Species sensitivity distribution model

Reliable chronic toxicity data were available for 21 species of plants (13 phytoplankton and 8 macrophytes) and 15 species of animals. The species sensitivity distributions (SSDs) for atrazine chronic toxicity (no observed effect concentrations [NOECs]) to plants and animals are shown in Figure 4.4. A log-normal distribution model was fitted to each SSD by least-squares regression. 

The exposure distribution and species sensitivity distributions were integrated to generate risk curves for chronic effects. From the 504 000 values in the exposure exceedence curve, annual maximum concentrations corresponding to each 0.5th percentile were determined. The percentage of plant or animal species whose chronic NOEC would be exceeded at each of these concentrations was calculated from the log-normal SSD model. The percentage of plant or animal species affected at each exposure exceedence percentile was plotted as shown in Figure 4.5. 

SimpleBox was created as a research tool in environmental risk assessment. Simple-Box (Brandes et al. 1996) is implemented in the regulatory European Union System for the Evaluation of Substances (EUSES) models (Vermeire et al. 1997) that are used for risk assessment of new and existing chemicals. Dedicated SimpleBox 1.0 applications have been used for integrating environmental quality criteria for air, water, and soil in The Netherlands. Spreadsheet versions of SimpleBox 2.0 are used for multi-media chemical fate modeling by scientists at universities and research institutes in various countries. SimpleBox models exposure concentrations in the environmental media. In addition to exposure concentrations, SimpleBox provides output at the level of toxic pressure on ecosystems by calculating potentially affected fractions (PAF) on the basis of species sensitivity distribution (SSD) calculus (see Chapter 4). 

An analysis of regularities observed in species sensitivity distributions (SSD) fitted on acute and chronic aquatic toxicity data for a large number of organic and inorganic toxicants is provided by De Zwart (2002). The log-logistic sensitivity model he used is characterized by the parameter a, which is the mean of the observed loglO-transformed L(E)C50 or NOEC values over a variety of test species, and /3, a scale parameter proportional to the standard deviation of the loglO-transformed... 

It is noteworthy that comparisons of existing assessment schemes reveal dissimilarities in the use of extrapolation methods and their input data between different jurisdictions and between prospective and retrospective assessment schemes. This is clearly apparent from, for example, a set of scientific comparisons of 5% hazardous concentration (HC5) values for different substances. Absolute HC5 values and their lower confidence values were different among the different statistical models that can be used to describe a species sensitivity distribution (SSD Wheeler et al. 2002a). As different countries have made different choices in the prescribed modeling by SSDs (regarding data quality, preferred model, etc.), it is clear that different jurisdictions may have different environmental quality criteria for the same substance. Considering the science, the absolute values could be the same in view of the fact that the assessment problem, the available extrapolation methods, and the possible set of input data are (scientifically) similar across jurisdictions. When it is possible, however, to look at the confidence intervals, the numerical differences resulting from different details in method choice become smaller because confidence intervals show overlap. 

Wheeler JR, Grist EPM, Leung KMY, Morritt D, Crane M. 2002a. Species sensitivity distributions data and model choice. Mar Poll Bull 45 192-202. 

Figure 5.10 Difference in the dose-effect models for humans and species assemblages (species sensitivity distribution [SSD], right). Threshold-type curves are used for many compounds it is assumed that below a certain daily intake there will be no effects. Nonthreshold chemicals (i.e. certain types of carcinogens) lead to increased probability of cancer, and for this a linear model is assumed in the relevant concentration range. Species sensitivities are assumed to follow a non-linear curve (the SSD), relating the exposure to the fraction of species affected, with a maximum of 100% of the species affected.

Maltby et al. (2002) and Van den Brink et al. (2006a) compared SSDs based on acute and chronic laboratory toxicity data for aquatic test species exposed to pesticides. The SSDs were constructed with toxicity data for the most sensitive taxonomic group, because of the specific toxic mode of action of the pesticides selected. The SSDs were used to calculate the hazardous concentration to 5% of the species (HC5) by means of a log-normal distribution model, and comparisons were performed for 2 insecticides and 7 herbicides (Table 6.4). The log-normal model did not fit the diuron (herbicide) short-term L(E)C50 data or the atrazine (herbicide) long-term NOEC data. Consequently, the L(E)C50 HC5 value for diuron and the NOEC HC5 value for atrazine should be interpreted with caution, as well as their acute HC5-chronic... 

One can differentiate between three types of transformation products of environmental pollutants. First, environmental pollutants can be metabolized during the toxicokinetic phase of uptake/metabolism/distribution/elimination in organisms (Table 1). Here, the observed effect is actually due to the combined effect of different metabolites. Taking these transformation reactions into account will help to understand mechanisms of toxicity, species sensitivity differences, and time dependency of effects. Lee and Landrum [8,9] developed a model to describe the mixture effects of PAH and their metabolites in Hyalella azteca. This combined toxicokinetic/toxicodynamic models convincingly demonstrated the importance of accounting for metabolite formation and how different mixture toxicity concepts can be incorporated into such models. 

Figure 9.4 Risk assessment for an aquatic environment based on a probabilistic procedure into which the concept of varying sensitivity in multispecies communities is incorporated (Nendza, Volmer and Klein, 1990). Exposure and effects are determined separately from experimental or, if not available, QSAR data. Physico-chemical data and information on bioaccumulation and biotransformation are the input for computer simulations of transport and distribution processes that estimate the concentrations of a potential contaminant in a selected river scenario, using, for example, the EXAMS model (Bums, Cline and Lassiter, 1982). For the effects assessment, the log-normal sensitivity distribution is calculated from ecotoxicological data and the effective concentrations for the most sensitive species are determined. The exposure concentrations and toxicity data are then compared by analysis of variance to give a measure of risk for the environment. Modified from Nendza, Volmer and Klein (1990) with kind permission from Kluwer Academic Publishers, Dordrecht.

---