Water, Air, & Soil Pollution

, 225:1694

Mercury and Methylmercury Dynamics in the Hyporheic Zone of an Oregon Stream

Authors

    • U.S. Geological Survey
  • Kenneth E. Bencala
    • U.S. Geological Survey
  • Dennis A. Wentz
    • U.S. Geological Survey
  • David P. Krabbenhoft
    • U.S. Geological Survey
Article

DOI: 10.1007/s11270-013-1694-y

Cite this article as:
Hinkle, S.R., Bencala, K.E., Wentz, D.A. et al. Water Air Soil Pollut (2014) 225: 1694. doi:10.1007/s11270-013-1694-y

Abstract

The role of the hyporheic zone in mercury (Hg) cycling has received limited attention despite the biogeochemically active nature of this zone and, thus, its potential to influence Hg behavior in streams. An assessment of Hg geochemistry in the hyporheic zone of a coarse-grained island in the Coast Fork Willamette River in Oregon, USA, illustrates the spatially dynamic nature of this region of the stream channel for Hg mobilization and attenuation. Hyporheic flow through the island was evident from the water-table geometry and supported by hyporheic-zone chemistry distinct from that of the bounding groundwater system. Redox-indicator species changed abruptly along a transect through the hyporheic zone, indicating a biogeochemically reactive stream/hyporheic-zone continuum. Dissolved organic carbon (DOC), total Hg, and methylmercury (MeHg) concentrations increased in the upgradient portion of the hyporheic zone and decreased in the downgradient region. Total Hg (collected in 2002 and 2003) and MeHg (collected in 2003) were correlated with DOC in hyporheic-zone samples: r2 = 0.63 (total Hg-DOC, 2002), 0.73 (total Hg-DOC, 2003), and 0.94 (MeHg-DOC, 2003). Weaker Hg/DOC association in late summer 2002 than in early summer 2003 may reflect seasonal differences in DOC reactivity. Observed correlations between DOC and both total Hg and MeHg reflect the importance of DOC for Hg mobilization, transport, and fate in this hyporheic zone. Correlations with DOC provide a framework for conceptualizing and quantifying Hg and MeHg dynamics in this region of the stream channel, and provide a refined conceptual model of the role hyporheic zones may play in aquatic ecosystems.

Keywords

MercuryMethylmercuryHyporheic zoneGroundwater/surface water interactionsOregonWillamette River

1 Introduction

Mercury (Hg) is a potent neurotoxin, is widespread in aquatic ecosystems, and, in its methylated form, biomagnifies through the food chain (U.S. Environmental Protection Agency 1997). Hg contamination is so widespread in the USA that all 50 states have fish consumption advisories due to elevated Hg levels (U.S. Environmental Protection Agency 2011). Hg in streams often is dominated by contributions from nonpoint sources, especially atmospheric deposition (United Nations Environment Programme 2008) and erosion of soils, sediment, and rocks; point sources, such as mine tailings and industrial waste, also can be important. Hg is removed from streams by volatilization, sorption onto geologic materials, and biologic uptake. The process of Hg exchange between surface water and underlying/adjacent geologic material can be one that either reduces or increases aqueous Hg concentrations. The direction of these exchanges may vary in space and time, and the direction and extent of these exchanges may be affected by dissolved organic carbon (DOC) dynamics because DOC forms strong complexes with Hg (Ravichandran 2004). DOC can exert strong controls on Hg mobilization and transport in streams (Brigham et al. 2009), lakes (Chadwick et al. 2006), wetlands (Hall et al. 2008), pore water (Marvin-DiPasquale et al. 2009), and groundwater (Barringer et al. 2010).

Exchange of Hg between streambed (including wetland) sediment and stream water has received considerable attention (e.g., Domagalski 2001; Marvin-DiPasquale et al. 2009; Flanders et al. 2010), in part because of the importance of streambed sediment in the methylation of Hg (conversion of inorganic Hg to the more toxic methylmercury, or MeHg, form) in some stream systems (Morel et al. 1998; Benoit et al. 2003; Wiener et al. 2003). However, the identification of and distinction between diffusional and advective exchange at the streambed/stream-water interface is seldom made, with the exception of studies focusing on groundwater discharge through streambed sediment (e.g., Krabbenhoft et al. 1995; Barringer et al. 2010). In particular, the role of advective flow through the hyporheic zone—the region underneath and adjacent to streams where stream water, groundwater, and sediment interact—has received little attention in the study of Hg dynamics. Such attention is warranted because the hyporheic zone is an environment with large volume and surface area that is important for the exchange of water (Bencala 1993), heat (Arrigoni et al. 2008), and solutes, including O2 (dissolved oxygen) (Greig et al. 2007), nutrients (Hill et al. 1998), DOC (Findlay et al. 2003), and trace elements, such as As, Co, Ni, Pb, and Zn, associated with mineralization or mining (Fuller and Harvey 2000; Palumbo-Roe et al. 2012; Moldovan et al. 2013). The timescales of hyporheic processes can be comparable to those of stream water and stream solute transport (Edwards 1998). Cyclical advection of water and solutes back and forth between a stream and the biogeochemically active hyporheic zone creates the potential for compounded effects on stream chemistry (Bencala et al. 2011), and these effects have the potential to be important for Hg cycling in aquatic systems.

In two comprehensive investigations of Hg cycling in watersheds, Stoor et al. (2006) and Creswell et al. (2008) observed Hg mobilization and Hg methylation in streambed sediment and attributed these reactions to hyporheic-zone processes. It appears, however, that the term “hyporheic zone” was used to indicate a near-stream environment receiving groundwater discharge. Working at the same sites studied by Creswell et al. (2008), both Krabbenhoft et al. (1995) and Meyer (2004) reported steady discharge of regional groundwater through streambed sediment into the stream. Similarly, groundwater discharge through streambed sediment also was the mechanism of Hg transport in the watershed studied by Stoor et al. (2006). These facts do not diminish the value of these contributions to the understanding of watershed Hg cycling, but they do serve to illustrate the distinction between (1) a near-stream environment receiving primarily unidirectional water movement (e.g., Stoor et al. 2006; Creswell et al. 2008), and (2) a near-stream environment in which stream water alternatingly enters and leaves a hyporheic zone (e.g., this paper).

In this contribution, Hg and other chemical data from a transect through a hyporheic zone at an island in a Hg-mine-impacted stream are presented. The purpose of this effort is to test the hypotheses that (1) the hyporheic zone can be a spatially dynamic zone of Hg cycling, and (2) DOC or other analytes can help explain patterns of Hg mobilization and attenuation in the hyporheic zone. Hyporheic-zone-water and stream-surface elevations were measured to delineate hyporheic flow regimes. Aqueous-chemical data [filtered total Hg (FTHg), filtered MeHg (FMeHg), DOC, major-ion concentrations, and field parameters] and sediment data (total Hg and MeHg concentrations, percentage loss on ignition, and percentage <62.5 μm) were collected from the hyporheic zone and associated stream to resolve our hypotheses.

2 Study Design and Methods

A small island (approximately 40 m length in July 2003) in the Coast Fork Willamette River near London, OR, contains a hyporheic zone that is the focus of this paper (Fig. 1). This island is situated at river kilometer 56.6 (river mile 35.2, USGS “Cottage Grove Lake” 7.5′ topographic map), at an elevation of 250 m above mean sea level. The island is composed of gravel and cobbles in a fine-grained matrix. Water-level measurements indicate that stream water flows through this coarse-grained island in response to hydraulic gradients established by adjacent stream elevations (Fig. 1). Average linear velocities through the island, based upon hydraulic conductivities of 40 to 200 m/day (Conlon et al. 2005), the hydraulic gradient observed on July 24, 2003 (Fig. 1), and a porosity 0.30, would be 0.75 to 3.7 m/day and would equate to a time-of-travel along the center axis of the island of 7 to 33 days. Flow directions and time-of-travel vary over time in response to changing stream surface elevations and to a lesser extent direct additions of precipitation, but the direction of flow can be expected to approximately parallel the direction of streamflow under most conditions, especially those encountered during the low-precipitation conditions that dominate Oregon in summer.
https://static-content.springer.com/image/art%3A10.1007%2Fs11270-013-1694-y/MediaObjects/11270_2013_1694_Fig1_HTML.gif
Fig. 1

An island in the Coast Fork Willamette River was the site of hyporheic zone investigations in 2002 and 2003; locations of year 2003 wells and the July 24, 2003 water-table contours are shown in Fig. 1; year 2002 wells were in slightly different locations (Island located approximately 2 km N of town of London, OR)

Sediment accumulations are thin at and near the island, as evidenced by bedrock outcrops in and along the stream near the study site. This bedrock is minimally fractured volcanic tuff with permeability several orders of magnitude smaller than that of near-stream sand and gravel (Conlon et al. 2005). The Coast Fork Willamette River at this site drains 195 km2 of forested land in a basin composed of volcanic and pyroclastic rocks. An abandoned cinnabar (Hg ore) mine (Ambers and Hygelund 2001) is located 10 km upstream of the site. Hg concentrations are elevated in stream and reservoir sediment in the basin (Wentz et al. 1998; Ambers and Hygelund 2001). Hg contamination in the Coast Fork Willamette River basin has resulted in exposure effects in birds that feed on aquatic macroinvertebrates (Henny et al. 2005). Hope (2006) discussed the development of a Hg Total Maximum Daily Load for the Willamette River following health advisories triggered by MeHg in fish.

Data collection occurred in 2002 and 2003. Water and sediment samples from the hyporheic zone and stream were collected September 16–18, 2002 (except streambed sediment: August 14, 2002), and again July 21–23, 2003 (Table 1). Samples for analysis of FMeHg were collected in 2002, but inadvertently not analyzed (laboratory oversight); however, FMeHg samples were collected and analyzed in 2003. Transient flow conditions in 2002 were caused by a rain event (3.6 cm at Cottage Grove Dam, 8 km to the north) (National Climatic Data Center 2012) that began late on September 16 and lasted through the next day, resulting in a stream elevation rise of 11.6 cm (stream discharge increasing from 0.24 m3/s to an estimated 1.7 m3/s) from the beginning of the storm to a point close to peak stream stage. Dry weather and stable flow conditions were present for the 2003 effort. Five shallow dug wells (approximately 10 cm below water table) along a hyporheic-zone transect were installed, developed (extensive pumping), and sampled for aqueous hyporheic-zone samples in both years. Five shallow dug holes for collection of sediment also were installed adjacent to these water-sample wells in both years. Locations of wells sampled in 2003, and the water-table contours of July 24, 2003, are shown in Fig. 1; wells sampled in 2002 were in slightly different locations (Table 1). Additional hyporheic-zone samples were collected from streambed sediment beneath the stream just upstream from the upgradient end of the island in 2003 (“sub-stream samples,” collected from 5.1 and 9.4 cm depths below the streambed). Stream-water samples were collected in 2002 and 2003. A spring adjacent to the island (north bank of stream; Table 1), issuing from the contact between permeable alluvial materials and underlying low-permeability volcanic tuff, represents shallow groundwater and also was sampled in 2002 and 2003. A nearby supply well tapping volcanic tuff represents deep groundwater and was sampled in 2002 (Table 1); this 37-m-deep well serves a small, rural school.
Table 1

Analytical data for hyporheic water, stream water, spring water, groundwater, and sediment samples (island located approximately 2 km N of London, OR; supply well located 104 m NW of island)

Site

Sampling date

Distance from upstream edge of water (m)

FTHg (ng/L)

FMeHg (ng/L)

STHg (ng/g)

SMeHg (ng/g)

Sediment, loss on ignition (%)

Sediment, dry sieved, % < 62.5μm

Field pH (Standard Units)

Field specific conductance (μS/cm)

DOC (Dissolved organic carbon) (mg/L)

Coast Fork Willamette River spring nr London

September/17/2002

a

5.19

a

a

a

a

a

7.2

62

8.9

Coast Fork Willamette River nr London

August/14/2002

a

a

a

625

4.73

11.9

39

a

a

a

Coast Fork Willamette River nr London (0.24 m3/s)

September/16/2002

a

3.42

a

a

a

a

a

7.6

107

1.3

Hyporheic zone: Island well 2002-1

September/18/2002

2.17

15.7

a

332

0.41

3.8

4.7

6.5

106

3.1

Hyporheic zone: Island well 2002-2

September/18/2002

3.34

9.99

a

233

<0.12

3.8

2.7

6.6

153

2.9

Hyporheic zone: Island well 2002-3

September/17/2002

8.34

7.25

a

200

0.24

3.6

8.5

5.9

176

3.0

Hyporheic zone: Island well 2002-4

September/17/2002

15.48

5.69

a

230

0.21

4.0

7.8

6.2

141

1.9

Hyporheic zone: Island well 2002-5

September/16/2002

21.45

3.62

a

145

0.34

3.8

5.3

6.0

172

1.5

Groundwater from deep (37 m) supply well

September/19/2002

a

0.25

a

a

a

a

a

6.7

279

<0.3

Coast Fork Willamette River spring nr London

July/23/2003

a

1.25

<0.04

a

a

a

a

7.4

74

1.7

Coast Fork Willamette River nr London (0.39 m3/s)

July/21/2003

a

3.40

0.25

421

2.54

9.5

12

8.4

94

1.4

Hyporheic zone: Sub-stream, 0.051 m deep

July/23/2003

a

15.8

0.45

a

a

a

a

7.4

94

1.8

Hyporheic zone: Sub-stream, 0.094 m deep

July/23/2003

a

12.6

0.35

a

a

a

a

7.2

97

1.4

Hyporheic zone: Island well 2003-1

July/22/2003

1.07

18.3

0.69

117

1.26

3.5

3.7

6.8

130

2.8

Hyporheic zone: Island well 2003-2

July/23/2003

3.29

7.54

0.37

239

<0.10

3.1

2.5

6.5

187

1.7

Hyporheic zone: Island well 2003-3

July/22/2003

7.50

3.32

0.12

142

<0.10

4.5

3.8

6.4

118

1.2

Hyporheic zone: Island well 2003-4

July/23/2003

13.29

4.31

0.11

11.5

<0.10

5.2

3.7

6.4

108

1.0

Hyporheic zone: Island well 2003-5

July/22/2003

21.52

4.17

0.08

102

<0.10

3.7

1.8

6.7

130

0.9

Site

O2 (Field dissolved oxygen) (mg/L)

Ca2+ (mg/L)

Mg2+ (mg/L)

Na+ (mg/L)

K+ (mg/L)

Fe (μg/L)

Mn (μg/L)

SiO2 (mg/L)

SO42 -(mg/L)

Cl- (mg/L)

HCO3- (mg/L)

Coast Fork Willamette River spring nr London

7.0

7.1

1.76

7.38

1.02

49

6

5.75

0.5

3.69

36

Coast Fork Willamette River nr London

a

a

a

a

a

a

a

a

a

a

a

Coast Fork Willamette River nr London (0.24 m3/s)

10.1

12.5

2.51

6.31

0.53

45

4

5.29

2.3

3.32

54

Hyporheic zone: Island well 2002-1

1.3

13.1

2.54

5.40

0.65

40

307

8.71

2.6

3.22

54

Hyporheic zone: Island well 2002-2

0.1

17.2

3.26

6.09

0.77

21

310

9.16

1.8

3.92

74

Hyporheic zone: Island well 2002-3

1.2

18.7

3.52

6.01

0.69

255

179

11.1

1.9

6.92

74

Hyporheic zone: Island well 2002-4

0.3

17.5

3.32

6.01

0.71

225

747

10.5

1.2

4.37

76

Hyporheic zone: Island well 2002-5

0.3

21.2

3.88

7.34

0.87

178

692

10.1

0.4

3.85

94

Groundwater from deep (37 m) supply well

4.2

12.8

2.38

42.9

0.77

<5

5

5.56

1.2

42.3

83

Coast Fork Willamette River spring nr London

6.5

7.5

1.76

3.71

0.52

8

<4

12.8

0.4

1.83

42

Coast Fork Willamette River nr London (0.39 m3/s)

8.7

11.8

2.27

3.36

0.55

50

4

8.83

1.6

2.52

50

Hyporheic zone: Sub-stream, 0.051 m deep

a

11.5

2.19

3.35

0.61

34

157

8.91

1.7

2.73

48

Hyporheic zone: Sub-stream, 0.094 m deep

a

12.0

2.31

3.40

0.63

16

75

9.00

1.6

2.63

50

Hyporheic zone: Island well 2003-1

0.5

17.0

3.10

3.63

0.97

36

324

10.2

1.3

2.72

70

Hyporheic zone: Island well 2003-2

0.3

25.9

4.56

7.25

1.04

173

414

10.7

0.5

2.62

103

Hyporheic zone: Island well 2003-3

0.4

16.2

2.99

3.38

0.78

645

465

10.0

0.7

2.75

65

Hyporheic zone: Island well 2003-4

0.7

14.4

2.70

3.37

0.67

174

438

9.47

0.6

2.69

61

Hyporheic zone: Island well 2003-5

0.3

17.3

3.21

3.60

0.73

5

213

9.79

0.5

3.79

72

FTHg filtered total Hg; FMeHg filtered MeHg; STHg total Hg in sediment; SMeHg MeHg in sediment; m meters; mg milligrams; μg micrograms; ng nanograms; uS/cm microsiemens per centimeter at 25 C

aNot analyzed

2.1 Aqueous Hg Sample Collection and Processing Procedures

Collection and processing of water samples for Hg (FTHg in 2002; both FTHg and FMeHg in 2003) were done using clean-sampling techniques similar to those described by Brigham et al. (2008), but adapted to hyporheic-zone sampling, as described below. Clean-sampling techniques included acid washing of sample collection materials followed by copious flushing with native water; the use of plastic-shelled collection and processing chambers; and the wearing of new, disposable, shoulder-length polyethylene gloves under new, disposable, powderless, nitrile gloves. Field equipment blanks for FTHg and FMeHg documented that the extent of contamination attributable to collection, processing (and analysis) of samples was negligible: FTHg at 0.09 ng/L (2002) and 0.06 ng/L (2003); FMeHg at <0.04 ng/L (2003).

Hyporheic-zone samples were collected from shallow wells dug with plastic shovels (Krabbenhoft and Babiarz 1992). Samples were collected using fluorocarbon polymer tubing and a peristaltic pump (short section of C-Flex peristaltic tubing through pump head), filtered into sample bottles inside wind- and dust-sheltered collection chambers using in-line filtration with fluorocarbon polymer filter units containing 0.7-μm, baked (550 °C) quartz-fiber filters. Samples were acidified in wind- and dust-sheltered sample preservation chambers in a dedicated water-quality sample processing vehicle using 6N HCl to produce a sample solution of approximately 1 % HCl, and were stored in the dark until analysis. Additional hyporheic-zone samples, collected from sediment beneath the streambed, were collected from a fluorocarbon polymer drivepoint inserted into streambed sediment and connected by fluorocarbon polymer components to the sample collection apparatus described above.

Samples from the stream, spring, and supply well were collected on-site as raw (unfiltered) water samples using either a factory-sealed 2-L polyethylene terephthalate copolyester, glycol-modified bottle or a 3-L fluorocarbon polymer bottle. Stream samples were collected by dipping a sample collection bottle into the centroid of flow. Filtering and preservation were done inside chambers in the processing vehicle immediately following collection. Filtering was by vacuum filtration through fluorocarbon polymer filter units [containing 0.7-μm, baked (550 °C) quartz-fiber filters] directly into sample bottles.

2.2 Sample Collection and Processing Procedures for DOC, Major Ions, and Field Parameters

Collection and processing of aqueous samples for analysis of DOC (0.7 μm filtered with 450 °C-baked glass fiber filters) and major ions (plus Mn, Fe, Si; 0.45 μm filtered with one-time-use sealed capsule filters) were done according to published procedures (U.S. Geological Survey 1999). DOC samples were kept in ice following filtering and during transport to the analytical laboratory. Measurements of O2, pH, and specific conductance were measured either in-line (flow-through cell) or in situ (following collection of aqueous samples), and alkalinity (0.45 μm filtered) was titrated in the field (U.S. Geological Survey 1999).

2.3 Sediment Sample Collection and Processing Procedures

Streambed sediment samples were collected as composite (integrated) samples; they were collected from the 0–2-cm depth interval at five to ten undisturbed, submerged depositional areas along the stream reach. Hyporheic-zone sediment was collected from the uppermost 2 cm of saturated sediment in dug wells. Sediment was collected using a fluorocarbon polymer scoop, composited in polypropylene or glass jars, and subsampled using fluorocarbon polymer utensils. Samples for determination of Hg and loss on ignition were frozen immediately by placing on dry ice. Samples to be analyzed for size fraction were stored in a cool, dark location. Greater detail is provided by Lutz et al. (2008).

2.4 Analytical Techniques

Analyses for FTHg and FMeHg (water), and total Hg and MeHg (sediment), were performed at the USGS Mercury Research Laboratory in Middleton, WI (http://wi.water.usgs.gov/mercury-lab/). Total Hg analysis followed U.S. EPA Method 1631. MeHg analytical procedures were according to DeWild et al. (2002), which includes standard distillation and ethylation procedures followed by chromatographic separation and CVAFS detection. Loss on ignition (a proxy for organic carbon content) in sediment was also performed at the USGS Mercury Research Laboratory, by decomposition at 550 °C (Heiri et al. 2001). DOC was determined at the USGS National Water Quality Laboratory in Lakewood, CO, by the methods of Brenton and Arnett (1993). Alkalinity, pH, and specific conductance were measured on-site according to U.S. Geological Survey (1999), following daily on-site calibration. Dissolved oxygen was measured by amperometry using a polarographic sensor. The dissolved-oxygen meter was compared against an anoxic solution prior to each field campaign (reading 0.01 mg/L in 2002 and 0.03 mg/L in 2003), and calibrated daily on-site with an air-calibration chamber in water. Major ion analyses were completed at USGS National Research Program laboratories in Menlo Park, CA, using methods described by Fishman and Friedman (1989). Dry sieving of sediment samples to determine percentage <62.5 μm was done at the USGS Cascade Volcano Observatory in Vancouver, WA, using methods of Guy and Norman (1970).

3 Results and Discussion

Hyporheic flow through the island is evident in the water-table geometry (Fig. 1) and major-ion chemistry (Table 1). The relative proportions of major ions in hyporheic-zone samples and stream samples were similar (Fig. 2), with cations dominated by the divalent cations Ca2+ and Mg2+. Biogeochemical reactions in the hyporheic zone led to slight increases in the concentrations of most ions, particularly HCO3 and the divalent cations Ca2+ and Mg2+ (Fig. 3 and Table 1). These changes are part of classic biogeochemical evolution of groundwater (Bricker and Garrels 1967; Garrels and Mackenzie 1967) that includes plant respiration through roots and bacterial decomposition of organic carbon, which increase HCO3 concentrations; dissolution/alteration of primary minerals and volcanic glass coupled with precipitation/formation of secondary minerals and amorphous phases, which generally increases cation and SiO2 concentrations; redox reactions (discussed below); and possibly minor ion exchange and/or concentration by evapotranspiration. Some heterogeneity was observed in the biogeochemical progression along this hyporheic zone (Fig. 3), possibly reflecting the mixing of different hyporheic-zone flow components and/or temporally variable chemistry of stream inputs to the hyporheic zone. However, the consistency in the relative proportions of major ions among hyporheic and stream samples (Fig. 2) indicates the presence of a flow system isolated from the regional groundwater system and is consistent with the flow system controls expected from a geologic framework consisting of an island of highly permeable fluvial materials resting upon a low-permeability bedrock surface.
https://static-content.springer.com/image/art%3A10.1007%2Fs11270-013-1694-y/MediaObjects/11270_2013_1694_Fig2_HTML.gif
Fig. 2

Trilinear diagram showing major-ion character of the Coast Fork Willamette River, hyporheic zone water samples, spring samples, and a deep groundwater sample (37-m-deep well adjacent to hyporheic zone site)

https://static-content.springer.com/image/art%3A10.1007%2Fs11270-013-1694-y/MediaObjects/11270_2013_1694_Fig3_HTML.gif
Fig. 3

The geochemical evolution of water moving from the Coast Fork Willamette River into and through the hyporheic zone of the island is reflected in the concentrations of several major ions in 2002 (left) and 2003 (right); hyporheic-zone samples plotted relative to distance downgradient from beginning of hyporheic-zone transect (Fig. 1); stream samples are arbitrarily located 5 m to the left of the zero-distance mark; sub-stream samples (hyporheic-zone samples collected from beneath the streambed at depths of 5.1 and 9.4 cm) are plotted 2.25 m to the left of the zero-distance mark, reflecting the location of these samples relative to the edge of the island at the location of the hyporheic-zone transect (SO42− data shown in Fig. 4 alongside data for other redox indicator species)

Neighboring samples included deep groundwater, represented by a sample from a single nearby well (37 m below land surface), and spring water (Table 1). Deep groundwater was distinct from stream and hyporheic-zone water, having a monovalent-cation signature (Fig. 2)—typical for regional bedrock groundwater in the southern Willamette Basin. For example, monovalent cations comprised 81 ± 21 % (mean ± 1σ) of major cations in ten bedrock groundwater samples from the southern Willamette Basin reported by Frank (1973). In contrast, water from a spring issuing from alluvium on the north bank of the Coast Fork Willamette River had a divalent-cation character, yet was more dilute than hyporheic-zone and stream water (Table 1) and differed from stream and hyporheic samples in having slightly greater proportions of monovalent cations and smaller proportions of divalent cations (Fig. 2).

Taken together, stream, hyporheic, and groundwater data describe a chemically dynamic stream/hyporheic-zone continuum operating within a physical setting controlled primarily by stream-water fluxes through the island and chemically distinct from other nearby groundwater and spring water regimes. However, although major-ion chemistry changed only gradually and exhibited uniform major-ion proportions along the hyporheic-zone transect, concentrations of redox-indicator species changed abruptly, reflecting a biogeochemically active hyporheic zone (Fig. 4 and Table 1). Conditions evolved from initially oxic conditions in the stream water to reduced conditions in the hyporheic zone. Measured O2 concentrations in hyporheic-zone samples were mostly <1 mg/L, whereas Mn and Fe concentrations generally were elevated. Consumption of O2 and mobilization of Mn and Fe indicate the presence of terminal electron acceptor processes, likely involving oxidation of sedimentary organic carbon (McMahon and Chapelle 2008). With no apparent hyporheic-zone sources, SO42− concentrations at the upgradient hyporheic-zone sites mirrored stream SO42− concentrations and then steadily decreased with increasing distance downgradient. SO42− loss may indicate SO42− reduction, although this cannot be corroborated as H2S was not measured. This biogeochemical evolution to (or possibly beyond) Fe-reducing conditions occurred in spite of the coarse texture of hyporheic-zone sediment, and probably reflects the abundance of sedimentary organic carbon as indicated by sediment loss-on-ignition data (Table 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs11270-013-1694-y/MediaObjects/11270_2013_1694_Fig4_HTML.gif
Fig. 4

Redox indicator species, DOC, FTHg and FMeHg concentrations in 2002 (left) and 2003 (right); hyporheic-zone samples plotted relative to distance downgradient from beginning of hyporheic zone transect (Fig. 1); plotting convention for stream and sub-stream samples (hyporheic-zone samples collected from beneath the streambed at depths of 5.1 and 9.4 cm) defined in Fig. 3 (FMeHg was not analyzed in 2002)

DOC concentrations along the hyporheic-zone transect also demonstrated the presence of biogeochemically active conditions (Fig. 4). DOC concentrations in the furthest upgradient portion of the hyporheic zone increased relative to stream DOC concentrations, and then steadily decreased downgradient as DOC was oxidized (contributing to the increases in HCO3) and possibly also lost to sorption (Battin 1999). These patterns of DOC occurrence could reflect spatially variable concentrations or reactivity of sedimentary organic carbon, with fast rates of DOC production/mobilization relative to DOC attenuation (oxidation or sorption) in upgradient portions of the hyporheic zone, followed by slower rates of DOC production/mobilization and/or greater DOC attenuation in the downgradient zone. Although a source of abundant or reactive organic carbon in the upgradient portion of this hyporheic zone has not been identified, it is possible that DOC is mobilized from recently deposited organic matter such as decaying algae or leaf litter that settles on streambed surfaces and becomes incorporated into shallow streambed sediment. Such organic-rich streambed sediment would be encountered by stream water upon first entering hyporheic flowpaths (Battin 1999). Supporting this hypothesis are the loss-on-ignition data, which indicate greater organic carbon concentrations in streambed sediment than in island hyporheic zone sediment (Table 1). Certainly, reversals in the balance between solute production/mobilization and solute attenuation in hyporheic zones, such as those observed here, are not unexpected. For example, Zarnetske et al. (2011) observed a similar pattern of solute dynamics in a hyporheic zone in an agricultural area, where an increase in NO3 concentrations (nitrification) in the upgradient region was followed by a decrease (denitrification) downgradient. An alternative hypothesis for the observed variations in DOC concentrations along the hyporheic transect could be based on temporally variable stream DOC inputs to the hyporheic zone. However, it would be unlikely that both the late summer synoptic in 2002 and the early summer synoptic in 2003 would be preceded by steadily increasing DOC concentrations in streams, only to be followed at the time of the synoptics by sudden decreases in stream DOC concentrations.

Patterns of Hg along the hyporheic-zone transect indicate that Hg responds to the biogeochemically dynamic conditions of this hyporheic zone. Concentrations of FTHg and FMeHg in the Coast Fork Willamette River (FTHg, 3.42 and 3.40 ng/L; FMeHg, 0.25 ng/L; Table 1) were greater than concentrations measured by Brigham et al. (2009) in 46 sets of samples collected between 2002 and 2006 from two Oregon streams: Lookout Creek (a forested site) and Beaverton Creek (an urban site). Hg concentrations in the Coast Fork Willamette River likely are elevated due to historical upstream mining activities (Ambers and Hygelund 2001) (although it is noted that these aqueous concentrations are low compared to those observed in some other mining-impacted streams, e.g., Domagalski 2001). In hyporheic zone samples, FTHg and FMeHg concentrations exhibited initial increases relative to stream water, followed by decreases along the hyporheic-zone transect (Fig. 4). These data demonstrate that the hyporheic zone can be an active zone of Hg cycling. Furthermore, DOC was correlated with FTHg (Fig. 5a) and FMeHg (Fig. 5b). Values of r2 for these correlations were 0.63 for DOC/FTHg for late summer 2002; 0.73 for DOC/FTHg for early summer 2003; and 0.94 for DOC/FMeHg for early summer 2003 (no FMeHg data from 2002). The observed correlations between DOC and either FTHg or FMeHg are consistent with a hypothesis that DOC production and subsequent oxidation or sorption exerts a major control on total Hg and MeHg mobilization, transport, and fate in this hyporheic zone. Additionally, it is possible that the different relations between FTHg and DOC in 2002 and 2003 (greater total Hg mass per unit mass of DOC in 2003; Fig. 5a) reflected seasonal differences in hyporheic-zone DOC reactivity (Canuel and Martens 1996). The strength of DOC/Hg complexes is a function of DOC character (Babiarz et al. 2001; Shanley et al. 2008; Dittman et al. 2009) and the stronger association of Hg with DOC in early summer 2003 than in late summer 2002 could reflect temporally variable DOC character present at different times of year in this hyporheic zone. However, in spite of some degree of difference in relations between FTHg and DOC in 2002 and 2003, similar biogeochemical evolution was observed in field investigations occurring in different years and under different hydrologic conditions. These observations suggest that overall patterns of redox progression, DOC production and subsequent DOC attenuation, and Hg mobilization/attenuation are at least occasional features of this hyporheic zone, and underscore the potential importance of the hyporheic zone for Hg cycling in stream systems.
https://static-content.springer.com/image/art%3A10.1007%2Fs11270-013-1694-y/MediaObjects/11270_2013_1694_Fig5_HTML.gif
Fig. 5

Relations between a FTHg and DOC, and b FMeHg and DOC for hyporheic-zone samples; the relation between FTHg and DOC for year 2002 samples (late summer; r2 = 0.63, blue diamonds) differs from that of year 2003 samples (early summer; r2 = 0.73, red circles), possibly indicating seasonal differences in DOC reactivity; FMeHg data (b) were available for year 2003 only

DOC dynamics appear to control increases and subsequent decreases in aqueous Hg and MeHg concentrations in this hyporheic zone. These linkages between DOC and Hg/MeHg dynamics are consistent with previous findings in other aquatic compartments (Fitzgerald and Lamborg 2003, Chadwick et al. 2006, Hall et al. 2008, Brigham et al. 2009, Marvin-DiPasquale et al. 2009, Barringer et al. 2010). SO42− reduction (possibly indicated by decreasing SO42− concentrations; Fig. 4) might be occurring along this hyporheic transect, and SO42−-reducing bacteria could be methylating Hg (e.g., Morel et al. 1998), but SO42− reduction does not appear to control the mobilization of MeHg into solution at this site (Fig. 4). Although DOC complexation appears to exert a fundamental control over Hg and MeHg processing in this hyporheic zone, alternative explanations for the observed increases in FTHg and FMeHg in the upgradient portions of this hyporheic zone could conceivably include (1) Hg contributions from mixing with underlying regional groundwater, (2) Hg introduced by atmospheric deposition, (3) Hg introduced by island flooding during high stream discharge events, and (4) Hg added during sample collection/processing/analysis. Each of these four alternative explanations appears unlikely for the following reasons:
  1. 1.

    Hg contributions from mixing with underlying regional groundwater. The apparent isolation of this hyporheic zone from regional groundwater, based on the geologic framework and evidenced in the major-ion chemistry as discussed above, argues against a regional groundwater source of Hg to this hyporheic zone. Furthermore, the low concentration of Hg in groundwater from the nearby supply well (FTHg = 0.25 ng/L; Table 1) suggests that regional groundwater would be an unlikely source of the elevated Hg.

     
  2. 2.

    Hg introduced by atmospheric deposition. Precipitation at the National Atmospheric Deposition Program/Mercury Deposition Network Lookout Creek site (southern Willamette Basin) contains precipitation volume-weighted total Hg on the order of 3–4 ng/L and MeHg on the order of 0.1–0.2 ng/L (Brigham et al. 2009), too low to account for the elevated Hg concentrations observed in this hyporheic zone. Atmospheric deposition also cannot explain the observed pattern of elevated Hg concentrations in the upgradient portion and lower Hg concentrations in the downgradient portion of the island during repeated sampling events (Fig. 4).

     
  3. 3.

    Hg introduced by island flooding during high stream discharge events. Hg transport during island flooding does not explain the observed pattern of elevated FTHg concentrations in the upgradient portion and lower FTHg concentrations in the downgradient portion of the island during repeated sampling events (Fig. 4). Furthermore, our measured Hg concentrations for (dry season) stream samples (FTHg, 3.42 and 3.40 ng/L; FMeHg, 0.25 ng/L) are similar to, but slightly greater than, typical concentrations present in seasonal (quarterly) samples collected by Oregon Department of Environmental Quality (ODEQ) at a nearby site (Coast Fork Willamette River above Cottage Grove Reservoir, 4 km downstream from our site) between October 2002 and June 2003 (FTHg, 2.66 ± 0.97 ng/L; FMeHg, 0.147 ± 0.063 ng/L; mean ± 1σ; Agnes Lut, ODEQ, written communication 2012).

     
  4. 4.

    Hg added during sample collection/processing/analysis. Field equipment blanks, containing FTHg at 0.06 to 0.09 ng/L and FMeHg at <0.04 ng/L, demonstrated that non-contaminating Hg sampling, processing, and analytical techniques were in place.

     
Correlations between FTHg and hyporheic-zone sediment characteristics were notably weaker than correlations between FTHg and DOC, and were not statistically significant (Table 2). This could indicate that aqueous Hg concentrations in this hyporheic zone are controlled to a greater extent by mobilization (e.g., complexation with DOC) and attenuation (DOC oxidation or sorption) processes than by the strength of solid-phase Hg sources (mass available for dissolution or desorption) or sinks (sites available for sorption). However, weak correlations between FTHg and sediment characteristics could also reflect the fact that whereas Hg and DOC are kinematically transported through the hyporheic zone in flowing water, the solid-phase variables represent sediment qualities near sampling sites and are not necessarily representative of the sequence of solid phases contacted by water prior to arrival at sampling sites. Thus, the absence of strong correlations between FTHg concentrations and sediment characteristics does not mean that sediment characteristics, such as the concentration or abundance of Hg, organic carbon, or silt-and-clay-sized particles, do not contribute to Hg source and fate.
Table 2

Correlation of FTHg with DOC and with solid-phase variables for hyporheic zone samples (data from 2002 and 2003 combined)

Correlation between FTHg and:

Slope of correlation

r2

p

n

Aqueous variable:

    

 DOC

Positive

0.36

0.04

12

Solid-phase variables:

    

 Total Hg in sediment

Positive

0.17

0.24

10

 Sediment loss on ignition (measure of organic carbon content)

Negative

0.17

0.23

10

 Sediment, dry sieved, % < 62.5μm

Negative

<0.01

0.85

10

The highly dynamic nature of FTHg and FMeHg concentrations along the hyporheic transect has important, albeit uncertain, implications for stream Hg budgets. The FMeHg concentration at the furthest downgradient site, 0.08 ng/L, was about 30 % of the FMeHg concentration in the stream. Apparent attenuation of about 70 % of the FMeHg along the hyporheic transect might be construed as evidence that the hyporheic zone can serve as a sink for MeHg in streams. On the other hand, the maximum FMeHg concentration along the hyporheic transect, 0.69 ng/L, was nearly 280 % of the FMeHg concentration in the stream. Had the stream cut through the island near this site, it is possible that discharge from the hyporheic zone to the stream could have resulted in the hyporheic zone being a MeHg source, rather than a sink. The maximum increase in FTHg along the hyporheic transect was even greater, rising to 460 % of stream concentrations in 2002 and 540 % in 2003 before declining to concentrations close to those of the stream concentrations near the end of the hyporheic transect. Such spatial variability in FTHg and FMeHg concentrations and uncertainty regarding the magnitude of water fluxes through this hyporheic zone present challenges to the quantification of the effects of hyporheic-zone processes on Hg mobilization, transport, and fate in this mining-impacted stream system. Transient flow conditions would add another dimension of variability. It is not clear how greater streamflow and higher stage would affect flow and geochemical reactions (e.g., rates of Hg methylation or Hg complexation with DOC) in this hyporheic zone, what effects higher stage would have on island hyporheic zone flushing, and the degree to which such perturbations would affect Hg loading to the stream. Yet hyporheic zones are a fundamental component of watersheds because the spatial scales of the hyporheic zone frequently encompass much of the near-stream environment from headwater streams (Triska et al. 1990) to large rivers (Hinkle et al. 2001) and because the timescales for water and solute fluxes through the hyporheic zone can be similar to the timescales of advection in streams (Choi et al. 2000). At the scale of watersheds, the effects of hyporheic-zone processes on solute mobilization, transport, and fate can accumulate and become important to public policy issues (Hester and Gooseff 2010; Bencala et al. 2011; Grant and Marusic 2011). Watershed investigations that include assessments of hyporheic-zone processes have the potential to account for the effects of hyporheic-zone processes on stream chemistry and function.

This study has demonstrated that Hg can be both mobilized and attenuated in the hyporheic zone, that Hg gradients in the hyporheic zone can be steep, and that Hg biogeochemistry in the hyporheic zone likely is linked to DOC biogeochemistry. The strong correlations observed for FTHg and FMeHg with DOC in this hyporheic zone provide a framework for a more general understanding of total Hg and MeHg dynamics in the hyporheic zone. We anticipate that recent advances in elucidating DOC cycling in hyporheic zones (Sobczak and Findlay 2002; Findlay et al. 2003; Williams et al. 2010; Zarnetske et al. 2011, Trimmer et al. 2012) will facilitate understanding of Hg mobilization and fate in hyporheic zones and the watersheds in which they reside. Determination of organic carbon mass and character (e.g., bioavailability, aromaticity, hydrophobic/hydrophilic proportions) in dissolved, particulate, and sedimentary compartments could be particularly useful in future work (Chapelle et al. 2009; O’Donnell et al. 2010). The detailed observations and modeling analysis of Schelker et al. (2011), finding that concentrations of Hg, MeHg, and DOC follow different patterns at a catchment outlet during snowmelt-dominated flushing, emphasize the need for additional characterization of Hg reactions and transport pathways in the various hydraulically connected components of the stream-catchment system, including the hyporheic zone.

Acknowledgments

Marisa Cox was instrumental in designing and implementing the hyporheic sampling program and performed the major-ion analyses. This effort was funded by the USGS National Water-Quality Assessment Program, Toxic Substances Hydrology Program, and National Research Program. Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the US Government.

Copyright information

© Springer Science+Business Media Dordrecht (outside the USA) 2013