Applying Species-Sensitivity Distributions in Ecological Risk Assessment: Assumptions of Distribution Type and Sufficientnumbers of Species

Abstract

Species-sensitivity distribution methods assemble single-species toxicity data to predict hazardous concentrations (HCps) affecting a certain percentage (p) of species in a community. The fit of the lognormal model and required number of individual species values were evaluated with 30 published data sets. The increasingly common assumption that a lognormal model best fits these data was not supported. Fifteen data sets failed a formal test of conformity to a lognormal distribution; other distributions often provided better fit to the data than the lognormal distribution. An alternate bootstrap method provided accurate estimates of HCp without the assumption of a specific distribution. Approximate sample sizes producing HC5 estimates with minimal variance ranged from 15 to 55, and had a median of 30 species-sensitivity values. These sample sizes are higher than those suggested in recent regulatory documents. A bootstrap method is recommended that predicts with 95% confidence the concentration affecting 5% or fewer species.

10th Anniversary Critical Review: The tissue-residue approach for toxicity assessment: concepts, issues, application, and recommendations

“…Often, the lower 95% CI for the HC 05 point estimate is selected as the concentration for protection. 70,71 If the data are lognormally distributed and an SSD approach is used, appropriate algorithms for these percentiles must be used.…”

The Consequences of Nonrandomness in Species-Sensitivity in Relation to Functional Traits for Ecosystem-Level Effects of Chemicals

“…These correlations between species’ sensitivity and functional traits can cause ecosystem-level effects of chemicals to differ strongly between systems with similar species sensitivity distributions, affecting the representativeness of species-level effects for ecosystem-level effects. Methods that infer ecosystem-level effects from single-species bioassays, such as the SSD approach, have often been criticized for ignoring potential effects of species interactions. ,,,,, Here, it is shown that the correlation between species sensitivities and functional traits can partly account for this lack of information (Table and SI Table 5). In addition, inferring ecosystem-level effects requires measuring species-level effects that are relevant to both the aggregated ecosystem function and exposure scenario under assessment (Figures and ).…”

“…Most risk assessment procedures worldwide, however, still rely on single-species bioassays. − Hence the reliability of the ecosystem-level effects that are inferred from the species-level effects measured in these bioassays, strongly depends on the assumptions made on how species-level and ecosystem-level effects are linked. − Environmental risk assessment procedures generally need to balance pragmatism and environmental realism due to time or monetary constraints. , Therefore, simple theoretical models, such as the cumulative species sensitivity distribution (SSD), have increasingly been used for both regulatory and scientific purposes since the 1990s. − SSDs are obtained by fitting a statistical distribution, generally a log-normal or log–logistic distribution, to the single-species toxicity data. , Environmental threshold concentrations are subsequently derived based on the adversely affected fraction of species that is considered acceptable, i.e. without putting the structure and functions of ecosystems at risk (e.g., 5% in EU legislation). ,, The SSD approach hence requires that the species from which it is derived are representative for all species in the system, and that a certain degree of functional redundancy between species exists so that ecosystem-level effects do not exceed species-level effects. ,, A variety of statistical and ecological effects can cause violations of these assumptions, which can consequently cause observed effects on ecosystem structure and function to deviate from those expected based on single species bioassays. ,,− …”

“…Changes in ecosystem structure by chemical exposure can arise through both direct effects on reproduction or survival rates, as well as through indirect effects by density changes in other species as a result of species interactions. ,, The direct effects measured in single-species bioassays thereby allow inferring the concentration of the chemical at which species start to become affected, and changes in ecosystem structure thus start to arise. Still, the correct inference of direct species-level effects in the system requires that the set of species exposed in bioassays in lab conditions is a sufficiently large, random sample of the species present in the ecosystem. ,,, If not, changes in ecosystem structure may start to occur at lower or higher chemical concentrations than expected. In addition, species-level effects observed in bioassays can also be unrepresentative because of differences in sensitivity between lab and field conditions. , However, due to indirect effects through species interactions, effects on ecosystem structure can exceed the direct species-level effects measured in bioassays. ,, The magnitude of indirect effects, and thus the overall change in ecosystem structure due to chemicals exposure, thereby depends on the type and strength of species interactions within the system. , Positive interactions (e.g., facilitation or mutualism) cause direct negative effects on one species to result in additional indirect negative effects on another species.…”

Risk assessment of butyltins based on a fugacity-based food web bioaccumulation model in the Jincheng Bay mariculture area: II. Risk assessment

“…† If multiple toxicity values were available for a single species, the geometric mean was used. 15 Sufficient toxicity data of MBT were not available, because of its low toxicity potential, and its risk to marine species was thus not quantied here.…”