Sediment and plant samples were also collected at the PCW inlet, midpoint, and outlet. Sediment samples were collected as described above.All samples were transported on ice prior to laboratory storage at 4 °C. Sediment samples were dried and ground with a mortar and pestle before extraction. Plant samples were thoroughly washed with DI water to remove any attached sediment particles and biofilms. The cleaned plant samples were dried before pulverizing the tissue in the presence of liquid nitrogen using a mortar and pestle. The TSS, sediment, and plant samples were extracted by pressurized fluid extraction on a Dionex ASE 350 using a method similar to Brennan et al. and Maul et al.. Briefly, sample cells were filled with 1:1 dichloromethane:acetone , heated to 100 °C, and extracted at 1500 psi for two 5 min cycles before being flushed with 60% solvent for 1 min. For the TSS samples, the entirety of the solids from each water sample was extracted. Aliquots of the prepared sediment and whole plant samples were extracted and subjected to in-cell cleanup with Cu powder packed between cellulose filters. All extracts were evaporated to approximately 2 mL under a gentle nitrogen stream. Each sample then underwent clean-up on a Florisil cartridge preconditioned with hexane and was eluted with 9:1 hexane:acetone. Cleaned extracts were again evaporated under a gentle stream of nitrogen to approximately 0.5 mL and reconstituted in 1.0 mL hexane for final analysis. Samples were analyzed using an Agilent 6890N/5973N GC/MSD equipped with a 30 m x 0.25 mm x 0.25 µm DB-5MS column.
Additional analytical information is provided in the SI. Following sample quantification,ebb flow tray statistical analyses were performed using SAS® 9.4. Numerous steps were taken to ensure the accuracy and quality of analysis. Instrumental controls during analysis included running a calibration standard every 10 samples, adding 13C4- 15N2-fipronil and d5-bifenthrin as internal standards to each extract, and determining method detection limits using EPA Method 40 CFR Part 136, Appendix B. Calculated MDLs were 2 μg L-1 for all analytes except for fipronil desulfinyl, which had an MDL of 1 μg L-1. In addition, several procedural controls were utilized. Reagent blanks were analyzed with every set of 7 liquid-liquid extraction samples and 10 pressurized fluid extraction samples. Reagent blanks revealed no presence of the target analytes. Matrix spike samples were analyzed to measure extraction efficiencies, which are listed in the SI. In addition, ethiprole and decachlorobiphenyl were added to all samples prior to extraction to assess surrogate recoveries, which were 105 ± 23% and 121 ± 18% for liquid-liquid extraction, respectively, and 108 ± 21% and 116 ± 18% for pressurized fluid extraction, respectively. Additional quality control measures are described in the SI. The concentrations of fiproles and pyrethroids in water samples collected from the PCW are depicted in Figure 3.1. In general, mean whole water concentrations progressively decreased on a spatial level when moving from the inlet to midpoint to outlet for all compounds of interest. In addition, mean outlet concentrations were always lower than mean inlet concentrations. Furthermore, statistically significant differences existed for the majority of inlet and outlet concentration comparisons. When the differences between inlet and outlet concentrations were not statistically significant, inlet concentrations were found at low levels and with relatively high variability.
Since water and the associated TSS is all that enters and leaves the PCW, these findings provide evidence indicating that statistically significant removal of fiproles and pyrethroids occurred as a result of treatment in the PCW. Previous research by agricultural and WWTP effluent-polishing CWs has similarly shown removal of fiproles and pyrethroids. The mean whole water concentrations of fiproles and pyrethroids also exhibited distinct temporal trends during the sampling period. Mean fipronil concentrations peaked in June 2018, gradually decreased until October 2018, and then increased until January 2019. Mean whole water concentrations of fipronil desulfinyl, fipronil sulfide, and fipronil sulfone incrementally increased from June-August 2018, decreased, and then increased until January 2019. The peak mean whole water concentrations for fipronil desulfinyl and fipronil sulfone occurred in January 2019 while the peak for fipronil sulfide occurred in August 2018. Taken together, these trends correspond to high fipronil application rates in the summer months, gradual decrease via degradation of the parent compound and delayed formation of degradates, and increased transport of applied and degraded residues due to rainfall events in the winter months of 2018-2019. The mean whole water concentrations of bifenthrin and cyfluthrin, on the other hand, steadily increased from June-August 2018, gradually decreased until November 2018, and then increased in December 2018 and January 2019. The mean bifenthrin concentrations peaked in August 2018 while mean cyfluthrin concentrations peaked in January 2019. These trends indicate high use of pyrethroids in the summer months, followed by a period of reduced use and transport, and finally an increase due to runoff from the rain events in the winter that could have transported additional residues through the CW and/or caused resuspension of residues from the sediment bed. Fipronil and cyfluthrin were detected at the highest levels in whole water PCW samples, with mean concentrations ranging from 13.5-369 and 13.8-455 ng L-1 , respectively.
Fipronil sulfone and bifenthrin were detected at moderate mean concentrations of ND-32.5 and ND-63.9 ng L-1, respectively. The compounds detected at the lowest levels, fipronil desulfinyl and fipronil sulfide, had mean whole water concentrations of ND-2.48 and ND-4.09 ng L-1, respectively. The water concentrations of fiproles and pyrethroids measured in this study were similar to previous measurements in California in urban runoff. Figure 3.2 shows the sediment concentrations of fiproles and pyrethroids in the PCW. As was the case with mean levels in whole water samples,flood and drain tray mean sediment concentrations for all compounds followed a decreasing spatial trend from the inlet to midpoint to outlet. Since the same spatial pattern was observed for mean water concentrations, this suggests that sediment binding was partially responsible for the dissipation of fiproles and pyrethroids from the water. This finding was to be expected due to the hydrophobicity and strong affinity of these two insecticide classes, particularly for pyrethroids. This was in agreement with previous studies where sediment binding was shown to be an important removal mechanism for pyrethroids in flow-through wetlands receiving agricultural drainage. However, based on this spatial trend alone, it was unclear whether the contaminants were temporarily retained by the sediment and available for partition back into the passing water or if they were subsequently degraded in the sediment. The mean sediment concentrations of fipronil desulfinyl, fipronil sulfide, and bifenthrin followed a similar trend over time. These contaminants generally increased in concentration from June-August 2018, when they reached a peak for the entire study, followed by a decreasing trend until November 2018 before gradually increasing until January 2019. Fipronil in the sediment increased from June-July 2018 when it reached its peak level, decreased until November 2018, and then increased until January 2019. Fipronil sulfone and cyfluthrin peaked in January 2019 and June 2018, respectively, but they followed the same trend as the other compounds. An inspection of Figure 3.1 and Figure 3.2 reveals similar seasonal temporal trends for fiproles and pyrethroids in both whole water and sediment samples. In other words, when analytes were present at high levels in whole water samples, they tended to also be present at high levels in the sediment. This finding, combined with the fact that sediment concentrations did not continuously increase over the duration of the study, indicates that fiproles and pyrethroids were likely actively degraded once partitioned into the sediment phase. Another possibility is that the contaminated sediment particles underwent resuspension and were carried out of the PCW; however, the generally lower whole water concentrations at the outlet relative to the inlet suggested that the contribution of this process was likely negligible. Among the six compounds, fipronil and cyfluthrin were detected at the highest mean sediment concentrations during the study period. Fipronil sulfone and bifenthrin were present at moderate levels ranging from 0.166-4.42 and ND-5.40 ng g-1, respectively.
The lowest mean levels were found for fipronil desulfinyl and fipronil sulfide, at ND-0.740 and ND-0.718 ng g-1, respectively. This pattern was also reflected in the whole water concentrations. These results again suggest that fiprole and pyrethroid residues in PCW water partitioned into the sediment and underwent degradation on site, rather than accumulated over time. Fiprole and pyrethroid concentrations in whole plant samples are shown in Figure 3.3. Mean plant concentrations follow the same spatial trend as whole water or sediment concentrations, with levels generally decreasing from inlet to midpoint to outlet for all compounds. Since the spatial trends of fiproles and pyrethroids are the same in whole water and plant samples, it appears that plant uptake played a role in the removal of these insecticides in the PCW. Fipronil is a systemic insecticide, so some degree of plant uptake of fiproles was to be expected. Detection of pyrethroid residues in whole plant samples was an unexpected finding due to the hydrophobicity of these insecticides and their consequent affinity for sediment. However, there are studies that have documented detection of pyrethroids in plant samples, either by uptake or by apparent irreversible sorption to plant tissues. Different temporal trends were observed for fipronil, fipronil degradation products, and the pyrethroids in plant tissues. Fipronil mean plant concentrations initially decreased from June-July 2018, increased to peak levels in August 2018, gradually decreased until November 2018, and progressively increased until January 2019. This trend of fipronil concentrations over time coincided with the temporal trends of fipronil in whole water and sediment samples, providing more evidence that plant uptake contributed to the removal of fipronil in the PCW. Moreover, it is likely that fipronil initially adsorbed to wetland sediment and was then absorbed into macrophyte roots. The mean plant concentrations of fipronil desulfinyl, fipronil sulfide, and fipronil sulfone all gradually increased from June 2018-January 2019, indicating some degree of accumulation in plant tissues over time. However, since fipronil did not follow this temporal trend of accumulation in wetland macrophytes, it is likely that some of the parent compound was metabolized into these derivatives upon uptake. Fipronil sulfone was present at higher concentrations in plant samples than the other degradation products, which was in agreement with previous studies showing that in vivo plant oxidation is a major metabolic pathway for absorbed fipronil. Bifenthrin and cyfluthrin displayed no temporal trend in plant tissues since they were both only detected in inlet samples at one time point. This suggests that plant adsorption or absorption did not play a major role in the removal of pyrethroids by the PCW. Fipronil was detected at the highest levels in PCW plants, with mean concentrations of 4.70-194 ng g-1. Moderate mean concentrations of ND- 17.7 ng g-1 were observed for fipronil sulfone. The lowest mean plant concentrations were measured for fipronil desulfinyl , fipronil sulfide , bifenthrin , and cyfluthrin. The results of plant tissue analysis reveal that plant uptake played an important role for the removal, degradation, and storage of fipronil, but did not contribute substantially to the removal of fipronil degradation products or pyrethroids. However, it must be noted that the dense vegetation was essentially slowing down the flow and filtering off suspended solids, contributing greatly to the removal through sedimentation. In addition, microbial activity in the rhizosphere of plant roots likely facilitated the degradation of these chemicals in the sediment, further contributing to the overall pesticide removal. The concentration-based removal values of fiproles and pyrethroids from water flowing through the PCW are given in Table 3.1. Over the entire course of the study, removal values for fipronil desulfinyl, fipronil sulfide, fipronil, fipronil sulfone, bifenthrin, and cyfluthrin were 100%, 99.7-100%, 57.8-88.1%, 75.6-100%, 74.7-100%, and 36.6-82.2%, respectively. The compounds with the highest removal values were fipronil desulfinyl, fipronil sulfide, fipronil sulfone, and bifenthrin, while fipronil and cyfluthrin showed the lowest removal. It is important to note that only fipronil, bifenthrin, and cyfluthrin were detected every month, and fipronil and cyfluthrin were detected at higher levels than all the other compounds. Previous studies have similarly shown that the average CW removal rates of fipronil and pyrethroids were 44% and 52-94%, respectively, for other CW systems.