A User-Friendly Kinetic Model Incorporating Regression Models for Estimating Pesticide Accumulation in Diverse Earthworm Species Across Varied Soils

Abstract

Existing models for estimating pesticide bioconcentration in earthworms exhibit limited applicability across different chemicals, soils and species which restricts their potential as an alternative, intermediate tier for risk assessment. We used experimental data from uptake and elimination studies using three earthworm species (Lumbricus terrestris, Aporrectodea caliginosa, Eisenia fetida), five pesticides (log K_ow 1.69-6.63) and five soils (organic matter content = 0.972-39.9 wt %) to produce a first-order kinetic accumulation model. Model applicability was evaluated against a data set of 402 internal earthworm concentrations reported from the literature including chemical and soil properties outside the data range used to produce the model. Our models accurately predict body load using either porewater or bulk soil concentrations, with at least 93.5 and 84.3% of body load predictions within a factor of 10 and 5 of corresponding observed values, respectively. This suggests that there is no need to distinguish between porewater and soil exposure routes or to consider different uptake and elimination pathways when predicting earthworm bioconcentration. Our new model not only outperformed existing models in characterizing earthworm exposure to pesticides in soil, but it could also be integrated with models that account for earthworm movement and fluctuating soil pesticide concentrations due to degradation and transport.

Keywords: dermal; elimination; gut; lipid; porewater; risk assessment; uptake.

Figure 1

Evaluation of the predictive performance of our new kinetic model based on porewater concentrations against the independent data for phenanthrene, pyrene, and other organic compounds. “Others” (in green) is the full evaluation data set excluding data for phenanthrene and pyrene. “p,p (3)” and “p,p (7–21)” and “phenanthrene, pyrene (Day 3)” and “phenanthrene, pyrene (Day 7–Day 21)” are the evaluation data sets for phenanthrene and pyrene measured in the uptake phase at Day 3 (red triangles) and for Days 7 to 21 (blue squares), respectively. The central black solid line represents a perfect model fit (1:1 line). The gray dotted, gray dashed, and outer solid lines represent a 3-fold, 5-fold, and 10-fold difference between the predicted and observed values, respectively.

Figure 2

Evaluation of the predictive performance of our new kinetic model based on porewater concentrations against the independent data at different exposure time periods. Data for phenanthrene and pyrene are excluded. “1–6”, “7–20”, and “21–42” and “Day 1–Day 6”, “Day 7–Day 20”, and “Day 21–Day 42” are the subsets of the evaluation data set for uptake from Days 1 to 6 (green circles), Days 7 to 20 (red triangles), and Days 21 to 42 (blue squares), respectively. The central black solid line represents a perfect model fit (1:1 line). The gray dotted, gray dashed, and outer solid lines represent a 3-fold, 5-fold, and 10-fold difference between the predicted and observed values, respectively.

Figure 3

Evaluation of the predictive performance of our new kinetic model based on porewater concentrations against the independent data for different earthworm species. Data for phenanthrene and pyrene are excluded. “EF/EA” (green circles), “LT” (red triangles), and “AC” (blue squares) represent data for the earthworm species E. fetida, E. andrei, L. terrestris, and A. caliginosa, respectively. The central black solid line represents a perfect model fit (1:1 line). The gray dotted, gray dashed, and outer solid lines represent a 3-fold, 5-fold, and 10-fold difference between the predicted and observed values, respectively.