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 Google Scholar
Google Scholar
Right arrow Articles by Ross, K. E.
Right arrow Articles by Bidwell, J. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Ross, K. E.
Right arrow Articles by Bidwell, J. R.
Agricola
Right arrow Articles by Ross, K. E.
Right arrow Articles by Bidwell, J. R.
Related Collections
Right arrow Heavy Metals
Right arrow Other Models
Right arrow Water Pollution
Published in J. Environ. Qual. 32:1677-1683 (2003).
© 2003 ASA, CSSA, SSSA
677 S. Segoe Rd., Madison, WI 53711 USA

TECHNICAL REPORTS

Heavy Metals in the Environment

Assessing the Application of an Additive Model to Estimate Toxicity of a Complex Effluent

Kirstin E. Ross*,a and Joseph R. Bidwellb

a Wildlife and Fisheries Sciences, Texas A&M University, 210 Nagle Hall, 2258 TAMU, College Station, TX 77843-2258
b Department of Zoology, Oklahoma State University, Stillwater, OK 74078

* Corresponding author (keross{at}wfscgate.tamu.edu).

Received for publication June 20, 2002.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
A number of industries monitor levels of chemicals in their effluent, but few have undertaken prolonged biological monitoring of this wastewater. The focus of the present study was to determine whether past chemical data for effluent from a lead smelter could be used to estimate its past toxicity. Since the interactive effects of metals in effluents are often assumed to be additive, it was hypothesized that an additive model, 100/[{Sigma}(metal concentration in effluent/EC50 for individual metal)], could be used to generate an EC50 from chemical data (where EC50 is the concentration of test material that affects 50% of the test organisms). To test the approach, a larval development toxicity test with the marine polychaete, Galeolaria caespitosa, was used to test 26 separate samples of effluent from a lead smelter, generating empirical EC50 values. EC50 values for each individual metal in the effluent were also generated using the larval development toxicity test. The concentrations of trace metals in each effluent sample were determined and, using the additive model, EC50 values were calculated. For the majority of effluent samples tested, the additive model underestimated toxicity, suggesting the presence of additional unidentified contaminants in the effluent samples. Additionally, a nonlinear rather than linear regression curve was found to best describe the relationship between the model and empirically derived EC50 values. This relationship was then used to estimate past trends in toxicity of the smelter effluent. Forty-eight percent of the variability in measured toxicity was explained by the model, with the model underestimating toxicity in the majority of samples.

Abbreviations: EC50, concentration of test material that affects 50% of the test organisms


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
ENVIRONMENTAL PROTECTION LEGISLATION has, for many years, required chemical analysis of effluents before their release (Hart, 1974; Verschoor and Reijniders, 2001). However, until recently, particularly outside North America and Europe, there has been little or no requirement for biological testing of effluents. While the need for incorporation of whole-effluent toxicity testing is now gaining recognition (Organisation for Economic Cooperation and Development, 1986; Environment Canada, 1992; USEPA, 1993, 1994; Australian and New Zealand Environment and Conservation Council, 2001), the lack of historical toxicity data makes it difficult to assess long-term trends in effluent quality. This type of temporal assessment is especially important when determining whether implementation of pollution control measures has been successful in reducing the toxicity associated with these outputs.

A number of studies of model-derived toxicity estimates based on chemical data have reported strong relationships between predicted and observed toxicity (Maltby et al., 2000; Sarakinos et al., 2000). This indicates that past chemical data for an effluent could be used to calculate toxicity endpoints and provide an estimate of changes in effluent quality over time.

Metals in mixture theoretically interact in additive (sum of the effects is equal to that of the individual metals), synergistic (metals in mixture have a greater effect than the individual components), or antagonistic (toxicity of one or more metal is reduced in mixture) modes. In studies of three or more metals, an additive effect was most often reported (Spehar and Fiandt, 1986; Enserink et al., 1991; Kraak et al., 1994).

It was therefore hypothesized that an additive model could be used to generate toxicity values from measured effluent chemical concentrations. To test the approach, a larval development toxicity test with the marine polychaete, Galeolaria caespitosa, was used to generate EC50 values for effluent samples from a lead smelter (empirical EC50 values) and for each of the effluent constituent metals. An additive model was then used to calculate EC50 values from chemical data for the effluent samples (model EC50 values), and the empirical and model EC50s were compared to assess the predictive value of the model. If successful, the approach could then be used to determine whether pollution reduction measures at the smelter have been successful, or whether increases in production over the years have invalidated these reduction measures.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Toxicity Test Procedure
The G. caespitosa larval development test procedure followed that described by Ross and Bidwell (2001). Galeolaria caespitosa were collected from locations on the South Australian coast. Collection was undertaken by dislodging a cluster of polychaetes of approximately 10 x 10 cm from jetty pilings or rocks at low tide. The clusters were transferred to the laboratory in polyethylene buckets covered with moistened toweling and held in a flow-through tank system (130 L) until test initiation. Tank holding conditions were 7 to 9 mg/L dissolved oxygen, salinity of 32 to 36 g/L, pH of 7.8 to 8.6, and temperature of 22 ± 2°C. Illumination was provided in a light–dark cycle of 16 and 8 h. The polychaetes were kept completely immersed in seawater until test initiation. Polychaetes were held for up to 96 h before test initiation. Previous observations had indicated that viability of gametes was maintained for at least 10 d when the polychaetes were held in their cases in aerated tanks. Tests were conducted in 50-mL acid-rinsed glass beakers. Four replicate beakers were used for each test solution and the controls. Individuals were removed from their cases and transferred to 200 µL of filtered (30 µm sand-filtered) seawater. Release of gametes began almost immediately and was allowed to continue for 15 min, after which the animals were removed. Ten to fifteen males and 50 to 60 females were used per toxicity test. Sperm were pooled and diluted to 60000 sperm/µL. Sperm (100 µL) were added to 30 mL of test solution. During this time eggs were pooled and washed three times with filtered seawater (by allowing the eggs to settle and removing the supernatant). Eggs were then concentrated to 6 mL (again by settling, and removal of supernatant), leaving an egg solution of approximately 20 eggs/µL, and 100 µL was added to the test solutions. Test solutions were incubated at 17 ± 1°C for 48 h under a 16 and 8 h light–dark regime.

Developed larvae and undeveloped eggs were concentrated to 2 mL by filtering away excess seawater through a 3-µm filter. A drop of this solution was examined under 200x magnification and the first 100 organisms encountered scored as either developed or undeveloped. Development was determined by visible internal organization and active cilia. Undeveloped larvae were either unfertilized eggs or undifferentiated cells (Ross and Bidwell, 2001). For tests to be considered acceptable, greater than 90% development in the controls was required. The data were expressed as the proportion of undeveloped eggs in each concentration and EC50 values were calculated using the trimmed Spearman–Karber method (Hamilton et al., 1977).

Composite Effluent Toxicity Tests
Larval development toxicity tests were conducted with 26 composite effluent samples from the smelter. Effluent samples were collected over a 72-h period and stored in 1-L acid-washed polyethylene containers at 4°C for transport to the University of South Australia. The salinity of the effluent was approximately 45 g/L. This was diluted to 32 g/L with Milli-RO water (Millipore, Bedford, MA) before serial dilutions. Undiluted effluent was also used in each toxicity test and hypersaline brine was used as an additional control for these toxicity tests. Hypersaline brine was prepared by evaporating filtered seawater in a 25°C temperature control cabinet until the salinity reached 50 g/L. This was then diluted to match salinity of the undiluted effluent.

Filtered seawater (30 µm sand-filtered) was used as the diluent and seawater controls in all experiments, unless otherwise stated. The seawater was collected from SARDI, West Beach, South Australia, in polyethylene 25-L carboys.

Reference toxicity tests were conducted with each of the effluent toxicity tests using copper, as CuCl2·2H2O (Sigma Chemical, St. Louis, MO).

Dissolved oxygen, pH, salinity, and temperature were measured in each test concentration at the start and end of the toxicity tests with a TPS 90FL water quality meter (Enviroequip, East Malvern, VIC, Australia).

Individual Metal Toxicity Tests
Eight metals (lead, zinc, cadmium, copper, manganese, nickel, arsenic, and selenium) had been reported to be the major constituents of the effluent (Ward et al., 1984; Pasminco Port Pirie Smelters, personal communication, 1998). Bioassays were conducted with each metal to determine the toxicity of each metal alone. Chemical concentrations of all dilutions were measured and EC50 values calculated based on measured concentrations.

Chemical Analyses
Levels of metals in each composite effluent sample were measured using an inductively coupled plasma optical emission spectrometer (Liberty 200; Varian, Palo Alto, CA) following microwave digestion in dilute HNO3. This method was also used to determine the concentration of each metal in the individual metal toxicity tests. This analytical method determines the acid-extractable metals—the concentrations of metals in solution after treatment of an unfiltered sample with hot dilute acid (American Public Health Association, 1992).

Model EC50 and Historical EC50 Value Calculations
The following additive model was used to calculate EC50 values:

[1]

Equation [1] calculates the percent dilution of the effluent sample required to attain an EC50 value (model EC50 value). Model EC50 values were generated for each effluent sample and compared with the value derived from the toxicity tests. Historical EC50 values were calculated with the additive model (Eq. [1]) using individual metal EC50 values and past metal concentrations in the effluent. These data were then transformed according to the relationship found between the model and empirical EC50 values determined for the 26 composite effluent samples (see Empirical and Model EC50 Values for Effluent Samples, below).

Artificial "Effluents"
To further assess the application of an additive model, toxicity tests with artificial "effluents," with known metal concentrations, were conducted. Eight metal mixtures were prepared (Table 1), at concentrations over a range similar to that found in the effluent. Numbers 1 and 2 were based on the highest monthly mean for 1998–2000, 3 and 4 on the lowest monthly mean for 1998–2000, and 5, 6, 7, and 8 on one effluent sample taken in 1999. Additional "effluents" were prepared with one or more of the metals removed. Seawater or intake water (the cooling water used at the smelter that forms the diluent of the final effluent) was used as the diluent in the preparation of these metal mixtures. The metal mixtures were tested using the larval development toxicity test, and empirical EC50 values calculated for each mixture. The additive model (Eq. [1]) was then used to calculate model EC50 values for each metal mixture.


View this table:
[in this window]
[in a new window]
  		Table 1. Concentrations of metals used to prepare artificial effluent solutions.

 
EDTA Treatment
Effluent samples were also treated with ethylenediaminetetraacetic acid (EDTA) to determine whether observed toxicity could be attributed to cationic metals, using a procedure modified from that described in USEPA (1991). Initial toxicity tests were conducted with EDTA (disodium-EDTA; Sigma Chemical) to determine the NOEC (no observed effect concentration—the highest test material concentration for which there was no statistically significant difference from the control) for the G. caespitosa larval development toxicity test (50 mg/L). Each dilution of the composite effluent samples was then treated with EDTA to achieve a final concentration of 50 mg/L before being tested.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Individual Metal EC50 Values
Larval development EC50 values for each constituent metal present in the smelter effluent are presented in Table 2. Copper was the most toxic, with an EC50 value of 19 µg/L, followed by zinc, cadmium and lead, with EC50 values of 762, 1530, and 3320 µg/L, respectively. Selenium, manganese, nickel, and arsenic were the least toxic (in that order), with EC50 values of 208000, 202000, 11900, and 11300 µg/L, respectively.


View this table:
[in this window]
[in a new window]
  		Table 2. Galeolaria caespitosa larval development toxicity test EC50 values (concentration of test material that affects 50% of the test organisms) for individual metals (with associated 95% confidence intervals).

 
Empirical and Model EC50 Values for Effluent Samples
Toxicity tests were conducted with 26 separate effluent samples. Empirical EC50 values ranged from 1 to 23% effluent, while EC50 values calculated with the additive model ranged from 3 to 34% (Fig. 1) . Empirical EC50 values were greater than the model EC50 values in four comparative tests and less than the model EC50 values 16 times. The confidence intervals of the empirical EC50 values overlapped the model EC50 values six times (Fig. 1). Estimated toxicity of the effluent was therefore less than actual toxicity in the majority of samples. Notwithstanding, a significant linear relationship was found between empirical and model EC50 values (r2 = 0.25) (y = 0.5x + 10.3) (P < 0.05) (Fig. 2) . Furthermore, a number of transformed regression models were tested (those available using SigmaPlot [SPSS, 2003] and Excel Excel [Microsoft Corporation, 2003] software) and the one demonstrating the strongest relationship between empirical and calculated EC50 values was chosen to transform the historical data (r2 = 0.48) (y = 5.5x0.44) (P < 0.01) (Fig. 2). This relationship was used in addition to the additive toxicity model to calculate historical EC50 values (see Historical Toxicity Associated with Effluent, below).



View larger version (14K):
[in this window]
[in a new window]
  		Fig. 1. Empirical and model Galeolaria caespitosa larval development EC50 values (concentration of test material that affects 50% of the test organisms) for 26 effluent samples from a lead smelter (95% confidence intervals are shown for the empirical EC50 values).

 


View larger version (18K):
[in this window]
[in a new window]
  		Fig. 2. Empirical EC50 values (concentration of test material that affects 50% of the test organisms) plotted against model EC50 values for 26 effluent samples from a lead smelter. The fine lines indicate the two trend lines (with associated formulas).

 
Artificial Effluent Toxicity Tests
As noted above, it was found that the additive model repeatedly underestimated toxicity of the effluent samples collected from the smelter. To further elucidate the possible cause of this underestimation, metal mixtures were prepared with seawater and intake water (Table 1). The EC50 values for the effluents prepared with intake water were slightly higher (less toxicity) than those prepared with the same metal concentrations in seawater. Additionally, removal of zinc from the metal mixtures reduced toxicity to undetectable levels (Fig. 3) . Metal mixtures were prepared to represent a range of possible concentrations in the effluent. The additive model accurately predicted the EC50 values for those metal mixtures with higher concentrations of metals (both seawater and intake water), and overestimated toxicity for the metal mixtures containing lower concentrations of metals (Fig. 3). This contrasts with the results found for the effluent samples from the smelter, for which the model tended to underestimate toxicity.



View larger version (12K):
[in this window]
[in a new window]
  		Fig. 3. Empirical and model EC50 values (concentration of test material that affects 50% of the test organisms) for eight artificial "effluent" metal mixtures (metal concentrations listed in Table 1) (95% confidence intervals are shown for the empirical EC50 values). Note: asterisk points indicate that the mixture was not toxic enough to generate an EC50 value.

 
EDTA Addition
Addition of EDTA to the effluent samples removed toxicity completely, with development in the highest concentrations not statistically significantly different from controls. This indicates that cations readily chelated by EDTA (Al, Ba, Cd, Co, Cu, Fe, Pb, Mn2+, Ni, Sr, and Zn) (Stumm and Morgan, 1981) are responsible for the majority of observed toxicity. Results from the toxicity tests with individual metals indicate that the major ions causing toxicity in the effluent samples were Cu, Zn, Cd, and Pb (in that order).

Reference Toxicity Tests
The reference toxicity tests conducted with copper run concurrently with the effluent tests indicated that the between-test variability in test species response was within acceptable levels (Cherr et al., 1994).

Dissolved oxygen, pH, salinity, and temperature were measured in each test concentration at the start and end of the toxicity tests and ranged from 7 to 9 mg/L, 7.5 to 8.9, 32 to 35 g/L, and 16 to 18°C, respectively, apart from the hypersaline brine and the undiluted smelter effluent, which ranged from 42 to 48 g/L. The wider pH range than is usual for seawater is the result of effluent pH variability.

Historical Toxicity Associated with Effluent
Historical toxicity values were calculated with the additive model and transformed using the formula:

This curve was found to best explain the variability in measured toxicity (Fig. 4) . Historically, toxicity was highest in the mid-1980s, and gradually decreased throughout the 1990s. The figure also shows a number of toxicity peaks (probably caused by spills within the smelter) in recent years.



View larger version (18K):
[in this window]
[in a new window]
  		Fig. 4. Modeled historical toxicity associated with the smelter effluent over the past 20 years.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
Empirical Toxicity Compared with Model Toxicity
The linear correlation coefficient (r2 = 0.25) found between empirical and model EC50 values is the same as that reported in a similar study (Sarakinos et al., 2000) that assessed the relationship between model and empirical toxicity of a number of different types of effluent. This model was also based on the assumption of additive toxicity of the constituent chemicals. The authors of that study reported a stronger relationship after modeling the free metal ion concentrations (r2 = 0.43) for the effluents in which 15% or more of inferred toxicity was attributed to metals or anions. The analytical data required to calculate free ion concentrations were not available for the present study.

In this study, the relationship between empirical and model toxicity was found to be stronger when a nonlinear rather than linear regression curve was fitted to the data (r2 = 0.48). The relationship inferred from this curve indicates that at higher metal concentrations, toxicity is greater than was predicted by the additive model. At lower concentrations of metals, the EC50 value is still less than that predicted by the additive model, although the difference is less marked. This is probably the result of suspended solids or other ions present in the effluent that may influence bioavailability of metals (O'Donnel et al., 1985; Allen and Hansen, 1996; Wiese et al., 1997). Bioavailability and metal toxicity are related to metal partitioning, and it is generally thought that the metal ions, rather than the total metal concentration, better predicts toxicity (O'Donnel et al., 1985; Wiese et al., 1997; Allen and Hansen, 1996). Assuming the levels of suspended solids in the effluent remained reasonably constant (although there were no data available to determine whether this is the case), at higher metal concentrations all binding sites would be exhausted, resulting in a higher concentration of free metal ions. Since it is generally agreed that the free metal ions are the bioavailable fraction, and therefore the major determinant of toxicity associated with metals in solution (O'Donnel et al., 1985), this assumption of increasing bioavailability with increasing metal concentrations would seem well-founded.

The additive model in this study underestimated the toxicity of the majority of effluent samples. These results are in contrast with those of Dyer et al. (2000) and Sarakinos et al. (2000), as both of these studies found that additive models for metal toxicity tended to overestimate effect. The authors of these studies concluded that only the free metal ions were causing toxicity and that total metal concentrations overestimate toxic effect. There are several possible explanations for our contrasting results. First, there may have been additional, unidentified toxicants (metals or organics bound to particulates) in the effluent that were not accounted for in the model. Second, the intake water was treated with a biocide (Stabrex ST70, active ingredient sodium hypobromite; Ondeo-Nalco Chemical Company, Naperville, IL), and while not toxic to the G. caespitosa larvae in the concentration it was present in the effluent, this addition may have enhanced metal toxicity. Kraak et al. (1994) suggests that contaminants present at levels below the NOEC can cause a toxic effect when present in mixtures. As Mehendale (1994) notes, "it is now widely recognized that exposure to combinations or mixtures of chemicals may result in highly exaggerated toxicity even though the individual chemicals might not be toxic." Third, there could be synergistic interactions between the metals in the effluent. A study assessing the toxicity of Cu, Zn, and Cd mixtures (the three major metals in the effluent examined in this study) reported that either a greater than additive or less than additive response is elicited depending on the ratio of the concentrations of these metals (Wang et al., 1995). The variability in results of other studies that focused on the toxicity of metals in mixture (e.g., Preston et al., 2000; Reed-Judkins et al., 1997; Kraak et al., 1994; Pavicic et al., 1994) underscores the complexity of interactions that could occur between the components in a complex effluent.

The additive model either accurately predicted (Samples 1, 2, 5, 7) or overestimated (Samples 3, 4, 6, 8) toxicity of the artificial "effluent" metal mixture samples, both those prepared with seawater and those with intake water. This suggests that it is unlikely that the metals in their current concentrations are overtly synergistic. However, we could not exactly reproduce the chemical changes the effluent may undergo as it passes through the smelter, which excludes ruling out synergistic interactions between trace metals. This is also true of the biocide, which was present in the intake water used in the metal mixture toxicity tests.

Historical Toxicity Associated with Effluent
Further toxicity identification and evaluation (TIE) could be undertaken to determine the nature of the unknown toxicants that are causing the model-derived EC50 values to underestimate toxicity. However, since the concentrations of metals associated with the effluent are extremely high, their contribution (particular zinc, copper, and cadmium) to acute toxicity is of greatest concern. Additionally, given that scientifically sound toxicity data have been established for less than 0.1% of listed chemicals (Wharfe, 1996), the effort and expense in determining the unknown compound(s) seems unjustified. As an alternative, it is possible to incorporate the contribution of the unknown toxicants into the additive model based on the relationship between empirical and calculated toxicity, and from this, estimate trends in historical toxicity. This was done by using the nonlinear relationship between empirical and calculated toxicity to estimate historical toxicity for each day over the past 20 years. From this, trends can be determined, and while the accuracy of these estimates is somewhat compromised by the use of transformed regression modeling to incorporate the unknown toxicants(s), there is value in these estimates, particularly identifying notable trends. Figure 4 presents the modeled historical toxicity associated with the smelter effluent for the past 20 years. Historically, toxicity was highest in the mid-1980s, and gradually decreased throughout the 1990s. The figure also shows a number of toxicity peaks (probably caused by spills within the smelter) in recent years. The most notable feature of this figure is that large amounts of analytical data have been reduced to a single, easily interpreted graph. However, the historical toxicity is only an estimate, based on the relationship between additive, predicted EC50 values and empirical EC50 values. Changes in the concentrations of metals (as determined by the model EC50 values) were found to be responsible for just under half the variability in the empirical EC50 values, and leaves 52% unaccounted for. Additionally, it is likely that analytical methods used at the smelter over the past 20 years have altered as equipment accuracy and precision have improved, although there is no way to test this assumption. There are examples of temporal water quality changes being a result of changes in analytical methods rather than actual changes in concentration (Chapman, 1996). Therefore, while the model allows some assessment of historical toxicity, it must be interpreted with caution.

Whole-Effluent Toxicity Testing
Effluents are complex mixtures, the composition (and therefore toxicity) of which can change over weeks, months, or even days (Dorn and van Compernolle, 1995). Toxicity identification and evaluation can identify the major toxicants associated with chemical mixtures, and from this, models can estimate the toxicity associated with mixtures based on measured concentrations of those components identified. However, whole-effluent toxicity (WET) testing is the only way to determine the true toxicity associated with chemical mixtures, since they act as integrators of the interactions of chemicals in combination (Chapman, 2000) and it is not always possible or necessary to determine all components present in chemical mixtures. Additionally, while there was reasonable agreement between empirical and model toxicity, 52% of toxicity associated with the effluent samples remained unexplained by chemical concentrations measured in each effluent sample. Because of the uncertainties associated with complex effluents, analytical data cannot be a substitute for whole-effluent toxicity testing. However, when whole-effluent toxicity testing is not possible (e.g., when only historical data are available), analytical data may provide a means to educe overall trends in toxicity.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
Many industrial facilities have extensive chemical data for their wastewater effluents. It is possible to use such data to estimate past trends in toxicity by calculating endpoints with an additive model. Trends associated with model predictions can be characterized by comparing model-derived EC50 values with those from actual toxicity tests, and incorporating those differences within the estimation model. This approach can then be used to evaluate how processing changes have influenced effluent quality over time.


   ACKNOWLEDGMENTS
 
The authors thank Dr. Chris Dyson (School of Geoinformatics, University of South Australia) for his statistical advice, and Naomi Cooper and Daniel Bellifemine for their assistance with toxicity tests.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2003 32: 1577-1582. [Full Text]  




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 Google Scholar
Google Scholar
Right arrow Articles by Ross, K. E.
Right arrow Articles by Bidwell, J. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Ross, K. E.
Right arrow Articles by Bidwell, J. R.
Agricola
Right arrow Articles by Ross, K. E.
Right arrow Articles by Bidwell, J. R.
Related Collections
Right arrow Heavy Metals
Right arrow Other Models
Right arrow Water Pollution

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Soil Science Society of America Journal Journal of Plant Registrations The Plant Genome