Systematic Consideration of Parameter Uncertainty and Variability in Probabilistic Species Sensitivity Distributions

Abstract

The calculation of a species sensitivity distribution (SSD) is a commonly accepted approach to derive the predicted no‐effect concentration (PNEC) of a substance in the context of environmental risk assessment. The SSD approach usually is data demanding and incorporates a large number of ecotoxicological values from different experimental studies. The probabilistic SSD (PSSD) approach is able to fully consider the variability between different exposure conditions and material types, which is of great importance when constructing an SSD for any chemical, especially for nanomaterials. The aim of our work was to further develop the PSSD approach by implementing methods to better consider the uncertainty and variability of the input data. We incorporated probabilistic elements to consider the uncertainty associated with uncertainty factors by using probability distributions instead of single values. The new PSSD method (named “PSSD+”) computes 10 000 PSSDs based on a Monte Carlo routine. For each PSSD calculated, the hazardous concentration for 5% of species (HC5) was extracted to provide a PNEC distribution based on all data available and their associated uncertainty. The PSSD+ approach also includes the option to consider a species weighting according to a typically constituted biome. We applied this PSSD+ approach to a previously published data set on C nanotubes and Ag nanoparticles. The evaluation of the uncertainty factor distributions and species weighting have shown that the proposed PSSD method is robust with respect to the calculation of the PNEC value. Furthermore, we demonstrated that the PSSD+ can handle both small and more comprehensive data sets because the PNEC distributions are a close representation of the data available. Finally, the sensitivity testing toward data set variations showed that the maximum variation of the mean PNEC was of a factor of about 2, so that the method is relatively insensitive to missing data points as long as the most sensitive species is included. Integr Environ Assess Manag 2020;16:211–222. © 2019 SETAC

A Meta‐analysis of Ecotoxicological Hazard Data for Nanoplastics in Marine and Freshwater Systems

“…The SSDs are a widely acknowledged method used in environmental protection, assessment, and management practices (Posthuma et al 2019). The present study performed the hazard assessment based on the PSSD method developed by Gottschalk and Nowack (2013), which was further refined by Wigger et al (2020) to include uncertainties and interlaboratory variation (PSSD+ method as described in the Data harmonization section). The PSSD+ method first calculates one probability distribution for each species for all available data points, and then all probability distributions are integrated into a single PSSD by a Monte‐Carlo routine (Wigger et al 2020).…”

Form‐Specific and Probabilistic Environmental Risk Assessment of 3 Engineered Nanomaterials (Nano‐Ag, Nano‐TiO2, and Nano‐ZnO) in European Freshwaters

“…For the hazard assessment, we built an ecotoxicological data set for each of the different forms of the three ENMs and calculated probabilistic species sensitivity distributions (SSDs) using the PSSD+ method (Wigger et al, 2020b). Finally, the environmental risks were determined by calculating the risk characterization ratio for each released form.…”

“…According to Wigger et al (2020b), UF t considers the exposure time of the experiment and thus converts acute toxicity into chronic toxicity. UF dd transforms a specific dose-descriptor to a NOEC value.…”

“…UF dd transforms a specific dose-descriptor to a NOEC value. The UF t values provided by Wigger et al (2020b) were complemented by including the types of species not considered in the previous study: amphibian, bacteria, hydroid and ciliated protozoan (see SI, Chapter 2).…”

Prospective environmental risk assessment of nanocellulose for Europe

“…47 We ran 10 000 iterations, resulting in a PNEC distribution rather than a single value. Further detail regarding the methodology can be found in Coll et al and Wigger et al 27,48…”