• Benthic impairment;
  • Stressor identification;
  • New Bedford Harbor;
  • Toxicity identification;
  • Nutrients


  1. Top of page
  2. Abstract
  8. Acknowledgements
  10. Supporting Information

Diagnosing the causes of impaired ecosystems in the marine environment is critical for effective management action. When ecological impairment is based on toxicological or biological criteria (i.e., degraded benthic community composition or toxicity test results), managers are faced with the additional problem of diagnosing the cause of impairment before plans can be initiated to reduce the pollutant loading. We evaluated a number of diagnostic tools to determine their ability to identify pollutants in New Bedford Harbor (NBH), Massachusetts (USA), using a modified version of the US Environmental Protection Agency's (USEPA) stressor identification (SI) guidance. In this study, we linked chemical sources and toxic chemicals in the sediment with spatial concentration studies; we also linked toxic chemicals in the sediment with toxicity test results using toxicity identification and evaluation (TIE) studies. We used geographical information systems (GIS) maps to determine sources and to aid in determining spatially integrated inorganic nitrogen (SIIN). The SIIN values of reference and test estuaries were quantified and compared. Using this approach, we determined that toxic chemicals continue to be active stressors in NBH and that a moderate nutrient stress exists, but we were unable to link the nutrient stressor with a source. Also excess sedimentation was evaluated, but it does not appear to be an active stressor in this harbor. The research included an evaluation of the effectiveness of tools under development that may be used to evaluate stressors in water bodies. We found that the following tools were useful in diagnosing active stressors: toxicity tests, toxicity identification and evaluation (TIE) methods, comparison of grain size-normalized total organic carbon (TOC) ratios with reference sites, and comparison of SIIN with reference sites. This approach allowed us to successfully evaluate stressors in NBH retrospectively; however, a limitation in using retrospective data sets is that the approach may underestimate current or newly emerging stressors. Integr Environ Assess Manag 2012; 8: 685–702. © 2012 SETAC


  1. Top of page
  2. Abstract
  8. Acknowledgements
  10. Supporting Information

Diagnosing the causes of impaired aquatic ecosystems is critical for effective management action. When impairment is based on toxicological or biological criteria, rather than specific pollutants, managers are faced with the additional problem of diagnosing the cause of impairment before plans can be initiated to reduce pollutant loads. In the United States, individual states are required to assess all water bodies, and list those that are impaired or threatened. They then must reduce the load of pollutants through the total maximum daily load process (TMDL) ( To successfully reduce the effects of pollutants, managers need tools to diagnose the causes of impairment. In Europe and Canada, the drivers-pressures-state-impacts-responses approach (DPSIR) ( also depends on effective methods to diagnosis impairment. Diagnosing the causes of impairment can be fraught with sources of uncertainty, including lack of accurate data, inadequate loading models, few definitive tests to demonstrate causal links, and insufficient diagnostic tools. Developing better methods for linking biological responses to environmental stressors, to identify the causes of impairments, is an important step in the development of accurate diagnostic tools that are critical for effective remediation or reduction of the pollutant load.

To assist in the development of effective tools for the diagnostic process, we have carried out an analysis with existing data sets from the Superfund site in New Bedford Harbor (NBH) in Massachusetts. Our approach builds on stressor identification (SI) and causal analysis diagnosis decision information system (CADDIS) methodologies (USEPA 2000b). SI/CADDIS uses a tiered iterative approach to identify and eliminate possible causes of impairment and focuses on developing stressor–effect relationships. In addition to tools to evaluate stressor–effect relationships, we evaluated tools to develop source–stressor relationships. Linking the source to the stressor and then to the effect is critical because management actions are often directed at sources, (e.g., an industrial effluent), not stressors or effects. In addition, by considering the source early and throughout the process of pollutant identification, and by linking the source to the stressor, it is possible to increase confidence that the correct pollutant is identified at the end of the process. Therefore, information about source tracking, identification, and ultimately reduction of source emissions allows managers to achieve their remediation and use attainment goals for a watershed or water body.

The objective of this research is to identify and evaluate diagnostic tools and approaches useful in determining the presence of a source, stressor, and effect, and then to link source to stressor, and stressor to effect to produce a definitive identification of the pollutant. We started with biological impairment and used a tiered, iterative design, which allowed us to develop and apply tools to determine the presence of sources, stressors, and effects that might result in impairment to marine and estuarine ecosystems (Figure 1). We evaluated and listed potential diagnostic tools and decision criteria to move between the different phases of the diagnosis (Tables 1 to 3). The tools used in the present study are potentially useful as diagnostic tools; however, the efficacy and reliability of some have yet to be proven within the process of stressor identification.

thumbnail image

Figure 1. Steps in the New Bedford Harbor retrospective analysis to diagnose stressors.

Download figure to PowerPoint

Table 1. Diagnostic tools used in this studya
PhaseStressorMeasureDiagnostic toolDiagnostic informationStatus of tool
  • a

    Adapted from USEPA (2006). DOC = dissolved oxygen concentrations; ERM = effect range median; ESBs = equilibrium partitioning models; GIS = geographical information systems; POTW = publicly owned treatment works; SIIN = spatially integrated inorganic nitrogen; TIEs = toxicity identification and evaluation; TOC = total organic carbon; TSS = total suspended solid; DIN = dissolved inorganic nitrogen.

1AllSpatially explicit data (e.g., population, land use)GIS mapsEvidence of possible sources.GIS maps are widely available. Data needs to be checked for accuracy.
    May be used to determine reference sites. 
    Data plotted on GIS maps may give spatial relationships. 
 AllExisting scientific literatureInformation can be obtained from peer-reviewed journals, technical reports, federal, state, and local websites, and citizens monitoring groups.Evidence of sources, stressors, effects, and their relationships with each other.Information exists but may be difficult to locate and interpret.
2NutrientsGrain size TOCGrain size-normalized TOC compared to a reference.Grain size-normalized TOC ratios outside of normal ranges may indicate an impacted benthic community. Elevated grain size-normalized TOC ratios may indicate eutrophication stress. This measure may be classified as either a stressor or an effect.

Potentially effective but needs further validation (Pelletier et al. 2011)

  DOCComparison of DOC to reference conditions or guidelines.Low DOC is an indicator of organic enrichment resulting in eutrophic conditions. This information may be classified as either a stressor or an effect.Sampling must be statistically representative to avoid erroneous conclusions due to natural factors.
  Mortality/survival or effects on test organisms.Comparison of survival in test sediment and water to control. Indicates the presence of a toxic stressor.Results are evidence of an effect and can serve to integrate all toxic chemicals in the benthic system.Toxicity tests are validated tools to detect the presence of toxic chemicals. Acute tests may not detect chronic, long-term effects, or be effective when the test organisms is not sensitive to the toxicant (e.g., Hg, dioxins).
     Increasing classes of species tested or exposure duration may increase sensitivity. In addition, chronic or more sensitive endpoints should be considered.
 Toxic chemicalsChemical analysesComparison of chemical concentrations with guidelines and effect concentrations allows us to determine if the stressor exists at a high enough concentration to cause effects.Evidence of a toxic chemical stressor.Chemical analyses are relatively sophisticated however methods may not exist for all compounds in all environmental matrices.
     Also, analyses of all possible chemical toxicants is impossible and impractical; therefore one may miss the presence of toxic stressors unless there is some knowledge of which toxic stressors to target.
3Toxic chemicalsToxicity test results of field sediments compared to sediments that have been manipulated to remove specific chemical classes so only one chemical class is active. This process is formally known as TIEs.TIE approachLinks stressor to effect by determining which toxic chemical(s) are causing effects.TIEs have been used in regulatory programs for over a decade. Some methods require further development and validation.
    Can identify active toxic chemicals or classes of chemicals in field samples.Because these methods use toxicity tests, the caveats of toxicity tests are inherent in TIEs (see Phase 2).
  Measurements of chemical concentrations and physical factors that influence their bioavailability (e.g., TOC, AVS).Compare measurements and factors that influence bioavailability with ESBs and other guidelines (ERMs).Links stressor to effect by determining which toxic chemical(s) may be responsible for observed effects.Correct interpretation of data is critical to avoid false positives.
    Identifies active toxic chemicals or classes of chemicals by comparison of chemical concentrations with ESBs or ERMs.New analytes need analytical methods development.
     Must use appropriate geochemical modifying factors.
 Excess clean sedimentsData from land use maps, areal photos that identify the presence or absence of sediment sources.Correlation of TSS measures with land use maps.Correlation of TSS with land use can be evidence for a link between source and stressor for excess clean sediments.Reference or historical deposition rates and associated grain size may not be available.
  TSS measures. May provide information on historical changes in types of deposition as a result of agricultural or construction activities.Criteria for acceptable levels of clean sediments in marine ecosystems have not been established.
  Measure rate of sedimentation in the system.Relate rate of sedimentation to benthic community condition compared to a reference condition or criteria levelCan link stressor to effect for excess sedimentation.Sedimentation rates are generally not available.
     Criteria for acceptable levels of clean sediments in marine ecosystems have not been established.
3AllModelsProvide information on mechanistic, organismal, and ecological pathways that support linkages among sources, stressors, and effects.Information obtained is specific to the model.Status of the tool is specific to the model.
 NutrientsNitrogen loads into the watershed (including riverine, atmospheric, and POTWs).Loading information can be integrated into SIIN that can be compared to “clean” and “eutrophic” reference systems and used as an indicator of normal or excessive DIN loadings.Links nutrient stressor with effect.Relationships need to be developed among DIN, and benthic effects.
  Volume of area.  Knowledge of normal levels of SIIN are difficult to determine.
  Residence time of area.  Reference estuarine systems may be difficult to acquire.
 NutrientsMonitoring of N, P, chlorophyll a, and DO over appropriate time periods.Provides temporal information on the amounts of nutrients and proximal stressors for nutrients present.Links nutrient source to stressor.Reference estuarine systems may be difficult to acquire.
     Nutrient criteria for marine systems have not been established.
Table 2. Decision criteria matrix for the diagnosis of stressors: Possible outcomes of Phase 2a
SourceStressorEffectGuidelines for action
  • a

    The consideration of source, stressor and effect would be performed for each potential class of stressors. — = data does not support the presence of source, stressor, or effect; DO = dissolved oxygen.

Eliminate stressor from consideration.
XXXMove stressor to Phase 3.
XXLook to a larger system to ensure all possible sources are considered.
   Look for a precursor to the stressor that may have a source.
   If the source is not found, move the stressor to Phase 3 to determine if a strong linkage can be developed between the stressor and the effect. If a strong linkage can be developed, invest more resources to determine if a source exists.
XXThe source may emit the stressor episodically or at certain times of the year; if possible, collect data targeting temporal events.
   The source may emit a different stressor with a similar effect; revisit list of possible sources to ensure all possibilities are included.
   Repeat Phase 2 with attention to all possible stressors; the source may be linked by a proximal stressor in the causal pathway (e.g., low DO or high chlorophyll events in the absence of nutrient concentrations).
   If ultimately no evidence of a stressor or proximal stressor is found, eliminate stressor from consideration.
XXEnsure that the endpoint measured is a sensitive effect endpoint and/or the stressor is bioavailable; given the preliminary nature of Phase 2 endpoints, possibly move to Phase 3.
   If no effect is found, eliminate stressor from consideration.
XEliminate stressor from consideration.
XEnsure that the endpoint measured is a sensitive effect endpoint.
   Consider lag time for the stressor to have a measurable effect.
   Look to a larger system to ensure all possible sources are considered.
   Repeat Phase 2.
XConsider other stressors that may have similar effect on endpoints.
   Consider an ephemeral or episodic stressor.
   Look to a larger system to ensure all possible sources are considered.
   Repeat Phase 2.
Table 3. Possible outcomes of Phase 3: Linkages among source, stressor, and effecta
Source and stressor linkedStressor and effect linkedGuidelines for action
  • a

    In all cases, use the strength of the weight of evidence along with professional judgment in elimination or diagnosis of a stressor. — = data does not support linkage between source, stressor, or effect.

Eliminate stressor from consideration, diagnosis completed.
XXLinks are established between source, stressor, and effect. Diagnosis completed.
  Confidence in the diagnosis is increases if 2 independent types of evidence point to the same stressor (e.g., laboratory and field measurements).
XLook to a larger system (increase spatial scale) to ensure that all possible sources are considered.
  Consider management actions that are directed at stressor removal or mitigation.
XA possible result of temporal loading; repeat Phase 3 with attention to temporal scales.
  Source may emit several different stressors; repeat Phase 3 with attention to all possible stressors.
  The stressor may not be at sufficient levels to produce the observed effect.
  If no stressor–effect link is ultimately found, eliminate stressor from consideration.


New Bedford Harbor is a highly contaminated industrialized estuary located in southeastern Massachusetts, in the United States (Figure 2A) (Weaver 1984; Pruell et al. 1990; USEPA 1996b). The focus of the current retrospective study is the upper harbor portion of the NBH Superfund site (Figure 2B). New Bedford Harbor was chosen for this retrospective diagnostic case study, because the site has an existing data set comprised of nearly 15 years of monitoring and experimental data and includes sediment and water chemistry, sediment toxicity, and benthic community analyses. New Bedford Harbor has high concentrations of PCBs and Cu in the sediments, as well as high nutrient inputs from the Fairhaven Wastewater Treatment Plant, which services a city of over 16 000. In addition it receives a myriad of contaminants and nutrients from combined sewage overflows (CSOs) and urban runoff. Under the TMDL program, NBH is listed as an impaired water body because of excess metals, nutrients, oil and grease, pathogens, organic enrichment and low DO, toxic and organic compounds that violate water quality standards. Concern over PCB effects on human and ecological health resulted in this site being placed on US Environmental Protection Agency's (USEPA) National Priorities List for remediation under the Superfund Amendments and Reauthorization Act (Superfund). Superfund sites are typically monitored before, during, and after the remediation process is initiated. Long-term monitoring at NBH (NBH-LTM) was started in 1993 (USEPA 1996b). For this retrospective study, we considered 3 categories of pollutants: toxic chemicals (organic chemical compounds and metals), nutrients (excess N), and sedimentation (excessive clean sediments entering the system). These 3 were chosen because of NBH's TMDL listing for toxic organics and metals, organic enrichment, and low DO, and because they are the pollutants most commonly listed by states as major categories of stressors in the TMDL process (Kurtz et al. 2006).

thumbnail image

Figure 2. New Bedford Harbor area map (A) and industry in the estuary (B).

Download figure to PowerPoint

Reference sites

To evaluate benthic communities, grain size-normalized total organic carbon (TOC), total suspended solids (TSS), and spatially integrated inorganic N (SIIN), we used the reference site comparison approach (Reynoldson et al. 1997; Bailey et al. 1998). The west branch of the Westport River in Massachusetts (WP) (Figure 2A) was chosen as a “clean” or “unimpacted” reference site for benthic community and grain size-normalized TOC comparisons; this system has no known anthropogenic point source inputs, and sediment analysis showed background levels and low concentrations (low ppb) of heavy metals and PCBs, respectively (Johnson et al. 2001). Greenwich Bay, Rhode Island (GB) (Figure 2A) was chosen as a nutrient “impacted” reference site for comparison of grain size-normalized TOC and SIIN. An impacted reference site is known to be impaired; in this case, the western portion of GB has documented events of low DO (Rhode Island Department of Environmental Management 2003; Bergondo et al. 2005), and high chlorophyll levels (Bergondo et al. 2005), which serve as proximal measures for nutrient enrichment. Greenwich Bay has a similar mean salinity (∼29.0 ppt) compared to NBH (∼28.8 ppt), and is within 50 miles of NBH, so the oceanic loads of N entering the 2 systems were assumed to be similar. Finally, we used a clean site in Narragansett Bay (NarrB) (Figure 2A) as a reference site for TSS. For this study, a change in benthic community, as measured by comparison to the Westport River, was designated as the impairment. The objective of the diagnostic method was to determine the cause of this benthic impairment in upper NBH.

The 4 phases

We used a multiphased diagnostic process (USEPA 2006) (Figure 1):

  • Phase 1 is the screening step during which we gathered and examined existing data and confirmed the initial assessment of biological impairment from the 303(d) listing.

  • Phase 2 establishes the presence of the source, stressor, and effect. We looked for evidence of a source, the presence of the stressor, and an effect that was consistent with the considered stressor. Phase 2 brings into consideration effects that may be more specific to the stressor class, i.e., toxicity.

  • Phase 3 establishes evidence of empirical, logical, or mechanistic links between the source, stressor, and effect. These linkages can include mechanistic evidence that shows the stressor could produce the effect, and/or evidence that manipulating the source output would change the stressor, or a change in stressor would change the observed effects.

  • Phase 4 is a confirmation step generally carried out postTMDL or postremediation. In the case of NBH, remediation is currently underway for toxic chemicals under the Superfund program. Because NBH is a Superfund site, monitoring will occur for decades to come (USEPA 1996b), which will allow us to confirm or deny our diagnosis at some point in the future.


  1. Top of page
  2. Abstract
  8. Acknowledgements
  10. Supporting Information

We illustrated this 4-phase approach (Fig. 1) by applying it to each of the 3 pollutant categories: toxic chemicals, nutrients and excess sedimentation. For each pollutant we described the steps taken in each phase of the method and discussed specific tools that could be used as evidence for the presence of source, stressor, and effect, or to link the source, stressor, and effect. When organizing data for Phase 3, we approached stressors individually and used a matrix to organize the information. Potential diagnostic tools are listed in Table 1 (USEPA 2006).

Toxic chemicals

Toxic chemicals—Phase 1: Screening

We evaluated existing temporally and spatially colocated data sets from the USEPA Environmental Monitoring and Assessment Program (EMAP), STORET database (, NBH- LTM program, the Commonwealth of Massachusetts, local municipality databases, and peer-reviewed literature (Brown and Wagner 1990; Pruell et al. 1990). Existing National Oceanographic and Atmospheric Administration bathymetric maps and GIS maps developed for the Superfund action, as well as published data, were also consulted for point source locations. Maps showing contaminant distributions indicated that the upper portion of NBH, north of the Coggeshall Street bridge was the most contaminated with both PCBs and Cu and had highest levels of toxicity and benthic community impairment (Figures 3A–D from USEPA [1996b]). Benthic community analysis compared to the WP reference site indicated a difference between the sites using species richness and evenness measures, as well as the dominance of Streblospio benedicti, a small polychaete worm commonly associated with organic enrichment and chemical contaminants (Rakocinski et al. 1997; Weisberg et al. 1997) (Appendix A, Benthic Community Analysis, provided in the Supplemental Data, contains further details on the analysis). Differences in the benthic structure suggested the presence of impairment; however, the cause of these benthic effects could not be discerned from this line of evidence. Results from toxicity testing and benthic community analysis confirmed that biological impairments existed for NBH and satisfied an objective of Phase 1.

thumbnail image

Figure 3. Gradients in upper, lower, and outer New Bedford Harbor (Nelson et al. 1996). (A) PCBs. (B) Cu. (C) Toxicity. (D) Number of species.

Download figure to PowerPoint

Toxic chemicals—Phase 2: Evidence of source, stressor, and effect

In Phase 2, we used sediment contaminant concentrations, toxicity test results, benthic community data, historical knowledge of the harbor, and GIS-based mapping to establish the presence of sources, stressors, and effects in the upper portion of NBH (Pruell 1984; Weaver 1984; USEPA 1996b; Ho et al. 1997) (Table 4).

Table 4. Data used to identify stressors in NBH
 NBH Station 114aNBH Station 117aWPbGBNarrBERM (ppm dry weight)c
  • NBH = New Bedford Harbor; ERM = effect range median; TOC = total organic carbon; WP = Westport River; GB = Greenwich Bay; NarrB = Narragansett Bay; NA = not available.

  • a

    USEPA 1996b.

  • b

    Johnson et al. 2001.

  • c

    Long et al. 1995.

  • d

    Perron et al. 2005.

  • e

    Estimated from Pruell et al. (1990).

  • f

    B. Howes, Massachusetts Department of Environmental Protection, personal communication.

  • g

    Hoagland et al. 1988.

  • h

    Latimer et al. 2003.

  • i

    Hartmann et al. 2005.

  • j

    Abdelrhman 2005.

TOC (%)9.811.54.04.0d  
Toxicity (% survival) 10-day Ampelisca abdita test52 ± 2335 ± 10NA100d  
% Silt–clay65.863.893.658.7d  
% Sand32.434.4NA44.2d  
PCB (total) (µg/g dry weight)190 000310 00019.0  0.180
PAH (µg/g dry weight)∼170e∼170eNA  44.80
Cu (µg/g dry weight)62282012.1  270
Cd (µg/g dry weight)1218NA  9.6
Pb (µg/g dry weight)40451615.0  218
Zn (µg/g dry weight)879107070.4  410
Total suspended solids (mg/L)7.9 ± 1.3f7.9 ± 1.3f4.3 ± 2.3g29.2 ± 41.1g35.87 ± 57.41d 
Sedimentation rate (cm/day)0.25–0.39h  0.5–0.85i  
Area (km2)j4.1  12  
Volume (m3)16.1 × 10e  28.4 × 10e  
Flushing time (days)12  7  
Load (kg N/d)135  333  
Chlorophyll (µg/l)10–20f  13.6 ± 8.7d  

Historical data indicated that a number of industries were located on the edge of NBH through the latter part of the last century (Pesch and Garber 2001). This information was used to create layers in a GIS-based map of potential toxic chemical sources (MassGIS, Landuse (Figure 2B). Most notably, 2 electronic parts manufacturers, Aerovox Corporation and Cornell Dubilier, directly discharged PCB-containing wastes into the harbor for over 30 years (Pesch and Garber 2001). In addition, Revere Copper and Brass and Atlas Tack were likely sources of heavy metal concentrations found in the sediments (Pesch and Garber 2001) (Table 4 and Figure 3B). From this information, we concluded there was strong evidence of a source of both organic and inorganic toxic chemicals. Published data indicated elevated concentrations of stressors, more specifically, PCBs, PAHs, and heavy metals, especially Cu, in the sediments of the upper bay (sites 114 and 117) relative to reference sediments (USEPA 1996b) (Table 4 and Figures 2B, 3A, and 3B). Ten-day whole sediment toxicity tests carried out with the amphipod Ampelisca abdita indicated an effect of significant toxicity at upper harbor stations 114 (52% ± 23% survival) and 117 (35% ± 10% survival) (USEPA 1996b). Thus we found evidence of source, stressor, and effect for toxic chemicals in the upper NBH estuary.

The high level of toxicity is a strong indicator of a bioavailable toxic chemical. Linking this effect to a stressor and a source would strengthen our conclusions and allow managers to take effective remediation actions. In addition to the evidence of toxicity, an impaired benthic community and elevated grain size-normalized TOC were present in NBH (see “Grain size-normalized TOC—A Tool for the Measure of Eutrophication” section and Appendix B, provided in the Supplemental Data). Both the impaired benthic community and the elevated TOC–grain size ratios are effects that may result from toxic action on organisms that either comprise a healthy community or mineralize organic C (Koepfler and Kator 1986; Zhou et al. 2009). If these organisms are compromised, there will most likely be a shift in the benthic community as well as a change in the mineralization rate, and therefore a shift in the level of organic C (Zhou et al. 2009) (Figure 4).

thumbnail image

Figure 4. Total organic C normalized by grain size. The straight lines are the 99% confidence intervals for 359 reference sediment samples from the 1990 to 1993 Virginian Province EMAP data set. Both NBH and our “impaired” reference, GB, fall outside the 99% confidence intervals for “unimpacted” reference sediments. Data were obtained from EMAP (

Download figure to PowerPoint

Table 2 identifies possible outcomes when considering Phase 2 evidence for the presence of source, stressor, and effect and suggests appropriate next steps. The results of Phase 2 for toxic stressors were organized in a matrix (Table 5). For toxic chemicals, there was evidence of sources (e.g., defunct Aerovox, Cornell Dubilier, and Revere Copper manufacturing plants), stressors (historically high sediment concentrations of PCBs and Cu), and effects (significant mortality in amphipod toxicity tests, impairment of the benthic community, and elevated grain size-normalized TOC). Table 2 guidelines indicate that, when there is evidence of all 3 conditions, the stressor should be moved to Phase 3 (USEPA 2006). Because all 3 lines of evidence (source, stressor, and effect) are positive for toxic chemicals (Table 5), we moved this stressor to Phase 3, where links between the source and stressor, or stressor and effect, could be verified (Table 2).

Table 5. Phase 2 summary of sources, stressors, and effects in NBH
  1. CSO = combined sewage overflows; GIS = geographical information systems; NBH = New Bedford Harbor; POTW = publicly owned treatment works; TOC = total organic carbon; DIN = dissolved inorganic nitrogen.

Toxic chemicalsHighly industrialized area (Aerovox, Revere Copper and Brass, Cornell Dubilier)Chemical concentrations in sedimentsToxicity (as evidenced by toxicity tests)
   Benthic community impairment
   Elevated grain size-normalized TOC
NutrientsSeptic systems (CSOs, POTWs)Riverine DIN may be elevatedBenthic community impairment
   Elevated grain size-normalized TOC
   Low dissolved oxygen
   High chlorophyll levels
SedimentationGIS and existing literature suggest no current sources (100 y) of sedimentationTotal suspended solids measuresBenthic community impairment
   Elevated grain size-normalized TOC
Toxic chemicals—Phase 3: Linking source, stressor, and effect

The objective of Phase 3 was to establish links among source, stressor and effect. Tools used to establish links between source and stressor include spatial chemical measures in sediments or waters (Table 1). For example, chemical concentrations that were higher closer to the source and lower farther away from the source would indicate that source and stressor were linked. Such was the case in NBH where chemical concentrations of PCBs were higher closer to the historic sources and lower farther away (Figure 3A). Corresponding amphipod toxicity and benthic community measures also showed this pattern (Figures 3C and D). To link toxic stressor and effect we used TIE methods, chemical analysis, and geochemical modifying factors, e.g., acid-volatile sulfide (Di Toro et al. 1990) and organic C concentrations (Di Toro et al. 1991) to determine bioavailable concentrations. Additional supporting evidence was obtained by using published guidelines (Long et al. 2001; USEPA 2003b, 2004, 2005; Wenning et al. 2005) (Table 1). We determined that the bioavailable contaminants were present at levels sufficient to account for observed toxicity (USEPA 2004, 2005).

Total PCBs, PAHs, Cu, Cd, Pb, and Zn were all above the appropriate effects range median (ERMs) levels (Table 4). Effects range medians are the concentrations of a contaminant above which harmful effects always or almost always occur (Long et al. 1995). Effects range medians are obtained from field mixtures of contaminants and it is not possible to ascertain the effect of a single chemical from the mixture; however, if the concentration of the stressors were all below the ERM values, this would cast doubt on the choice of stressor. The TIE results specifically ruled out metals and PAHs, indicating that PCBs were the major cause of toxicity in NBH sediments from the upper harbor (Ho et al. 1997). In a TIE, chemical analysis and toxicity testing are used in an iterative, phased approach to identify the chemical or class of chemicals causing toxicity (USEPA 1989, 1991, 1993, 1996a, 2007). Chemical analysis carried out as part of the TIE indicated that the sum of the PCB concentrations in interstitial waters from NBH sediments totaled approximately 40 ppb (Ho et al. 1997). This concentration was within the upper bounds of the 96-h LC50 concentrations for amphipods (10 to 40 ppb) and could account for the observed toxicity (Ho et al. 1997). The results of this TIE indicated that PCBs were responsible for the acute toxicity of the sediments effectively linking the PCB stressor and toxic effect.

In summary, for toxic chemicals, we linked source and stressor using concentration gradients of PCBs in NBH from GIS source maps (Figure 3A) and then linked stressor and effect using TIEs. Toxicity test results documented in TIEs (Ho et al. 1997) have been linked back to field and community effects in many studies (Swartz et al. 1982, 1994; Ferraro and Cole 2002). According to our guidance (Table 3) (USEPA 2006), links have been established between source, stressor, and effect and 2 lines of evidence (TIE and chemical analyses) point to the same stressor; therefore, the weight of evidence supports a diagnosis of toxic chemicals as an active agent causing impact to the benthic ecosystem of upper NBH (Figure 5A). The original sources of toxic chemicals in NBH were removed in the 1980s when the electronics parts, industries were closed. This situation now leaves the sediments, originally sinks, as the new sources of toxic chemicals. This information supports the remediation methods currently underway for toxic chemicals in upper NBH sediments (stations 114 and 117) via Superfund legislation.

thumbnail image

Figure 5. Matrix linking source, stressor, and effect in Phase 3 diagnosis New Bedford Harbor. (A) Toxic chemicals. (B) Nutrients. (C) Sedimentation.

Download figure to PowerPoint


Nutrients—Phase 1: Screening

The NBH watershed is primarily residential and industrialized, without significant agriculture. No large point sources of nutrients are evident in the upper harbor; however, numerous CSOs are located throughout the harbor, and the Fairhaven, MA wastewater treatment plant outfall is located in the southern portion of the lower harbor approximately 2 km south of the upper portion of NBH (Pesch and Garber 2001) and GIS sources (MassGIS, Landuse (Figure 2B). Because of mixing in the estuary we took the dissolved inorganic N from the Fairhaven wastewater treatment plant outfall into account in our loading calculations (Appendix C, provided in the Supplemental Data). As part of this study, we obtained data on sediment grain size, TOCs, N load values, flushing time, estuary area, and estuary volume for NBH and GB (Table 4).

Nutrients—Phase 2: Evidence of source, stressor, and effect

High nutrient input from the Fairhaven publicly owned treatment works (POTW) source was present (87.9 kg N/d) (L Lima, Fairhaven Wastewater Treatment Plant, personal communication). This input enters the lower harbor approximately 5 km away from the upper harbor. Circulation patterns indicate that water from the lower harbor is moved into the upper harbor (Abdelrhman 2006). High nutrient levels act as a stressor and could result in elevated grain size-normalized TOC (see “Grain size-normalized TOC—A tool for the measure of eutrophication” section) and/or impaired benthic communities when dense algal populations resulting from high nutrient concentrations senesce and sink to the bottom, increasing the organic matter load, decreasing dissolved O2, and increasing biological O2 demand. Excess nutrients can also encourage growth of opportunistic species of algae that change benthic communities either by blocking sunlight and therefore reducing beneficial submerged aquatic vegetation, or encouraging the growth of species that can smother benthic communities (Bischof et al. 2002, 2006). In NBH, the dominant macrobenthic species was Streblospio benedicti, a small polychaete worm commonly associated with organic enrichment and chemical pollution (Rakocinski et al. 1997; Weisberg et al. 1997). Although both elevated grain size-normalized TOC (see “Grain size-normalized TOC—A tool for the measure of eutrophication” section) and an impaired benthic community may result from nutrient stress, they are not specific indicators of it. More specific evidence of adverse nutrient effects would be documented low dissolved O2 (DO) events or algal blooms.

Grain size-normalized TOC—A tool for the measure of eutrophication

The grain size-normalized TOC in sediments (a ratio of the total organic C to the grain size of the sediment) has been proposed as an indicator of the long-term trophic status of benthic ecosystems (Mayer 1994a, 1994b). For the sake of this discussion, we defined eutrophication as the organic enrichment of ecosystems from anthropogenic activities (Odum 1971). In the absence of anthropogenic inputs, the C content of sediments is correlated to grain size (Mayer 1994a, 1994b; Pelletier et al. 2011). The surface area of sediment particles available for the adsorption of organic matter increases toward finer grain size. Based on the relationship between TOC and grain size (Pelletier et al. 2011), we determined the trophic status of estuarine sediments by comparing the expected concentration of C for sediments with the same average grain size. To this end, a linear regression relationship was developed between the square root of TOC (SQRTOC) and the percent silt–clay (%SC), a measure of average grain size, in 359 “clean” or reference sediment samples from the 1990–1993 Virginian Province EMAP, the 1997–1998 Mid-Atlantic Integrated Assessment, and the 2000–2001 National Coastal Assessment-Northeast data sets (Figure 6) (Pelletier et al. 2011). The relationship had a correlation coefficient of 0.86 (see Appendix B for details, provided in the Supplemental Data).

thumbnail image

Figure 6. Conceptual model diagram of nutrient inputs, sedimentation, and grain size-normalized total organic C measure. + signifies an increase; − signifies a decrease.

Download figure to PowerPoint

Grain size-normalized TOC data from NBH stations 114 and 117 fell outside the 99% confidence interval of the reference relationship (Figure 6). Although this indicator is evidence that the sediments were enriched in C, it does not identify the source of the C enrichment, which may be from excessive nutrient) or C loading. We examined the grain size-normalized TOC for the clean reference stations in the WP, which had background riverine nutrient inputs similar to NBH, but did not have CSO or POTW inputs. The grain size-normalized TOC levels in WP (1.70) fell within the bounds of the 99% confidence interval for reference conditions of sediments with silt–clay of 90.4% to 96.4% from the EMAP data set (Figure 6). In addition, we also found the grain-size normalized TOC for the western portion of GB (our impacted reference, or eutrophic, site) to be outside of the bounds of 99% confidence intervals—additional evidence for using grain size-normalized TOC ratios in delineating eutrophic areas. Along with contemporary loadings of excess C, NBH was exposed to excess C loads during the early nineteenth century (Pesch and Garber 2001).

Grain-size normalized TOC could be considered a stressor rather than an effect using evidence that it causes changes in the composition of the benthic community. However, in the context of linking source-stressor-effect, we found it useful to think in a logical continuum of a source (e.g., a wastewater treatment plant) that emits a stressor (e.g., excess particulate organic C) that causes an effect (e.g., elevated grain-size normalized TOC), which indicates that the ecosystem has been compromised (see conceptual model in Figure 4). Regardless of how the measure is classified, we believe it is a useful indicator of eutrophication and can be viewed as stressor or effect, depending on the usefulness of the classification in solving a problem.

As was true for toxic chemicals, there was reasonable evidence of nutrient sources, stressors, and effects in NBH (Table 5), and so we moved this stressor to Phase 3, where the links between source and stressor, or stressor and effect, would be determined.

Nutrients—Phase 3: Establishing links among source, stressor, and effect

For nutrients, linking source and stressor, or source and effect, is not as straightforward as for toxic chemicals. Nutrient stressors can be difficult to assess directly because nutrients are scavenged very quickly and recycled in a complex web of interactions in marine systems (USEPA 2001). Furthermore, the effect of the stressor may be many kilometers away or separated in time from the actual source because of hydrodynamic transport and the lag time characteristic of ecological interactions governing nutrient concentrations. We used low DO and elevated chlorophyll levels as proximal indicators of nutrient effects, and attempted to link source and effect by using monitoring data to discern corresponding patterns of these 2 measures. Environmental Monitoring and Assessment Program data indicated that DO levels ranged from 0 to 4 ppm in the lower harbor in NBH during the summer of 1990 ( Levels of DO continuously measured at 0.3 m above the sediment surface in the upper harbor ranged from approximately 4 to 6 ppm with slight dips below 4 ppm over a 3-month summer period in 2002 (B Howes, University of Massachusetts, Dartmouth, MA, personal communication). Chlorophyll levels for the same period ranged between 10 and 20 µg/L (B Howes, Massachusetts Department of Environmental Protection, personal communication). In general, these DO and chlorophyll levels indicate a moderate to severe nutrient example (USEPA 2000a, 2001, 2003a). For example USEPA (2000a) reports DO levels less than 2.3 mg/L represent a risk to aquatic life, and levels between 2.3 and 4.8 mg/L indicate that some organisms may be stressed. In addition, chlorophyll a levels above 10 µg/L are often found in estuarine waters classified as eutrophic (USEPA 2000a, 2001, 2003a.). To establish a definitive link between nutrient sources and in situ concentrations of the stressor, more detailed spatial monitoring and/or models of nutrient concentrations in the area would be required.

Spatially Integrated Inorganic Nitrogen (SIIN)—A tool for measuring nutrient stress

We linked the nutrient stressor with an effect by comparing SIIN between our test estuary and an estuary known to be nutrient stressed. The SIIN includes dissolved inorganic nitrate, nitrite, and ammonia. The SIIN in NBH was compared to the SIIN in GB, an estuary with known nutrient stress (Rhode Island Department of Environmental Management 2003; Bergondo et al. 2005) and high chlorophyll levels (Table 4). If the SIIN in NBH were equal to or higher than the SIIN in GB, we would conclude that there is nutrient stress in NBH. Recent toxicity tests indicate an absence of sediment toxicity throughout GB (Table 4), which further implicates low DO as the likely cause for the repeated fish kills observed in GB. For both NBH and GB, we calculated the SIIN of the entire estuary, because the data we used does not let us discriminate separate coves or areas of the estuaries.

In this approach, the calculation of SIIN in an estuary takes into account the volume and flushing time of the estuary as well as the load of N from rivers, atmospheric deposition, and other sources. The SIIN is adapted from Dettmann (2001) and is calculated using the formula

  • equation image

where SIIN is the spatially integrated concentration of dissolved inorganic N in the system, L is the load input by atmospheric deposition, rivers, and sewage treatment plants (kg/day), t is the freshwater flushing time (days), and V is the volume of the estuary (m3). Using the SIIN approach, we determined the SIIN of GB, the “dirty” reference station to be 5.86 µM/L N and the SIIN of NBH to be 7.19 µM/L N. (For details on the SIIN calculation, see Appendix C, provided in the Supplemental Data.) Because no reliable measures of variability were available from these data, we can only conclude that the level of SIIN in GB (5.86 µM/L N) was equal to or less than the level of SIIN in NBH (7.19 µM/L N). Thus, by this measure we expected NBH to be affected by nutrient stressors, because GB (an area known to have nutrient impacts) had nutrient concentrations equal to or less than NBH. Although this model indicated nutrient stress in both NBH and GB, it begs the question: Why do the effects of nutrient stress seem much greater in GB (documented fish kills, very low DO) relative to NBH? Reasons for this apparent difference may include: 1) advection of low DO water from upper NarrB into GB, which may place GB into a higher risk category for low DO events, 2) N loading into GB may be underestimated due to unmeasured N-rich groundwater flows, 3) toxic chemicals in NBH could prevent algal blooms from occurring to the extent they occur in GB, and 4) GB has an east–west orientation and NBH has a north–south orientation. Prevailing summer winds in this area are southerly, which would enhance mixing and therefore increase DO in the waters of NBH but not GB.

Additional evidence that supported our conclusions come from 2 other studies. Hauxwell et al. (2003) stated that estuarine N loads greater than 60 kg N ha−1 y−1 would result in changes in community structure, loss of eelgrass beds and changes in C and N cycling. If we normalized our loading rate to NBH (135.4 kg N d−1 by area [4.1 km2, Appendix C, provided in the Supplemental Data]) we obtained a N load of 120 kg N/ha−1 y−1, which implied that observed benthic community impairment could be caused by excess N. In addition, Latimer and Rego (2010) determined that shallow estuaries with N loading over 100 kg N ha−1 y−1 are susceptible to impairment that implied that NBH with a N load of 120 kg N/ha−1 y−1 would be susceptible.

Because more than 2 independent lines of evidence supported the same conclusion concerning a stressor, the weight of evidence links the stressor (excessive N inputs) and effect (observed impaired benthic communities) in NBH. The following 3 measures taken together indicate a moderate nutrient stress in NBH:

  • Monitored DO and chlorophyll a levels that indicated moderate nutrient stress,

  • high SIIN calculations compared to an impaired site, and

  • estuarine loading values determined from land use were comparable to those shown to cause impairment.

For excess nutrients, although we found that long-term DO monitoring showed relatively low DO readings in the lower harbor of NBH, we were unable to concretely link specific nutrient sources with effects to demonstrate benthic impairment. New Bedford Harbor, like many other estuaries, has more than 1 source of nutrients. To link nutrient effects with specific sources, time series DO and chlorophyll a measures would be needed throughout NBH; we were able to obtain data for upper NBH only and consequently were unable to compare the importance of nutrient sources in upper and lower NBH using gradients similar to PCB toxicity. Although we were unable to link the stressor to a specific nutrient source, it is clear that nutrient sources exist in both the upper and lower portions of NBH (Figure 2B), and effects (although not specific to nutrients) exist in the upper harbor. Based on this information that links stressor to effect, we concluded that there may be a moderate to severe nutrient stress in NBH and recommend monitoring to link a source with the stressor and effect (Table 3 and Figure 5B). Linking a source to the stressor and effect is key for TMDL actions as management activities are focused on sources, not stressors or effects.

Excess sedimentation

Excess sedimentation—Phase 1: Screening

We found no evidence for a source of excess sedimentation. However, evidence of a stressor (moderate TSS levels and sedimentation rates), and evidence of an effect (impaired benthic community and elevated grain size-normalized TOC ratio) existed. Although the evidence for the effect is not specific to sedimentation, we moved this stressor to Phase 2 to determine if the stressor can truly be linked to the observed effects. If so, we would invest more energy in looking for a possible source beyond the immediate water body.

Excess sedimentation—Phase 2: Evidence of source, stressor and effect

There was no recent evidence, within the last 100 years, of a significant sediment source such as clear cutting, or certain agricultural practices that would cause large sediment loads to enter NBH from the watershed (Pesch and Garber 2001). In addition, there was no evidence of unusual construction activity or stormwater runoff, which could increase sediment loads. Total suspended solid concentration and sedimentation rate, both measures of sedimentation, were available (Table 4). These measures were generally moderate and a more thorough comparison to reference sites was carried out in Phase 3. Evidence of an effect included an impaired benthic community and elevated grain size-normalized TOC ratio. As with nutrient stress, these effects may occur when excess organic sedimentation is present, but they are not specific indicators of excess sedimentation. Excess mineral sedimentation can result in a depression of the expected grain size-normalized C, which was not observed. Guidance in Table 2 suggests that if a stressor and an effect are present, but no sources are evident, one should look beyond the currently considered system to ensure that all possible sources have been considered (USEPA 2006). Land use GIS maps (MassGIS, Landuse showed no potential sources of sedimentation, e.g., logging or agriculture, and despite conducting these steps, no evidence of a source of excess sedimentation was found. Table 2 suggests that an investigator might want to move the stressor to Phase 3 to determine if a link can be made between the stressor and effect. If a strong link can be found between the stressor and the effect, there would be additional impetus to locate a source. For completeness of this study, we moved this stressor to Phase 3 for further evaluation of the link between stressor and effect.

Sedimentation—Phase 3: Linking Source, stressor, and effect

After following the guidance listed in Table 2 (USEPA 2006), we still could not find a sedimentation source. We then evaluated the strength of the stressor–effect (or stressor–response) signal in Phase 3. If the signal was very strong, it would make sense to invest more resources in seeking a source. We compared the TSS measures in NBH (7.88 ± 1.32 mg/L) (Table 4) with TSS measures from both the reference station in WP (4.3 ± 2.3 mg/L) and a unimpaired reference station in NarrB (35.87 ± 57.41 mg/L) (Table 4). The NarrB-TSS values were measured biweekly during the summer of 2004. The NBH-TSS measures were taken throughout July and September 1987 (Garten et al. 1996). New Bedford Harbor TSS measures were lower than NarrB but higher than the west branch of the WP. In general, all the numbers measured were relatively low compared to published Washington State guidelines for TSS (excessive TSS levels >50 mg/L; Because we observed no effects of excess sedimentation at the highest of our measured sites in our reference site at NarrB (unpublished data), we concluded that concentrations less than 36 mg/L of TSS would not cause an impairment of the benthic community. High TSS measures are 1 indication of excess sedimentation in an estuary.

Another method to determine if excessive sedimentation might be causing benthic community impairments is to compare sedimentation rates. Latimer et al. (2003) reported sedimentation rates in NBH to range between 0.24 to 0.39 cm/y. Hartmann et al. (2005) found that sedimentation rates in GB range between 0.5 and 0.85 cm/y. Although the levels in NBH are lower than in GB, there are no published sediment rate levels indicative of degraded or pristine areas. More baseline scientific data is needed before those levels can be determined. Although an effect in NBH (impaired benthic community and an altered grain size-normalized TOC level) that may be caused by changes in sedimentation was observed (Figure 4), the effect was general or nonspecific and might have resulted from a number of stressors. Given the lack of linkage between the excess sediment stressor and effect, and the lack of evidence of a source of sediment, we did not recommend moving this stressor forward into the TMDL or remediation stage (Figure 5C and Table 3).

Summary of evidence from Phases 1–3

In summary, given the strong evidence linking source, stressor, and effects for toxic chemicals, this analysis supports the Superfund remediation underway for chemical stressors. For nutrients, although a plausible link occurs between stressor and effect, based on available data, we were unable to link the nutrient source with the stressor. We recommend further monitoring of nutrients, and indicators of nutrient enrichment, not only to link the nutrient stressor to the source but also to ensure that a portion of the observed benthic impairment is related to nutrient stress. Sedimentation rates and TSS measures did not indicate excess sediments were entering the system.

In NBH, and in other marine systems, where more than 1 class of stressor has been found to be present, we showed it was useful to look across the stressors and evaluate each one based on the strengths of the linkages found in Phase 3. Different management actions might be taken depending on the results. For example, in NBH, managers may choose to remediate the toxic chemical stressor first; they might then proceed to collect more data on excess nutrients to determine if nutrient remediation was necessary.


  1. Top of page
  2. Abstract
  8. Acknowledgements
  10. Supporting Information

Our approach to diagnosing the causes of benthic impairment by bioavailable pollutants allowed us to organize the information systematically and succinctly and easily identify stressors and data gaps within the body of available information. Information on data gaps can be used to guide diagnostic tool development for DPSIR, CADDIS and the TMDL process. In cases where there is no data available to either support or disprove the presence of a source, stressor or effect, the data must be obtained through field sampling, monitoring or laboratory studies because it is unlikely that a supportable conclusion on causal stressors will be reached without high quality data.

For Phases 1 and 2 we found the use of GIS maps and existing scientific literature an invaluable resource to find sources, stressors and effects. For Phase 3, we found stressors could be linked to effects using a number of methods: 1) demonstration of known mechanistic routes of exposure and effect, for example, PAHs cause an enhanced photochemical reaction in organisms via oxidative stress; if this reaction can be observed in the laboratory or evidence of oxidative stress can be collected in the field or laboratory, then this is strong evidence that the stressor and effect are linked; 2) empirically through TIEs (USEPA 2007), which link stressor and effect through a series of laboratory manipulations that remove specific stressors and then evaluate organisms for changes in toxicity; and 3) demonstration of field effects correlated with concentration of specific stressors along a field gradient near a source. Additional information may be obtained through comparison of measured levels of the suspected toxicant with known effect levels (Long et al. 2001; USEPA 2003b, 2004; Wenning et al. 2005) (Table 1). In addition to the TIE results that formed the definitive link between stressor and effect for chemical pollutants, the tiered approach advocated in the TIE process and the guidelines of 2 independent lines of information as well as the absence of negative information (information that is not consistent with the suspected stressor) to reach a diagnosis was critical to coming to a conclusion.

Tools under development for Phases 2 and 3 include grain size-normalized TOC as an indicator of long term eutrophication (Pelletier et al. 2010), and the comparative SIIN watershed approach for comparing dissolved inorganic nitrogen (DIN) inputs to determine the relative loading of water bodies. Other tools not evaluated here such as ELISA, enzyme, or protein assays that indicate specific stressors (Moshe and Auslander 2005) and species patterns in benthic communities that may indicate classes of stressors (Christman and Dauer 2003) may be useful for linking stressor and effect. Also, the development of criteria or guidelines for nutrients, sedimentation, and sediment contaminants would be very useful.

We considered low DO a consequence and an indicator of a nutrient stressor; however, low DO can be considered an effect or a stressor, depending on the usefulness of the classification in solving a problem. In general, we have found it useful to be flexible in classifying measures to optimize their use in understanding causal pathways.

Although we have specifically discussed linking source and stressor, and stressor and effect, we do not mean to ignore links between source and effect. For example, greater toxicity observed closer to the source and less toxicity farther from the source provides additional evidence for diagnosing pollutants.


  1. Top of page
  2. Abstract
  8. Acknowledgements
  10. Supporting Information

This research is the first to take an impaired marine ecosystem through a causal diagnosis. In doing so, we evaluated diagnostic tools and diagnosed stressors causing impairment in an estuarine waterbody. New Bedford Harbor provided a great deal of existing high-quality data. Despite the amount of data that existed about NBH, we still did not have enough of the correct types of colocated temporal and spatial data to link nutrient sources to the stressor. In addition, stressor-response information on sedimentation rates and how they relate to benthic degradation was not readily available. Although it is difficult to determine how much data is “enough” to develop a cause-and-effect link, or source-to-stressor link, following TIE guidance of 2 independent lines of information pointing to the same conclusion is certainly a reasonable approach and of course, the more independent lines of information the greater the level of confidence. Much depends on the quality and nature of the data, as well as the professional judgment of the investigators. A limitation of retrospective analyses is that the data sets used are often older and may preclude the detection of emerging or current stressors in a system. Tools such as the grain size-normalized TOC and the comparative estuary approach using SIIN values to assess nutrient effects proved useful, but their efficacy needs to be evaluated further. In this retrospective study, we demonstrated the use of TIEs within a diagnostic approach and the use of sound decision criteria to move through the phases of a diagnosis. This study emphasizes the importance of further development of diagnostic tools and measures to identify stressors that cause impairment in marine environments.


  1. Top of page
  2. Abstract
  8. Acknowledgements
  10. Supporting Information

We thank Drs. Barbara Bergen and Skip Nelson for providing data from the New Bedford Harbor, Long Term Monitoring Program; Dr. Brian Howes for providing data on dissolved O2 and chlorophyll measures in NBH; Mr. Steve Granger for providing data on N loading in Narragansett Bay; and Dr. James Latimer for providing data on nutrient loading in NBH. We also thank Drs. Ed Dettman, William Nelson, Naomi Detenbeck, and Susan Norton for critical review of this manuscript. This is EPA contribution number AED-06-093 of the USEPA, Atlantic Ecology Division (AED), and has been technically reviewed by AED; however, it does not necessarily represent the views of the USEPA. No official endorsement of any aforementioned product should be inferred.


  1. Top of page
  2. Abstract
  8. Acknowledgements
  10. Supporting Information
  • Abdelrhman MA. 2005. Simplified modeling of flushing and residence times in 42 embayments in New England, USA, with special attention to Greenwich Bay, Rhode Island. Estuar Coast Shelf Sci 62:339351.
  • Abdelrhman MA. 2006. A method to incorporate ecology into residence time of chemicals in embayments: Local effect time. Integr Environ Assess Manag 2:247252.
  • Bailey RC, Kennedy MG, Dervish MZ, Taylor RM. 1998. Biological assessment of freshwater ecosystems using a reference condition approach: Comparing predicted and actual benthic invertebrate communities in Yukon streams. Freshwater Biol 39:765774.
  • Bergondo DL, Kester DR, Stoffel HE, Woods WL. 2005. Time-series observations during the low sub-surface oxygen events in Narragansett Bay during summer 2001. Marine Chem 97:90103.
  • Bischof K, Peralta G, Krabs G, van de Poll WH, Perez-Llorens JL, Breeman AM. 2002. Effects of solar UV-B radiation on canopy structure of Ulva communities from southern Spain. J Exp Bot 53:24112421.
  • Bischof K, Rautenberger R, Brey L, Perez-Llorens JL. 2006. Physiological acclimation to gradients of solar irradiance within mats of the filamentous green macroalga Chaetomorpha linum from southern Spain. Mar Ecol Prog Ser 306:165175.
  • Brown JF Jr, Wagner RE. 1990. PCB movement, dechlorination and detoxification in the Acushnet Estuary. Environ Toxicol Chem 9:12151233.
  • Christman CS, Dauer DM. 2003. An approach for identifying the causes of benthic degradation in Chesapeake Bay. Environ Monit Assess 81:187197.
  • Dettmann EH. 2001. Effect of water residence times on annual export and denitrification of nitrogen in estuaries: A model analysis. Estuaries 24:481490.
  • Di Toro DM, Mahony JD, Hansen DJ, Scott KJ, Hicks MB, Mayr SM, Redmond MS. 1990. Toxicity of cadmium in sediments: The role of acid volatile sulfide. Environ Toxicol Chem 9:14871502.
  • Di Toro DM, Zarba CS, Hansen DJ, Berry WJ, Swartz RC, Cowan CE, Pavlou SP, Allen HE, Thomas NA, Paquin PR. 1991. Technical basis for establishing sediment quality criteria using equilibrium partitioning. Environ Toxicol Chem 10:15411583.
  • Ferraro SP, Cole FA. 2002. A field validation of two sediment-amphipod toxicity tests. Environ Toxicol Chem 21:14231437.
  • Garten LS, Bonner JS, Ernest AN, Alutenrieth RL. 1996. Fate and transport of PCBs at the New Bedford Harbor Superfund site. Environ Toxicol Chem 15:736745.
  • Hartmann PC, Quinn JG, Cairns RW, King JW. 2005. Depositional history of organic contaminants in Narragansett Bay, Rhode Island, USA. Mar Pollut Bull 50:388395.
  • Hauxwell J, Cebrian J, Valiela I. 2003. Eelgrass Zostera marina loss in temperate estuaries: Relationship to land-derived nitrogen loads and effect of light limitation imposed by algae. Mar Ecol Prog Ser 247:5973.
  • Ho KT, McKinney RA, Kuhn A, Pelletier MC, Burgess RM. 1997. Identification of acute toxicants in New Bedford Harbor sediments. Environ Toxicol Chem 16:551558.
  • Hoagland MR, Kelley EF, Caldwell DW, Fitzgerald DM. 1988. Hydrogeology and contamination investigation of the West Branch of the Westport River. Boston (MA): Hydrogeology Research Group, Boston University.
  • Johnson RL, Perez KT, Davey EW, Cardin JA, Rocha KJ, Dettman EH. 2001. Discrimination of natural and non-point source (DIN) effects from anthropogenic effects as reflected in benthic state in three New England estuaries. In: 30th Annual Marine Benthic Ecology Meeting, University of New Hampshire, Durham (NH).
  • Koepfler ET, Kator HI. 1986. Ecotoxicological effects of creosote contamination on benthic microbial populations in an estuarine environment†. Toxicity Assess 1:465485.
  • Kurtz JC, Detenbeck ND, Engle VD, Ho K, Smith LM, Jordan SJ, Campbell D. 2006. Classifying coastal waters: Current necessity and historical perspective. Estuar Coast 29:107123.
  • Latimer JS, Boothman WS, Pesch CE, Chmura GL, Pospelova V, Jayaraman S. 2003. Environmental stress and recovery: The geochemical record of human disturbance in New Bedford Harbor and Apponagansett Bay, Massachusetts (USA). Sci Total Environ 313:153176.
  • Latimer JS, Rego SA. 2010. Empirical relationship between eelgrass extent and predicted watershed-derived nitrogen loading for shallow New England estuaries. Estuar Coast Shelf Sci 90:231240.
  • Long ER, Hong CB, Severn CS. 2001. Relationships between acute sediment toxicity in laboratory tests and abundance and diversity of benthic infauna in marine sediments: A review. Environ Toxicol Chem 20:4660.
  • Long ER, MacDonald DD, Smith SL, Calder FD. 1995. Incidence of adverse biological effects with ranges of chemical concentrations in marine and estuarine sediments. Environ Manag 19:8197.
  • Mayer LM. 1994a. Relationship between mineral surfaces and organic carbon concentrations in soils and sediments. Chem Geol 114:347363.
  • Mayer LM. 1994b. Surface area control of organic carbon accumulation in continental shelf sediments. Geochim Cosmochim Acta 58:12711284.
  • Moshe T, Auslander M. 2005. Transcript and protein environmental biomarkers in fish—A review. Chemosphere 59:155162.
  • Odum HT. 1971. An energy circuit language for ecological and social systems, its physical basis. In: Patten B. Systems analysis and simulation in ecology. New York (NY): Academic Press. p 139211.
  • Pelletier MC, Campbell DE, Ho KT, Burgess RM, Audette CT, Detenbeck NE. 2011. Can sediment total organic carbon and grain size be used to diagnose organic enrichment in estuaries? Environ Toxicol Chem 30:538547.
  • Perron MM, Pelletier MC, Ho KT, Burgess RM, Perez K, Cantwell MG, Rocha K, Cardin J, Johnson RL, Davey E. 2005. An approach for diagnosing impairments in estuaries: preliminary results of the Narragansett Bay case study. Annual Meeting of the Estuarine Research Federation, Annual Meeting of the Estuarine Research Federation, Norfolk, VA.
  • Pesch CE, Garber J. 2001. Historical analysis, a valuable tool in community -based environmental protection. Mar Poll Bull 42:339349.
  • Pruell RJ. 1984. The geochemistry of organic contaminants in Narragansett Bay sediments and the availability of these compounds to the blue mussel Mytilus Edulis [PhD thesis]. Kingston (RI): Univ Rhode Island.
  • Pruell RJ, Norwood CN, Bowen RD, Boothman WS, Rogerson PF, Hackett M, Butterworth BC. 1990. Geochemical study of sediment contamination in New Bedford Harbor, Massachusetts. Mar Environ Res 29:77101.
  • Rakocinski CF, Brown SS, Gaston GR, Heard RW, Walker WW, Summers JK. 1997. Macroresponses to natural and contaminant-related gradients in northern Gulf of Mexico estuaries. Ecol Appl 7:12781298.
  • Reynoldson TB, Norris RH, Resh VH, Day KE, Rosenberg DM. 1997. The reference condition: a comparison of multimetric and multivariate approaches to assess water-quality impairment using benthic macroinvertebrates. J North Am Benthol Soc 16:833852.
  • Rhode Island Department of Environmental Management. 2003. The Greenwich Bay fish kill- August 2003 causes, impacts and responses. Providence (RI): Rhode Island Department of Environmental Management.
  • Swartz RC, Cole FA, Lamberson JO, Ferraro SP, Schults DW, DeBen WA, Lee HI, 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.
  • Swartz RC, DeBen WA, Sercu KA, Lamberson JO. 1982. Sediment toxicity and the distribution of amphipods in Commencement Bay, Washington, USA. Mar Poll Bull 13:359364.
  • [USEPA] US Environmental Protection Agency. 1989. Methods for aquatic toxicity identification evaluations: Phase II toxicity identification procedures. Duluth (MN): USEPA. EPA/600-3-88/035.
  • [USEPA] US Environmental Protection Agency. 1991. Methods for aquatic toxicity identification evaluations: Phase I toxicity characterization procedures. 2nd ed. Duluth (MN): USEPA. EPA-600/6-91/003.
  • [USEPA] US Environmental Protection Agency. 1993. Methods for aquatic toxicity identification evaluations: Phase III toxicity confirmation procedures for samples exhibiting acute and chronic toxicity. Duluth (MN): USEPA. EPA/600/R-92/081.
  • [USEPA] US Environmental Protection Agency. 1996a. Marine toxicity identification evaluation (TIE) procedures manual: Phase I Guidance Document. Washington DC: USEPA/Office of Research and Development. EPA 600/R-96/054.
  • [USEPA] US Environmental Protection Agency. 1996b. New Bedford Harbor long term monitoring assessment report: Baseline sampling. Narragansett (RI): USEPA/National Health and Effects Research Laboratory, Atlantic Ecology Division. EPA/600/R-86/097.
  • [USEPA] US Environmental Protection Agency. 2000a. Ambient aquatic life water quality criteria for dissolved oxygen (saltwater): Cape Cod to Cape Hatteras. Washington, DC: Office of Water, Office of Science and Technology, Office of Research and Development. EPA-822-R-00-012.
  • [USEPA] US Environmental Protection Agency. 2000b. Stressor identification guidance document. Washington, DC: Office of Water, Office of Research and Development. EPA 822-B-00-025.
  • [USEPA] US Environmental Protection Agency. 2001. Nutrient criteria technical guidance manual estuarine and coastal marine waters. Washington (DC): Office of Water. USEPA. EPA-822-B-01-003.
  • [USEPA] US Environmental Protection Agency. 2003a. Ambient water quality criteria for dissolved oxygen, water clarity and chlorophyll a for the Chesapeake Bay and its Tidal Tributaries. Washington, DC: Region III Chesapeake Bay Program Office, Region III Water Protection Division, Office of. Water, Office of Science and Technology. EPA-903-R-03-002.
  • [USEPA] US Environmental Protection Agency. 2003b. Procedures for the derivation of equilibrium partitioning sediment benchmarks (ESBs) for the protection of benthic organisms: Endrin. Washington, DC: Office of Research and Development. EPA-903-R-03-002.
  • [USEPA] US Environmental Protection Agency. 2004. Procedures for the derivation of equilibrium partitioning sediment benchmarks (ESBs) for the protection of benthic organisms: PAH mixtures. Washington, DC: Office of Research and Development. EPA-600-R-02-013.
  • [USEPA] US Environmental Protection Agency. 2005. Procedures for the derivation of equilibrium partitioning sediment benchmarks (ESBs) for the protection of benthic organisms: Metal mixtures. Washington, DC: Office of Research and Development. EPA-600-R-02-011.
  • [USEPA] US Environmental Protection Agency. 2006. Conceptual models and methods to guide diagnostic research. Washington, DC: USEPA, Office of Research and Development, National Health and Environmental Effects Research Laboratory, Atlantic, Gulf, and Mid-Continent Ecology Divisions. EPA 600/R-06/024.
  • [USEPA] US Environmental Protection Agency. 2007. Sediment toxicity identification evaluation (TIE) Phases I, II, and III guidance document. Washington DC: Office of Research and Development. EPA/600/R-07/080.
  • Weaver G. 1984. PCB contamination in and around New Bedford, Mass. Environ Sci Technol 18:22A27A.
  • Weisberg SB, Ranasighe JA, Schaffner LC, Diaz RJ, Dauer DM, Frithsen JB. 1997. An estuarine index of biological integrity (B-IBI) for Chesapeake Bay. Estuaries 20:149158.
  • Wenning RJ, Bately GE, Ingersoll CG, Moore DW. 2005. Use of sediment quality guidelines and related tools for the assessment of contaminated sediments. Pensacola (FL): SETAC.
  • Zhou HW, Wong AH, Yu RM, Park YD, Wong YS, Tam NF. 2009. Polycyclic aromatic hydrocarbon-induced structural shift of bacterial communities in mangrove sediment. Microb Ecol 58:153160.

Supporting Information

  1. Top of page
  2. Abstract
  8. Acknowledgements
  10. Supporting Information

Additional Supporting Information may be found in the online version of this article.

ieam_1303_sm_SupplAppFigA1.tif143KSupplementary Appendix Figure A1
ieam_1303_sm_SupplAppFigA2.tif59KSupplementary Appendix Figure A2
ieam_1303_sm_SupplAppendix.doc56KSupplementary Appendix

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.