The role of monitored natural recovery in sediment remediation

Authors


Abstract

The long-term goal of monitored natural recovery (MNR) is to achieve ecological recovery of biological endpoints in order to protect human and ecological health. Insofar as ecological recovery is affected by surface-sediment-contaminant concentrations, the primary recovery processes for MNR are natural sediment burial and contaminant transformation and weathering to less toxic forms. This paper discusses the overall approach for effective implementation of MNR for contaminated sediment sites. Several lines of evidence that may be used to demonstrate natural recovery processes are summarized, including documentation of source control; evidence of contaminant burial; measurement of surface sediment mixing depths and the active sediment benthic layer; measurement of sediment stability; contaminant transformation and weathering; modeling sediment transport, contaminant transport, and ecological recovery; measuring ecological recovery and long-term risk reduction; knowledge of future plans for use and development of the site; and watershed and institutional controls. In general, some form of natural recovery is expected and should be included as part of a remedy at virtually all contaminated sediment sites. Further, MNR investigations and an understanding of natural recovery processes provide cost-effective information and support the evaluation of more aggressive remedies such as capping, dredging, and the use of novel amendments. The risk of dredging or capping may be greater than the risk of leaving sediments in place at sites where capping or dredging offer little long-term environmental gain but pose significant short-term risks for workers, local communities, and the environment.

EDITOR'S NOTE:

This paper is among 9 peer-reviewed papers published as part of a special series, Finding Achievable Risk Reduction Solutions for Contaminated Sediments. Portions of this paper were presented by the author at the Third International Conference on Remediation of Contaminated Sediments held in New Orleans, Louisiana, USA, in January 2005.

INTRODUCTION

Risk reduction is the long-term goal of contaminated sediment management. In aquatic environments affected by contaminated sediments, risk management strategies focus on either removing the contaminated material or interrupting exposure pathways by which contaminants pose an ecological or human health risk over time (pathway interdiction). This is generally achieved by dredging, isolation (capping), or monitored natural recovery (MNR). The implementation of a carefully planned MNR remedy at sediment sites can result in risk reduction either by pathway interdiction through natural processes such as contaminant burial and sequestration, or by mass removal via chemical weathering and biodegradation. Erosion, transport, and dispersion of particle-bound contaminants also can contribute to recovery, but threatens to increase contaminant loading and potential risks to downstream areas.

Monitored natural recovery involves leaving contaminated sediments in place and allowing ongoing aquatic sedimentary and biological processes to contain, destroy, or otherwise reduce the bioavailability of the contaminants in order to protect receptors (NRC 1997). A decision to implement MNR at a site does not represent a no-action decision for remediation; it is the result of a thoughtful decision-making process following careful site assessment and characterization (NRC 1997; NRC 2001; USEPA 2005). Because MNR relies on natural processes, it is likely to be most applicable to sites, or portions of sites, where human or ecological risks are not immediate or substantial. Site assessment should be designed to demonstrate that contaminants are sequestered or otherwise controlled to adequately reduce human health and ecological risks, using multiple lines of evidence to reduce uncertainty.

Several physical, chemical, and biological processes contribute to MNR (Magar 2001):

  • Contaminant burial: Natural deposition of increasingly clean sediments reduces the potential for contaminant exposure by gradual reductions in surface sediment concentrations over time (Brenner et al. 2001, 2004).

  • Reduced contaminant mobility: Sorption, metals precipitation, and other binding processes reduce contaminant mobility and bioavailability.

  • Chemical or biological transformation: Contaminants can be converted to less toxic forms through biotic and abiotic transformation processes (Stout et al. 2001; Magar, Brenner, et al. 2005; Magar, Johnson, et al. 2005).

  • Dispersion of particle-bound contaminants: Erosion, transport, and dispersion of particle-bound contaminants potentially benefit highly contaminated environments and lead to localized contaminant-concentration reductions; however, these same processes may increase contaminant loading and risk to downstream areas.

MNR investigations may be more costly than conventional site characterization and remedial investigations because MNR relies on multiple lines of evidence not always required for conventional investigations (e.g., geochronological age dating or detailed chemistry for chemical forensic analyses). Over the long term, however, MNR is cost effective because the information gained also supports the evaluation of more aggressive technologies including capping, dredging, and the use of novel amendments. As in the case of terrestrial groundwater contamination where natural attenuation processes affect the fate and transport of chemicals in all hydrologic systems (USEPA 1998), some form of natural recovery is likely at nearly all contaminated sediment sites. Therefore, MNR should be evaluated whether it is expected to be the primary remedy or a secondary remedy in combination with more active remedial approaches.

CURRENT REGULATORY CONSIDERATIONS

Monitored natural recovery is increasingly recognized by the US Environmental Protection Agency (USEPA) as an alternative for managing or remediating contaminated sediment sites (USEPA 2005). Most major environmental contamination problems rely on the use of multiple technologies and risk management strategies to treat, contain, and manage contaminants in the environment. In the United States, consistent with federal regulations, the USEPA “expects to use a combination of methods, as appropriate, to achieve protection of human health and the environment” (40 CFR §300.430[a][1][iii][C]).

When selecting a technology for site remediation, a primary goal of the technology screening and selection process is to optimize cleanup to achieve the greatest degree of protection at the most reasonable cost. Despite increasing recognition of MNR as a sediment management alternative (USEPA 2005), federal guidance is currently not available on how to evaluate MNR for risk management and integrate MNR with other remedial alternatives. Unfortunately, this has the potential to create uncertainty and inconsistency among practitioners about how to properly characterize contaminated sediment sites and optimize remediation by taking advantage of ongoing natural processes. If one assumes that some portion of contamination will remain in place, then it becomes increasingly necessary to determine where dredging/capping should end and where MNR should begin, at each site.

In the United States, the 1990 National Contingency Plan explicitly cites potential remedy implementation risk as an important consideration in the remedy selection process (40 CFR §300.430[f][1][ii][C][4]). Thus, to the extent that MNR can achieve remedial goals, provide long-term ecological protection, and reduce or even eliminate short-term risks imposed by more aggressive remedies, MNR is included among the remedial alternatives for most sediment sites. The US National Research Council (NRC) also recommends broadening the evaluation of remedy effectiveness to include consequences of remedy implementation, including indirect risk and consideration of “overall” or “net” risk reduction, in addition to specific risks (NRC 2001).

LINES OF EVIDENCE FOR NATURAL RECOVERY

As with virtually all facets of sediment management (e.g., characterization, risk assessment, and remediation), MNR relies on several lines of evidence to demonstrate sediment deposition and contaminant burial and, more importantly, long-term ecological recovery and risk reduction. Table 1 lists the different lines of evidence that should be considered in sediment investigations to support the implementation of MNR as a remedy option.

Documenting source control

Monitored natural recovery relies primarily on source control combined with natural sediment deposition and burial of contaminated sediments (Magar 2001; Brenner et al. 2004). Following source removal, natural sedimentation can result in the gradual recovery of surface sediments as deposited sediments become increasingly less contaminated with time (Magar 2001; Brenner et al. 2004; USEPA 2004). Source-control documentation relies primarily on historical records of former remedial investigations and cleanup efforts. If evidence of source control is unavailable, then it may be necessary to conduct a field investigation to demonstrate that source control has been achieved.

In addition to controlling primary sources, it is often necessary to distinguish “external” upland/watershed sources (e.g., outfalls or nonpoint sources) versus “internal” secondary sources associated with releases from legacy sediments (e.g., resuspension of contamination from historical releases). These distinctions may be particularly appropriate if the boundaries of the site under consideration are not well defined. Internal contaminant sources, including hotspots, can act as potential reservoirs that periodically release contaminants into the aquatic system and slow recovery processes. At some sites, it may be necessary to combine removal or capping of secondary sources with MNR.

The presence of multiple sources also can mask recovery. For example, urban runoff can be a consistent and ongoing source of petroleum hydrocarbons (PAHs) and may negatively impact surface sediment recovery, even after removal of a major point source (van Meter et al. 2000; Stout et al. 2001; USEPA 2004).

Evidence of contaminant burial and reduction of surface sediment concentrations

In net depositional environments, surface sediment concentration reductions rely on deposition of increasingly clean sediments over time, and corresponding contaminant burial. Historical contaminant releases and trends can be inferred from sediment core analyses (Brenner et al. 2001, 2004; USEPA 2004). Vertical contaminant profiles provide a temporal record of sediment deposition, contamination, and weathering. The approach to analyzing these vertical contaminant profiles involves collecting vertical sediment cores, segmenting the cores, analyzing segments for contaminants of concern, and age dating the sediments.

Coupling the vertical contaminant profile with radionuclide analyses (e.g., lead-210 [210Pb] and cesium-137 [137Cs]) provides information on the temporal variations of contaminant release and recovery, sediment accumulation rates, and surface mixing depths (van der Perk et al. 2002; Brenner et al. 2004). Radio-geochemistry analyses used to date sediment core segments have been combined with contaminant concentration profiles to estimate the time required to meet surface sediment target concentrations (Brenner et al. 2004) and may help reduce uncertainty in future predictions of MNR.

Table Table 1.. Lines of evidence that should be considered in sediment investigations and to support the use of monitored natural recovery at sediment sites
1. Documentation (and possibly confirmation) of source control.
2. Evidence of contaminant burial and reduction of surface sediment concentrations.
3. Measurement of surface sediment mixing to estimate the active sediment benthic layer and to determine the surface sediment depth to which remedial action objectives should be applied.
4. Measurement of sediment stability to assess the risk of contaminant resuspension under normal and high-energy events.
5. Evidence of contaminant transformation and risk attenuation.
6. Modeling of long-term recovery, including surface water, sediment, and biota.
7. Monitoring ecological recovery and long-term risk reduction.
8. Knowledge of future site use and institutional controls.

Work by others (van Metre and Callender 1997; van Metre et al. 1998, 2000; Brenner et al. 2004) demonstrates the efficacy of sediment core profiling to characterize natural recovery due to source control and natural sedimentation of increasingly clean sediments over time. For example, Figure 1 shows how sediment core profiles were used to establish vertical polychlorinated biphenyl (PCB) concentration profiles, age date sediments, and determine surface sedimentation rates and surface-sediment-contaminant reduction rates in cores collected in Lake Hartwell, South Carolina, USA (Brenner et al. 2004). PCB trends showed decreasing surface sediment concentrations starting in the late 1970s, when USEPA began to regulate the use of PCBs.

In lieu of coring, other methods may be used to monitor or characterize surface sediments and sedimentation rates. For example, surface sediment grab samples can be collected over time to monitor surface sediment recovery; however, spatial and temporal variability often overshadow short-term temporal trends and may require relatively large sample sets or long monitoring periods to achieve statistically representative changes in surface sediment concentrations (Connolly et al. 2005). Sediment traps can measure deposition and characterize recently deposited sediments; however, sediment traps also can be influenced by spatial and temporal variability, and it can be difficult to distinguish recently deposited and resuspended sediment fractions. Bathymetry monitoring identifies bed elevation changes with time; however, bathymetry measurements cannot always provide sufficient vertical resolution to confidently determine changes in sediment thickness, particularly those involving a few centimeters over short periods.

Reduction of contaminant concentrations in water and biota

Evaluating the effectiveness of MNR is no different from evaluating the effectiveness of any other sediment remedy (USEPA 2005). To support the proper evaluation of MNR, sufficient data must be available to demonstrate that potential risks posed by contaminants are alleviated to acceptable levels within an acceptable time frame. Most sites are expected to achieve sediment and surface water target concentrations that are protective of human health and the environment. These target concentrations typically are the basis for setting long-term sediment recovery goals. Ecological recovery is a more elusive and less predictable goal (particularly for bioaccumulative contaminants) than surface sediment recovery (Zarull et al. 2002; Chapman and Anderson 2005).

Figure Figure 1..

Surface sediment recovery in PCB-contaminated Lake Hartwell sediments (from Brenner et al. 2004).

In general, sediment restoration objectives fall into 1 of 3 categories: (1) reduction of contaminant conditions in sediment, (2) recovery of the sediment benthic community or specific species, or (3) protection of human health. Monitoring the recovery of aquatic species and populations, as well as the ecological community, requires an understanding of the nature of the impairments due to the contamination and the ability to distinguish site-related impairments from those caused by other sources (either anthropogenic or natural).

Biological effects most often attributed to contaminated sediments include increased contaminant body burdens, the absence or diminished presence of a biological community, and gross deformities and lesions (Wenning et al. 2005). Significant sources of uncertainty typically identified in ecological risk assessments and pertaining to the evaluation of biological effects from sediment contaminant exposure include, but are not limited to, factors such as size, age, sex, reproductive state, and behavior (i.e., competition for food) of receptor individuals; the percentage of time receptor species or their prey forage at the site; the diet of the receptor individual; and both predator and prey population dynamics such as migratory behavior (Linkov et al. 2002; von Stackelberg et al. 2002; Vorhees et al. 2002). Therefore, MNR generally must rely on long-term monitoring to demonstrate biological recovery years or decades after surface sediment recovery.

For human health protection, sediment remediation goals are often tied to achieving water quality goals or contaminant levels in edible fish tissues that are protective of anglers and support recreational uses of the site. A challenge in both risk assessment and monitoring the effectiveness of MNR is the difficulty in quantifying the links between sediment concentrations, ecological receptors of concern, and population- or community-level conditions.

Measurement of surface sediment mixing and the active benthic layer

Measurement of the active benthic layer (i.e., the surface sediment layer primarily influenced by mixing due to bioturbation from benthic organisms) is often necessary to predict long-term surface-sediment-contaminant concentration reductions (Thibodeaux 1996; Fitzgerald et al. 2001). Surface sediment mixing due to bioturbation and hydraulic disturbances, as well as depositional processes, control the rate of contaminant reduction to background levels, and determine whether years or decades are required for surface sediments to meet risk-based benchmark concentrations. Thus, even when source control is achieved, surface sediment recovery tends to be gradual due to surface sediment mixing processes. The active benthic layer also defines the surface sediment depth to which sediment cleanup levels should be applied. Common tools for measuring the active benthic layer include beryllium-7 (7Be) profiles (Fitzgerald et al. 2001), oxic/anoxic profiles, and sediment profile imaging. When measuring benthic activity and the depth of the active benthic layer, it is important to recognize and (if possible) measure the spatial variability of the benthic community and any bioturbation.

Measurement of sediment stability and contaminant resuspension potential.

Because contaminants are left in place with MNR, the risks of contaminant breakthrough or resuspension persist. Therefore, sediment stability is often an important component of MNR. Sediment erosion potential should be investigated to ensure that sediments are stable during normal to high-energy conditions, including storms, flood events, other natural events, and human disturbances, especially ship wake, which can be significant near shipping channels and in ports and harbors.

Critical shear stress measurements in sediment cores are used to create a vertical profile of erosion potential with sediment depth. Surface water hydrodynamic studies are used to measure bottom shear stresses from wave-induced and current-induced forces. The critical shear stress measurements and hydrodynamic bottom shear stress measurements can be used to determine the erosion potential and estimate a depth of scour during normal or high-energy events. Hydrodynamic data combined with sediment suspension data can then be used to determine contaminant flux and transport in the water column. When conducting hydrodynamic studies, it is important to measure across a range of hydrodynamic conditions, including storms, high flows, high winds, and other high-energy events. However, because such studies rarely are conducted during extreme events, usually it is necessary to estimate bottom shear stresses during extreme events using models.

Erosion rates vary with applied shear stress due to waves and currents and as a function of sediment bulk properties (Jones and Lick 2000). Measurable sediment bulk properties include particle size distribution, moisture content, organic carbon, and bulk density. Though these measurements can predict critical shear stress values of noncohesive sediments, they cannot directly predict critical shear stress for cohesive sediments (Jones and Lick 2000). Therefore, direct measurements of sediment shear strength may be necessary for more cohesive sediments. Sandy sediments tend to be noncohesive, whereas silts and clays are cohesive. Because sediment contaminants typically are associated with fine-grained cohesive sediments, site-specific critical shear stress measurements may be required to determine potential scour depths under high-flow events.

Critical shear stress can be measured directly using in situ or ex situ erosion measurement devices. In situ measurement devices include the Sea Carousel (Maa et al. 1995), or Inverted Flume (Ravens and Gschwend 1999); ex situ devices include the SED flume (McNeil et al. 1996; Jepsen et al. 1997; Roberts et al. 1998), ASSET flume (Roberts et al. 2003), SEAWOLF Flume (Jepsen et al. 2002), and Shaker/PES Flume (Tsai and Lick 1986). Differences among these flumes include whether they are field or laboratory deployed, in situ or ex situ, the applicable shear stress range (some are better at low shear stresses and others at high shear stresses), and the depth to which measurements can be made (some measure shear stresses at the sediment surface only, while others measure shear stress with depth).

Evidence of contaminant transformation and risk attenuation

Monitored natural recovery also involves consideration of long-term contaminant weathering and transformation processes. However, because weathering tends to be relatively slow compared to deposition and contaminant burial, natural capping typically provides more immediate risk reduction. Contaminant weathering in sediments consists of dilution, volatilization, chemical transformation, biotransformation and biodegradation, and sorption. Although slow, these processes contribute to the permanent reduction of contaminant concentrations and bioavailability.

Chemical transformation processes are different for different types of contaminants, even within a contaminant group, such as PAHs. For example, several PAHs have been shown to degrade aerobically, including naphthalene, acenaphthene, fluorene, dibenzothiophene, anthracene, phenanthrene, fluoranthene, pyrene, chrysene, benzo[a]anthracene, and benzo[a]pyrene (Prince and Drake 1999). Anaerobically, PAHs and other petroleum hydrocarbons are slower and more difficult to degrade. Naphthalene and acenaphthene have been shown to biodegrade under nitrate-reducing conditions (Milhelcic and Luthy 1991; Durant et al. 1995) and naphthalene and phenanthrene have been shown to biodegrade under sulfate-reducing conditions (Coates et al. 1996, 1997). The extent to which weathering facilitates natural recovery must be evaluated with regard to site-specific, chemical-specific, and sediment-specific weathering properties.

Hydrocarbon weathering is generally characterized by the progressive decrease in the percentage of low-molecular-weight hydrocarbons and the corresponding accumulation of high molecular-weight hydrocarbons (Brenner et al. 2001; Stout et al. 2001). This weathering pattern has positive short-term benefits, by reducing or eliminating acute exposures to lower molecular-weight compounds (e.g., naphthalene and methylnaphthalenes). However, high-molecular-weight compounds tend to persist over the long term.

The primary factors affecting PCB biotransformation are the number and pattern of chlorine substituents and the redox state of the medium. Aerobically, PCBs behave similarly to hydrocarbons and PAHs, where low-molecular-weight (i.e., less chlorinated) PCBs (≤3 chlorines) are preferentially weathered and degraded. The ability for bacteria to degrade PCBs decreases with increased chlorination; congeners with 5 or more chlorines are relatively recalcitrant to aerobic biodegradation (Furukawa et al. 1983; Bedard et al. 1986). Under anaerobic conditions, the primary metabolic pathway is reductive dechlorination, in which chlorine removal and substitution with hydrogen by bacteria results in a reduced organic compound with fewer chlorines (Mohn and Tiedje 1992; Bedard and Quensen 1995). In addition to lowering the overall toxicity of PCB-contaminated materials, dechlorination also decreases the tendency of the PCB mixture to bioaccumulate (Abramowicz and Olson 1995).

The positive impact of dechlorination (e.g., reduced mass and reduced toxicity) is tempered by the fact that dechlorination increases with sediment depth and age (Magar, Brenner, et al. 2005; Magar, Johnson, et al. 2005). Surface sediments have had the least amount of time to dechlorinate, yet they pose the greatest risk of environmental exposure. Consequently, dechlorination should be evaluated for long-term detoxification of PCB-contaminated buried sediments, whereas more immediate recovery must be provided by sediment burial.

The potential mobility and toxicity of metals and organometals, and the various site-specific features that influence the transport and behavior of metals, make it uniquely challenging to assess the efficacy of MNR for metals-contaminated sediments. The total concentration of metals in sediment typically is not a strong indicator of sediment toxicity, because of the effect of sediment chemistry on metals speciation and bioavailability (Berry et al. 1999, 2004; Boothman et al. 2001; Apitz et al. 2005). Redox chemistry and pH strongly influence the solubility and availability of metals, particularly those with multiple valence states. For some metals (e.g., chromium [Cr]), solubility, mobility, and toxicity decrease under anaerobic/low-redox conditions and in the presence of acid-volatile sulfide (AVS). The oxidized form of Cr (hexavalent Cr) is highly toxic, whereas its reduced form (trivalent Cr) is virtually nontoxic (DeLaune et al. 1998; Berry et al. 2004). The fact that sediments are commonly organic-rich and anaerobic contributes to the rapid detoxification, stabilization, and subsequent natural recovery of Cr-contaminated sediments.

For other cationic metals, including silver, cadmium, copper, nickel, lead, and zinc, AVS is a useful predictor of toxic effects in sediments (Ankley et al. 1996; Berry et al. 1999; Boothman et al. 2001). AVS is formed in sulfate-rich environments and in the presence of organic carbon, which serves as the electron donor for the biochemical reduction of sulfate to sulfide. Under anaerobic conditions and in the presence of AVS, cationic metals can form metal-sulfide precipitates. Most metal-sulfide complexes have low aqueous solubilities and some tend to resist dissolution even if later exposed to oxidizing conditions, thus permanently reducing their mobility and bioavailability. Seawater's high natural concentration of sulfur means that marine sediments are often rich in reduced sulfur. Consequently, most metals in reduced marine sediments are relatively stable as metal-sulfides.

Microorganisms can catalyze the methylation of metals and metalloids, thereby increasing mobility and resulting in significantly greater toxicity. Anaerobic sediments with high organic content and low sulfur content tend to favor methylation (Rassmussen et al. 1998). For example, organomercury species are best known for their ability to bioconcentrate; as mercury (Hg) bioaccumulates, the total mercury concentration may decrease, but the percent total mercury as methylmercury typically increases (Mikac et al. 1985). The potential methylation of Hg can compromise MNR, unless it can be demonstrated that Hg-contaminated sediments are buried and sequestered, and that the Hg is stable in an inorganic form (e.g., as HgS).

Modeling long-term site recovery

When used in the context of a sediment remedy, MNR relies on assessment, modeling, and long-term monitoring to verify recovery. After empirically demonstrating surface sediment recovery trends, sediment and contaminant transport models can be applied to estimate further long-term recovery and to evaluate long-term sediment resuspension potential. Models integrate rate-limiting sorption processes, sedimentation, sediment transport via hydrodynamic suspension, ongoing primary or secondary sources, and benthic mixing to better understand long-term behavior and transport of contaminant into the aquatic environment and to model sediment, water, and biota recovery. Long-term ecological models can be linked to chemistry and sediment models to predict ecological recovery over time. For bioaccumulative chemicals, ecological recovery may include reduced biological tissue concentrations in the food chain, whereas for non-bioaccumulative chemicals recovery will likely focus on benthic community recovery and acute toxicity reduction.

Figure Figure 2..

Vertical, 1-D contaminant transport model for PCBs in sediments. Kp = 1.2 × 105 L/kg; Dm = 10−7 cm2/s; Db = 10−7 cm2/s (left) and 10−5 cm2/s (right); Deposition = 0.5 cm/y (developed by Dr. Craig Jones; Sea Engineering; Santa Cruz, CA, USA).

Simple 1-dimensional models (e.g., Figure 2) can provide information on surface sediment recovery and chemical concentration-reduction rates; such models readily integrate sediment deposition, surface sediment mixing, and chemical transport and transformation processes. Figure 2 shows hypothetical model results that simulate diffusion, mixing, and deposition processes for PCB-contaminated sediments, over a 100-y period with 2 bioturbation rates (developed by C. Jones, Sea Engineering, Inc.; Santa Cruz, CA, USA). The model includes contaminant partitioning (Kp = 1.2 × 105 L/kg), diffusion (Dm = 10−7 cm2/s), the deposition rate (deposition = 0.5 cm/y), and the effective diffusion rate due to bioturbation (Db = 10−5 and 10−7 cm2/s); Db in this model decays with depth. Contaminant transport is dominated by deposition and surface mixing. Deposition drives the peak contaminant concentrations downward resulting in an additional 52-cm burial at 0.5 cm/y (the peak migrated from 23 cm to 75 cm below the sediment-water interface). Figure 2 demonstrates the effect of mixing on recovery; mixing results in the gradual reduction of surface sediment concentrations and the asymptotic approach to background levels at the sediment surface.

For larger, more dynamic sites, more complex models may be necessary to incorporate the combined effects of changes in contaminant loadings, biological and chemical degradation, and natural transport and mixing processes that occur over a wide range of spatial and temporal scales. They may be used to predict long-term reductions in exposure, or short-term changes in response to meteorological (e.g., storms) or anthropogenic (e.g., dam removal) events (Dekker et al. 2003). The appropriate modeling complexity will ultimately depend on the available data, project resources, site complexity, and the scope of decisions being made. Uncertainty exists when making future predictions based on past data trends, especially without a mechanistic understanding or verifiable model of contaminant fate processes. Site investigations should be designed to reduce uncertainty with multiple lines of evidence, using a well-constrained model that addresses major sources of uncertainty, and collecting site-specific data to limit or resolve uncertainty.

Monitoring ecological recovery and long-term risk reduction

Mounting evidence of ecological and human health risks associated with contaminated sediments requires careful management of contaminated sediments to control risks to human health and the environment. If contaminants were uniformly bioavailable and uniformly toxic, then contaminated sediments could be regulated, ranked, and managed based upon bulk sediment concentrations, in a manner similar to many water management programs (Apitz et al. 2005). However, this is not the case, and various factors already discussed can mitigate or enhance contaminant availability and toxicity in sediments. Consequently, sediments are evaluated based not only upon bulk concentration, which reflects potential exposure, but on ecological and toxic effects (Apitz et al. 2005).

Measurement of ecological recovery over time can be much more challenging than measuring sediment recovery over time, because ecological recovery tends to be slower and more difficult to quantify. For example, for bioaccumulative chemicals, the relatively long life spans of some fish and their ability to forage over relatively large areas can make it difficult to quantify ecological recovery using fish data. Notably, ecological recovery where contaminants are not significantly bioaccumulative tends to be easier to monitor, because the focus is on acute toxicity reduction.

Increases or decreases in higher-trophic-level animals may not necessarily indicate changes in surface sediment chemistry. Instead, they may indicate changes in exposures mediated through diet (Herbert et al. 2000; Fisk et al. 2001) or sources and environmental loadings not related to sediments. In other words, decreases in environmental loadings due to natural recovery may not be immediately reflected in decreased biological concentration trends, if these trends are overshadowed by other influences such as dietary loading or species age, sex, or size.

The measurement of contaminant concentration changes in biological endpoints is one of the greatest challenges to the implementation, regulatory acceptance, and public acceptance of MNR. The numerous confounding factors that affect ecological recovery of bioaccumulative chemicals make it difficult and sometimes impossible to measure statistically significant concentration changes in biological endpoints. Alternative techniques may better measure short-term ecological responses to contaminant concentration reductions in sediment. Such techniques include caged fish or clam deployments, collection of lower-trophic species (e.g., phytoplankton or macroinvertebrates), and short-term deployment of semipermeable membrane devices.

Knowledge of future site use and institutional controls

In some situations, it is possible that the implementation of natural recovery should be accompanied by institutional controls to ensure protection of public health. Institutional controls such as fish-consumption advisories, restricted access to contaminated environments, and restrictions on construction and development activities, however, tend to be difficult to implement and manage effectively. Even when institutional controls are effective, maintenance of these controls (including fish consumption advisories, property deed restrictions, and construction and dredging restrictions) over extended periods may be problematic.

Changes in future site use and potential changes to the watershed also must be considered. For example, recreational or industrial site uses could result in navigational dredging requirements that impact natural recovery. Critical changes to the watershed, such as increased urbanization or changes in agricultural practices, also can impact nonpoint source contaminant loadings and natural sediment loads. For example, Jaffe et al. (1998) report that from 1951 to 1983, much of San Francisco Bay, California, USA changed from being depositional to erosional as the sediment supply diminished and currents and waves continued to remove sediment from the bay. The decreased sediment supply was attributed to upstream flood-control and water-distribution projects that reduced peak flows, which were responsible for sediment transport in the bay.

Efforts to protect wetlands, shoreline habitat, and aquatic habitat increasingly restrict filling estuarine environments and changing existing bathymetric and shoreline contours. The San Francisco Bay Plan (BCDC 2003), for example, restricts filling of San Francisco Bay. The flexibility to dredge or cap in protected environments may be limited by such protective measures, and may require greater reliance on MNR, long-term monitoring, and deed restrictions, particularly for low-risk sites where the risk of remedy is greater than the need to protect existing natural resources.

CONCLUSIONS

This paper discusses the overall approach for effective implementation of MNR for contaminated sediment sites, and it presents lines of evidence that may be used to demonstrate natural recovery processes. The long-term goal of MNR is to achieve recovery of biological endpoints in order to protect human and ecological health. Ecological recovery is affected by surface-sediment-contaminant concentration reductions over time via natural sediment burial, and contaminant transformation and weathering.

The current state of the practice in sediment management relies on surface sediment recovery goals; that is, remediation usually targets surface-sediment chemical concentration endpoints and not long-term biological endpoints. The primary reason for this is practical: where contaminants are bioaccumulative, biological trends take decades to monitor and observe. Target surface-sediment chemical concentrations are more readily achievable, whether by MNR, capping, or dredging.

In the United States and elsewhere, regulatory guidance is needed to provide clear direction to remedial project managers, regulators, and remediation experts for implementing natural recovery. Such guidance should discuss how to assess and validate MNR and provide requirements for monitoring after MNR is implemented. Guidance documents also should recognize the site-specific factors that can influence the rate and extent of natural recovery processes, so that flexibility is afforded to practitioners. Uncertainty can be minimized by relying on multiple lines of evidence to demonstrate sediment deposition, contaminant burial, sediment stability, contaminant weathering, ecological recovery, and risk reduction.

Finally, more case studies demonstrating sediment and ecological recovery under natural conditions are needed. Although the fundamental contaminant transport mechanisms in sediments (e.g., sorption, sequestration, dissolution, volatilization, bioturbation, and biotransformation) are known, the extent to which these processes control contaminant migration in the natural environment, particularly in biological food webs, remains difficult to measure under in situ conditions. Engineering and environmental science communities are encouraged to publish results of sites where such evidence exists and such processes have been measured.

The limited resources available for sediment remediation require that MNR be evaluated and considered at all sediment sites, especially at low-risk sites. The impulse to remove all contaminated sediments from the environment by dredging, while understandable, does not always reflect proper environmental stewardship and risk management. MNR is most appropriate where chemicals in sediments will be effectively buried or transformed in an acceptable time frame and where dredging or capping will have greater negative impacts than leaving contaminants in place.

Acknowledgements

The authors acknowledge the helpful comments from anonymous reviewers and the coordinating efforts of the conference organizers at Battelle, Patricia White and Robert Olfenbuttel. This work and participation at the conference was funded entirely by the authors. Portions of this paper also were presented by Magar (2004) at the Contaminated Sediments Workshop sponsored by the Strategic Environmental Remediation and Development Program and the Environmental Security Technology Certification Program.

Disclaimer—The peer-review process for this article was managed by the Editorial Board without the involvement of its Editor-in-Chief, who appears as an author in this article.

Ancillary