Mg/Nasil ratios taken from whole rock chemical analyses. How-ever, these can be uncertain because fractionation caused byweatheringitself (Negrel et al., 1993) or inadequate study area coverage. It is there-fore important to apply Ca/Nasil and Mg/Nasil ratios of watershed-representative and relatively unaltered silicate bedrock samples. Toaccount for uncertainty in the calculated chemical weatheringrates, we utilize two sets of Ca/Nasil and Mg/Nasil ratios: (1) valuesof 0.63 (Ca/Nasil) and 0.92 (Mg/Nasil) corresponding to the geometricmean ionic ratios of silicate bedrock samples (equal proportions ofvolcanic and metamorphic rocks) collected in Central Idaho (datafrom McIntyre et al., 1982; Connor et al., 1991), and (2) values of1.50 (Ca/Nasil) and 2.06 (Mg/Nasil) which correspond to the geometricmean Ca/Na and Mg/Na values of the volcanic and metamorphic sourcewaters in the study area. These are assumed to be influenced only by sil-icate weathering without major contribution from secondary calcite dis-solution or atmospheric precipitation (mean precipitation Ca/Na andMg/Na = 0.17 and 0.30, respectively). The latter approach allows for amore localized end-member identification that takes into accountlocal ionic and isotopic considerations (Fig. 9a, b). Subsequent calculationsof chemical weathering and CO2 consumption rates are based on bothend-member ranges of Ca/Nasil and Mg/Nasil ratios.

Cation yields from CASW and CACW were computed as:

NaCASW ¼ Na�W–NaSASW ð9Þ

CaCASW ¼ NaCASW Ca=Nað Þsil ð10Þ

MgCASW ¼ NaCASW Mg=Nað Þsil ð11Þ

Fig. 8. 87Sr/86Sr vs. (Ca+Mg)/(Na+K) (a), Ca/Sr (b) and Na/Sr (c) inwaters collected in the Paand NADP (2014), respectively. Weathering pathways and water groups are indicated based ospanned by the mean tracer values of the volcanic, metamorphic and carbonate end-members

KCASW ¼ NaCASW K�=Na�

� �W ð12Þ

CaþMgð ÞCACW ¼ Ca� þMg�� �

W– CaþMgð ÞSACW– CaþMgð ÞSASW– CaþMgð ÞCASW:

ð13Þ

From these estimates, the chemical weathering rate (CWR) ofcarbonates (carb) and silicates (sil) were approximated by the totalmass of cations as:

CWRcarb ¼ CaþMgð ÞSACW þ CaþMgð ÞCACW ð14ÞCWRsil ¼ CaþMgþ Naþ Kð ÞSASW þ CaþMgþ Naþ Kð ÞCASW ð15Þ

and the CO2 consumption rate from carbonic acid weathering wascalculated as:

CO2rate ¼ 2Caþ 2Mgþ Naþ Kð ÞCASWQ=A ð16Þ

where Q is the Pahsimeroi River discharge at Ellis and A is the total wa-tershed area (data from USGS National Water Information System).

CWRcarb and CWRsil values range from 0.04–2.48 mmol/L and 0.04–1.45 mmol/L, respectively, using the whole rock Ca/Nasil and Mg/Nasilratios, and from 0–2.36 mmol/L and 0.07–2.58 mmol/L, respectively,using the water chemistry-deduced Ca/Nasil and Mg/Nasil values.Mean CWRcarb/CWRsil ratios (Fig. 7b) are generally highest (N1) in thecarbonate-dominated southwestern portion of the watershed andlowest (b0.1) in the metamorphic eastern portions of the watershed.Similar to the 87Sr/86Sr distribution,mean CWRcarb/CWRsil values exhib-it decreased silicate weathering rates in MSPR and adjacent springwater and groundwater sampled along the gaining stretches of theriver. This confirms an earlier hypothesis that northeasterly flowinggroundwater dilutes the MSPR silicate weathering signatures derivedfrom Lehmi Range tributaries. There is also a particular trend towardslower mean CWRcarb/CWRsil at higher elevations (and channel slopes)in the Lehmi Range stream waters, supporting observations by Hrenet al. (2007) that for monolithologic regimes, CWRsil fluxes are highestin areas of highest topographic drive.

Calculated CO2 consumption rates (Eq. (16)) at the MSPR outletrange from 0.074–0.117 Mmol km−2 a−1 using the whole rock Ca/Nasiland Mg/Nasil ratios and increase to 0.147–0.232 Mmol km−2 a−1 if thewater chemistry-deduced Ca/Nasil and Mg/Nasil ratios are applied. Thehigh sensitivity of calculated rates to the Ca/Nasil and Mg/Nasil ratios un-derscores the need for an accurate understanding of the lithologic com-position in the study area for reliable weathering rate assessments. Theherein presented rates are up to an order of magnitude higher thanthose of other rivers draining orogenic settings on the continent such asthe Yukon (0.019–0.023 Mmol km−2 a−1) or the Mackenzie (0.0004–0.067Mmol km−2 a−1) (Millot et al., 2003). These discrepancies likely re-late to the latitudinal effect (colder climate) and geologic setting withboth the Yukon and Mackenzie draining vast areas of stable basementshields constituted of resistant shales. More CO2 consumption data fromsmaller scalewatersheds in theRockyMountains are needed to systemat-ically evaluate the controlling factors on CO2 consumption in the area.

The herein reported CO2 consumption rates are similar to, if notslightly lower than theNorth American average of 0.21Mmol km−2 a−1

(Moosdorf et al., 2011) which is based on an empirical forward modelthat relates measured bicarbonate fluxes induced by rock-weatheringto catchment lithology and land cover. This is suprising given that re-gional forward models predict relatively high CO2 consumption ratesin Rocky Mountain streams. Given the methodological differences be-tween forward and inverse modeling approaches and the differentdatasets underlying both, it is not possible to reliably discern the causes

hsimeroiWatershed. Precipitation 87Sr/86Sr andmajor ion data are from Clow et al. (1997)n chemical and isotopic fingerprints. Also illustrated is the 3 end-member ternary domain(see text for details).

Fig. 9. 87Sr/86Sr vs. Ca/Na (a) and Mg/Na (b) in Pahsimeroi waters. Note the difference between geometric mean Ca/Nasil and Mg/Nasil ratios of whole rock samples in the area and thosededuced from the hydrochemistry of the silicate dominated (87Sr/86Sr N 0.713) volcanic and metamorphic source waters.

134 B. Hagedorn, R.B. Whittier / Chemical Geology 391 (2015) 123–137

for this discrepancy. Lower than average streamflow during the sam-pling campaigns (Fig. 2) could be one explanation; carbonate solute-enriched baseflow input at the Pahsimeroi outlet could be another. Inany case, the overall high variability in CWR and 87Sr/86Sr valuesthroughout thewatershed highlights the small scale variability of chem-ical weathering (and CO2 consumption) that can occur at thewatershedscale and this variability needs to be taken into consideration when ex-trapolating weathering rates to continental or even global scales(Hagedorn and Cartwright, 2009; Peucker-Ehrenbrinck and Fiske,2014). Further research should hence address CO2 consumption predic-tion model calibration. One means to achieve this goal could be usingthe outputs of local inverse modeling studies as calibration targets forlarger scale CO2 consumption models.

4.3. Quantifying the solute contribution of waters from volcanic, metamor-phic and carbonate terrains

To estimate the relative solute contributions of various lithologic re-gimes to the Pahsimeroi River, we linked chemical weathering finger-prints in MSPR waters to the various local bedrock end-members. Toachieve this, a mixing model was established using (a) 87Sr/86Sr, and(b) Na/Sr ratios as tracers. We considered these appropriate for thistask because they highlight the compositional variability of theindividual carbonate and silicate solute sources in the watershed. Theobservation that these tracers span a ternary domain that basicallyencompasses MSPR samples, spring waters and groundwater (Fig. 8c)supports this hypothesis. Assuming that atmospheric precipitation

135B. Hagedorn, R.B. Whittier / Chemical Geology 391 (2015) 123–137

affects all water samples to the same degree and the solute load of thePahsimeroi River is only controlled by input of metamorphic-, volcanic-and carbonate source waters, the contribution of each to the total soluteload can be calculated following Liu et al. (2004) as:

f 1 ¼C1river−C1

3

� �C22−C2

3

� �− C1

2−C13

� �C2river−C2

3

� �

C11−C1

3

� �C22−C2

3

� �− C1

2−C13

� �C21−C2

3

� � ð17Þ

f 2 ¼ C1river−C1

3

C12−C1

3

−C11−C1

3

C12−C1

3

f 1 ð18Þ

f 3 ¼ 1− f 1− f 2 ð19Þ

where ƒ1, ƒ2 and ƒ3 are the solute contributions from metamorphic,volcanic and carbonate sources, respectively, and C corresponds to thetracer concentration. The C superscript highlights the specific tracer(87Sr/86Sr for 1 and Na/Sr for 2) and the subscript denotes the component(i.e., river for the river sample, 1 for themetamorphic, 2 for the volcanic and3 for the carbonate end-member). Of allMSPR samples, only one headwa-ter sample plots outside the triangular domain possibly reflecting analyt-ical error and/or end-member uncertainty. For this outlier, a two end-member mixing solution solving the Pythagorean Theorem (Liu et al.,2004) was applied.

The biggest challenge in mixing assessments is accounting for end-member uncertainty (Genereux, 1998; Uhlenbrook and Hoeg, 2003). Inthis case, this is not a major issue for the carbonate or metamorphicsource waters, which each exhibit relatively low coefficients of variation(CV) for both 87Sr/86Sr and Na/Sr (b0.5). There is nevertheless a signifi-cant variability in the volcanic water Na/Sr values (CV: ~1), likely due tothe heterogeneity of the volcanic rocks in the area with compositionsranging from K-rich basalt to andesite (McIntyre and Hobbs, 1987). Weaccounted for this and other end-member variability by running themixing model iteratively using all possible combinations of end-members derived from either arithmetic mean or geometric mean valuesof the tracer data of the individual source water sampling groups(e.g., geomean tracer volcanics vs. arithmetic mean tracer carbonates vs.arithmeticmean tracermetamorphics, followed by arithmeticmean trac-er volcanics vs. geomean tracer carbonates vs. arithmetic mean tracermetamorphics, etc.). This procedure results in a total of 8model runs pro-viding a total possible range of input fractions illustrated in Fig. 10. Itshould be noted that we did not further assess the error caused by the

Fig. 10. Calculated input contributions from volcanic, carbonate and metamorphic source wateand in summer, metamorphic solute inflows are highest downstream and in summer and volc

different means of individual tracers (e.g., an end-member with an arith-meticmean 87Sr/86Sr value combinedwith a geometric mean Na/Sr valueor vice versa). Further research should address the integration ofmultiplemixing studies at the sub-watershed scale with more refined end-member tracer values to validate and refine the herein presented mixingassessment.

The results of this study show that the mean solute input from car-bonate terrains decreases from about 95.9% in the headwaters toabout 36.4% in the lowlands while the corresponding input from volca-nic and metamorphic sources increases from 3.63% to 51.0% and 0.45%to 12.6%, respectively. This indicates a clear dominance of water precip-itated over the Lost River Range on contributing to streamflow in thebasin. Although both bounding mountain ranges receive comparableannual precipitation (PRISM, 2010), the Lost River Range area is signif-icantly greater than that of the Lehmi Range and thus likely dominatesprecipitation, runoff and recharge volume. The herein presented datacould hence be useful for validating watershed models aimed at simu-lating the amount of streamflow generated from precipitation overthe Lost River vs. the Lehmi Ranges.

Seasonally, volcanic source input in the MSPR outlet samples islower in summer (mean for June, August and September samples:38.0%) than in winter (90.0% in February); a trend that is reversed forthe metamorphic (~15.0% vs. 5.68%) and carbonate (~47.1% vs. 4.39%)input. These seasonal patterns likely reflect the correspondingvariations in baseflow and sub-basin runoff. The increased meta-morphic and carbonate source contributions in summer (when MSPRstreamflow is low) could be indicative for increased baseflow inputof groundwater from fractured carbonate and, to a lesser degree,metamorphic rock aquifers. The increased volcanic rock solutesource input in winter (when MSPR streamflow is higher) could bedue to increased components of surface runoff enriched in solutesderived fromweathering of volcanic rock which make up the largestlithologic proportion of the watershed. However, more time-seriesgeochemical data particularly for the transitional spring and autumnmonths could strengthen our understanding of the hydrogeologicalcontrols on water and solute transport.

4.4. Assessing baseflow

There is a remarkable heterogeneity in alluvial spring andgroundwater chemistry in relation to landscape position and aquifer

rs along the Pahsimeroi River. Carbonate solute inputs are highest upstream (headwaters)anic solute inputs are highest downstream and in winter.

Fig. 11. Conceptualmodel of groundwater and surfacewater interaction in the PahsimeroiWatershed. Baseflow input increases downgradient as the basin decreases in depth andwidth. Localized baseflow input near McCoy Lane occurs where alluvial groundwaterreaches the decreased permeability of the footwall of the Goldburg Fault.

136 B. Hagedorn, R.B. Whittier / Chemical Geology 391 (2015) 123–137

mineralogy. ChurchWell (sample 43 in Fig. 1), for example, is locatedat themountain fronts of the Lemhi Range and sampled groundwaterdisplays 87Sr/86Sr vs. (Ca + Mg)/(Na + K), Ca/Sr and Na/Sr valuesmuch different to groundwater or spring water sampled closer tothe Lost River Range (e.g., IDF&G Well 2, sample 39 in Fig. 1). SulfurCreek Pump (sample 42 in Fig. 1), which is located in the basin interiorand in close proximity to the MSPR exhibits intermediate values(Fig. 8a–c) consistent with mixing and homogenization of carbonateand metamorphic mountain front end-members.

The overall variability of groundwater and spring water samplesprohibits a direct quantification of baseflow input to the Pahsimeroibased on simple end-member mixing analysis. Nonetheless, observedtrends allow some qualitative assessment of the extent of groundwater–surfacewater interaction in the basin. Upper reach to lowlandMSPR sam-ples show distinct 87Sr/86Sr vs. (Ca + Mg)/(Na + K), and Na/Sr valuesaligning over a relatively narrow range bracketed by the Circle Pi UpperSpring (sample 34 in Fig. 1),Dowton Spring (sample 32 in Fig. 1) and SulfurCreek Pump end-members (Fig. 8). This suggests a common solute sourceand/or ongoingmixingwith baseflowstarting at least atMcCoy Lane. Thisalso confirms the seepage runs by Williams et al. (2006) that character-ized gaining segments upstream of McCoy Lane and Furey Lane (Fig. 3).MSPR headwater samples, on the other hand, plot outside the alluvialspring water/groundwater range (Fig. 8) which is indicative for a lack ofhydraulic communication between surface water and groundwater (los-ing segment) and also consistent with Williams et al.'s (2006) observa-tions. That these distinct signatures of the headwater samples are notfully reflected in the 87Sr/86Sr vs. Ca/Sr plot is likely due to the overlappingCa/Sr values of waters characterized by carbonate and Ca-silicateweathering. Unfortunately, no MSPR data was collected from the secondreported losing segment of the MSPR between McCoy and Furey Lane,so the nature of this segment cannot be confirmed in the present study.However, unlike the gaining stretch upgradient of Furey Lanewhich is as-cribed mainly to topographic control (baseflow input increases as thebasin shallows and decreases in width), the gaining stretch at McCoyLane occurs where the basin is widest and deepest. Liberty et al. (2006)linked this segment to the Goldburg Fault (Fig. 1) forcing easterly flowinggroundwater upwards when it encounters impervious material at thefootwall. This general conceptual model is depicted in Fig. 11. More spa-tially distributed data particularly from the upper reaches of thePahsimeroi would allow better constraining the location and time atwhichMSPR 87Sr/86Sr vs. Na/Sr and (Ca+Mg)/(Na+K) values align out-side of the alluvial spring and groundwater rangewhichwould indicate achange from gaining to losing conditions.

5. Conclusions and implications

Geochemical data indicate that the Pahsimeroi's solute content ismainly derived from the volcanic-rock dominated sections of theLost River Range making these sections and the tributaries thatdrain them most important with regards to anthropogenic(e.g., fertilizer-derived) contamination. The question whether theinflow component of the volcanic source waters reflects increasedtributary/baseflow input from the respective sub-watersheds or aclear dominance of volcanic rock-derived sediments in the basindepocenter needs to be further investigated. Tributary flow datafrom the various sub-watersheds and petrographic information onbasin fill sediments could potentially shed more light on the respec-tive controls on the stream solute load.

In agreement with previously performed seepage runs, geochemicaldata depict a trend towards increasing downstream baseflow inputsthat sustain perennial flow downgradient of McCoy Lane. Quantitativebaseflow estimates, however, cannot be provided due to the chemicalvariability of alluvial groundwater and spring water samples. 222Rn in-formation may provide useful constraints on the extent and rate ofbaseflow input if groundwater and surface water end-members can bereliably identified.

Current and future water diversions for agricultural or other usesshould focus on the downgradient sections (i.e., beyond Furey Lane)and the distinct upgradient stretch (i.e., around McCoy Lane) of thePahsimeroi where continuous baseflow input prevents the river fromrapidly drying up. However, future development should include timeseries streamflow and water quality monitoring to trace the zonewhere the water table intercepts the river channel altitude and, moreimportantly, to documenthow this zone responds to increasing ground-water extraction and irrigation.

Acknowledgments

This study was funded by the U.S. Department of Interior, Bureauof Land Management (grant numbers: DLA060120, DLA060223) andthrough a startup grant (grant number: 2013GF00100337) to B.H.from California State University, Long Beach. The authors wouldlike to thank the numerous land owners in the Pahsimeroi Valleyfor their assistance during the field campaigns and for their sharingof experiences from years of working in the area. Joe Lichwa, DougPyle, Ken Rubin and SIRFER staff are acknowledged for their supportwith the geochemical analyses. We also thank David Hilton, AndrewJackson and an anonymous reviewer for their helpful comments andsuggestions that improved this manuscript.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.chemgeo.2014.10.031.

References

Amiotte-Suchet, P., Probst, J.L., 1995. A global-model for present-day atmospheric soil CO2

consumption by chemical erosion of continental rocks (GEM-CO2). Tellus 47B,273–280.

Amiotte-Suchet, P., Probst, J.L., Ludwig, W., 2003. Worldwide distribution of conti-nental rock lithology: implications for the atmospheric/soil CO2 uptake by conti-nental weathering and alkalinity river transport to the oceans. Glob. Biochem.Cycles 17, 1048.

Anibas, C., Buis, K., Verhoeven, R., Meire, P., Batelaan, O., 2011. A simple thermal mappingmethod for seasonal spatial patterns of groundwater–surface water interaction. J.Hydrol. 397, 93–104.

Berner, R.A., Lasaga, A.C., Garrels, R.M., 1983. The carbonate silicate geochemical cycle andits effect on atmospheric carbon dioxide over the past 100 million years. Am. J. Sci.283, 641–683.

Blandford, T.K., Humes, B., Harshburger, B., Moore, B., Walden, V., Ye, H., 2008. Seasonaland synoptic variations in near-surface air temperature lapse rates in a mountainousbasin. J. Appl. Meteorol. Climatol. 47, 249–261.

Cartwright, I., Hoffmann, H., Sirianos, M., Weaver, T.R., Simmons, G.T., 2011. Geochemicaland 222Rn constraints on baseflow to the Murray River, Australia, and timescales forthe decay of low-salinity groundwater lenses. J. Hydrol. 405, 333–343.

137B. Hagedorn, R.B. Whittier / Chemical Geology 391 (2015) 123–137

Cartwright, I., Hoffmann, H., Gilfedder, B., Smyth, B., 2014. Understanding parafluvial ex-change and degassing to better quantify groundwater inflows using 222Rn: the KingRiver, southeast Australia. Chem. Geol. 380, 48–60.

Clow, D., Mast, M.A., Bullen, T.D., Turk, J.T., 1997. Strontium 87/strontium 86 as a tracer ofmineral weathering reactions and calcium sources in an alpine/subalpine watershed,Loch Vale, Colorado. Water Resour. Res. 33, 1335–1351.

Connor, J.J., Evans, K.V., Briggs, P.H., Crock, J.G., Hopkins, J.R.T., Kennedy, K.R., O'Leary, R.M.,Roushey, B.H., Slaughter, K.E.,Welsch, E.P., 1991. Chemical analyses of rock samples col-lected in east-central Idaho during a study of mineralization in the northern. U.S. Geo-logic Survey Open File Report 90-0655. Lemhi Range, Lemhi County, Idaho, p. 32.

Craig, H., 1961. Isotopic variations in meteoric waters. Science 133, 1702-1702.EPA, 2014. Environmental Protection Agency — National Primary Drinking Water

Regulations. http://water.epa.gov/drink/contaminants/.ESRI, 2010. ArcInfo 10, Redlands, Ca.Fisher, F.S., McIntyre, D.H., Johnson, K.M., 1992. Geologic map of the Challis 1° × 2°

quadrangle. Geologic Survey Miscellaneous Investigations Series Map I–1819.Gaillardet, J., Dupre, B., Louvat, P., Allegre, 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., 1999. Weathering processes in the Ganges–Brahmaputra basinand the riverine alkalinity budget. Chem. Geol. 159, 31–60.

Galy, A., France-Lanord, C., Derry, L., 1999. The strontium isotopic budget of HimalayanRivers in Nepal and Bangladesh. Geochim. Cosmochim. Acta 63, 1905–1925.

Genereux, D.P., 1998. Quantifying uncertainty in tracer-based hydrograph separations.Water Resour. Res. 34, 915–919.

Genereux, D.P., Pringle, C., 1997. Chemical mixing model of streamflow generation at LaSelva biological station, Costa Rica. J. Hydrol. 199, 319–330.

Grosbois, C., Negrel, P., Grimaud, D., Fouillac, C., 2001. An overview of dissolved andsuspended matter fluxes in the loire river basin: Natural and anthropogenic inputs.Aquat. Geochem. 7, 81–105.

Guida, M., Guida, D., Guadagnuolo, D., Cuomo, A., Siervo, V., 2013. Using Radon-222 as aNaturally Occurring Tracer to investigate the streamflow-groundwater interactions in atypical Mediterranean fluvial-karst landscape: the interdisciplinary case study of theBussento river (Campania region, Southern Italy). WSEAS Trans. Syst. 12, 85–104.

Hagedorn, B., Cartwright, I., 2009. Climatic and lithologic controls on the temporal andspatial variability of CO2 consumption via chemical weathering: an example fromthe Australian Victorian Alps. Chem. Geol. 260, 234–253.

Hagedorn, B., Cartwright, I., Raveggi, M., Maas, R., 2011. Rare earth element and strontiumgeochemistry of the Australian Victorian Alps drainage system: evaluating the dom-inance of carbonate vs. aluminosilicate weathering under varying runoff. Chem. Geol.284, 105–126.

Harrington, G.A., Herzceg, A.L., 2003. The importance of silicate weathering of a sedimen-tary aquifer in Central Australia indicated by very high Sr-87/Sr-86 ratios. Chem. Geol.199, 281–292.

Herzceg, A.L., Simpson, H.J., Mazor, E., 1993. Transport of soluble salts in a large semiaridbasin: River Murray, Australia. J. Hydrol. 144, 59–84.

Hren, M.T., Chamberlain, C.P., Hilley, G.E., Blisniuk, P.M., Bookhagen, B., 2007. Major ionchemistry of the Yarlung Tsangpo–Brahmaputra river: chemical weathering, erosion,and CO2 consumption in the southern Tibetan plateau and eastern syntaxis of theHimalaya. Geochim. Cosmochim. Acta 71, 2907–2935.

Janecke, S., 1993. Structures in segment boundary zones of the Lost River and Lemhifaults, east central Idaho. J. Geophys. Res. 98, 16223–16238.

Lambs, L., 2004. Interactions between groundwater and surface water at river banks andthe confluence of rivers. J. Hydrol. 288, 312–326.

Langhoff, J.H., Rasmussen, K.R., Christensen, S., 2006. Quantification and recogionalizationof groundwater–surface water interaction along an alluvial stream. J. Hydrol. 320,342–358.

Li, D.D., Jacobson, A.D., McInerney, D.J., 2014. A reactive-transport model for examiningtectonic and climatic controls on chemical weathering and atmospheric CO2 con-sumption in granitic regolith. Chem. Geol. 365, 30–42.

Liberty, L.M., Hess, S., Beukelmann, G., 2006. Stratigraphic and structural controls ofground water flow in the Pahsimeroi Basin, Idaho: insights from geophysical data.Technical Report BSU CGISS 06-01 Prepared for the Bureau of Land Management.Center for Geophysical Investigation of the Shallow Subsurface (CGISS). Boise StateUniversity, Boise, ID, p. 15.

Link, P.K., Janecke, S., 1999. Geology of east-central Idaho: geologic roadlogs for the Bigand Little Lost River, Lemhi, and Salmon River Valleys. In: Hughes, S.S., Thackray,G.D. (Eds.), Guidebook to the Geology of Eastern Idaho. Museum of Natural History,Pocatello, Idaho, pp. 295–334.

Liu, F., Williams, M., Caine, N., 2004. Source waters and flowpaths in a seasonally snow-covered catchment, Colorado Front Range, USA. Water Resour. Res. 40, W09401.

Luce, C.H., Holden, Z.A., 2009. Declining annual streamflow distributions in the PacificNorthwest United States, 1948–2006. Geophys. Res. Lett. 36, L16401.

Luce, C.H., Abatzoglou, J.T., Holden, Z.A., 2013. The missing mountain water: slower west-erlies decrease orographic enhancement in the Pacific Northwest USA. Science 342,1360–1364.

Maret, T.R., Hortness, J.E., Ott, D.S., 2005. Instream flow characterization of upper SalmonRiver Basin streams. Central Idaho, 2004, U.S. Geological Survey Scientific Investiga-tions Report 2005–5212, p. 124.

McCallum, J.L., Cook, P.G., Brunner, P., Berhane, D., 2010. Solute dynamics during bankstorage flows and implications for chemical base flow separation. Water Resour.Res. 46, W07541.

McIntyre, D.H., Hobbs, S.W., 1987. Geologic map of the Challis Quadrangle, Custer andLemhi Counties, Idaho. U.S. Geological Survey Geologic Quadrangle Map GQ-1599,scale 1:62,500.

McIntyre, D.H., Ekren, E.B., Hardyman, R.F., 1982. Stratigraphic and structural frameworkof the Challis volcanics in the eastern half of the Challis 1° × 2° Quadrangle, Idaho. In:Bonnichsen, B., Breckenridge, R.M. (Eds.), Cenozoic Geology of Idaho. Idaho Bureau ofMines and Geology Bulletin 26, pp. 3–22 (Boise, ID).

Meinzer, O.E., 1924. Groundwater in Pahsimeroi Valley, Idaho. Idaho Bureau of Mines andGeology, Pamphlet 9, p. 36.

Millot, R., Gaillardet, J., Dupre, B., Allegre, C.J., 2003. Northern latitude chemicalweathering rates: clues from the Mackenzie River Basin, Canada. Geochim.Cosmochim. Acta 67, 1305–1329.

Moosdorf, N., Hartmann, J., Lauerwald, R., Hagedorn, B., Kempe, S., 2011. Atmospheric CO2

consumption by chemical weathering in North America. Geochim. Cosmochim. Acta75, 7829–7854.

Moye, F.J., Hackett, W.R., Blakeley, J.D., Snider, L.G., 1988. Regional geologic setting andvolcanic stratigraphy of the Challis Volcanic Field, Central Idaho. In: Link, P.K.,Hackett, W.R. (Eds.), Guidebook to the Geology of Central and Southern Idaho.Idaho Geologic Survey Bulletin 27, pp. 87–97.

NADP, 2014. National Atmospheric Deposition Program. http://nadp.sws.uiuc.edu/.Negrel, P., Petelet-Giraurd, E., 2005. Strontium isotopes as tracers of groundwater-in-

duced floods: the Somme case study (France). J. Hydrol. 305, 99–119.Negrel, P., Allegre, C.J., Dupre, B., Lewin, E., 1993. Erosion sources determined by inversion

of major and trace element ratios and strontium isotopic ratios in river water: theCongo Basin case. Earth Planet. Sci. Lett. 120, 59–76.

Negrel, P., Petelet-Giraurd, E., Barbier, J., Gautier, E., 2003. Surface water–groundwater in-teractions in an alluvial plain: chemical and isotopic systematics. J. Hydrol. 277,248–267.

Norman, M.D., Mertzman, S.A., 1991. Petrogenesis of Challis volcanics from central andsouthwestern Idaho: trace element and Pb isotopic evidence. J. Geophys. Res. 96,13279–13293.

NRCS, 2008. National Resource Conversation Service: Pahsimeroi— 17060202 HydrologicUnit Profile (March 2008). , p. 39.

Peucker-Ehrenbrinck, B., Fiske, G., 2014. Systematic biases in estimates of continental87Sr/86Sr runoff. 2014 Goldschmidt Conference Presentation and Abstract. Sacramen-to, USA.

PRISM, 2010. Parameter-elevation Relationships on Independent Slopes Model —

precipitation maps. Northwest Alliance for Computational Science and Engineering,(http://www.prism.oregonstate.edu/).

Rieman, B.E., Isaak, D., 2010. Climate change, aquatic ecosystems, and fishes in the RockyMountain West: Implications and alternatives for management. Gen. Tech. Rep.RMRS-GTR-250. U.S. Department of Agriculture, Forest Service, Rocky Mountain Re-search Station, Fort Collins, CO, p. 46.

Rieman, B.E., Isaak, D., Adams, S., Horan, D., Nagel, D., Luce, C., Myers, D., 2007. Anticipatedclimate warming effects on bull trout habitats and populations across the interiorColumbia River Basin. Trans. Am. Fish. Soc. 136, 1552–1565.

Rood, S.B., Samuelson, G.M., Weber, J.K., Wywrot, K.A., 2005. Twentieth-centurydecline in streamflows from the hydrographic apex of North America. J. Hydrol.306, 215–233.

Ryu, J.-S., Jacobson, A.D., 2012. CO2 evasion from the Greenland Ice Sheet: a new carbonclimate feedback. Chem. Geol. 320–321, 80–95.

SOI-DFG, 2014. State of Idaho, Department of Fish and Game — State fish hatcheries.http://fishandgame.idaho.gov/public/fish/?getPage=84.

Sophocleous, M.A., 2002. Interactions between groundwater and surface water: the stateof the science. Hydrogeol. J. 10, 52–67.

Spence, J., Telmer, K., 2005. The role of sulfur in chemical weathering and atmosphericCO2 fluxes: evidence from major ions, δ13CDIC and δ34SSO4 in rivers of the CanadianCordillera. Geochim. Cosmochim. Acta 69, 5441–5458.

Stallard, R.F., Edmond, J.M., 1981. Geochemistry of the Amazon 1. Precipitation chemistryand the marine contribution to the dissolved load at the time of peak discharge. J.Geophys. Res. 86, 9844–9858.

Stellato, L., Petrella, E., Terrasi, F., Belloni, P., Belli, M., Sansone, U., Celico, F., 2008. Somelimitations in using 222Rn to assess river–groundwater interactions: the case of Casteldi Sangro alluvial plain (central Italy). Hydrogeol. J. 16, 701–712.

Tang, C., Crosby, B.T., Wheaton, J.M., Piechota, T.C., 2012. Assessing streamflow sensitivityto temperature increases in the Salmon River Basin, Idaho. Glob. Planet. Chang.88–89, 32–44.

Uhlenbrook, S., Hoeg, S., 2003. Quantifying uncertainties in tracer-based hydrographseparations: a case study for two-, three- and five component hydrograph separa-tions in a mountainous catchment. Hydrol. Process. 17, 431–453.

Whiteman, C.D., Zhong, S., Shaw, W.J., Hubbe, J.M., Brian, X., Mittelstadt, J., 2001. Coldpools in the Columbia basin. Weather Forecast. 16, 432–447.

Williams, C.J., McNamara, J.P., Whittier, R.B., 2006. Surface-water and ground-water inter-action in the Pahsimeroi Valley, Idaho, USA. Technical Report BSU CGISS 06-02 Boise,Idaho, p. 73.

Winter, T.C., Harvey, J.W., Franke, O.L., Alley,W.M., 1998. Ground water and surface watera single resource. U.S. Geologic Survey Circular 1139, p. 79.

WRCC, 2014. Western Regional Climate Center — Online Database. http://www.wrcc.dri.edu/.

Wu, W., Xu, S., Yang, J., Yin, H., Tao, X., 2009. Sr fluxes and isotopic compositions in theheadwaters of the Yangtze River, Tongtian River and Jinsha River originating fromthe Qinghai–Tibet Plateau. Chem. Geol. 260, 63–72.