DECIDING WHEN TO INTERVENE

Almost all the Great Lakes Areas of Concern have documented sediment contamination. Current sediment guidelines are based on the comparison of chemical concentrations at a site with guideline concentrations that have been established as representing a perceived safe concentration on a chemical by chemical basis. However, current practitioners generally acknowledge that chemical concentration alone is insufficient to determine sediment contamination and that biological information is also essential to determine sediment conditions. The purpose of environmental assessment and management is ultimately the maintenance of biological integrity, so it is our view that the setting of water and sediment quality objectives should include biological targets together with chemical surrogates. This approach is the basis of the sediment quality triad proposed by Chapman and co-workers (Chapman et al. 1992) and strongly endorsed in two International Joint Commission (1987; 1988) reviews of assessing sediment problems in the most contaminated areas of the Great Lakes. Both the sediment quality triad and the IJC promote the incorporation of laboratory and field biological assessment in identifying contaminated sediment. In both cases the use of invertebrate assemblage structure is suggested as the appropriate field component and toxicity testing as the laboratory component. This paper describes the development of numeric target values for these biological measures.

There are two basic assumptions behind these biological sediment guidelines. The first is that the effects of sediment contamination on biological processes are the primary concern, and therefore, assessment of biological effects is paramount. The second is that the complexity of the sediment matrix makes chemical concentration a poor predictor of the biological availability of contaminants.

The biological sediment guidelines incorporate the structure of benthic invertebrate communities by using predictive models that relate site habitat attributes to an expected community, and functional responses (survival, growth, and reproduction) in four sediment toxicity tests (bioassays) with benthic invertebrates using ten test endpoints. For both community structure and toxicity, guidelines have been established that allow determination of the community as either unstressed, potentially stressed, stressed, or severely stressed, and the sediment as either non-toxic, potentially toxic, or toxic.

To simplify the assessment process, the BEAST software has been developed, which incorporates the complex multivariate analysis required by this approach and presents the user with straightforward categories of sediment quality on a site by site basis. Designed for the BEnthic Assessment of SedimenT, the software automates the methodology and employs the RAISON Mapping and Analysis package from Environment Canada as a foundation. BEAST combines new methods with a simple, straightforward software user interface. The result is a powerful new tool for sediment assessment.

Reference condition concept

Until recently, the development of numeric biological targets was considered too difficult due to the temporal and spatial variability inherent in biological systems. However, over the past 15 years, methods developed in the United Kingdom (Wright et al. 1984) and elsewhere (Corkum and Currie 1987) have demonstrated the ability to predict the biological response in clean (or 'uncontaminated') sites using simple habitat and water quality parameters. In all these studies, the biological attributes of choice have been invertebrate assemblages. This approach allows appropriate site-specific biological objectives to be set for ecosystems from measured habitat characteristics, and also provides an appropriate reference for determining when degradation at a site due to anthropogenic contamination is occurring. The acceptance by regulatory agencies of biological water and sediment quality objectives has been slow, but is now being given serious consideration as shown by current work in Canada (Reynoldson and Zarull 1993; Reynoldson et al. 1995), the USA (Hunsaker and Carpenter 1990), the United Kingdom (the RIVPACS method; Wright et al. 1984) and Australia (Parsons and Norris 1996).

This paper describes the development of biological guidelines for sediment in nearshore fine-grained habitats in the Laurentian Great Lakes. These guidelines have been developed for invertebrate assemblages and benthic invertebrate laboratory tests using a modification of the technique developed in the UK (Wright et al. 1984) and now described as the reference condition concept (for more detail, see Reynoldson et al. 1997). The choice of invertebrate assemblages was made on the basis of the fact that these organisms are in direct contact with the contaminants associated with the sediment, and are therefore most likely to exhibit effects. The use of laboratory tests was supported to confirm that any responses observed in the field are due to sediment and not other environmental stressors. In selecting the test organisms and endpoints, it was the view that infaunal invertebrate species would be most appropriate, and that ecologically relevant (growth and reproduction) chronic, as well as acute, endpoints should be used.

Fundamental to the scientific method is the use of controls or control conditions against which results obtained under test conditions can be compared. In field comparisons, attempts are made to choose test and control sites that are as similar as possible. The variable of interest can then be manipulated, but uncontrolled variables are assumed to fluctuate. The actual choice of separate sites in the field that are similar in all aspects and that can be divided into control and experimental sites is difficult. Traditionally, this problem has been solved in aquatic studies by choosing adjacent sites in streams (i.e., upstream and downstream comparisons), dividing lakes into halves, using artificial enclosures or mesocosms, or by locating sites thought to be similar at an appropriate distance from any source of contamination. Such approaches have several problems, especially the problem of "pseudoreplication". In the reference condition approach, a wide range of minimally disturbed sites are sampled and organized by selected physical, chemical, and biological characteristics to form one or more reference conditions. These reference conditions then serve as the control(s) against which individual test sites can be compared. The notion of a reference condition is therefore really a description of best available condition.

Using the reference condition approach in developing biological guidelines for the Great Lakes involves the following steps:Data collection. Collection of data on invertebrate assemblages, sediment toxicity tests, and habitat descriptors from reference sites that describe the broadest range of natural variation in fine-grained sediment from the nearshore of the Great Lakes must occur.

Site classification and model building. Reference sites are organized into groups with similar biological attributes based either on the composition of their invertebrate fauna or the response in the laboratory test endpoints. The characteristics of these community groups and the test endpoint ranges form the bases for the guidelines.

Predictive models are developed that relate a set of habitat attributes to the groups of sites formed from the biological data. The models are used to determine the probability of a test site belonging to individual reference site groups.

The data collection, reference site classification, and model building are required to develop the guidelines and are a one time effort. However, the models can be refined and periodically upgraded as further data are collected.

The following steps are used in the assessment of sediment quality using the biological sediment guidelines:

Selection of reference sites for comparison. A statistical technique, discriminant function analysis (DFA), with physio-chemical variables is used to determine the probability of a test site belonging to one or more of the reference groups.

Test site assessment. This is the step that defines whether the biological response at a test site meets expectation, and compares the biological attributes of the test site with the normal range observed at the appropriately matching reference sites.

In the Great Lakes study (Reynoldson and Day 1998), a large database was assembled from 312 locations in Lakes Ontario, Erie, Michigan, Superior, and Huron and analyzed to establish reference conditions. Information from each site included: the responses of four species of benthic invertebrates (Hyalella azteca, Chironomus riparius, Hexagenia sp. and Tubifex tubifex) exposed in the laboratory, the structure of the benthic invertebrate community, and selected environmental variables from the same site.

Using multivariate statistical methods, 6 community assemblages have been characterized in the Great Lakes. Using discriminant function analysis with 12 habitat descriptors, reference sites can be predicted to one of these 6 assemblages with confidence (average error rate 12%). Thus the invertebrate community at any test site in the Great Lakes can be predicted from the measurement of 12 easily and inexpensively measured habitat attributes (latitude, longitude, depth, alkalinity, pH, TN, TOC, K2O, CaO, MgO, MnO, SiO2). Assessment of the condition of the invertebrate community at a test site involves a simple comparison of the community at the test site with the communities occurring at the group of reference sites to which the test site is predicted as belonging.

The actual comparison is done by reducing the species matrix to ordination vectors. This is because 162 species have been identified in the Great Lakes, and species by species comparison is impractical. Therefore, the reference communities are described by three ordination axes and can be plotted graphically as a site "cloud" in ordination space. By plotting the test site with the reference sites "cloud", the similarity to the reference sites can be determined by using probability ellipses for the reference sites only and examining the position of the test site relative to the reference site ellipses. A site is defined as "equivalent to reference" if it is within the 90% probability ellipse, "possibly different" if it is between the 90 and 99% ellipse, "different" if between the 99 and 99.9% ellipse and "very different" if outside the 99.9% ellipse (i.e., less than a 1 in 1000 chance of error).

Determining toxicity was done in the same way, even though individual toxicity test endpoints showed little variability and single guideline values can be established for each endpoint, the overall assessment of sediment toxicity uses this multivariate approach as it permits the integration of multiple endpoints.

Cornwall

Twelve sampling sites were selected to assess the potential biological effects of sediment associated contaminants. Nine sites were located within depositional areas investigated in 1994, two sites (175, 179) within a deposit north of Cornwall island, and one site (167) at a boat launch deposit. In 1977, the IJC identified this part of the St. Lawrence River as an Area of Concern based upon input from the Ontario Ministry of the Environment. Restrictions to beneficial use included possibly impaired benthic invertebrate communities and restrictions on dredging because of sediment contamination (Dreier et al. 1997).

The National Water Research Institute (NWRI) of Environment Canada, with the support of the Cornwall RAP team and Ontario Region of Environment Canada, conducted sampling of the sediment in Cornwall in October 1997 from which data were used in assessment of community structure and sediment contamination with biological sediment guidelines developed by NWRI and Ontario Region of Environment Canada (Reynoldson and Day 1998).

Invertebrate Community Structure. All 12 sites were predicted to reference Group 2 based on the habitat attributes, with the probability of group membership ranging from 51.5% (site 175) to 82.0% (site 127).

Reference Group 2 is characterized by the fingernail clam Pisidium casertanum and the amphipod Diporeia hoyi, and this reference group represented more oligotrophic conditions (Reynoldson and Day 1998). Nine taxa were common (>50% occurrence) at the reference sites. Of these, all but one, the tubificid worm Potamothrix vejdovskyi, was found at the Cornwall sites. The occurrence of 120 genera at the 39 reference sites comprising Group 2 was compared with the 12 test sites sampled in Cornwall. Two very simple descriptors of community structure are taxa richness and total abundance. The range observed at the reference sites is shown in Table 1 together with the results from the Cornwall sites. These data show the overall diversity and abundance to be well within the range observed at reference sites and a general trend to greater diversity and abundance. The results of the multivariate analysis are summarized in Table 1. Four sites (127, 164, 175, 179) were outside the 90% probability ellipse, and none are outside the 99% ellipse. The 90% ellipse is defined as an area where sites may be considered as different to the reference sites, with a Type 1 error of 10%. The sites showing possible stress all have high diversity and abundance.

Table 1. Summary of taxonomic composition of benthic invertebrates at Group 2 reference sites and 12 Cornwall test sites (Occurrence at reference sites is based on percentage of those sites at which a taxon was present; abundance is expressed in terms of numbers per 34.2 cm²)

Site	Taxa richness No. of genera mean (SD)	Total abundance No. per 34.2 cm2 mean (SD)	Abundance and occurrence of: Procladius	Abundance and occurrence of: Pisidium	Abundance and occurrence of: Piona	BEAST assessment community
Reference (n=39)	20.5 (9.8)	60.6 (46.1)	1.94 (84.6%)	6.24 (82.1%)	0.0 (0.0)
105	19	38.25	5.2	4.6	1	Unstressed
109	19	35.2	7.2	4.2	0.2	Unstressed
117	20	32.4	6.8	4.4	0.8	Unstressed
127	24	98.6	3.8	0.6	1.2	Possibly stressed
128	37	106.4	2.6	12.6	0.6	Unstressed
131	20	51.8	5.8	10	0.8	Unstressed
132	25	48	4.4	8	0	Unstressed
156	26	49.8	2.2	3.6	0	Unstressed
164	17	30	3.6	0	1.2	Possibly stressed
167	28	74	4.8	9.4	0.8	Unstressed
175	24	81.6	11	0	0.4	Possibly stressed
179	26	97.8	14.4	0	0.6	Possibly stressed

Sediment toxicity. The results from the 10 test endpoints showed only two endpoints to be below the warning levels derived from the reference sites. These are both related to reproduction in the tubificid oligochaete Tubifex tubifex. Five sites (105, 127, 156, 164, 167) had a reduced rate of cocoon hatching, suggestive of impairment in the embryogenesis of the worm eggs. All five sites, not unexpectedly, had reduced young also. However, a further two sites (117 and 128) had reduced young per adult. The other three species show no evidence of sediment toxicity.

Comparison of community and toxicity data. The assessments of community and toxicity effects are summarised in Table 2. In addition, the habitat attributes that either exceed Ontario sediment criteria (Persaud et al. 1992) or are outside the range observed at the reference sites are also identified.

In general there is no strong evidence for either impaired invertebrate communities or any associated sediment toxicity. While several variables exceed the OMOE low effect level criteria, this also occurs at reference sites. At six sites, Hg exceeded the Severe Effect Level and at Site 109, Zn also exceeded the severe effect concentration (Table 2). However, this site and three of the other six showed no indication of either toxicity or impaired community structure. There is some effect on reproduction of the worm Tubifex. This may account for the absence of the worm P. vejdovski, a species present at many (53%) of the reference sites. However, immature worms were found at all but one site (164) and the apparent absence of Potamothrix is likely due to the absence of identifiable mature animals. In conclusion, these data do not indicate a sediment contamination problem associated with the samples taken form the Cornwall area.

Table 2. Summary of sediment quality based on invertebrate community structure, sediment toxicity, and sediment chemistry

Site	BEAST Assessment		Variables Exceeding OMOE Sediment Criteria		Variables > 2 SD than reference
	Community	Toxicity	Low	Severe
105	Unstressed	Possibly toxic	TP, TOC, Cr, Ni, Cu, Zn, Pb, Hg		Zn
109	Unstressed	Non toxic	TP, TOC, Cr, Ni, Cu, Zn, Pb	Zn, Hg	Cu, Zn, Pb
117	Unstressed	Possibly toxic	TP, TOC, Cu, Zn, Pb, Hg		Zn
127	Possibly stressed	Possibly toxic	TP, TOC, Cu, Zn	Hg	Gravel, 25% PS
128	Unstressed	Non toxic	TP, TOC, Cu, Zn	Hg
131	Unstressed	Non toxic	TP, TOC, Cr, Ni, Cu, Zn, Pb	Hg	Cu, Zn
132	Unstressed	Non toxic	TP, TOC, Cr, Ni, Cu, Zn, Pb	Hg
156	Unstressed	Possibly toxic	TP, TOC, Cu, Zn, Hg
164	Possibly stressed	Non toxic	TP, TOC, Cr, Ni, Cu, Zn, Pb	Hg	Cu, Zn
167	Unstressed	Possibly toxic	TP, TOC, Ni, Cu, Zn, Hg
175	Possibly stressed	Non toxic	TP, TOC, Cr, Ni, Cu, Zn
179	Possibly stressed	Non toxic	TP, TOC, Cr, Ni, Cu, Zn

References

Chapman, P. M., Power, E. A. and G.A. Burton, Jr. 1992. "Integrative Assessments in Aquatic Ecosystems." Sediment Toxicity Assessment. G.A. Burton, Jr., ed. Lewis Publishers: Chelsea, MI. pp. 313-340.

Corkum, L. D. and D. C. Currie. 1987. "Distributional Patterns of Immature Simuliidae (Diptera) in Northwestern North America." Freshwater Biology. 17:201-221.

Dreier, S., Anderson, J., Biberhofer, J., Eckersley, M., Helliar, R., Hickey, M. B. C., Richman, L., Stride, F. and the St. Lawrence River RAP Public Advisory Committee. 1997. Remedial Action Plan for the St. Lawrence River (Cornwall) Area of Concern Stage 2 Report: The recommended plan. Ontario Ministry of Environment and Environment Canada. Toronto, Ontario.

Hunsaker, C. T. and D. E. Carpenter, eds. 1990. Ecological Indicators for the Environmental Monitoring and Assessment Program. EPA 600/3-90/060. U.S. Environmental Protection Agency, Office of Research and Development. Triangle Park, N.C.

International Joint Commission. 1988. Procedures for the Assessment of Contaminated Sediment Problems in the Great Lakes. Report to the Great Lakes Water Quality Board. Windsor, Ontario. 140 p.

International Joint Commission. 1987. Report on Great Lakes Water Quality. Report to the Great Lakes Water Quality Board. Windsor, Ontario. 236pp.

Johnson, R. K. and T. Wiederholm. 1989. "Classification and Ordination of Profundal Macroinvertebrate Communities in Nutrient Poor, Oligo-mesohumic Lakes in Relation to Environmental Data." Freshwater Biology. 21: 375-386.

Parsons, M. and R. H. Norris. 1996. "The Effect of Habitat-specific Sampling on Biological Assessment of Water Quality Using a Predictive Model." Freshwater Biology. 36: 419-434.

Persaud, D., Jaagumagi, R. and A. Hayton. 1992. Guidelines for the Protection and Management of Aquatic Sediment Quality in Ontario. ISBN 0-7729-9248-7. Ontario Ministry of the Environment, Water Resources Branch. Toronto, Ontario.

Reynoldson, T. B. and K. E. Day. 1998. Biological Guidelines for the Assessment of Sediment Quality in the Laurentian Great Lakes. NWRI Report No. 98-232. 119 pages + appendices.

Reynoldson, T. B., Rosenberg, D. R, Day, K. E., Norris, R. H., and V. H. Resh. 1997. "Use of the Reference Condition Concept in Water Quality Assessments Using Benthic Invertebrates." Journal of North American Benthic Soc. 16:833-852.

Reynoldson, T. B., Day, K. E., Bailey, R. C. and R. H. Norris. 1995. "Biological Guidelines for Freshwater Sediment Based on Beenthic Assessment of SedimenT (the BEAST) Using a Multivariate Approach for Predicting Biological State." Australian Journal of Ecology. 20: 198-219.

Reynoldson, T. B. and M. A. Zarull. 1993. "An Approach to the Development of Biological Sediment Guidelines." Ecological Integrity and the Management of Ecosystems. S. Woodley, J. Kay and G. Francis (eds). St. Lucie Press: Del Ray Beach, Florida. pp. 177-200.

Wright, J. F., Moss, D., Armitage, P. D. and M. T. Furse. 1984. "A Preliminary Classification of Running-water Sites in Great Britain Based on Macroinvertebrate Species and the Prediction of Community Type Using Environmental Data." Freshwater Biology. 14:221-256.