What's causing toxicity in sediments? Results of 20 years of toxicity identification and evaluations

Authors

  • Kay T. Ho,

    Corresponding author
    • National Health and Environmental Effects Research Laboratory, Atlantic Ecology Division, Office of Research and Development, US Environmental Protection Agency, Narragansett, Rhode Island, USA
    Search for more papers by this author
  • Robert M. Burgess

    1. National Health and Environmental Effects Research Laboratory, Atlantic Ecology Division, Office of Research and Development, US Environmental Protection Agency, Narragansett, Rhode Island, USA
    Search for more papers by this author

Errata

This article is corrected by:

  1. Errata: Corrigendum Volume 33, Issue 3, 718, Article first published online: 14 February 2014

Address correspondence to ho.kay@epa.gov.

Abstract

Sediment toxicity identification and evaluation (TIE) methods have been used for 20 yr to identify the causes of toxicity in sediments around the world. In the present study, the authors summarize and categorize results of 36 peer-reviewed TIE studies (67 sediments) into nonionic organic, cationic, ammonia, and “other” toxicant groups. Results are then further categorized according to whether the study was performed in freshwater or marine sediments and whether the study was performed using whole-sediment or interstitial-water TIE methods. When all studies were grouped, nonionic organic toxicants, either singly or in combination with other toxicants, were implicated in 70% of all studies. When studies were divided into interstitial-water TIE methodology compared with whole-sediment TIE methodology, results indicated that studies performed using interstitial-water TIE methods reported nonionic organic toxicity slightly more often than toxicity from cationic metals (67% compared with 49%). In contrast, studies using whole-sediment TIE methods report nonionic organic chemical toxicity, either singly or in combination with another toxicant, in 90% of all sediments tested. Cationic metals play a much smaller role in whole-sediment TIE studies—fewer than 20% of all sediments had a metals signal. The discrepancy between the 2 methods can be attributed to exposure differences. Contrary to earlier findings, ammonia generally plays only a minor role in sediment toxicity. Environ Toxicol Chem 2013;32:2424–2432. © 2013 SETAC. This article is a US Government work and is in the public domain in the USA.

INTRODUCTION

Globally, sediments are a major sink for many types of anthropogenic contaminants [1-7]. When sufficiently elevated, these contaminants can cause toxicological effects to benthic and epibenthic organisms. For example, Long et al. [8] estimated that large sections of US estuarine sediments were acutely toxic to benthic organisms. In addition to toxicity, severely contaminated sediments can cause ecological [9-11] and economic [12] damage to aquatic resources around the globe. These damages include declines in important sport and commercial fisheries, loss of recreational areas, degradation of habitat for valued species, and costly remediation and disposal actions. Knowledge of the presence and magnitude of sediment toxicity is important, but identification of the toxicants causing the adverse effects is critical for effective selection of contaminated sediments management options. These options may include choosing appropriate remediation techniques; evaluating disposal options for dredged material; assessing important stressors in impaired benthic communities; developing stressor–response relationships for ecological risk assessments; or identifying sediment-associated stressors and, ultimately, stressor sources for total maximum daily loading (TMDL) actions.

For all of these reasons, methods and approaches have been developed to identify active toxicants in contaminated sediments. The approach called sediment toxicity identification and evaluation (TIE) consists of both whole-organism toxicity testing and chemical manipulations used iteratively and in parallel until classes of toxicants or specific toxicants are identified [13]. The TIE approach was first developed in North America in the 1980s for use with municipal and industrial effluents as part of the US Clean Water Act effort to remove toxic chemicals from effluents discharged to the environment [14-18]. In the 1990s, TIE methods were modified for use with sediment interstitial waters [19]; and by the mid-2000s, methods were also available for whole sediments [17]. The TIE manipulations are conceptually used in a three-phase approach, in which phase I characterizes the toxicant into classes (i.e., cationic metals—such as cadmium, copper, nickel, lead, zinc—or nonionic organic chemicals), phase II identifies the specific toxicants (e.g., copper, pyrethroids), and phase III confirms the findings of phases I and II.

Sediment TIEs have been successfully performed in many countries around the world over the last 20 yr [20-28]. The present study builds on a 2002 article [29] that analyzed the available set of data at the time (n = 18). In the present study, data from more than 67 sediments have been included. Trends that appeared in the earlier, smaller study have been corrected, and new trends were identified. These TIEs have identified a variety of toxicants, including nonionic organic chemicals, cationic metals, ammonia, sulfides, ionic imbalance, pH, and mixtures of these toxicants. Most of these TIEs were phase I characterizations. The 3 most frequently characterized classes of toxicants were nonionic organics, cationic metals, and ammonia. The present article reviews the findings of these studies and determines the major causes of toxicity in field sediments as identified by interstitial-water and whole-sediment TIE methods in freshwater and marine systems. Based on the results of these studies, we can summarize the major causes of sediment toxicity and determine whether most toxicity was caused by a single class of toxicant or by mixtures.

MATERIALS AND METHODS

Studies were grouped by toxicant into categories of nonionic organic chemicals, cationic metals, ammonia, and mixtures of these 3 toxicants with hydrogen sulfide, ionic imbalance, and pH. Results were separated into studies performed in interstitial waters or whole sediments and then further subdivided by freshwater and marine systems. To determine the probability of a specific class of toxicants causing toxicity, we summed the percentage of occurrences among the different classes. The number of TIEs performed on whole sediments and those performed on interstitial waters does not sum to 67 because 15 sediments had both interstitial-water and whole-sediment TIEs performed and were reported separately in 2 sections. When calculating the results of all sediments, we used the whole-sediment results because we believe the whole-sediment exposures are more environmentally realistic. Because the objective of the present analysis is to perform a survey of field sediments (not journal articles), when studies looked at more than 1 site, each site was counted separately. Conversely, if more than 1 published study considered the same site, it was counted as only 1 site.

The main criterion for inclusion into the present study of peer-reviewed articles was that the study must have included methods designed to characterize at least 3 major groups of contaminants in sediments: nonionic organic chemicals, metals, and ammonia. Studies may have included more specific methods in addition to those 3 but must have looked at least for these 3 broad categories. Targeted TIEs [30-32] designed to look for a specific class of chemicals were not included, because they may fail to identify broader classes of possible toxicants. For example, many of the targeted TIEs were designed to identify pyrethroid, organophosphate, or carbamate pesticides. These methods are effective at detecting these compounds; however, when used in isolation, they may not give a complete picture of other causes of sediment toxicity (i.e., the role of cationic metals, ammonia). For example, when farmers apply pesticides, they often apply fertilizers or fungicides as well (http://www.hort.purdue.edu/fruitveg/2012ID168.pdf), which may result in elevated ammonia or metal concentrations that would not be detected by methods designed exclusively for pyrethroid, carbamate, or organophosphate pesticides. In addition, studies that used frozen interstitial water for conducting interstitial-water TIEs were not included. Freezing interstitial waters can cause myriad chemical changes to the water and its constituents, with unpredictable effects on toxicity [1]. Finally, TIEs performed on farm ponds or holding areas for agricultural field runoff [33] were not included, because the sediments appeared to be highly artificial and anthropogenically created, which could bias the interpretation of the present study of contaminants that cause toxicity in natural sediments. In addition, these sediments are not subjected to state or federal rules under the Clean Water Act.

RESULTS AND DISCUSSION

Overall

In the peer-reviewed literature, TIE results for 67 sediments have been reported (Figure 1 and Table 1; Supplemental Data, Table S1). Among all sediments tested, more than half (63%) had toxicity from a single class of toxicants. Toxicity from nonionic organic chemicals played the largest role in causing sediment toxicity. Approximately 42% of all TIE studies indicated nonionic organic chemicals as the sole source of toxicity, and another 28% reported that this class of toxicants caused toxicity within a mixture of other toxicants; therefore, nonionic organic chemicals were implicated as contributing to toxicity in 70% of all sediments tested. Cationic metals were the next largest single category of causal toxicant at 16%, with metals in a mixture with other toxicants at 19%, for a total of 35% of all sediments that had metals contributing to toxicity. Fewer than 5% of all sediment TIE studies indicated ammonia as a single class of toxicant, and approximately 21% of all TIEs reported ammonia contributing to toxicity in combination with another toxicant. Somewhat surprisingly, unlike what was reported in earlier studies [29], ammonia did not play a large role in causing toxicity; only approximately 26% of all TIE results indicated that ammonia, either singly or in combination with another toxicant, was responsible for causing toxicity. Among all TIE studies performed, only 1 sediment (∼2%) was unable to be characterized, and another 3% of the studies reported unknown toxicants combined in a mixture of known toxicants.

Figure 1.

Summary of the findings of all peer-reviewed interstitial-water and whole-sediment freshwater and marine toxicity identification evaluations (n = 67). Toxicants are categorized as nonionic organic chemicals (organics), cationic metals (metals), ammonia, hydrogen sulfide (sulfides), ionic imbalance, and unknown. Values in the figure represent the number of sediments actually tested.

Table 1. Studies grouped by the type of toxicant reported and the matrix in which the toxicity identification and evaluation (TIE) was performeda
ToxicantWhole sedimentInterstitial water
MarineFreshwaterMarineFreshwater
  1. aSee Reference list for full citation.
Organics[13, 23, 29, 40, 69, 70][13, 24, 36, 39, 42, 69, 71, 72][13, 22, 26, 73-77][13, 20, 24, 28, 36, 38, 39, 78-83]
Metals[69][69, 71][25, 76, 77, 84, 85][13, 20, 68, 78, 79, 86-88]
Ammonia [69][25-27, 74, 77, 89][13, 27, 78, 79, 82, 86, 88, 90]

Interstitial-water TIEs

Figure 2a shows the causes of toxicity in both freshwater and marine sediments as identified by interstitial water TIEs. Fifty-two interstitial-water TIEs were performed, and more than half of all the sediments analyzed in these TIEs had a single toxicant responsible for toxicity. Nonionic organic chemicals and cationic metals were the 2 largest categories. Approximately 65% of all of the interstitial-water TIE results showed a nonionic organic toxicant signal, approximately 46% of the sediments had a cationic metal signal, and 31% of the sediments had an ammonia signal either singly or in combination with other toxicants.

Figure 2.

Summary of the findings of peer-reviewed interstitial-water freshwater and marine toxicity identification evaluations (TIEs; n = 52). Toxicants are categorized as nonionic organic chemicals (organics), cationic metals (metals), ammonia, hydrogen sulfide (sulfides), ionic imbalance, and unknown. Values in the figure represent the number of sediments actually tested. (a) All interstitial-water TIEs; (b) freshwater interstitial-water TIEs; (c) marine interstitial-water TIEs.

For freshwater sediments analyzed by interstitial-water TIE methods, approximately 76% of all sediments had a nonionic organic chemical, 57% had a cationic metal signal, and 24% had an ammonia signal either singly or in some combination (Figure 2b). Marine interstitial-water TIE results indicated that approximately 61% of sediments had a nonionic organic chemical signal, whereas 39% had a cationic metal signal, and 35% had an ammonia signal either singly or in combination with other toxicants (Figure 2c).

In summary, the interstitial water TIE results indicated that nonionic organic chemicals played a larger role in sediment toxicity than cationic metals, with ammonia contributing to nearly one-third of the toxicity. Freshwater systems appeared to have a larger cationic metal signal (57%) compared with marine systems (39%). Interestingly, our earlier analysis of interstitial-water TIEs [29], performed with a smaller sample (n = 5), showed a lack of cationic metal toxicity in marine sediments, which is not supported in the present study of a larger data set.

Whole-sediment TIEs

The 30 whole-sediment TIEs revealed a simpler pattern of toxicity (Figure 3). Ninety percent of all whole-sediment TIEs reported nonionic organic chemical toxicity either singly or in combination with another toxicant. Cationic metals and ammonia played a much smaller role in whole-sediment TIEs—17% and 7%, respectively. When results were divided by freshwater (n = 18) or marine (n = 12) whole-sediment TIEs, the same trends emerged. Approximately 90% from each displayed a nonionic organic chemical signal, and only 8% of the marine and 22% of the freshwater sediments had a cationic metals signal either singly or in combination with other toxicants. It is interesting to note that ammonia was never identified as a toxicant causing marine whole-sediment toxicity and was identified as a toxicant acting in concert with nonionic organic chemicals in only 11% of the sediments (2 sediments) in freshwater whole-sediment TIEs.

Figure 3.

Summary of the findings of peer-reviewed whole-sediment freshwater and marine toxicity identification evaluations (TIEs; n = 30). Toxicants are categorized as nonionic organic chemicals (organics), cationic metals (metals), ammonia, hydrogen sulfide (sulfides), ionic imbalance, and unknown. Values in the figure represent the number of sediments actually tested. (a) All whole-sediment TIEs; (b) freshwater whole-sediment TIEs; and (c) marine whole-sediment TIEs.

Differences between whole-sediment and interstitial-water TIEs

The most obvious difference between the interstitial-water and whole-sediment TIEs is that nonionic organic chemical toxicity plays a much larger role in whole-sediment TIEs (90% of all toxicity reported resulted from a nonionic organic chemical signal) compared with interstitial-water TIES (65% of all toxicity reported included nonionic organic chemicals). Cationic metals contributed to 46% of the observed toxicity in interstitial-water TIEs; but in whole-sediment TIEs, metals contributed to toxicity in only 17% of the sediments. Ammonia was found to act only in combination with other toxicants (7%), not singly, in whole-sediment TIEs, whereas in interstitial waters, ammonia was found as a single toxicant 6% of the time and contributed to toxicity in mixtures (31%). These differences most likely are a result of the exposure methodologies used in interstitial-water and whole-sediment TIEs. For example, because of the difficulty of obtaining large volumes of interstitial water for performing TIEs, the exposures are often performed in small volumes (10–20 mL). Therefore, the possibility is greater that nonionic organic chemicals, particularly those with high octanol–water partition coefficient (log KOW) values, may adsorb to the surfaces of exposure containers. Once the contaminants are adsorbed to the surfaces, the environmentally relevant equilibrium concentrations of nonionic organic chemicals cannot be re-established because the sediments are not present in interstitial-water tests to replace adsorbed chemical; therefore, nonionic organic chemical toxicity is very likely to be underestimated in an interstitial water exposure relative to whole sediments. Also, depending on the organism being tested, interstitial-water TIEs tend to expose the organisms to more toxicants than in a whole-sediment exposure. For example, in a whole-sediment exposure, the tube building amphipod Ampelisca abdita irrigates its tube with cleaner overlying water [34] and therefore has less exposure to more contaminated interstitial water. When these organisms are placed in 100% interstitial water, their exposure to all contaminants, especially the water-soluble contaminants such as ammonia or metals, may be exaggerated relative to a whole-sediment exposure [35].

Interstitial-water TIEs appear to have identified many more types and combinations of toxicity, including pH, sulfides, and chlorides. This may be a result of the history of TIEs—interstitial methods were developed first, allowing researchers to work with them longer and develope more varied methods that identify different types of toxicants. In addition, the results might occur because water is an easier matrix to chemically manipulate and extract than are whole sediments. On the other hand, whole sediments have a great deal more buffering and binding capacity than water, and many toxicants may be bound and not bioavailable as they are in interstitial-water exposures.

Several researchers have performed both interstitial-water and whole-sediment TIEs on the same sediments [13, 24, 29, 36]. These 15 sediments were included in both the whole-sediment and interstitial-water results in the present study. In 5 sediments, differences between whole-sediment and interstitial-water results were found [13, 24, 36]. In 4 cases, metal and nonionic organic compound toxicity was identified in the interstitial-water TIEs, and only nonionic organic compound toxicity was identified in the whole-sediment TIEs. In 1 sediment, researchers were unable to determine the cause of toxicity in whole sediments but determined the cause of toxicity (pyrethroid pesticides and metals) in interstitial waters [24]. Although this sample size is too small to make generalizations about these findings, operational and mechanistic differences as discussed above support differing results between the 2 methods when used on the same sediment.

Differences between freshwater and marine sediments

Sulfides acting in concert with other toxicants were found to play a part in approximately 6% of marine interstitial-water TIEs and never in marine whole-sediment TIEs. Given the widespread occurrence of sulfides in marine sediments [37], the low percentage of reported incidences of sulfide toxicity is a bit unusual. It may be a methodological artifact, because toxic sulfides (e.g., H2S) may oxidize to less toxic forms (e.g., SO4−2) in the preparation of interstitial-water testing performed in an oxic atmosphere, or they may be volatilized by aeration during the 24-hr equilibration step in whole-sediment TIEs.

In freshwater interstitial-water TIEs, pH was indicated as a toxicant in only 1 sediment, but it was never indicated in marine sediments. This most likely was due to the strong CO2 buffering capacity in marine systems, which makes substantial changes in seawater pH unlikely.

Causes of nonionic organic toxicity

Among the 44 sediments in which nonionic organic compounds contributed to toxicity, only 11 of the studies identified specific compounds using TIE methods [13, 29, 36, 38-42]. In 9 of the 11 cases, the organic toxicants identified were pyrethroid and organophosphate pesticides. These sediments were found in the agricultural and urban areas of California. In 2 cases, polycyclic aromatic hydrocarbons [40] and polychlorinated biphenyls [41] were identified as the specific toxicants. In many other cases, the contaminant mixture was too complex, methods to identify the organic compounds did not exist, or researchers performed only phase 1 manipulations. We discuss the need for improved phase II methods to identify specific organic compounds responsible for toxicity in the Future research section.

Limitations of TIE

Although it may appear somewhat obvious, it is reasonable and accurate to state that TIE methods can identify only toxic chemicals. Detrimental changes in benthic community composition may be due to toxic chemicals and/or other stressors such as eutrophication, invasive species, changes in hydrological regime, navigational dredging, or changes in sedimentation rates. These other stressors cannot be identified by TIE methods. Other approaches exist to identify multiple stressors in aquatic systems [43-45]. Because TIE combines aquatic toxicology and chemistry, limitations related to the toxicity testing or analytical methods used with TIE manipulations can define the overall limitations of the TIE. For example, using organisms that are insensitive to cationic metals in a toxicity test paired with TIE manipulation may not indicate an effect unless the metal concentrations are very high, despite the fact that the TIE manipulations for metals are effective. Additionally, if analytical methods to identify a toxicant do not exist or are not used by the researcher, the TIE may not successfully identify the specific toxicant, even though the TIE may indicate the correct toxicant or contaminant class and the test organism indicates that toxicity exists.

Currently, much of the discussion of TIE methodology development is on the topic of identifying specific organic toxicants. With complex natural mixtures, it is often difficult to discern specific, individual toxicants, particularly in the large class of nonionic organic chemicals [23]. Methods to identify specific pesticides such as pyrethroids or organophosphates have been successfully employed [13, 30, 31, 46-50]; however, developing methods to identify other emerging contaminants—including toxic ionic organic chemicals such as many of the pharmaceuticals, personal care products, and degradation products of pesticides and herbicides—as well as the milieu of legacy organic compounds still remains a challenge. Finally, to our knowledge, no TIE methods have been developed to characterize toxicity caused by the carbon and metal nanomaterials now being introduced into the environment [51, 52].

Future research

Toxicity identification and evaluation development and application remain an active field. Future developmental efforts should focus on new tools for classes of sediment contaminants not currently addressed. As discussed, TIE methods for ionic organic chemicals are not readily available. An example of this need is manifested by the category of pollutants considered emerging contaminants. As noted earlier, this extremely wide-ranging category of contaminants includes pharmaceuticals, personal care products, and nanomaterials. The only characteristic these chemicals appear to share is that their ecological fate and effect data are only nascent, and TIE methods are essentially nonexistent. Furthermore, many of the emerging contaminants are not easily categorized as nonionic organic or cationic metal but share complex chemistries. This suggests that predictive chemical characteristics, such as the log KOW, used to categorize toxicants initially may not be as relevant with these contaminants [53].

Along with the need to develop new TIE methods is the need to advance the area of toxicological endpoints. As noted in the Introduction, most sediment toxicity assessments, including TIEs, focus on mortality and a limited number of sublethal endpoints [17]. Many of the emerging contaminants, for example, are not acutely toxic but have effects that require much more sensitive endpoints than we now commonly apply in TIEs. This limitation holds true not just for emerging contaminants; many of the historical pollutants such as dioxins and mercury are also not necessarily acutely toxic and require an array of toxicological endpoints different from those currently being employed. The use of genomics is not strictly a toxicological endpoint, but the combination of genomic biomarkers that can identify patterns of codons or genomic subunits affected by specific toxicants could be a useful tool with TIEs to identify toxicants or classes of toxicants [54, 55].

Finally, with regard to TIE methods, based on the studies surveyed and our current understanding of the types of contaminants adversely affecting benthic environments, nonionic organic contaminants represent the greatest ecological risk. However, although our phase I TIE methods are effective for discriminating between classes of toxicants, our phase II methods to identify specific toxicants within classes are limited. For the last few years, research has been underway to develop better phase II methods for nonionic organic contaminants. For example, Heinis et al. [56] and Perron et al. [57] describe a method involving the preparation of a surrogate contaminant phase that emulates the organic carbon in sediments. This method offers an approach for generating artificial interstitial water that can be manipulated for toxicity identification purposes under controlled conditions for which whole-sediment methods do not allow and interstitial water collection techniques may compromise. Complementary to this research are recent advances in passive dosing in which fairly simple polydimethylsiloxane (PDMS)-based systems have been developed and optimized to provide consistent concentrations of nonionic organic chemicals for exposure systems, including aqueous systems, that can be applied to toxicity testing [58-61]. Furthermore, by incorporating the elegant and sophisticated chemical fractionation and analysis methods developed in effects-directed analysis [62], the specificity of phase II TIEs would be greatly enhanced. Ultimately, these types of research and development will allow us to further distinguish between and identify specific organic toxicants.

As discussed, the first TIEs were used in the 1980s in the United States as part of the regulations under the Clean Water Act for effluents and receiving waters. Since then, TIEs have been used around the world for a variety of applications [22, 26]. More recently in the United States, the TMDL process has been recognized as the next significant step in achieving the goals of the Clean Water Act. The TMDL approach to water quality protection involves evaluating the effects of both point and nonpoint sources of water pollution that result in environmental impairment. A primary component of the TMDL is diagnosing what stressor is causing the impairment. When the impairment is sediment toxicity, TIEs may be invoked to characterize the category of toxicant or identify the specific toxicant. The continued refinement of approaches for integrating the use of sediment TIEs into the TMDL process is an ongoing challenge requiring further research. The National Academy of Sciences [43] has stated that water bodies that have been categorized as impaired under the TMDL process should be assessed through an integrated approach. The whole-sediment TIE approach is ideal for inclusion in this type of diagnostic assessment.

SUMMARY

Nonionic organic chemicals play the largest role in sediment toxicity regardless of freshwater or marine environments. This is observed to a greater extent in whole-sediment TIEs compared with interstitial-water TIEs. In contrast, ammonia plays a surprisingly small role as a toxicant, particularly in whole-sediment exposures. While ammonia may not be a major player in this nation-wide survey of sediments, it may play a role in localized areas [63].

Approximately 5% of all sediments in the present TIE survey were reported to have an unknown toxicant. Reasons for the categorization of unidentifiable toxicants include the need for method development for additional classes of toxicants not currently covered, particularly ionic organic chemicals. This large class of chemicals includes emerging contaminants such as pharmaceutical and personal care products. Also, many of the earlier studies reviewed for this survey might not have used the more recently developed TIE methods for specific pesticides [36, 48, 49, 64-67].

Finally, this relatively simplistic breakdown of phase I TIE results into nonionic organic chemicals, cationic metals, and ammonia categories belies the complexity of the many peer-reviewed studies from which conclusions have been drawn (Table 1; Supplemental Data, Table S1). For example, many of these elegant studies successfully used TIE methods outlined in the present study and in the references cited to identify and confirm specific toxicants such as diazinon, chloropyrifos, cyhalothrin, permethrin, bifenthrin, arsenic and manganese [13, 68]. To fully understand the capability of the TIE approach, we recommend reading the original papers.

SUPPLEMENTAL DATA

Table S1. (54 KB DOCX).

Acknowledgment

The authors thank R. McKinney, J. Serbst, L. Portis, and A. Kuhn for their comments. This is contribution AED-07-018 of the US Environmental Protection Agency, Office of Research and Development, National Health and Environmental Effects Research Laboratory, Atlantic Ecology Division. This paper has been technically reviewed by the Atlantic Ecology Division; however, it has not been subject to Agency-wide peer review and therefore does not necessarily represent the views of the US Environmental Protection Agency. No official endorsement of any aforementioned products should be inferred.

Ancillary