The goals of protecting natural ecosystems are typically defined with reference to population size and viability, or to measures of community structure, composition, and function. In the case of ecological risk assessment (ERA), however, such protection arises usually by extrapolation from laboratory studies that depend, in turn, on single species or highly simplified communities of test organisms that are assumed to represent field responses (Hommen et al. 2010).
Laboratory investigations benefit from controlled conditions, although their realism is clearly questionable. In contrast, field surveys can provide realistic descriptions of biotic communities or functions. Although they lack the close control of laboratory investigations, field surveys deserve a critical or at least complementary role in ERAs (Iwasaki and Ormerod 2012). It is thus surprising that field approaches to ERA are infrequent in the literature by comparison with laboratory and modeling studies.
In this Learned Discourse (LD), we highlight the usefulness of field approaches for assessing the relevance of laboratory results to natural populations or communities. We do this relative to recent studies on metals and riverine macroinvertebrates. We also highlight further needs to improve understanding of field studies. This LD has wider relevance to natural communities exposed to stressors than just metals and stream macroinvertebrates.
The estimation of the “safe” concentrations of substances below which unacceptable effects will not occur, such as the predicted no effect concentration, is a critical step in ERAs. Generally, such safe concentrations are derived from laboratory toxicity tests or/and their statistical extrapolation to community-level consequences (e.g., species sensitivity distributions). Assessment of field relevance of laboratory estimates is a necessary step in the ERA process (Crane et al. 2007). River macroinvertebrates are frequently used in routine biomonitoring and field surveys to accomplish such needs. Although the impacts of metals on river macroinvertebrates are well documented, until recently such field studies had not necessarily provided useful information on safe concentrations largely due to high natural variability. Two approaches to deriving safe concentrations are possible, however, depending on the quantity and quality of data obtained.
The first approach estimates safe concentrations of individual metals using quantile regression analysis (Cade and Noon 2003) using large field data sets (e.g., >250 sites; Crane et al. 2007). Iwasaki and Ormerod (2012) used this approach to suggest the likely ranges of environmental quality standards for metals using field data on macroinvertebrates. The second approach uses smaller data sets and reduces variability in those field data by sampling exposed and reference sites with similar physicochemical conditions and carefully considering the ecology of target organisms. Using such methodology, Iwasaki et al. (2011) suggested that the Japanese water quality standard for Zn is likely to be overprotective for macroinvertebrate diversity.
Although these 2 approaches are useful, improved interpretation of the field effects of chemical substances is needed to provide more reliable information to ERA. There are 2 requirements in this regard: identification of the appropriate predictors of field effects of metals; and, characterization of the most appropriate spatiotemporal scales at which invertebrates respond to metals.
There are limitations to using metal concentrations in ambient water as an effects predictor. Xie and Buchwalter (2011) noted “Because we know so little about the true toxicity of trace metals to aquatic insects, it is impossible to determine whether current water quality criteria are protective.” Improved mechanistic understanding of the effects of metals in real field environments is required. Consideration of metal bioavailability and multiple-metal effects would be valuable, although evidence showing how the effects of metal bioavailability should be incorporated into models of responses of common lotic macroinvertebrates (e.g., mayflies and caddisflies) remains limited (Iwasaki et al. 2011). A few studies address these issues using field data. Schmidt et al. (2011), for example, used the amount of metal bioaccumulation observed in aquatic insects to predict effects on invertebrate populations and communities in the field—the tissue residue approach.
Equally importantly, macroinvertebrate responses at different spatiotemporal scales must be considered, including: behavioral responses of invertebrate individuals at short time scales, effects of larval exposure on adult insects and their reproduction, and evolutionary responses of populations. For example, at one extreme, macroinvertebrate populations may have adapted to metal contamination over many generations, particularly in extensively contaminated areas. At the other extreme is the long-standing problem of separately assessing the effects of metals in short, brief exposures to high concentrations during hydrological events versus more prolonged low-level exposure. Macroinvertebrates can drift downstream in response to metal exposure or also drift from uncontaminated to contaminated sites. Both processes would lead to a mismatch between faunal composition and base-flow metal concentrations.
Although there is still a need to pay careful attention to the interpretation of field data, we believe field studies provide crucial evidence to evaluate the relevance of laboratory-based estimates in natural environments. This is an unquestionably important step in providing useful and relevant ERAs. Further research is strongly encouraged to increase the value of field data to ERAs.