A Review of Mechanistic Models for Predicting Adverse Effects in Sediment Toxicity Testing

Abstract

Since recognizing the importance of bioavailability for understanding the toxicity of chemicals in sediments, mechanistic modeling has advanced over the last 40 years by building better tools for estimating exposure and making predictions of probable adverse effects. Our review provides an up-to-date survey of the status of mechanistic modeling in contaminated sediment toxicity assessments. Relative to exposure, advances have been most substantial for non-ionic organic contaminants (NOCs) and divalent cationic metals, with several equilibrium partitioning-based (Eq-P) models having been developed. This has included the use of Abraham equations to estimate partition coefficients for environmental media. As a result of the complexity of their partitioning behavior, progress has been less substantial for ionic/polar organic contaminants. When the EqP-based estimates of exposure and bioavailability are combined with water-only effects measurements, predictions of sediment toxicity can be successfully made for NOCs and selected metals. Both species sensitivity distributions and toxicokinetic and toxicodynamic models are increasingly being applied to better predict contaminated sediment toxicity. Furthermore, for some classes of contaminants, such as polycyclic aromatic hydrocarbons, adverse effects can be modeled as mixtures, making the models useful in real-world applications, where contaminants seldomly occur individually. Despite the impressive advances in the development and application of mechanistic models to predict sediment toxicity, several critical research needs remain to be addressed. These needs and others represent the next frontier in the continuing development and application of mechanistic models for informing environmental scientists, managers, and decisions makers of the risks associated with contaminated sediments. Environ Toxicol Chem 2024;43:1778-1794. © 2023 SETAC. This article has been contributed to by U.S. Government employees and their work is in the public domain in the USA.

Keywords: Acute and sublethal toxicity; Cationic divalent metals; Mechanistic modeling; Passive sampling; Polar and nonpolar organic chemicals; Sediment toxicity; Species sensitivity distributions.

FIGURE 1:

Cartoon illustrating the distribution of non-ionic organic contaminants between sediment particle (i.e., partitioned to the organic carbon) and interstitial water (i.e., freely dissolved) phases. In this example, the contaminants are three-dimensional representations of the pesticide DDT (available to the public domain via Wikimedia Commons).

FIGURE 2:

Components of a basic conceptual model for estimating sediment toxicity including an exposure component for predicting the partitioning and resulting bioavailability of the contaminant(s) combined with an effects component dedicated to providing toxicity data. Combined the components result in an estimate of sediment toxicity. The components are similar, if slightly more complicated, for ionic and metallic contaminants. Recent enhancements to these components include incorporation of passive sampler data to provide concentrations of bioavailable non-ionic organic contaminants along with estimates of mixture toxicity.

FIGURE 3:

Visualization of phases and partitioning of non-ionic organic contaminants used to predict exposure to aquatic organisms (e.g., a benthic amphipod) in sediments. The colloidal organic carbon phase and related partitioning are not explicitly illustrated in this simple conceptual model. Amphipod image from Shutterstock.

FIGURE 4:

Comparison of the fifth percentile values from species sensitivity distributions across individual modes of action on a critical target site basis. The normalized $Z$ score is used to allow compression of the magnitude of the relative toxicities of the different modes of action (on the x axis). Based on Table 4 from Boone & Di Toro (2019).

FIGURE 5:

Visualization of phases and partitioning of ionic or polar organic contaminants used to predict exposure to aquatic organisms (e.g., a benthic amphipod) in sediments. The colloidal organic carbon phase and related partitioning is not explicitly illustrated in this simple conceptual model. Amphipod image from Shutterstock.

FIGURE 6:

Visualization of phases and primary partitioning in the (A) simultaneously extracted metal-acid volatile sulfide and (B) biotic ligand models used to predict cationic metal exposure and bioaccumulation to aquatic organisms in sediments. The larger the arrows between the phases, the greater the exposure relative to other phases. Amphipod image from Shutterstock.

FIGURE 7:

Hypothetical species sensitivity distribution for a sediment contaminant and derivation of a hazard concentration 5% (HC05) value. For simplicity, adversely affected organisms are grouped by family or class.