SEARCH

SEARCH BY CITATION

Keywords:

  • Sediment toxicity;
  • Benthic infauna

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION AND CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

Autesedimenttoxicity testshave become important in regulatory, monitoring, and scientificprograms, partly because it has been assumed that they are indicative of ecological damage to benthic infaunal resources. Data from tests of sediment toxicity and measures of benthic community structure were examined from > 1,400 saltwater samples to determine the relationships between acute toxicity and changes in the abundance and diversity of infauna resources. Data were compiled from studies conducted along portions of the Atlantic, Gulf of Mexico, and Pacific coasts of the United States. There was considerable variability among the data sets in the relationships between laboratory results and benthic measures. However, in 92% of the samples classified as toxic, at least one measure of benthic diversity or abundance was <50% of the average reference value. In 67% of these samples, at least one measure of benthic infauna abundance or diversity was <10% of average reference conditions. No amphipods were found in 39% of samples that were classified as toxic, whereas amphipods were absent from 28% of the nontoxic samples. In many survey areas, the abundance of crustaceans (notably the amphipods) decreased in the infauna as amphipod survival decreased in the laboratory tests. There appeared to be a break point in the data indicating that, generally, amphipod abundance in the field was lowest when survival in the laboratory tests dropped below 50% of controls. Based on the weight of evidence from all the data analyses, we conclude that ecologically relevant losses in the abundance and diversity of the benthic infauna frequently corresponded with reduced amphipod survival in the laboratory tests.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION AND CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

Laboratory tests to determine the toxicity of sediments have become components of assessments of sediment quality frequently used in many monitoring, regulatory, and scientific programs [1]. Sediment toxicity tests can be performed with relatively simple bioassays, can be designed to minimize the effects of naturally occurring properties of sediments, and can provide a rapid and integrated measure of the toxicological significance of sediment-bound contaminants [2]. In dredged material assessments, they represent a key factor in determining whether sediments are suitable for unconfined, open-water disposal. The operational manual for dredged material assessments states, “Thus, a statistically significant result in… the dredged material in question causes a direct and specific biological effect under test conditions and, therefore, has the potential to cause an ecologically unacceptable impact” [3]. Furthermore, it states, “Although bioassays are not precise predictors of environmental effects, they are regarded as the best methods available for integrating the effects of multiple contaminants.” Risk assessors at hazardous waste sites are encouraged to select biological tests, such as mortality in sediment toxicity tests, that reflect the ecological end points of concern at the site, including risks of mortality and losses of abundance of sensitive taxa [4].

However, despite the increased emphasis on using these tests, their ecological relevance has not been determined empirically in controlled, cause-effect experiments [5]. It is uncertain whether results of acute tests, such as those used in dredged material assessments, protect against adverse population-level effects [6]. There are no theoretical ecological models with which to predict ecological effects with toxicological data, and as a result, “ultimately, sediment toxicity test end points must be shown to be predictive of community- and ecosystem-level responses” [2]. Users of the dredged material testing manuals are cautioned [3] that “mortality of a certain percent of the organisms of a particular species in a laboratory test does not imply that the population of that species around the disposal site would decline by the same percent if the proposed disposal takes place.” It has been assumed by some [7] that laboratory toxicity tests are primarily responsive to naturally occurring sedimentological properties, such as grain size and therefore they are of limited value in regulatory and monitoring programs.

Very little information has been analyzed with which to actually evaluate the relationships between measures of acute toxicity and the abundance and diversity of benthic assemblages that the tests were intended to protect. Using an experimental approach, it was determined that amphipods responded the same way to doses of cadmium in laboratory tests and in field chambers; however, actual measures of the abundance or diversity of any taxonomic group in the field were not reported [8]. Data from a few field studies have indicated that amphipods are among the first taxa to either diminish in abundance or disappear from benthic communities when exposed to pollution [9]. In a study conducted off southern California, USA, for example, good correspondence between chemical contamination of sediments, decreased survival of amphipods in laboratory tests, and decreased amphipod abundance in the benthic infauna was reported [10]. An initial analysis of matching toxicity and benthic infaunal data from a few U.S. estuaries suggested that the numbers of species often decreased in the infauna, usually as a result of the losses of crustaceans, as survival in amphipod tests decreased in the laboratory [11].

The purpose of this paper was to determine the relationships between results of acute tests of toxicity in laboratory bio-assays and the frequency and degree of changes, if any, in benthic diversity and abundance. In this study, a number of independent data sets developed in surveys conducted in U.S. estuaries were compiled and reviewed to determine these relationships. Multiple data sets were analyzed to provide a weight of evidence applicable to all three U.S. coastlines and to satisfy as many criteria as possible for establishing reasonable cause-effect relationships [12]. In these analyses, we attempted to determine whether measures of diversity and/or abundance in the benthic infauna decreased, and the degree to which they decreased, as survival decreased in the laboratory tests. In addition, we attempted to discover whether there was a break point in percentage survival below which benthic resources indicated the most severe losses.

The purpose of this study was not to provide a field validation of laboratory tests of sediment toxicity. We are in agreement with Chapman [13] that such field validations are not necessary.

In this review, we fully recognized the importance of natural variables that can influence the structure of benthic infaunal communities. The data we analyzed were invariably collected in harbors, bays, and estuaries or along the continental shelves in which there were differences among sampling locations in salinity, depth, sediment texture, organic carbon content, and many other factors that could have influenced the structure of the benthos. The diversity and abundance reported in the benthic samples could have been influenced by predation, distance from brood stocks, near-bottom hypoxia, long-term succession, and storm events. We were reminded of the drawbacks of taking this correlative, empirical approach [8]. Nevertheless, we chose to explore this approach to determine objectively whether there were any relationships between survival as measured in the laboratory and the abundance and diversity of the infauna in the field and the nature of those relationships.

These data analyses were performed with results from amphipod survival tests. Amphipods are the most commonly used taxa in tests of sediment toxicity in North America [1,2,5]. Therefore, by focusing on data from these tests, we were assured of encountering information from many geographical areas and representing a wide range in environmental conditions. Data from chronic (life cycle) tests or sublethal end points were insufficient to warrant equivalent analyses.

METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION AND CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

Most data were obtained from surveys conducted by either the U.S. Environmental Protection Agency (U.S. EPA) or the National Oceanic and Atmospheric Administration (NOAA) or their partners. Both agencies and their partners used similar methods in the field and laboratory, thereby ensuring that our analyses would be performed on comparable data sets. Data sets were analyzed only if they indicated a wide range in survival and if they included data from analyses of infauna collected at the same sites and at the same times. The possible effects of temporal or spatial heterogeneity were minimized by synoptically collecting samples for both toxicity tests and infaunal analyses. Except in a few studies noted in Results, amphipod test results were treated as control-adjusted percentage survival. Except in a few studies noted in Results, mean control-adjusted amphipod survival of less than 80% was used as the benchmark for classifying samples as toxic [14].

The benthic endpoints reported often differed among studies as a function of individual objectives and study designs. Generally, benthic data were available for measures of total abundance, total numbers of species, proportion of total abundance contributed by sensitive taxonomic groups such as crustaceans or amphipods, or multiparameter benthic indices. Unless specified otherwise, all the benthic data were reported on a per sample basis (e.g., total species or total abundance per sample).

Four types of data analyses were conducted on each data set. First, to determine whether measures of benthic community structure were concordant with bioassays of survival, we calculated the averages of benthic community parameters within ranges of percentage survival for each case study. These data were illustrated either in tables or, where sufficient data warranted, in histograms. In some of the smaller data sets, benthic data were compared only between samples classified as toxic and nontoxic. Second, to determine how frequently benthic indices were depressed when samples were toxic in the laboratory, we calculated the proportion of toxic samples in which at least one measure of benthic diversity or abundance was less than either 50 or 10% of the average values determined for reference samples (see the following explanation). Third, to determine whether amphipod mortality predicted the absence of amphipods in the benthos, we calculated the proportion of toxic samples in which amphipods were absent: either the species used in the toxicity tests or all species, depending on what was reported. Fourth, to determine whether amphipod abundance in the benthos universally corresponded with amphipod mortality, we calculated average abundances of benthic amphipods among 11 ranges in amphipod survival with the data combined from all studies.

In the second type of analysis, we wanted to contrast benthic characteristics in toxic samples with those that represented the most abundant and diverse assemblages. Therefore, reference conditions for benthic parameters were calculated for each case study as the averages of the highest 10% of the values for each benthic measure or, in the case of small studies, the top three samples, whichever provided the largest sample size. This approach was used to identify the most abundant and diverse samples in the study as the basis for comparisons, recognizing that such samples may inflate the definition of reference conditions relative to what might have been determined by a cluster analysis or with a reference envelope approach. Unlike the alternative approaches, we believed that the selected approach would yield results that would be comparable among studies while recognizing that the selection of the top 10% or top three samples seemed arbitrary. Uses of either the chemical data or toxicity results to define reference conditions for the benthos seemed circular in logic. That is, we elected to define reference benthic conditions with the benthic data, not with other kinds of data that may or may not have been correlated with the benthic measures.

Data were evaluated from 14 survey areas along the Atlantic, Pacific, and Gulf of Mexico coasts of the United States (Fig. 1). Data from a total of 1,463 samples were reviewed. The data sets ranged in size from nine samples to 634 samples. There were differences in the methods among the studies that could have led to differences in the data. In the U.S. EPA and NOAA studies, as well as most other studies, samples were collected with 0.01 to 0.04 m2 grab samplers at each station and the benthos retained on 0.5-mm sieves. However, in other studies, samplers ranged in size from 0.015 to 0.1 m2. In some cases, stacked 0.5- and 1.0-mm sieves were used. In all cases, no measures of within-site variance were reported.

thumbnail image

Figure Fig. 1.. Locations of survey areas in U.S. estuaries and marine bays within which data were collected for analysis.

Download figure to PowerPoint

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION AND CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

NOAA data from Puget Sound, Washington, and San Francisco Bay, California, USA

Chemical, bioassay, and benthic data were developed for 14 sites scattered throughout portions of Puget Sound, Washington, during a triad study done in 1983 [15]. Equal numbers of samples were classified as toxic (significantly different from controls, p < 0.05) and nontoxic, although percentage survival was not reported. The data indicated that the frequency of toxicity in amphipod survival tests (Rhepoxynius abronius) increased with increasing chemical concentrations [11]. The average abundances of arthropods, echinoderms, and amphipods were higher by factors of roughly 2 to 10 in nontoxic samples relative to those that were toxic (Table 1). Notably, there were no phoxocephalid amphipods (the family that includes R. abronius) in the toxic samples.

A similar pattern was observed among data compiled for 17 samples collected during two triad studies performed in several regions of San Francisco Bay (northern California) [16,17]. Those data indicated that the abundance of all animals and crustaceans decreased with decreasing amphipod survival, but average numbers of species did not change (Table 2). The nontoxic samples (survival >90%) had abundant crustacean populations, whereas those that were toxic (survival <80%) were dominated by polychaetes and mollusks and included very few amphipods and other crustaceans [11].

U.S. EPA data from Puget Sound, Washington; southern California; and San Francisco Bay (northern California)

Results of chemical, bioassay (using R. abronius), and benthic analyses were reported by the U.S. EPA for Commencement Bay adjoining Puget Sound, the Palos Verdes shelf off southern California, and Richmond Harbor adjoining San Francisco Bay [10,18,19]. Similar patterns in the data were apparent among all three studies: The abundance of all amphipods and phoxocephalid amphipods decreased in samples that were toxic relative to those that were nontoxic. In all three studies, the average abundance of these groups in toxic samples (significantly different from controls, p < 0.05) was about one-half to one-fifth of that in nontoxic samples. An analysis of data collected in Commencement Bay [18] indicated that average densities (number per 0.1 m2) of all amphipods and phoxo-cephalids were 8.3 and 3.1, respectively, in nontoxic samples and decreased to 4.7 and 0.0, respectively, in samples classified as toxic (significantly different from controls, p < 0.05) [11]. Average total abundance and numbers of species in the three toxic (significantly different from controls, p < 0.05) samples from Palos Verdes [10] were about one-quarter of that in the six nontoxic samples, and no phoxocephalids were observed in the toxic samples (Table 3). In Richmond Harbor [19], average abundance of amphipods (excluding Grandidierellajaponica) was 49 organisms per 0.1 m2 in nontoxic samples and nine organisms per 0.1 m2 in samples that were toxic to R. abronius [11]. Unexpectedly, the abundance of G. japonica in the benthos increased among samples as chemical concentrations and toxicity increased.

Table Table 1.. Average (± standard deviation) percentage contribution of taxonomic groups to total benthic abundance in samplesa from Puget Sound, Washington, USA, that were either toxic or nontoxic in tests with Rhepoxynius abronius [15]
 % Total abundance
Amphipod survivalPolychaetes and mollusksArthropodsEchinodermsPhoxocephalids
  1. a Total number of samples = 14.

  2. b Significantly different from controls, p < 0.05.

Nontoxic, n = 769.8 ± 32.219.2 ± 26.010.0 ± 15.11.9 ± 2.8
Toxicb, n = 795.5 ± 2.23.7 ± 2.10.2 ± 0.40.0
Table Table 2.. Average (± standard deviation) total abundance, species richness, and crustacean abundance (as percentage of total abundance) within three ranges in amphipod (Rhepoxynius abronius) survival in samplesa from San Francisco Bay, California, USA (data combined from [16,17])
AmphipodTotal abundanceTotal no.Abundance of crustaceans
  1. a Total number of samples = 17.

survival(no./0.1 m2)of species(% of total)
>90% (average 95.8%, n = 5)2,591.5 ± 1,508.113.5 ± 4.691.2 ± 5.5
80–89.9% (average 82.2%, n = 3)1,590 ± 1,659.011.8 ± 7.260.4 ± 41.8
<80% (average 48.8%, n = 9)666.6 ± 921.116.4 ± 6.337.3 ± 32.4

Chesapeake Bay, Maryland, USA

Data collected for 20 samples by the U.S. Fish and Wildlife Service and University of Maryland (Baltimore, MD, USA) for northern Chesapeake Bay (including Baltimore harbor) showed a strong association between the concentrations of polynuclear aromatic hydrocarbons in sediments and toxicity in tests performed with Leptocheirus plumulosus [20]. The abundance of the test species was reduced by approximately two orders of magnitude in the highly toxic samples (survival <60%) relative to the nontoxic samples (survival >91%) (Table 4). The average abundance of L. plumulosus in the highly toxic samples was about 0.1% of that in the reference samples (the three samples in which abundance was highest).

Hudson-Raritan estuary, New York/New Jersey, USA

In NOAA's survey of the Hudson-Raritan estuary [21], matching bioassay (using Ampelisca abdita) and benthic data were reported for 73 stations scattered throughout the region. There was no apparent relationship between percentage amphipod survival and either total abundance or total numbers of species (Table 5). The data indicated that many samples with low survival had an abundant population of worms, such as Capitella capitata, thus masking the relationship between toxicity and total abundance. However, although there was considerable variability in results (as noted with the large standard deviations), the average abundance of arthropods and the test species (A. abdita) decreased in samples as survival decreased. Sixty-five percent of samples in which survival equaled or exceeded 80% were devoid of amphipods in the field, and amphipods were absent in 64% of samples in which survival was 50 to 79.9%. In contrast, 100% of the 17 samples in which survival was <50% were devoid of the amphipods A. abdita.

Biscayne Bay, Florida, USA

Data collected in Biscayne Bay by NOAA during 1995 [22] were assembled for the region near and within the Port of Miami, lower Miami River, and vicinity. Samples were tested for toxicity with A. abdita. The data from analyses of 22 samples (Table 6) showed that average total abundance, total numbers of species, arthropod abundance, amphipod abundance, and ampeliscid abundance were very high among 15 samples in which amphipod survival was high (≥77%, average of 96%). In seven samples that were toxic (<50% survival, average of 25%), total abundance was very similar to abundance in nontoxic samples. However, the numbers of benthic species decreased markedly along with order-of-magnitude losses in abundance of all arthropods, amphipods, and ampeliscids. There were no ampeliscids observed in the highly toxic samples.

New York/New Jersey Harbor, USA

Data were collected by U.S. EPA Region 2 for many regions of the New York/New Jersey Harbor during a survey in 1993 [23]. Bioassay data for 169 samples (using A. abdita) showed a wide range in response. Control-adjusted survival was less than 80% in 24 samples collected from locations scattered throughout the survey area. However, none of the indices of benthic community abundance or diversity indicated strong concordance with amphipod survival (Table 7). Although total numbers of species were lowest in samples in which amphipod survival was less than 30%, none of the measures of abundance diminished as amphipod survival decreased. The average abundance of most groups was highest when survival was intermediate (between 30 and 80%). The relatively high stanphipod dard deviations in the benthic results illustrate the variability among these data.

Table Table 3.. Average (± standard deviation) total abundance, total numbers of species, and abundance of all amphipods and phoxocephalid amphipods in samplesa from Palos Verdes shelf, California, USA, that were either toxic or nontoxic in amphipod tests with Rhepoxynius abronius [10]
   Abundance
Amphipod survivalTotal abundance (no./0.1 m2)Total no. of speciesAmphipodsPhoxocephalids
  1. aTotal number of samples = 9.

  2. b Significantly different from controls, p < 0.05.

Nontoxic (average 91.2%, n = 6)1,352.8 ± 1,241.270.9 ± 16.141.2 ± 23.829.1 ± 13.8
Toxicb (average 79.0%, n = 3)446.7 ± 153.820.9 ± 5.30.8 ± 1.10.0
Table Table 4.. Abundance (average ±M standard deviation) of Leptocheirus plumulosus within three ranges in amphipod (L. plumulosus) survival in samplesa from Baltimore Harbor, Maryland, USA, and vicinity [20]
 Average amphipod abundance
Amphipod survivalNo./m2% Reference
  1. a Total number of samples = 20.

>91% (average 99.8%, n = 12)8,244 ± 10,47232.2
75–85% (average 80.6%, n = 3)2,937 ± 3,38911.5
<60% (average 43.4%, n = 5)18 ± 130.1

The apparent lack of concordance between amphipod survival and amphipod abundance in these samples was further illustrated by examining samples with abundances less than 1% of reference. In 71% (17 of 24) of the toxic samples (survival <80%), amphipod abundance was <1% of reference conditions. Similarly, amphipod abundance was <1% of reference in 75% (108 of 144) of the nontoxic samples. Amphi-pods were absent in 25% of the toxic samples and in 19% of the nontoxic samples.

Environmental Monitoring and Assessment Program- Lousianian Province

As part of the Environmental Monitoring and Assessment Program (EMAP) along the Gulf of Mexico, the U.S. EPA [24] analyzed 634 samples for amphipod survival (using A. abdita) and benthic community composition (V. Engle, personal communication). Amphipod survival was very high (>80%) in most of the samples. Averages for total numbers of species and the abundance of all biota, arthropods, and amphipods were compared among eleven ranges in amphipod survival (Table 8). The average numbers of taxa in each sample were highly variable in each category of amphipod survival and did not indicate a clear pattern in concordance with survival. Similarly, the average abundance of all biota did not show concordance with survival. However, the abundance of amphipods and, to a lesser extent, all arthropods generally decreased as amphipod survival dropped below 80%. The relationship between amphipod survival in the laboratory bio-assays and amphipod abundance in the benthos is illustrated in a histogram (Fig. 2). Average abundance of amphipods was 20 to 54 per sample in each category when amphipod survival was 80% or greater. There were no benthic amphipods found in 60% of the samples in which survival was <80%, whereas 48% of the samples were devoid of amphipods when survival equaled or exceeded 80%. Average amphipod abundance in the benthic samples decreased to 9 to 16 per sample when survival dropped to 50 to 79.9% and dropped again to ≤7 per sample when survival was less than 50%. However, the numbers of samples decreased markedly as percentage survival decreased.

The EMAP data were difficult to analyze because they were collected from many different estuarine and coastal areas. These areas included bayous, open bays, marshes, and the mouth of the Mississippi River. Therefore, sedimentological and benthic conditions would be expected to differ naturally among these areas, thus making it difficult to define what would constitute a reference condition of any particular estuary. By selecting the average of the samples with the highest 10% of the abundance counts as “reference,” we may have inflated the expectations of an abundant benthic infauna.

New Bedford Harbor, Massachusetts, USA

Data were available (K.J. Scott, personal communication) for 76 samples collected within New Bedford Harbor, where there was a wide range in response in the amphipod survival tests (A. abdita). Benthic metrics that were evaluated were total numbers of species, amphipod abundance, and a multiparameter index of benthic health. The index of benthic health was calculated as an amalgam of numerous individual measures of diversity, species abundance, and either the presence or the absence of selected indicator species as in the EMAP surveys [24]. Values from calculations of this index of <2.0 are considered to indicate a degraded benthic community. Species richness did not uniformly decrease as survival decreased (Table 9). However, the benthic index values and amphipod abundance decreased sharply with decreasing survival. Average benthic index values were positive in samples in which amphipod survival exceeded 100% of the controls. The benthic index values became −2.0 or greater, indicating “degraded” conditions, when amphipod survival decreased to <90%. The average abundance of amphipods was 16 to 20 per sample when survival was 90% or greater and dropped to <6 per sample when survival decreased to <90% (Fig. 3). There was an average of two amphipods per sample with survival at 80 to 90%. Amphipods were absent in 71% of benthic samples when survival was <80% (toxic samples), whereas 19% of samples were devoid of amphipods when survival equaled or exceeded 80% (nontoxic samples).

Table Table 5.. Average (± standard deviation) numbers of species and abundance of all biota, arthropods, and Ampelisca abdita within three ranges in amphipod (A. abdita) survival in samplesa from the Hudson-Raritan estuary New York/New Jersey, USA [21]
   Total Abundance(%oftotal)
Amphipod survivalabundance (no./m2)Total no. of speciesArthropodsA. abdita
  1. aTotal number of samples = 73.

> 80% (average 98.4%, n = 42)372 ± 40714 ± 626 ± 2815 ± 24
50–79.9% (average 69.0%, n = 14)500 ± 38619 ± 615 ± 1810 ± 17
< 50% (average 20.0%, n = 17)462 ± 58319 ± 911 ± 190 ± 0
Table Table 6.. Average (± standard deviation) species richness and abundance of all species, arthropods, amphipods, and ampeliscids within two ranges in amphipod (Ampelisca abdita) survival in samplesa from central Biscayne Bay, Florida, USA [22]
   Abundance
Amphipod survivalTotal abundance (no./m2)Total no. of speciesArthropodsAmphipodsAmpeliscids
  1. aTotal number of samples = 22.

>77% (average 96%, n = 15)1,569 ± 1,220206 ± 95508 ± 846183 ± 3177 ± 16
<50% (average 25%, n = 7)1,412 ± 1,13218 ± 531 ± 3819 ± 320 ± 0

Delaware Bay, Delaware, USA

Unpublished data were available for 80 samples collected during NOAA's survey of Delaware Bay. Amphipod survival was determined in tests using A. abdita. Although there were only four samples in which amphipod survival was <80% (Table 10), these four samples were highly toxic (average of 30% survival). Measures of total numbers of species, species diversity (H'), total abundance per sample, and amphipod abundance in the toxic samples generally were about one-half those in the nontoxic samples. For example, average total abundance in the three nontoxic categories ranged from 205 to 371 animals per sample. In contrast, there was an average of 96 animals in the toxic samples. The abundance of A. abdita was variable among the nontoxic samples with an average of only one per sample when survival ranged from 90 to 100%. These animals were absent in the four toxic samples.

San Francisco Bay, California

Toxicity and benthic data for 97 samples (compiled by the San Francisco Estuary Institute, Richmond, CA, USA, and B. Thompson, personal communication) were reviewed (Table 11). In tests performed with Eohaustorius estuarius, 35 samples were classified as toxic. Neither the total numbers of species nor the total abundance decreased as amphipod survival decreased. The average abundance of echinoderms, arthropods, amphipods, and A. abdita initially tended to increase as survival decreased, reaching peaks when survival ranged from 70 to 80%. Then the abundance of these animals decreased when amphipod survival dropped below 70%. No echinoderms were found in samples in which amphipod survival was less than 50%. Average amphipod abundance in 33 of the 35 toxic samples was less than 50% of that in the reference samples (average of top 10% of 97 samples) and less than 10% of the average reference conditions in 19 of 35 samples. Seven of the 35 toxic samples (20%) were devoid of amphipods. However, amphipods also were absent in 16% of the 62 nontoxic samples.

Southern California bays

A considerable amount of matching chemical, toxicological, and benthic data were generated as a part of the state of California's Bay Protection and Toxic Cleanup Program (Sacramento, CA, USA). Most of the toxicity tests were performed with Rhepoxynius abronius (Table 12). These tests were performed on 102 samples from San Pedro Bay [25], 64 samples from San Diego Bay [26], and 43 samples from Newport Harbor (B.M. Phillips et al., unpublished data). We excluded the counts of the amphipod G. japonica from these data because this species increased in abundance in Richmond Harbor with increasing chemical contamination and toxicity [19]. As observed in the San Francisco Bay data, it appeared that the abundance of many groups initially increased as survival decreased. Total abundance peaked when survival ranged from 40 to 70%, then dropped off as survival decreased below 40%. The abundance of all amphipods and all crustaceans followed the same pattern. Total numbers of species, polychaete abundance, and amphipod abundance were relatively low in samples in which amphipod survival was less than 30%. One amphipod was observed in the four samples with <20% survival. Mollusk abundance did not show any clear patterns relative to amphipod survival. Many of the samples in which amphipod survival was greater than 100% of the control responses were collected in coarse sands, thus possibly accounting for low infaunal abundance.

Table Table 7.. Average (± standard deviation) numbers of species and abundance of major taxonomic groups within seven ranges in amphipod (Ampelisca abdita) survival in samplesa from New York/New Jersey Harbor, USA [23]
  Abundance per sample
Amphipod survivalTotal no. of speciesTotalAnnelidsMollusksArthropodsAmphipodsAmpeliscids
  1. aTotal number of samples = 169.

≥ 100% (average 102%, n = 24)27 ± 11871 ± 788514 ± 541153 ± 259153 ± 336148 ± 33593 ± 300
90.0–99.9% (average 95.5%, n = 100)25 ± 121,310 ± 1,901655 ± 868179 ± 277416 ± 1,283408 ± 1,281333 ± 1,079
80.0–89.9% (average 85.7%, n = 21)23 ± 15843 ± 877518 ± 77387 ± 208107 ± 20490 ± 19636 ± 155
70.0–79.9% (average 73.8%, n = 6)24 ± 112,889 ± 3,464550 ± 465182 ± 2982,014 ± 3,2982,004 ± 3,2961,443 ± 2,356
50–69.9% (average 60.3%, n = 6)27 ± 164,846 ± 5,927947 ± 1,069126 ± 1763,673 ± 5,2273,662 ± 5,2211,977 ± 2,821
30.0–49.9% (average 41.6%, n = 5)30 ± 162,807 ± 4,267465 ± 383130 ± 1202,084 ± 4,0592,056 ± 4,0641,079 ± 2,148
<30% (average 11.6%, n = 7)15 ± 101,059 ± 1,811389 ± 410323 ± 699314 ± 750303 ± 736299 ± 731
Table Table 8.. Average (± standard deviation) numbers of species and abundance of major taxonomic groups per sample within 11 ranges in amphipod (Ampelisca abdita) survival in samplesa from the Louisianian Province of the Gulf of Mexico (V. Engle, personal communication)
  Abundance
Amphipod survivalTotal no. of speciesTotalArthropodsAmphipods
  1. a Total number of samples = 634.

≥100% (average 102%, n = 150)22 ± 22291 ± 44726 ± 10020 ± 93
90–99.9% (average 95.3%, n = 337)25 ± 25354 ± 47824 ± 8517 ± 78
80–89.9% (average 86.5%, n = 80)27 ± 26487 ± 95766 ± 27854 ± 266
70–79.9% (average 75.6%, n = 31)22 ± 31278 ± 35014 ± 369 ± 34
60–69.9% (average 65.5%, n = 12)15 ± 25403 ± 46532 ± 6816 ± 44
50–59.9% (average 55.7%, n = 6)37 ± 53440 ± 37319 ± 3311 ± 18
40–49.9% (average 44.4%, n = 4)7 ± 729 ± 253 ± 20.1 ± 0.1
30–39.9% (average 36.0%, n = 1)23300
20–29.9% (average 26.4%, n = 3)28 ± 181,016 ± 1,3027 ± 73 ± 7
10–19.9% (average 15.4%, n = 6)14 ± 1045 ± 36861 ± 927 ± 61
0–9.9% (average 2.6%, n = 4)13 ± 8165 ± 1324 ± 20.4 ± 4

Tests of amphipod survival were conducted with Eohaus-torius estuarius on 18 samples from San Pedro Bay, California [25]. Amphipod survival ranged from >100% of controls to 10% of controls, providing a maximum of only three samples within each range in survival (Table 13). Because of the small sample sizes, clear patterns in the data were difficult to identify. However, the abundances of amphipods, all crustaceans, and mollusks were relatively low in samples with survival <60%.

Data generated for 25 samples tested with A. abdita from a survey of Newport Bay also showed a general lack of concordance between survival and the abundance and diversity of the infauna (Table 14). The abundance of most taxonomic groups was greatest in two samples in which survival was the lowest (<10%).

Data summaries

Following the approach used in analyses of data from Elliott Bay, Washington, [27], we assembled the data from each of the studies reviewed to determine the frequencies and degrees of benthic changes when samples were classified as toxic in the amphipod survival tests. In the following discussion, we pooled the data described previously to identify any patterns of concordance between laboratory results and infaunal structure. The data were summarized despite known differences in sampling methods. Three different analyses were conducted with summarized data.

thumbnail image

Figure Fig. 2.. Average abundance of amphipods within 11 ranges in amphipod survival determined for samples from the Louisianian Province (V. Engle, personal communication).

Download figure to PowerPoint

thumbnail image

Figure Fig. 3.. Average abundance of amphipods within nine ranges in amphipod survival determined for samples from New Bedford Harbor, Massachusetts (K.J. Scott, personal communication).

Download figure to PowerPoint

First, we determined the frequency of 50 and 90% changes (losses) in benthic diversity or abundance when samples were toxic. In this analysis, the numbers of samples classified as toxic (control-adjusted survival <80%) were determined for each study and for all studies combined. Then we determined the proportion of those samples in which at least one measure of benthic abundance or diversity was less than 50% of the average reference condition or less than 10% of reference for each study (Table 15). For each study area, the benthic end point(s) that showed the greatest number of stations with reduced abundance is (are) shown. The EMAP data were excluded from this table because samples were collected within and among many different estuaries, making it difficult to determine what would constitute reference conditions for each area within the large study provinces.

With the data from the regional surveys combined together (n = 829, excluding the EMAP data for the Louisianian Province), there were 231 samples in which control-adjusted survival was less than 80% (Table 15). In 92% of these samples, at least one measure of diversity or abundance was <50% of the respective reference value. In 67% of these samples, at least one measure of benthic infauna abundance or diversity was <10% of reference conditions when the samples were classified as toxic. Frequently, the abundance of arthropods, crustaceans, or amphipods appeared to be the most sensitive measure of infaunal structure. In all toxic samples from Puget Sound (WA), Palos Verdes (CA), Chesapeake Bay (MD), Biscayne Bay (FL), and Delaware Bay (DE), at least one of the benthic measures was <50% of the respective reference values. In addition, all the toxic samples from Puget Sound, Palos Verdes, and Biscayne Bay had benthic measurements less than 10% of reference.

Because the laboratory tests of survival are performed with amphipods that often constitute an important portion of the local benthic infauna [28], it is reasonable to assume that these animals would be less abundant or absent in locations where sediments were determined to be toxic [6]. In many of the data sets analyzed in this review, indeed, the average abundance of amphipods was substantially lower in toxic samples than in nontoxic samples. Therefore, in the second analysis of summarized data, we calculated the percentages of samples in which amphipods (either all species or the test species, depending on what was reported) were absent in toxic samples versus nontoxic samples. In most cases, there was a higher proportion of samples devoid of amphipods when toxicity was observed than when it was not observed (Table 16). Such differences were most apparent in the relatively small data sets from Puget Sound, Palos Verdes, San Francisco Bay (U.S. EPA data), and Biscayne Bay. In the seven samples from Puget Sound that were toxic to R. abronius, there were no amphipods found in the infauna. Similarly, there were no A. abdita in the four samples from Delaware Bay that were toxic. Based on the data compiled from these studies, amphipods were not found in 39% of the toxic samples and in 28% of the nontoxic samples.

The relatively good agreement between low amphipod survival and lack of amphipods in the small data sets was masked by the much larger EMAP and NOAA data sets, in which there was relatively poor agreement. Therefore, although a greater proportion of samples that were toxic were devoid of these animals, amphipods also were absent from many of the non-toxic samples because of factors other than toxicity.

In the third analysis of summarized data, we calculated the average abundances of amphipods within 11 ranges of amphipod survival to determine whether abundance tended to diminish as survival decreased. In this analysis, data were compiled from 1,145 samples (Table 17). In addition, the data from the 209 samples from California tested with R. abronius were kept separate but treated the same way. The numbers of samples in each category are shown in the first column for the combined data set and the California data.

Table Table 9.. Average (± standard deviation) numbers of species, benthic index values, and amphipod abundance per sample within nine ranges in amphipod (Ampelisca abdita) survival in samplesa from New Bedford Harbor, Massachusetts, USA (K.J. Scott, personal communication)
Amphipod survivalEMAPb benthic health indexTotal no. of speciesAmphipod abundance
  1. a Total number of samples = 76.

  2. b EMAP = U.S. Environmental Monitoring and Assessment Program.

>100% (average 103.6%, n = 13)0.9 ± 1.453 ± 2416 ± 22
90–99.9% (average 95.1%, n = 20)−0.2 ± 3.656.1 ± 27.220 ± 26
80–89.9% (average 87.8%, n = 9)−2.1 ± 2.430 ± 202 ± 3
70–79.9% (average 74.6%, n = 7)−7.3 ± 8.325 ± 110.4 ± 0.7
60–69.9% (average 64.5%, n = 5)−3.5 ± 1.418 ± 50
50–59.9% (average 53.9%, n = 3)−4.6 ± 2.120 ± 51 ± 0.8
40–49.9% (average 44.6%, n = 6)−3.8 ± 4.025 ± 80
30–39.9% (average 33.6%, n = 4)−3.3 ± 5.149 ± 346 ± 7
<30.0% (average 6.1%, n = 9)−2.0 ± 1.931 ± 205 ± 10

In the combined data set, there was an average of 332 amphipods in samples when survival was greater than the controls (Table 17). The large standard deviation (2,583) indicates the natural variability in the abundance of amphipods in the infauna. The averages for each category ranged from 103 to 888 when amphipod survival was greater than or equal to 50% and from 4 to 33 when survival was below 50%. Average amphipod abundance initially appeared to increase as amphipod survival decreased, reaching a maximum when survival was 50 to 70%, then decreased as survival continued to decrease. Again, the relatively large standard deviations indicated that there was wide variability in the data within each category of survival. However, average amphipod abundance decreased remarkably in samples in which survival dropped below 50% and differed from that in samples in which survival was <50% by one to two orders of magnitude. In contrast, the percentages of samples in which amphipods were absent increased inconsistently and by only small amounts as survival decreased. These data, therefore, suggest that the absence of amphipods in the benthos corresponded poorly with amphipod survival, whereas their numerical abundance tracked with survival much better.

The data from the California bays appeared to indicate that amphipods were considerably less abundant there than in other areas (Table 17). However, these samples were much smaller in size (three 0.075-m2 samples) than in most other studies (e.g., three 0.04-m2 samples in EMAP), which probably contributed to some portion of these differences. Nevertheless, the relationship between amphipod abundance and amphipod survival in California seemed to parallel that observed with the other data. Average abundance increased initially, reached a maximum when survival was 50 to 70%, then decreased to a minimum when survival was less than 20%. In both data sets, the proportion of samples in which amphipods were absent did not show any consistent trends.

Table Table 10.. Average (± standard deviation) species richness, diversity, and total abundance within four ranges in amphipod (Ampelisca abdita) survival in samplesa from Delaware Bay, Delaware, USA (unpublished data)
   Species Abundance (no./sample)
Amphipod survivalTotal no. of species(H′)TotalAmphipodsA. abdita
  1. a Total number of samples = 80.

>100% (average 103.1%, n = 36)12 ± 92.8 ± 0.6255 ± 45964 ± 9385 ± 331
90.0–99.9% (average 96.1%, n = 28)15 ± 71.8 ± 0.5205 ± 29878 ± 1011 ± 2
80.0–89.9% (average 86.8%, n = 12)15 ± 71.5 ± 0.7371 ± 432123 ± 13172 ± 178
< 80% (average 29.7%, n = 4)9 ± 41.4 ± 0.796 ± 6438 ± 390 ± 0
Table Table 11.. Average (± standard deviation) numbers of species and abundance of major taxonomic groups within seven ranges in amphipod (Eohaustorius estuarius) survival in samplesa from San Francisco Bay monitoring programs (San Francisco Estuary Institute, Richmond, CA, USA, unpublished data)
   Total Abundance (no./0.05 m2 sample)
Amphipod survivalTotal no. of speciesabundance (no./0.05 m2)EchinodermsArthropodsAmphipodsbAmpelisca abdita
  1. aTotal number of samples = 97.

  2. bExcluding Grandidierella japonica.

> 100% (average 107.1%, n = 16)18 ± 16583 ± 1,1391.1 ± 2.1296 ± 845277 ± 821216 ± 776
90.0–99.9% (average 94.6%, n = 24)19 ± 17669 ± 1,0800.1 ± 0.4213 ± 422125 ± 37463 ± 274
80.0–89.9% (average 84.9%, n = 22)27 ± 16623 ± 7990.7 ± 2.5392 ± 717314 ± 705127 ± 374
70.0–79.9% (average 76.3%, n = 13)25 ± 141,537 ± 1,8250.2 ± 0.51,064 ± 1,460599 ± 2,217545 ± 1,145
60.0–69.9% (average 66.5%, n = 5)22 ± 101,243 ± 1,2590.2 ± 0.4514 ± 482358 ± 49424 ± 45
50.0–59.9% (average 55.3%, n = 9)18 ± 12394 ± 2310.4 ± 1.3150 ± 157103 ± 15384 ± 152
<50% (average 18.2%, n = 8)12 ± 121,447 ± 2,1570 ± 0100 ± 11548 ± 8636 ± 88

DISCUSSION AND CONCLUSIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION AND CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

A summary of the results from EMAP studies of the Atlantic and Gulf of Mexico coasts was compiled to quantify the incidence of degraded benthic conditions with the EMAP, multiparameter index [29]. The combined data sets indicated that 72% of samples had benthic conditions classified as degraded when survival (using A. abdita) was less than 80%. In addition, 84% of samples had degraded benthos when survival was less than 60%.

In a study performed at a polycyclic aromatic hydrocarbon- contaminated site in Elliott Bay (WA), the U.S. EPA [27] demonstrated the relationship between losses of either total abundance, total numbers of species, or dominance in the benthos and the survival of amphipods in the laboratory tests (using R. abronius). As average survival decreased, the percentage loss of the indices increased relative to values determined for nontoxic reference areas in Puget Sound (WA). Losses of benthic resources ranged from 50 to 53% when survival dropped to 0%.

Data examined in this review indicated a general but highly variable pattern of concordance between acute toxicity tests and various indices of abundance and diversity of benthic assemblages. Generally, diversity as measured by the total numbers of species and the abundance of crustaceans, notably amphipods, decreased as amphipod survival in the laboratory dropped below 80%. The changes in abundance of amphipods were apparent in the data from Puget Sound, San Francisco Bay, Palos Verdes shelf, Baltimore Harbor, Hudson-Raritan estuary, Biscayne Bay, New Bedford Harbor, and the Louisianian Province of the Gulf of Mexico.

In many case studies and with data from many studies combined, average abundance of resident amphipods initially showed considerable variability among samples in which survival was relatively high. Such variability was probably a function of the many natural environmental factors that would be expected to cause differences in abundance among both toxic and nontoxic samples. Samples with relatively low amphipod abundance probably were collected in sandy habitats, which would not be expected to accumulate toxic levels of contaminants. Sandy habitats would provide less food and soft substrates for burrowing and tube-building macroinvertebrates than fine-grained sediments. Average amphipod abundance peaked when amphipod survival dropped to about 50 to 70%, probably reflecting the presence of more fine-grained particles that would tend to accumulate higher toxicant concentrations than sandy materials. Such sediments also would tend to accumulate sufficient food material to support amphipods able to adapt to the presence of toxicants. Amphipod abundance then decreased sharply as amphipod survival continued to decline, probably as a result of the toxicity of increasing chemical concentrations and/or declining oxygen concentrations that often covaries with high chemical contamination in industrialized harbors. A similar pattern of increasing benthic biomass, abundance, and diversity, followed by sharp decreases over spatial and temporal scales in inputs of organic matter into sediments, was reported in studies of Scandinavian fjords [30].

Table Table 12.. Average (± standard deviation) numbers of species and organism abundance per station of major taxonomic groups within 11 ranges in amphipod (Rhepoxynius abronius) survival in samplesa from all Southern California bays USA (R. Fairey, personal communication)
   Abundance (no./sample)
Amphipod survival (%)No. of samplesTotal no. of speciesTotalPolychaetesCrustaceansMollusksAmphipodsb
  1. aTotal number of samples = 209.

  2. b Excluding Grandidierella japonica.

≥100515.5 ± 844.6 ± 2837.0 ± 254.7 ± 42.5 ± 22.7 ± 2
90–99.96016.0 ± 679.7 ± 5953.3 ± 4114.3 ± 195.6 ± 118.1 ± 12
80–89.97614.2 ± 675.8 ± 2853.0 ± 3912.7 ± 186.3 ± 176.1 ± 9
70–79.93311.9 ± 659.3 ± 2839.8 ± 259.7 ± 144.6 ± 134.1 ± 7
60–69.91313.7 ± 8117.0 ± 12962.0 ± 4118.3 ± 456.1 ± 511.8 ± 30
50–59.9914.4 ± 8129.4 ± 15871.8 ± 6038.0 ± 7911.0 ± 1916.7 ± 31
40–49.9317.9 ± 6154.9 ± 9467.6 ± 2116.9 ± 814.2 ± 1011.8 ± 7
30–39.9417.4 ± 787.5 ± 7654.2 ± 429.3 ± 86.8 ± 25.9 ± 8
20–29.928.0 ± 096.4 ± 1518.0 ± 119.9 ± 89.4 ± 99.0 ± 8
10–19.9110.061.018.07.035.00.0
0–9.937.3 ± 629.3 ± 3614.3 ± 155.3 ± 89.0 ± 130.3 ± 0.5
Table Table 13.. Average (± standard deviation) numbers of species and organism abundance per station of major taxonomic groups within 11 ranges in amphipod (Eohaustorius estuarius) survival in samplesa from San Pedro Bay, California (R. Fairey, personal communication)
   Abundance (no./sample)
Amphipod survival (%)No. of samplesTotal no. of speciesTotalPolychaetesCrustaceansMollusksAmphipodsb
  1. a Total number of samples =18.

  2. b Excluding Grandidierella japonica.

  3. c ND = no data.

≥100117.348.737.79.02.06.3
90–99.9327.3 ± 5119.1 ± 4992.5 ± 5319.9 ± 174.8 ± 214.2 ± 12
80–89.9327.0 ± 5209.6 ± 136186.8 ± 11912.1 ± 112.7 ± 24.9 ± 4
70–79.9326.2 ± 5174.9 ± 8494.4 ± 1837.1 ± 4917.9 ± 212.4 ± 2
60–69.9319.3 ± 1529.3 ± 263277.3 ± 57206.0 ± 2302.3 ± 180.7 ± 88
50–59.9315.7 ± 3263.0 ± 148215.3 ± 10510.0 ± 82.0 ± 20.3 ± 0.5
40–49.9112.070.054.011.00.00.0
30–39.90NDcNDNDNDNDND
20–29.90NDNDNDNDNDND
10–19.90NDNDNDNDNDND
0–9.9119.01,196.0372.0811.013.0100.0

The relationships between toxicity and changes to the benthos were poorest in data sets gathered from some large, multiestuary systems (New York/New Jersey Harbor, southern California bays). Often, they were clearer with smaller data sets from areas known to be contaminated (e.g., Baltimore Harbor in northern Chesapeake Bay, MD, and Richmond Harbor in San Francisco Bay, northern California) and over which clearer pollution gradients were observed.

The indices or measures of benthic structure that appeared most correlated with toxicity differed among some locations. The abundances of polychaetes and mollusks often did not decrease with decreasing survival of the amphipods. Measures of total abundance often were not concordant with toxicity gradients, probably as a result of species substitutions among locations. In contrast, the abundance of arthropods, crustaceans, and amphipods often were more concordant with amphipod survival. Thus, it appears that acute toxicity in laboratory bioassays often, but not always, was associated with losses of benthic resources of one kind or another, usually losses in abundance and diversity of crustacean groups.

The high variability in the results suggests that factors other than toxicity probably contributed significantly to differences among sites in the abundance and diversity of the benthos. The abundance, diversity, and taxonomic structure of estuarine benthic communities can be controlled or influenced by numerous environmental factors, such as water depth, sediment texture, salinity, scouring, current speed, dissolved oxygen concentrations, succession, and predation. These factors may cause changes in benthic communities with or without the simultaneous effects of toxic chemicals. Therefore, toxicological factors that induce responses in acute laboratory bioassays may not be the same as those that cause changes in the resident benthos. Some portion of the variability in these results also may be attributable to the toxicity tests. Toxicity in the amphipod tests is not always observed in highly contaminated samples, nor is the lack of toxicity always ensured in uncon-taminated sediments [11].

Table Table 14.. Average (± standard deviation) numbers of species and organism abundance per station of major taxonomic groups within 11 ranges in amphipod (Ampelisca abdita) survival in samplesa from Newport Bay, California (R. Fairey, personal communication)
   Abundance (no./sample)
Amphipod survival (%)No. of samplesTotal no. of speciesTotalPolychaetesCrustaceansMollusksAmphipodsb
  1. a Total number of samples = 25.

  2. b Excluding Grandidierella japonica.

  3. c ND = no data.

>100515.6 ± 7163.5 ± 7790.3 ± 4737.4 ± 3825.5 ± 424.8 ± 4
90–99.91017.1 ± 8106.5 ± 4957.7 ± 3520.1 ± 136.4 ± 67.6 ± 8
80–89.9615.0 ± 8110.0 ± 5858.0 ± 3426.1 ± 315.9 ± 95.8 ± 9
70–79.918.755.345.07.71.03.0
60–69.90NDcNDNDNDNDND
50–59.90NDNDNDNDNDND
40–49.90NDNDNDNDNDND
30–39.90NDNDNDNDNDND
20–29.9131.0176.7106.018.324.310.4
10–19.90NDNDNDNDNDND
0–9.9219.9 ± 1537.0 ± 15153.0 ± 48216.5 ± 4335.3 ± 24109.4 ± 7
Table Table 15.. Numbers of toxic samples in which measures of benthic diversity or abundance were less than 50% or less than 10% of average reference conditions in each study location (USA)
LocationNo. of toxic samplesSamples with any measure of benthic diversity or abundance <50% of reference%Samples with any measure of benthic diversity or abundance <10% of reference%Benthic endpoint(s)
  1. a Unpublished data from National Oceanic and Atmospheric Administration.

  2. b Data from U.S. Environmental Protection Agency.

  3. C B. Thompson, personal communication.

  4. d Data from tests with Rhepoxynius abronius only.

Puget Sound, WAa77100.07100.0Abundance of arthropods, amphipods, echinoderms
Palos Verdes shelf, CA33100.03100.0Species richness, total abundance, amphipod abundance
San Francisco Bay, CAb4375.0375.0Abundance of amphipods
San Francisco Bay, CAC353394.31954.3Abundance of amphipods
San Francisco Bay, CAC9888.9444.4Total abundance
Biscayne Bay, FL77100.07100.0Species richness, total abundance, arthropod abundance
Hudson-Raritan estuary, NY/NJ312787.12064.5Abundance of arthropods
Delaware Bay, DE44100.0375.0Abundance of amphipods
New York/New Jersey REMAP,NY/NJ 242083.31875.0Abundance of amphipods
New Bedford Harbor, MA332987.92472.7Abundance of amphipods
Chesapeake Bay, MD66100.0350.0Abundance of amphipods
Newport Bay, CAd232295.71356.5Abundance of mollusks
San Pedro Bay, CAd292896.62275.9Abundance of amphipods
San Diego Bay, CAd161593.8850.0Abundance of crustaceans
Totals231212 154  
Percentage of total  91.8 66.7 
Table Table 16.. Numbers of infauna samples in which amphipods were absent when laboratory tests indicated they were toxic versus nontoxic for each sampling location and for all locations (USA)
LocationNo. of toxic samplesToxic samples without benthic amphipods%No. of nontoxic samplesNontoxic samples without benthic amphipods%Measure of amphipod abundance
  1. a Data from tests with Rhepoxynius abronius only.

Puget Sound, WA77100.07457.1All amphipods
Palos Verdes shelf, CA3266.7400.0All amphipods
San Francisco Bay, CAb4250.0400.0All amphipods
San Francisco Bay, CAc35720.0621016.1All amphipods
Chesapeake Bay, MD6116.71516.7Leptocheirus plumuosus
Biscayne Bay, FL7457.11516.7Ampelisca abdita
Hudson-Raritan estuary, NY/NJ312683.9422661.9A. abdita
Delaware Bay, DE44100.0765369.7A. abdita
New York/New Jersey REMAP, NY/NJ24625.01442819.4All amphipods
New Bedford Harbor, MA33927.34237.1All amphipods
Southern Californiac681927.91412920.6All amphipods
Totals22287 552155  
Percentage of total  39.2  28.1 

In many of the data sets we examined (e.g., Delaware Bay, DE, Biscayne Bay, FL), samples were collected in areas influenced by discharges from rivers and canals, resulting in definable salinity gradients. In some samples, we suspect that organic enrichment may have driven the concentrations of dissolved oxygen down to critical levels and hydrogen sulfide and/or ammonia up to critical levels. In analyses of data from estuaries of the southeastern United States, significant statistical relationships were reported between indices of benthic composition and the concentrations of chemical toxicants calibrated to numerical sediment quality guidelines [31]. However, the benthic indices in the polluted sites also were correlated with sediment texture, organic carbon content, and sampling depths. Thus, it was likely that the concentrations of toxicants covaried with the natural variables. On the other hand, such correlations were not observed in clean areas. Therefore, it appeared that in polluted areas the composition of the benthos was driven largely by the presence of the chemical toxicants [31].

Table Table 17.. Average abundance of amphipods (± standard deviation) and percentages of samples without amphipods within 11 ranges in amphipod survival. Dataa were compiled from many studies (n = 1,145) and southern California (n = 209)
 Abundance of amphipodsPercentages without a of samples mphipods
Amphipod survivalAll otherSouthern CaliforniaAll otherSouthern California
  1. a Data from San Francisco Bay, Chesapeake Bay, Biscayne Bay, Hudson-Raritan estuary, Delaware Bay, New York/New Jersey REMAP, New Bedford Harbor, and Louisianian Province of EMAP.

≥100% (n = 262 and 5)332 ± 2,5833 ± 342.740.0
90.0–99.9% (n = 526 and 60)148 ± 1,0208 ± 1237.521.7
80.0–89.9% (n = 154 and 76)103 ± 3606 ± 938.618.2
70.0–79.9% (n = 68 and 33)411 ± 1,5384 ± 757.524.2
60.0–69.9% (n = 27 and 13)888 ± 2,88412 ± 3244.430.8
50.0–59.9% (n = 31 and 9)766 ± 2,53117 ± 3335.533.3
40.0–49.9% (n = 18 and 3)8 ± 1612 ± 966.70.0
30.0–39.9% (n = 10 and 4)9 ± 216 ± 960.025.0
20.0–29.9% (n = 9 and 2)33 ± 849 ± 1244.40.0
10.0–19.9% (n = 16 and 1)5 ± 8056.3100.0
<10% (n = 24 and 3)4 ± 100.3 ± 0.670.867.0

In some of the study areas where changes in benthic communities co-occurred with low amphipod survival, there were only very small differences among stations in water depths and proximity to river mouths. Data from the Palos Verdes shelf were collected offshore along a consistent isobath on the Continental Shelf where a huge range in the concentrations of DDT was measured. Data from Richmond Harbor, several other areas in San Francisco Bay, Baltimore Harbor in Chesapeake Bay, and several locations in Puget Sound were not collected in areas influenced by riverine conditions. Huge differences (orders of magnitude) in chemical concentrations were reported among stations in these areas. In the Hudson-Raritan estuary, Baltimore Harbor, San Francisco Bay, and central Bis-cayne Bay, decreased amphipod survival was highly correlated with the concentrations of mixtures of toxicants in the sediments. In these areas, therefore, it is highly probable that contaminant-induced toxicity contributed significantly to losses of benthic resources.

In view of the data examined and the precautionary caveats, the collective weight of evidence from our analyses suggests that toxicity in laboratory bioassays of amphipod survival often but not always was accompanied by ecologically significant alterations to resident infaunal populations. In the large majority of the samples determined to be toxic, the diversity and/or abundance of the benthos were markedly reduced relative to samples that had the highest diversity and/or abundance. Notably, the average abundance of resident amphipods in the benthic infauna dropped remarkably when amphipod survival decreased below 50% of controls.

The precise numerical relationship between percentage survival in the laboratory and amphipod abundance in the field cannot be described or modeled with the data that are currently available. All the data we reviewed were collected during field surveys intended for other purposes. Although the data from these surveys suggest that amphipod abundance and other measures of diversity and abundance tended to diminish with decreasing survival in some areas, these relationships were not apparent in all areas and all studies. Cause-effect experiments, in which natural environmental variables are controlled as well as possible, are needed to quantify the actual effects of contaminant-induced toxicity on the abundance of these animals in the infauna and, therefore, to satisfy all the criteria [12] for establishing causality. In such experiments, the weight of data from chemical, toxicity, and benthic results should be used in multivariate statistical analyses to define a population of reference stations. Then metrics of benthic abundance and diversity in toxic samples should be compared against those measures in the population of reference samples in statistical analyses. In lieu of such experiments, we conclude that it is still necessary to conduct benthic community analyses to determine empirically that benthic resources are adversely affected at sites determined to be toxic in laboratory tests.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION AND CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

Data for this review were kindly provided by Virginia Engle, John Scott, Bruce Thompson, Beth McGee, and Russell Fairey. Initial versions of the manuscript were reviewed by Jawed Hameedi and Jeffrey Hyland and an anonymous reviewer. All provided helpful and constructive comments.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION AND CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES
  • 1
    Swartz RC., 1989. Marine sediment toxicity tests. In Contaminated Marine Sediments-Assessment and Remediation. National Research Council, National Academy Press, Washington, DC, USA, pp 115129.
  • 2
    Lamberson JO, DeWitt TH, Swartz RC., 1992. Assessment of sediment toxicity to marine benthos. In BurtonGAJr, ed, Sediment Toxicity Assessment. Lewis, Boca Raton, FL, USA, pp 183240.
  • 3
    U.S. Environmental Protection Agency. 1991. Evaluation of dredged material proposed for ocean disposal. EPA-503/8–91–001. Washington, DC.
  • 4
    U.S. Environmental Protection Agency. 1998. Guidelines for ecological risk assessment. EPA/630/R-95/002F. Washington, DC.
  • 5
    Ingersoll CG, et al., 1997. Workgroup summary report on uncertainty evaluation of measurement endpoints used in sediment ecological risk assessments. In IngersollCG, DillonT, BiddingerGR, eds, Ecological Risk Assessment of Contaminated Sediments. Special Publication. Society of Environmental Toxicology and Chemistry, Pensacola, FL, USA, pp 271296.
  • 6
    Voorhees DJ, Kane Driscoll SB, Bridges TS., 1998. Improving dredged material management decisions with uncertainty analysis. DOER-3. Technical Report. U.S. Army Corps of Engineers, Vicksburg, MS.
  • 7
    Spies RB., 1989. Sediment bioassays, chemical contaminants and benthic ecology: New insights or just muddy water? Mar Environ Res 27: 7375.
  • 8
    DeWitt TH, Hickey CW, Morrisey DJ, Nipper MG, Roper DS, Williamson RB, Dam LV, Williams EK., 1999. Do amphipods have the same concentration-response to contaminated sediment in situ as in vitro? Environ Toxicol Chem 18: 10261037.
  • 9
    American Society for Testing Materials. 1993. Standard guide for conducting solid phase, 10-day, static sediment toxicity tests with marine and estuarine infaunal amphipods. E 1367–92. Philadelphia, PA.
  • 10
    Swartz RC, Cole FA, Schults DW, DeBen WA., 1986. Ecological changes on the Palos Verdes Shelf near a large sewage outfall: 1980–1983. Mar Ecol Prog Ser 31: 113.
  • 11
    Long ER., 2000. Degraded sediment quality in U.S. estuaries: A review of magnitude and ecological implications. Ecol Appl 10: 338349.
  • 12
    Sindermann CJ., 1997. The search for cause and effect relationships in marine pollution studies. Mar Pollut Bull 34: 218221.
  • 13
    Chapman, PM., 1995. Do sediment toxicity tests require field validation? Environ Toxicol Chem 14: 14511453.
  • 14
    Thursby G, Heltshe BJ, Scott KJ., 1997. Revised approach to toxicity test acceptability criteria using a statistical performance assessment. Environ Toxicol Chem 16: 13221329.
  • 15
    Long ER, Chapman, PM., 1985. A sediment quality triad: Measures of sediment contamination, toxicity and infaunal community composition in Puget Sound. Mar Pollut Bull 16: 405415.
  • 16
    Chapman PM, Dexter RN, Long ER., 1987. Synoptic measures of sediment contamination, toxicity and infaunal community composition (the Sediment Quality Triad) in San Francisco Bay. Mar Ecol Prog Ser 37: 7596.
  • 17
    Long ER, et al., 1990. A comparative evaluation of five toxicity tests with sediments from San Francisco Bay and Tomales Bay, California. Environ Toxicol Chem 9: 11931214.
  • 18
    Swartz RC, DeBen WA, Sercu KA, Lamberson JO., 1982. Sediment toxicity and the distribution of amphipods in Commencement Bay, Washington, USA. Mar Pollut Bull 13: 359364.
  • 19
    Swartz RC, Cole FA, Lambertson JO, Ferraro SP, Schults DW, DeBen WA, Lee H II, Ozretich RJ., 1994. Sediment toxicity, contamination and amphipod abundance at a DDT- and dieldrin-contaminated site in San Francisco Bay. Environ Toxicol Chem 13: 949962.
  • 20
    McGee BL, Fisher DJ, Turley S., 1999. Assessment of sediment contamination, acute toxicity, and population viability of the es-tuarine amphipod Leptocheirus plumulosus in Baltimore Harbor, Maryland, USA. Environ Toxicol Chem 18: 21512168.
  • 21
    Long ER, Wolfe DA, Scott KJ, Thursby GB, Stern EA, Peven C, Schwartz T., 1995. Magnitude and extent of sediment toxicity in the Hudson-Raritan estuary. NOAA Technical Memorandum NOS ORCA 88. National Oceanic and Atmospheric Administration. Silver Spring, MD.
  • 22
    Long ER, et al., 1999. Magnitude and extent of chemical contamination and toxicity in sediments of Biscayne Bay and vicinity. NOAA Technical Memorandum NOS ORCA 141. National Oceanic and Atmospheric Administration. Silver Spring, MD, USA.
  • 23
    Adams DA, O'Connor JS, Weisberg SB., 1998. Sediment quality of the NY/NJ Harbor system. EPA/902-R-98–001. U.S. Environmental Protection Agency, Region 2, New York, NY.
  • 24
    Summers JK, Macauley JM., 1993. Statistical summary: EMAP-Estuaries Louisianian Province-1991. EPA/600/R-93–001. U.S. Environmental Protection Agency, Washington, DC.
  • 25
    Anderson BS, et al., 2001. Sediment quality in Los Angeles Harbor: A triad assessment. Environ Toxicol Chem (in press).
  • 26
    Fairey RC, et al., 1998. Assessment of sediment toxicity and chemical concentrations in the San Diego Bay region, California, USA. Environ Toxicol Chem 17: 15701581.
  • 27
    Ferraro S, Cole F., 1997. Field verification of sediment-amphipod toxicity tests and calibration of PAH effects on the macrobenthos. Abstracts, 18th Annual Meeting, Society of Environmental Toxicology and Chemistry, San Francisco, CA, USA, 16–20 November, p 21.
  • 28
    Swartz RC, DeBen WA, Jones JKP, Lamberson JO, Cole FA., 1985. Phoxocephalid amphipod bioassay for marine sediment tox-icity. In CardwellRD, PurdyR, BahnerRC, eds, Aquatic Toxicology and Hazard Assessment: Seventh Symposium. STP 854. American Society for Testing Materials, Philadelphia, PA, pp 284307.
  • 29
    Scott KJ., 1998. Relationship between Ampelisca sediment tox-icity test response and parameters of benthic community conditions. Abstracts, 19th Annual Meeting, Society of Environmental Toxicology and Chemistry, Charlotte, NC, USA, 15–19, November, p 88.
  • 30
    Pearson TH, Rosenberg R., 1978. Macrobenthic succession in relation to organic enrichment and pollution of the marine environment. Oceanogr Mar Biol Annu Rev 16: 229311.
  • 31
    Hyland JL, Van Dolah RF, Snoots TR., 1999. Predicting stress in benthic communities of southeastern U.S. estuaries in relation to chemical contamination of sediments. Environ Toxicol Chem 18: 25572564.