| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a U.S. Geological Survey, Water Resources Division, 345 Middlefield Road, Mail Stop 480, Menlo Park, CA 94025
b Dep. of Geology, 245 Natural History Bldg., 1301 W. Green St., Univ. of Illinois, Urbana, IL 61301
c Dep. LAWR, One Shield Avenue, Univ. of California, Davis, CA 95616
* Corresponding author (mjherbel{at}usgs.gov)
Received for publication August 31, 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
) between surface water and reduced Se species in the underlying sediments. Selenium in aquatic macrophytes was very similar isotopically to the NaOH and Na2SO3 sediment extracts designed to recover soluble organic Se and Se(0), respectively. For the integrated on-farm drainage management sites, evaporite salts were slightly (approximately 0.6) enriched in the heavier isotope relative to the inferred parent waters, whereas surface soils were slightly (approximately 1.4) depleted. Bacterial or chemical reduction of Se(VI) or Se(IV) may be occurring at these sites, but the small isotopic contrasts suggest that other, less isotopically fractionating mechanisms are responsible for accumulation of reduced forms in the sediments. These findings provide evidence that Se assimilation by plants and algae followed by deposition and mineralization is the dominant transformation pathway responsible for accumulation of reduced forms of Se in the wetland sediments.
Abbreviations: IFDMS, integrated on-farm drainage management system • TLDD, Tulare Lake Drainage District
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
A major concern for these drainage water management strategies is onsite accumulation of Se in sediments and waters through evapoconcentration or biological and chemical sequestering mechanisms. For example, the Tulare Lake Drainage District (TLDD) is testing the ability of constructed flow-through wetlands with various vegetation regimes to specifically remove Se from irrigation water prior to disposal into evaporation ponds (Drainage Water Treatment Technical Committee, 1999). Sediments and aquatic plants in these wetlands accumulate Se well above background levels. Integrated on-farm drainage management systems (IFDMS), which are being tested to reuse the saline drainage water on salt-tolerant crops, eucalyptus trees, and halophytes to minimize the volume of water requiring impoundment or disposal (Cervinka, 1994; Imhoff et al., 1993; Tanji and Karajeh, 1993), will evapoconcentrate and bioaccumulate both major ions and trace elements. Throughout the region, evaporation basins collect nonreusable saline drainage water as a temporary disposal option, and many trace elements absorb onto the sediments or accumulate in the evaporite minerals (Tanji et al., 1992).
The biogeochemical cycling and long-term fate of Se in these drainage water management systems are still not fully understood. Selenium can exist in multiple oxidation states (VI, IV, 0, -II) and as both inorganic and organic compounds (McNeal and Balistrieri, 1989). The mobility and bioavailability of Se in the various compounds ranges from the highly soluble Se(VI) (or SeO2-4) and volatile methylated selenides (e.g., dimethyl selenide) to the immobile elemental Se [Se(0)] and metal–selenide precipitates (Elrashidi et al., 1989). Consequently, evaluation of the primary reaction processes and accumulation patterns within each site is important, but usually this requires Se speciation and concentration analyses for many samples over prolonged periods of time.
Selenium stable isotope ratios are useful as potential indicators for Se oxyanion reduction in the environment (Johnson et al., 1999). Isotope ratios of lighter elements, such as C, N, and S, have long been used as indicators of solute sources and chemical reactions in geochemical studies. Many chemical reactions fractionate the isotopes according to their mass. For example, lighter isotopes of sulfur are more reactive than the heavier isotopes during sulfate reduction, and 34S/32S ratios thus provide evidence of sulfate reduction in ground water (Habicht and Canfield, 1997; Kaplan and Rittenberg, 1964; Strebel et al., 1990). The 15N/14N ratios are similarly useful as indicators of denitrification (Boettcher et al., 1990; Lund et al., 2000; Mariotti et al., 1981). Dissimilatory reduction of Se(VI) and Se(IV) oxyanions to more reduced forms [Se(IV) and Se(0)] by pure bacteria cultures is associated with significant (up to 8.4) fractionation (Herbel et al., 2000). Reduction of Se oxyanions by Fe(II) + Fe(III) hydrous oxides similarly induces large (up to 12) isotopic discriminations (Johnson and Bullen, 2000). Biological reduction of Se oxyanions in slurries using TLDD, San Luis Drain, and San Francisco Bay estuary sediments revealed 80Se/76Se shifts of 2.8 ± 0.2 for Se(VI) reduction and 5.6 ± 0.2 for Se(IV) reduction (Ellis et al., 2002).
In contrast, other processes are associated with minimal (<1) isotopic fractionation (Johnson et al., 1999). Oxidation of reduced Se in sediments to soluble Se(VI) and Se(IV) results in <1 fractionation, and adsorption processes, such as Se(IV) adsorption onto amorphous ferric hydroxides fractionates Se by approximately 0.5. Preliminary results from algae cultures indicate that Se assimilation ranges from 1 to 2.5 fractionations, with one algae sample grown in San Joaquin Valley agricultural waste water exhibiting 1.0 fractionation (Hagiwara, 2000). Volatilization of Se from plants and sediments, which is known to be important in Se-affected environments (Frankenberger and Karlson, 1994; Karlson and Frankenberger, 1990; Terry and Zayed, 1998), induces only minor isotopic fractionation (Johnson et al., 1999). Consequently, only reduction of Se oxyanions by chemical or bacterial dissimilatory processes significantly fractionates the stable isotopes of Se, and isotopic shifts in nature should provide evidence for these processes.
The goals for this study were to (i) use the stable isotope approach to better distinguish the relative importance of chemical or dissimilatory reduction versus plant and algae assimilation in the removal of Se from the agricultural drainage water at TLDD flow-through wetland and IFDMS sites, (ii) analyze plant samples to determine the isotopic fractionation associated with macrophyte Se uptake, and (iii) gather reconnaissance isotope ratio data from these and other Se-affected sites in the western San Joaquin Valley of California. This investigation augments other research conducted at these sites (Gao et al., 2000; Tanji, 1999; Terry et al., 1999) in that it specifically examines the mechanisms affecting Se accumulation in TLDD or IFDMS sediments.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
At the Mendota IFDMS site, shallow ground waters (water table approximately 2 m below surface) collected from monitoring wells were processed similarly to the TLDD surface and pore waters. Evaporite samples were dissolved in distilled, deionized water, filtered to remove insoluble material, and acidified with HNO3. Sediment core sections (0 to 25 cm) were collected in plexiglass tubes and frozen at -4°C until later analyses.
Selenium Concentration Determination and Extraction Techniques
The dried and ground plants collected from the TLDD flow-through wetland site were digested with a modified HNO3–H2O2 method (Zarcinas et al., 1987). The plant material (0.2 g) was refluxed at 40°C with 10 mL of concentrated HNO3, with periodic additions of 200 µL 30% H2O2, over a period of 5 d. The solutions were evaporated to near dryness, then dissolved with 6 M HCl. Subsamples were heated to 100°C for 45 min and analyzed by inductively coupled argon plasma mass spectrometry (ICP–MS) (Soltanpour et al., 1996).
Speciation of the soluble Se in shallow ground waters, surface waters, and pore waters was determined by hydride generation atomic absorption spectroscopy (HGAAS) (Cutter, 1986). The Se(IV) concentrations were determined by direct analyses with HGAAS, and Se(VI) was determined by difference from subsamples heated at 100°C with 6 M HCl for 45 min [Se(VI) + Se(IV)] and Se(IV) concentrations. Total dissolved Se concentrations were determined after HClO4–HNO3 digestion (Zasoski and Burau, 1977) followed by 6 M HCl reduction at 110°C for 45 min. Dissolved organic Se was considered to be the difference between the total dissolved Se and [Se(VI) + Se(IV)].
For the TLDD site, Se(VI) was selectively removed from solution for 
80/76Se analyses due to the greater variability in Se speciation of the waters in the wetlands, and because potential isotopic fractionation from reduction processes would be more evident in this species. To separate Se(VI) from other dissolved Se species, a modified ferric hydroxide–coprecipitation technique (Nakashima, 1979) was used to isolate Se(VI) for later mass spectrometry analyses. The acidified water samples were mixed with 5 mg Fe(III) as FeNH4(SO4)2·12H2O. The solution was slowly titrated to pH 4.0 with NH4OH, which induced amorphous ferric hydroxide to precipitate. Less soluble forms of Se, such as Se(IV), were coprecipitated and removed from solution by filtration. The filtrate containing Se(VI) was then further purified for isotope ratio analyses.
To test for possible isotopic variation in the Se that accumulated in sediments, sections of sediment cores were either completely dissolved with HClO4–HNO3 digestion (Zasoski and Burau, 1977) or sequentially extracted to remove the Se from the different sediment compartments. Selenium associations in the TLDD sediments were subdivided as previously described (Gao et al., 2000). Briefly, the wet sediments were extracted in the order: (i) 0.25 M KCl to remove soluble Se, (ii) 0.1 M K2HPO4 at pH 8.0 to remove adsorbed Se, (iii) 1.0 M sodium acetate (NaOAc) to remove carbonate-associated Se, (iv) 0.1 M NaOH to remove organic matter–associated Se, and (v) 1.0 M Na2SO3 to remove the Se(0). A portion of the original wet sediment was digested with the HClO4–HNO3 method for total Se analyses (Zasoski and Burau, 1977). All extracts were analyzed for Se by digesting subsamples with HClO4–HNO3 or K2S2O8–H2O2 followed by reduction of Se(VI) to Se(IV) and hydride generation atomic absorption spectroscopy analyses. Portions of the digested sediments and sediment extracts were then analyzed for 80/76Se as described below.
The IFDMS and evaporation pond sediment cores were extracted in the order (i) 0.001 M Na2HPO4 to remove soluble and adsorbed Se, (ii) 0.02 M NaOH to remove organic Se, and (iii) HClO4–HNO3 digestion of the remaining sediments to account for refractory Se. Sediment cores were thawed at room temperature and extruded onto aluminum foil. The oxidation–reduction potential of the sediments was measured within 5 min of extrusion. A 15- to 18-g portion of sediment was removed by slicing the top 0 to 5 cm of the core vertically, and transferred into 200 mL 0.001 M Na2HPO4 at pH 8. The sediment suspension was shaken for 24 h at 22°C, then centrifuged at 7000 RCF for 30 min. The supernatant, containing the soluble and adsorbed Se fractions from the sediment, was passed through a 0.45-µm cellulose acetate filter and refrigerated. The sediments were then suspended in 100 mL 0.02 M NaOH, placed in an 85°C water bath and shaken every 20 min for 2 h, then centrifuged. The Na2HPO4 and NaOH extracts were evaporated on a hotplate at 35°C to concentrate the solution prior to HClO4–HNO3 digestion. The remaining sediment was digested with HClO4–HNO3 (Zasoski and Burau, 1977) to recover refractory Se, which consists of Se(0) and other reduced forms tightly bound to oxides and aluminosilicates. Samples of the original sediments were similarly digested to obtain total Se.
Stable Isotope Ratio Methodology
Purification
Thermal ionization mass spectrometry (TIMS) analysis requires that the Se be separated from matrix components and purified. This was done with a modified NaBH4–HNO3 purification method (Tanzer and Heumann, 1991). Samples were diluted to 100 mL with 4 M HCl in a Teflon reaction vessel and bubbled with N2 at 350 mL min-1 to remove any O2. A 10% NaBH4 solution was pumped into the reaction vessel at a flow rate of 1.25 mL min-1 for 10 min. The H2Se formed by reduction of Se(IV) was swept out of the reaction vessel by N2, and trapped by bubbling into concentrated HNO3. This method was used for all samples except the TLDD water samples, which were purified with a continuous flow HCl–NaBH4 reaction vessel in place of the batch reaction vessel (Ellis et al., 2002).
The HClO4–HNO3 sediment digests typically contained abundant Fe(III), which interferes with NaBH4 purification. To remove Fe(III), the samples were diluted to 0.7 M HCl and passed through a cation exchange column containing BioRad (Hercules, CA) AG50W-X8 resin in the H+ form. An additional 10 mL of 0.7 M HCl was passed through to rinse any residual Se from the resin. The Se(IV) standards that were mixed with 100 mg L-1 Fe(III) in a 0.7 M HCl matrix were processed in this manner and yielded the expected values.
Double Isotope Spike
Significant, variable shifts in 80Se/76Se ratios that are known to occur during purification of Se via hydride generation and thermal ionization mass spectrometry analyses are corrected with the double isotope spike technique. The double spike contains 82Se and 74Se in a known ratio; a predetermined amount of this solution, with Se speciation matching that of the sample, is added to each sample. The measured 82Se/74Se ratio can be used to calculate and correct for any isotopic fractionation that occurs after the spike is added. Details of this technique are described in a recent paper (Johnson et al., 1999). The double spike only works correctly if the chemical forms of the spike and the sample Se to be analyzed are the same. Accordingly, the double spike was added as early as possible in the sample preparation process, but always after the targeted Se fraction(s) was known to be entirely Se(VI) or Se(IV). For example, speciation of Se in the raw sediment extracts or total sediment digestions was not known, and in these cases, the double spike was added after all Se had been converted to Se(IV). Any isotopic fractionation occurring prior to spike addition is not detected, but fractionation should not occur in these steps. Significant loss of isotopically fractionated Se must occur for this to happen, and these methods have been shown to retain Se quantitatively (Gao et al., 2000; Tokunaga et al., 1991; Zhang and Moore, 1996; Zhang et al., 1999).
Mass Spectrometry
The Se isotope ratios were measured with thermal ionization mass spectrometry, in which ions generated on a hot metal filament were accelerated, separated by a magnetic sector, and measured to very high precision. After NaBH4 purification, the concentrated HNO3 samples containing the purified Se were evaporated to dryness and loaded onto filaments for mass spectrometry (Johnson et al., 1999). In some instances, volatile organic contaminants from the digested and reduced samples were trapped in the concentrated HNO3 along with the H2Se. These organics were decomposed by adding 100 µL concentrated HNO3 and 50 µL 30% H2O2 to the evaporated sample, and again heating to dryness. The cleaned samples were suspended in 1 to 5 µL distilled, deionized water and mixed with colloidal graphite suspension. The sample containing 0.25 to 1.0 µg total Se was loaded on top of dried Ba(OH)2 (2.0 µL saturated solution) on a rhenium filament. The 74Se/76Se, 76Se/80Se, and 82Se/80Se ratios were measured on a Finnigan MAT 261 thermal ionization mass spectrometer (Thermo-Finnigan, San Jose, CA) in negative ion mode, and these raw data were processed with an iterative double spike data reduction routine (Johnson et al., 1999). This routine uses the measured isotope ratios to separate the contributions of the spike and sample, calculate and correct for the isotopic shifts in preparation and mass spectrometry, and calculate 80/76Se values for the samples. The 80/76Se values are defined as:
80/76Se values reported are averages of multiple (3 to 8) blocks of data, each block containing 10 measurements of each ratio. Error bars represent the larger value of (i) the internal precision (2
= 95% confidence interval) of each compiled analyses, which could be as low as 0.05 under optimal conditions, or (ii) 0.25, the estimated external precision determined by replication. Samples selected for duplicate purification and Se isotope ratio analyses typically resulted in good reproducibility, with average absolute differences from the reported values being 0.22 with a standard deviation of 0.16 (n = 10). The running average and 2 value for nonhydride processed MH495 Se(IV) standard (n = 18) was 0.05 ± 0.06, while hydride processed MH495 Se(IV) standard (n = 12) was 0.09 ± 0.08. |
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
80/76Se values for macrophytes, waters, and sediment sequential extracts are shown in Fig. 3, 4, 5, and 6
, respectively. The 80/76Se values also were grouped according to sample type for statistical analyses using T tests (Table 1)
. Sample types included (i) surface and pore waters, (ii) macrophytes, (iii) 0- to 5-cm sediment extracts, and (iv) 5- to 10-cm sediment extracts.
|
|
|
|
|
Five samples were analyzed for 80/76Se. The total Se in the plant samples showed little isotopic variation, with 80/76Se values ranging from +2.63 for the saltmarsh bulrush roots to +3.15 for the widgeon grass (Fig. 3). Baltic rush and saltgrass yielded insufficient Se for 80/76Se determinations. These results indicate that macrophyte uptake of soluble Se results in minor isotopic fractionation, as 80/76Se values in the plant digests averaged only 0.74 less than the Se(VI) in the measured surface waters and 0.46 less than the mean of all the sediment extracts (Table 1). This pattern occurs in all of the cells, despite the fact that the cells have differing macrophyte populations with varying abilities to accumulate Se. The small but consistent isotopic fractionation upon Se oxyanion uptake by macrophytes suggested by the data of the present study is consistent with data for sulfate and nitrate uptake. Sulfate uptake and assimilation by plants and algae is associated with only minor shifts in 34S/32S (Kemp and Thode, 1968; Thode, 1991), and assimilation of nitrate to organic N by macrophytes in a flow-through wetland showed minimal fractionation of 15N/14N (Lund et al., 2000). Thus, assimilation of Se, N, and S oxyanions involves little isotopic fractionation, presumably because reduction reactions occur deep within the plants where they cannot influence the overall isotopic composition of the assimilated flux. The results presented here for Se confirm preliminary assertions that macrophyte uptake involves little Se isotope fractionation (Johnson et al., 1999, 2000). Assimilation of Se oxyanions by algae similarly involves a small isotopic fractionation of about 1 (Hagiwara, 2000).
Surface Waters
The relative proportions of Se(VI), Se(IV), and organic Se changed as water flowed from the inlet to the outlet across Cells 1 and 3 (Fig. 4); these findings are similar to previous observations (Gao et al., 2000). At the inlet of each cell, Se(VI) dominated, reflecting a highly oxidized source of drainage water. In Cell 1, total Se concentrations decreased as the water flowed across the cell, and by the time it reached the outlet, Se(VI) dropped to 24% of total dissolved Se, organic Se increased to 63%, and Se(IV) remained a minor component (13%). Similar trends were observed for Cell 3. Total dissolved Se dropped from inlet to outlet by 23% in Cell 1 and 12% in Cell 3. Total dissolved Se concentrations were less in waters collected from shallow wells at a 100-cm depth in Cell 1 (1A-100) (66% decrease relative to overlying surface water) and a 50-cm depth in Cell 3 (3A-50) (53% decrease), with the latter having organic Se as the dominant Se species.
The mechanism of Se(VI) removal from the surface waters is reflected in the Se isotope data. The 80/76Se values of Se(VI) in Cell 1 showed little variation, ranging from +3.63 to +3.93 (Fig. 4). For Cell 3, two samples yielded +3.75 and +3.07. Constancy in the 80/76Se values of Se(VI) across the cells suggests that dissimilatory or chemical reduction was not the dominant Se(VI) removal mechanism at the time of sampling or that the isotopic fractionations induced by these reactions were somehow not manifested in the surface water Se(VI). Dissimilatory reduction involves an enrichment of the lighter isotopes in the reaction products; Se (VI) reduction by TLDD sediments induces a 80Se/76Se shift of 2.8 ± 0.2 (Ellis et al., 2002). As reduction proceeds, the remaining Se(VI) becomes progressively enriched in the heavier isotopes. In Cell 1, the surface water lost more than 75% of its Se(VI) in transit from inlet to outlet, and yet no significant shift in 80/76Se occurred. This provides strong evidence that dissimilatory reduction is not responsible for Se(VI) reduction by the TLDD wetlands.
Sediments
The TLDD sediments have accumulated Se primarily in the shallow 0- to 5- and 5- to 10-cm intervals; considerably less Se reaches lower sediment depths (Fig. 5). Only minor amounts of Se in these sediments are in the soluble form (indicated by 0.25 M KCl extraction). Cores 1A and 1C have nearly equivalent amounts of Se in the K2HPO4 (adsorbed Se), NaOH (organic Se), and Na2SO3 [Se(0)] extracts, with negligible amounts in the NaOAc (carbonate Se) extracts. Cores 3C and 5C showed greater proportions of Se in the NaOH and Na2SO3 extracts.
The range of 80/76Se values in the sequential extracts from the TLDD site was small, with most extracts being slightly enriched in 76Se relative to surface water Se(VI) (Fig. 6 and Table 1). These isotopic contrasts were minor in comparison with isotopic shifts previously measured for Se(VI) and Se(IV) reduction by sediments from this site, which were 2.8 for Se(VI) reduction to Se(IV) and 5.6 for Se(IV) reduction to Se(0) (Ellis et al., 2002). The 80/76Se values for the Na2SO3 extracts, which recover primarily Se(0), were 0.4 to 1.0 less than the Se(VI) in the water column. Lack of consistent, large (e.g., 5) isotopic differences between these extracts and the Se(VI) in the overlying water column suggests that the dominant pathway responsible for the accumulation of reduced Se involves little or no isotopic fractionation. Notably, the mean 80/76Se of the Na2SO3 extracts is indistinguishable from that of the macrophytes.
Selenium in phosphate and hydroxide extracts, which are designed to extract adsorbed and organically bound fractions, respectively, was also isotopically similar to the surface waters and macrophytes, though a few of the deeper samples yielded variations as great as 2 from the mean (Table 1). The 0- to 5-cm K2HPO4 extracts were isotopically identical to the surface waters (means: +3.63 and +3.65, respectively). However, one deeper (5 to 10 cm) extract was lower by approximately 3.0 (Fig. 6). This isotopic excursion indicates that at this location, Se oxyanion reduction must be substantial (see below).
The 80/76Se values of the 0- to 5-cm NaOH extracts were slightly lower than those of the surface water Se(VI), and were indistinguishable from the macrophytes (Table 1). This is not surprising, as decaying macrophyte tissues provide the dominant source of soluble, organically bound Se in the sediments. However, the two NaOH extracts from the deeper (5 to 10 cm) interval were enriched in the heavier isotopes. This Se could not have been derived entirely from macrophyte tissues, which have distinct isotopic values, and must have been partially affected by Se oxyanion reduction (see below).
Integrated On-Farm Drainage Management System and Evaporation Pond Samples
At the Mendota IFDMS site, Se has evapoconcentrated and accumulated in the sediments and waters, often exceeding 1 mg L-1 or 1 mg kg-1 (Table 2)
. The measured oxidation–reduction potentials (+418 to +441 mV), electrical conductivity (11.4 to 35.6 dS m-1), pH (7.5 to 8.0), and dissolved Se species in the shallow ground waters indicate that the near-surface environment is very oxidized and highly saline. Selenate accounts for at least 80% of the total dissolved Se species, with Se(IV) and organic Se being present in smaller amounts. This also is reflected in the evaporite salts, where Se(VI) makes up 90% of total Se in the dried minerals. The sediment cores are moderately to highly oxidized (Table 3)
. Minor to moderate amounts (4 to 36%) of Se are found in the soluble or adsorbed fractions (Na2HPO4 extract) and organic fractions (10 to 16%), while the majority of the total Se resides in the refractory fraction (53 to 88%). This refractory fraction must be dominated by reduced forms of Se (i.e., elemental Se, aluminosilicate Se, and insoluble organically associated Se).
|
|
80/76Se between shallow ground waters, evaporites, and soils are small (Tables 2 and 3). Shallow ground waters ranged from +2.93 for the halophyte field (MWHC) to +3.65 for the eucalyptus zone (MWB1). The highest 80/76Se values, approximately +4.2, were observed for two evaporite salt samples, which are produced by evaporation of the final recycled water. The 80/76Se values for total Se in sediments were slightly lower than the shallow ground waters, with sediments from the higher-salinity halophyte field (+2.54, MHB2) being only slightly greater than the lower-salinity eucalyptus zone (+1.97, MNW). The 80/76Se measurements were obtained for two sequential extracts from the halophyte soil (MHB2): the Na2HPO4 extract containing soluble and adsorbed Se was +2.75 while the refractory Se was +2.23. The sequential extracts of the MHB2 sediment yielded similar 80/76Se values to those found for two evaporation pond sediments, the exception being the NaOH extract for Lost Hills evaporation pond sediments (+5.06).
The lack of strong variation in these 80/76Se values has implications for Se dynamics in these oxidized environments. The 80/76Se values of the soils are only slightly less than the shallow ground waters, while 80/76Se values for the evaporite salts are slightly greater (approximately 1) than the shallow ground waters. Reduced Se species that dominate the sediments must have been generated predominantly by plant uptake and deposition of dead plants, because this process is not expected to induce isotopic fractionation (see below). Chemical and dissimilatory bacterial reduction of Se would lead to enrichment of heavier isotopes in the residual Se(VI) in the recycled water and the terminal evaporites and enrichment of lighter isotopes in the sediments. However, the small isotopic contrasts between the Se in the ground waters and evaporites versus Se in the sediments suggest that this later mechanism does not dominate Se uptake and removal from the IFDMS waters.
Selenium Cycling Scenarios
A conceptual diagram of the Se cycle based on Se concentration determinations and 80/76Se measurements for the TLDD flow-through wetland system is shown in Fig. 7
. A large proportion (e.g., 30%) of soluble Se(VI) from the inlet water is assimilated by algae and macrophytes in Cells 1 and 5, and by algae in Cell 3. This is reflected in the high Se concentrations in the saltmarsh bulrush and rabbitsfoot grass, which contribute greatly to the organic matter in the sediments. Algae in Cell 3 also assimilates Se and is probably the primary contributor to the organic component of the sediment Se. For Cell 1, the average 80/76Se for the saltmarsh bulrush (shoots plus roots) was +2.8, which is also the average value for the four Na2SO3 extracts for sediments collected at cores 1A and 1C. Losses of Se via volatilization by the macrophytes, microbes, or algae may occur, yet in this scenario the small degree of isotopic fractionation caused by volatilization does not dramatically alter the 80Se/76Se in the plants, waters, or sediments (Johnson et al., 1999).
|
80/76Se values of the living macrophytes and the Na2SO3 extracts support this interpretation.
If production of Se(0) in these sediments were mostly the result of reduction of Se(VI) diffusing into the sediments from the surface water, the Se(0) should have been enriched in the lighter isotope relative to the Se(VI). The sediments are indeed slightly enriched in the lighter isotope, but this offset is quite small. The average difference between the surface water 80/76Se values and those of the Na2SO3 extracts is approximately 0.7. This suggests that bacterial dissimilatory or chemical reduction of Se(VI) oxyanions to Se(IV) or Se(0) is not the dominant Se(0)-producing mechanism, as compared with mineralization of organic matter.
The only samples that show evidence for the larger isotopic fractionations that might be expected for Se oxyanion reduction are one K2HPO4 extract and a few of the NaOH extracts. For the TLDD Core 1A, 5- to 10-cm depth, K2HPO4 extract, the 80/76Se value was +0.60, which is substantially lighter than the other extracts. The K2HPO4 primarily removes adsorbed Se(IV), which is the expected intermediate species during bacterial dissimilatory reduction of Se(VI) to Se(0), if it is occurring in the sediments. The 80/76Se values in this extract are expected to be controlled by two potentially simultaneous reactions: Se(VI) reduction to Se(IV) and Se(IV) reduction to Se(0), which both have been observed in batch experiments with pure bacterial cultures (Herbel et al., 2000) and sediments from this site (Ellis et al., 2002). The Se(IV) produced by Se(VI) reduction would be enriched in lighter isotopes relative to the surface water. However, Se(IV) reduction induces a greater isotopic fractionation than Se(VI) reduction, and it would thus tend to enrich the pool of Se(IV) in the sediments in heavier isotopes. Accordingly, the enrichment of the one K2HPO4 extract in lighter isotopes suggests that in this location, Se(VI) reduction was occurring at the time of sampling, but reduction of the produced Se(IV) was limited. The reader will note that this extract represents a small fraction of the total Se in the core, and that other K2HPO4 extracts from TLDD yielded 80/76Se values close to those of the overlying water and Na2SO3 extracts. If Se oxyanion reduction reactions were occurring, we would expect that K2HPO4 extracts of the deeper 5- to 10-cm sediment cores would have greater 80/76Se values, but this was not observed in Cores 1A or 3C.
A few NaOH extracts were 1.9 to 2.5 enriched in the heavier isotope relative to the average sedimentary 80/76Se values: these were TLDD Cores 1A and 3C, 5- to 10-cm depths, and a core from the Lost Hills evaporation pond. These results may represent partial reduction, and therefore an isotopically shifted Se(IV) fraction that became strongly adsorbed onto organic matter. It is known that Se(IV) can be strongly bound to or coated by organic compounds in sediments (Martens and Suarez, 1997; Tokunaga et al., 1991), thus, the organic matter solubilized by the NaOH extracts could contain a strongly sorbed Se(IV) component along with organic Se(-II). This scenario explains the aberrant data, but does not apply to most of the sediments, as most of the NaOH extracts were not isotopically un-usual.
Similar cycling of Se occurs at the Mendota IFDMS site. The 80/76Se values suggest that the dominant Se removal pathway is uptake of Se(VI) from the reapplied tile-drained water by crops, eucalyptus trees, and halophytes followed by plant deposition and organic Se mineralization to refractory forms [Se(0) and reduced Se associated with sediment minerals]. Although not directly measured in this study, uptake of Se oxyanions by eucalyptus trees [leaves up to 1270 mg Se kg-1 dry weight (Tanji, 1999)] and halophytes is expected to result in only minor isotopic fractionation, similar to that observed with macrophyte uptake at the TLDD flow-through wetlands. Decomposition and mineralization of eucalyptus tree and halophyte detritus to refractory Se forms are the dominant Se sequestering mechanisms in the sediments, even under highly oxidizing conditions. Further oxidation of organic Se and Se(0) to more soluble Se(VI) and Se(IV) oxyanions is a slow process (Dowdle and Oremland, 1998; Losi and Frankenberger, 1998; Zawislanski and Zavarin, 1996); nevertheless, this is similarly associated with only minor isotopic fractionation (Johnson et al., 1999, 2000).
Chemical and dissimilatory Se oxyanion reduction by the native bacteria, which require more anaerobic conditions (Myneni et al., 1997; Oremland and Stolz, 2000), are probably inhibited at the Mendota IFDMS site due to the highly oxidized nature of the sediments and waters. If these reduction processes were occurring, we would expect lower Se(VI) concentrations and higher 80/76Se values for the shallow ground waters in the halophyte zone (MWHC and MWHE) compared with the eucalyptus zone (MWB1 and MWE1). Only the former was observed. Any Se(VI) not removed by plant uptake that remains in the highly evaporated tile-drained water accumulates in the evaporite salts in the solar evaporators. The minor increases in 80/76Se observed in the evaporite salts compared with the shallow ground waters and sediments may indicate that a small degree of chemical or dissimilatory reduction of Se oxyanions occurred in the tile-drained water from the halophyte field, but the 80/76Se offset was much smaller than what would be expected if these processes were more dominant.
Additional Factors to Consider
The 80/76Se values in the sediments, particularly for Se(0) in the Na2SO3 extracts, probably reflect averaging of variable conditions over different seasons. The inflows, outflows, and water depths have varied, and vegetation has slowly increased coverage of the different cells from year to year. The dominance of different species of algae that colonized the wetlands may shift from year to year, and from season to season, as the macrophyte vegetation becomes established. Roots may exude oxygen during times of intense photosynthesis activity (Velinsky and Cutter, 1991). The varying redox status of the sediments may affect dissimilatory reduction in the sediment profile, and may even allow for transient reoxidation of Se(0). Despite the potential variation in the geochemical processes, the Se(0) isotope ratios vary little across the site. This is expected if plant and/or algal Se uptake and decomposition is the dominant Se(0)-forming mechanism.
It is also difficult to predict accurately the size of the isotopic difference that should occur in a system where the dominant source of Se(0) in the sediments is chemical or dissimilatory reduction of Se(VI) in surface water. The isotopic fractionations observed by Ellis et al. (2002) provide an estimate of the fractionation that occurs during the reduction reactions, but exchange of Se(VI) and Se(IV) between pore waters and surface water plays an additional role in determining the actual isotopic difference between Se(0) produced in sediments and the parent Se(VI). For example, a spatial pattern is expected whereby the deeper pore waters would be enriched in 80Se relative to the shallower pore waters because 76Se has been preferentially removed by reduction. This demands that the deeper Se(0) is enriched in 80Se relative to the shallower Se(0). At the sediment–water interface, the Se(0) is enriched in 76Se by the full amount of the fractionation induced by reduction, whereas in the deeper sediments, this difference is less. Thus, while the overall effect is still an enrichment of lighter isotopes in the sediments, the size of that enrichment is somewhat less than the fractionation induced by reduction. We do not currently know how much less, and a modeling study is needed to explore the theoretical aspects of this.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
80/76Se data presented here for several agricultural drainage water processing sites in the San Joaquin Valley of California improve our understanding of Se stable isotope systematics and the cycling of Se in these settings. Uptake of Se(VI) and Se(IV) oxyanions from the water column and sediment pore waters by macrophytes at the TLDD flow-through wetland was associated with only minor isotopic fractionation (approximately 0.7), with the assimilated Se in the plants slightly favoring the lighter 76Se. Little isotopic variation was observed among four of the more dominant plants. This result confirms our hypothesis that of the major biogeochemical processes affecting Se, reduction of Se oxyanions is the only one that induces 80/76Se shifts of more than 1.
Because Se stable isotopes are sensitive primarily to reduction of Se oxyanions, 80/76Se measurements can be used as reduction indicators. Sediment extracts designed to recover Se(0) had 80/76Se values only slightly lower than those of Se(VI) in the surface waters, and nearly indistinguishable from those of the macrophytes. If the Se(0) were generated by reduction of Se(VI) to Se(IV) followed by reduction of Se(IV) to Se(0), it should be enriched in the lighter isotope by several per mil. We conclude that dissimilatory reduction within the reducing environment of the sediments is not the dominant Se(0)-producing mechanism. The isotope data suggest that the actual mechanism is release of Se from decaying vegetation, followed by oxidation to Se(0). The success of wetlands in removing Se from wastewater thus apparently derives more from growth of plants and/or algae and less from diffusion of Se(VI) into reducing sediments.
The 80/76Se values of most of the phosphate and hydroxide extracts designed to extract adsorbed and organic fractions, respectively, were also within 1.0 of the surface or shallow ground water values. This further suggests that Se(VI) reduction does not dominate the Se budget of the sediments overall. However, deviations in 80/76Se of roughly 2 from the surface water value occurred in a few of the deeper TLDD extracts and suggest that limited reduction of Se oxyanions occurred in some locations.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. M. Johnson and T. D. Bullen Mass-Dependent Fractionation of Selenium and Chromium Isotopes in Low-Temperature Environments Reviews in Mineralogy and Geochemistry, January 1, 2004; 55(1): 289 - 317. [Full Text] [PDF]
|
|
| |||||||||||||||||||||
