Preliminary analyses of 0.6N HCl-extractable metals were conducted on 14 cores from a group of 273 gravity cores collected in 1990 and 1991. Detailed analyses were conducted on six of the 14 cores and, for comparison, on one core from Tomales Bay. Tomales Bay is 45 km north of San Francisco Bay and is neither industrialized nor urbanized, although there is a small, inactive mercury mine in its watershed. The six cores from San Francisco Bay included replicate cores from two major regions of the Bay. All six had bimodal depth profiles of HCl -extractable metal concentrations (higher concentrations in the near surface sediments than at depth) and were from locations of net sediment deposition, as indicated by differences in bathymetry between 1955 and 1990 (Ogden Beeman & Assoc., 1992)
Core RB92-3 was collected at the mouth of Richardson Bay, a 2-km wide, wind-protected embayment near the mouth of San Francisco Bay. A ship building facility was operational during World War II within Richardson Bay; otherwise, local sources of contaminants are minimal. Richardson Bay sediments should integrate contamination from both North and South Bay (Krone, 1979), but are probably also strongly influenced by marine processes (Conomos, 1979). Cores CB90-9 and CB90-12 were from Central Bay which is also strongly influenced by marine inputs (Conomos, 1979). In addition, these sites are approximately 6 km from the Alcatraz Island dredge spoil dumping site, where dredging waste has been deposited since 1975 (AHI and Williams Associates, 1990). Three cores from further landward in the estuary were analyzed in detail; two from San Pablo Bay (SP90-2 and SP90-8) and one from the Grizzly Bay arm of Suisun Bay (GB90-6). Together these are termed North Bay cores. Water chemistry and sediment transport in these locations are influenced by inputs from the San Joaquin and Sacramento Rivers (Conomos, 1979). Salinities less than 1 can occur during the highest river flows; low flow salinities are typically 17 to 24 in northern San Pablo Bay. Previous studies have documented enriched concentrations of Cr, Ni, V, Cu, and Cd in water and bivalves in Suisun and northern San Pablo Bay as a result of a variety of industrial discharges (Flegal et al., 1991; Abu-Saba and Flegal, 1995; Brown and Luoma, 1995).
Field and Sample Preparation and Analytical Procedures
All gravity cores were collected from the R/V David Johnston using a corer with a 363 kg weight sound. The cores were 9 cm in diameter and ranged from 0.5 - 2.5 meters in length. The core barrel was steel, with a polybutyrate liner. In addition to the 1990-91 sampling, a gravity core and box core were obtained from the mouth of Richardson Bay in August, 1992 (RB92-3) (Fuller et al., 1998). Comparison of isotope profiles (Fuller et al., 1998) and organic contaminant distributions (Venkatesan et al., 1998) between the box core and the surface sediments of the gravity core verified that no significant loss of surface materials or distortion of surface profiles occurred during gravity coring (Crusius and Anderson, 1991). The core from Tomales Bay was collected in 1993 using a diver-operated piston corer (Sansone et al., 1994).
After collection the cores were X-rayed, split into a working half and an archive half, and stored in a cold room (2-3°C) until sampling. Sand/silt ratio was determined on all samples. Sediment samples were wet-sieved using an acid-cleaned nylon-mesh screen into a tared 100 ml beaker to <64 µm in ultra-clean deionized water and dried at 70°C.
The <64 µm sediments were analyzed for metals in all cores (these are the data reported here, unless otherwise noted); bulk analyses were conducted on selected samples. Sieving effectively reduces the most important grain size biases that can affect comparisons (Salomons and Forstner, 1984; Luoma, 1990). Each sediment sample was homogenized using a mortar and pestle, split into 0.5 g replicate aliquots, and placed into a scintillation vial. For the weak-acid digest, two replicate 0.5 g sediment aliquots were digested at room temperature for two hours in 0.6 N HCl. The sample was filtered with a 0.45_m filter and analyzed by Inductively Coupled Argon Plasma Emission Spectroscopy (ICAPES). For near-total metal analyses, replicate sub-samples from each horizon and procedural blanks were digested using the concentrated nitric acid reflux method described by Luoma and Bryan (1981). Sediment aliquots of approximately 0.5 g were placed into 22 ml scintillation vials. Ten ml of concentrated trace metal grade nitric acid was added to each, a reflux bulb was placed on the vial, and the sample was left at room temperature overnight. Samples were then refluxed at 150°C for approximately one week, until clear. Reflux bulbs were removed and the samples were evaporated to dryness. The residue was reconstituted in 0.6 N trace metal grade hydrochloric acid, then filtered through 0.45 _m filters. Decomposition with concentrated nitric acid reflux is comparable with procedures previously employed on Bay sediments (San Francisco Bay Estuary Inst., 1994). It is indicative of metals sufficiently mobile to be of potential toxicological interest, but it has the disadvantage of not providing a complete dissolution of the sediment.
Total decomposition was conducted on a full suite of samples from RB92-3 and SP90-8, and on selected samples from CB90-12, in order to compare trends to those observed by near total decomposition. One ml of concentrated HClO4 and 2 ml of concentrated HF were added to sub-samples of 0.2 g, with selected replicates, in a Teflon vial. The samples were placed on an aluminum heat block preset at 110°C, and taken to dryness. One ml of HClO4 was added and then ultra-clean deionized water added to bring the Teflon vial to half full. Samples were returned to the hot plate for evaporation, cooled, and reconstituted to 10 ml in 0.6 N HCl. The vials were capped and heated at 90°C for 1 hour.
Samples for Hg analyses were reacted at 100°C in aqua regia followed by 10% nitric/dichromate reconstitution; 3% NaBH4 (in 1% NaOH) was added as a reductant before analysis by cold vapor AAAS.
Concentrations of Al, Cr, Cu, Fe, Mn, Ni, V, and Zn in the sediment were analyzed by ICAPES, after careful correction for peak interferences in the sediment digest matrix. Concentrations of Ag were analyzed by Graphite Furnace Atomic Absorption Spectroscopy (GFAAS) using Zeeman background correction with calibration by the method of standard additions. Lead concentrations were analyzed by flame AAS. The ICAPES was profiled and standardized according to normal operating procedures, then a quality control (QC) standard was run every 10-15 samples to ensure consistent performance of the instrument. Procedural blanks were analyzed as an unknown, but no blank subtraction was necessary. The instrument limit of detection (LOD) and limit of quantitation (LOQ) were determined by 10 or more analyses of a standard blank (0.6 N HCl) throughout each analytical run (Keith et al., 1983). All data reported here fall above the LOQ. If readings from replicate values of a solution were of low precision (relative standard deviation >10%), the readings were not used.
Recoveries from standard reference materials (SRM Sediment Standard 1646 and 2709) are reported. Because Pb analyses by ICAPES had an uncorrectable bias from Al, Pb (HNO3 digest) was analyzed by AAS. Recoveries of Pb from SRM 2709 were low in HNO3. As a second test of recoveries, Pb in selected horizons of core sediments were analyzed by both AAS (HNO3 digest) and isotope dilution by mass spectrometry of totally decomposed sediments. These two methods compared within 5% in both uncontaminated and contaminated horizons, suggesting a high fraction of Pb recovery in San Francisco Bay sediment.
Analyses of 137Cs, 210Pb, 239,240Pu, and 234Th were conducted on cores RB92-3 and SP90-8 to derive chronologies (Fuller et al, 1998). Additional analyses of 10Be constrained early human activities on each core and 14C was used to identify the oldest sediments in RB92-3 (van Geen et al., 1998). Sediments in RB92-3 appeared to be continuously deposited since well before significant anthropogenic activity began in the watershed (van Geen et al., 1998). The linear sedimentation rate at the surface of RB92-3 was 0.89 cm/yr and the core was vertically mixed to 33 cm depth (Fuller et al., 1998). Dates of sediment deposition were determined by numerical simulation of 210Pb profiles. The dates on individual horizons are the minimum age of sediments at that depth. The deposition rate in SP90-8 averaged 4.1 cm/yr based on 137Cs and 239,240Pu activity maxima and 210Pb profiles. Profiles of 137Cs were also determined in three additional cores, in order to estimate the depth of sediment deposition since 1952 ± 2 (method described by Fuller et al., 1998).
Jaffe et al. (1998) reconstructed depositional processes at SP90-8 by comparing five detailed bathymetry surveys conducted since 1850. A discontinuity in chemical concentrations observed at ~120 cm depth in this core appears to coincide with a depositional hiatus that extended from 1880 to 1950. Both depositional history and Sr/Nd isotopic signatures suggest sediments between 150-250 cm originated from hydraulic mining activities and were deposited between 1850 and 1880 (Jaffe et al., 1998; Bouse et al., 1996). Thus, sediments deposited before 1850 lie directly beneath sediments deposited in the 1950's.
Inventories of ‘excess’ metal (mass per area deposited in excess of baseline) were determined in GB90-6, SP90-8, and RB92-3 by integrating contaminant metal inputs downcore. The µg/cm2 of metal in interval a (Ma) was determined by
Ma = (Ca-B) * pz * _z (1)
where Ca is the metal concentration in µg/g in interval a, B is the baseline metal concentration that occurred before anthropogenic activities began in the watershed (see later discussion), pz is the bulk density of the sediment, and z is interval thickness in cm. Fuller et al. (1998) reported pz for each horizon in RB92-3; an average bulk density of 1.1 g/cm2 was used for SP90-8. The integrated inventory for sediments (_) was determined by
for all intervals. Intervals not sampled were assigned values by linear interpolation of the concentration (mass of metal/g) of adjacent intervals. Long-term mean excess metal flux, _, was compared among cores with different sedimentation rates. To do so, _ was divided by the number of years of excess metal input (e.g., normalized to the period of human disturbance) using age estimates from Fuller et al. (1998).