Environmental fate of pyrethroids in urban and suburban stream sediments and the appropriateness of Hyalella azteca model in determining ecological risk



According to several recent studies using standard acute Hyalella azteca sediment bioassays, increased pyrethroid use in urban and suburban regions in California has resulted in the accumulation of toxic concentrations of pyrethroids in sediments of area streams and estuaries. However, a critical review of the literature indicates that this is likely an overestimation of environmental risk. Hyalella azteca is consistently the most susceptible organism to both aqueous and sediment-associated pyrethroid exposures when compared to a suite of other aquatic taxa. In some cases, H. azteca LC50 values are less than the community HC10 values, suggesting that the amphipod is an overly conservative model for community- or ecosystem-level impacts of sediment-associated pyrethroids. Further, as a model for responses of field populations of H. azteca, the laboratory bioassays considerably overestimate exposure, because the amphipod is more appropriately characterized as an epibenthic organism, not a true sediment dweller; H. azteca preferentially inhabit aquatic macrophytes, periphyton mats, and leaf litter, which drastically reduces their exposure to contaminated sediments. Sediment-bound pyrethroids are transported via downstream washing of fine particulates resulting in longer range transport but also more efficient sequestration of the chemical. In addition, site-specific variables such as sediment organic carbon content, grain size, temperature, and microbial activity alter pyrethroid bioavailability, degradation, and toxicity on a microhabitat scale. The type and source of the carbon in particular, influences the pyrethroid sequestering ability of sediments. The resulting irregular distribution of pyrethroids in stream sediments suggests that sufficient nonimpacted habitat may exist as refugia for resident sediment-dwelling organisms for rapid recolonization to occur. Given these factors, we argue that the amphipod model provides, at best, a screening level assessment of pyrethroid impacts and can correctly identify those sediments not toxic to benthic organisms but cannot accurately predict where sediments will be toxic. Integr Environ Assess Manag 2011;7:325–335. © 2011 SETAC


Since the decision by the US Environmental Protection Agency (USEPA) to curtail use of organophosphate insecticides, the subsequent use of pyrethroid insecticides has increased, as these compounds present a valuable alternative to toxic organophosphates. Specifically in the case of residential and home use, pyrethroid insecticides such as bifenthrin and permethrin have replaced the previously popular organophosphates chlorpyrifos and diazinon. Although the pyrethroids offer lower toxicity to human applicators and nontarget mammals and birds, they are toxic to invertebrates and fish, and have been demonstrated to exhibit different environmental behavior and fate. Specifically, the hydrophobicity of pyrethroid insecticides has raised concerns regarding persistence in soils and sediments following this increased use.

Recent research has shown that increased pyrethroid use in urban and suburban regions in California has resulted in sediments toxic to resident benthic invertebrates. Amweg and Weston (2007) report that more than 20% of California sediment samples contain pyrethroid concentrations exceeding the acute toxicity threshold for the amphipod Hyalella azteca. Similarly, Holmes et al. (2008) reported that 100% of urban creek sediment samples collected from sampling sites in the San Francisco Bay area contained bifenthrin concentrations ranging from 2.19 to 219 ng/g dry weight; bifenthrin is 1 of the pyrethroid compounds that is more toxic to H. azteca, with reported 10-d LC50 values of less than 9 ng/g (Amweg et al. 2005).

Although recent laboratory and field evaluations have indicated that pyrethroid levels in California sediments have reached levels toxic to H. azteca, the use of these data as an indicator for larger ecological impact in California's aquatic ecosystems is questionable. First, the environmental fate and toxicity of pyrethroids in suburban and urban California water bodies is affected by a number of site- and compound-specific variables, such as dissolved organic carbon (DOC) content in the water column, sediment organic carbon content and grain size variability, adsorption affinity of specific pyrethroids, site temperature, and microbial activity, among others. This has been demonstrated to result in highly variable distribution and bioavailability of pyrethroids in sediments. Second, even though the linkage between effect in H. azteca and pyrethroid concentrations in field-collected sediments has improved with the development of Toxicity Identification Evaluation (TIE) tools, H. azteca is a poor indicator species for greater community effects, as these amphipods exhibit uniquely high sensitivity to pyrethroid insecticides, and not true sediment dwellers or feeders. Consequently, effects observed in H. azteca populations are not predictive of effects for other species. This is borne out by mesocosm studies of population responses and recoveries, species sensitivity distributions and site-specific field investigations of community-level pyrethroid effects. Taken together, the available literature data on pyrethroid environmental fate and toxicity indicate that toxicity, especially to sediment-dwelling invertebrates, is site-specific and that distribution of toxic concentrations of pyrethroids will be interspersed with areas of relative “cleanness.” Consequently, data generated using field-collected sediment and standard laboratory H. azteca assays comprise at best a screening-level assessment of pyrethroid aquatic ecosystem effects, and further investigation is required before a community-level assessment of effects is possible.


Given the highly hydrophobic behavior of pyrethroid insecticides, their environmental fate depends largely on physical characteristics within the stream, such as the organic carbon quality and content of the system, prevalence of vegetation, and flow rate, among others. In general, pyrethroids entering surface waters either 1) rapidly adhere to organic materials (You et al. 2008, Giddings et al. 2001), or 2) are associated with terrestrial particulate or organic material that is washed into surface waters (Lee et al. 2004, Ortego and Benson 1992). Consequently, the transport and fate of pyrethroid contamination in surface sediments depends heavily on sedimentation rates and the movement of particulates through the system. This may partially explain why laboratory studies of water-only exposures of H. azecta overpredict effects observed in the field (Giddings et al. 2001, Kedwards et al. 1999a). The strong binding of pyrethroids to organic carbon and particulates reduces the bioavailable fraction, and particulate-bound pyrethroids are patchily distributed within a stream, leaving sections of substrate relatively free of pyrethroids potentially to serve as refugia for sensitive benthos.

Gan et al. (2005) described the distribution and fate of sediment-bound pyrethroids along a 260-m runoff channel draining a large nursery near Irvine, California, which regularly used both bifenthrin and permethrin for pest control. Runoff water drained first to a settling pond and then along a drainage channel. Both bifenthrin and permethrin sediment concentrations became increasingly enriched further downstream from the source. For example, bifenthrin pond sediment concentrations averaged 0.33 mg/kg and increased to greater than 10 mg/kg in sediment samples collected 145 m downstream. However, the transport of these and other synthetic pyrethroids depends on particle size. Heavy particulates settle out quickly while those compounds associated with more mobile DOC and dissolved organic matter (DOM) likely will be transported much further (Liu et al. 2004). In fact, the amplified downstream sediment concentration of bifenthrin and permethrin observed by Gan et al. (2005) resulted from a greater rate of transport of small particles with high organic carbon and clay content, which had the highest binding affinity for synthetic pyrethroids. This was also observed by Lao et al. (2010) who reported that fine-grained, high organic carbon sediments in a Los Angeles, California urban estuary contained the highest concentrations of detected pyrethroids, including bifenthrin, permethrin and cypermethrin. However, as discussed in the next section, pyrethroids bound to DOC or DOM are significantly less bioavailable to benthic invertebrates, resulting in the counterintuitive conclusion that the stream sediments with the highest pyrethroid concentrations had the lowest toxicity (Gan et al. 2005).

Lack of mobility of pyrethroids bound to sediments occurs in some instances. For example, aerial applications of cypermethrin to a cotton field surrounding an experimental pond resulted in significant concentrations of the pyrethroid insecticide in deep sedimented zones, while littoral areas contained lower cypermethrin residues (Kedwards et al. 1999b). This distribution may have resulted from increased fine sedimentation in the deeper areas, or from more rapid biodegradation or photodegradation in the shallower, more biologically productive areas. Because littoral zones provide habitat to a greater diversity of organisms than do the deeper pelagic areas of ponds or lakes, this indicates that most of the invertebrate and fish taxa inhabiting the pond are likely exposed to much lower concentrations of pyrethroids than would be predicted by simple dilution models.

Aquatic and rooted macrophytes are another major sink for pyrethroids introduced directly in the water column. For example, the concentration of bifenthrin adhered to plants can exceed the maximum sediment bifenthrin concentration by a factor of 100 (Bennett et al. 2005). Mesocosm studies with BaythroidTM (a formulation containing 12.7% cyfluthrin) also indicated that aquatic macrophytes are important sinks for surface water pyrethroid contamination (Heimbach et al. 1992).

Continuous runoff can result in accumulation of sediment-bound pyrethroids in receiving water bodies, but this scenario is likely to be punctuated by significant rainfall events that alter the deposition rate and locations, or wash out previous deposits of contamination (Yang, Spurlock, et al. 2006). This results in patchy distributions of pyrethroids throughout a system, with relatively nonimpacted benthic habitats serving as refugia for sensitive invertebrates. In suburban (Roseville, California, USA) streams, multiple sediment samples that were toxic to Hyalella in laboratory sediment assays were not uniformly distributed throughout a stream (Weston et al. 2005). The majority of toxic sediments were situated near storm drain inputs and tributary entry points, which may have resulted from a change in sediment transportation at these junctions. However, it is probable that as localized habitat conditions improve (e.g., through microbial degradation of pyrethroids, a reduction in deposition of contaminated sediments, or by the flushing of contaminated hotspots), H. azteca will be able to recolonize these areas from local source populations. Consequently, even urban sites receiving high inputs of pyrethroids are likely to be characterized at worst by localized depletions of the sensitive invertebrate, rather than pervasive and persistent depauperate benthic and aquatic communities. In suburban and urban waterways, addition of vegetation, settling ponds, and sedimentation traps are low-cost techniques for remediating pyrethroid-contaminated sediment transport (Bennett et al. 2005, Gan et al. 2005).


The bioavailability of hydrophobic pyrethroid compounds to pelagic and benthic organisms is heavily influenced by DOC or DOM. In general, increased water column organic content is correlated with decreased pyrethroid toxicity to water column organisms, likely as a result of reduced bioavailability (Muir et al. 1994, Yang et al. 2007). However, recent studies indicate that specific physical-chemical properties of organic matter also influence pyrethroid bioavailability, to the point that the “quality” of DOM may be more important than the “quantity” of DOM (Yang et al. 2007). Consequently, traditional analytical methods have tended to overestimate the bioavailable fraction of pyrethroids in porewater.

Influence of DOC/DOM on aquatic toxicity

Rainbow trout uptake of deltamethrin, cypermethrin, and permethrin was significantly reduced (by 63, 59, and 41%, respectively) by the addition of humic acid into the aqueous exposure system (Muir et al. 1994). Similarly, addition of suspended sediment decreased the toxicity of bifenthrin and cypermethrin to Ceriodaphnia dubia by more than an order of magnitude. This indicates that sediment-adsorbed pyrethroids are unavailable for uptake and have no toxic effects on water column invertebrates (Yang, Spurlock, et al. 2006).

The bioavailability of sediment-associated pyrethroids is influenced by a number of variables, including particle size, sediment organic carbon content and variety, and degree of aging. You et al. (2008) examined different sediments with high pyrethroid concentrations but unexpectedly low toxicity to H. azteca. Although a standard toxic unit method was used to predict toxicities of sediment, H. azteca exhibited variable responses that were not predicted by the model. Discrepancies between modeled and observed toxicities were a result of differences in pyrethroid bioavailability not quantified during chemical analysis. Bioavailability was strongly correlated with sediment grain size distribution; increased coarse grain content (e.g., sand) reduced bioavailability, and decreased H. azteca toxicity (You et al. 2008). Consequently, the pyrethroid toxic unit method will overestimate the toxicity of pyrethroids in sandy sediments.

Fleming et al. (1998) also noted the importance of the type and quality of organic carbon in regulating the toxicity of pyrethroids to sediment-dwelling invertebrates. Chironomus riparius larvae were more likely to survive permethrin exposure in peat-containing sediments versus those inhabiting sediments with α-cellulose as the dominant organic carbon source. Further, binding affinities of pyrethroids are compound-specific; for example, the derived DOC binding constant (KDOC) for cyfluthrin is twice the value for permethrin (Yang et al. 2007).

Mortality of C. riparius in artificially constructed sediments was reduced by increased clay and carbon content in both peat and α-cellulose sediments (Fleming et al. 1998). The authors concluded that predictions of toxicity based on artificial sediments were likely to overestimate effects in natural systems because of the high natural variety of organic carbon in streams. This was corroborated by Maund et al. (2001) who conducted field studies and showed that the bioavailability of cypermethrin was related to local sediment organic carbon content. They also demonstrated that both epibenthic (inhabiting the water-benthic interface) and benthic (sediment-dwelling) organisms bioaccumulated a similar amount of sediment-associated pyrethroids, and concluded that ingestion of or direct contact with pyrethroid-contaminated sediments did not significantly contribute to bioaccumulated body burdens (Maund et al. 2001). This indicates that uptake and bioaccumulation is driven by cuticular uptake of the dissolved fraction, and further that dietary exposures to pyrethroids are unlikely to contribute to uptake and toxicity.

Pyrethroid compounds can behave differently under the same environmental conditions, making it difficult to extrapolate results from 1 chemical to another. For example, even small amounts of DOM in surface waters (concentrations of 320 mg/L) are sufficient to inhibit the uptake of cyfluthrin by pelagic invertebrates such as Daphnia magna and C. dubia, while permethrin uptake is not significantly influenced by DOM concentrations (Yang et al. 2007). Bifenthrin also is much more likely to be bound to DOM than is permethrin (Yang, Gan, et al. 2006). As with sediments, the physicochemical properties of the DOM may be more critical than DOM concentrations in the sequestration of pyrethroid insecticides. Ortego and Benson (1992) showed that the source of humic acid significantly affected the aqueous toxicity of pyrethroid insecticides. The relationship between DOM characteristics and their adsorption affinity for various chemicals has been previously established with other hydrophobic compounds. For pyrethroid compounds, DOM carboxyl group content is likely to be a critical driver in chemical sequestration (Yang et al. 2007).

Measurement of bioavailable fraction

Gan (2006) reported that the conventional methods of quantifying bioavailable pyrethroids in sediments often overestimated bioavailability and therefore the toxicity of the sediment-associated contamination. However, when the uptake of hydrophobic contaminants from sediment by solid-phase microextraction (SPME) fibers was compared to that of resident biota, the concentrations were demonstrated to closely correlate (You et al. 2006), suggesting this may be a more appropriate technology for quantifying actual exposures (i.e., the bioavailable fraction).

Phase-distribution analysis of water column pyrethroid contamination via SPME demonstrated that more than 98% of the bifenthrin, cis-permethrin, and trans-permethrin spiked into river water samples was bound to DOM (Liu et al. 2004). Similarly, when runoff water was collected from an Orange County, California, nursery outlet, more than 86% of the bifenthrin, cis-permethrin, and trans-permethrin concentrations were determined to be adsorbed to DOM. In sediments, organic carbon content greater than 3% resulted in pyrethroid pore water concentrations 1 to 3 orders of magnitude lower than laboratory-spiked sediments low in organic matter (Hunter et al. 2009).

Given the importance of site-specific sediment characteristics in determining pyrethroid bioavailability, sediment toxicity will be highly variable from site to site, even if traditional analytical methods indicate that sediments contain similar pyrethroid concentrations. Aging and sediment turnover, in particular, have been shown to decrease the bioavailable fraction of pyrethroid contamination. Recently developed SPME methodologies have also proven more successful in quantifying this fraction (You et al. 2006), and dietary studies indicate that ingestion of sediment-sorbed pyrethroids is unlikely to cause toxic effects at naturally occurring levels (Maund et al. 2001).


Although adsorption to sediment reduces the bioavailability and toxicity of pyrethroids to nontarget aquatic organisms, sediment-binding has also been shown to inhibit degradation (Lee et al. 2004), causing concerns regarding the potential accumulation of pyrethroids in sediments. However, recent evidence shows that pyrethroid degradation does occur under certain environmental conditions (Åkerblom et al. 2008, You et al. 2009).

Over the course of a sediment partitioning study, significant hydrolysis of both cyfluthrin and cypermethrin was observed prior to adsorption to particulates (Hunter et al. 2009). Consequently, in the case of freely dissolved pyrethroids entering surface waters, significant degradation may occur prior to sediment sequestration or uptake by non-target organisms.

Biodegradation studies have demonstrated that pyrethroids bound to sediment or other particulate matter are more resistant to microbial breakdown than freely dissolved pyrethroids (Lee et al. 2004). No information exists about the potential relationship between specific sediment characteristics and microbial degradation rates of bound pyrethroids. However, the degradation rate of deltamethrin was demonstrated to be much higher in natural sediment systems than in artificial sediments, perhaps because of increased microbial activity in field-collected sediments (Åkerblom et al. 2008).

In tests with the sediment-dwelling oligochaete Lumbriculus variegatus, a significant portion of the pyrethroid insecticides bifenthrin and permethrin in sediment was biotransformed and metabolized by the worms (e.g., approximately 50% of parent compound transformed for each) (You et al. 2009). This is surprising given that USEPA has previously considered the oligochaete only a weak degrader of hydrophobic contaminants. Hence, field biodegradation may be higher than previously thought, because both microbes and macrobenthos contribute to biodegradation, in effect reducing the residence times of sediment-associated pyrethroids.

Enantioselectivity of pyrethroid toxicity, uptake, and biodegradation

Synthetic pyrethroid compounds are often composed of various isomer compounds, which are known to exhibit different toxicities. In some cases, uptake, degradation, and transport can also differ significantly for isomers. This increases the difficulty of extrapolating observed system responses among the different pyrethroids, as significant differences in effects (or effect thresholds) may occur within 1 compound, depending on the mixture of enantiomers (isomers exhibiting the same chemical bond structures, but different geometric structures).

For instance, significant enantioselectivity was noted in both the toxicity of bifenthrin and in environmental fate and degradation pathways. Ceriodaphnia dubia were twice as sensitive to the 1R-cis-bifenthrin isomer as to the racemic mixture composed of both 1R-cis-bifenthrin and 1S-cis-bifenthrin (Liu et al. 2005). This is likely a result of different rates of uptake, metabolism, or toxic activity influenced by structural variability between enantiomers. Similarly, the 7-d survival lowest-observable-effect concentration for Daphnia magna was 20-fold higher when the invertebrates were exposed to 1S-cis-bifenthrin versus 1R-cis-bifenthrin. Much of this discrepancy appears to be a result of reduced bioaccumulation of 1S-cis-BF by D. magna (Zhao et al. 2009); Japanese medaka (Oryzias latipes) have also been shown to preferentially bioaccumulate the 1R-cis-bifenthrin enantiomer (Wang et al. 2007).

The variable toxicity of pyrethroid enantiomers may also indicate that enantiomers exhibit different affinities for suspended sediment and DOM, but no information is available on this topic. Microbial degradation of pyrethroids, in particular, can also proceed at different rates for different isomers, depending on bacterial strain and specific pyrethroid compound. A Yersinia frederiksenii culture preferentially degraded 1R-cis-permethrin over the 1S-cis isomer, but none of the 6 pyrethroid-degrading bacterial strains tested demonstrated increased degradation of either the 1R-cis or 1S-cis bifenthrin isomers (Liu et al. 2005). However, in field situations, variations in pyrethroid isomer fate, toxicity, and degradation may be significant.


The environmental fate and bioavailability data discussed above, in combination with differential toxicities of pyrethroids and life history traits of in situ organisms, suggest that the consequences of exposure to pyrethroid-containing sediments will vary not only among species but even between sites with ostensibly similar pyrethroid loads. Data from some field sites in California support this theory and demonstrate the difficulty of predicting field effects from single species (H. azteca) laboratory bioassays (You et al. 2008, Weston and Amweg 2007).

Sediment collected from Del Puerto Creek, California, USA, with a bifenthrin concentration equivalent to 35 toxic units, was diluted to 6%, giving toxic units of 2.1, assuming homogeneity of sediment samples. Resulting H. azteca survival averaged 56%. However, contrary to what might be expected, H. azteca survival was also 56% in sediment collected from Kaseberg Creek, California, USA, determined to contain bifenthrin equivalent to 0.64 toxic units (Weston and Amweg 2007). This is compelling evidence that the magnitude of toxicity of sediment-bound pyrethroids may be highly site-specific and dependent on the qualities of the localized organic matter. It also indicates a potential ameliorative effect of sediment turnover and mixing.

As a shredder species, H. azteca can preferentially inhabit leaf material and other large organic detritus in heterogeneous benthic sediment matrices; consequently, inclusion of variation in materials in the sediment provides additional realism in modeling effects likely seen in field situations. Maul et al. (2008) exposed H. azteca to bifenthrin in sediment-only, leaf-only, and mixed sediment and leaf matrices. Organic carbon-normalized lowest-observable-effect concentration values demonstrated that the mixed matrix was 3 times less toxic to the benthic invertebrate than sediment-only exposures. Hyalella azteca bifenthrin body residues also were lower when exposed in the mixed sediment and leaf matrix, demonstrating that uptake was reduced in a more realistic environment. The authors surmised that 2 factors contributed to this phenomenon: 1) the sediment sequestered significantly more bifenthrin mass, and 2) H. azteca preferentially inhabited leaf material over sediment. This effectively reduced the concentration of bifenthrin to which the invertebrates were exposed (Maul et al. 2008). The lower KOC value for leaf material versus sediment or suspended particulates also indicates that dietary exposure to pyrethroids is less likely.

You et al. (2008) investigated California sediments that exhibited unexpectedly low toxicity to H. azteca in laboratory bioassays based on the concentration of pyrethroids in samples. In this case, sediments with similar pyrethroid toxic units exhibited variable toxicities to H. azteca. It was concluded that site-specific sediment characteristics altered the bioavailability of pyrethroids; organic content and quality, as well as grain size, were demonstrated to affect toxicity at a site-specific level. Consequently, even in pyrethroid-influenced systems, benthic habitats with rich organic content or high proportions of sand may prove suitable for some sensitive invertebrates, as the site-specific pyrethroid toxicity is likely to be reduced in these areas.


Development of novel TIE tools for sediment-associated pyrethroids has assisted in the identification of sites where toxicity to H. azteca is driven by these insecticides (as opposed to sites containing toxic loads of heavy metals or organophosphate insecticides). Two main TIE methodologies have been advanced to distinguish toxic levels of pyrethroids: piperonyl butoxide addition (increases pyrethroid toxicity) and carboxylesterase addition (decreases pyrethroid toxicity). Piperonyl butoxide (PBO), a relatively nontoxic pyrethroid synergist, simultaneously reduces the toxicity of organophosphate insecticides, making it a questionable tool for distinguishing between contributions to toxicity in sediments containing both organophosphates and pyrethroids. Often, this TIE would require the quantification of pyrethroid and organophosphate toxic units to elucidate results of PBO addition. In a weight-of-evidence approach, this tool was used to identify bifenthrin toxicity in field-collected sediments, and was found to increase toxicity by a factor of 2.5 to 3.7 (Amweg and Weston 2007). In some cases, however, the toxicity of chlorpyrifos (an organophosphate) was also increased by a factor of 1.8 to 2.6. Therefore, attributions of causation based solely on this test are inconclusive at best. Phillips et al. (2006) used both PBO and carboxylesterase TIE tools to evaluate the toxicity of field collected sediments, and found that pyrethroids were the likely cause of sediment toxicity in some instances. Carboxylesterase rapidly degrades pyrethroid insecticides, thus increasing amphipod survival, and PBO addition reduces amphipod survival (Wheelock et al. 2008). Weston and Amweg (2007) also used the carboxylesterase TIE in evaluating the toxicity of field collected sediments, and found that the addition of the esterase increased H. azteca survival in bifenthrin-contaminated sediment by as much as a factor of 30, and survival in both permethrin- and cypermethrin-contaminated sediment by as much as a factor of 15.

Temperature modification has been introduced as a 3rd step in the TIE for pyrethroids, as a distinct increase in pyrethroid toxicity to benthic invertebrates at lower temperatures has been observed (Wheelock et al. 2008). This is in contrast to other contaminants, particularly organophosphate insecticides, which exhibit a marked increase in toxicity with a rise in ambient temperature. In general, higher temperatures have been theorized to cause heightened toxicity via increased uptake of contaminant. However, in the case of pyrethroids and DDT, toxicity increases at lower temperatures, even though contaminant uptake is slowed. The observed increase in toxicity may occur as a result of diminished metabolic capacity of invertebrates at lower temperatures, resulting in an accumulation of the more toxic parent compounds; conversely, reduced metabolism likely limits the conversion of organophosphates to the more toxic oxon forms (Harwood et al. 2009). This temperature modification method was used to examine 30 sediment samples from urban creeks across California (Los Angeles, North Coast, Central Coast, San Francisco, Santa Ana, San Diego, Lake Tahoe, and Central Valley regions). In 18 of the samples, H. azteca survival was reduced when sediment toxicity bioassays were conducted at 15 °C versus 23 °C, indicating that either pyrethroids or organochlorine insecticides were likely present at toxic levels (Holmes et al. 2008). At both temperatures, however, the sum of pyrethroid toxic units closely correlated with H. azteca mortality, and although organochlorine insecticides were detected in more than 75% of samples, it was concluded that pyrethroids were the most likely causative agent.

Given the development of useful TIE methods for ascribing observed sediment toxicity to pyrethroid contamination, there is persuasive evidence that, for a number of sites, pyrethroids do contribute to the reduced H. azteca survivorship observed during standard toxicity bioassays. However, TIE results are most useful in the identification of causes of toxicity in laboratory experiments, and the connection between these results and potential field effects is more tenuous. Therefore, it is questionable whether these laboratory results can be used to accurately predict conditions in the field. In fact, several factors indicate that the standard approach of laboratory H. azteca toxicity tests conducted with field-collected sediment provides a poor model for predicting in situ community-level effects of pyrethroid exposure, and that these tests are more appropriately used as a 1st-tier screening tool for urban and suburban sites.

Hyalella as predictors of community effects

Available data indicate that H. azteca bioassays significantly overestimate the impacts of pyrethroid contamination on aquatic ecosystems, as this species is exceptionally sensitive to pyrethroid insecticides. Microcosm investigations of pyrethroid effects, in particular, bear this out.

Giddings et al. (2001) and Kedwards et al. (1999b) demonstrated that amphipods responded to lower pyrethroid concentrations than did other resident invertebrates in mesocosm ponds. Giddings et al. (2001) reported that exposures of 30 ng/L cypermethrin and greater reduced populations of resident amphipods, whereas other taxa, specifically Chironomidae, Ephemeroptera, Cladocera, and Copepoda, exhibited increased abundances. Cypermethrin surface water concentrations of 100 ng/L and greater resulted in decreased abundances of these less sensitive taxa, but concentrations of up to 1,000 ng/L had no effect on odonates, snails, and fish. Except for the amphipods, all species showed signs of recovery within 10 weeks following cessation of exposure. Similarly, Kedwards et al. (1999b) treated mesocosms with cypermethrin to give sediment concentrations ranging from 2 to 25 µg/kg. Populations of Gammaridae and Asellidae were depressed, and Chironomidae, Planorbidae, Hirudinae, and Lymnaeidae increased in numbers. These authors also noted that amphipods were the most sensitive organisms. Heimbach et al. (1992) and Hoagland et al. (1993) also documented increased rotifer populations during mesocosm experiments.

In addition to differences in species sensitivity, Heimbach et al. (1992) noted that initially depressed populations recovered so rapidly after cyfluthrin exposure that the effects on the mesocosm communities were limited, and that diversity indices were unaffected during the course of the study. Sherratt et al. (1999) determined that recovery rates of impacted invertebrate populations depended not only on relative sensitivity but also on species-specific reproductive rate. Giddings et al (2001) also noted that even with modest immigration rates, populations of sensitive amphipods would be able to recover rapidly following pyrethroid exposures. Consequently, under normal field conditions, it is likely that even extremely pyrethroid-sensitive organisms will be able to recover quickly if they exhibit a high reproductive output (faster generation time, higher number of offspring, and so forth). Hyalella azteca, in particular, demonstrate rapid population growth in situ, with an average daily population turnover of 2.9% measured in a Michigan, USA lake (Cooper 1965). Even without significant immigration, species with high rates of intrinsic increase, such as H. azteca, were shown to recover faster from population reductions than species with longer generation times and fewer offspring (Stark et al. 2004).

Recovery of a benthic invertebrate population following pyrethroid exposure was observed in field ponds by Conrad et al. (1999). Populations of C. riparius residing in ponds sprayed with PicketTM (a permethrin insecticide) were initially reduced, but abundances recovered rapidly with no indication of prolonged sublethal effects. The authors reported that this finding was in line with reports of fast recovery of nontarget aquatic arthropods, even after greater than 50% reduction in abundances, following pyrethroid exposure. More recently, Rudolph et al. (2009) demonstrated that H. azteca toxicity was not predictive of indices of community structure and vigor in field streams. Although pyrethroid concentrations in sediments collected from streams in the San Diego, California, area were sufficient to reduce H. azteca survivorship in laboratory bioassays, the test results showed no correlation to in situ indices of resident macroinvertebrate diversity. In fact, alterations to the macroinvertebrate community structure were most closely correlated to indices of urbanization. More importantly, even at sites demonstrated to be uninhabitable by H. azteca because of pyrethroid contamination, no significant effect on the larger invertebrate community could be established.

Relative sensitivity of Hyalella

It is likely that the lack of correlation between positive responses of H. azteca in laboratory bioassays and in situ invertebrate communities is at least partially explained by the high sensitivity of amphipods to pyrethroids. When compared to other aquatic organisms, H. azteca are notably more sensitive to aqueous bifenthrin and permethrin exposures. The amphipod is 1 to 2 orders of magnitude more susceptible to aqueous exposures than the next most sensitive taxon, the cladoceran Ceriodaphnia dubia (Figures 1 and 2), and can be 1 to 2 orders of magnitude more sensitive than Chironomus tentans in sediment exposures (Figure 3). Hyallela azteca is also orders of magnitude more sensitive to pyrethroids than amphibian and fish species, including some endangered taxa (Figure 2). In fact, for aqueous exposures, H. azteca toxicity values are less than the community HC10 value (the concentration of contaminant considered hazardous to 10% of the community's species). Giddings et al (2001) concluded that basing the ecological risk characterization on the 10th percentile for community species sensitivity would incorporate sufficient conservatism.

Figure 1.

Species sensitivity distribution for aqueous exposure to bifenthrin (96-h LC50 values). Line indicates community HC10 value. Data from Anderson et al. (2006), Liu et al. (2005a), and Liu et al. (2005b).

Figure 2.

Species sensitivity distribution for aqueous exposure to permethrin (96-h LC50 values). Line indicates community HC10 value. Data from Anderson et al 2006, Bridges et al. (2002), Conrad 1999, Dwyer et al. (2005), Hemmer et al 1992, Liu et al. (2005b), and Yang et al. (2007).

Figure 3.

Species sensitivity distribution for organic carbon normalized bifenthrin sediment toxicity (10-d LC50 values). Data from Amweg et al. (2005), Amweg and Weston (2007), and Maul et al. (2008).

In terms of sediment toxicity, the sensitivity of H. azteca to sediment-bound bifenthrin, cypermethrin, and permethrin is greater than that of Chironomus spp., and H. azteca toxicity values are noticeably clustered at the sensitive ends of both scales (e.g., not interspersed with Chironomus spp. toxicity values) (Figures 3 to 5). However, these data sets are relatively limited, because only 2 species—H. azteca and Chironomus spp.—are regularly used to assess sediment toxicity. Given that no inherent differences in pyrethroid sensitivity have been detected between water column- and sediment-dwelling organisms (Warinton et al. 2006, Maund et al. 2001), it is likely that species sensitivity distributions for sediment-dwellers would approximate those for water column organisms if the same diversity of taxa could be included.

Figure 4.

Species sensitivity distribution for organic carbon normalized cypermethrin sediment toxicity (10-d LC50 values). Data from Aweg and Weston (2007) and Maund et al. (2002).

Figure 5.

Species sensitivity distribution for organic carbon normalized permethrin sediment toxicity (10-d LC50 values). Data taken from Amweg et al. (2005), Conrad et al. (1999), Harwood et al. (2009), and Maul et al. (2008).

Consequently, these data support the conclusion that H. azteca are uniquely susceptible to pyrethroid exposure, and that responses in these populations are not likely to be indicative of effects on larger community structure or function. In fact, this has been cited as a potential benefit of the use of H. azteca in pyrethroid toxicity testing, as their sensitivity increases the chances of detecting toxicity in laboratory bioassays (Weston and Lydy 2010). Although this approach provides an excellent 1st-tier bioassay screening useful for large or complex sites, sediment toxicity results from H. azteca bioassays provide little information concerning the population- or community-level effects of sediment-adsorbed pyrethroids. Certainly, reductions in a few sensitive taxa are unlikely to cripple ecosystem functionality (Giddings et al. 2001, Giller et al. 2004), as less sensitive but functionally redundant organisms may proliferate and inhabit the ecological niche left available by the impacted sensitive species. Further, patchy chemical distributions and habitat heterogeneity result in uncontaminated refugia from which impacted species can recolonize habitats depleted by pyrethroids (Giddings et al. 2001). Maund et al. (2001) demonstrated that under environmentally realistic conditions pyrethroid exposure is considerably less than what would be predicted by 1st-tier assessments, and therefore the risk posed to field populations of fish and invertebrates by pyrethroids is small. Consequently, ecosystem sensitivity is poorly represented by the responses of the most sensitive taxa. Giddings et al. (2001) proposed that the USEPA's assumption when setting water quality criteria of protection of 90% of the species in an aquatic community is a valid and sufficiently protective management goal.

Influences on in situ species vulnerability

Vulnerability of a species to chemicals is influenced by its relative sensitivity (discussed above), and its exposure rate. Exposures are, in turn, influenced by life history traits that place the organism in areas of highest (or lowest) concentration of the chemical, the bioavailability of the chemical to specific organisms, the transport processes in the environment that allow the chemical to accumulate in areas where organisms may be most (or least) exposed. This section examines how these factors affect the vulnerability of the benthic macroinvertebrate community as a whole, as well differences among specific genera.

Hyalella in situ Behavior and Habitat Preference

To ensure reliable bioassay results, heterogeneity of laboratory test systems is often reduced (e.g., sediment homogenization and removal of large debris in sediment bioassays). Although this may ensure equivalent individual exposures, it often provides an imprecise estimation of likely field responses. In the case of H. azteca sediment bioassays, the standard laboratory set-ups may serve to artificially increase exposure above levels that would be observed under field conditions.

In the field, Hyalella populations demonstrate a strong preference for aquatic plants, algae, and other organic detritus over truly benthic habitat, and are often found burrowing into macrophyte root mats and epibenthic leaf material rather than sediments (Cooper 1965). In fact, spatial variability associated with field populations of H. azteca is largely influenced by the distribution and quantity of rooted aquatic macrophytes or by chlorophyll concentration of the epibenthic layer, rather than sediment characteristics (Cooper 1965, Wang et al. 2004). Consequently, exposure to sediment-associated contamination is likely to be greatly reduced under heterogeneous field conditions.

According to Wang et al. (2004), H. azteca burrowing behavior observed during laboratory assays is most likely a disturbance response to light or jostling, or is a result of the absence of rooted plants. When observed undisturbed in a natural environment, H. azteca is more likely to seek refuge in branched macrophtyes than to burrow into sediments. In a bifenthrin sediment assay conducted with H. azteca, senesced leaf material was provided as additional substrate. During the course of the study, individuals were observed preferentially inhabiting leaf material over sediments, and H. azteca survival was significantly increased in the presence of leaf detritus (Maul et al. 2008). The amphipods were noted both clinging to and actively shredding the allochthonous material, suggesting that this substrate also provided a richer food source.

Dietary preferences of H. azteca also suggest that the species is more accurately characterized as an epibenthic omnivore, rather than a true sediment feeder such as larval Chironomus species. Wang et al. (2004) report that results of a 14C dietary study indicated a clear difference in the dietary preferences of chironomids (e.g., sediment-based sources) and H. azteca (algal and exogenous leaf material). Consequently, these data are a strong indication that, in the field, H. azteca is more appropriately considered an epibenthic or water column species, and as such is unlikely to be frequently exposed to contaminated sediments in situ. Given this, access to appropriate food sources may be a common confounding factor in standard laboratory H. azteca sediment bioassays. In a series of laboratory studies, Ankley et al. (1994) determined that H. azteca survival in uncontaminated sediment was highly dependent on access to food, whereas survival of sediment-dwelling Chironomus tentans and Lumbriculus variegatus was not. Ankley et al. (1994) surmised that this may be a result of the amphipod's dietary preference for diatoms and green algae, while the dipteran and oligochaete more frequently consume sediment detritus as a significant food source. This led to the conclusion that reduced food availability is likely to increase the chances of false positive results in laboratory H. azteca sediment bioassays.

The significant differences between bioassay and field conditions, along with the habitat preferences of H. azteca, suggest that behavior under laboratory conditions may artificially increase exposure to sediment-associated contaminants and physiological responses when compared with environmentally relevant exposures. Because of this, results of standard sediment bioassays may be most appropriately used to distinguish nontoxic field sediments (e.g., a first-tier screening tool) rather than to identify toxic sediments. Pyrethroid toxicity to field populations likely cannot be satisfactorily determined through a sediment-only H. azteca bioassay.

Life history traits

Life history attributes such as age, life stage at time of exposure, and nutritional state have been demonstrated to affect sensitivity of benthic invertebrates to pyrethroid exposure. For example, grass shrimp (Palaemonetes pugio) embryos are 130 times less sensitive to phenothrin than are larvae or adults (Key et al. 2009). Unfed flagfish (Jordanella floridae) are 8 times as sensitive as fed fish to permethrin at 2 days old, and they are nearly 3 times as sensitive as fed fish at 4 days old (Holdway and Dixon 1988). Similarly, unfed white suckers (Catostomus commersonii) are 92, 10, and 21 times as sensitive as fed white suckers at age 13, 20, and 26 days (Holdway and Dixon 1988) and Japanese medaka (Oryzias latipes) embryos are less susceptible to cypermethrin exposure than were the larval or juvenile life stages (Kim et al. 2008). Experimentation with H. azteca indicated that acute toxicity values could vary by as much as 50% between different age classes, but that age-related sensitivity was inconsistent among various toxicant classes (Collyard et al. 1994). Consequently, it is likely that pyrethroid sensitivity fluctuates with age class, but it is unclear which life stage would exhibit the most sensitivity. It may be that lack of an appropriate food source is confounding age-related sensitivities (Ankley et al. 1994).

The source of the exposed organisms can also influence metabolic capacity for and sensitivity to insecticides. For example, a laboratory-derived Chironomus riparius population was demonstrated to be 13 to 250 times as sensitive to a range of insecticides as a field-collected C. riparius population (Hoffman and Fisher 1994). In general, this suggests that ecological and biological conditions alter the susceptibility of field organisms in ways that will not be predicted by H. azteca laboratory models.


Complex natural systems result in a nonhomogenous distribution of pyrethroid concentrations, allowing for the potential establishment of refugia, and indicating that localized shifts in resident benthic fauna may occur, but that extensive, long-term depletion of invertebrate populations or alteration of ecological services is unlikely. In contrast to sediment-rich agriculturally influenced streams, suburban and urban streams can contain high levels of allochthonous organic matter and vegetation, producing a highly variable benthic habitat that provides a rich and energetically preferable refuge for H. azteca, in particular. Consequently, the distribution and effects of pyrethroids resulting from urban and suburban residential use is not likely to cause widespread impacts on the structure or function of aquatic ecosystems, despite preliminary toxicity results from H. azteca sediment bioassays. A significant body of literature shows that pyrethroid toxicity is altered by adsorption to organic carbon content in surface sediments and waters, and that, while adsorbed, it is effectively sequestered and therefore does not pose a significant risk to nontarget organisms, even when ingested.

When the dispersion of particulate-associated pyrethroid contamination was studied in a model California stream system, distribution and distance traveled was highly dependent on particle size (Gan et al. 2005). Also, the adsorption coefficient for highly mobile pyrethroid-contaminated particulates was high, indicating very low bioavailability while sorbed. This may partially explain the noted occurrences of field sediment with high pyrethroid concentrations and low H. azteca toxicity and clarify incidences in which the toxic unit method for estimating toxicities of pyrethroid mixtures significantly overestimated risk to H. azteca (You et al. 2008). Because of the site-specific variables that can ameliorate aspects of pyrethroid toxicity, the distribution of toxic pyrethroid-contaminated sediments would be expected to be patchy and uneven in suburban and urban stream systems. In fact, Weston et al. (2005) noted that populations of H. azteca persisted in stretches of a suburban Sacramento, California, stream system in which areas of sediments contained toxic amounts of pyrethroids.

Species sensitivity distributions for aqueous bifenthrin and permethrin exposures demonstrate that under laboratory conditions H. azteca is 1 to 2 orders of magnitude more sensitive than other aquatic invertebrate and fish species. Further, a species sensitivity distribution based on sediment-associated bifenthrin toxicity also indicated that H. azteca were approximately an order of magnitude more sensitive than other benthic invertebrates, and in all cases, H. azteca toxicity values were lower than those for all other taxa. Evidence from mesocosm studies also demonstrated the unique susceptibility of amphipods, because lower pyrethroid exposures elicited greater effects in these organisms than in any other taxa monitored. Most significantly, recent field research in San Diego County confirmed that no link could be established between sediment-associated pyrethroid toxicity to H. azteca (as determined in standard laboratory bioassays) and any indices of macroinvertebrate community health. This leads us to conclude that although the H. azteca laboratory bioassay is a useful screening tool, data garnered with the toxicity test do not provide constructive information concerning greater effects on community structure and abundance.

In general, extrapolating the results of laboratory toxicity tests to predict effects in the field is difficult, as variables such as population age structure, behavioral avoidance, and immigration can offset the impacts of exposure to toxicants. Invertebrate populations with stable age distributions were demonstrated to be significantly less susceptible to pesticide exposures than populations consisting of a single age (Stark and Banken 1999). Further, acute toxicity tests were shown to be a poor predictor of long-term effects on population growth and structure (Stark 2005). In laboratory bioassays, the homogenous environment is a limitation for test organisms; given a selection of habitats, H. azteca is more likely to select macrophyte or leaf material substrate that will effectively shelter the amphipods from the highest pyrethroid exposures (Cooper 1965, Wang et al. 2004, Maul et al. 2008). Additionally, for hydrophobic compounds in particular, environmental conditions are a significant driver of bioavailability.

In conclusion, available evidence suggests that H. azteca bioassays can be effectively employed as screening tools for pyrethroid contamination, but that no information concerning the effects on the larger ecosystem can be concluded from bioassay results. The complexity of natural systems and site-specific variables alter pyrethroid bioavailability, degradation, and toxicity on a microhabitat scale, resulting in an irregular distribution of pyrethroid concentrations in urban and suburban stream habitats. Further, H. azteca biological data indicate that the organisms are unlikely to be heavily exposed to contaminated sediments under field conditions, while toxicological data suggest that under laboratory conditions this species exhibits a uniquely high sensitivity not representative of community-level effects. Given these concerns, we conclude that the H. azteca model can be appropriately utilized as a screening tool to identify those sediments not toxic to benthic organisms but cannot accurately predict in-field sediment toxicity.