Risk-based approach to appraise valve closure in the
clam Corbicula fluminea in response to waterborne metals
Chung-Min Liao
a,)
, Li-John Jou
a,b, Bo-Ching Chen
ca
Department of Bioenvironmental Systems Engineering, National Taiwan University, Taipei, Taiwan 10617, ROC
b
Department of Biomechatronic Engineering, National Ilan University, Ilan, Taiwan 260, ROC
c
Department of Post-Modern Agriculture, Mingdao University, Changhua, Taiwan 52345, ROC Received 30 April 2004; accepted 15 October 2004
A model was developed to link valve closure in clams to concentrations of metals in water. Abstract
We developed a risk-based approach to assess how the valve closure behavior of Asiatic clam Corbicula fluminea responds to waterborne copper (Cu) and cadmium (Cd). We reanalyzed the valve closure response data from published literature to reconstruct the response time-dependent dose–response profiles based on an empirical three-parameter Hill equation model. We integrated probabilistic exposure profiles of measured environmental Cu and Cd concentrations in the western coastal areas of Taiwan with the reconstructed dose–response relationships at different integration times of response to quantitatively estimate the valve response risk. The risk assessment results implicate exposure to waterborne Cu and Cd may pose no significant risk to clam valve activity in the short-time response periods (e.g., !30 min), yet a relative high risk for valve closure response to waterborne Cu at response times greater than 120 min is alarming. We successfully linked reconstructed dose–response profiles and EC50-time relationships associated with the fitted daily valve opening/closing rhythm characterized by a three-parameter lognormal function to predict the time-varying bivalve closure rhythm response to waterborne metals. We parameterized the proposed predictive model that should encourage a risk-management framework for discussion of future design of biological monitoring systems.
Ó 2004 Elsevier Ltd. All rights reserved.
Keywords:Valve closure; Corbicula fluminea; Risk assessment; Probabilistic; Cadmium; Copper
1. Introduction
In 1986, a mass mortality of oysters caused by high levels of copper pollution, known as the green oyster incident, occurred in the mariculture beds located along
the western coastal areas of Taiwan (Lee et al., 1996).
Since then, the pollution of various rivers and coastal areas by heavy metal has been received increasing
atten-tion in Taiwan. Jeng et al. (2000)carried out a 3-year
Asia/Pacific Mussel Watch project (1995–1998) to investigate the heavy metal concentrations in various organisms collected from several coastal areas of Taiwan, and proposed that several species of clams are potential candidates for monitoring copper (Cu), zinc (Zn), lead (Pb), and cadmium (Cd) in the marine
environment. Hung et al. (2001) further analyzed the
correlations among the trace metal concentrations in bivalves, water and sediments collected in these areas, and indicated that good correlations were observed between bivalves and water for Zn, Cd, Pb, and Cu.
Doherty et al. (1987)reported that the bioconcentration factors are especially higher for Cu and Cd of the Asiatic clam Corbicula fluminea. Therefore, monitoring of
) Corresponding author. Tel.: C886 2 2363 4512; fax: C886 2 2362 6433.
E-mail address:cmliao@ntu.edu.tw(C.-M. Liao).
0269-7491/$ - see front matterÓ 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.envpol.2004.10.015
behavioral activities of the clam as well as their body residues can provide information regarding the metal levels in the environment.
For the water quality management in aquatic ecosystems, it is important to be able to predict the impact of chemical and toxic effects on aquatic species. Building such a predictive capability requires continu-ous, real-time monitoring for environmental toxicity. Conventional monitoring methods that are performed with discrete sampling followed by chemical analysis in the laboratory, however, are usually characterized as non-continuous processes, and the analytical results are
available after a certain time interval (Borcherding and
Wolf, 2001; Kramer and Foekema, 2001). Conse-quently, biological early warning systems, which are designed to detect developing toxicity by continuously tracking behavioral or physiological change of a tested organism, have been increasingly applied to water
quality control during the last few decades (Borcherding
and Jantz, 1997; van der Schalie et al., 2001).
Many aquatic organisms including fish, mussel, invertebrate, and algae are selected as biological early warning indicators because of their sensitive, immediate responses to the occurrence of contaminants at concen-trations which could threat the organisms therein (Baldwin and Kramer, 1994; Sluyts et al., 1996; van der Schalie et al., 2001). One of the more successful biological early warning systems is the valvometric technique that employs bivalves as sentinel organisms for monitoring the concentration of selected pollutants
in the environment (Curtis et al., 2000; Tran et al.,
2003a,b).
Bivalves are commonly available organisms that are abundant in the freshwater as well as the marine environment. They have been suggested as ideal contamination indices in aquatic ecosystems because of their wide distribution, extensive population, seden-tary nature and the ability to accumulate contaminants (Jeng et al., 2000; El-Shenawy, 2004). Moreover, they close their shells for an extended period of time as an escape behavior to exclude themselves from the outside
environment when exposed to a contaminant (Wildridge
et al., 1998; Kadar et al., 2001). Continuous monitoring of the frequency and duration of valve closure in bivalves can thus provide a reliable means for estimating
the level of pollution in the environment (Sluyts et al.,
1996; Heinonen et al., 2003).
In this present work, we develop a systematic and quantitative risk assessment framework, which is most needed to interpret the significance of the reported exposures. A major complication in predicting or estimating risks for aquacultural species is the high degree of uncertainty resulting from the lack of dose– response information and the large environmental variability in exposures among individuals. As a result, formal risk assessments are scarce regarding the
aquacultural species. We focus on the risk of physio-logical and behavioral changes of clam exposed to waterborne Cu and Cd because evidence for this type of adverse effect has been presented in numerous studies (Doherty et al., 1987; Markich et al., 2000; Borcherding and Wolf, 2001; Kadar et al., 2001; Heinonen et al., 2001; Markich, 2003). Thus, knowledge of the potential risks associated with exposure to Cu and Cd is essential for the effective formulation of conservation and management plans.
The objectives of this study are twofold: (1) to con-duct an environmental risk assessment based on the USEPA methodology and (2) to develop a mathematical
model to better describe the
concentration–time-response relationships for the clam exposed to water-borne Cu and Cd. Based on the pharmacodynamic concept, we reanalyze the valve closure response data of the Asiatic clam C. fluminea exposed to waterborne Cu
and Cd proposed byTran et al. (2003a,b)to reconstruct
the temporal, concentration–response profiles. We use a probabilistic approach to address the uncertainties of Cu and Cd concentrations presented in clam farms along the western coastal areas of Taiwan. We combine the probabilistic exposure profiles with the concentra-tion–response relationships at different integration times of response of 30, 60 and 120 min to arrive at the risk characterization that allows us to quantitatively estimate the valve closure response risk for clams exposed to waterborne Cu and Cd. Incorporating modeled daily opening/closing rhythm into dose–time-response profiles enables us to predict the time-varying bivalve closure response and to assess behavioral endpoint for clams exposed to waterborne metals.
2. Materials and methods
Our risk assessment approach is divided into 5 phases (Fig. 1) and is described in the subsequent sections. 2.1. Problem formulation
The problem formulation of this study is the phase where the assessment endpoint is defined, analyses for associating Cu/Cd contamination with the assessment endpoint are planned, and a conceptual model is
developed (Fig. 1). The conceptual model is developed
starting from the estimation of predicted environmental contamination of Cu/Cd in terms of measured environ-mental concentrations in clam farms in that the major database were adopted from the studies conducted by
Jeng et al. (2000) and Hung et al. (2001).
We define the assessment endpoint for clam is the valve closure response risk. The conceptual model is based on a number of assumptions. These assumptions are necessary mainly because of lack of data, and to
keep the model simple yet reasonably functional. Most of the assumptions are stated in the parameter descriptions in the subsequent sections. The major assumptions are as follows. The pathway of Cu/Cd exposure of farmed clams is only via water ingestion. Farmed clams are contaminated with Cu/Cd through aquacultural water that is obtained from surface water along the western coastal areas of Taiwan.
2.2. Exposure analysis
Measured environmental metal concentrations for Cu and Cd were obtained from a large field study of
effluent monitoring data collected along the western
coastal areas of Taiwan (Jeng et al., 2000; Hung et al.,
2001). We selected the major C. fluminea farming sites at
Chunghua, Yunlin, Chaiyi, and Tainan located at the western coastal areas of Taiwan as our study clam farms. In these areas, since the land-subsidence caused by overusing groundwater for aquaculture was serious, the major farming strategies of C. fluminea are poly-culture by mixing seawater and freshwater to reduce
aquaculture freshwater demand and groundwater
dependence. The selected clam farms had similar feeding strategies. We reanalyzed the measured environmental Cu/Cd concentrations through statistical tests. There are Exposure Analysis
Measured Environmental metal
concentrations
Modeling daily rhythm functions Changes of daily
opening/closure rhythm
Effect analysis: PD-based Hill model
Reconstructed time-varying
dose-response profiles EC50 - time profiles
Risk Characterization Response - risk exceedence probabilistic curves
Bivalve closure response predictions
Uncertainty analysis Sources of uncertainty /variability Model parameterization
Problem formulation
Define endpoint for farmed clam C. fluminea Plan for probabilistic analysis
Develop a conceptual model
Fig. 1. A conceptual diagram showing a risk-based approach including problem formation, exposure analysis, effect analysis, risk characterization, and uncertainty analysis developed to assess Corbicula fluminea valve closure response risk exposed to waterborne metals.
multiple sources of variability and uncertainty to be considered during distribution development for concen-trations of Cu and Cd in water from measured values. Therefore, data were log-transformed when necessary to meet the assumptions of statistical tests. All statistical
analyses were conducted using the StatisticaÒ software
package (StatSoft, Tulsa, OK, USA).
Thanks to the excellent recordings of typical daily valve opening/closing activities of C. fluminea from
the previous researchers (Doherty et al., 1987; Tran
et al., 2003a), distributions of the daily valve opening/ closing rhythm in C. fluminea have a well-established
sequence framework. Tran et al. (2003a) used 71
C. fluminea(average fresh weight without the shell was
0.76 G 0.03 g) over a period of 50 d at water
temper-ature of 15 G 0.5 C with pH ranged from 7.8 to 8.0 and
fed continuously with a unicellular algae Scenedesmus
subspicatus to determine the distribution of daily valve
opening/closing rhythm. We used the maximum-likeli-hood method to fit the distribution of the daily rhythm of clam valve opening/closing. We used TableCurve 2D (Version 5, AISN Software Inc., Mapleton, OR, USA) to optimal fit the distribution data of the opening/ closing valve daily rhythm in C. fluminea obtained from
Tran et al. (2003a). A value of p ! 0.05 was judged significant.
2.3. Effect analysis
Time-varying percentage of response events (closure or brief opening followed by closure) in relation to occasional additions of Cd or Cu to the water were fitted using an empirical three-parameter Hill equation model (Lalonde, 1992; Bourne, 1995) based on the previously
published data of % response versus mg L1Cd or Cu in
water from Tran et al. (2003a,b). In fitting the Hill
equation model to the observed data of % valve response as a function of Cd/Cu concentration in water, the dose–response profile can be expressed as,
RZRmax!C n w Kn 0:5CCnw ; ð1Þ
where R is the measured response (% response), K0.5is
the metal concentration yielding half of maximal
response of Rmax(mg L1), Cwis the metal concentration
in water (mg L1), and the exponent n is a fitted average
value or referred to as the Hill coefficient which is a measure of cooperativity. A value of n O 1 indicates positive cooperativity. The EC50 values were calculated as the dose where R Z 50% response in that relation between EC50 values of C. fluminea to Cd or Cu concentrations and response time were adopted from
Tran et al. (2003a,b) and incorporated into the reconstructed dose–response models to determine the effect profiles.
We reanalyzed the published data obtained from
Tran et al. (2003a,b)regarding the effect concentrations of Cd and Cu causing 50% of total valve closure response of clam at different integration times of response of 30, 60 and 120 min. The Hill mathematical model was used because it allows for comparison of cooperativity among different response time periods that validates the observations of published studies and shows that bivalve closure response is dependent on the response time and metal concentration.
We treated EC50 values in Eq. (1) probabilistically
and the cumulative distribution function (cdf) of predicted bivalve closure response function for a given metal concentration, F(R|C ), could be expressed sym-bolically as a conditional cdf, FðRjCÞZF Rmax!Cn ðEC50ÞnCCn ; ð2Þ
where F( ) is the cumulative standard normal distribu-tion.
2.4. Risk characterization
Risk characterization is the phase of risk assessment where the results of the exposure and quantitative effect assessments are integrated to provide an estimate of risk for the population under study. Risk at a specific metal concentration to bivalve closure response can be calculated as the proportion of clam expected to expose to that metal concentration multiplied by the condi-tional probability of bivalve closure response, given a certain concentration. This results in a joint probabil-ity function (JPF) or exceedence profile, which describes the probability of exceeding the concentration associ-ated with a particular degree of effect. This can be expressed mathematically as
RðCÞZFðCÞ!FðRjCÞ; ð3Þ
where R(C ) is the risk for a valve closure at con-centration C, and F(C ) is the cdf of having metal concentration C in water.
A risk curve was generated from the cumulative distribution of simulation outcomes. Each point on the risk curve represents both the probability that the chosen proportion of clam will be affected and also the frequency with which that level of effect would be exceeded. The x-axis of the risk curve can be interpreted as a magnitude of effect (a percentage of the given clam expected to suffer the adverse effect), and the y-axis can be interpreted as the probability that an effect of at least that magnitude will occur. These probabilities are based on the current exposure data so at each point on the JPF we can also interpret as: under current conditions, x% of clam will be effected and that this proportion of clam would be affected by y% of the current observations.
2.5. Uncertainty analysis 2.5.1. Model parameterization
Parameterization of the model involved selecting data sets and deriving input distributions. Data were sorted by reported statistical measure, e.g., mean, standard deviation, standard error, etc. We used the chi-square
(c2) and the Kolmogorov–Smirnov (K-S) statistics
to optimize the goodness-of-fit of distributions. We
employed StatisticaÒ to analyze data and to estimate
distribution parameters. The selected distribution type and parameters were based on statistical criteria and comparisons of distribution parameters.
2.5.2. Measured environmental
metal concentration: Cw
Distributions of measured environmental water Cu
and Cd concentrations (Cw) were fit to the polled field
observations obtained from selected clam farms located at the western coastal Taiwan region. We determined that the lognormal distribution model fits the observed data of Cu concentrations in Chunghua, Yunlin, Chaiyi, and Tainan clam farms favorably. All variables modeled as the lognormal distributions from which geometric mean and geometric standard deviation for each
vari-able was calculated (Table 1). The results of data
reanalysis give the Cd distributions in Chaiyi and Tainan clam farms ranged from 0.3 to 0.5 and 0.5 to
0.7 ng mL1. We used uniform distributions with these
ranges for Cd concentration distributions (Table 1).
2.5.3. Parameter in Hill equation model: EC50
In applying dose–response relationships derived from
experimental studies adopted fromTran et al. (2003a,b),
we must consider the limitations of the data and account for the inherent uncertainty that arises from a number of sources, including the limited number of observations
and limited sample size within treatment sets. To account this uncertainty, we constructed distributions for the input variables of EC50 values of Hill model
derived dose–response function in Eq. (1). We
deter-mined a normal distribution for EC50 values based on K-S statistics, and incorporated the distributions into the Monte Carlo simulation to obtain 2.5- and 97.5-percentiles as the 95% confidence interval for recon-structed dose–response profiles.
2.5.4. Monte Carlo analysis
In order to quantify this uncertainty and its impact on the estimation of expected risk, we implemented a Monte Carlo simulation that includes input distribu-tions for the parameters of the derived dose–response function as well as for estimated exposure parameters. To test the convergence and the stability of the numerical output, we performed independent runs at 1000, 4000, 5000, and 10 000 iterations with each parameter sampled independently from the appropriate distribution at the start of each replicate. Largely because of limitations in the data used to derive model parameters, inputs were assumed to be independent. The result shows that 5000 iterations are sufficient to ensure
the stability of results. We employed Crystal BallÒ
software (Version 2000.2, Decisioneering, Inc., Denver, Colorado, USA) to implement the Monte Carlo simulation.
3. Results
3.1. Exposure assessment
Fig. 2shows the box plots of interquartile and 50th-percentile associated with whisker plots indicating 2.5-and 97.5-percentile of Cu 2.5-and Cd concentrations in study clam farms in that the median Cu and Cd levels in
water range from 1.3 to 5.74 and 0.4 to 0.6 ng mL1,
respectively.
Optimal statistical models were selected on the basis of least squared criterion from a set of generalized linear and nonlinear autoregression models provided by Table Curve 2D package fitted to the daily rhythm distribu-tions of valve opening/closing of C. fluminea. The optimal fitted nonlinear regression function was the three-parameter lognormal distribution, given by fðt; a; b; cÞZa exp " 0:5 lnt b c 2# Cd; ð4Þ
where a is the amplitude, b is the value of t at maxi-mum, c can be derived from the area-under-curve
(AUC) Z abcpffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2p expðc2Þ; and d is the displacement
in y-axis. Here we denote the valve opening function as j(t; a,b,c) and f(t; a,b,c) for valve closing function.
Table 1
Model input distributions of measured metal concentrations in selected clam farms
Study clam farm Measured environmental
metal concentration in water (ng mL1) Distribution Copper concentration Chunghua 3.0 G 0.8a LN(2.90, 1.30)c Yunlin 1.3 G 0.1 LN(1.30, 1.08) Chaiyi 6.3 G 0.4 LN(5.54, 1.66) Tainan 6.0 G 1.7 LN(5.77, 1.32) Cadmium concentration Chaiyi 0.3–0.5b U(0.3, 0.5)d Tainan 0.5–0.7 U(0.5, 0.7) a Mean G SD. b Minimum–maximum. c
Lognormal distribution with geometric mean and geometric standard deviation.
d
Fig. 3 demonstrates the fitted three-parameter lognor-mal models for the observations of daily rhythm of valve opening and closing.
A bimodal distribution of the daily rhythm of valve opening/closing is separated at 7 AM that is based on
the suggestion byTran et al. (2003a). The fitted
three-parameter lognormal models for daily rhythm of valve opening and closing have the forms, respectively, as
where j(t) and f(t) are the daily rhythm functions of valve opening and closing proportions at any given time t, respectively.
3.2. Effect analysis
As can be seen inFig. 4, dose–response profiles were
in a time-dependent fashion. The Hill model and a 5000 Monte Carlo simulation provided an adequate fit for the
data (c2 goodness-of-fit, P O 0.5). The n and EC50
values clearly show that there were profound differences in sensitivity to Cu and Cd in different integration times of response. Regression lines from the nonlinear Hill three-parameter model transformations of percent valve closure response versus metal concentrations in water
curves had good fit as judged by high r2values (0.991–
0.997, p ! 0.05). The Hill coefficient (n) for Cu concentration to valve response (1.27–1.71) was
in-dicative of positive cooperativity. In the case of Cd-response profiles, n Z 1.43 for Cd-response time less than 30 min, whereas for response time greater than 60 min,
n values all less than 1.
Based on our fitted concentration–response model (Fig. 4), the estimated EC50 values in Cu-response
profiles were 21.4, 10.9, and 8.1 ng mL1 for valve
response times of 30, 60, and 120 min, respectively; whereas in Cd-response relationships were 114.2, 42.72,
and 25.91 ng mL1, respectively, for 30, 60, and
Chunghua Yunlin Chaiyi Tainan Clam farm 0 2 4 6 8 10 12 14 16 Cu concentration in water (ng/mL) Cd in water (ng/mL) 2.5 -97.5 25 -75 Median Chaiyi Tainan 0 0.2 0.4 0.6 0.8 1.0 B A
Fig. 2. Box and whisker plot representations of (A) water Cu concentrations at four selected clam farm locations and (B) water Cd concentrations at two selected clam farm locations.
jðtÞZ 8 < : j1ðtÞZ6:4 exp 0:5ln0:47ð3:2Þt 2 C1:9; 0%t%7; r2 Z0:96; j2ðtÞZ16:8 exp 0:5ln0:24ð10:8Þt 2 C1:7; 7!t%24; r2 Z0:96; ð5Þ fðtÞZ 8 < : f1ðtÞZ12:3 exp h 0:5ln0:20ð4Þt 2i C3:8; 0%t%7; r2 Z0:84; f2ðtÞZ14:8 exp 0:5ln0:083ð18:2Þt 2 C3:6; 7!t%24; r2 Z0:92: ð6Þ
120 min. Therefore, low concentrations of Cu and Cd caused a significant change in the valve position, suggesting that valve position is suitable for a biologi-cally sensitive endpoint.
3.3. Risk estimates
Risk curves shown in Fig. 5 indicate the estimated
probabilistic of effects of differing magnitude for clam for each selected clam farm locations exposed to Cu and Cd in water. The plotted probabilities, calculated from the outcome of the Monte Carlo simulation followed
a JPF shown in Eq. (3) describing the exceedence cdfs
(Fig. 5) associated with a metal-specific dose–response
relationship (Fig. 4), take into account the uncertainty
in estimating risk derived from variability and uncer-tainty in model parameters.
Table 2 gives the probabilities that 10% or more of the valve opening response of clam is affected by waterborne Cu and Cd (risk Z 0.1) for 30 and 120 min
response times at different clam farms. Table 2
demonstrates that proportion of clam closing response affected by Cu and Cd in 30 min response time ranged from 1.0 to 23.1% and 0.04 to 0.07%, respectively, indicating that it poses no threat to valve response behavior. For 120 min response time, however, the proportions of clam response affected by Cu at Chaiyi (58.5%) and Tainan (50.7%) are relative high. The results of the risk assessment indicate that waterborne Cu and Cd concentrations pose no significant risk to clam valve activity in the short-time response periods
(e.g., !30 min), yet a relative high risk for valve closure response to waterborne Cu concentration at response times greater than 120 min in Chaiyi and Tainan clam farms is alarming.
3.4. Bivalve closure response predictions
We link the reconstructed time-varying dose–
response profiles (Fig. 4) and EC50-time relationships
provided by published data associated with the fitted
model of daily rhythm valve opening/closing (Fig. 3) to
predict the bivalve closure rhythm response to water-borne Cu and Cd. The mathematical description using an analysis of the dose–response-type curve that integrates time of any detection mechanisms therefore can be expressed as,
jðtCDtÞZjðt; 0Þ!FðRjCwÞZjðt; 0Þ½1 RðDt; CwÞ;
ð7Þ where j(t C Dt) is the daily rhythm function of valve opening at any given incremental time Dt (h), j(t,0) is the daily rhythm function of valve opening exposed to
unpolluted water, F(R|Cw) is the cdf of predicted bivalve
opening response function for a given metal
concentra-tion Cw, and R(Dt,Cw) is the Hill model-based dose–
response function at any given Dt. The daily rhythm function of valve opening exposed to unpolluted water
j(t,0) has the same form as given in Eq.(5).
The Hill model-based dose–response function at any given Dt can be expressed as,
RðDt; CwÞZ
100CnðDtÞ
w
½EC50ðDtÞnðDtÞCCnðDtÞw
; ð8Þ
in that EC50(Dt) and n(Dt) are time-dependent metal-specific functions. The function of EC50(Dt) for Cu can
be adopted fromTran et al. (2003b)as,
EC50ðDtÞCuZ3:76 exp 91:9 23:99CDt ; r2Z0:996: ð9Þ
We fitted a nonlinear regression to EC50-time
relationships derived from Hill model (Fig. 4B) to
obtain the function of EC50(Dt) for Cd,
EC50ðDtÞCdZ2:92C50:19Dt1:212; r2Z0:988: ð10Þ
The function of n(Dt) for Cu and Cd can be obtained, respectively, by a nonlinear regression fitting the
response time-dependent n values indicated in Fig. 4A
and B, and results in
nðDtÞCuZ1:396 exp Dt 0:379 C1:284; r2 Z0:97; ð11Þ 0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (hr) 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (hr) Valves opening ( ) Valves closing ( ) B
A : Observations: Fitted model
Fig. 3. Selected three-parameter lognormal functions fitted to observations of daily rhythm of Corbicula fluminea valve (A) opening and (B) closing histograms.
nðDtÞCdZ1:45 exp Dt 1:21 C0:417; r2 Z0:94: ð12Þ
Fig. 6 demonstrates the simulations of daily rhythm of valve opening subjected to different waterborne Cu
(20–100 ng mL1) and Cd (100–500 ng mL1)
concen-trations at different exposure time periods. The pro-posed simulation model can quantitatively describe the bivalve behavioral activity when clam exposed to metals. Moreover, the bivalve ability to close its shell as an alarm signal when exposed to a contaminant can be successfully tested in terms of the potential and the limitations of using bivalve as a rapid response and/or sensitive biosensor for different waterborne contami-nants.
4. Discussion
A number of studies in modeling environmental fate, bioaccumulation, and toxicity have recognized the need for body burden–effect relationships and advocated a body burden-based approach in environmental
toxicity and risk assessments (Bartell et al., 1988;
McCarty and Mackay, 1993; Liao et al., 2002). As the body residue of xenobiotic metal elevated, C. fluminea reduces its rate of metabolism, heartbeat, and oxygen
consumption with the onset of valve closure (Aunaas
and Zachariassen, 1994; Curtis et al., 2000; Ortmann and Grieshaber, 2003). Consequently, the establishment of dose–response profile based on the internal body burden concept is necessary while assessing the potential 0 20 40 60 80 100 A 30 min n = 1.71 ± 0.26 EC50 = 21.4 (95 CI: 8.9-34.2) r2 = 0.989, p < 0.05 0 100 200 300 400 500 Cu concentration in water (ng/mL) 0 20 40 60 80 100 0 100 200 300 400 500 00 100 200 300 400 500 20 40 60 80 100 120 min n = 1.27 ± 0.24 EC50 = 8.1 (4.2-16.0) r2 = 0.977, p < 0.05 60 min n = 1.32 ± 0.15 EC50 = 10.9 (5.5-21.2) r2 = 0.991, p < 0.05 0 20 40 60 80 100 0 200 400 600 800 1000 1200 Cd concentration in water ( ng/mL) Response ( ) Response ( ) 0 20 40 60 80 100 0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200 n = 0.80 ± 0.22 EC50 = 25.91 (16.9-34.9) r2 = 0.994, p < 0.05 120 min 0 20 40 60 80 100 n = 0.92 ± 0.19 EC50 = 42.72 (34.5-50.9) r2 = 0.994, p < 0.05 60 min B 30 min n = 1.43 ± 0.15 EC50 = 114.2 (95 CI: 101.2-127.2) r2 = 0.997, p < 0.05
Fig. 4. Reconstructed dose–response profiles with 95% confidence interval for the percentage of valve response as a function of (A) Cu and (B) Cd concentrations characterized by a nonlinear three-parameter Hill equation model at different integration times of response of 30, 60, and 120 min. The measurements are shown with open circles. Error bars represent one standard deviation from the mean.
risk of C. fluminea exposed to metals. In the present study, however, valve response can only be expressed as a function of waterborne Cd/Cu concentration because
of lack of body burden data in C. fluminea.Graney et al.
(1983), Baudrimont et al. (1997), and El-Shenawy (2004)
indicated that simple linear correlations exist between internal and external concentration while exposing
clams to higher levels of waterborne metals such as Cd
and Cu.Doherty et al. (1987) also pointed out that the
valve closure responses of C. fluminea would not interfere with the bioaccumulation process in operation
at low levels of Cd contaminations (!0.1 mg mL1). In
order to respond to the realistic situations, further research works will undoubtedly continue a trend
0 10 20 30 40
0 10 20 30 40
50 60 70 80
0 10 20 30 40 50 60 70 80 90
Proportion of clam affected by Cu (%)
0 5 10 15 20 25 30 35 40
Proportion of clam affected by Cu (%) 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 1 2 3 4 5 B 120min 30min Yunlin 120min 30min Chunghua A 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Proportion of clam affected by Cu (%)
Exceedence Risk Exceedence Risk
Exceedence Risk Exceedence Risk Exceedence Risk Exceedence Risk
C
120min Chaiyi
100 0 10 20 30 40 50 60 70 80 90 100
Proportion of clam affected by Cu (%) D
120min Tainan
Proportion of clam affected by Cd (%)
0 2 4 6 8 10 12 14
0 2 4 6 8 10 12 14
Proportion of clam affected by Cd (%) 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0.03 0.05 0.07 0.09 0.11 0.13 120min F 30min Tainan E 120min 0 0.02 0.04 0.06 0.08 0.1 30min Chaiyi 30min 0 20 40 60 80 100 0 0.2 0.4 0.6 0.8 1 0 10 20 30 40 50 60 30min 0 0.2 0.4 0.6 0.8 1
Fig. 5. Exceedence risk curves with 95% confidence interval at 30 and 120 min response times for selected clam farms at (A) Chunghua, (B) Yunlin, (C) Chaiyi, and (D) Tainan exposed to waterborne Cu, whereas (E) Chaiyi and (F) Tainan were exposed to waterborne Cd.
towards a better understanding regarding the relation-ships between external and internal metal concentra-tions at low exposure levels in clam farms.
It is generally recognized that an alarm must be set by a biological early warning system within 30 min after the
addition of a toxicant (Borcherding and Jantz, 1997).
The mathematical descriptions we developed in the present study, which combine the natural valve opening/
closing rhythm and the valve behavior changes of
C. fluminea exposed to various Cd/Cu contaminations,
to arrive at an exhaustive comprehension of concentra-tion–response relationships at any given time. We also show that the concentration-time–response profiles proposed here can be served as a powerful tool to quantitatively and continuously monitor the ambient water condition in which the clams farmed. A major
Table 2
The proportion of valve closure response of C. fluminea affected by waterborne Cu and Cd for exceedence risk is equal to 0.1 at selected clam farms
Response time Chunghua Yunlin Chaiyi Tainan
Proportion of clam valve response affected by Cu (%)
30 min 5.5 (2.5–20.7)a 1.0 (0.4–4.2) 23.1 (11.9–57.4) 16.5 (8.1–46.9)
120 min 29.0 (14.8–48.9) 9.9 (4.4–20.2) 58.5 (37.2–76.5) 50.7 (30.2–70.3)
Proportion of clam valve response affected by Cd (%)
30 min 0.04 (0.03–0.05) 0.07 (0.06–0.08)
120 min 4.0 (3.2–5.5) 5.2 (4.2–7.2)
a 95% confidence intervals were computed from 2.5 and 97.5-percentile of 5000 Monte Carlo simulations.
0 5 10 15 20 25 30 35 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (hr) 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (hr) Valves opening ( ) Valves opening ( ) 0 4 8 12 16 20 7 9 11 13 15 4 3 2 1 0 2 4 6 8 10 14 16 18 20 4 3 2 1 2 1: Cw = 0 ng/mL 2: Cw = 20 (Cu) ng/mL 3: Cw = 50 (Cu) ng/mL 4: Cw = 100 (Cu) ng/mL 1: Cw = 0 ng/mL 2: Cw = 100 (Cd) ng/mL 3: Cw = 200 (Cd) ng/mL 4: Cw = 500 (Cd) ng/mL t=16:00 t=10:00 0 2 4 6 8 1 4 3 2 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 t=03:00 10 A 0 5 10 15 20 25 30 35 B 0 4 8 12 16 20 7 9 11 13 15 3 4 2 1 0 4 8 12 16 20 12 14 16 18 20 3 4 2 1 t=11:00 t=14:00 0 2 4 6 8 10 4 3 2 1 t=02:00
Fig. 6. Simulations of Corbicula fluminea valve daily opening rhythm subjected to different waterborne metal contaminations at various exposure time periods: (A) starting exposure times were 03:00, 10:00, and 16:00, respectively, subjected to waterborne Cu levels ranged from 20 to 100 ng mL1 and (B) starting exposure times were 02:00, 11:00, and 14:00, respectively, subjected to waterborne Cd levels ranged from 100 to 500 ng mL1.
challenge in transferring laboratory data to the field is that organisms are simultaneously or sequentially
exposed to multiple stressors in ecosystems (Eggen
et al., 2004). We believe that the methodology estab-lished in this work can be extrapolated to other species and other contaminants to simultaneously monitor different types of contaminants with various kinds of aquatic organisms in the field.
We believe that a probabilistic risk-based framework – probability distributions and risk diagrams such as
Fig. 5– is an effective representation of state-of-the-art results of scientific assessments for bivalve closure response to waterborne contaminants and has potential for use in biological early warning systems design. To our knowledge, this risk-based framework has not been addressed until now. Despite great uncertainty in many aspects of integrated assessment, e.g., the problem of physical and chemical variables in water such as temperature, pH, turbidity, oxygen level, which may modify the daily rhythm of valve opening/closing (Sluyts et al., 1996; Simon and Garnier-Laplace, 2004; Fournier et al., 2004), cautious interpretation of observations obtained from optimized-controlled labo-ratory can substantially reduce the likelihood.
Although the suitability and effectiveness of techni-ques for presenting uncertain results is context-dependent, we believe that such probabilistic methods are more valuable for communicating an accurate view of current scientific knowledge to those seeking infor-mation for decision-making than assessments that do not attempt to present results in probabilistic frame-work. We suggest that our probabilistic framework and methods be taken seriously because they produce general conclusions that are more robust than estimates made with a limited set of scenarios or without proba-bilistic presentations of outcomes, and our bivalve closure response modeling technique offers a risk-management framework for discussion of future bio-monitors design in biological early warning systems.
References
Aunaas, T., Zachariassen, K.E., 1994. Physiological biomarkers and the trondheim biomonitoring system. In: Kramer, K.J.M. (Ed.), Biomonitoring of Coastal Waters and Estuaries. CRC Press, Boca Raton, FL, USA, pp. 107–133.
Baldwin, I.G., Kramer, K.J.M., 1994. Biological early warning systems (BEWS). In: Kramer, K.J.M. (Ed.), Biomonitoring of Coastal Waters and Estuaries. CRC Press, Boca Raton, FL, USA, pp. 1–28.
Bartell, S.M., Gardner, R.H., O’Neill, R.V., 1988. In: Adams, W.J., Chapman, G.A., Landis, W.G. (Eds.), Aquatic Toxicity and Hazard Assessment, vol. 10. American Society for Testing and Materials. ASTM STP 971, Philadelphia, pp. 261–274.
Baudrimont, M., Metivaud, J., Maury-Brachet, R., Ribeyre, F., Boudou, A., 1997. Bioaccumulation and metallothionein response in the Asiatic clam (Corbicula fluminea) after experimental
exposure to cadmium and inorganic mercury. Environ. Toxicol. Chem. 16, 2096–2105.
Borcherding, J., Jantz, B., 1997. Valve movement response of the mussel Dreissena polymorpha – the influence of pH and turbidity on the acute toxicity of pentachlorophenol under laboratory and field conditions. Ecotoxicology 6, 153–165.
Borcherding, J., Wolf, J., 2001. The influence of suspended particles on the acute toxicity of 2-chloro-4-nitro-aniline, cadmium, and pentachlorophenol on the valve movement response of the zebra mussel (Dreissena polymorpha). Arch. Environ. Contam. Toxicol. 40, 497–504.
Bourne, D.W.A., 1995. Mathematical Modeling of Pharmacokinetic Data. Technomic Publishing Company, Inc., Lancaster, Penn, 139 pp.
Curtis, T.M., Williamson, R., Depledge, M.H., 2000. Simultaneous, long-term monitoring of valve and cardiac activity in the blue mussel Mytilus edulis exposed to copper. Mar. Biol. 136, 837–846.
Doherty, F.G., Cherry, D.S., Cairns Jr., J., 1987. Valve closure responses of the Asiatic clam Corbicula fluminea exposed to cadmium and zinc. Hydrobiologia 153, 159–167.
Eggen, R.I.L., Behra, R., Burkhardt-Holm, P., Escher, B.I., Schwei-gert, N., 2004. Challenges in ecotoxicology. Environ. Sci. Technol. 38, 58A–64A.
El-Shenawy, N.S., 2004. Heavy-metal and microbial depuration of the clam Ruditapes decussatus and its effect on bivalve behavior and physiology. Environ. Toxicol. 19, 143–153.
Fournier, E., Tran, D., Denison, F., Massabuau, J.-C., Garnier-Laplace, J., 2004. Valve closure response to uranium exposure for a freshwater bivalve (Corbicula fluminea): quantification of the influence of pH. Environ. Toxicol. Chem. 23, 1108–1114. Graney, R.L., Cherry, D.S., Cairns Jr., J., 1983. Heavy metal indicator
potential of the Asiatic clam (Corbicula fluminea) in artificial stream systems. Hydrobiologia 102, 81–88.
Heinonen, J., Kukkonen, J.V.K., Holopainen, I.J., 2001. Temperature-and parasite-induced changes in toxicity Temperature-and lethal body burdens of pentachlorophenol in the freshwater clam Pisidium amnicum. Environ. Toxicol. Chem. 20, 2778–2784.
Heinonen, J., Penttinen, O.P., Holopainen, I.J., Kukkonen, J.V.K., 2003. Sublethal energetic responses by Pisidium amnicum (Bivalvia) exposed to pentachlorophenol at two temperatures. Environ. Toxicol. Chem. 22, 433–438.
Hung, T.C., Meng, P.J., Han, B.C., Chuang, A., Huang, C.C., 2001. Trace metals in different species of mollusca, water and sediments from Taiwan coastal area. Chemosphere 44, 833–841.
Jeng, M.S., Jeng, W.L., Hung, T.C., Yeh, C.Y., Tseng, R.J., Meng, P.J., Han, B.C., 2000. Mussel watch: a review of Cu and other metals in various marine organisms in Taiwan, 1991–98. Environ. Pollut. 110, 207–215.
Kadar, E., Salanki, J., Jugdaohsingh, R., Powell, J.J., McCrohan, C.R., White, K.N., 2001. Avoidance responses to aluminium in the freshwater bivalve Anodonta cygnea. Aquat. Toxicol. 55, 137–148.
Kramer, K.J.M., Foekema, E.M., 2001. The ‘‘MusselmonitorÒ’’ as
biological early warning system. In: Butterworth, F.M., Gunatilaka, A., Gonsebatt, M.E. (Eds.), Biomonitors and Bio-markers as Indicators of Environmental Change 2. Kluwer Academic/Plenum Publishers, NY, USA, pp. 59–87.
Lalonde, R.L., 1992. Pharmacodynamics. In: Evans, W.E., Schentag, J.J., Jusko, W.J. (Eds.), Applied Pharmacokinetics. Lippincott Williams and Wilkins, NY, USA, pp. 4-1–4-33. Lee, C.L., Chen, H.Y., Chuang, M.Y., 1996. Use of oyster,
Crassostrea gigas, and ambient water to assess metal pollution status of the charting coastal area, Taiwan, after the 1986 green oyster incident. Chemosphere 33, 2505–2532.
Liao, C.M., Chen, B.C., Lin, M.C., Chiu, H.M., Chou, Y.H., 2002. Coupling toxicokinetics and pharmacodynamics for predicting
survival of abalone (Haliotis diversicolor supertexta) exposed to waterborne zinc. Environ. Toxicol. 17, 478–486.
Markich, S.J., Brown, P.L., Jeffree, R.A., Lim, R.P., 2000. Valve movement responses of Velesunio angasi (Bivalvia: Hyriidae) to manganese and uranium: an exception to the free ion activity model. Aquat. Toxicol. 51, 155–175.
Markich, S.J., 2003. Influence of body size and gender on valve movement responses of a freshwater bivalve to uranium. Environ. Toxicol. 18, 126–136.
McCarty, L.S., Mackay, D., 1993. Enhancing ecotoxicological modeling and assessment. Environ. Sci. Technol. 27, 1719–1728. Ortmann, C., Grieshaber, M.K., 2003. Energy metabolism and valve
closure behaviour in the Asian clam Corbicula fluminea. J. Exp. Biol. 206, 4167–4178.
Simon, O., Garnier-Laplace, J., 2004. Kinetic analysis of uranium accumulation in the bivalve Corbicula fluminea: effect of pH and direct exposure levels. Aquat. Toxicol. 68, 95–108.
Sluyts, H., van Hoof, F., Cornet, A., Paulussen, J., 1996. A dynamic new alarm system for use in biological early warning systems. Environ. Toxicol. Chem. 15, 1317–1323.
Tran, D., Ciret, P., Ciutat, A., Durrieu, G., Massabuau, J.C., 2003a. Estimation of potential and limits of bivalve closure response to detect contaminants: application to cadmium. Environ. Toxicol. Chem. 22, 914–920.
Tran, D., Fournier, E., Durrieu, G., Massabuau, J.C., 2003b. Copper detection in the Asiatic clam Corbicula fluminea: optimum valve closure response. Aquat. Toxicol. 65, 317–327.
van der Schalie, W.H., Shedd, T.R., Knechtges, P.L., Widder, M.W., 2001. Using higher organisms in biological early warning systems for real-time toxicity detection. Biosens. Bioelectron. 16, 457–465. Wildridge, P.J., Werner, R.G., Doherty, F.G., Neuhauser, E.F., 1998. Acute effects of potassium on filtration rates of adult zebra mussels, Dreissena polymorpha. J. Great Lakes Res. 24, 629–636.
