An investigation of acidic head-water streams in the Judith Mountains, Montana, USA

Applied Geochemistry xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Applied Geochemistry journal homepage: www.elsevier.com/locate/apgeochem

An investigation of acidic head-water streams in the Judith Mountains, Montana, USA George P. Williams a, Katelyn Petteys b, Christopher H. Gammons b,⇑, Stephen R. Parker a a b

Department of Chemistry and Geochemistry, Montana Tech of The University of Montana, 1300 W. Park St., Butte, MT 59701, United States Department of Geological Engineering, Montana Tech of The University of Montana, 1300 W. Park St., Butte, MT 59701, United States

a r t i c l e

i n f o

Article history: Available online xxxx

a b s t r a c t Acid rock drainage exists in three headwater streams (Armells Creek, Collar Gulch, Chicago Gulch) in the Judith Mountains, central Montana, USA. The streams drain opposing sides of a central, pyrite rich, hydrothermally altered, granite-porphyry intrusion. Although significant production of precious metals has occurred elsewhere in the Judith Mountains, no present day mining is taking place and the drainages of interest have not been heavily impacted by historical mining activities. All three streams are acidic (pH < 4) in their headwaters and become pH-neutral with distance downstream due to the influx of alkaline groundwater and tributary flows. Concentrations of dissolved aluminum, cadmium, copper, lead, thallium, zinc and fluoride ion are locally well above regulatory standards. Detailed synoptic sampling using the continuous tracer injection method shows that metal loading is diffuse and is associated with weathering of the pyrite-rich bedrock, with negligible contributions of acid or metals from legacy mine waste. Each stream is precipitating hydrous ferric and aluminum oxides in discrete zones due to a combination of oxidation and acid-neutralization processes. An abundance of terraced ferricrete deposits shows that acid rock drainage existed prior to human disturbance. Furthermore, a comparison of the trace-metal content (Cu/Fe ratio) of modern, in-stream precipitates vs. ancient (undated) ferricrete deposits suggests that the chemistry of the streams has not changed significantly due to anthropogenic activity. The geochemistry of headwater streams in the Judith Mountains may provide a useful baseline comparison to nearby regions in central Montana with similar geology that have been heavily disturbed by historic and modern mining. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction Acid rock drainage (ARD) refers to the release of protons, dissolved sulfate, and associated metals and metalloids during weathering of pyrite and other sulfide minerals (Nordstrom, 1982a,b; Alpers et al., 1994; Nordstrom and Alpers, 1999). In most cases, ARD has been initiated or amplified by mining activities (Wireman and Stover, 2011). ARD is widely considered to be the most challenging environmental problem that the modern mining industry must face, and is also a major concern for other stakeholders, such as government agencies, land managers, and the general public. ARD has potentially serious effects on terrestrial and aquatic ecosystems which give rise to a host of multifaceted issues ranging from remediation and treatment strategies to human exposure and health concerns. These types of problems have long-term consequences and can be expensive to solve.

⇑ Corresponding author. Tel.: +1 406 496 4763. E-mail address: [email protected] (C.H. Gammons).

Although usually discussed within the context of modern or historical mining activities, ARD may also occur in streams that have had little or no human disturbance. Examples of natural ARD have been reported from Colorado, Montana, and New Mexico in the western USA (McKnight and Bencala, 1990; Furniss et al., 1999; Posey et al., 2000; Bird, 2003; Sjostrom et al., 2004; Verplanck et al., 2009). Some of these streams are in pristine, undisturbed settings (McKnight and Bencala, 1990; Bassett et al., 1992), whereas others have been influenced by mining but show evidence of pre-modern acidic drainage (e.g., Fernández-Remolar et al., 2003). In the latter case, it is often a challenge to determine pre-mining water-quality conditions (Runnells et al., 1992; Bird, 2003; Runkel et al., 2007). Many streams that experienced pre-modern ARD are characterized by the presence of ferricrete (Verplanck et al., 2007). Ferricrete typically consists of alluvial or colluvial gravel and cobbles cemented by a matrix of iron oxide or hydroxide (e.g., goethite, FeOOH). Ferricrete may crop out on the banks of streams as a ledge or terrace of erosionally-resistant material, often at some elevation above the modern streambed (Verplanck et al., 2009). Certain ferricrete deposits dated to

http://dx.doi.org/10.1016/j.apgeochem.2015.05.012 0883-2927/Ó 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Williams, G.P., et al. An investigation of acidic head-water streams in the Judith Mountains, Montana, USA. Appl. Geochem. (2015), http://dx.doi.org/10.1016/j.apgeochem.2015.05.012

2

G.P. Williams et al. / Applied Geochemistry xxx (2015) xxx–xxx

>1000 y in age have been used as a record of climate change during the Holocene Epoch (Furniss et al., 1999; Sjostrom et al., 2004). Furthermore, Nimick et al. (2009) have shown that a comparison of the trace element concentration of fresh Fe-oxide precipitates forming in a stream vs. adjacent ferricrete deposits may be used to compare the water quality of acidic streams before and after human disturbance. This study investigates the extent and origin of acidic drainage in three headwater streams (Armells Creek, Collar Gulch, and Chicago Gulch) in the Judith Mountains of central Montana, USA. Although there has been significant mining of precious metals elsewhere in the Judith Mountains (Weed and Pirsson, 1898; Robertson, 1950; Zhang and Spry, 1994; Woodward, 1995), historic mining has been minimal within the watersheds of this study. To augment routine water quality monitoring, additional field experiments were conducted in two of the streams (Armells Creek, Collar Gulch): (1) detailed synoptic sampling using the continuous tracer injection method (Kimball et al., 2002) to quantify longitudinal changes in metal concentrations and loads; (2) diel (24-h) sampling to examine short-term changes in the concentrations of trace metals (Gammons et al., 2005; Parker et al., 2007; Nimick et al., 2011); and (3) collection of a longitudinal transect of ancient ferricrete samples for trace metal analysis, to compare with modern HFO in-stream precipitates (Nimick et al., 2009).

2. Site description The Judith Mountains are located near the geographic center of the State of Montana, USA, about 15 km to the northeast of the city of Lewistown (Fig. S1, Supplemental Files). The range forms an arc

109o13’0”W

B

approximately 32 km long and 16 km wide and consists of a large number of late Cretaceous to early Tertiary plutons that have intruded into originally flat-lying Paleozoic and Mesozoic sedimentary rock (Wallace, 1953; Goddard, 1988; Porter and Wilde, 1993). The intrusions of the Judith Mountains belong to the Central Montana Alkalic Province (Marvin et al., 1980; Baker and Berg, 1991), which includes a number of other ‘‘island mountain ranges’’, including the Crazy Mountains, the North and South Moccasins, the Bears Paw Mountains, the Big Snowy Mountains, and the Little Rocky Mountains (Fig. S1, Supplemental Files). In all of these ranges, intrusions have caused uplift and doming of the surrounding sediments, contact metamorphism, and localized precious- and base-metal mineralization. Over time, due to natural weathering processes, the plutons have become exposed along the crest of the mountain ranges. The geology of the immediate study area is dominated by porphyry intrusions of alkali granite, quartz monzonite, and syenite (Fig. 1, Goddard, 1988; Kohrt, 1991). Within these igneous rocks, centered on Judith Peak and Red Mountain (the two highest peaks in the central Judith Mountains at 1960 m and 1882 m, respectively), a prominent zone of hydrothermal alteration and pyrite mineralization exists. This alteration may reflect the presence of a buried porphyry Cu–Mo or carbonatite deposit (Hall, 1977; Lindsey and Fisher, 1985). In addition to pyrite, fluorite is widespread within the altered zone (Hall, 1977). All streams draining the zone of hydrothermal alteration on Judith Peak and Red Mountain are acidic in their headwaters, but become pH-neutral at lower elevation as the geology transitions into Paleozoic and Mesozoic sedimentary rock (Fig. 1). Both Armells Creek and Collar Gulch contain large, cliff-forming outcrops of the Mississippian Madison Limestone in their lower reaches. In

109o11’0”W

TKqm

State of Montana

Sample Site Locaons

A

TKqm = Quartz monz. porphyry Tg = Alkali granite porphyry 0

TKqm

0.4

0.8

1.2

1.6

Km’s

47014’0”N

C Red Mountain 1867m

pH < 3.5 3.5 < pH < 5 5 < pH < 7 pH > 7

Tg 1823 m Judith Peak 1959m

Red Mountain 1867m

47o12’0”N

Judith Peak 1959m

Tai Holt Mine

1823m

1867m

pyrite-rich

N 1 km

TKqm

1867m

Collar Gulch

Fig. 1. Judith Mountains field area showing: (A) location within Montana; (B) sampling locations and geologic information; (C) the radial pattern of acid streams draining the central region of pyrite-rich alteration on Judith Peak and Red Mountain (C).

Please cite this article in press as: Williams, G.P., et al. An investigation of acidic head-water streams in the Judith Mountains, Montana, USA. Appl. Geochem. (2015), http://dx.doi.org/10.1016/j.apgeochem.2015.05.012

G.P. Williams et al. / Applied Geochemistry xxx (2015) xxx–xxx

Chicago Gulch, the Madison Group is buried beneath younger Cretaceous shales and sandstones. Armells Creek, Collar Gulch, and Chicago Gulch have similar drainage areas (4.3 km2, 4.4 km2, and 3.4 km2, respectively) and form a radial pattern around the zone of pyrite-rich alteration on Judith Peak and Red Mountain (Fig. 1; see also Lindsey and Fisher, 1985). The streams are small, with typical discharge values of 10–100 L/s in late summer and fall baseflow conditions, although spring snowmelt and/or heavy rain can swell flows to >103 L/s for short periods of time. The steep, upper reaches of each stream gain groundwater, are acidic, and have streambeds coated with hydrous ferric and aluminum precipitates which limit subsurface flow (see Fig. S2 of Supplemental Data for field photographs). In contrast, the pH-neutral, lower reaches of each stream lack cementing hydrous oxides and lose water to the subsurface. In particular, the lower reaches of Collar Gulch typically dry up by mid-summer. Extensive deposits of ferricrete are present in the upper reaches of each stream, forming erosionally-resistant benches up to 3 m in thickness through which the active stream has incised a modern channel (Fig. S2, Supplemental Files). Active deposition of hydrous ferric oxide is evident in off-channel springs that have built up ferricrete mounds with a distinctive terrace structure. The only mine within the study area with recorded production (36.2 kg gold and 20.1 kg silver from 3  106 kg of ore, Robertson, 1950) is the Tail Holt mine, located in the northwest part of the Collar Gulch watershed (Fig. 1). This mine, located several hundred meters in elevation above Collar Gulch, has a collapsed adit with a small discharge (<0.1 L/s) of pH-neutral water which soaks into the ground shortly after surfacing. The ridgeline that connects Judith Peak and Red Mountain has been disturbed by construction of roads, communication towers, and a radar station (now dismantled). Despite these human modifications, the majority of the study area exemplifies a natural setting and gives the impression of quintessential wilderness.

3. Methods 3.1. Field methods Armells Creek, Collar Gulch, and Chicago Gulch were visited on ten different dates for collection of water quality data between September 2011 and August 2013. For greater precision and spatial resolution in streamflow and metal loads, a detailed synoptic sampling in conjunction with continuous tracer injection (Kimball et al., 2002, 2010; Runkel et al., 2013) was carried out in Armells Creek (June, 2012) and Collar Gulch (August, 2013). Location maps for samples collected in these detailed synoptic events are available in Figs. S3 and S4 (Supplemental Data). A stock solution made from 2.5 kg of KBr in approximately 20 L of stream water was injected into the furthest upstream location of each creek at a slow, continuous rate (10 mL/min) for >24 h using a Fluid Metering piston pump with DC power source. KBr was selected as the tracer due to the naturally low (below detection) concentrations of Br in the streams and the conservative behavior of Br in an acidic environment (Runkel et al., 2013). The injection rate was checked several times during the experiment with a volumetric flask and stopwatch. Tracer breakthrough and plateau concentrations were monitored at two downstream locations using automated sample collectors. After the tracer plateau was achieved, a set of water samples was collected at 10–15 locations along the main channel, as well as for each tributary. Discharge at each main stream location was calculated based on the measured dilution of the KBr tracer (Kimball et al., 2002). To check for diel (24-h) variations in trace metal concentration, water samples collected in the automated

3

samplers in Armells Creek were later acidified and measured for total (unfiltered) metal concentrations using ICP-AES, and a second MS5 datasonde was deployed to collect hourly readings in field measurements. Concurrent with the detailed synoptic sampling events in Armells Creek and Collar Gulch, a longitudinal set of in-stream precipitates and ferricrete samples was collected. Following the methods of Nimick et al. (2009), rocks and stream sediment laden with hydrous oxide coatings were placed in a five gallon bucket and subsequently agitated to release the fresh precipitates. The water was then decanted into a plastic bag and later filtered using an aspiration apparatus and Whatman 40 ashless filter paper. The precipitates were dried overnight at 40 °C prior to digestion. Ferricrete samples were collected with a rock hammer from representative outcrops as close as possible to the location where the corresponding in-stream precipitate sample was collected. Several specimens of ferricrete from each outcrop were crushed to a coarse (>1 cm) size and pieces rich in oxide-cemented matrix material (shown by X-ray diffraction to be goethite) were hand-picked. This material was crushed to 100 mesh and dried overnight (40 °C). The in-stream precipitates and ferricrete samples were digested following EPA method 3050 using concentrated aqua regia, with hydrogen peroxide added to oxidize organic matter. The final solutions were diluted with deionized water prior to chemical analysis. During routine sampling and detailed synoptic sampling, similar protocols were followed for collection and preservation of water samples. Field measurements, including temperature (°C), pH, dissolved oxygen (DO; % saturation and mg/L), electrical potential (Eh, mV) and specific conductivity (SC, lS/cm), were collected with a Hydrolab Minisonde 5 (MS5) calibrated on the morning of each sampling event. Zobell’s solution was used to calibrate the redox electrode to the standard hydrogen electrode. Raw water samples with pH > 4.5 were titrated in the field with H2SO4 and bromocresol green–methyl red pH-indicator dyes to determine alkalinity (mg/L as CaCO3). Three types of water samples were collected: unfiltered (raw)-acidified (RA), filtered-acidified (FA), and filtered-unacidified (FU). FA and RA samples were collected in 60 mL high density polyethylene (HDPE) bottles that had been pre-soaked in 5% HNO3 and triple-rinsed with deionized water prior to field work, and were preserved with 0.6 mL of Trace Metal grade HNO3 within 12 h of sampling. Filtered samples (FA, FU) were collected with a 60 mL HDPE syringe and disposable, 0.2 lm polyethersulfone (PES) syringe filters. Syringes and sample bottles were rinsed with sample water prior to filling the bottles. 3.2. Analytical methods Quantification of major and trace metal concentrations of acid-preserved samples was determined by inductively coupled plasma-atomic emission spectroscopy (ICP-AES) at the Environmental Biogeochemistry Laboratory at the University of Montana using EPA method 200.7. The detailed synoptic FA and RA water samples and the ferricrete/precipitate samples were later re-analyzed by inductively coupled plasma mass spectrometry (ICP-MS) at the Montana Bureau of Mines and Geology (method EPA 200.8). All samples were analyzed for a wide range of solutes, including Al, As, Ca, Cd, Co, Cr, Cu, Fe, K, Mg, Mn, Mo, Na, Pb, S, Tl and Zn. Anion concentrations (bromide, chloride, fluoride, sulfate, nitrate and phosphate) in FU samples were measured via ion chromatography (EPA method 300.0) at Montana Tech. 3.3. Geochemical modeling The computer code PhreeqcI (Parkhurst and Appelo, 1999) was used to calculate chemical speciation and saturation indices

Please cite this article in press as: Williams, G.P., et al. An investigation of acidic head-water streams in the Judith Mountains, Montana, USA. Appl. Geochem. (2015), http://dx.doi.org/10.1016/j.apgeochem.2015.05.012

4

G.P. Williams et al. / Applied Geochemistry xxx (2015) xxx–xxx

and Zn (up to 340 lg/L). Concentrations of Pb and Cd are especially high in Collar Gulch despite the fact that upper Collar Gulch has higher pH than the headwaters of the other streams (Table 1). Thallium (Tl), a very rare but highly toxic trace metal, is present at concentrations >1 lg/L in the headwaters of each stream. In terms of major ions, there is a general trend from Ca–K–SO4 dominated to Ca–HCO3–SO4 dominated with distance downstream. High K+ in the headwaters is derived from the K-rich intrusive rocks on Judith Peak and Red Mountain. Because Tl is known to substitute for K in alkalic igneous rocks (Calderoni et al., 1985; Peter and Viraraghavan, 2005), the unique geology may also explain the presence of dissolved Tl in the drainages. Concentrations of F ion are also notably high in the headwaters (up to 4.5 mg/L), probably a result of weathering of fluorite from the hydrothermally altered intrusions. Increases in Ca2+, Mg2+ and HCO 3 with distance downstream (Table 1) are attributed to the influx of pH-neutral groundwater and tributary streams that originate from unaltered intrusions and surrounding sedimentary rock.

(SI = log [ion activity product/solubility constant]) for possible solid phases. Results are presented below for amorphous Fe(OH)3, amorphous Al(OH)3, basaluminite (Al4(SO4)(OH)105 H2O), and fluorite (CaF2). A log Ksp value of 22.7 for basaluminite (Singh and Brydon, 1969; Singh, 1980; Gustafsson, 2010) was added to the PhreeqcI database. The saturation index for this phase was divided by 4 to better compare SI values with other Al solids that contain one Al atom per formula unit. Calculations for redox-sensitive compounds (e.g., aqueous Fe species) were based on the measured temperature, pH, and Eh. As discussed by many authors (e.g., Langmuir, 1997), this approach to speciating redox-sensitive compounds is less reliable than direct measurement of solute concentrations in each valence state. Nonetheless, the calculations provide a useful framework for later discussion. 4. Results Chemical analyses for selected sampling sites in the upper, middle, and lower reaches of each drainage are given in Table 1. The complete datasets for the detailed synoptic sampling events in Armells Creek and Collar Gulch are given in the Supplementary Data.

4.2. Detailed synoptic sampling 4.2.1. Armells Creek During the synoptic sampling/tracer experiment the pH in the upper portion of Armells Creek decreased after the confluence with an acidic tributary, referred to here as the ‘‘red tributary’’, and then increased steadily down gradient to near-neutral values (Fig. 2A). The SC increased continuously over most of the experimental reach but leveled off below the confluence of the East Fork of Armells Creek. Streamflow likewise increased steadily to the East Fork, below which point it leveled off. The redox potential (Eh) changed

4.1. General trends in stream chemistry The three drainages in the study area are acidic in their headwaters but have near-neutral pH in their lower reaches (Table 1). Concentrations of several solutes indicative of acid rock drainage are elevated in the upper stream reaches, including Al (3– 5 mg/L), Fe (1–5 mg/L), SO2 4 (up to 80 mg/L), Cd (up to 3.2 lg/L), Cu (up to 908 lg/L), Mn (up to 557 lg/L), Pb (up to 1310 lg/L),

Table 1 Representative water chemistry for upper, middle, and lower sampling sites in each drainage. All samples were filtered and have concentrations of mg/L (Ca through F-) or lg/L (As through Zn). Ca

Mg

Na

K

Si

Al

Fe

HCO 3

SO2 4

Cl

F

Collar Gulch (August, 2013) Upper 3.8 Middle 7.3 Lower 7.3

5.5 21 26

1.2 2.8 3.5

2.3 3.1 3.3

4.2 1.7 1.6

21 12 8.4

3.81 0.34 0.01

2.30 0.025 0.002

0.0 44 72

42 30 22

0.36 1.02 1.64

4.52 2.74 1.75

lg/L

As

Cd

Ce

Co

Cu

La

Mn

Pb

Sr

Tl

U

Zn

Upper Middle Lower

0.26 <0.2 0.23

3.2 1.2 0.3

13.6 0.5 <0.2

3.7 1.4 <0.5

908 8.9 1.6

10.9 0.6 <0.2

185 87.8 <1

1310 46 1.4

75 512 596

4.9 1.4 0.8

5.1 1.4 1.8

296 143 21.7

Ca

Mg

Na

K

Si

Al

Fe

HCO 3

SO2 4

Cl

F

mg/L

pH

Armells Creek (June, 2012) mg/L pH a

a

Upper Middle Lower

3.2 4.1 7.2

6.3 20 31

2.1 3.4 4.6

2.7 5.8 4.8

5.5 3.3 3.0

23.3 18.4 14.1

4.26 2.40 0.08

4.22 2.79 0.13

0.0 0.0 23

81 73 66

0.45 0.40 0.84

2.6 2.8 2.3

lg/L

As

Cd

Ce

Co

Cu

La

Mn

Pb

Sr

Tl

U

Zn

Uppera Middle Lower

<0.2 1.7 0.5

0.4 <0.2 <0.2

1.1 0.4 <0.2

4.7 2.6 1.7

10.1 9.6 1.1

0.5 0.2 <0.2

557 363 352

<0.2 <0.2 <0.2

153 366 322

2.3 1.0 0.5

1.2 0.3 0.3

340 148 45

Chicago Gulch (April, 2012) mg/L pH

Ca

Mg

Na

K

Si

Al

Fe

HCO 3

SO2 4

Cl

F

Upper Middle Lower

3.0 6.2 6.6

4.0 11 15

0.8 1.9 2.3

1.9 2.1 2.3

3.9 1.5 1.3

15 8.6 7.7

4.22 0.07 0.07

1.86 0.07 0.01

0.0 8 28

66 28 27

0.3 0.6 1.0

0.9 0.6 0.5

lg/L

As

Cd

Ce

Co

Cu

La

Mn

Pb

Sr

Tl

U

Zn

Upper Middle Lower

<0.2 <0.2 <0.2

0.5 0.3 <0.2

1.5 0.2 <0.2

3.9 0.9 0.4

18 1.1 0.8

1.6 0.7 <0.2

341 114 57

1.1 2.1 0.5

58 153 208

1.5 0.3 <.2

1.5 <.2 0.2

159 48 37

Upper site in Armells is the ‘‘Red trib’’ in the upper portion of the drainage (Fig. 1).

Please cite this article in press as: Williams, G.P., et al. An investigation of acidic head-water streams in the Judith Mountains, Montana, USA. Appl. Geochem. (2015), http://dx.doi.org/10.1016/j.apgeochem.2015.05.012

5

G.P. Williams et al. / Applied Geochemistry xxx (2015) xxx–xxx

mostly in an inverse fashion to pH (Fig. 2B). The concentrations of Fe, Al, Mn, Zn, and Tl showed a stepwise increase at the confluence of the red tributary, but then continued to increase due to diffuse Distance downstream (m) 1800

A

2400 240

SC

pH

160

pH

5

120

RT pH=3.22

B

80

Flow

45

ORP

Temp

10

300

ORP (mV)

400

o

Temp ( C)

12

EF pH=6.96

30

8

Flow (L/s)

6

SC (µS/cm)

200

15

200 6

0 0

600

1200

1800

2400

Distance downstream (m) Fig. 2. Downstream profiles from injection site (distance 0) during the Armells Creek synoptic sampling/tracer injection study: pH and specific conductivity (A), and temperature, oxidation–reduction potential (Eh) and flow (B). Location of the ‘‘red trib’’ (RT) and east fork (EF) are show for reference including pH measurement’s before the confluence with Armells Creek.

2

C. Mn

0.8

0.6

1

0.4

1 0.2

0

120

D. Zn 24

E. Cu East fork

Red trib

160

0.0 Red trib

East fork

0

F. Tl 0.9

16

0.6

80 8

Aq-C

40

East fork

Concentration, µg/L

East fork

2

B. Al Red trib

East fork

Red trib

Concentration,

g/L

3

A. Fe

East fork

1200

Red trib

7

600

Red trib

0

groundwater inputs to a point roughly 800 m below the tracer injection point (Fig. 3). With the exception of Mn, concentrations of most trace elements decreased below the East Fork confluence. A comparison of filtered and non-filtered concentrations shows a greater tendency for metals to partition into suspended solids with distance downstream, especially below the East Fork confluence. Visually, the streambed of Armells Creek was dominated by orange-red hydrous ferric oxide (HFO) above the East Fork, with a mixture of HFO and milky-white hydrous Al oxide (HAO) below the East Fork. In addition, black deposits of Mn-oxide-cemented gravel were locally observed where recent floods had eroded a new channel up to 2 m below the base level of the stream (Fig. S2, Supplemental Files). The Montana chronic aquatic life standards are shown (red lines in Fig. 3) for Zn and Cu (MDEQ, 2012) and are dependent on water hardness. Cu and Zn concentrations are above these standards between the red tributary and the East Fork (Fig. 3D and E). Zn (total and dissolved) dropped below the standard before the stream left the study area while total Cu remained near the standard. The deviation between the dissolved and total lines below the East Fork confluenceindicates partitioning of trace metals into freshly formed HFO and HAO precipitates in the approximate order (Fe, Al, Cu) > Zn > (Mn, Tl). Montana DEQ does not list an aquatic life standard for Tl, but the current human health standard for Tl in surface water in Montana is 0.24 lg/L (MDEQ, 2012). The low standard for thallium reflects the high toxicity of this element (Peter and Viraraghavan, 2005; EPA, 2009). Concentrations of Tl were well above this standard for all of Armells Creek (Fig. 3F). Thallium exists entirely in the dissolved phase until the confluence of the East Fork, below which point it begins to partition into suspended solids (Fig. 3F). With the exception of Cu, the loads of all trace metals increased substantially after the confluence of the red tributary (Fig. 4). Increases in loads at distances of 500–750 m below the tracer injection site are attributed to diffuse groundwater seepage. The

HH

0.3

Aq-C

0

0 0

600

1200

1800

0.0 0

600

1200

1800

0

600

1200

1800

Distance below tracer injection (m) Fig. 3. Concentration profiles for Fe, Al, Mn, Zn, Cu and Tl from the Armells Creek synoptic sampling/tracer injection study as a function of distance downstream from injection site (distance 0). Filled circles (green) show the dissolved concentration and the open squares show the total acid recoverable concentration. The Montana chronic aquatic life standard (Aq-C) for Zn and Cu are calculated as a function of water hardness (D and E) and the Montana human health standard for surface water (HH) for Tl is 0.24 lg/L (F). Also, shown are the two main tributaries; ‘‘red trib (RT)’’ and east fork (EF). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Please cite this article in press as: Williams, G.P., et al. An investigation of acidic head-water streams in the Judith Mountains, Montana, USA. Appl. Geochem. (2015), http://dx.doi.org/10.1016/j.apgeochem.2015.05.012

G.P. Williams et al. / Applied Geochemistry xxx (2015) xxx–xxx

300

200

120

East Fork

East Fork

300

C. Mn

B. Al

Red trib

Red trib

load, g/hour

400

A. Fe

East Fork

500

Red trib

6

80

200 100

40

0

0

100

D. Zn East Fork

1.5

0.10 1.0

East Fork

10

Red trib

15

0.15

5

0.05

0.5

0

0.00

0.0 0

F. Tl

East Fork

E. Cu 2.0

Red trib

load, g/hour

20

Red trib

0

500

1000

1500

2000

0

500

1000

1500

0

2000

500

1000

1500

2000

Distance below tracer injection, meters Fig. 4. Profiles of changes in load (conc.  flow) for Fe, Al, Mn, Zn, Cu and Tl with distance downstream during the Armells Creek synoptic sampling/tracer injection study. Solid green circles show the dissolved load and the open squares show the total acid recoverable load. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

4.2.2. Collar Gulch Fig. 5 summarizes synoptic changes in pH, SC, temperature, and streamflow in Collar Gulch. Three tributaries in the study reach are

Distance downstream (m) 8

A

800 T3

T2 T1

1600 pH

2400 200

SC

160

pH

6 120

SC (µS/cm)

0

80

4

40 20

B Flow

15

o

13 10

Flow (L/s)

14

Temp ( C)

East Fork increased the total loads of all trace metals. Below the East Fork confluence, total loads of all trace metals remained approximately constant whereas dissolved loads either decreased (Fe, Al, Zn, Cu) or remained constant (Mn, Tl). The persistence of total metal loads where the majority of the metal resides as suspended particles indicates minimal settling of HFO and HAO in the fast-moving water of Armells Creek. During the detailed synoptic study, samples were collected hourly near the lower end of the study reach to test for diel fluctuations in trace metal concentrations. Concentrations of total Zn decreased from an early morning maximum of 63 lg/L to a late afternoon low of 44 lg/L (data in Supplemental File). This type of diel fluctuation in Zn concentration is typical of small, pH-neutral streams draining abandoned mine lands, and is thought to be caused by 24-h changes in pH and water temperature coupled with adsorption to metal oxide and biotic surfaces (Nimick et al., 2011). The 24-h ranges in pH and temperature in lower Armells Creek were 0.15 pH units and 3.7 °C, respectively (Supplemental File), both relatively small ranges compared to streams that typically display much larger diel variations in trace metal concentration. Among the other metals of interest, a weak diel pattern with large scatter was noted for total Al and total Fe, with higher concentrations at night and lower concentrations during the day. Nighttime increases in the concentrations of particulate metals have been noted in other streams draining abandoned mine lands (Nimick et al., 2011). Overall, the diel cycles noted in Armells Creek are relatively small in magnitude, and should not cause problems for the interpretation of synoptic data described above. Although a similar attempt to monitor diel changes in Collar Gulch failed due to equipment malfunction, it is likewise expected that 24-h changes in trace metal concentration would be minimal given the similar geology, hydrology, and chemistry of the two streams.

Temp

12

5

11 0

800

1600

0 2400

Distance downstream (m) Fig. 5. Downstream profiles from injection site (distance 0) during the Collar Gulch synoptic sampling/tracer injection study; pH and specific conductivity (A), and temperature and flow (B). Also, shown are the three main tributaries; T3, T2 and T1 as vertical dashed lines, left to right, respectively.

labeled T1, T2, and T3, proceeding in an upstream direction. The watershed of T3 includes the historic Tail Holt mine and adjacent hydrothermally altered bedrock, whereas T1 and T2 originate

Please cite this article in press as: Williams, G.P., et al. An investigation of acidic head-water streams in the Judith Mountains, Montana, USA. Appl. Geochem. (2015), http://dx.doi.org/10.1016/j.apgeochem.2015.05.012

7

G.P. Williams et al. / Applied Geochemistry xxx (2015) xxx–xxx

Concentration, mg/L

A. Fe

B. Al

1

C. Mn

0.3

1 0.2

0.1 SO4/100

0.1 0.1

0.01 0.01

Concentration, mg/L

0.001

0.0

D. Pb

1

E. Cu

0.1

F. Zn

0.4 0.3

0.1 0.01

0.01

0.2

Aq-C

Aq-C

Aq-C

0.1

0.001 0.001

0.0

Concentration, µg/L

4

G. Cd

H. Tl

5

I. Ce

15

3 4 10 2

3

Aq-C

2

5

1 1 HH

0

0 0

1000

2000

0

3000

0

1000

2000

3000

0

1000

2000

3000

Distance below tracer injection, meters Fig. 6. Concentration profiles for Fe, Al, Mn, Pb, Cu, Zn, Cd, Tl and Ce from the Collar Gulch synoptic sampling/tracer injection study as a function of distance downstream from injection site (distance 0). Open squares show the total acid recoverable concentration and green filled circles show the dissolved concentration. Triangle symbols in panel A show concentrations of dissolved sulfate (divided by 100) for reference. The Montana chronic aquatic life standard (Aq-C) for Pb, Cu, Zn and Cd are calculated as a function of water hardness (D-G) and the human health standard (HH) for Tl is 0.24 lg/L (H). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

40

100

A. Fe

B. Al 80

30

Dissolved Total

60

20

4 40

SO4/100

10

Load, g/hour

C. Mn 6

2

20

0

0

30

D. Pb

2.5

Dissolved Total

20

0

E. Cu

F. Zn 8

2.0 6 1.5 4

1.0

10

2

0.5 0.0

0 0

1000

2000

3000

0 0

1000

2000

3000

0

1000

2000

3000

Distance below tracer injection, meters Fig. 7. Profiles of changes in load (conc.  flow) for Fe, Al, Mn, Pb, Cu, and Zn with distance downstream during the Collar Gulch synoptic sampling/tracer injection study. Open squares show the total acid recoverable load and the filled circles (green) show the dissolved load. Triangle symbols in panel A show loads of dissolved sulfate (divided by 100) for reference. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Please cite this article in press as: Williams, G.P., et al. An investigation of acidic head-water streams in the Judith Mountains, Montana, USA. Appl. Geochem. (2015), http://dx.doi.org/10.1016/j.apgeochem.2015.05.012

8

G.P. Williams et al. / Applied Geochemistry xxx (2015) xxx–xxx

outside the zone of hydrothermal alteration and pyrite mineralization. Consequently, T3 had an acidic pH (4.75) on the date of the detailed synoptic investigation, whereas T1 and T2 contained pH-neutral water with higher alkalinity than the main stem. Most of the increase in flow in Collar Gulch occurs from diffuse groundwater inflow, with relatively small contributions from the tributary streams (Fig. 5B). Concentrations of trace metals, including Fe, Al, Mn, Pb, Cu, Zn, Cd, Tl and Ce, all decreased in Collar Gulch with distance downstream from the injection site (Fig. 6A–H). Fe, Al, Pb, and Cu show some partitioning into the particulate phase after the confluence with T3 while the other metals were mostly in the dissolved phase. Concentrations of Pb, Cu, Zn, and Cd were above the chronic aquatic life standards (MDEQ, 2012) in the upper reaches of the study area but fell below the standard at the downstream end. Concentrations of Pb were particularly high in upper Collar Gulch, being >1 mg/L. As was the case for Armells Creek, concentrations of Tl in Collar Gulch greatly exceeded the human health standard of 0.24 lg/L (MDEQ, 2012).

Loads of trace metals in Collar Gulch increased steadily to a point roughly 500 m downstream of the tracer injection point due to the influx of acidic groundwater and then decreased sharply over the remainder of the study reach (Fig. 7). Although T3 was mildly acidic, the main effect of the confluence of T3 on the main stem was an increase in pH (Fig. 5A), and a decrease in metal loads (Fig. 7). 4.3. Ferricretes and in-stream precipitates The ferricrete samples and in-stream precipitates (ISPs) from Armells Creek and Collar Gulch were digested in acid to determine the major and trace element concentrations in the solids. Selected results are summarized in Tables 2 and 3, with the complete data set in the Supplementary Material. Concentrations of most trace elements were similar in ISPs compared to ferricrete samples collected near the same location in the creek. This relationship is further quantified below. Based on XRD analysis, the dominant cementing material in the pre-modern ferricrete deposits is

Table 2 Chemistry of ferricrete and ISP samples from Armells Creek, collected July 2012. km below tracer inj.

a b

Stream pH

Al (g/kg)

Fe (g/kg)

0.51 0.65 0.77 0.93 1.13 1.26 1.34 1.53

pH (est.)a 4.44 5.17 4.88 4.67 3.97 3.75 5.24 5.93

Ferricrete samples 5.8 240 5.3 129 5.1 198 5.3 415 2.9 464 31.3 399 5.3 119 8.9 199

0.51 0.65 0.77 0.93 1.13 1.26 1.34 1.53 1.85 2.12 2.38

pH (meas.)b 6.09 6.14 4.25 3.82 3.84 5.10 5.35 5.71 6.21 6.40 6.48

In-stream precipitates (ISPs) 104 132 77.9 144 8.6 320 10.0 338 6.0 376 19.4 341 22.9 333 80.2 227 60.1 144 49.3 127 50.5 122

As (mg/kg)

Cu (mg/kg)

Mn (mg/kg)

P (mg/kg)

Pb (mg/kg)

Zn (mg/kg)

282 <50 1040 85 <50 <50 <50 <50

86 149 144 240 26 <20 153 699

29 47 36 39 31 3240 3860 24,800

7960 849 5430 1070 265 <200 631 487

<50 <50 <50 <50 <50 <50 <50 <50

56 79 48 572 699 2820 1050 1280

262 166 95 <50 302 162 150 94 <50 <50 60

1970 1150 158 96 83 208 275 615 332 291 292

441 632 46 105 86 273 170 304 4000 5490 8460

3560 2710 1070 728 1220 651 812 460 565 709 720

<50 <50 <50 <50 <50 <50 <50 <50 <50 <50 <50

545 673 74 106 58 193 206 534 1920 2140 2820

Pb (mg/kg)

Zn (mg/kg)

Estimated pH of stream at time that ferricrete formed. Measured pH of modern stream.

Table 3 Chemistry of ferricrete and ISP samples from Collar Gulch, collected August 2013. km below tracer inj. 0.00 0.14 0.24 0.52 0.85 1.03 0.00 0.14 0.32 0.47 0.84 0.94 1.03 a b

Stream pH

Al (g/kg)

Fe (g/kg)

pH (est.)a 3.90 3.04 3.91 4.08 5.89 7.36 pH (meas.)b 3.80 3.75 4.08 5.58 5.65 6.52 7.33

Ferricrete samples 9.67 241 7.97 336 9.90 177 12.6 165 76.8 27.8 57.1 23.6 In-stream precipitates (ISPs) 19.5 174 25.9 90.0 18.7 115 55.7 93.9 123 44.5 196 38.5 125 36.5

As (mg/kg)

Cu (mg/kg)

Mn (mg/kg)

P (mg/kg)

35 22 666 48 35 52

290 66 219 278 688 2940

8 <5 1330 776 20,600 18,900

232 118 1170 185 428 268

2120 5490 7590 188 141 386

68 66 288 1760 1660 897

46 131 91 79 46 51 44

207 108 154 1000 1970 3580 2430

50 114 82 201 232 292 520

643 391 381 561 395 390 439

2730 1540 4450 18,100 22,800 41,700 27,900

63 116 106 196 501 2160 3750

Estimated pH of stream at time that ferricrete formed. Measured pH of modern stream.

Please cite this article in press as: Williams, G.P., et al. An investigation of acidic head-water streams in the Judith Mountains, Montana, USA. Appl. Geochem. (2015), http://dx.doi.org/10.1016/j.apgeochem.2015.05.012

9

G.P. Williams et al. / Applied Geochemistry xxx (2015) xxx–xxx

goethite. Similar attempts to determine the mineralogy of ISPs were inconclusive, due to the very fine-grained and amorphous character of the modern precipitates. 5. Discussion 5.1. Major controls on trace metal concentrations

3þ

Al

Fig. 8 summarizes changes in chemical processes with distance downstream that are believed to be common to all three streams of interest. Groundwater in the upper reaches of each drainage that has been in contact with the mineralized porphyry intrusions is weakly acidic (pH 4–4.5) and rich in Al, ferrous iron (Fe2+), and other trace metals (e.g., Cd, Cu, Mn, Pb, Zn, Tl). When this groundwater emerges as seeps and springs, Fe2+ is oxidized and hydrous ferric oxide (HFO) precipitates, eventually forming ferricrete deposits. The key reactions can be summarized as follows:

Fe2þ þ 1=4O2 ðgÞ þ Hþ ! Fe3þ þ 1=2H2 O

ð1Þ

Fe3þ þ 3H2 O $ FeðOHÞ3 ðsÞ þ 3Hþ

ð2Þ

with the overall reaction being:

Fe2þ þ 1=4O2 ðgÞ þ 5=2H2 O ! FeðOHÞ3 ðsÞ þ 2Hþ

ð3Þ

The net effect of reaction (3) is the formation of HFO precipitates and a decrease in pH of the stream to values <4. This is consistent with an initial decrease in pH in the upper reaches of Armells and Collar Gulch (Figs. 2A and 5A) coincident with an increase in Fe load, formation of HFO along the streambed, and a decrease in the saturation index (SI) of ferrihydrite (amorphous Fe(OH)3). (saturation index calculations are summarized in Figs. S5 and S6 of the Supplemental Files). Further downstream, increases in pH due to the influx of alkaline groundwater and tributaries causes precipitation of any remaining dissolved Fe as HFO, via reactions (2) and (3). At the low pH values where the headwater ferricretes form, most of the cationic trace metals (Cd2+, Cu2+, Mn2+, Pb2+, Tl+, Zn2+) do not adsorb significantly onto HFO (Gadde and Laitinen, 1974; Dzombak and Morel, 1990; Smith, 1999; Trividi and Axe,  2001), whereas anionic solutes, such as SO2 4 , arsenate (H2AsO4 ) and phosphate (H2PO 4 ), should adsorb strongly (Dzombak and Morel, 1990; Wilkie and Hering, 1996; Smith, 1999). This explains the relatively high concentrations of As and P measured in headwater ferricrete and ISP samples (Tables 2 and 3) whereas dissolved concentrations at the same locations were near or below detection (Table 1).

• • •

Upper Reaches 2+ Oxidation of Fe to HFO pH drops to 3.5 • Sorption of As, P, SO4 • •

pH 4.5 groundwater

As the streams go down the valley, pH rises due to influx of alkaline tributaries and upwelling groundwater that has not come in contact with the hydrothermally altered bedrock. As pH rises above 4, the dissolved Fe load drops off quickly as HFO precipitates. With a further rise in pH to 4.5–5, dissolved Al precipitates as HAO (Alpers et al., 1994):

þ 3H2 O $ AlðOHÞ3 ðsÞ þ 3Hþ

or as a hydrous Al-sulfate, such as basaluminite: 3þ

4Al

þ þ SO2 4 þ 15H2 O $ Al4 ðSO4 ÞðOHÞ10  5H2 OðsÞ þ 10H

ð5Þ

Based on geochemical modeling (Figs. S5 and S6, Supplemental Files), the stream waters in these middle reaches are closer to equilibrium with basaluminite as opposed to amorphous Al(OH)3. Basaluminite has recently been shown to be an important control on the mobility of Al in acidic, sulfate-rich waters, including mining pit lakes (Sánchez-España et al., 2011) and acid sulfate soils (Jones et al., 2011). However, more detailed solids characterization is needed to verify the mineralogy of the white precipitates in this study. In Collar Gulch, the rise in pH is gradual with distance downstream and there is a spatial separation of Fe and Al in the streambed, with red-coated rocks in the headwaters and white-coated rocks in the middle reaches. In Armells Creek and Chicago Gulch, abrupt rises in pH occur due to the confluence of alkaline tributaries, and both HAO and HFO precipitate in the same reach. As pH of the streams rises above 5, the cationic trace metals begin to adsorb onto HFO and HAO in the general order: Pb > Cu > Zn > Mn P Tl. This is evident by the partitioning of Pb and Cu into suspended particles in the mid-reaches of each stream, whereas Zn, Mn, and Tl remain primarily in the dissolved phase (Figs. 3 and 6). Although oxidation of Mn2+ to hydrous manganese oxides (HMO) is thermodynamically favorable in the oxygenated surface water, this is a very slow reaction at pH 4–7 (Hem, 1981; Balzer, 1982; Davison, 1993). In the Armells Creek synoptic study, stream velocities were high, and it took less than 2 h for the KBr tracer to travel the entire distance from the injection point to the lowest sampling point. This amount of time was probably insufficient for much HMO to form, which explains the persistence of elevated Mn concentrations and loads to the end of the study reach (Fig. 3). Nonetheless, at locations where the active stream channel had scoured down from recent floods, patches of black, HMO-cemented alluvium (i.e., manganocrete) were observed (Fig. S2, Supplemental Files). These deposits most likely formed

Middle Reaches Influx of alkaline water raises pH of stream > 5 Deposition of HFO and HAO Sorption of Cu, Pb, (Zn, Cd)

ferricrete

2+

ð4Þ

• • •

Lower Reaches Stream loses water to subsurface Precipitation of HMO Sorption of Zn, Cd, Tl

Fe -rich, anoxic

Porphyry intrusion hydrothermally altered

pH 7.5 groundwater HCO3-rich, anoxic

unaltered ephemeral stream

Al-rich ferricrete manganocrete

Shale

Limestone perennial stream

Fig. 8. Conceptual diagram summarizing hydrogeochemical processes in a typical headwater stream in the central Judith Mountains.

Please cite this article in press as: Williams, G.P., et al. An investigation of acidic head-water streams in the Judith Mountains, Montana, USA. Appl. Geochem. (2015), http://dx.doi.org/10.1016/j.apgeochem.2015.05.012

10

G.P. Williams et al. / Applied Geochemistry xxx (2015) xxx–xxx

in the subsurface of the stream by slow-moving hyporheic water that had sufficient dissolved oxygen and time to precipitate HMO. Any remaining trace metals in solution could have adsorbed or co-precipitated with this HMO. ICP-MS analysis of a single digested manganocrete sample from lower Armells Creek showed elevated concentrations of Zn (3300 mg/kg), Cu (990 mg/kg), Cd (20 mg/kg), and Tl (20 mg/kg). Thallium(I) has been shown to strongly adsorb onto HMO (Jacobson et al., 2005), although the mechanism may involve a redox process (Gadde and Laitinen, 1974). In contrast, ICP-MS analysis of two digested ferricrete samples from Armells Creek had Tl and Cd concentrations near or below the detection limit of 1 mg/kg. In the Collar Gulch synoptic study, the lower gradient of the stream combined with the lower streamflow conditions meant that velocities were much slower, with the tracer taking >9 h to reach the lowest elevation of the study reach. Because of the longer residence time, most of the dissolved Mn2+ was removed in the middle and lower reaches of Collar Gulch (Fig. 6). Unlike Al and Fe, there is no separation of filtered and total Mn concentrations in Collar Gulch (Fig. 6), suggesting that Mn did not adsorb onto suspended particles. Instead, it is hypothesized that Mn was removed by precipitation of Mn-oxides onto the streambed and/or on the undersides of cobbles and boulders in the streambed. Concentrations of Cd, Tl and Zn in both Armells Creek and Collar Gulch closely follow the synoptic trends in Mn much more than those for Fe and Al. This is indirect evidence that Cd, Tl and Zn mainly co-precipitated with HMO, rather than adsorbing onto HAO or HFO.

are broadly similar to the concentrations of F measured by Bove et al. (2009) in headwater streams draining hydrothermally altered volcanic rock in the upper Animas River watershed, Colorado. Bove et al. (2009) suggested that stream water with F/Cl molar ratio >10 could be used as an indicator of proximity to mineralization associated with F-rich, high-silica volcanic rocks. In the two detailed synoptic datasets of this study, F/Cl molar ratios fell in the range of 5–12 in upper Armells Creek, and 19–26 in upper Collar Gulch (Supplemental Data). The close association of high F with high dissolved metal content (e.g., Cu, Pb, Zn) suggests the presence of mineralized veins containing fluorite and metal sulfide minerals on the flanks of Red Mountain and Judith Peak. Indeed, some mineralization of this type was encountered in an exploration drill hole located on the saddle between the two peaks (Hall, 1977). Fluoride forms strong aqueous complexes with dissolved Al, and PhreeqcI predicts that >90% of the dissolved Al in upper Armells Creek and Collar Gulch is present as species of the type AlFn3n (n = 1–3). Based on trends in load with distance downstream (Figs. S5 and S6, Supplemental Files), fluoride appears to behave more or less conservatively. This agrees with the computed saturation indices for fluorite which were consistently negative (Figs. S5 and S6, Supplemental Files). In the Armells drainage, fluoride loads approximately doubled at the confluence of the East and West Forks, and remained steady to the end of the study area. In Collar Gulch, fluoride loads decreased about 20% below the confluence of the ‘‘clean’’ T1 and T2 tributaries. Given the precipitous drop in Al load through this same reach, the decrease in F load may have been due to incorporation of F into the OH sites of rapidly depositing basaluminite or amorphous Al(OH)3.

5.2. Metal sorption onto HFO 5.4. Comparison of modern and pre-modern stream chemistry

5.3. Fluoride geochemistry

Data collected from the synoptic sampling investigations were used to construct best-fit power functions describing the relationship between stream pH and the Fe/Cu ratio of in-stream precipitates in Armells Creek and Collar Gulch (Fig. 9). In both cases, an increase in pH resulted in a decrease in Fe/Cu ratio. The equations in Fig. 9 were then used to estimate what the ‘‘pre-modern’’ pH values of Armells Creek and Collar Gulch were as a function of distance downstream, based on the Fe/Cu ratio of the ferricrete deposits. The profiles of the estimated pre-modern pH values and the measured pH values are broadly similar, especially for Collar Gulch (Fig. 10B). Differences in the estimated and measured pH values for upper Armells Gulch (Fig. 10A) could reflect spatial

8 Armells -0.125 y=12.078*x R2=0.72

7

Collar (acid digestion) y=8.0054*x-0.301 R2=0.91 Collar (XRF) y=8.697·x-0.114 2 R =0.92

6

pH

The adsorption of Cu and Zn onto HFO surfaces during the Armells tracer test was simulated with a surface-complexation model using PhreeqcI (Parkhurst and Appelo, 1999). Each chemical analysis for samples collected during the tracer experiment was equilibrated with a hypothetical HFO surface adapted from Dzombak and Morel (1990) with a surface area of 300 m2/g, a site density of 1  104 mol site/g substrate, and an HFO/water ratio of 0.001 g/L. The latter value falls within the range of suspended (particulate) Fe measured in Armells Creek during the tracer study (0.13–2.9 mg/L Fe, average of 1.2 mg/L Fe). Previous experimental studies have demonstrated that adsorption of cations and anions to metal oxide surfaces is strongly temperature dependent (e.g., Machesky, 1990; Rodda et al., 1996; Trivedi and Axe, 2000, 2001). To refine the simulation of metal adsorption, we added enthalpy values for adsorption onto HFO to the PhreeqcI database of 38 kJ/mol for H+ (Machesky, 1990), +85 kJ/mol for Cu2+ and +66 kJ/mol for Zn2+ (Parker et al., 2007). Each solution from the synoptic sampling set was equilibrated with the model HFO surface and the moles of Cu and Zn sorbed to the surface were computed by the model. The mass of particulate Fe in each solution was determined by the difference between Fe measured in unfiltered minus filtered samples and was used with the modeled mass of sorbed Cu and sorbed Zn to calculate the ratio of Fe/Cu and Fe/Zn. As shown in Figs. S7 and S8 (Supplemental Files), the model-generated ratios follow the same general trend of measured Fe/Cu and Fe/Zn ratios observed for the ISPs and ferricrete samples collected along Armells Creek.

5

4

3 0

As discussed above, the altered bedrock contains widespread fluorite which is most likely the source of measureable F concentrations in the Armells, Chicago and Collar watersheds (Table 1). Fluoride concentrations in the headwaters of these streams locally exceed the human health standard of 4 mg/L (MDEQ, 2012), and

1500

3000

4500

6000

Fe(mg)/Cu(mg) Fig. 9. Cross plot showing the relationship of stream pH and Fe/Cu ratios from ISPs collected in Armells Creek and Collar Gulch. Power function equations for Armells Creek are derived from acid digestions of ISPs. Equations for Collar Gulch are derived from both acid digestions and XRF analysis.

Please cite this article in press as: Williams, G.P., et al. An investigation of acidic head-water streams in the Judith Mountains, Montana, USA. Appl. Geochem. (2015), http://dx.doi.org/10.1016/j.apgeochem.2015.05.012

G.P. Williams et al. / Applied Geochemistry xxx (2015) xxx–xxx

variability in ferricrete composition in the extreme headwaters of the watershed. Overall, the similar pH profiles are consistent with the concept that the stream water has similar historical and present day chemical signatures as it leaves the study area. Overall, the results of this study help to validate the approach used by Nimick et al. (2009) to use the trace element composition of in-stream precipitates and ferricretes to compare the water quality of modern and pre-modern streams impacted by acid rock drainage. As discussed by Nimick et al. (2009), this method assumes that the trace element concentrations of the solids do not change during transformation of the amorphous HFO precipitates to goethite. Based on the weight of evidence reviewed by Nimick et al. (2009), as well as the results of the current study, this appears to be a safe assumption for Cu and possibly Zn. However, examination of the data in Tables 2 and 3 suggests that other trace elements may behave non-conservatively during recrystallization. In particular, the concentration of Pb in ISPs of lower Collar Gulch exceeds 4 wt%, whereas the Pb content of ferricrete at the same location in the stream is <400 mg/L. Conversely, concentrations of Mn in the ferricretes in lower Collar Gulch were much higher than in the ISP samples (Table 3). More detailed mineralogical work is needed to determine the precise residency of Pb, Mn, and other trace elements in the ISP and ferricrete samples.

5.5. Application to other field areas The hydrothermally-altered porphyry rocks at the headwaters of Armells Creek resemble similar rocks exposed 90 km to the north in the Little Rocky Mountains. Between 1979 and 1998, the Zortman-Landusky open pit mines extracted gold from

Downstream distance (km) 0.6 7

A

0.9

1.2

1.5

RT

1.8 EF

Armells stream pH Calc pH

pH

6

5

4

3

B

pH

6.0

4.5

3.0

Collar stream pH Calc pH (acid digest) Calc pH (XRF)

1.5 0.0

0.3

0.6

0.9

Distance downstream (km) Fig. 10. Downstream profiles of estimated pre-mining and measured pH for Armells Creek (A) and Collar Gulch (B).

11

hydrothermally altered syenite porphyry near the crest of the Little Rockies (Wilson and Kyser, 1988). After mine closure, serious acid-rock drainage problems were recognized in a small stream named Swift Gulch, which drains the Landusky area to the northwest (Kill Eagle et al., 2009). Because Swift Gulch contains ancient ferricrete deposits, it was one of field sites used in the original study of Nimick et al. (2009) to compare modern vs. pre-modern water-quality conditions. Based on their analysis, Nimick and others concluded that the water quality of Swift Gulch is substantially worse today compared to pre-modern times, although the stream has probably been weakly acidic (e.g., pH 4.9–6.1) for thousands of years (one outcrop of ferricrete near the bottom of Swift Gulch was dated by 14C at 10.36 ± 0.06 ka BP, Gabelman et al., 2005). The Judith Mountains also share geologic similarities to the closed Kendall gold mine in the North Moccasin Mountains, located just 20 km west of the study area. At Kendall, gold occurs in brecciated limestone and dolomite cut by numerous porphyry intrusions (Lindsey, 1985; Lindsey and Fisher, 1985). Due to the abundance of carbonate rock, acid mine drainage is not a problem at Kendall. However, groundwater that has undergone prolonged contact with waste rock from the mine has very high concentrations of dissolved Tl, locally exceeding 500 lg/L (Mueller, 2001). Thallium removal technologies are currently being evaluated. Based on the results of our own work, it is possible that background concentrations of Tl in the local groundwater were elevated prior to open pit mining at Kendall, although unlikely to the levels that are observed today.

6. Conclusions The geology of the headwaters of Armells Creek, Collar Gulch and Chicago Gulch is dominated by hydrothermally-altered and pyrite-rich intrusive rocks, whereas the lower stretches of the drainages flow through outcrops of Paleozoic limestone and younger sediments (Fig. 1A). This geological transition explains the overall evolution of each stream from acidic and metal-rich in its headwaters, to pH-neutral and relatively metal-poor near its mouth. No large mines occur in the study area, although some surface disturbances exist, including road construction and installation of radio and radar towers on the summit of Judith Peak. However, detailed synoptic sampling shows that the metal and acidity loads in Armells Creek and Collar Gulch are coming from hillslopes that have been minimally disturbed. Concentrations of Cu and Zn are well above chronic standards for aquatic life in upper Armells Creek, and concentrations of Tl, although very low (around 1 lg/L), are still well above human health standards for surface water. In Collar Gulch, the concentrations of Fe, Al, Mn, Pb, Cu, Zn, Cd, Tl and Ce all decreased with distance downstream from the headwaters. Pb, Cu, Zn, and Cd have concentrations above the chronic aquatic life standards in the upper reaches of the study area but fall below the standard at the downstream end most likely due dilution from surface and groundwater inputs. As in Armells Creek, concentrations of Tl in Collar Gulch exceed the human health standard of 0.24 lg/L (MDEQ, 2012). Our study has provided a good test of the method proposed by Nimick et al. (2009) to compare pre-modern vs. modern water-quality conditions in streams impacted by acid rock drainage. Analyses of modern in-stream precipitates vs. pre-modern ferricrete deposits suggest that the hydrogeochemical conditions present today in upper Armells Creek and Collar Gulch are similar to what was present during ferricrete formation. The abundance of ferricrete in the headwaters of Armells Creek, Collar Gulch and Chicago Gulch, all of which drain different sides of Red Mountain, suggests that the acidic drainage is natural. The results

Please cite this article in press as: Williams, G.P., et al. An investigation of acidic head-water streams in the Judith Mountains, Montana, USA. Appl. Geochem. (2015), http://dx.doi.org/10.1016/j.apgeochem.2015.05.012

12

G.P. Williams et al. / Applied Geochemistry xxx (2015) xxx–xxx

of the current study may be applicable to the estimation of pre-mining water-quality conditions at mine sites throughout the central Montana alkalic province where no pre-mining baseline data exist. Acknowledgements This study was funded by the U.S. Bureau of Land Management and we owe special thanks to Chad Krause (BLM) and the land owners for access to the study sites. H. Reid and the 2013 Montana Tech Field Hydrogeology class helped with field work. Heiko Langner and Matt Young (Univ. of Montana, Environmental Geochemistry Lab), and Jackie Timmer (MBMG) assisted with analytical work. The manuscript was improved by detailed comments of Rob Runkel and an anonymous reviewer.

Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.apgeochem.2015. 05.012. References Alpers, C.N., Blowes, D.W., Nordstrom, D.K., Jambor, J.L., 1994. Secondary minerals and acid mine-water chemistry. In: Blowes, D.W., Jambor, J.L., Alpers, C.N. (Eds.), The Environmental Geochemistry of Sulfide Mine Wastes, vol. 22. Miner. Soc. Canada, Short Course, pp. 247–270. Baker, D.W., Berg, R.B., 1991. Guidebook of the Central Montana Alkalic Province; Geology, Ore Deposits and Origin, vol. 100. Montana Bur. Mines Geol., Special Publ. Balzer, W., 1982. On the distribution of iron and manganese at the sediment/water interface. Thermodynamic versus kinetic control. Geochim. Cosmochim. Acta 46, 1153–1161. Bassett, R.L., Miller, W.R., McHugh, J., Catts, J.G., 1992. Simulation of natural acid sulfate weathering in an alpine watershed. Water Resour. Res. 28, 2197–2209. Bird, D.A., 2003. Characterization of anthropogenic and natural sources of acid rock drainage at the Cinnamon Gulch abandoned mine land inventory site, Summit County, Colorado. Environ. Geol. 44, 919–932. http://dx.doi.org/10.1007/ s00254-003-0835-9. Bove, D., Walton-Day, K., Kimball, B., 2009. The use of fluoride as a natural tracer in water and the relationship to geological features: examples from the Animas River Watershed, San Juan Mountains, Silverton, Colorado. Geochem. Explor. Environ. Anal. 9, 125–138. Calderoni, G., Ferri, T., Giannetti, B., Masi, U., 1985. The behavior of thallium during alteration of the K-alkaline rocks from the Roccamonfina volcano (Campania, southern Italy). Chem. Geol. 48, 103–113. Davison, W., 1993. Iron and manganese in lakes. Earth Sci. Rev. 34, 119–163. Dzombak, D.A., Morel, F.M.M., 1990. Surface Complexation Modeling: Hydrous Ferric Oxide. Wiley-Interscience, New York. EPA, 2009. Factsheet on: Thallium. U.S. Environ. Protection Agency. (retrieved 22.07.14). Fernández-Remolar, D.C., Rodriguez, N., Gómez, F., Amils, R., 2003. Geological record of an acidic environment driven by iron hydrochemistry: the Tinto River system. J. Geophys. Res. 108, 5080. http://dx.doi.org/10.1029/2002JE001918. Furniss, G., Hinman, N.W., Doyle, G.A., Runnells, D.D., 1999. Radiocarbon-dated ferricrete provides record of natural acid rock drainage and paleoclimate changes, New World Mining District, MT, U.S.A. Environ. Geol. 37, 102–106. Gabelman, J., Furniss, G., Shaw, S., Osborne, T., Adams, J., 2005. Pre-mining Water Quality Evaluation of Swift Gulch, Phillips County, Montana. Unpublished Report Prepared by HydroSolutions Inc., for the Bureau of Land Management, Billings, Montana. Gadde, R.R., Laitinen, H.A., 1974. Studies of heavy metal adsorption by hydrous iron and manganese oxides. Anal. Chem. 46, 2022–2026. Gammons, C.H., Nimick, D.A., Parker, S.R., Cleasby, T.E., McCleskey, R.B., 2005. Diel behavior of Fe and other heavy metals in a mountain stream with acidic to neutral pH: Fisher Creek, Montana, USA. Geochim. Cosmochim. Acta 69, 2505– 2516. http://dx.doi.org/10.1016/j.gca.2004.11.020. Goddard, E.N., 1988. Geologic Map of the Judith Mountains, Fergus County, Montana. U.S. Geol. Surv., Misc. Inv. Series, Map I-1729. Gustafsson, J.P., 2010. Visual MINTEQ ver. 3.0b. KTH, Department of Land and Water Resources Engineering, Stockholm, Sweden. Hall, R.J., 1977. Petrology of Diamond Drill Core from Judith Peak-Red Mountain Area, Fergus County, Montana. M.S. Thesis, Eastern Washington Univ., Cheney, Washington. Hem, J.D., 1981. Rates of manganese oxidation in aqueous systems. Geochim. Cosmochim. Acta 45, 1369–1374.

Jacobson, A.R., McBride, M.B., Baveye, P., Steenhuis, T.S., 2005. Environmental factors determining the trace-level sorption of silver and thallium to soils. Sci. Total Environ. 345, 191–205. Jones, A.M., Collins, R.N., Waite, T.D., 2011. Mineral species control of aluminum solubility in sulfate-rich waters. Geochim. Cosmochim. Acta 75, 965–977. Kill Eagle, J.L., Gammons, C.H., Weight, W.D., Babcock, J., Jepson, W., Langner, H.W., 2009. Results and lessons learned from a continuous tracer injection test in a small mountain stream receiving acid mine drainage. Mine Water Environ. 28, 182–193. Kimball, B.A., Runkel, R.L., Walton-Day, K., Bencala, K.E., 2002. Assessment of metal loads in watersheds affected by acid mine drainage by using tracer injection and synoptic sampling: Cement Creek, Colorado, USA. Appl. Geochem. 17, 1183–1207, PII: S0883-2927(02)00017-3. Kimball, B.A., Runkel, R.L., Walton-Day, K., 2010. An approach to quantify sources, seasonal change, and biogeochemical processes affecting metal loading in streams: facilitating decisions for remediation of mine drainage. Appl. Geochem. 25, 728–740. http://dx.doi.org/10.1016/j.apgeochem. 2010.02.005. Kohrt, P.B., 1991. Alkalic rocks of the Judith Mountains, central Montana. In: Baker, D.W., Berg, R.B. (Eds.), Guidebook of the Central Montana Alkalic Province, vol. 100. Mont. Bur. Mines Geol. Special Publ., pp. 77–85. Langmuir, D., 1997. Aqueous Environmental Geochemistry. Prentice Hall, Upper Saddle River, New Jersey. Lindsey, D.A., 1985. A Gold-Mineralized Breccia Zone at Kendall, North Moccasin Mountains, Fergus County, Montana. U.S. Geol. Surv., Prof. Paper 1301-C, pp. 42–56. Lindsey, D.A., Fisher F.S., 1985. Mineralized Breccias in Intrusive Complexes of Late Cretaceous and Paleocene Age, North-Central Montana. U.S. Geol. Surv. Prof. Paper 1301-A, pp. 1–34. Machesky, M., 1990. Influence of temperature on ion adsorption by hydrous metal oxides. In: Bassett, L., Machoir, D.C. (Eds.), Chemical Modeling of Aqueous Systems I: American Chemical Symposium Series, vol. 416. Oxford University Press, pp. 282–292. Marvin, R.F., Hearn Jr., B., Mehnert, H.H., Naesser, C.W., Zartman, R.E., Lindsey, D.A., 1980. Late Cretaceous-Paleocene-Eocene igneous activity in north-central Montana. Isochron-West 29, 5–25. McKnight, D.M., Bencala, K.E., 1990. The chemistry of iron, aluminum, and dissolved organic material in three acidic, metal-enriched, mountain streams, as controlled by watershed and in-stream processes. Water Resour. Res. 26, 3087–3100. MDEQ, 2012. Montana Numeric Water Quality Standards. Montana Dept. Environ. Qual., Helena, Montana, Circular DEQ7. . Mueller, R.F., 2001. Microbially mediated thallium immobilization in bench scale systems. Mine Water Environ. 20, 17–29. Nimick, D.A., Gurrieri, J.T., Furniss, G., 2009. An empirical method for estimating instream pre-mining pH and dissolved Cu concentration in catchments with acidic drainage and ferricrete. Appl. Geochem. 24, 106–119. http://dx.doi.org/ 10.1016/j.apgeochem.2008.11.007. Nimick, D.A., Gammons, C.H., Parker, S.R., 2011. Diel biogeochemical processes and their effect on the aqueous chemistry of streams: a review. Chem. Geol. 283, 3– 17. http://dx.doi.org/10.1016/j.chemgeo.2010.08.017. Nordstrom, D.K., 1982a. Aqueous pyrite oxidation and the formation of secondary iron minerals. In: Kittrick, J.A., Fanning, D.S., Hossner, L.R. (Eds.), Acid Sulfate Weathering, vol. 10. SSSA Spec. Publ., pp. 37–56. Nordstrom, D.K., 1982b. The effect of sulfate on aluminum concentrations in natural waters: some stability relations in the system Al2O3–SO3–H2O at 298 K. Geochim. Cosmochim. Acta 46, 681–692. Nordstrom, D.K., Alpers, C.N., 1999. Geochemistry of acid mine waters. In: Plumlee, G.S., Logsdon, M.J. (Eds.), The Environmental Geochemistry of Mineral Deposits, Part A: Processes, Techniques, and Health Issues, vol. 6A. Soc. Econ. Geol. Rev. Econ. Geol., pp. 133–160. Parker, S.R., Gammons, C.H., Jones, C.A., Nimick, D.A., 2007. Role of hydrous iron and aluminum oxide formation in attenuation and diel cycling of dissolved trace metals in an ARD-affected stream. Water Air Soil Pollut. 181, 247–263. http:// dx.doi.org/10.1007/s11270-006-9297-5. Parkhurst, D.L., Appelo, C.A.J., 1999. User’s Guide to PHREEQC (Version 2)—A Computer Program for Speciation, Batch-Reaction, One-Dimensional Transport, and Inverse Geochemical Calculations: U.S. Geol. Surv., Water-Res. Invest. Rept. 99-4259. Peter, A.L., Viraraghavan, T., 2005. Thallium: a review of public health and environmental concerns. Environ. Int. 31, 493–501. Porter, K.W., Wilde, E.M., 1993. Geologic Map of the Lewistown 300  600 Quadrangle, Central Montana. Mont. Bur. Mines Geol, Open File Rept. 308. Posey, H.H., Renkin, M.L., Woogling, J., 2000. Natural acid drainage in the upper Alamosa river of Colorado. In: Proc. 5th Intl. Conf. Acid Rock Drainage. Soc. Mining Metall. Explor. (SME), Littleton, CO, USA, vol. 1, pp. 485–498. Robertson, F., 1950. Mines and Mineral Deposits (Except Fuels), Fergus County, Montana. U.S. Bur. Mines, Inf. Circ. 7544, 76 pp. Rodda, D.P., Johnson, B.B., Wells, J.D., 1996. Modeling of the effect of temperature on adsorption of lead(II) and zinc(II) onto goethite at constant pH. J. Colloid Interface Sci. 184, 365–377. Runkel, R.L., Kimball, B.A., Walton-Day, K., Verplanck, P.L., 2007. A simulation-based approach for estimating premining water quality: Red Mountain Creek, Colorado. Appl. Geochem. 22, 1899–1918.

Please cite this article in press as: Williams, G.P., et al. An investigation of acidic head-water streams in the Judith Mountains, Montana, USA. Appl. Geochem. (2015), http://dx.doi.org/10.1016/j.apgeochem.2015.05.012

G.P. Williams et al. / Applied Geochemistry xxx (2015) xxx–xxx Runkel, R.L., Walton-Day, K., Kimball, B.A., Verplanck, P.L., Nimick, D.A., 2013. Estimating instream constituent loads using replicate synoptic sampling, Peru Creek, Colorado. J. Hydrol. 489, 26–41. Runnells, D., Shepherd, T., Angino, E., 1992. Metals in water – determining natural background concentrations in mineralized areas. Environ. Sci. Technol. 26, 2316–2323. Sánchez-España, J., Yusta, I., Diez-Ercilla, M., 2011. Schwertmannite and hydrobasaluminite: a re-evaluation of their solubility and control on the iron and aluminum concentration in acidic pit lakes. Appl. Geochem. 26, 1752– 1774. Singh, S.S., 1980. Thermodynamic properties of synthetic basic aluminite [Al4(OH)10SO45H2O] from solubility data. Can. J. Soil Sci. 60, 381–384. Singh, S.S., Brydon, J.E., 1969. Solubility of basic aluminum sulfates at equilibrium in solution and in the presence of montmorillonite. Soil Sci. 107, 12–16. Sjostrom, D.J., Hren, M.T., Chamberlain, C.P., 2004. Oxygen isotope records of goethite from ferricrete deposits indicate regionally varying halocene climate change in the Rocky Mountain region, U.S.A. Quatern. Res. 61, 64–67. http:// dx.doi.org/10.1016/j.yqres.2003.08.008. Smith, K.S., 1999. Metal sorption on mineral surfaces: an overview with examples relating to mineral deposits. In: Plumlee, G.S., Logsdon, M.J. (Eds.), The Environmental Geochemistry of Mineral Deposits, Part A: Processes, Techniques, and Health Issues, vol. 6A. Soc. Econ. Geol. Rev. Econ. Geol., pp. 161–182. Trivedi, P., Axe, L., 2000. Modeling Cd and Zn sorption to hydrous metal oxides. Environ. Sci. Technol. 34, 2215–2223.

13

Trivedi, P., Axe, L., 2001. Predicting divalent metal sorption to hydrous Al, Fe, and Mn oxides. Environ. Sci. Technol. 35, 1779–1784. Verplanck, P.L., Yager, D.B., Church, S.E., Stanton, M.R., 2007. Ferricrete classification, morphology, distribution, and carbon-14 age constraints. U.S. Geol. Surv. Prof. Paper 1651, 721–744. Verplanck, P.L., Nordstrom, K.D., Bove, D.J., Plumlee, G.S., Runkel, R.L., 2009. Naturally acidic surface waters draining porphyry-related mineralized areas of the Southern Rocky Mountains, Colorado and New Mexico. Appl. Geochem. 24, 255–267. http://dx.doi.org/10.1016/j.apgeochem.2008.11.014. Wallace, S.R., 1953. The Petrology of the Judith Mountains, Fergus County, Montana. Ph.D. Diss., Univ. of Michigan, Ann Arbor, Mich. Weed, W.H., Pirsson, L.V., 1898. Geology and Mineral Resources of the Judith Mountains of Montana. U.S. Geol. Surv., Folio. Wilkie, J.A., Hering, J.G., 1996. Adsorption of arsenic onto hydrous ferric oxide: effects of adsorbate/adsorbent ratios and co-occurring solutes. Colloids Surf. A: Physicochem. Eng. Aspects 107, 97–110. Wilson, M.R., Kyser, T.K., 1988. Geochemistry of porphyry hosted Au–Ag deposits in the Little Rocky Mountains Montana. Econ. Geol. 83, 1329–1346. Wireman, M., Stover, B., 2011. Hard-rock mining and water resources. Ground Water 49, 310–316. Woodward, L.A., 1995. Metallic Mineral Deposits of the Judith Mountains, Central Montana. Mont. Bur. Mines Geol., Memoir 67, 78 pp. Zhang, X., Spry, P.G., 1994. Petrological, mineralogical, fluid inclusion, and stable isotope studies of the Gies gold-silver telluride deposit, Judith Mountains, Montana. Econ. Geol. 89, 602–627.

Please cite this article in press as: Williams, G.P., et al. An investigation of acidic head-water streams in the Judith Mountains, Montana, USA. Appl. Geochem. (2015), http://dx.doi.org/10.1016/j.apgeochem.2015.05.012