JEQ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (5)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Zawislanski, P.T.
Right arrow Articles by Wong, H.C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Zawislanski, P.T.
Right arrow Articles by Wong, H.C.
Agricola
Right arrow Articles by Zawislanski, P.T.
Right arrow Articles by Wong, H.C.
Related Collections
Right arrow Water Quality
Right arrow Wetlands and Aquatic Processes
Right arrow Other Environmental Contamination
Right arrow Water Pollution
Journal of Environmental Quality 30:1080-1091 (2001)
© 2001 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America

TECHNICAL REPORT
Wetlands and Aquatic Processes

Selenium Distribution and Fluxes in Intertidal Wetlands, San Francisco Bay, California

P.T. Zawislanski, H.S. Mountford, E.J. Gabet, A.E. McGrath and H.C. Wong

Earth Sciences Division, Mail Stop 90-1116, Lawrence Berkeley National Lab., Berkeley, CA 94720

Corresponding author (PTZawislanski{at}lbl.gov)

Received for publication January 5, 2000.

   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 Sediment and Selenium Dynamics
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY
 REFERENCES
 
Selenium (Se) concentrations exceeding ecological guidelines for sediments and suspended particulate matter (SPM) have been observed in the northern reach of the San Francisco Bay estuary. Long-term availability of elevated Se in wetland sediments depends in part on the fluxes controlling Se distribution. The relative contribution of sedimentary vs. post-depositional Se fluxes in two San Francisco Bay intertidal wetlands was estimated. Selenium concentrations on surface wetland sediments were compared with levels on SPM, and with previously established background levels in San Francisco Bay sediments. Sediment Se fluxes to the wetlands were measured directly using sediment traps. Although dissolved Se concentrations are higher than particulate Se concentrations in San Francisco Bay water, sediment input into the system provides the major flux of Se. Strong correlation between Se and C on SPM (r2 = 0.81) indicates the importance of organic particulate deposition. Dependence on sediment texture was qualitatively established by measuring Se on particle-size separates. Normalization to Al showed that 65% of Se spatial variability is related to sediment texture. Selenium is further enriched in the marsh via post-depositional inputs, probably due to in situ adsorption from overlying water and chemical reduction. According to sediment flux measurements, enrichment in the marsh is equivalent to 20 to 25% of the particulate Se flux, thereby defining the marsh as a Se sink. These findings highlight the need for more intensive monitoring of SPM as the major source of Se to intertidal wetlands.

Abbreviations: BSP, Benicia State Park • MRP, Martinez Regional Park • OM, organic matter • SPM, suspended particulate matter


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
Sediment and Selenium Dynamics
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY
 REFERENCES
 
THE Carquinez Strait connects the Suisun Bay and the San Pablo Bay in the northern reach of the San Francisco Bay estuary, in northern California (Fig. 1). Historical releases of selenium (Se) in oil refinery effluent, as well as natural inputs of Se over geologic time (Presser and Ohlendorf, 1988), have resulted in elevated Se in local waters (Cutter, 1989), sediment (Johns et al., 1988), bivalves (Johns et al., 1988), diving ducks (California Department of Fish and Game, 1988), and nonmarine birds, such as the endangered Clapper rail (Lonzarich et al., 1992), in and around Carquinez Strait. The highest Se levels in bivalves have been found near the outfall point for a refinery in eastern San Pablo Bay (San Francisco Bay Regional Water Quality Control Board [SFBRWQCB], unpublished report, 1992). Similarly, areas in which the highest sediment Se levels have been observed are near oil refineries in the Carquinez Strait (SFBRWQCB, unpublished report, 1992). At present, oil refineries contribute as much as 75% of the total Se load flowing into the San Francisco Bay during low riverine flow periods, and approximately 50% of the Se load during high riverine flow periods (SFBRWQCB, unpublished report, 1992). Refineries that process crude oil from seleniferous geologic formations in the San Joaquin Valley produce effluent with Se concentrations about one order of magnitude higher than refineries using Alaskan North Slope crude (SFBRWQCB, unpublished report, 1992). Three such refineries currently operate along the shores of the Carquinez Strait.



View larger version (62K):
[in this window]
[in a new window]
  		Fig. 1. The San Francisco Estuary (central and northern) and location of field sites. Inset maps show the locations of sediment traps, sampling points, and vertical cross-sections of the Benicia State Park (BSP) and Martinez Regional Park (MRP) field sites.

 
Dissolved Se concentrations in San Francisco Bay are generally around 0.1 to 0.2 µg L-1 (Cutter, 1989). Aqueous Se speciation is dominated by Se(IV) and Se(VI), though organo-Se species may also be present (Cutter, 1989). Although the drinking water standard for Se is 5 µg L-1, and there have been no Se-related adverse effects observed to date, there are concerns that chronic exposure to elevated Se concentrations will result in bioaccumulation and eventual toxicosis, especially in birds and fish. There are two major Se bioaccumulation pathways: one through phytoplankton, where Se may be magnified as much as 100- to 1000-fold (Saiki and Lowe, 1987), and the other through bottom-feeders, such as bivalves, which ingest phytoplankton and sediment. Ecological guidelines for Se in suspended particulate matter (SPM) and bottom sediments are 1.5 mg kg-1 and 0.7 mg kg-1, respectively (SFBRWQCB, unpublished report, 1992), and have been exceeded on occasion in specific "hot-spots." Certain bivalve species contain up to 10 mg kg-1 Se compared with an ecological assessment guideline of 3.2 mg kg-1. Since bivalves ingest particulate matter, Se concentrations in sediment and on SPM need to be scrutinized (Luoma et al., 1992). The importance of understanding Se cycling and concentrations in shallow sediments is evident, given that bivalves, which are the primary food source for diving ducks, reside either at or just below the sediment–water interface. Furthermore, the success or failure of a plan to reduce Se in the biota by decreasing inputs of dissolved Se to the estuary may depend on the availability of sediment-bound Se. The exchange of trace elements between bottom sediments and the overlying water is a flux that is commonly measured in the subtidal environment (Santschi et al., 1990). It is a measure of the potential for sediments to either release or take up trace elements, thereby acting as either a source or sink of potential contaminants. Much less work has been done on such fluxes in intertidal zones. Due to the tidal flooding and drying of this habitat, tools such as benthic chambers and peepers are impractical.

The purpose of this study was to measure the relative contribution of particulate-bound and dissolved Se fluxes to intertidal wetlands. By examining factors that determine the magnitude and direction of Se fluxes, we aimed to define the intertidal wetlands as Se sinks or sources. The distribution of Se in two marsh–tidal flat systems was characterized and compared with Se concentrations on SPM, and with previously established background levels in San Francisco Bay sediments. The distributions of Al, Fe, organic matter (OM) content, and selected trace metals were also established. Selenium dependence on sediment texture was qualitatively established by measuring Se on particle-size separates. Sediment Se fluxes to the wetlands were measured directly using sediment traps. Because the study was conducted over a period of 15 mo, it does not provide a definitive measurement of long-term sedimentation or erosion rates. The complexities of sediment dynamics in both the salt marsh (Pethick, 1981; Stumpf, 1983) and on tidal flats (Pestrong, 1972) cannot be described by one year of sampling and measurement. However, seasonal trends in sediment flux and sediment-bound Se flux can provide insight into the relative contribution of SPM to Se distribution in these intertidal wetlands.


   Sediment and Selenium Dynamics
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Sediment and Selenium Dynamics
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY
 REFERENCES
 
The northern reach of the San Francisco Estuary receives most of its water and SPM from the Sacramento River (San Francisco Estuary Project, unpublished data, 1992), which also supplies the estuary with more than half of the mass of waterborne metals (Luoma and Phillips, 1988). The remaining metal mass comes from local point sources and urban runoff. The flow of water and SPM varies substantially, both seasonally and from year to year. During the 2-yr period from October 1995 to September 1997, the mean monthly flows through the Carquinez Strait varied from lows around 500 m3 s-1 in the late summer–early fall periods, to more than 2500 m3 s-1 in the winter (based on U.S. Geological Survey stream gauge measurements). According to long-term averages (1955–1990), approximately 55% of the sediment flowing into San Francisco Bay is retained within the Bay, with suspended sediment loads proportional to discharge (Ogden Beeman and Associates, unpublished data, 1992). Depositional patterns are complex and are dependent on bathymetry and flow velocities. Bay-wide average depositional rates do not represent depositional rates in nearshore areas, such as wetlands.

Sedimentation in intertidal wetlands is complicated by the seasonal resuspension of tidal flat sediments and their deposition on the marsh plain surface (Childers and Day, 1990). Since fine-textured suspended sediments tend to carry higher metal concentrations (Gambrell, 1994), it is expected that wetland areas dominated by clays and silts will also contain higher metal levels. The deposition of fine silt and clay in the marsh occurs due to two factors: (i) the relatively lower energy of the system, allowing physical settling and (ii) biological trapping by plants and filter feeders (Stumpf, 1983; Orson et al., 1992). In contrast, the tidal flats are dominated by silts and sands.

Dissolved Se retention in the marsh environment is anticipated due to the geochemical conditions prevalent in wetland soils. Selenium immobilization in the marsh can occur via several mechanisms. The most stable form, zero-valent Se [Se(0)], occurs under reducing conditions (Masscheleyn et al., 1991) but its oxidation kinetics are very slow (Zawislanski and Zavarin, 1996), leading to its long-term stability under suboxic and even oxic conditions. Research has shown that Se(0) forms primarily through microbial reduction of Se(IV) and Se(VI) (Oremland et al., 1990; Macy et al., 1989), though more recent work has shown abiotic Se reduction in the presence of Fe(II) (Myneni et al., 1997). The presence of Se(0) in both marsh and tidal flat sediments (Velinsky and Cutter, 1991; Zawislanski and Crawley, 1997) has been documented. However, the process of Se(0) formation has not been studied in situ and it is not clear whether the redox potential at the very surface of the marsh (i.e., at the sediment–water interface) is low enough to support Se-reducing microorganisms. Instead, less stable Fe-oxyhydroxides may control metal retention (Gambrell, 1994). Selenium can also be immobilized through the adsorption of Se(IV) onto oxide surfaces (Hamdy and Gissel-Nielsen, 1977), clay minerals (Bar-Yosef and Meek, 1987), and soil organic matter (SOM) (Yläranta, 1983). Selenium can be incorporated into the sediment system via biological uptake by rooted plants, algae, or benthic fauna (Masscheleyn and Patrick, 1993). Finally, Se can be lost from the system via microbial and plant volatilization (Duckart et al., 1992; Frankenberger and Karlson, 1994).


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Sediment and Selenium Dynamics
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 SUMMARY
 REFERENCES
 
Site Description
The study was conducted at two field sites, one in Martinez Regional Park (MRP), approx. 1.5 km downstream of the Shell Refinery in Martinez, and the other in Benicia State Park (BSP), approx. 4 km downstream of the Exxon Refinery in Benicia (Fig. 1). Each site is a brackish-water intertidal zone, with an extensive tidal flat and a marsh consisting of a saltgrass [Distichlis spicata (L.) Greene]–pickleweed (glasswort) (Salicornia spp.) dominated marsh plain and a bulrush (Scirpus spp.)–cattail (Typha latifolia L.)–cordgrass (Spartina spp.) dominated lower marsh. Due to the differences in elevation between the marsh and the tidal flat, these habitats are subjected to varying tidal influence. The tidal flats are submerged nearly 100% of the time at their lowest point and less than 50% of the time near the marsh–tidal flat interface. The marsh plain is submerged for as little as 10% of the time. Due to local relief and the presence of a thick plant cover, the marsh remains under standing water for much longer than would be calculated based on the tidal cycle (as long as 25% of the time in some local depressions). In addition, seasonal rainfall often results in increased flooding of the marsh. The MRP site is a relatively high-energy shoreline because of its exposed location on the Carquinez Strait and its proximity to shipping channels. Strong winds from the northwest as well as passing ships occasionally produce high-energy waves, up to 0.5 m in amplitude. In contrast, the BSP site is nestled within Southampton Bay, approx. 0.5 km from the main channel of the Carquinez Strait, resulting in a relatively low-energy shoreline.

Sample Collection and Preparation
Sampling of soil and sediment took place on 4 and 5 Dec. 1995 at MRP and BSP, respectively. Additional samples were collected at BSP in July of 1996. Samples were taken along two transects perpendicular to shore at each site (Fig. 1). Twenty-one locations were sampled at MRP, and 18 at BSP, with locations being generally 10 to 20 m apart. A 20-cm-deep core was collected at each sample point using a stainless steel corer. The location and elevation of each sample point was laser-surveyed. Samples were taken at low tide to enable sampling of the largest expanse of tidal flat possible. The tidal flat sediments at BSP have a slurry-like consistency, making it impossible to traverse them on foot. Therefore, samples of BSP tidal flat sediment were collected on 16 July 1996 from a boat, through approximately 0.5 m of overlying water. Cores were sampled through, and stored in butyrate liners, immediately sealed with plastic end-caps and tape, thereby minimizing exposure to air. Sample compaction was minor because of the small length of the core.

Samples were frozen in the laboratory and cut into thin horizontal slices: 0 to 1 cm, 1 to 2 cm, 2 to 3 cm, 3 to 5 cm, 5 to 7.5 cm, 7.5 to 10 cm, 10 to 15 cm, and 15 to 20 cm. Frozen sediment sections were thawed immediately prior to digestion to minimize exposure to air. The physical appearance of the sectioned core intervals was quickly described, after which each section was centrifuged (10000 rpm; 15 min) to remove any excess porewater. The sample was manually homogenized with a stainless steel blade. Part of the homogenized sample was split for digestion, while another subsample was used for a sequential extraction procedure. The remainder of the bulk sample was freeze-dried and stored for additional analyses.

Water and SPM in the Carquinez Strait were sampled approximately every 2 wk at the MRP site from a nearby pier (650 m east of site), where the water depth is approximately 2 to 4 m. Samples were collected by lowering an acid-rinsed polyethylene container just below the surface of the water, thereby avoiding touching the bottom to prevent sediment resuspension. At each collection, 21-L volumes were taken from the field to the laboratory where sediments were separated via centrifugation (12000 rpm; 15 min) and filtration (0.45-µm nitrocellulose filter), and subsequently freeze-dried. The SPM load was defined as the mass of particulate matter normalized to water volume.

Sedimentation Rate Measurement and Trapped Sediment Collection
Twelve erosion–sedimentation pins and sediment traps were installed at the MRP site and 11 at the BSP site. The pins and traps were installed along transects perpendicular to the shoreline, spanning all of the habitats of interest (Fig. 1), with the exception of the BSP tidal flat, which is inaccessible. At MRP, the sediment traps were within 2 to 5 m of Transect 2 and are compared with those concentrations only. At BSP, the traps were between 5 and 10 m from Transect 1, and data from both transects were used. The pins, constructed of 60-cm lengths of stainless steel rod, were used to measure local changes in surface elevation (Pestrong, 1965; Gabet, 1998). They were inserted into sediments so that 30 cm protruded above the ground surface. The sediment traps were static versions of the traps used by Lawson and Brockett (personal communication, 1993). These were 10- x 10-cm squares of plastic with long bolts at each corner to secure the trap onto the sediment surface. The top of each trap was roughened by gluing fine sediment onto it to approximate surface roughness. These were then installed flush with the ground surface. The pins and traps were monitored monthly, between February 1996 and June 1997. The exposed portion of the pins was measured to the nearest millimeter. Sediment from each trap was collected by rinsing its contents into a plastic freezer bag with distilled water, and was subsequently freeze-dried and weighed. The sediment was ground to a powder, acid-digested, and subsequently analyzed for total Se.

Particle Size Separation
Surface sediment samples collected from the MRP site were used to measure the concentrations of Se associated with different particle size fractions. Seven samples were collected at low tide, three from the marsh plain, two from the cordgrass zone of the lower marsh, and two from the tidal flat. Any visible plant detritus was cleared from the surface, after which sediment was collected over an area of 0.25 m2 to a depth of 1 cm, using a soil spatula. The samples were placed in plastic freezer bags and transported immediately to the laboratory. Each sample was roughly homogenized and subsampled. Plant pieces, mainly roots, which did not pass a 2-mm sieve, were removed from the bulk sample. The remainder of the main sample was then thoroughly homogenized and a particle size analysis was carried out using the hydrometer method (Gee and Bauder, 1986). After the measurement of the clay fraction (<=2 µm), water in each settling column was carefully siphoned off, without disturbing the coarser fraction, which had already settled out. The clay and the sand + silt fractions were then centrifuged and all of the fractions, including plant fragments and the bulk sediment, were freeze-dried, acid-digested, and analyzed for Se. Only minor losses of soluble Se during the settling process are expected, as previous work has shown that less than 5% of total Se in these sediments is soluble (Zawislanski and Crawley, 1997).

Digestion and Analytical Methods
A strong acid digest procedure (Zawislanski and Zavarin, 1996) was used to remove Se and trace metals. A dry, powdered (425-µm mesh) sample was digested using hot, concentrated HNO3 and 30% H2O2 for 24 h. The residue was then refluxed, using hot 6 M HCl, and washed several times with HCl. Supernatant solutions were passed through a 0.45-µm, nitrocellulose filter immediately after extraction. Sediment digests were analyzed for Se using hydride generation atomic absorption spectrometry (HG–AAS; Perkin Elmer [Norwalk, CT] Model 3030) (Weres et al., 1989). Carquinez Strait water samples were analyzed using either a cold vapor HG–AAS method (Cutter, 1978) or a lanthanum coprecipitation HG–AAS method developed by Tao and Hansen (1994). Cobalt, Ni, Cu, Zn, Mo, Cd, and Pb were analyzed using inductively coupled plasma–mass spectrometry (ICP–MS) (VG Elemental PQIII ICP–MS with S-Option [ThermoElemental, Cheshire, UK]). Although this digestion method was developed for Se, analyses of National Institute of Standards and Technology (NIST) standards (NIST 2709, San Joaquin Soil; NIST 1646, Estuarine Sediment; NIST 1646a, Estuarine Sediment) were used to establish the method's ability to remove trace metals. The results of total elemental analysis of NIST reference materials are shown in Table 1. Recovery ranged from 88 to 113% over a wide range of concentrations, with all but three values falling between 90 and 110%, exhibiting the ability of the strong acid procedure to effectively solubilize these elements. Iron, Mg, and Al were analyzed using meta-borate fusion, followed by inductively coupled plasma–atomic emission spectrometry (ICP–AES) (Chemex Labs, Reno, NV). Organic matter was calculated based on loss-on-ignition (1000°C furnace), using the regression derived by Howard and Howard (1990). Carbon was measured using a Carlo Erba (Milan, Italy) NA 1500 C and N analyzer, on dry, finely ground samples.


View this table:
[in this window]
[in a new window]
  		Table 1. Analysis of Se and trace metals in National Institute of Standards and Technology (NIST) reference materials.

 

   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Sediment and Selenium Dynamics
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
SUMMARY
 REFERENCES
 
Selenium in Carquinez Strait Water and Suspended Particulate Matter
No consistent time trends were observed in either dissolved or particulate Se concentrations in Carquinez Strait water sampled between December 1995 and January 1997. Dissolved Se concentrations fell in the range of 0.07 to 0.35 µg L-1. The SPM-Se concentrations averaged 0.93 mg kg-1, with a range of 0.60 to 1.59 mg kg-1. With the exception of the four highest concentrations (out of 23), these values overlap with data collected in this area by Cutter (1989) in April and September of 1986. When the SPM-Se concentrations are normalized to water volume, the range is equivalent to 0.01 to 0.38 µg L-1. On average, SPM-Se comprises approximately 25% of the total Se load.

Suspended particulate matter selenium is strongly correlated with C content (r2 = 0.81, n = 19), with C concentrations falling between 1.5 and 3.5%. Sample size limitations did not permit organic carbon (OC) determination for most SPM samples, but a comparison between C and OC measurements on a composite SPM sample showed close agreement (3.3 vs. 3.8%, respectively). The strong correlation between SPM-Se and SPM-C indicates the importance of the biogenic fraction of SPM to Se partitioning in the Carquinez Strait, though from this alone one cannot infer the presence of organo-Se compounds. Organic coatings on SPM are considered an important factor in their enrichment with respect to trace metals (Luoma, 1990) and Se (Cohen et al., 1992).

Elemental Distribution in Intertidal Sediments
Concentration profiles of Se, OM, Al, Fe, Cu, and Mo in the marsh plain, lower marsh, and tidal flat habitats are shown in Fig. 2, along with the mean (±1 SD) of elemental concentrations measured on SPM. Copper was chosen as representative of metals likely present in cationic form in the dissolved state. Molybdenum, like Se, is expected to occur primarily in anionic forms. Aluminum is indicative of the clay content of the sediments (Zwolsman et al., 1993).



View larger version (51K):
[in this window]
[in a new window]
  		Fig. 2. Depth profiles of Se, organic matter (OM), Al, Fe, Cu, and Mo, averaged by habitat, at the Martinez Regional Park (MRP, upper frames) and Benicia State Park (BSP, lower frames) sites. Error bars, shown on one side of the mean values for the purpose of clarity, are equal to 1 SD. Vertical lines and shaded areas represent the mean elemental concentration and ±1 SD on suspended particulate matter (SPM), respectively.

 
Selenium concentrations (Fig. 2a) are lowest in the tidal flats where a slight increase with depth is observed, with near-surface values ranging from 0.5 to 0.8 mg kg-1. Selenium concentrations increase landward, with highest values (averaging around 1.2 mg kg-1) observed in the MRP marsh plain. Differences among habitats are greater at MRP than BSP. Organic matter content (Fig. 2b) increases from the tidal flats to the marsh plain, where it can be as high as 30%. Organic matter trends are indicative of varying biomass production and organic carbon mineralization in the three habitats. There is prodigious carbon input in the marsh from emergent plants and their root systems, but no such input in the tidal flats. This is reflected by the OM content of marsh sediments vastly exceeding the OM content of SPM. Organic matter content decreases with depth in marsh sediments, probably due to carbon mineralization. It should be noted that OM is higher in BSP than MRP tidal flats, probably because of the much lower energy at the BSP site, allowing for the trapping of both finer-textured and organic particulate matter. Aluminum data (Fig. 2c) agrees with the textural description of the sediments, in that Al (and clay content) increase from the tidal flats to the marsh plain at site MRP, whereas those differences are poorly defined at BSP. Iron shows similar trends (Fig. 2d), though a sharp increase near the BSP marsh plain surface probably indicates redox-controlled Fe redistribution (Santschi et al., 1990). Copper concentrations (Fig. 2e) do not vary much with depth, except in the BSP marsh sediments, where an increase below 3 cm is observed. Molybdenum trends (Fig. 2f) resemble Se, which may be attributed to their similar reactivity in soil and sediment systems. Both elements, primarily present as anions in water, have been shown to strongly sorb onto Fe and Mn oxides, via ligand exchange (Hingston et al., 1972; Goldberg et al., 1996). More importantly, Se and Mo mobility is strongly dependent on their oxidation state, and both are less soluble under reducing conditions. Specifically, Mo(VI) and Mo(V) occur as highly soluble ions, whereas under reducing conditions, Mo(IV) forms insoluble phases (Brookins, 1988). The reduction of Mo by sulfate-reducing bacteria has been shown to occur in anaerobic environments (Tucker et al., 1997), apparently by processes similar to those that have been shown to immobilize Se (Oremland et al., 1990).

Selenium decreases in near-surface tidal flat sediments (Fig. 2a) may be indicative of diffusive losses of Se to overlying water. Note that these values are somewhat lower than those observed on SPM. Although there are no significant depth trends in marsh Se concentrations, OM and Fe profiles indicate early diagenetic changes in chemical parameters, which may affect Se distribution. Therefore, our attention is hereinafter focused on the top 1 cm of sediments, which is not likely to have been affected by diagenetic processes and represents most recent sediment deposition. Selenium in the top 1 cm of sediment is well correlated with OM (r = 0.820), Al (r = 0.806), and Fe (r = 0.853), as are most of the trace metals.

Selenium on Particle-Size Fractions
The high correlation between Se and Al is indicative of the anticipated dependence of Se concentrations on fine particle content. The distribution of Se on clay, sand + silt, and plant fractions of sediment collected from the MRP marsh and tidal flat is compared with their bulk concentration (Fig. 3). In tidal flat and cordgrass marsh sediments there are consistent differences in concentrations between the clay and sand + silt fractions. The clay fraction is enriched as much as twofold relative to the sand + silt fraction. Selenium concentrations on the clay fraction of the tidal flat and cordgrass marsh sediments are around 0.6 to 0.7 mg kg-1, in contrast with levels on the sand + silt fraction, which are 0.4 mg kg-1 or less. In the marsh plain, Se concentrations on both the clay and the sand + silt fraction increase to around 1.0 to 1.2 mg kg-1. The similarity in the concentrations on the two size fractions can be explained by the presence of organic coatings (Cohen et al., 1992), as well as mineral–organic aggregates (Zonta et al., 1994), which would settle out in the sand + silt fraction. The plant fraction contains between 0.55 and 0.85 mg kg-1 Se, less than the other fractions. Selenium concentrations in plant tissue should be interpreted cautiously, as the efficacy of cleaning clay- and colloid-size particles from plant fragments was not tested. In the first sample shown in Fig. 3, the bulk Se concentration significantly exceeds Se on the other fractions. In this case, the mass does not balance out. This may be due to errors in the fraction sampling method or analytical error. Overall, these results show that there are significant grain-size effects on Se concentrations, that Se is further enriched in the marsh plain, and that the presence of plant fragments and organic coatings may play a role in Se enrichment.



View larger version (63K):
[in this window]
[in a new window]
  		Fig. 3. Selenium concentrations on particle-size separates and plant fraction in surface sediments from the Martinez Regional Park. The average median particle size in each habitat is shown.

 
Selenium and Metals Normalized to Aluminum
Samples containing higher concentrations of Se and metals correlate qualitatively with finer-grained sediments (e.g., predominantly silty sands and sandy silts of the MRP tidal flat, vs. clayey silts of the MRP marsh plain). An inverse relationship between metal and Se concentrations and sediment size fractions is well documented (Singh et al., 1981; Takayanagi and Wong, 1983; Christensen et al., 1989; Luoma, 1990). Particularly high trace metal and Se levels are found associated with the clay and colloidal fractions (Takayanagi and Wong, 1983; Benoit et al., 1994). Aluminum concentrations are often used as an indication of clay content and therefore metal concentrations are commonly normalized to Al to account for variations in grain size (Salomons and Förstner, 1984; Luoma, 1990). Selenium in the top 1 cm of each core was normalized to Al and compared with the Se to Al ratio for SPM (Fig. 4). In addition, Se concentrations from a core from Richardson Bay (Hornberger et al., 1999; M.I. Hornberger, personal communication, 1999), which pre-dates major industrial inputs, were normalized to Al. This comparison shows that Se is enriched in the marsh beyond the levels supplied on SPM and beyond what may be considered natural background. Conversely, Se in the tidal flats falls in the range defined by SPM and Richardson Bay data. Based on the same approach, Cu was also higher in the marsh than in the tidal flats, but not beyond the Cu to Al ratio on SPM. Conversely, Mo was greatly enriched in the marsh relative to the tidal flat and to values on SPM, and exhibited trends similar to Se.



View larger version (57K):
[in this window]
[in a new window]
  		Fig. 4. Selenium normalized to Al in the top 1 cm sediment samples from both Martinez Regional Park (MRP) and Benicia State Park (BSP). Mean Se to Al ratios (±1 SD) on suspended particulate matter (SPM) and in Richardson Bay (RB) sediment cores are shown. Richardson Bay data represent pre-anthropogenic levels.

 
The Al-normalized values were further normalized to SPM. The resultant values, along with the OM fraction, are shown in Fig. 5. This representation further illustrates the enrichment of marsh sediments and depletion of tidal flat sediments with respect to Se and Mo, relative to concentrations supplied on SPM. The qualitative correlation of these Al-normalized trends to OM distribution suggests either that the enrichment of marsh sediments results from dissolved Se sorption and reduction onto organic surfaces, or that Se-enriched organic particulate matter is an important component of SPM settling out in the marsh. The distinction between these two mechanisms can be made by direct measurement of Se on settling particulate matter.



View larger version (28K):
[in this window]
[in a new window]
  		Fig. 5. Elemental to Al ratios in Martinez Regional Park (MRP) and Benicia State Park (BSP) surface sediments, normalized to the mean suspended particulate matter (SPM) elemental to Al ratio. The organic matter fraction is also shown for both sites.

 
Sediment Flux Estimates
The direct measurement of Se fluxes was made using a combination of sediment traps and erosion–sedimentation pins. Data from site MRP include the marsh and the tidal flat, whereas only the marsh was investigated at site BSP. The annual sediment flux for the period February 1996–February 1997 is shown in Fig. 6. Trap measurements were converted to a sedimentation rate based on surface sediment bulk density data (not shown). Sedimentation rates decreased from the tidal flats to the marsh, showing a positive correlation with tidal range (Harrison and Bloom, 1977). A comparison of sediment trap and erosion–sedimentation pin data illustrates the agreement between the two methods, which establishes the reliability of this approach. The limitations of using sediment traps to measure sediment flux in an area that experiences both erosion and sedimentation are apparent in parts of the MRP tidal flat, where net erosion of 1 to 2 cm was observed during this period. Aside from this difference, the trap- and pin-based sediment flux estimates are in good agreement. In the marsh plain, the pin data shows apparently higher sedimentation, but the tide-related swelling and shrinking of the marsh sediment observed in this environment makes the pin data less reliable. As much as 20% of the sediment trapped at -16 m at MRP consisted of plant debris that was difficult to separate from the bulk sample. At the BSP site, lower rates but similar patterns were observed, with the sedimentation highest in the lower marsh due to sediment entrapment. The combination of trap and pin data provides a semi-quantitatively consistent measurement of sediment fluxes. The annual sediment flux to the MRP and BSP marshes agrees well with previous findings from the San Francisco Bay (Pestrong, 1965; 0 to 2 cm yr-1), and other estuarine marshes (Pethick, 1981; Griffin et al., 1989; Orson et al., 1990), where most rates fell within 0.5 to 1.0 cm yr-1.



View larger version (26K):
[in this window]
[in a new window]
  		Fig. 6. Total annual sediment flux at the Martinez Regional Park (MRP) and Benicia State Park (BSP) sites, based on sedimentation traps and sedimentation–erosion pins (February 1996–February 1997). Pin data from BSP are not shown.

 
Selenium, Organic Matter, and Metals on Trapped Particulate Matter
Monthly measurements of Se, OM, and Al concentrations on trapped sediment at sites MRP and BSP were averaged seasonally, and are presented in Fig. 7 through 9, respectively. Selenium concentrations (Fig. 7) do not show much seasonal variability, except in the marsh plain where concentrations were consistently higher in the summer and fall of 1996 and lowest in spring 1997 for both sites. Selenium was consistently higher in the marsh plain than in the tidal flat. Similar spatial trends were observed in the distribution of OM and Al (Fig. 8 and 9). In particular, the Se and OM trends at site MRP are in good agreement. There are large seasonal fluctuations in OM levels on sediment trapped in the marsh, indicative of the biological component of settling particulate matter. However, aside from the trap at -16 m at MRP, very little macroscopic plant debris was trapped. Overall, Se is well correlated with OM (r = 0.806). This could be the result of the strong Se-C correlation on SPM. When data from the trap at -16 m is removed from the data set, there is good correlation between Se and Al (r = 0.765), similar to that found in surface core sediments. This suggests that both clay content and OM control Se concentrations on settling SPM.



View larger version (34K):
[in this window]
[in a new window]
  		Fig. 7. Seasonal average concentrations of Se on sediment trapped in the marsh and tidal flats of Martinez Regional Park (MRP) (a) and Benicia State Park (BSP) (b).

 


View larger version (34K):
[in this window]
[in a new window]
  		Fig. 9. Seasonal average concentrations of Al on sediment trapped in the marsh and tidal flats of Martinez Regional Park (MRP) (a) and Benicia State Park (BSP) (b).

 


View larger version (32K):
[in this window]
[in a new window]
  		Fig. 8. Seasonal average organic matter (OM) content on sediment trapped in the marsh and tidal flats of Martinez Regional Park (MRP) (a) and Benicia State Park (BSP) (b).

 
The lower winter concentrations in the marsh may be due to the resuspension and landward migration of coarser sediments from the tidal flats during storm events. During the low sedimentation months in the marsh (i.e. summer and fall), it is likely that coarse particles do not reach the marsh plain. Formation of bioflocculants on the BSP marsh plain in the summer may be a factor in the elevated Se and OM concentrations. The pelletization of clay and silt by suspension and bottom feeders is considered an important process in salt marshes (Frey and Basan, 1985), whereby some indigested organic material is mixed with residual mineral matter and excreted.

Selenium concentrations were weight-averaged over a period of 12 mo to account for seasonal variability in the mass of trapped sediment. The results are compared with concentrations found in the top 1 cm of sediment and with the mean Se concentration (±1 SD) on SPM (Fig. 10). In the MRP tidal flats, Se concentrations on trapped sediment were lower than on SPM but did not differ greatly from surface sediments. However, Se on sediments trapped in the marsh was lower than in the surface sediments. Most marsh sediments contain more than 1 mg kg-1 Se, while most trapped sediment and SPM-Se values fall below 1 mg kg-1. The average difference of 0.2 to 0.3 mg kg-1 is equivalent to about a 25% increase and indicates post-depositional enrichment. Trends were similar in the BSP marsh (Fig. 10b), where trapped sediments contained around 0.7 to 0.9 mg kg-1 Se, whereas concentrations in surface sediments were consistently higher, in the range of 0.9 to 1.3 mg kg-1. Therefore, the post-depositional enrichment in the BSP marsh may be even greater, an average of 35%. Since we were not able to collect trapped sediments from the BSP tidal flat, a comparison with surface sediments there is not possible.



View larger version (28K):
[in this window]
[in a new window]
  		Fig. 10. Weight-averaged annual mean Se concentrations in trapped sediment and in the top 1 cm of sediment at Martinez Regional Park (MRP) (a) and Benicia State Park (BSP) (b), compared with the mean annual suspended particulate matter (SPM)-Se concentration. The standard deviations represent seasonal variation.

 
Finally, trapped sediment Se data were normalized to Al and compared with Se to Al ratios on SPM (Fig. 11). The influence of high OM accumulated at -16 m at site MRP is apparent. Aside from this data point, most of the trapped Se data falls in the range observed on SPM. This is interpreted as evidence of the strong dependence of Se concentrations on SPM texture and suggests that the higher trapped Se values found on the marsh plain are related to very fine-textured sediment. Selenium has been found to be intimately associated with the colloidal component of SPM (Takayanagi and Wong, 1983), which may not settle out under most hydrodynamic conditions (Stumpf, 1983), but may do so in the upper reaches of the marsh plain. Senescence of lower marsh plants in the fall produces a blanket of plant debris, which has been shown to significantly contribute to the trapping of fine particulate matter (Orson et al., 1992).



View larger version (51K):
[in this window]
[in a new window]
  		Fig. 11. Weight-averaged annual mean Se concentrations in trapped sediment normalized to the mean Al concentration, and compared with mean Se to Al ratio (±1 SD) on suspended particulate matter (SPM) and in Richardson Bay (RB) sediment cores.

 

   SUMMARY
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Sediment and Selenium Dynamics
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY
REFERENCES
 
Selenium concentrations in sediments from two intertidal wetlands increase from the tidal flats to the marsh. Analysis of particle-size fractions shows a grain-size dependence of Se concentrations, with further enrichment in the marsh plain, probably due to the presence of macroscopic plant fragments and organic coatings. Although there is a marked influence of sediment texture on Se concentrations, Se is apparently enriched in marsh sediments, implying post-depositional fluxes due to reduction and/or adsorption, thereby defining the marsh as a Se sink. Selenium tidal flat profiles suggest a net loss at the sediment–water interface, possibly due to diffusion between porewater and tidal waters (Zawislanski and Crawley, 1997). However, changes in sedimentation patterns and/or actual decreases in Se concentrations on SPM over the last decade cannot be ruled out. Selenium is strongly correlated with OM, with which it is known to form complexes (Cohen et al., 1992). High OM in the marsh environment helps promote reducing conditions near the sediment surface and is likely to influence Se reduction and immobilization. The correlation between Se and Fe could be indicative of Se sorption onto Fe-hydroxides. Comparison of trapped sediment, SPM, and the top 1 cm of sediment leads to the conclusion that most of the Se mass flux is derived from settling particulate matter. There appears to be post-depositional enrichment of the marsh surface with respect to Se and trace metals. Approximately 25% of the Se in marsh sediments appears to be authigenic.

Volatilization by marsh plants and microbes is a potential flux of Se to the atmosphere (Duckart et al., 1992). However, in the few instances where Se volatilization rates have been measured in the field, the relative rates are low. Hansen et al. (1998) report that 10 to 30% of Se in a waste stream flowing through a treatment wetland was volatilized, but the losses they describe are relative to a soluble (20–30 µg L-1) Se inventory. Given the small fraction of dissolved Se at MRP and BSP (Zawislanski and Crawley, 1997), it is unlikely that methylation results in a significant Se flux.

The potential for sediment resuspension needs to be considered in the interpretation of Se fluxes to the marsh and especially the tidal flat. As is apparent from the annual sediment fluxes at site MRP (Fig. 6), the tidal flat is a highly dynamic environment, in which both sedimentation and erosion occur at different times of the year. Resuspension and subsequent trapping of resuspended sediment undoubtedly occur in the tidal flats throughout the year. However, the distinct and spatially consistent seasonal fluctuations in elemental concentrations, as seen in Fig. 7 through 9, are evidence of real changes in the composition of settling particles. The marsh surface is much more stable due to the presence of rooted plants and minor wave action. Good agreement between marsh sediment fluxes calculated based on trapped sediment mass and measured using sedimentation–erosion pins (Fig. 6) shows that little or no erosion occurred on the marsh plain. Therefore, the SPM trapped in the marsh is a reliable estimate of the net Se and metal flux in this environment.

Year-to-year changes in Se concentrations on SPM need to be considered as a potential source of trapped sediment variability. A long-term study of particulate-bound Se and metal deposition would be required. However, the vertical distribution of Se and most of the metals (Fig. 2) indicates consistent net inputs over recent years to decades.


   ACKNOWLEDGMENTS
 
The authors thank A.A. Brownfield, A.F. Haxo, S.J. Rodriguez, S.C.B. Myneni, D. King, J. Oldfather, and T. Alusi for assistance in the field and the laboratory. We are grateful to M.I. Hornberger for providing Se data from Richardson Bay. Comments made by M. Zavarin, T.K. Tokunaga, T.M. Johnson, and S.M. Benson substantially improved the manuscript. We acknowledge the support of the Department of Energy (A. Hartstein, E. Zuech, A.B. Crawley, and G. Walker) and the San Francisco Bay Regional Water Quality Control Board (M. Carlin, K. Taylor). This study was supported by the Assistant Secretary for the Fossil Energy, Office of Oil, Gas and Shale Technologies of the U. S. Department of Energy under Contract no. DEAC03-76SF00098.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 Sediment and Selenium Dynamics
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY
 REFERENCES
 
E.J. Gabet, current address: Dep. of Geology, Univ. of California, Santa Barbara, CA 93106. A.E. McGrath, current address: SECOR International, Oakland, CA 94612. H.C. Wong, current address: DepoMed, Foster City, CA 94404.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Sediment and Selenium Dynamics
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY
 REFERENCES
 


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2001 30: 677-682. [Full Text]  



This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
H. H. Zhang, Z. F. Wu, C. L. Yang, B. Xia, D. R. Xu, and H. X. Yuan
Spatial Distributions and Potential Risk Analysis of Total Soil Selenium in Guangdong Province, China
J. Environ. Qual., May 1, 2008; 37(3): 780 - 787.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (5)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Zawislanski, P.T.
Right arrow Articles by Wong, H.C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Zawislanski, P.T.
Right arrow Articles by Wong, H.C.
Agricola