Introduction

This guide is intended to assist with characterizing injury to freshwater benthic macroinvertebrates (BMIs) in Natural Resource Damage Assessment and Restoration (NRDAR) cases. The contents are narrowly focused on insects, crustaceans, snails, and other invertebrate fauna that are typically considered part of BMI communities and are not intended to address studies of injury to larger benthic taxa such as freshwater mussels, crayfish, or benthic fishes or amphibians. Although some percentage of the community functions as predators, BMIs are predominantly primary consumers (for example, scrapers, shredders, and filterer/gatherer feeding groups [detailed in the “Identification and Metrics” section]) that play an essential role in converting carbon and nitrogen from plant tissues into animal biomass for higher-order consumers, especially in flowing waters. Aquatic contaminants can disrupt the quantity and quality of energy transferred (ecosystem function) by reducing invertebrate biomass and diversity. Additionally, the accumulation of toxic residues in invertebrate tissues may be a source of exposure leading to adverse effects in higher trophic levels. The goal of NRDAR BMI assessments is to establish direct linkages of contaminant exposure to injuries reflected by changes in community structure (for example, reduced density and taxa richness) or by effects at the individual population level (for example, survival, growth, and reproduction). BMIs are infrequently the U.S. Department of Interior (DOI)-managed resource in a NRDAR case, with managed resources more frequently including migratory birds, fish, or other insectivorous vertebrates. Therefore, it is critical to have clearly defined objectives for evaluating BMIs and an understanding of how invertebrate data relate to the quantification of injuries to the DOI-managed resource; some of these pathways are shown in figure 1. This guide is intended to assist decisions on whether or not to proceed with BMI studies, use of existing information and data for screening purposes, and what types of studies can support a BMI-injury determination. Examples of the latter are illustrated in figure 2, and these broader categories of study types form the basis for the sections that follow. This document is intended to provide general considerations and best practices for assessing BMIs. Relevant guidance and references are listed throughout the report as sources for specific methods and analysis.

The following components of (BMI) studies are covered in the following guide:The “Characterizing Chemical Exposure” section discusses the following:

The “Benthic Community Surveys” section discusses the following: The “Toxicity Testing” section discusses the following: The “Data Analysis” section discusses the following: The “Monitoring Restoration Success” section discusses the following:

Characterizing Chemical Exposure

Characterizing injury to an aquatic ecosystem requires understanding of the nature of chemical contaminants present and their concentrations in environmental media. This information is important whether the BMI community (or a particular taxon) is the receptor of interest or is viewed as a potential source of contaminants to organisms at higher trophic levels. Armed with an initial understanding of the range of chemicals present in water or sediment, it is important to identify which chemicals and exposure levels pose substantial hazard of toxicity and (or) bioaccumulation. It is also important to understand factors like mobility and persistence of these chemicals for efficient targeting of samples. For example, toxic effects of ammonia are often related to acute effects of dissolved ammonia; however, toxic effects of persistent pollutants like polychlorinated biphenyls and metals are more likely a result of chronic exposure by way of water, diet, and sediments. In addition to understanding the source of contaminant exposures, knowledge of chemical cofactors that influence contaminant bioavailability may be necessary. Bioavailability is the amount of a chemical that is taken up from the environment and able to elicit an adverse effect. These cofactors may include, but are not limited to, sediment organic matter, acid-volatile sulfide, pH, dissolved organic carbon, alkalinity, and hardness.

Although measuring contaminants in environmental media is essential for estimating injury, measuring contaminant concentrations in BMI tissue may also be necessary for demonstrating exposure of the invertebrate. Measuring contaminant concentrations in invertebrates can be difficult because of the large dry mass of tissue required to achieve acceptable analytical detection limits; often, large numbers of organisms must be composited to produce a single sample. If the primary interest is in food chain transfer from invertebrates to higher trophic levels (for example, fish or amphibians), and the diet of the higher levels is well characterized, it may be more efficient to focus on concentrations in the higher trophic level receptors. However, in some cases where tissue concentrations are measured, there may be sufficient information to relate tissue concentrations in invertebrates to expected effect levels (Jarvinen and Ankley, 1999). Additionally, some persistent contaminants including polychlorinated biphenyls, mercury, and metals can remain in the tissues of emergent adult insects, which can transfer aquatic contaminants into terrestrial ecosystems (Kraus and others, 2014).

Characterizing chemical exposure with a limited scope can also be a useful reconnaissance tool for practitioners. Reconnaissance is a preliminary survey of the area and can help a practitioner determine the potential next steps. A preliminary sample collection during reconnaissance may be done in-house with limited resources, and it does not need to necessarily meet the rigor reserved for the NRDAR sample collection and preparation. However, data generated during reconnaissance could assist in directing future sampling efforts should they be deemed useful.

Benthic Community Surveys

The process of sampling and evaluating of BMI communities is roughly divided into three major activities: field sampling, sorting and counting of the BMI, and identification of BMI and calculation of metrics. Issues related to these major activities are discussed further in the following sections.

Toxicity Testing

Assessment of injury to benthic communities based on field surveys alone has some limitations, notably the possibility that habitat degradation not associated with oil or hazardous substance releases may modify invertebrate communities. Associations between exposure to chemical pollutants and community responses can be further documented with controlled laboratory toxicity tests. These tests may focus on documenting contaminant effects in field-collected water or sediment compared to reference sites, or they may focus on characterizing site-specific toxicity thresholds for contaminants of concern.

Laboratory toxicity tests typically rely on standardized methods that provide a high level of confidence in testing results. As discussed in the “Characterizing Chemical Exposure” section, the environmental fate of the oil or hazardous substance of interest will determine exposure route and the type of toxicity test (that is, surface water, pore water, or sediment) that should be performed. Successful laboratory tests can provide rigorous proof of cause-and-effect relations between chemical exposure and responses of invertebrates. The most widely applicable strategy for evaluating toxic effects on benthic invertebrates is laboratory tests with field-collected sediments (American Society for Testing and Materials, 2020). Tests for toxicity of effluents, which are also applicable to testing ambient waters, are designed to test fish and planktonic invertebrates; however, they may be adapted for testing benthic invertebrates as well. Standard test methods can be used for repeatedly testing sediment from sites undergoing remediation to quantify the recovery of invertebrate populations (Steevens and others, 2020). This repeated laboratory testing is most valuable if it is performed in coordination with ongoing biomonitoring of invertebrate communities.

Laboratory tests have limitations as well, such as difficulty of maintaining exposure conditions (for example, water quality, food quality, and dietary exposure) comparable to conditions in the field. Comparisons of laboratory and field responses may be hampered by the difficulty of calibrating individual responses (survival, growth, and fecundity) from toxicity tests with aggregate responses (species richness, age/size structure, and abundance) from community surveys. One approach that may reduce this laboratory-field disconnect is the modification of standard methods to allow testing with resident species from the study area or closely related surrogates.

Studies with test organisms exposed to chemicals in situ have been recommended to avoid some of the limitations of both laboratory toxicity tests (for example, lack of site specificity) and field surveys (for example, limited evidence of causality) (Burton and others, 2005). Test organisms deployed in the field presumably experience more realistic exposure conditions, making their individual-level responses more directly comparable with responses of resident invertebrate communities. In situ assessments are especially well-suited for documenting effects of episodic or transient pollution events. However, the rarity of in situ tests reflects the limitations of the approach because cages may create unfavorable holding conditions that obscure site effects. In situ tests are subject to greater risks from stochastic events such as vandalism, sedimentation, and high- or low-flow events. The biggest limitation of in situ testing with invertebrates likely is the lack of standard methods and the difficulty of finding laboratories or investigators with the ability to perform in situ tests.

Data Analysis

In all situations regarding collecting data for NRDAR, data must meet quality-assurance and -control guidelines to be legally appropriate for the NRDAR case. Barbour and others (1999) and U.S. Environmental Protection Agency (2016) among others provide guidance on appropriate quality-assurance procedures for activities ranging from sample labelling to taxonomic identification. Other data-quality questions include the following:

Data from field and laboratory studies with benthic invertebrates may be analyzed with a variety of statistical tools, depending on the objectives of the study. Multivariate statistical methods, such as ordination and cluster analysis, are useful as exploratory data analyses to look for spatial or temporal patterns in community data and associations of these data with environmental conditions. Results of these analyses may be helpful for identifying chemicals of concern and sensitive taxa for toxicity studies.

Significant observed toxic effects are often determined by the analysis of variance (ANOVA) or related techniques, and the responses in contaminated field sites or higher laboratory exposure levels can be compared to baseline responses. The nature of these comparisons differs between laboratory and field studies. In laboratory toxicity tests, the baseline response is estimated by a negative control that receives no added chemical but is known to support survival, growth, or reproduction of the test organisms. Controls are not available in field studies, so investigators may select one of the following:

Because it is normal for baseline responses to vary among low-chemical reference sites, determinations of impairment are typically made by comparison to the range of responses from multiple reference samples. For example, a common approach is to define a deviation from reference condition as a high or low quantile (for example, 10th percentile) of the reference distribution (Hawkins and others, 2010). Specifically, low quantiles are used for metrics that decline with increased perturbation and high quantiles are used for those that increase with increased perturbation. Further background on the reference condition and considerations useful in selecting reference sites can be found in Hughes (1995), Reynoldson and others (1997), Barbour and others (1999), Stoddard and others (2006), and the U.S. Environmental Protection Agency (2006, 2016). Another potential approach for showing departure from the baseline is the Before-After-Control-Impact model, which requires temporal baseline data prior to the discharge or release of oil or a hazardous substance (Smith and others, 1993).

If ANOVA detects significant toxicity in a laboratory or field study, the next step in the data analysis is often to resort to concentration-response modeling that utilizes nonlinear regression to estimate threshold concentrations for toxic effects. Based on ANOVA, the concordance of spatial patterns of toxicity with exceedance of toxicity thresholds, which are estimated by concentration-response models, provides strong evidence of exposure and effects. The estimation of thresholds by regression methods to estimate a specified level of effect is a widely accepted approach for demonstrating departure from control response(s). Thresholds for different levels of effect may be appropriate to meet the objectives of NRDAR studies. If conditions for laboratory tests are comparable to those in the field, modeled effect concentrations may serve as benchmarks for assessing toxicity in sites that have not been tested directly. Pathways for appropriate assessment tools and endpoints for BMIs, given the relation between the BMI and the DOI-managed resource, are provided in figure 1. Table 1 expands on figure 1 with more details on the tools, endpoints, and references that are potentially useful in injury quantification and claim development.

Monitoring Restoration Success

A similar suite of techniques used for assessing baseline and injury can be used to monitor remediation or restoration efforts and document movement along the restoration trajectory. For BMIs, careful consideration must be given to points made in the “Key Considerations in Selecting Methods for BMI Community Sampling and Analysis” sidebar, particularly the seasonality of sampling during restoration monitoring. For example, after remediation or restoration, it would be most common to do the following:

In some instances, when an active restoration effort is not considered advantageous for the system, managers may prefer to monitor over time (months, years, or decades) to assess recovery rates to support adaptive management decisions (Magar and others, 2009). Frequency of monitoring after restoration is case-specific; however, take note of the following considerations: (1) in some cases, remedial efforts could initially make things worse for invertebrate communities; (2) if remediation of the release or discharge is successful and physical habitat is rehabilitated, invertebrate communities may respond quickly; and (3) stochastic events like drought(s) or flooding may necessitate changes in a monitoring plan. A higher frequency of sampling may be necessary until the spatial and temporal variations in data are characterized.