The fate of antimony in a major lowland river system, the Waikato River, New Zealand

Applied Geochemistry 24 (2009) 2283–2292

Contents lists available at ScienceDirect

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

The fate of antimony in a major lowland river system, the Waikato River, New Zealand Nathaniel Wilson a,*, Jenny Webster-Brown b a b

School of Geography, Geology and Environmental Science, University of Auckland, Auckland, New Zealand Department of Chemistry, University of Auckland, Auckland, New Zealand

a r t i c l e

i n f o

Article history: Available online 18 September 2009

a b s t r a c t Antimony is an element that is becoming of increasing concern as an environmental contaminant. Geothermal systems are a source of Sb into some fresh waters of New Zealand’s North Island. The purpose of this research was to determine the factors controlling the behaviour of geothermally-derived Sb in the large lowland Waikato River system. The Waikato River is New Zealand’s longest and most utilised river. Antimony in the system exhibited mainly conservative behaviour, and seasonally variable dilution was found to be the most important control on Sb concentrations. The most significant potential removal process was identified as adsorption of Sb onto suspended particulate material (SPM). The adsorption of Sb onto the SPM is enhanced at low (<5) pH conditions, and in the anoxic base of stratified lakes. There was evidence that the adsorption of Sb is mainly onto Fe oxides in SPM, and changes with changing Fe concentrations. Therefore, Sb adsorption was higher in winter (when Fe concentrations in SPM were higher) than in summer. In Lake Ohakuri, which was stratified during the late summer/early autumn of 2007, there was also potential for removal of Sb as Sb2S3 in the presence of sulfide formed in the anoxic layer. The behaviour of Sb was conservative through the estuary at the mouth of the river. Antimony was compared to As, a metalloid often assumed to exhibit behaviour similar to Sb in aquatic environments. It was found that while the removal processes affecting Sb will also affect As, the inverse did not necessarily apply. Arsenic will adsorb more readily to SPM than Sb and, while there was evidence for bioaccumulation of As by freshwater macrophytes, there was no such evidence for Sb. Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction Antimony is a group V metalloid that the US EPA recognises as a priority contaminant in freshwaters. The behaviour of Sb in aquatic systems is not well understood (Filella et al., 2002a,b), but Sb is assumed to exhibit similar behaviour to As, a metalloid that has been responsible for the poisoning of millions of people in SE Asia, and has received considerably more attention in aquatic environments (Hindmarsh, 2000). In New Zealand, Sb is present in both epithermal and mesothermal ore deposits, but New Zealand’s steep topography has meant that transport of mining derived Sb has been restricted to relatively minor waterways (Ashley et al., 2006). However, New Zealand’s geothermal fields, such as those in the Taupo Volcanic Zone (TVZ) are also a significant source of aqueous Sb (Ritchie, 1961). The chemistry of the Waikato River, which flows through the north-western section of the TVZ (Fig. 1), is directly affected by such geothermal inputs (Aggett and Aspell, 1980). The presence of As in the Waikato River has encouraged multidisciplinary research on the uptake of As by plants (Robinson et al., * Corresponding author. Fax: +64 9 373 7434. E-mail address: [email protected] (N. Wilson). 0883-2927/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.apgeochem.2009.09.016

2003) and fish (Robinson et al., 1995), dissolved As behaviour in stratified lakes (Aggett and Kriegman, 1988), and the overall fate of aqueous As in the Waikato River system (Webster-Brown et al., 2000). Of particular interest is the seasonal variation in dissolved As concentrations, which are typically lower in winter than in summer, even when factors such as dilution have been taken into account (McLaren and Kim, 1995). Various explanations have been offered for the seasonal changes, the most recent is that winter As concentration minima are caused by the dilution of geothermal fluids during the wetter winter months, in combination with increased adsorption of As onto Fe-oxide-rich SPM, which is more prevalent in winter (Webster-Brown and Lane, 2005a). In comparison, little has been published regarding Sb in the Waikato River. McLaren and Kim (1995) reported Sb concentrations in the lower Waikato were <2 lg/kg, but thorough investigations have not been published. The purpose of this research was to identify what processes affect Sb concentrations in the Waikato River, and to determine just how similar Sb is to As. The major sources of Sb into the Waikato River are geothermal, but the processes identified should be applicable to rivers contaminated by Sb from other sources world-wide, such as the mine-derived Sb in the Macleay River in Australia (Ashley et al., 2006) or the Orb River in France (Casiot et al., 2007).

2284

N. Wilson, J. Webster-Brown / Applied Geochemistry 24 (2009) 2283–2292

Fig. 1. The Waikato River and the location of sites sampled during this investigation. Map adapted from Beard (2008).

2. Regional setting The Waikato River (Fig. 1) is the longest river in New Zealand (425 km, including the headwaters), and more than 300,000 people live within its catchment. The river originates from Lake Taupo, which lies in the southern half of the Taupo Volcanic Zone, and the upper reaches of the river receive a number of geothermal inputs. Lake Taupo itself is the product of a series of rhyolitic eruptions, and pumice characterises the geology of the upper reaches of the Waikato River (Reid, 1983). Downstream from Ohakuri, tephra deposits become more common, and from Arapuni north, unconsolidated sediments dominate the geological terrain (McDowall, 1996). Lake Taupo is oligotrophic, but the Waikato River itself has been heavily modified and has a wide range of anthropogenic discharges (Chapman, 1996). At least 12 sewage treatment plants discharge into the river, along with a number of major industrial discharges, and intensive dairying occurs along the entire length of the river (Bilinska, 2005). The river is also a source of drinking water for more than 30 communities, including Hamilton, a city in excess of 120,000 people (McLaren and Kim, 1995). Eight major hydro-

electric dams, marked on Fig. 1, have been built along the Waikato River, creating a series of lakes, some of which can stratify over summer/autumn months. The river discharges into the Tasman Sea at the Port Waikato estuary. At the head of the river (at Taupo Gates, site WR 1 on Fig. 1), the base flow of the Waikato River is 48 m3/s, near the river mouth (at Mercer, site WR 14 on Fig. 1) the base flow is 185 m3/s (Rutherford, 2005). The flow at the Taupo Gates is controlled, as are the flows from the series of hydro-lakes along the river’s length. The flowcontrolled mean river flows are not necessarily responsive to catchment rainfall. Two geothermal power stations lie on the banks of the Waikato River (Wairakei power station and Oha¯ki power station). Geothermal inputs into the Waikato River, both natural and anthropogenic, contain elevated concentrations of elements such as B, Li and As, and some of the geothermal inputs contain elevated Sb concentrations as well. The most significant input of geothermal fluid into the Waikato River is the discharge of waste bore water from the Wairakei Power Station, which has historically contributed about 45% of the river’s Cl, Li, As and B (Timperley and Huser, 1996). Under current consent conditions 60,000 tonnes/day of untreated

N. Wilson, J. Webster-Brown / Applied Geochemistry 24 (2009) 2283–2292

waste bore fluids are discharged into the Waikato, contributing 30% of the dissolved As found downstream (Contact Energy, 2007; Webster and Nordstrom, 2003). The untreated waste bore discharge also contains elevated concentrations of Sb, and represents the largest single source of Sb into the river.

3. Materials and methods In 2006 and 2007, the Waikato River was sampled in February and August at 13 sites from Lake Taupo (WR 1) to Mercer Bridge (WR 14), as shown in Fig. 1. Lake Maraetai (WR 7) was not included in these surveys. Samples were also collected monthly from site WR 14 from February (late summer) 2006 until January 2008. Additional samples were taken from eight of the nine hydro-lakes situated along the upper half of the Waikato River in April 2008 (WR 1–2, WR 5–10). A series of samples were collected on an out-going tide from the Port Waikato estuary at the mouth of the Waikato River (WR 15) in December 2007. Filtered (0.45 lm) and unfiltered water samples, suspended particulate material (SPM) and bed sediment were collected. Flow data for site WR 14 was provided by Environment Waikato, the local regulatory authority; flow data were not available for all the sites sampled along the river. Dissolved O2, pH and conductivity were measured in situ using standard portable meters. Suspended particulate material samples were not collected from the hydro-lakes during April 2008, but macrophytes were sampled, and depth profiles were determined for three of the nine lakes (Lakes Ohakuri (WR 5), Maraetai (WR 7) and Arapuni (WR 9)).

3.1. Sampling procedures All aqueous samples were collected into polypropylene (PP) or high density polyethylene (HDPE) containers (50–1000 mL). Water samples were generally not preserved by chemical additives, except during the depth profile sampling of lakes, when samples collected for total sulfide determinations were preserved with 1 M zinc acetate, and samples collected for Fe and Mn analysis were preserved with HCl (pH < 2). Other samples, which were analysed for Sb, As and Li, were left unpreserved in order to ensure consistency with samples collected for analogous research on the behaviour of Sb in geothermal fluids and surface features (Wilson, 2008; Wilson et al., 2007). Furthermore, there is evidence to suggest that acid-preservation techniques used for metal cations are not necessarily appropriate for anions such as those  formed by Sb (e.g. H6 SbO6 ) or As (Wilson, 2008). Water samples collected at depth from lakes were collected using a purpose-built ultrahigh-molecular-weight HDPE sampler (described in Webster, 1994). In order to collect SPM, known volumes of water (500– 2000 mL) were filtered upon collection through 0.45 lm nitrocellulose filters and the filters retained. Bed sediment (oxic, <5 cm deep) was collected by dragging 250 mL plastic containers along the bottom of the river. Replicates were collected at each site, and no differences in sediment size fraction distributions were observed between replicates for any site sampled. Bed sediment samples were left to settle, the supernatant was removed, and then these samples were homogenised and placed in a 40 °C oven until dry (up to one week). Dry samples were then sieved to collect the <63 lm fraction. Macrophyte samples were collected by cutting submerged plants at their stem-base; no roots were collected or analysed. Macrophyte samples were washed in distilled water and then dried at 40 °C. Samples were then homogenised and sieved to <0.5 mm before digestion.

2285

3.2. Analytical methods Samples were analysed for Sb and As using hydride generation atomic absorption spectroscopy (HGAAS). Prior to analysis, 10 mL of concentrated reagent-grade HCl and 2.5 mL of 10% (w/v) KI were added to sample aliquots, diluted to 50 mL with MilliQ water, and left for at least 1 h to ensure complete reduction of all Sb species to SbIII (and all As to AsIII). A GBC HG3000 continuous-flow hydride generator was employed, using 0.6% (w/v) NaBH4 in 0.6% (w/v) NaOH to induce the reduction of Sb3+ to Sb3 (and As3+ to As3), with samples carried in concentrated reagent-grade HCl. The hydride generator was coupled to a GBC Avanta Atomic Absorbance Spectrometer, and integrated peak areas were read for absorbance at 217.6 nm. The practical detection limit was 0.1 lg/kgsolute Sb (equivalent to lg/L for the fresh waters sampled). Dissolved SbIII and AsIII were determined without the pre-reduction step, so as to exclude AsV and SbV from the analyses. A solution of 6% w/v citric acid was used as a carrier instead of HCl (Andreae et al., 1981). The practical detection limits were 0.3 lg/kg SbIII and 0.5 lg/ kg AsIII. Iron and Mn were measured using either flame atomic absorption spectroscopy (FAAS), or graphite furnace coupled with atomic absorption spectroscopy (GF-AAS), depending on the sample concentration. The detection limits for FAAS were 1 mg/kg for Fe and 200 lg/kg for Mn. For GF-AAS analyses, the detection limits were 5 lg/kg for Fe and 1 lg/kg for Mn. The analytical error was typically <5%, and for results approaching detection limits were, on occasion, slightly higher (up to 10%). In order to analyse sediment digests, which contained 2 M HCl, Mn standards for digest analyses were also made up in 2 M HCl, to account for the formation of MnCl complexes. Recoveries were 100 ± 5% for both elements. Total sulfide concentrations (as H2S) were measured using the Methylene Blue method, with a detection limit of 0.05 mg/L total sulfide (APHA, 1998). Lithium was determined by atomic emission spectroscopy (AES) at 670.7 nm. Measureable concentrations of Li were found in all samples and detection limits for Li were not quantified. Detection limits for AES are relative, and change depending on the concentration of the top standard. Bed sediment and SPM samples were digested using an open hot aqua-regia digestion process, using a method modified from Tighe et al. (2004). Digestion recoveries were tested by using the LKSD set of standard reference materials. For LKSD-2 and LKSD-3, recoveries of 100 ± 10% for Fe, Mn and As were achieved. Recoveries were not as good for Sb, ranging from 60% to 140%, and overall precision was 25%. However, there was a lack of practicable alternatives; for example, concentrations of Sb following microwave digestions were too dilute. Biotic samples were incinerated after being dosed with MgO and MgNO3, based on a method developed by Pahlavanpour et al. (1981) and modified by Almela et al. (2006). Recoveries for As from the tomato leaf reference standard (NIST-SRM 1573a) used were 96 ± 17%. However, for Sb recoveries were 185 ± 31%. Such recoveries indicated the digestions were contaminated, but the source of the contamination (i.e. the crucibles or one of the reagents) could not be determined. Nonetheless, this method gave an indication of the degree of biological uptake, even if the accuracy of the analyses were relatively poor. 3.3. Experimental methods A sequential extraction procedure was used to determine the binding preference of Sb in SPM collected from site WR 15. The method chosen used a 5-step process, developed for As by Wenzel et al. (2001) and first applied to Sb by Müller et al. (2007). Because other studies have reported that organically-bound Sb can account for more than 10% of the total Sb in sediment (Brannon and Patrick,

2286

N. Wilson, J. Webster-Brown / Applied Geochemistry 24 (2009) 2283–2292

1985), a single step extraction using EDTA was also employed (Lintschinger et al., 1998). Adsorption edge experiments were performed using a method based on that described in Bibby and Webster-Brown (2006). Freeze-dried SPM was resuspended in 0.01 M NaNO3 at a concentration of 100 mg/L. The solution was left to re-hydrate at 4 °C for 48 h, and then left to equilibrate to 25 °C (typically 4 h). The pH of the solution was raised to 10 using NaOH, and constantly mixed using an acid-washed magnetic stirrer, before being spiked with Sb (either SbIII or SbV). Once pH had stabilised, an aliquot (15 mL) was extracted and then the pH of the bulk solution was lowered by about 0.5 pH units and the process repeated. The end pH was typically 3. The extracted aliquots were placed on a mixing wheel in a temperature controlled room (25 °C) and left to rotate in darkness. After 24 h, the samples were removed from the wheel. Half of each aliquot was filtered (0.45 lm) and the pH was re-measured in the remaining solution to determine the final pH of each sample. Control experiments, one without SPM, and one without Sb added, showed that no measurable adsorption onto or release of Sb from the PP vial walls occurred, nor was there any evidence for removal of Sb through precipitation of Sb solid phases.

4. Results 4.1. Aqueous Sb concentrations The Waikato River was oxic and circumneutral at all sites (Table 1). Dissolved Sb (and As) concentrations were higher in summer than in winter, as shown in Fig. 2, and were at maxima immediately downstream of the discharge from Wairakei Power station (WR 2). This indicates, as has been observed for As concentrations (Webster-Brown and Lane, 2005b), that the station’s discharge is the most significant source of Sb to the river. Both Sb and As showed elevated concentrations from site WR 2 downstream in winter 2007, which were unlikely to be related to low flows because river flow at WR 2 during that time (270 m3/s) was above average (200 m3/s) and higher than those measured in the summer of the same year (230 m3/s) (Beard, 2008). Dissolved Sb typically constituted >90% of the total measured concentrations, and concentrations of SbIII were below detection limit (0.3 lg/kg) at all sites. At WR 15, for which good flow data were available, Sb concentrations were converted into Sb fluxes, multiplying the concentration data by the mean daily flow collected at the site. The results, shown in Fig. 3, indicate Sb fluxes were relatively stable, apart from temporarily elevated loads in May 2006 and June/July 2007.

Fig. 2. Dissolved concentrations of (a) Sb and (b) As in the Waikato River.

Fig. 3. Sb flux with time at site WR 15. The shaded area indicates one standard deviation from the mean (n = 24).

Table 1 Mean results for aqueous samples collected from the Waikato River 2006–2007. Site WR WR WR WR WR WR WR WR WR WR WR WR WR

1 2 3 4a 5 6 8a 9 10a 11 12a 13 14

Temp (°C)

DOa (mg/kg)

pH

Sbdiss (lg/kg)

Sbtot (lg/kg)

Asdiss (lg/kg)

Astot (lg/kg)

Li (lg/kg)

16.3 18.7 18.2 16.1 17.7 17.7 15.6 17.4 15.4 16.6 15.0 16.1 15.2

9.6 11.6 9.8 9.4 10.1 9.8 9.4 11.0 10.0 9.5 10.0 9.4 8.8

7.07 7.57 6.98 6.59 6.96 7.11 6.78 7.03 6.86 7.01 6.83 6.92 6.86

0.3 0.9 0.9 0.8 0.9 0.7 0.6 0.7 0.6 0.7 0.5 0.6 0.4

0.3 0.9 0.9 0.9 0.9 0.7 0.6 0.7 0.7 0.7 0.6 0.6 0.4

9.2 30.3 25.8 22.5 24.1 21.0 18.7 19.6 18.1 19.3 17.6 15.1 9.7

9.8 31.0 29.8 29.3 26.8 22.3 21.5 20.8 20.1 20.2 19.4 16.1 11.9

46.4 129.4 135.4 119.9 129.5 113.6 98.3b 107.5 94.0 104.7 100.5 92.4 64.9

n = 4 Unless noted otherwise. a n = 3. b n = 2.

N. Wilson, J. Webster-Brown / Applied Geochemistry 24 (2009) 2283–2292

accumulating As, but were not accumulating Sb, relative to the sediments.

Table 2 Concentrations of Sb and As in bed sediment of the Waikato River. Site WR WR WR WR WR WR WR WR WR WR WR WR WR WR WR WR

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

2287

Distance (km)

Sb (mg/kg)

As (mg/kg)

Fe (wt.%)

Mn (wt.%)

1 13 40 49 79 100 105 130 155 180 205 225 250 288 300 325

1.1 0.6 2.6 4.9 8.3 6.7 6.7 2.4 3.3 5.6 1.4 1.4 1.2 3.3 0.8 1.9

11.3 8.9 64.3 334.8 383.1 263.5 225.3 208.5 194.2 78.9 207.8 52.6 42.3 71.6 105.6 48.9

2.71 0.91 1.94 1.60 4.17 3.35 2.87 2.54 2.59 1.78 2.60 2.82 3.07 2.66 2.67 2.68

0.03 0.02 0.05 0.06 0.15 0.14 0.17 0.25 0.27 0.06 0.11 0.08 0.18 0.11 0.19 0.11

4.2. Antimony in SPM and bed sediment Antimony concentrations in the SPM collected from the Waikato River in summer were similar to those measured in bed sediment, particularly in the upper catchment (Table 2). However, winter Sb-SPM concentrations were more elevated, compared to summer samples, in samples collected along the first 130 km of the river (Fig. 4).

4.4. Depth profiles The three deepest lakes along the Waikato River, Lakes Ohakuri, Maraetai and Arapuni, were sampled at sites as close as possible to maximum lake depth in April 2007. The three lakes were initially sampled at three depths, (the surface, 20 m and 40 m depth), with additional DO readings at depths between 20 m and 40 m if there was any evidence for chemical stratification. Chemical stratification was only apparent in Lake Ohakuri, as shown in Fig. 5a. Lake temperatures were warmer at the surface than they were at 40 m, but the difference was only 2 °C for Lake Ohakuri and just 0.5 °C within Lakes Maraetai and Arapuni. Dissolved Sb concentrations were compared with the depthprofile concentrations of other elements (As, Fe and Mn), as shown in Fig. 5b. It is clear that the behaviour of Sb is different to these elements: Fe and Mn appear to be being released from Fe- and Mn-rich sediments in the anoxic base of Lake Ohakuri, as was also observed by Aggett and Roberts (1986). Arsenic concentrations increased slightly, but rather than being released from sediments into O2-depleted lake waters, Sb appears to be removed from solution instead. Detectable concentrations of aqueous H2S (24 lg/L) were also measured at the bottom of Lake Ohakuri, but were not found in any other samples. 4.5. Adsorption experiments

4.3. Macrophyte uptake A sample of the most common species of macrophyte at each site was collected from the eight lakes surveyed. From Lakes Taupo to Ohakuri (WR 1 and 2, WR 5), Egeria densa dominated. Further downstream (WR 6–10), Ceratophyllum demersum was more prevalent. Macrophyte uptake of As in the Waikato is well established (e.g. Robinson et al., 2003; Gibbs and Costley, 2000), and the results for As concentrations in the macrophyte samples collected as part of this study (listed in Table 3) were in good agreement with previous studies. Macrophyte concentrations of Sb were 2–3 orders of magnitude lower than those measured for As. Enrichment factors (EFs) were calculated as

EF ¼

½Mmacrophyte ½Mbedsediment

where M is either Sb or As; an EF > 1 indicates the enrichment of the metalloid in macrophytes relative to the sediment they grow in. The results of these calculations (Table 3) indicate macrophytes were

Fig. 4. Antimony concentrations in SPM and bed sediment collected from the Waikato River.

One aspect of Sb behaviour for which there are little data currently available, is the adsorption of dissolved Sb onto natural SPM. Studies of Sb adsorption onto bed sediments (Li et al., 1984) or flood-plain soils (Tighe et al., 2005) indicate that Sb will bind readily to available Fe-oxide surfaces. However, there is no information for dissolved Sb interaction with the solid phase it is most likely to encounter in the aquatic environment, SPM. In order to address this imbalance, Sb adsorption experiments were conducted using WR 15 SPM as the adsorbing substrate. Samples of WR 15 SPM were collected in August 2003 (winter) and in November 2006 (spring), by filtering 20–40 L of collected water through 0.45 lm filters, then air-drying and freeze-drying the sediment. The general characteristics of the 2006–2008 WR 15 SPM and the winter 2003 SPM are listed in Table 4. The results show that the 2003 winter SPM contained concentrations of Sb and As comparable to those measured in 2006–08, and was therefore valid for use in the experiments. Because there was more winter SPM collected than spring SPM, the winter SPM was also selected for a sequential extraction, in order to characterise how naturally adsorbed Sb was bound to the material. The results of the adsorption experiments are shown in Fig. 5. The adsorption of SbIII onto winter SPM was much higher than SbV adsorption onto similar SPM (Fig. 6a). Results for SbIII adsorption were consistent with the results for SbIII adsorption onto metal-oxyhydroxides reported by Thanabalasingam and Pickering (1990). For SbV there were no significant differences between results for winter SPM and spring SPM, nor was there evidence for any adsorption of SbV above pH 5. Consideration was given to whether Sb adsorption to SPM might be a slow process, and that 24 h may not be enough time for the system to reach equilibrium. This was especially pertinent given the lack of available Sb adsorption data, although McComb et al. (2007) have reported Sb adsorption reaching a maximum in less than 2 h (using amorphous Fe oxide at pH 3). In order to measure the rate of adsorption, two 10 day experiments were set up. In the first experiment, the pH of samples was fixed at 3, where maximum adsorption had been shown to occur. In the second,

2288

N. Wilson, J. Webster-Brown / Applied Geochemistry 24 (2009) 2283–2292

Table 3 Concentrations of Sb and As in lake bed sediments and macrophytes (dry weight). Lake

Site

Taupo Aratiatia Ohakuri Whakamaru Maraetai Waipapa Arapuni Karapiro

WR WR WR WR WR WR WR WR

Antimony

1 2 5 6 7 8 9 10

Arsenic

Macrophyte (mg/kg)

Sediment (mg/kg)

EF

Macrophyte (mg/kg)

Sediment (mg/kg)

EF

0.1 0.3 0.3 0.7 0.5 0.1 0.2 0.1

1.1 0.6 8.3 6.7 6.7 2.4 3.3 5.6

0.12 ± 0.9 0.43 ± 0.32 0.03 ± 0.02 0.10 ± 0.08 0.08 ± 0.06 0.06 ± 0.05 0.07 ± 0.05 0.02 ± 0.02

22 52 105 210 520 478 455 220

11 8.9 380 260 230 210 190 79

2.0 ± 0.6 5.9 ± 1.8 0.27 ± 0.08 0.80 ± 0.24 2.3 ± 0.7 2.3 ± 0.7 2.3 ± 0.7 2.7 ± 0.8

EF = enrichment factor.

Fig. 5. Results from depth profile sampling: (a) Oxygen profiles for Lakes Ohakuri, Maretai and Arapuni, and (b) Dissolved metal and metalloid concentration profiles for Lake Ohakuri.

Table 4 Results of Sb sequential extraction on SPM collected from WR 15 (following Müller et al. (2007)). Bound fraction

Total Sb bound (%)

Non-specific sorption onto outer sphere complexes Specific sorption onto inner sphere complexes Organic/sulfide bound Amorphous Fe/Mn oxide bound Crystalline oxide Fe/Mn bound Residue

2.0 ± 1 0 7±1 46 ± 3 17 ± 1 28 ± 4

the pH of samples was fixed at 7, where no adsorption had been observed, and which represented conditions more typical of the Waikato River. The results of the timed experiment, shown in Fig. 6b, indicate that Sb adsorption onto Waikato River SPM reaches equilibrium at pH 3 in about 5 days (120 h), much slower than for adsorption onto synthetic Fe-oxide surfaces (McComb et al., 2007). At pH 7, very minor adsorption onto natural SPM occurred only after 7 days (168 h). While there is evidence therefore to suggest such adsorp-

Fig. 6. Adsorption edges for Sb adsorption onto Waikato River SPM from site WR 15: (a) SbIII and SbV vs pH (b) SbV adsorption over time.

tion experiments require longer than the 24 h period used in this research to achieve equilibrium, the overall findings are nonetheless robust; Sb adsorption onto Waikato River SPM is much more significant at low pH than in neutral conditions. 4.6. Sequential extraction The results of sequential extraction and single organic/sulfide extraction of winter SPM are presented as Table 5. The Sb bound to the easily exchangeable fraction (specific and non-specific sorption) was minor, accounting for just 2% of the total Sb bound, and only a further 7% was bound to organic or sulfide binding sites. Most of the Sb (63%) was bound to oxide phases or present as residual Sb (28%). The latter would include Sb bound to silicate phases, and any Sb-bearing minerals present in the SPM. 4.7. Port Waikato estuary The final aspect of Sb behaviour in the Waikato River to be investigated was Sb behaviour in the estuary at the mouth of the

N. Wilson, J. Webster-Brown / Applied Geochemistry 24 (2009) 2283–2292 Table 5 Typical characteristics of SPM collected from WR 15 in 2006 and 2007(n = 18). The characteristics of specific SPM collected for adsorption experiments in August 2003 (winter) and November 2006 (spring) are also included.

Mean Maximum Minimum Spring SPM Winter SPM a

Sb (mg/kg)

As (mg/kg)

Fe (wt.%)

Mn (wt.%)

1.7 ± 0.8 2.8 1.0 1.3 1.0

180 ± 30 350 64 110 190

3.4 ± 0.5 4.9 1.35 4.40 4.20–4.51a

0.23 ± 0.04 0.38 0.13 0.13 0.16–0.29a

Concentrations for SPM samples collected in August 2006 and 2007 (n = 2).

river. There are few published studies for the behaviour of Sb in estuaries, and the consensus appears to be that the element’s behaviour depends upon the nature of the estuary (Filella et al., 2002a). Sampling the estuary (WR 16) at different stages of the tidal cycle showed that Sb concentrations, for the most part, plotted on a linear mixing curve with respect to salinity (Fig. 7). The point lying above the linear mixing line in Fig. 7 may indicate a release of Sb from sediments, but is more likely to be the result of sample contamination. Analytical errors for this sample were higher for those measured at other salinities. Overall, in Port Waikato dissolved Sb appears to be largely unaffected by changing chemical conditions and is not removed from solution during the mixing of freshwater with seawater. Instead, Sb behaves conservatively. In a study by Webster-Brown and Webster (2000), it was shown that, overall, As also exhibited conservative behaviour in this estuary. Bed sediment and SPM data was also collected for Port Waikato, to see if the sediment might be enriched in Sb, despite the aqueous results. At 0.8 mg/kg, Sb concentration in WR 16 SPM was lower than recorded at WR 15 (Table 2). The Sb concentration of bed sediment was higher at WR 16 (1.9 mg/kg) than at WR 15 (0.8 mg/kg), but Sb concentrations in bed sediment were very variable, and further upstream at WR 14 were higher (3.3 mg/kg). There is therefore no case for Sb enrichment of estuarine sediment.

5. Discussion 5.1. Removal of Sb from the Waikato River A mixture of point (e.g. the Wairakei Power station) and non-point geothermal sources (e.g. groundwater) are the major contributors of Sb into the Waikato River, and as a result, Sb concentrations are higher in the upper reaches of the river (within the TVZ), than

Fig. 7. Dissolved and total concentrations of Sb with respect to salinity in the Port Waikato estuary at site WR 16.

2289

downstream where tributaries are not influenced by geothermal activity. Changes in dissolved Sb concentrations with distance in the Waikato River are similar to those observed for Sb concentrations in Waikato River SPM and concentrations of Sb in the river’s bed sediment, and enrichment factors in lake macrophytes did not indicate any bioaccumulation of Sb into biota. Overall, the evidence suggests that Sb is, for the most part, behaving conservatively in the river. Positive correlations between Sb concentrations and Li concentrations provide further evidence that the behaviour of Sb in the Waikato River is mainly conservative (Fig. 8). However, the Li:Sb ratios for the 2007 winter data were slightly different to the ratio observed for the rest of the data. This difference in ratios indicates the discharge from the Wairakei Power Station, the probable source of elevated Sb concentrations in August 2007, may have contained relatively more Li at this time. While seasonal differences in SPM Fe concentrations may affect Sb behaviour, seasonal differences between Sb concentrations with respect to Li concentrations are relatively small. Dilution is therefore the most significant cause of decreases of Sb concentrations in the Waikato River (especially downstream of the hydro-lakes). While Sb behaviour appears to be generally conservative, there is evidence that adsorption processes and the chemical stratification of lakes are also viable mechanisms for the removal of Sb from the Waikato River. 5.2. Dilution The decrease in Sb concentrations measured in the Waikato River with distance is almost certainly related to dilution, particularly in the lower reaches, where the Waikato River is joined first by the Waipa River and then by the sediment-laden Mangawara Stream (Fig. 1). Flows in the Waikato River upstream of Cambridge (approximately 165 km downstream from Lake Taupo) are relatively constant all year round because of the series of hydro-electric power station dams and Taupo Gates control dam. Further downstream, the influence of tributaries such as the Waipa River varies with catchment rainfall. In summer, tributary flows are lower, and therefore the decline in Sb concentrations with distance is less. In the wetter winters, swollen tributaries lead to more significant dilution of Sb (and As) concentrations. The most likely reason for the elevated concentration of Sb in winter 2007 (Fig. 2a) was that the Wairakei Power station was discharging more spent geothermal fluids into the Waikato River than normal, because rainfall in August 2007 was relatively high and the river was swollen along its entire length. Arsenic concentrations were similarly elevated (Fig. 2b). Whether an increased discharge actually occurred could not be confirmed.

Fig. 8. Correlations between dissolved Li concentrations and dissolved Sb concentrations in the Waikato River.

2290

N. Wilson, J. Webster-Brown / Applied Geochemistry 24 (2009) 2283–2292

The relatively constant Sb flux measured at site WR 15 further emphasises this (Fig. 3). Apart from the apparent spikes during May 2006 and June and July 2007, the consistency of the flux determinations suggest that the apparent differences in concentrations between summer and winter (Fig. 2) have more to do with the amount of rainfall and river flow than any active removal processes. Rather than being aberrations however, the spikes detected in Fig. 3 do suggest that there are processes upstream that affect the amount of Sb in the Waikato River which may be unrelated to changes in river flow. The upstream release of Sb adsorbed onto SPM or temporarily precipitated may offer possible explanations.

5.3. Adsorption of Sb onto SPM McComb et al. (2007) investigated the adsorption of SbV onto amorphous Fe oxide and found that maximum adsorption occurred at pH 6 3, and it is likely that the adsorption observed for WR 15 SPM was also onto amorphous Fe hydroxides. The adsorption profiles presented in Fig. 6 may therefore explain, in part, why aqueous Sb generally behaves conservatively in natural waters, because conditions in freshwaters such as the Waikato River (oxic and pH 6–8) simply do not favour adsorption of SbV onto Fe or Mn oxide surfaces in the SPM. In contrast, any dissolved SbIII that reaches the Waikato River could adsorb to SPM and eventually be removed from the water column, but SbIII is typically a minor constituent of the total Sb in oxic waters (Filella 2002a). Furthermore, Fe- and Mn-oxyhydroxides have been shown to oxidise SbIII species to SbV species over time (Belzile et al., 2001), so such a removal process is unlikely to ever be significant in naturally oxic waters (Filella et al., 2002a). Above pH 2.7, antimonate dissociates from H7SbO6 to form neg atively charged H6 SbO6 (Jain and Banerjee, 1961), so adsorption to positively charged surfaces might be expected. However, Leuz et al.  (2006) argue that H6 SbO6 may form ion pairs with alkali metals, and this could explain why adsorption begins to decrease at pH 5, while the surface charge on amorphous Fe oxide does not change from negative to positive until pH 7–8 (Dzombak and Morel, 1990). For example, adsorption of SbV to goethite decreases above pH 6–7 (depending on ionic strength), and SbV adsorption to amorphous Fe oxide decreases with pH above pH 3 (Leuz et al., 2006; McComb et al., 2007). Therefore, even weak ion pairing in more complex systems, such as freshwaters, may be sufficient to prevent SbV adsorption onto SPM. The results of this study do not entirely agree with those found for the adsorption of dissolved Sb onto humic acids, Fe-oxyhydroxides, or mine-derived flood-plain soils presented by Tighe et al. (2005). However, Tighe et al. (2005) used much higher concentrations of Sb (28–11,300 lg/kg Sb) and substrate (400–50,000 mg/ kg) than are ever likely to be present in a system like the Waikato River. A pattern of decreasing adsorption with decreasing SbV concentrations was evident (Tighe et al., 2005). Humic acid concentrations were not considered, but Buschmann and Sigg (2004) have showed that dissolved SbIII will adsorb to humic acids at pH > 6 in laboratory conditions. However, it has been shown elsewhere that binding of dissolved SbIII to humic acids is only significant if SbIII concentrations are greater than 10 lM, the equivalent of 1.2 mg/kg Sb (Oelkers et al., 1998), orders of magnitude higher than the observed Sb concentrations in freshwaters studied in this research. The sequential extraction experiment showed that most of the Sb in the winter SPM is bound to Fe and/or Mn oxide sites, similar to results found for estuarine and anaerobic bed sediments in the United States (Brannon and Patrick, 1985; Crecelius et al., 1975), and river sediments in Europe (Leleyter and Probst, 1999). Therefore it is likely that the adsorption edges developed for Waikato

River SPM collected at WR 15 should be able to be applied to SPM from other river systems. 5.3.1. Adsorption of Sb onto Waikato River SPM To establish whether a greater degree of Sb association with SPM occurs when there is a higher Fe-oxide content, kd values for Sb in the Waikato River, calculated as dissolved [Sb]/SPM [Sb], were compared to SPM Fe concentrations (Fig. 9a). With the exception of two summer results (for sites WR 1 and WR 2), kd values for Sb in the Waikato River do appear to be positively correlated with SPM Fe concentrations. A similar correlation was also observed for As (Fig. 9b) and these results for As are in agreement with those of Webster-Brown and Lane (2005a). Therefore, while the degree of Sb adsorption onto Waikato River SPM is typically close to an order of magnitude less than for As, the SPM Fe-oxide content appears to be an important factor influencing adsorption of both Sb and As onto Waikato River SPM. Preferential adsorption of SbIII could explain the apparent removal of dissolved Sb from the basal O2-depleted layer of Lake Ohakuri when it was stratified. Fig. 6a shows that SbIII adsorbs to Waikato SPM even at neutral pH, and an O2-depleted environment may mean that SbIII species may be stable. Modelling of the system using PHREEQCi (Parkhurst and Appelo, 1999) and the MINTEQ v.4 database (updated with Sb2S3 data from Zotov et al., 2003; see Wilson et al., 2007) predicts that dissolved Sb at the bottom of Lake Ohakuri should be predominantly present as SbIII, under the depleted O2 conditions. 5.4. Precipitation of SbIII minerals in Lake Ohakuri The precipitation of Sb as stibnite (Sb2S3) could be an alternative Sb removal mechanism occurring at the bottom of stratified Lake Ohakuri. The concentration of H2S detected (24 lg/kg) is sufficient to at least approach Sb2S3 saturation conditions (log10SI = 0.24, PHREEQCi model (Parkhurst and Appelo, 1999;

Fig. 9. Comparisons between kd values for (a) Sb and (b) As with SPM Fe concentrations in the Waikato River.

N. Wilson, J. Webster-Brown / Applied Geochemistry 24 (2009) 2283–2292

Fig. 10. Correlations between dissolved Sb concentrations and dissolved As concentrations in the Waikato River.

Sb2S3 data from Zotov et al., 2003). Whether such precipitation was actually occurring could not be tested, as the amount of Sb2S3 produced would be too small (ng/g or pg/g total sediment) to observe. The removal of dissolved Sb from the basal layer of Lake Ohakuri, whether by adsorption or precipitation, may also explain the anomalous Sb flux peaks shown in Fig. 3. Following the re-establishment of an O2-rich environment at the bottom of Lake Ohakuri, during lake turnover, the adsorbed (and/or precipitated) Sb may be mixed into surface waters and the downstream transport of redissolved Sb and Sb-enriched SPM from Lake Ohakuri could then cause the observed flux peaks. It would require unusual seasonal conditions (a very warm and relatively windless autumn) if lake turnover did not occur until in winter (June/July 2008), but such a combination is not impossible. Unfortunately, it was not known when Lake Ohakuri actually overturned in 2007 or 2008. 5.5. Comparisons with arsenic behaviour The dissolved As and Sb found in the Waikato River are both predominantly derived from geothermal sources, and therefore exhibit similar patterns of distribution in the upper reaches of the system. However, correlations between the two metalloids suggest that seasonal differences have a much more significant impact on As concentrations than on Sb concentrations. In winter, there is a positive correlation between dissolved Sb and dissolved As in the Waikato River, as shown in Fig. 10, further evidence that increased Fe concentrations in winter SPM are a significant influence upon the behaviour of both metalloids. In summer though, there is little evidence of a correlation between Sb and As concentrations. This suggests that some process is occurring in summer that affects one of the metalloids, but not the other. The positive correlations between Sb and Li, regardless of season (Fig. 8), suggest that the behaviour of Sb is mainly conservative, whereas some further process must be influencing As behaviour in summer that does not similarly affect Sb. Interactions between As and algae such as Asterionella Formosa, which bloom in summer and make up a substantial component of summer SPM in the Waikato River (Hegan, 2005), may not similarly occur for Sb. The greater affinity for As for Fe oxide in SPM, and the accumulation of As in macrophytes and in algae are all factors that can, in part, explain the observed differences between the two metalloids. 6. Conclusions From an environmental management perspective, detectable levels of dissolved Sb in the Waikato River effectively classify the river as being Sb-impacted. However, Sb concentrations should not be a cause for concern because Sb concentrations in the river

2291

are below even the most stringent of water quality guidelines and present no known risk to ecosystems, human health, or agriculture. When compared to As, which reaches concentrations in excess of 40 lg/kg, Sb is only a minor geothermal contaminant in the Waikato River, and most of the Sb transported into the Waikato River is eventually transported to the Tasman Sea. Most of the Sb in the Waikato River reaches the Tasman Sea because aqueous Sb generally behaves conservatively. There is a strong correlation between Sb and Li, there is no evidence for any bioaccumulation of Sb by aquatic macrophytes, and the Sb flux measured monthly at WR 15 was relatively stable. However, the potential exists for minor Sb adsorption onto Fe-oxide surfaces in SPM or bed sediment and for precipitation of Sb2S3 in anoxic lake waters. Adsorption of Sb onto Fe oxides in Waikato River SPM is enhanced by low pH conditions, or reducing conditions during which SbIII is the dominant dissolved species. While the pH of the Waikato River is rarely lower than 6, anoxic (reducing) conditions may develop when hydro-lakes such as Lake Ohakuri chemically stratify over the summer months. Occasional Sb flux peaks may be observed at site WR 15, and these are possibly caused by lake turnover events, releasing Sb into the water column that had been temporarily removed from solution by adsorption and precipitation processes.

Acknowledgements This study was funded by a Bright Futures Enterprise Scholarship, jointly supported by Mighty River Power and the Tertiary Education Committee of New Zealand, and by the University of Auckland. We would also like to thank Sara Aprea, Peter Crossley, Joe Davies, Timothy Dee and Xueqiang Lu for their assistance in the field, and Dr. Rebecca Bibby and Vincent Lane for their assistance with analyses in the laboratory. Our thanks also to Dr. Kevin Brown, for guidance and insight throughout this project. We are also grateful to Environment Waikato for supplying flow data.

References Aggett, J., Aspell, A., 1980. Arsenic from geothermal sources in the Waikato catchment. New Zeal. J. Sci. 23, 77–82. Aggett, J., Kriegman, M.R., 1988. The extent of formation of arsenic(III) in sediment interstitial waters and its release to hypolimnetic waters in Lake Ohakuri. Water Res. 22, 407–411. Aggett, J., Roberts, L.S., 1986. Insight into the mechanism of accumulation of arsenate and phosphate in hydro lake sediments by measuring the rate of dissolution with ethylenediaminetetraacetic acid. Environ. Sci. Technol. 20, 183–186. Almela, C., Jesus Clemente, M., Velez, D., Montoro, R., 2006. Total arsenic, inorganic arsenic, lead and cadmium contents in edible seaweed sold in Spain. Food Chem. Toxicol. 44, 1901–1908. Andreae, M.O., Asmode, J.F., Foster, P., Vantdack, L., 1981. Determination of antimony(III), antimony(V), and methylantimony species in natural-waters by atomic-absorption spectrometry with hydride generation. Anal. Chem. 53, 1766–1771. APHA, 1998. Standard Methods for the Examination of Water and Wastewater, 20th ed. APHA, Washington, DC. Ashley, P.M., Craw, D., Tighe, M.K., Wilson, N.J., 2006. Magnitudes, spatial scales and processes of environmental antimony mobility from orogenic gold–antimony mineral deposits, Australasia. Environ. Geol. 51, 499–507. Beard, S., 2008. Waikato River Water Quality Monitoring Programme: Data Report 2007. Environment Waikato Technical Report 2008/20. Hamilton. Belzile, N., Chen, Y.W., Wang, Z.J., 2001. Oxidation of antimony(III) by amorphous iron and manganese oxyhydroxides. Chem. Geol. 174, 379–387. Bibby, R.L., Webster-Brown, J.G., 2006. Trace metal adsorption onto urban stream suspended particulate matter (Auckland Region, New Zealand). Appl. Geochem. 21, 1135–1151. Bilinska, K., 2005. Thermal stratification in the Waikato Hydro Lakes. Centre for Water Research, Unpublished Honours Dissertation, University of Western Australia, Perth. Brannon, J.M., Patrick, W.H., 1985. Fixation and mobilization of antimony in sediments. Environ. Pollut. Ser. B Chem. Phys. 9, 107–126. Buschmann, J., Sigg, L., 2004. Antimony(III) binding to humic substances: influence of pH and type of humic acid. Environ. Sci. Technol. 38, 4535–4541.

2292

N. Wilson, J. Webster-Brown / Applied Geochemistry 24 (2009) 2283–2292

Casiot, C., Ujevic, M., Munoz, M., Seidel, J.L., Elbaz-Poulichet, F., 2007. Antimony and arsenic mobility in a creek draining an antimony mine abandoned 85 years ago (Upper Orb Basin, France). Appl. Geochem. 22, 788–798. Chapman, M.A., 1996. Human impacts on the Waikato River System. New Zeal. Geo. J. 40, 85–99. Contact Energy, 2007. Contact Energy Sustainability Report 2007. Contact Energy Limited. Crecelius, E.A., Bothner, M.H., Carpenter, R., 1975. Geochemistries of arsenic, antimony, mercury, and related elements in sediments of Puget Sound. Environ. Sci. Technol. 9, 325–333. Dzombak, D.A., Morel, F.M.M., 1990. Surface Complexation Modelling. New York, Wiley. Filella, M., Belzile, N., Chen, Y.W., 2002a. Antimony in the environment: a review focused on natural waters I. Occurrence. Earth-Sci. Rev. 57, 125–176. Filella, M., Belzile, N., Chen, Y.W., 2002b. Antimony in the environment: a review focused on natural waters II. Relevant solution chemistry. Earth-Sci. Rev. 59, 265–285. Gibbs, M., Costley, K., 2000. Heavy Metals in Waikato Hydro Power Station WeedDump Sites. NIWA, Report Number MRP01231. Hegan, A.F., 2005. A preliminary investigation into the role of diatoms in the binding of copper and arsenic onto SPM. Unpublished M.Sc. Thesis, Univ. Auckland, Auckland. Hindmarsh, J.T., 2000. Arsenic, its clinical and environmental significance. J. Trace Elem. Exp. Med. 13, 165–172. Jain, D.V.S., Banerjee, A.K., 1961. On the structure of antimonates. J. Inorg. Nucl. Chem. 19, 177–179. Leleyter, L., Probst, J.-L., 1999. A new sequential extraction procedure for the speciation of particulate trace elements in river sediments. Int. J. Environ. Anal. Chem. 73, 109–128. Leuz, A.K., Monch, H., Johnson, C.A., 2006. Sorption of Sb(III) and Sb(V) to goethite: influence on Sb(III) oxidation and mobilization. Environ. Sci. Technol. 40, 7277– 7282. Li, Y.-H., Burkhardt, L., Buchholtz, M., O’Hara, P., Santschi, P.H., 1984. Partition of radiotracers between suspended particles and seawater. Geochim. Cosmochim. Acta 48, 2011–2019. Lintschinger, J., Michalke, B., Schulte-Hostede, S., Schramel, P., 1998. Studies on speciation of antimony in soil contaminated by industrial activity. J. Environ. Anal. Chem. 72, 11–25. McComb, K.A., Craw, D., McQuillan, A.J., 2007. ATR-IR spectroscopic study of antimonate adsorption to iron oxide. Langmuir 23, 12125–12130. McDowall, R.M., 1996. Volcanism and freshwater fish biogeography in the northeastern North Island of New Zealand. J. Biogeogr. 23, 139–148. McLaren, S.J., Kim, N.D., 1995. Evidence for a seasonal fluctuation of arsenic in NewZealand’s longest river and the effect of treatment on concentrations in drinking-water. Environ. Pollut. 90, 67–73. Müller, K., Daus, B., Morgenstern, P., Wennrich, R., 2007. Mobilization of antimony and arsenic in soil and sediment samples – evaluation of different leaching procedures. Water Air Soil Pollut. 183, 427–436. Oelkers, E.H., Sherman, D.M., Ragnarsdottir, K.V., Collins, C., 1998. An EXAFS spectroscopic study of aqueous antimony(III)-chloride complexation at temperatures from 25 to 250 °C. Econ. Geol. 151, 21–27. Pahlavanpour, B., Thompson, M., Thorne, L., 1981. Simultaneous determination of trace amounts of arsenic, antimony and bismuth in herbage by hydride generation and inductively coupled plasma atomic-emission spectrometry. Analyst 106, 467–471. Parkhurst, D., 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. US Geol. Surv. Water Resour. Invest. Rep. 99-4259. Reid, F., 1983. Origin of the rhyolitic rocks of the Taupo volcanic zone, New-Zealand. J. Volcanol. Geoth. Res. 15, 315–338. Ritchie, J.A., 1961. Arsenic and antimony in some New Zealand thermal waters. New Zeal. J. Sci. 4, 218–229. Robinson, B.H., Brooks, R.R., Outred, H.A., Kirkman, J.H., 1995. Mercury and arsenic in trout from the Taupo volcanic zone and Waikato River, North-Island, NewZealand. Chem. Spec. Bioavailab. 7, 27–32. Robinson, B., Duwig, C., Bolan, N., Kannathasan, M., Saravanan, A., 2003. Uptake of arsenic by New Zealand watercress (Lepidium Sativum). Sci. Total Environ. 301, 67–73. Rutherford, K., 2005. The Effect on Waikato Water Quality of Changes to River Flow and Catchment Land Use, Including the December Appendum. Report Prepared by NIWA for Environment Waikato, Hamilton. Environment Waikato Technical Report, 2005/03. Thanabalasingam, P., Pickering, W.F., 1990. Specific sorption of antimony(III) by the hydrous oxides of Mn, Fe, and Ai. Water Air Soil Pollut. 49, 175–185. Tighe, M., Lockwood, P., Wilson, S., Lisle, L., 2004. Comparison of digestion methods for ICP-OES analysis of a wide range of analytes in heavy metal contaminated soil samples with specific reference to arsenic and antimony. Commun. Soil Sci. Plant Anal. 35, 1369–1385. Tighe, M., Lockwood, P., Wilson, S., 2005. Adsorption of antimony(V) by floodplain soils, amorphous iron(III) hydroxide and humic acid. J. Environ. Monitor. 7, 1177–1185. Timperley, M.H., Huser, B.A., 1996. Inflows of geothermal fluid chemicals to the Waikato River catchment, New Zealand. New. Zeal. J. Mar. Freshwater Res. 30, 525–535. Webster, J.G., 1994. Trace-metal behaviour in oxic and anoxic Ca-Cl Brines of the Wright Valley Drainage, Antarctica. Chem. Geol. 112, 255–274. Webster, J.G., Nordstrom, D.K., 2003. Geothermal arsenic. In: Welch, A.H., Stollenwerk, K.G. (Eds.), Arsenic in Ground Water. Kluwer Academic Publishers, Boston, pp. 101–126 (Chapter 4). Webster-Brown, J.G., Lane, V., 2005a. Modeling seasonal arsenic behaviour in the Waikato River, New Zealand. In: O’Day, P., Vlassopoulos, D., Meng, X., Benning, L. (Eds.), Advances in Arsenic Research: Integration of Experimental and Observational Studies and Implications for Mitigation. Am. Chem. Soc. Symp. Ser, vol. 915. American Chemical Society, Washington, pp. 253–266. Webster-Brown, J.G., Lane, V., 2005b. The environmental fate of geothermal arsenic in a major river system, New Zealand. In: Proc. World Geothermal Congress 2005, Antalya, Turkey. Webster-Brown, J.G., Webster, K.S., 2000. Trace Metals in the Waikato River. Environmental Science and Research, Report CO3081. Webster-Brown, J.G., Webster, K.S., Lane, V., 2000. Arsenic in the Waikato River: an update. In: Proc. 22nd New Zealand Geothermal Workshop, Auckland, New Zealand. Wenzel, W.W., Kirchbaumer, N., Prohaska, T., Stingeder, G., Lombi, E., Adriano, D.C., 2001. Arsenic fractionation in soils using an improved sequential extraction procedure. Anal. Chim. Acta 436, 309–323. Wilson, N.J. 2008. The Behaviour of Antimony in Geothermal Systems and their Receiving Environments. Unpublished Ph.D. Thesis, Univ. Auckland. Wilson, N.J., Webster-Brown, J., Brown, K., 2007. Controls on stibnite precipitation at two New Zealand geothermal power stations. Geothermics 36, 330–347. Zotov, A.V., Shikina, N.D., Akinfiev, N.N., 2003. Thermodynamic properties of the Sb(III) hydroxide complex Sb(OH)3(aq) at hydrothermal conditions. Geochim. Cosmochim. Acta 67, 1821–1836.