Advancing the Prediction of PAH Bioaccumulation in Earthworms and Plants for Historically Contaminated Soils Using Chemical In-Vitro Methodologies

Abstract

<p>Contamination of soil with polycyclic aromatic hydrocarbons (PAHs) by a range of anthropogenic activities causes concern for human and environmental health. Understanding PAH bioavailability provides regulatory decisions for contaminated sites with a realistic assessment of risks associated with the contaminants, and as such a range of approaches have been developed for assessment of contaminant bioavailable fractions. Application of these methods to predict PAH bioaccumulation in terrestrial organisms and plants from historically contaminated (> 50 years) soils has not been broadly studied. The broad aim of this thesis was to understand how a range of well-used in-vitro chemical predictive methods vary in the measurement of PAH bioaccessibility in historically contaminated soils and how effectively the methods can predict PAH bioavailability to different receptors (earthworms and a model plant).</p> <p>Six in-vitro chemical predictive methods were applied to determine PAH bioaccessible fractions in four historically contaminated manufactured gas plant (MGP) soils. The methods included depletive non-exhaustive extractions (butanol (BuOH), non-buffered and buffered 2-hydroxypropyl-β-cyclodextrin extractions (HPCD, Buf-HPCD), potassium persulfate oxidation (KPS), solid phase extraction using Tenax resin (Tenax)), and a non-depletive method (polyoxymethylene solid-phase extraction (POM)). The chemical extractions were compared with bioassays using different receptors: three representative earthworm ecotypes (<i>Amynthas</i> sp., <i>Eisenia fetida</i>, and <i>Lumbricus terrestris</i>) and ryegrass (<i>Lolium multiflorum</i>) as the model plant.</p> <p>The PAH bioaccessible fraction measured using different in-vitro methods varied significantly across the methods and also among the soils, with the highest quantity of bioaccessible ∑16 PAHs measured using KPS, and the lowest quantity using POM. Soil properties were generally inconclusive as indicators of bioaccesibility for the historically contaminated soils. Moreover, toxicity values and biodegradation endpoints for the contaminated soils derived using bioaccessibility-biodegradability linear regression models on the PAH extracted using the different methods were highly variable, with significant implications for predicted ecotoxicity or success for bioremediation when applied in risk management.</p> <p>Earthworm bioaccumulation studies revealed distinct differences between epigeic surface dwelling standard test species, <i>E. fetida</i>, and the relevant epi-endogeic and anecic burrowing species, <i>Amynthas</i> sp. and <i>L. terrestris</i>, with accumulated ∑16 PAH concentration following the order <i>Amynthas</i> sp. > <i>L. terrestris</i> > <i>E. fetida</i>. Results indicated that gut uptake was important for PAH partitioning into earthworm lipids, specifically for burrowing species and the desorption-resistant higher molecular weight (HMW) PAHs in historically contaminated soils. Calculated bioconcentration factors (BCFs) indicated stronger relationships with octanol water partition coefficients (KOW) for the burrowing species than for <i>E. fetida</i>. This together with the results of predicted BSAFs from measured pore water concentrations demonstrated that EqPT does not consider differences in uptake between organisms, nor changing bioavailability through different exposure routes (e.g. gut uptake), and caution is required when using KOW values as a proxy for BCF in EqPT especially for <i>E. fetida</i></p> <p>The efficiency of the in-vitro methods for predicting PAH bioaccumulation in the earthworm species was investigated by comparing the method-derived bioaccessibility results with measured bioaccumulation in earthworms directly and using a calculation approach based on EqPT with a combination of different partitioning parameters (KOW values and organic carbon water partition coefficients (K_OC values)) to improve predictions. While Tenax and POM showed the best estimates of bioaccumulation in <i>E. fetida</i> by direct comparison, Buf-HPCD provided the best approximation for <i>Amynthas</i> sp. and <i>L. terrestris</i>. The predictions using the calculation approach depended on the receptor and the partitioning parameters used. Integrating K_OC values derived from historically contaminated soils into calculations improved predictions for <i>E. fetida</i> using all methods (except Tenax), but this was not the case for soil-burrowing species. In general, despite promising results obtained using HPCD-based, Tenax and POM extractions, no one method accurately predicted measured bioaccumulation in MGP soils for the range representative species.</p> <p>In the last part of this project, PAH bioaccumulation in ryegrass (<i>Lolium multiflorum</i>) roots grown in MGP soils was compared with bioaccessibility outcomes of the in-vitro methods both directly and through the EqPT-based calculation approach using different partitioning parameters and different root components (lipid, carbohydrates, and total) regulating PAH sorption. Accumulation of 16 PAHs in <i>L. multiflorum</i> was estimated within a factor of 5 using direct comparison for all bioaccessibility extraction methods, with Buf-HPCD providing the closest estimate. Outcomes using the calculation approach depended on the K_OC, KOW values and root components used in calculations. Using K_OC values derived from historically contaminated soils improved accuracy of prediction of total root accumulation although precision was poor. Comparative assessment for different root components showed that the combined contribution of PAH in lipid and carbohydrate root components overestimated accumulation and a lipid-based approach using generic partitioning parameters provided more accurate and precise approximation of <i>L. multiflorum</i> bioaccumulation in these soils. Generally, Tenax, Buf-HPCD and POM extractions were promising for the prediction of <i>L. multiflorum</i> root accumulation by different approaches</p> <p>This work significantly extends current knowledge for integrating simple chemical extractions into ecological risk assessment frameworks. The results demonstrated clearly that predicting bioavailability needs to be adapted to the target receptor, and the success of including sitespecific K_OC values in modelling depended on organism. Generally, Tenax and POM extractions provided the optimum prtedictions for <i>E. fetida</i> and ryegrass; and HPCD-based extractions for <i>Amynthas</i> sp. and <i>L. terrestris</i>. Expanding the work performed in this project to a wider range of historically contaminated soils from different sites by including suggested predictive methods tested against different plant species (including model crop species), and other earthworm species from each ecotype (such as <i>E. andrei</i> from epigeic and <i>Aporrectodea caliginosa</i> from epi-endogeic ecotypes) would advance the modelling and prediction of PAH bioavailability and improve risk assessment for ecological and human exposure.</p>