Introduction
Following
five years of RMP bioaccumulation monitoring, we determined that
a more in-depth analysis of the database might enable us to assess
how well this monitoring component met its original goals, and how
it might evolve in the next few years to meet the new RMP objectives
and help answer relevant management questions (see Chapter
2: Review Implementation). The purpose of this article is to
continue synthesis of the growing bioaccumulation database in order
to stimulate discussion for design improvements, including those
related to monitoring contaminant effects (a new RMP objective).
In 1998, the Steering Committee decided to modify the monitoring
objectives to include a description of general sources and loadings
of contamination to the Estuary and measurements of contaminant
effects on selected parts of the Estuary ecosystem.
The
initial goals of the RMP bivalve monitoring component were to:
1.
Measure the bioavailable portion of contaminants in the water
column.
2.
Evaluate which contaminants may be transferred to higher trophic
levels of the food web, and thus to what extent certain contaminants
may pose health risks to wildlife and humans.
These
general goals implicitly address the overall original RMP objectives
of determining seasonal and long-term trends in chemical and biological
water quality. Unlike the "snapshot" in time of contamination
obtained from water sampling three times each year, the bioaccumulation
component provides an integrative measure of water contamination,
since exposure to varying concentrations during the three-month
deployment is reflected in their tissues. Also, measuring dissolved
and total/near-total contaminant concentrations in water and sediment
alone does not reveal how likely it is for various contaminants
to enter the food web and pose risks to higher-order consumers.
Bivalves are very good trend indicators for many contaminants, particularly
lipophilic compounds such as chlorinated hydrocarbons and PAHs,
because their contaminant body burdens equilibrate with corresponding
contaminants in the surrounding environment relatively quickly
(Russell and Gobas, 1989; Stephenson, 1992). However, not all
contaminants are bioaccumulated in the same way by bivalves, and
bivalve species differ in their bioaccumulation characteristics.
Overall, oysters accumulate trace metals to a greater degree than
mussels and clams, while mussels accumulate PCBs to higher concentrations
than oysters and clams. We have also learned that bivalves are unsuitable
indicators for mercury bioaccumulation, although we know that methylmercury
is highly accumulative in other species and is rapidly magnified
in the food web, as evidenced in fish tissue levels that are of
human health concern (see Chapter 6: Pilot and
Special Studies). Similarly, bivalves do not appear to bioaccumulate
arsenic and are likely of limited use in determining bioavailability
of this contaminant, for trend monitoring, or as a diagnostic tool
for identifying potential problem areas.
Time
series of raw trace substance concentrations in bivalves (not normalized
for tissue lipid content) for the last ten sampling events starting
in 1993 (with the exception of Corbicula fluminea) are depicted
in Figures 5.19a, b,
c, d,
& e, and 5.20a,
& b in Bivalve Monitoring
Trends. As with water and sediment concentration trends, numerous
environmental variables influence bivalve concentrations. In most
cases, the raw data essentially show no trends, and the noise surrounding
the signal of interest (tissue concentrations over time) is so large
that any changes over time or spatial patterns require many years
of measurements before any definitive conclusions can be drawn.
Gunther
et al. (in press) and Gunther and Davis (1997) analyzed bivalve
data by combining the databases of the RMP and the State Mussel
Watch Program, thus increasing the size of the data set. They found
statistically significant declines in silver in both Central and
South Bay reaches, and less pronounced declines in mercury and lead
concentrations. They also demonstrated that lipid normalization
of chlorinated hydrocarbon concentrations in bivalves reveals patterns
that otherwise may not be apparent. The combined databases normalized
to tissue lipid content show dramatic initial declines in concentrations
of chlorinated hydrocarbons, such as PCBs and DDTs after use restrictions
were implemented. When 1997 data are added to the trend lines at
Coyote Creek (BA10), Yerba Buena Island (BC10), and San Pablo Bay
(BD20), unnormalized bivalve PCB concentrations show consistent
declines at all stations between 1994 and 1997, but lipid-normalized
PCB concentrations indicate a decline only at Yerba Buena Island
(Figure 5.21). We have expanded
this type of analysis to explore how water quality parameters, such
as temperature, salinity, dissolved oxygen, suspended sediment,
and chlorophyll a concentrations, might affect tissue concentrations
and bivalve condition.
Bivalves
as Tools for Meeting New RMP Objectives
As
part of designing the RMP so it can answer the management questions
formulated in 1998 (see Chapter 2: Review
Implementation), we examined the possible role of biomonitoring
with bivalves in meeting the new RMP objectives and answering
some of the management questions. Bivalve measurements can serve
more purposes than this RMP element was originally designed for.
They have the potential or have been shown to contribute to the
following assessments:
1. Bivalve tissue concentrations are probably the most suitable
indicator for long-term trends of many contaminants in the Estuary.
2. Bivalves are suitable as a diagnostic tool for problem identification
and prioritization of follow-up action, and for the identification
of most bioaccumulative substances.
3. Studies by the U.S. Geological Survey and others (Luoma and
Linville, 1996; Salazar and Salazar, 1995, 1998) have shown the
potential of bivalves as indicators of pollutant effects.
4. Bivalve tissue concentrations can represent a "substitute"
or enhancement of water measurements, since they integrate water
concentrations over long periods of time.
5. They represent a tool to estimate contaminant transfer to higher
trophic levels to be used by others for ecosystem risk assessments.
6. Bivalves can serve as a tool for prioritizing problem watersheds
or sites that may contribute contaminants of concern to the Estuary
(pollutant source/pathway indicator).
If
the potential of bivalves in meeting these goals is to be recognized
and evaluated for incorporation into the new RMP design, the kinds
of analyses summarized in this article are a necessary first step.
Data
Analysis
The
analyses described in this article proceeded in three phases. First,
we determined what quantitative relationships exist between bivalve
data (i.e., trace substance concentrations and indicators of bivalve
health) and key environmental factors. These quantitative relationships
were then used to statistically adjust the bivalve data to remove
the suggested effects of these environmental factors. This enabled
us to determine the magnitude of the noise surrounding the signal
that the environmental factors are likely to contribute. Second,
we examined in more detail whether there were statistically significant
spatial and temporal trends where the same species was deployed
and whether these trends were affected by adjusting the bivalve
data for the suggested effects of the environmental factors. This
analysis was performed to determine whether there were significant
trends and whether the trends were more or less apparent after the
data had been adjusted. Third, we compared bivalve concentrations
of trace substances, with and without adjustment for the suggested
effects of environmental factors, with water concentrations in the
particulate and dissolved fractions. The purpose of this analysis
was to demonstrate the value that bivalve measurements add to the
RMP, and to determine whether adjustment of bivalve data improved
this value.
The
initial step was to determine whether bivalve measurements may be
affected by natural water quality parameters in ways that confound
our ability to describe spatial and temporal trends in bioavailable
contaminants. The influence of various water quality parameters
on invertebrate bioaccumulation has been demonstrated in numerous
studies (Absil et al., 1994; Hutchins et al., 1996; Luoma and Bryan,
1982; Magni, 1993; Wang et al., 1995; Wright and Zamuda, 1987).
Although the bivalve bioaccumulation method has been used worldwide
to determine spatial and temporal variation in contaminants, RMP
data and other studies have shown that the San Francisco Estuary
provides unique challenges because of the very high spatial and
temporal variation in natural water quality parameters.
Statistical
analyses were performed to determine whether chlorophyll, dissolved
oxygen, salinity, temperature, and total suspended solids might
be affecting the bivalves and their accumulation of trace substances.
U.S. Geological Survey (USGS) water quality data, collected as part
of the RMP base program and supported by both RMP and Department
of Interior funds, are recorded on approximately monthly intervals
(http://sfbay.wr.usgs.gov/access/wqdata/archive). We obtained these
data for stations near seven RMP bivalve sites (Figure
5.22). The USGS data from the three 1-m intervals that bracketed
our bivalve deployment depths were averaged across all the USGS
cruises that occurred during each bivalve deployment period to estimate
the conditions experienced by the bivalves for that site and deployment.
The
potential effects of water quality parameters on bivalves were examined
for four indicators of bivalve health (condition, tissue growth,
percent tissue lipid, and survival) and the tissue concentrations
of trace metals and selected organic contaminants (totals for PAHs,
PCBs, DDTs, chlordanes, and HCHs). It should be noted that, based
on initial analysis by Gunther and Davis (1997), all trace organic
contaminants were normalized to lipid concentrations prior to these
analyses. Oysters (Crassostrea gigas) were examined at Coyote
Creek and Davis Point; mussels (Mytilus californianus) were examined
at Redwood Creek, Alameda, Red Rock, and Pinole Point; and clams
(Corbicula fluminea) were examined at Sacramento River (Figure
5.22).
The
statistical procedures involved backward stepwise regressions using
the water quality parameters as independent variables and the bivalve
parameters as dependent variables. These procedures enable determination
of which independent variables account for most of the variation
in each dependent variable. The backward stepwise procedure initially
begins with all of the independent variables included in the analysis,
and the variables that account for the least variation are successively
removed at each step until the remaining independent variable(s)
account for most of the remaining variation in the dependent variable.
The resulting regression coefficient approximates the percentage
of the variation in the dependent variable that is due to the independent
variables. The probability (P) indicates whether the resulting regression
line for the relationship between the independent and dependent
variables is statistically significantly different from zero (i.e.,
P < 0.05 is significantly different from zero). Because data
were occasionally missing for some water quality parameters, whenever
the stepwise procedures found a slope significantly different from
zero, multiple regression was performed using only the important
independent variables. The residuals from the multiple regressions
(i.e., the distance of each data point from the regression line)
were used to correct the dependent variables (i.e., bivalve data)
for the effects of the water quality parameters using the method
of Hebert and Keenleyside (1995).
Following
the application of any appropriate corrections to the bivalve measurements,
we examined in more detail bivalve concentrations of copper, mercury,
PAHs, and determined whether spatial and temporal trends were statistically
significant. These four trace substances were selected because of
their regulatory importance. Regressions of bivalve contaminant
concentrations against time were tested to determine whether temporal
trends were significant and whether trends differed among sites.
Analyses of variance (ANOVA) were performed for the aggregate of
all sites within each bivalve species to determine whether overall
differences among years were significant. ANOVA was also performed
to determine whether sites with the same species differed.
Caution
must be used when interpreting the results of the regression analyses.
Regression analyses assume the independent variables (i.e., chlorophyll,
dissolved oxygen, salinity, temperature, total suspended solids,
and time) affect the dependent variables (i.e., bivalve health and
trace substance concentrations). While we have used the regression
analyses to establish whether there are systematic relationships
or correlations between the independent and dependent variables,
true cause and effect relationships can only be confirmed through
experimentation. Regression analyses were more advantageous for
our purposes than calculations of simple correlations because the
resulting regression equations provide the means to adjust the bivalve
data for the suggested effects of environmental factors.
Effects
of Water Quality Parameters on Bioaccumulation
Numerous
bivalve measurements are significantly related to chlorophyll, dissolved
oxygen, salinity, temperature, and total suspended solids (Tables
1, 2, and 3
in Appendix E). Thirteen out of 18
bivalve measurements were significantly related to these water quality
parameters for Mytiluscalifornianus (Table
1 in Appendix E), 11 out of 18
were significantly related for Crassostrea gigas (Table
2 in Appendix E), and five out of
18 were significantly related for Corbicula fluminea (Table
3 in Appendix E). This finding in
and of itself is not surprising and was expected.
Mussels
Health and Survival
Condition,
tissue growth, percent lipid, and survival of M. californianus
were all significantly positively related to various combinations
of chlorophyll, dissolved oxygen, and salinity. The suggested
effects of dissolved oxygen on condition, tissue growth, and percent
lipid are consistent with the super-saturated dissolved oxygen
concentrations prevalent in the surf zone where these bivalves
naturally live. The sharp decline in survival below salinities
of 1820 parts per thousand (Figure
5.23) was fit best with a second-order polynomial regression.
This relationship between survival and salinity is also consistent
with the open coast habitat of this species.
Bioaccumulation
All
the water quality parameters, either singly or in combination,
were significantly related to bioaccumulation of silver, cadmium,
lead, nickel, zinc, PAHs, PCBs, chlordanes, and HCHs. Only chlorophyll
and temperature were consistent regarding the direction of their
effects, with bioaccumulation of cadmium and zinc being negatively
related to chlorophyll, and bioaccumulation of silver, lead, and
zinc being positively related to temperature.
Oysters
Health and Survival
Condition
and tissue growth were negatively related to temperature, with
the negative relationship between survival and temperature also
being nearly significant. Percent lipid was positively related
to dissolved oxygen, salinity, and total suspended solids.
Bioaccumulation
All
the water quality parameters, either singly or in combination,
also were significantly related to bioaccumulation in oysters.
Only dissolved oxygen and temperature were consistent in the direction
of their suggested effects, with bioaccumulation of cadmium, PAHs,
and PCBs being negatively related to dissolved oxygen, and bioaccumulation
of chromium and lead being positively related to temperature.
Clams
Health and Survival
Condition
and survival were negatively related to temperature and chlorophyll,
respectively, while condition was positively related to salinity.
Bioaccumulation
Only
three contaminants, silver, PCBs, and HCHs, were significantly
related to water quality parameters, probably because of the low
sample size related to using data from a single site. All three
contaminants were negatively related to chlorophyll. Silver and
HCHs were also negatively related to total suspended solids and
temperature, respectively.
These
findings confirm the common wisdom that water quality variables
influence bivalve parameters, although this is the first time that
RMP data were subjected to this kind of analysis. It is now possible
to determine whether the adjustments to bivalve data for the suggested
effects of water quality variables reveal spatial or temporal trends
that are not apparent using unadjusted data. These analyses also
reveal that some water quality parameters in the Estuary are outside
optimum levels for the bivalves and may thus affect bioaccumulation.
For example, dissolved oxygen concentrations in the Estuary seem
to affect the "health" (as defined by tissue growth, condition,
and percent lipid) of Mytilus californianus. This species also survived
poorly where salinities averaged less than 20. Summer temperatures
in the Estuary also may exceed those that are optimal for Crassostrea
gigas. This is not to say that these bivalves are inappropriate
for bioaccumulation monitoring in the Estuary, but that the ultimate
data users need to be clear regarding the limitations of these indicators
and the uncertainties surrounding the data. Although these transplanted
bivalves experience environmental stress in the Estuary, resident
bivalves may also experience stress at certain times of the year.
Nevertheless, drawing conclusions about the absolute biomagnification
potential of a trace substance based solely on transplanted bivalves
may not be appropriate, and other bivalve species that are better
adapted to Estuary conditions may be more suitable for contaminant
transfer estimates.
Spatial
and Temporal Trends and the Effects of Analyzing Adjusted Bivalve
Data
Numerous
spatial and temporal trends occurred for copper, mercury, PAHs,
and PCBs in the three species of bivalves. If significant regressions
were not found between tissue contaminant concentrations and the
natural variables (Tables 1, 2,
and 3 in Appendix
E), trends are described for unadjusted data (i.e., measured
concentrations of metals and lipid-normalized concentrations of
organic contaminants). But, whenever possible, data are used that
have been adjusted for the suggested effects of the natural variables.
Mussels
ANOVA
results indicated relatively little spatial variation in the bioaccumulation
of copper, mercury, PAHs, and PCBs (Table
5.1). Only PCBs indicated a significant difference (Table
5.1), with mussels deployed at Redwood Creek bioaccumulating
greater amounts of PCBs than did mussels deployed at Pinole Point
or Red Rock. This is in agreement with previous conclusions drawn
from sediment and water data comparing Estuary reaches, with the
South Bay exhibiting higher PCB concentrations than the Central
Bay reach.
ANOVA
results also indicated relatively little temporal variation in the
bioaccumulation of copper, mercury, PAHs, and PCBs (Table
5.1). Tissue concentrations of copper were significantly greater
in 1996 and 1997 than in 1993 and 1994, suggesting increases through
time. Unadjusted PCBs were significantly lower in 1995 and 1997
than in 1994. There were no significant differences among years
for mercury or PAHs.
Regression
analyses using unadjusted data revealed that significant temporal
trends were site-specific (Figures
5.24, 5.25, 5.26
& 5.27). The increase of copper
through time was significant at Pinole Point and nearly significant
at Redwood Creek, but not at Alameda or Red Rock. The very slight
decline in mercury through time was nearly significant at Alameda,
but not at any other site. Increases in PAH concentrations were
nearly significant at Red Rock and Redwood Creek, but not at Alameda
or Pinole Point. Decreases in PCB concentrations were nearly significant
at Alameda and Redwood Creek, but the decreases at Red Rock and
Pinole Point were much less pronounced and insignificant.
Adjustment
of tissue concentrations of PAHs and PCBs for suggested effects
of environmental variables provided contrasting results (Figures
5.28 and 5.29). In the case
of PAHs, adjustment of tissue data made no difference in the ANOVA
results, except that with unadjusted data, 1994 had the lowest mean,
and with adjusted data, 1995 had the lowest mean (Table
5.1). Probabilities were also lower with adjusted data indicating
reduced variation within sites and years. Use of adjusted data in
the analysis of temporal trends indicated much less dramatic increases
through time at each site than were seen with the unadjusted data
(Figure 5.26), suggesting that
the increases seen in the unadjusted data may be related to differences
in dissolved oxygen and total suspended solids. P values for trend
lines based on adjusted data were substantially greater, indicating
that by adjusting PAH tissue concentrations, any hint of increases
through time was even less pronounced than for unadjusted data (Figures
5.26 and 5.28; Table
5.2). In the case of PCBs, adjustment of tissue data made no
difference in the ANOVA test for differences among sites, although
the probabilities were lower (Table
5.1). The ANOVA test for differences among years gave different
results for adjusted and unadjusted data, with adjusted data suggesting
more consistent decreases from year to year. Use of adjusted PCB
data in the analysis of temporal trends suggested decreases through
time that were more significant than for unadjusted data (Figures
5.27 and 5.29; Table
5.2).
Oysters
Unlike
for mussels, ANOVAs for oysters revealed no significant differences
among sites or years for any of the four contaminants (Table
5.3). Unadjusted tissue data also revealed no significant trends
through time (Figures 5.30, 5.31,
5.32, & 5.33).
Adjustment
of oyster tissue data changed the slope of some trend lines, but
all remained insignificantly different from zero (Figures
5.34, 5.35, 5.36,
& 5.37). For instance, the
insignificant decrease in copper at Coyote Creek for unadjusted
data became an insignificant increase with adjusted data (Figures
5.30 and 5.34; Table
5.2) and the insignificant increase in PAHs at Davis Point for
unadjusted data became an insignificant decrease with adjusted data
(Figures 5.32 and 5.36;
Table 5.2). Use of adjusted data
caused the increase and decrease in mercury at Coyote Creek and
Davis Point, respectively, to become less insignificant in each
case. The use of adjusted data for PCBs caused temporal trend lines
to be slightly less insignificant.
Clams
Because
only the Sacramento River site was used in the statistical tests
for clams, it is not possible to evaluate spatial variation, although
there were indications of temporal variation. Although ANOVA results
revealed no significant differences among years (Table
5.4), trendlines based on unadjusted data indicated significant
increases in copper, nearly significant increases in mercury, insignificant
increases in PAHs, and significant decreases in PCBs (Figures
5.38, 5.39, 5.40,
& 5.41). Adjustment of PCB
data for the suggested effects of environmental variables made the
decrease in PCBs insignificant (Figure
5.42 and Table 5.2).
We
can conclude from the ANOVAs and trendlines that there are spatial
and temporal differences in trace substance accumulation by bivalves
in the Estuary. For example, there were higher concentrations of
PCBs in mussels from South Bay sites, which are consistent with
RMP water and sediment data. The decreases in PCBs at most sites,
although generally not significant in unadjusted data, are also
consistent with previous findings (Gunther et al., in press).
The increases in copper also appear to be regional because they
were evident at every mussel site and the clam site. The absence
of significant spatial and temporal differences in the oysters indicates
that either there is a high degree of spatial variation between
the oyster sites and nearby mussel sites, or the bioaccumulation
trends are species-specific.
Adjustment
of bivalve data for suggested effects of environmental variables
often reduces variation in the data and improves our ability to
detect spatial and temporal trends. In eight of the 13 cases in
which ANOVAs were performed on both adjusted and unadjusted data,
the adjusted data had lower probabilities (i.e., the results were
either more significant or less insignificant; Table
5.1, 5.3, and 5.4).
Analysis
of bivalve data that have been adjusted for the suggested effects
of environmental variables may also lead to different conclusions
than would be drawn from analyzing raw, unadjusted data. For example,
the slopes for copper and PAH trendlines in oysters at Coyote Creek
and Davis Point, respectively, differed between adjusted and unadjusted
data. Also, the greater significance of decreases in mussel PCBs
and the disappearance of significance in decreases in clam PCBs
after adjusting the tissue data could lead to different conclusions
regarding temporal trends in trace substances in the Estuary. Such
conclusions have important ramifications in the assessment of the
health of the Estuary and the evaluation of regulatory requirements.
Nevertheless, only in the cases of copper and PAHs at Coyote Creek
and Davis Point, respectively, would conclusions about the direction
of trends be affected.
Comparisons
of Bivalve Bioaccumulation and Water Contaminant Concentrations
Backward
stepwise regressions were performed to determine how well
the concentrations of contaminants in bivalves tracked concentrations
of dissolved and particulate water contaminants. More specifically,
the data were analyzed to shed light on the following questions:
1.
Do the data from one or two water measurements during a bivalve
deployment account for significant variation in the bivalve data?
In other words, are high or low water concentrations reflected
by corresponding bivalve tissue concentrations?
2.
Do the adjustments to tissue concentrations for suggested effects
of water quality parameters improve the correspondence between
tissue measurements and water measurements?
3. Do the bivalves consistently bioaccumulate certain contaminants
from either the dissolved or particulate fractions?
Mussels
The
tissue concentrations of very few contaminants were significantly
related to either dissolved or particulate water fractions (Table
4 in Appendix E). The measured
tissue concentrations of copper, mercury, and zinc were negatively
related to particulate or dissolved fractions and five of the remaining
20 possible regressions indicated non-significant negative regressions,
suggesting no effect of water measurements on the tissue measurements.
Adjustment of tissue data for the effects of water quality parameters
actually reduced the correspondence between water measurements and
tissue measurements for silver and chlordane, although adjustment
did provide a significant regression between tissue and the dissolved
water fraction for chlordane. Adjusted tissue concentrations improved
the correspondence between tissue and water PCBs, with a significant
regression for the dissolved fraction.
No
consistent relationship existed between either the adjusted or unadjusted
tissue trace substances and the dissolved or particulate water fractions,
although the dissolved fraction appeared more important. Eleven
out of 23 possible regressions indicated either significant or non-significant
positive correlations with the dissolved fraction, and four indicated
significant or non-significant positive correlations with the particulate
fraction. The remaining regressions indicated negative relationships
with either dissolved or particulate fractions.
Oysters
Oysters
were similar to mussels in the paucity of significant regressions
between tissue concentrations and either water fraction (Table
5 in Appendix E). Only four contaminants
had significant regressions for either adjusted or unadjusted tissue
contaminants. The tissue concentrations of both copper and mercury
exhibited negative correlations with the particulate fraction. Six
of the remaining tissue trace substances indicated non-significant
negative correlations with either dissolved or particulate fractions.
Adjustment of tissue data for suggested effects of environmental
variables decreased the correspondence to water measurements for
PCBs, although both adjusted and unadjusted tissue data indicated
positive correlations with dissolved water concentrations. None
of the four cases of significant regressions indicated improved
correspondence between water measurements and adjusted tissue concentrations.
Neither
the dissolved nor the particulate fractions were predominant in
their suggested effects on tissue concentrations. Three tissue trace
substances indicated significant or non-significant positive correlations
with the dissolved fraction and two indicated significant or non-significant
positive correlations with the particulate fraction.
Clams
There
were only four trace substances in clam tissues (mercury, selenium,
PAH, PCB) that were significantly correlated with either dissolved
or particulate water fractions on tissue concentrations (Table
6 in Appendix E), and one of them
(selenium) was negatively correlated with the dissolved. Four tissue
trace substances indicated non-significant negative correlations
with either dissolved or particulate fractions. Adjustment of tissue
data for suggested effects of environmental variables improved the
correspondence to water measurements for mercury and PCB, although
the significant regression for adjusted PCB included positive effects
of the particulate fraction and negative effects of the dissolved
fraction.
Unlike
with the mussels and oysters, tissue trace substances were more
often positively correlated with the particulate fraction than with
the dissolved fraction. Seven tissue trace substances had either
significantly or non-significantly positive correlations with the
particulate fraction and only two had significantly or non-significantly
positive correlations with the dissolved fraction.
Findings
and Conclusions
1. Bivalves are effective tools for monitoring long-term trends,
especially for bioaccumulative trace organics.
2. Bivalves are of limited use in monitoring trends for those
trace elements that do not accumulate in tissues, as integrators
of water contamination for mercury and arsenic, or for estimating
mercury transfer to higher levels of the food web.
3. The comparisons of tissue and corresponding water concentrations
reveal that time-integrated bioaccumulation of contaminants by
bivalves generally does not correspond well to water measurements
of contaminants made on one or two occasions during bivalve deployments.
Although this conclusion is not necessarily surprising, it indicates
that bivalves are important sampling devices and add information
that water or sediment data alone would not supply.
4. Bivalves are but one of many tools to determine the transfer
and potential magnification of contaminants to higher trophic
levels. The current use of non-resident species appears suboptimal
in this regard.
5. The bivalve data indicate spatial and temporal trends in contaminants
that have important implications for management of the Estuary.
Although PCB tissue concentrations seem to be decreasing at some
stations, overall Estuary trends are not yet clear. For PCBs,
the removal of natural environmental variables that may influence
tissue trace substance data may reveal different patterns from
the unadjusted data (e.g., temporal trends for mussels and clams).
Other trace substances, when the suggested effects of environmental
variables are statistically removed from tissue concentrations,
may exhibit clearer trends than PCBs and will be investigated
in the future. Tissue concentrations of PCBs are higher in the
South Bay reach than in other reaches, thus mirroring the findings
in water and sediment. Both mussels and clams indicate increases
in copper in the Estuary. Whether this increase is due to increased
copper loading to the Estuary from runoff or other causes is not
immediately apparent. Perhaps most interestingly, the spatial
and temporal trends evident with the mussels were not apparent
in the oyster data. This emphasizes the importance of species
selection in view of the management issues important for the Estuary.
Bivalves serve as useful biomonitors for site comparisons (provided
the same species can be deployed) in efforts to determine general
pollutant sources or pathways.
6. The bivalve monitoring component includes measurements that
theoretically lend themselves to evaluate contaminant effects
on these indicators, such as growth, condition, and survival.
While we are currently using bivalves merely as contaminant integrators
and surrogates for pollutant measurements in the water column,
they might also serve as response indicators to pollutants. Whether
or not bivalves are an effective tool for evaluating pollutant
effects remains to be assessed and will likely be introduced in
the RMP re-design discussion.
Recommendations
for Consideration in Redesign
1.
Maintain the approach of using transplanted bivalves for long-term
trend monitoring and as a relatively simple diagnostic tool of
emerging pollutant problems, provided that the current analyte
list is expanded to include bioaccumulative substances and other
contaminants that are currently not quantified but which are suspected
to cause environmental problems.
2. Determine the potential application of a variety of bivalve
species in pollutant source and pathway identifications.
3. Determine if bivalves are useful in the determination of pollutant
effects.
4. Continue to explore the effects of environmental variables
on bivalve health and bioaccumulation by collecting water data
near bivalve deployment sites at the same depths as the bivalves.
5. Evaluate which indicator species should be used to assess contaminant
transfer to higher trophic levels.
References
Absil,
M.C.P., J.J. Kroon, and H.T. Wolterbeck. 1994. Availability of copper
from phytoplankton and water for the bivalve Macoma balthica.
2. Uptake and elimination from 64Cu-labelled diatoms
and water. Marine Biology 118(1): 129135.
Gunther,
A.J. and J.A. Davis. 1997. An evaluation of bioaccumulation monitoring
with transplanted bivalves in the RMP. In 1996 Annual Report:
San Francisco Estuary Regional Monitoring Program for Trace Substances.
San Francisco Estuary Institute, Oakland, CA. pp. 187200.
Gunther,
A.J., J.A. Davis, D.D. Hardin, J. Gold, D. Bell, J.R. Crick, G.M.
Scelfo, J. Sericano, and M. Stephenson. 1999. Long-term bioaccumulation
monitoring with transplanted bivalves in the San Francisco estuary.
Marine Pollution Bulletin 38(3):170181.
Hebert,
C.E. and K.A. Keenleyside. 1995. To normalize or not to normalize?
Fat is the question. Envir. Tox. and Chem. 14:801807
Hutchins,
D.A., I. Stupakoff, and N.S. Fisher. 1996. Temperature effects on
accumulation and retention of radionuclides in the sea star Asterias
forbesi: Implications for contamination of northern waters.
Marine Biology 125(4):701706.
Luoma,
S.N. and G.W. Bryan. 1982. A statistical study of environmental
factors controlling concentrations of heavy metals in the burrowing
bivalve Scrobicularia plana and the polychaete Nereis
diversicolor. Estuarine, Coastal and Shelf Science 15(1):95108.
Luoma,
S.N. and R. Linville. 1996. A comparison of selenium and mercury
concentrations in transplanted and resident bivalves from north
San Francisco bay. In 1995 Annual Report: San Francisco Estuary
Regional Monitoring Program for Trace Substances. San Francisco
Estuary Institute, Oakland, CA. pp. 160170.
Magni,
P. 1993. Effect of oxygen concentration on the bioavailability of
copper for the bivalve Macoma balthica. Instituut voor Oecologisch
Onderzoek, Heteren, Netherlands. NTIS No. PB94110707XSP. 33p.
Russell,
R.W. and F.A.P.C. Gobas. 1989. Calibration of the Fresh Water Mussel
Elliptio complanata for quantitative biomonitoring of hexachlorobenzene
and octachlorostyrene. Bull. Environ. Contam. Toxicol. 43:576582.
Salazar
M.H and S.M. Salazar. 1995. In situ bioassays using tranplanted
mussels: 1. Estimating chemical expposure and bioeffects with bioaccumulation
and growth. In J.S. Hughes, G.R. Biddinger, and E. Mones,
(eds.), Third Symposium on Environemntal Toxicology and Risk Assessment,
STP 1218. American Society for Testing and Materials, Philadelphia,
1994, pp. 216241.
Salazar
M.H and S.M. Salazar. 1998. Using cages bivalves as part on an expposure-dose-response
triad to support and integrated risk assessment strategy. In
Proceedings Ecological Risk Assessment: A meeting of policy and
science. A. de Peyster and K. Day, (eds.). SETAC Press, pp. 167192.
Stephenson,
M. 1992. A report on bioaccumulation of trace metals and organics
in bivalves in San Francisco Bay. Submitted to California Regional
Water Quality Control Board, San Francisco Bay Region. California
Department of Fish and Game, Moss Landing Marine Labs, Moss Landing,
CA.
Wang,
W-X, N.S. Fisher, and S.N. Luoma. 1995. Assimilation of trace elements
ingested by the mussel Mytilus edulis: Effects of algal food
abundance. Marine Ecology Progress Series 129:165176.
Wright,
D.A. and C.D. Zamuda. 1987. Copper accumulation by two bivalve molluscs:
Salinity effect is independent of cupric ion activity. Marine Environmental
Research 23(1):114.
|
RMP:
RMP
Home | Program Information
| Publications | RMP
Data
