Mind the Exposure Gaps—Modeling Chemical Transport in Sediment Toxicity Tests

Abstract

Chemical exposure in flow-through sediment toxicity tests can vary in time, between pore and overlying water, and amid free and bound states, complicating the link between toxicity and observable concentrations such as free pore (C free,pore), free overlying (C free,over), or the corresponding dissolved concentrations (C diss, free + bound to dissolved organic carbon, DOC). We introduce a numerical model that describes the desorption from sediments to pore water, diffusion through pores and the sediment–water boundary, DOC-mediated transport, and mixing in and outflow from overlying water. The model explained both the experimentally measured gap between C free,over and C free,pore and the continuous decrease in overlying C diss. Spatially resolved modeling suggested a steep concentration gradient present in the upper millimeter of the sediment due to slow chemical diffusion in sediment pores and fast outflux from the overlying water. In contrast to continuous decrease in overlying C diss expected for any chemical, C free,over of highly hydrophobic chemicals was kept relatively constant following desorption from DOC, a mechanism comparable to passive dosing. Our mechanistic analyses emphasize that exposure will depend on the chemical’s hydrophobicity, the test organism habitat and uptake of bound chemicals, and the properties of sediment components, including DOC. The model can help to re-evaluate existing toxicity data, optimize experimental setups, and extrapolate laboratory toxicity data to field exposure.

Comparing 10‐ and 28‐Day Sediment Toxicity and Bioaccumulation of Fluoranthene in Hyalella azteca Using Passive Sampling Techniques

“…We measured the C diss in overlying water (C diss,over ), C diss,pore , C free,over,L , C free,over,U , and C free,pore of fluoranthene using spiked-sediment exposure tests (Figure 2). The following relationship was found: C diss,pore ≫ C free,pore > C diss,over ≈ C free,over,L , which was generally consistent with the findings of our previous studies using several HOCs in the same sediment toxicity test system (Fischer et al, 2021;Hiki et al, 2021). Values of C diss and C free in porewater were higher than those in overlying water by factors of 42 ± 20 and 5.5 ± 0.6, respectively, on day 10 and by factors of 76 ± 73 and 5.1 ± 0.8, respectively, on day 28.…”

“…Note that C diss,pore was highly variable (Figure 2), which was possibly due to the incomplete removal of fluoranthene bound to colloidal particles by ultracentrifugation. The difference in C free between pore and overlying water was comparable with a factor of 5, which was estimated using our mathematical kinetic model with DOC-mediated transport (Fischer et al, 2021) for a chemical with the log of the K OC of 5 (log K OC of fluoranthene 5.3 ± 0.1). These results indicate that the previously proposed mechanisms of mass distribution within the test system apply to fluoranthene as well; thus, the dilution rate achieved by the semi-flow-through waterexchange system surpasses the diffusive influx from sediment to overlying water and suppresses the concentration in the overlying water in comparison with the porewater.…”

“…We compared C free measured using fibers at two different positions in overlying water (Figure 1) and found that C free,over,L was on average 1.4 times that of C free,over,U for any tested concentration. Considering the result from our previous study, which demonstrated that the overlying water was well mixed from 1 cm above the sediment surface to the water surface (Fischer et al, 2021; Hiki et al, 2021), C free,over,L was probably elevated because the PDMS fiber absorbed the chemical in the unstirred layer of <1 cm near the sediment–water interface, where C free was substantially higher than that in the bulk overlying water. The difference between C free,over,L and C free,over,U was reasonable considering the large C free difference between the porewater and overlying water (i.e., a factor of >5) and the small part of the PDMS fiber exposed to the unstirred layer (i.e., <1 cm of the whole 7‐cm fiber).…”

“…Despite the temporal fluctuation, the time‐weighted mean concentration of C diss,over differed by up to 35% between the initial 10 days and the entire 28 days. These temporal trends were reproduced using the mathematical model developed in our previous study (Fischer et al, 2021). The model showed that C diss,over continuously decreased over the entire 28 days but almost reached a plateau after approximately day 10 (Supporting Information, Figure S4).…”

“…Because sediment–water test systems are highly complex and benthic organisms are affected via multiple exposure pathways, understanding the spatial distribution and speciation of the test chemical is important for the assessment of its bioavailability, bioaccumulation, and toxicity (Dorn et al, 2021; Driscoll et al, 1997). In our previous studies (Fischer et al, 2021; Hiki et al, 2021), we demonstrated that there was a large gap between the overlying water and porewater concentrations of hydrophobic organic chemicals (HOCs), which occurred as a result of the continuous exchange of overlying water and slow mass transfer of the chemical at the sediment–water interface. Moreover, the overlying water concentration was not constant over a test period of 10 days (Hiki et al, 2021).…”