Integrating bioavailability measurements in persistence testing of partially biodegradable organic chemicals in soil

Overestimation of risk is one of the main problems in environmental risk assessments if only total concentration of organic pollutants is considered. In this study, we integrated bioavailability measurements into persistence testing of pollutants in soil to show that it is the key to have a more realistic environmental risk assessment (ERA). To this integration, two standardized methods were used: OECD 307, as persistence test, and ISO 16751: 2020, to bioavailability measurements based on 20 h extractions with a strong adsorbent (Tenax), using pyrene and carbamazepine as model test substances. Because the ISO method was initially designed for nonpolar compounds with log K_ow > 3, a slight adaptation was necessary for carbamazepine (log K_ow = 2.7), assuming this also as an extension of the applicability domain of the method. During the biodegradation of these compounds, the mineralization extents did not exceed 4 %, giving rise to transformation products. Therefore, the bioavailability measurements covered both the parent compound and the metabolites produced. In the case of pyrene, the partial transformation carried out by a specialized microbial inoculum accounted for, respectively, 32 % or 40 % of the initial concentration (4 mg kg^-1) in unamended or compost-amended soil. Only 1 % was present as hydrophilic transformation products that were not trapped by Tenax, but partitioned into the water. The nonbioavailable residue increased in both soils after biodegradation. The distribution of chemicals in the different phases of the system was the key to assess more realistically the shifts in bioavailability during persistence testing. The same procedure was carried out for carbamazepine where an additional desorption study showed a slower desorption rate of the parent compound after incubation. In this case, 25 % of transformation products was mobilized to the aqueous phase. Our results show that bioavailability measurements provide valuable information when integrated in persistence testing, therefore contributing to realism in prospective ERA scenarios.