Study Region

Six sampling sites were chosen following consultation with the cooperator and based on several stream and watershed criteria. Sites were chosen in watersheds with suitable salmonid habitat, known applications of pesticides for commercial forestry or rights of way maintenance, the presence of cannabis grow sites, or combinations of those attributes. Sampler and field personnel safety also were key considerations when specific sampler deployment locations were chosen. Five of the six sites were sampled five times each by USGS personnel (table 1; figs. 1, 2). The North Fork Calaveras River near Mokelumne Hill, California (USGS site 11307550), was inaccessible for one out of five sampling events, so a seventh, upstream site on the North Fork Calaveras River below Jesus Maria Road Bridge near Paloma, California (USGS site 381713120393701), was sampled instead (U.S. Geological Survey, 2021).

Streams sampled in the study region drain watersheds in the foothills of the western slope of the Sierra Nevada in northern California. The streams flow from east to west, eventually draining into tributaries of the Sacramento–San Joaquin Delta (not shown) in California’s Central Valley. Forests, grasslands, and shrublands are the predominant land cover types, although there are differences in their proportions in the six watersheds. Developed and agriculture areas make up a small percentage of total land area (appendix 1, table 1.1; U.S. Geological Survey, 2019a).

Precipitation in the study region typically occurs during the winter and spring; summer and fall are usually dry (Menne and others, 2012; fig. 3). Streams were sampled with the goal of capturing the first fall or winter storm event (first flush) after a dry summer or fall, respectively; the first spring storm event; and a dry season event. Site accessibility was another factor that restricted sample collection between December and April.

Pesticide Use

Pesticide-use reporting in California can be grouped into three categories: (1) agricultural use by licensed applicators that is reported by date at the Public Land Survey System level (a grid system made up of cells that are approximately 1 mile by 1 mile); (2) applications by licensed applicators for non-agricultural uses such as landscape maintenance, rights of way maintenance, and structural pest control that are reported at the county level by month; and (3) applications by homeowners and most industrial uses that are not reported (California Department of Pesticide Regulation, 2020a). Pesticide use at the section level was summarized for the study region watersheds because the approximately 1 square mile (mi²) resolution of section-level data was suitable for estimating pesticide use in watersheds with areas ranging from 31 to 450 mi². Some county-level pesticide use was not apparent in the section-level pesticide summaries. Pesticide application at the county level was not summarized because of the high level of uncertainty in approximating pesticide use in the study region watersheds based on county-level data.

Most registered, section-level pesticide applications inside the study region were during the spring and summer based on 2018 pesticide-use data (fig. 4; California Department of Pesticide Regulation, 2020a); 2019 pesticide-use data were not available when this report was written. Application sites for pesticide use at the section level varied among the watersheds. The Bear River, North Fork Calaveras River, Oregon Creek, and South Fork American River watershed areas within the study region only had registered pesticide use for timberland sites. The North Fork Cosumnes River and South Honcut Creek watershed areas within the study region had registered pesticide applications for a variety of uses (fig. 5). County-level pesticide applications were not included on figures 4 and 5, but thousands of pounds of pesticides were applied at the county level in study region counties for uses such as landscape maintenance, public health, and rights of way maintenance (California Department of Pesticide Regulation, 2020a). Figures 4 and 5 include pesticide-use totals from pesticides not analyzed for during the study. Pesticide-use totals on figure 5 are limited to the 155-pesticide analyte list.

Registered pesticide-use data showing only those compounds that were analyzed for in this study indicate that the greatest amount and variety of pesticides with the potential to be detected were applied in the North Fork Cosumnes River watershed (table 1; fig. 6). Other pesticides such as glyphosate, imazapyr, triclopyr, clopyralid, aminopyralid, and 2,4-dichlorophenoxyacetic acid were applied in study region watersheds or counties but were not analyzed because of incompatibility with the analytical methods used in this study (McWayne, 2020).

Passive Sampler Devices

Chemcatcher passive sampling devices (hereafter referred to as “Chemcatchers”) were used to passively sample for pesticides and pesticide degradates in surface water for this study. Briefly, a Chemcatcher functions by allowing target compounds in water to pass through a diffusion-limiting membrane (DLM) and accumulate in a receiving phase (Kingston and others, 2006). The Chemcatcher (T.E. Laboratories, Tullow, Ireland) consists of a weighted stainless-steel cage with two tiers of disk holders (fig. 7). Each tier can secure three polytetrafluoroethylene (PTFE) disk holders for a maximum of six disks deployed concurrently within a single cage. The receiving phase is a 47-millimeter (mm) solid-phase extraction (SPE) disk. An optional DLM can be inserted between the SPE disk (hereafter referred to simply as “disk”) and the retaining ring (fig. 8). Polyethersulfone (PES) DLMs with a pore size of 0.2 micrometers (µm) were used for disk configurations deployed with a DLM in this study (McWayne, 2020).

Silicone bands also were tested as simple, inexpensive, passive sampling devices for use inside the Chemcatcher cages at multiple sites during the study period. Several studies have demonstrated the effectiveness of silicone as a passive sampler for many of the pesticide compounds analyzed for in this study (Emelogu and others, 2013; Ahrens and others, 2015; Swanson and others, 2018). The silicone bands tested were obtained from 24hourwristbands.com with the following specifications: 4.40 grams, 5 cubic centimeters in volume, 203.2 mm in circumference, 12.7 mm in height, 2 mm thick, and 100 percent silicone.

Disk Configuration

Chemcatchers are typically deployed with the goal of measuring compounds with a targeted range of physio-chemical properties and using compound-specific sampling rates to estimate time-weighted average concentrations. The target analyte list for this study includes 155 pesticide compounds with a wide variety of physio-chemical properties including a range of logarithm of octanol–water partition coefficient (log K_ow) values from −0.5 to 7.0. One of the challenges in adapting the use of Chemcatchers as a broad screening tool for such a diverse range of pesticide compounds is that no single disk configuration will be ideal for all the compounds analyzed for in this study. As such, multiple configurations of SPE disks and DLMs were tested with the goal of detecting the presence/absence of the widest range of compounds possible. Identifying pesticide occurrence was prioritized over optimizing the SPE disk configuration for the most ideal uptake rates.

One decision to make when planning a deployment configuration is whether to use a DLM between the receiving phase (SPE disk) and the PTFE retaining ring. A meta-analysis by Charriau and others (2016) showed that the use of DLMs can substantially restrict the sampling rate of some compounds for a given receiving phase. The sampling rate of hydrophobic compounds was found to be especially limited using DLMs. However, deploying Chemcatchers without a DLM can cause the receiving phase to reach equilibrium too quickly (less than 10 days) and make the calculation of water concentrations unreliable for longer deployments (Shaw and others, 2009). Deploying the disks without a DLM also limits measurements to whole-water concentrations only as opposed to dissolved concentrations that can be measured when a DLM is used over the disk. In the context of this study (approximately 14-day deployments and a 155-compound analyte list), the ability to detect a wider range of pesticide compounds was prioritized over calculating water concentrations, so samplers were deployed without DLMs when possible.

Several 47-mm SPE disks were available to use as receiving phases with various DLM materials. Disk types tested during this study are listed in table 2. The original study design called for testing the Atlantic hydrophobic lipophilic balance receiving phase (HLB) and Empore styrene divinylbenzene-reverse phase sulfonate (SDB-RPS) SPE disk types based on the results of an earlier, unpublished pilot study by the USGS and the Central Valley Regional Water Quality Control Board. However, the Empore SDB-RPS was discontinued by 3M shortly after sample collection for this study, forcing a change in SPE disk configurations. The Resprep C18 SPE disk was added to the study and was used as a replacement for the Empore SDB-RPS. Although not initially considered for this study, the AttractSPE HLB (Affinisep) and AttractSPE SDB-RPS (Affinisep) SPE disks were added later as potential alternatives as they became available. A comparison study initiated because of the discontinuation of the Empore SDB-RPS SPE disk type showed good comparability between the Empore SDB-RPS, the AttractSPE HLB, and AttractSPE SDB-RPS for a set of nine pesticides (Becker and others, 2021). The thin, soft membrane material of the AttractSPE HLB and SDB-RPS SPE disks was resilient enough to withstand deployment without a DLM, whereas the Atlantic HLB SPE disks tended to disintegrate without a DLM.

Field Methods

Chemcatchers were transported to sample sites on wet ice with transport lids affixed to the PTFE disk assemblies, wrapped in foil, then sealed in plastic bags. Chemcatcher cages and disks were installed by two-person field teams. Upon arrival at a sample location, one team member (dirty hands) put on clean gloves, rinsed the stainless steel Chemcatcher cage in native stream water, then secured the cage in an appropriate location. The other team member (clean hands) removed the still-sealed PTFE disk assembly from a cooler and brought it to the stream bank. Using clean gloves, the clean hands team member removed the PTFE disk assembly from the protective plastic bag and foil wrapper. The clean hands team member then removed the transport lid and handed the uncovered PTFE disk assembly to the dirty hands team member, who installed the disk assembly in the submerged stainless-steel cage, taking care to minimize the amount of time the disk was exposed to air.

For passive sampler field blanks, disks were transported to the field sites in their PTFE disk assemblies, exposed to air by the clean hands team member for about 10 minutes (min) while the environmental disk assemblies were being deployed, then re-sealed, and transported back to the laboratory on wet ice. The field blank disk assemblies were stored in a freezer at −20 degrees Celsius (°C) between deployment and retrieval. The process was then repeated using the same field blank disk during the retrieval of the corresponding environmental passive sampler.

Passive samplers were deployed in streams for about 14 days. Retrieval of the Chemcatchers was the reverse of the deployment process described earlier in the text: the dirty hands team member retrieved the cage from the stream, removed the disk assembly from the stainless-steel cage, then handed the disk assembly to the clean hands team member. The clean hands team member quickly sealed the disk assembly with a clean PTFE transport lid; then wrapped the assembly in foil; and placed the disk assembly in a clean plastic bag labeled with the sample site, deployment date and time, retrieval date and time, and disk configuration. The sealed disk assemblies were then transported to the analyzing laboratory on wet ice for processing and analysis by gas chromatography with tandem mass spectrometry (GC-MS/MS) and liquid chromatography with tandem mass spectrometry (LC-MS/MS).

Silicone band passive sampling devices were deployed in the same cage as the Chemcatchers passive samplers. The silicone bands were placed above the upper tier disk holder inside the Chemcatcher cages. Cleaned silicone bands were handled in the same manner as Chemcatchers when transported to and from sample locations and when deploying and retrieving the bands. The clean hands team member removed the aluminum foil-wrapped silicone band from the plastic bag, removed the silicone band, then handed to silicone band to the dirty hands team member. The plastic bag and foil were labeled and saved for re-use during retrieval. The dirty hands team member then inserted the silicone band into the cage above the upper disk holder around the center post.

Surface-water grab samples were collected by submerging two narrow-mouthed, amber-glass bottles below the stream surface until the bottles were full and there was no headspace remaining. One bottle was collected for analysis by gas chromatography with mass spectrometry (GC-MS), and one bottle was collected for analysis by LC-MS/MS. Surface-water samples were filtered through a 0.7-µm glass fiber filter before extraction. Grab samples were collected at the time of passive sampler deployment, at the time of passive sampler retrieval, or collected at the time of deployment and retrieval.

Laboratory Methods

Before deploying, PTFE disk holders were cleaned, and disks were cleaned and conditioned in the laboratory. The PTFE disk holders were cleaned with warm water and detergent, rinsed with tap water, rinsed with methanol, then rinsed with organic-free water (OFW) before use. Disks for use in Chemcatchers were cleaned and conditioned using solvents outlined in table 4. Disk cleaning was done using 47-mm filter holders (Diskcover-47, Restek, Bellefonte, Pennsylvania) on a vacuum manifold. Conditioning solvents were allowed to soak into the disk for 1 min before being drawn though the disk under gentle vacuum. Cleaned disks were then inserted into the PTFE disk holders until use, and then the PTFE disk holders were sealed (using a threaded PTFE cap) to keep the SPE media wet until deployment. At this stage for disk configurations where a DLM was used, the DLM was inserted between the receiving phase and the PTFE retaining ring. The PES DLMs were soaked in methanol overnight then transferred to OFW until assembly.

After deployment and recovery, the SPE disks were stored in a freezer at −20 °C until they were ready to be processed. Each SPE disk was freeze-dried to remove remaining water before elution, then it was eluted using a 47-mm filter holder on a vacuum manifold with a collection tube placed below the filter holder to catch the eluent. After inserting the freeze-dried disk into the filter holder, the disk was spiked with 50 microliters (µL) of a 2 nanogram per microliter (ng/µL) surrogate solution containing 13C3-atrazine, di-N-propyl-d14-trifluralin, and 13C4-fipronil and 50 µL of a 1 ng/µL solution containing d4-imidacloprid. The disk was then eluted with two 10 milliliter (mL) aliquots of a 1:1 volume/volume (v/v) mixture of methanol and acetonitrile. The extract was then concentrated to near dryness and reconstituted with an addition of 200 µL of acetonitrile. Before transferring the extract to an autosampler vial for analysis, 20 µL of a 10 ng/µL solution containing d10-acenaphthene and d10-phenanthrene and 20 µL of a 5 ng/µL solution containing d3-clothianidin were added as internal standards. Sample extracts were stored in a freezer at −20 °C until analysis (up to 30 days).

Silicone band passive sampler devices were prepared for use by conditioning them using a series of solvent soaks. Silicone bands were soaked for 2.5 hours in a 1:1 v/v mixture of ethyl acetate and hexane, then soaked for 2.5 hours in a new 1:1 v/v solution of ethyl acetate and hexane, and then soaked for 2.5 hours in a 1:1 v/v mixture of ethyl acetate and methanol. After the three solvent soaks, the bands were loosely wrapped in aluminum foil, and residual solvent was allowed to evaporate overnight. Cleaned bands were wrapped in solvent-rinsed aluminum foil and then placed in a sealed plastic bag until deployment.

After deployment and retrieval, the silicone bands were gently rinsed with OFW and then air-dried. After drying, the weight of each band was recorded, and the bands were stored at −20 °C until extraction. The silicone bands were extracted by shaking the bands in flasks filled with ethyl acetate. Before adding ethyl acetate to completely submerge the silicone bands, the surface of the bands was spiked with 50 µL of a 2 ng/µL surrogate solution containing 13C6-permethrin, di-N-propyl-d14-trifluralin, 13C12-dichlorodiphenyldichloroethylene (p,p’-DDE), and 13C4-fipronil; and then spiked with 50 µL of a 1 ng/µL solution containing d4-imidacloprid. After adding the surrogate solutions, 50 mL of ethyl acetate was added to each flask and then shaken for 2 hours. After 2 hours, the extract was decanted from the flask into an evaporator tube, and a second addition of 50 mL of ethyl acetate was poured into each flask. Flasks were shaken for another 2 hours, and the second extract was then decanted into the same evaporator tube as the first extract. The combined extracts were concentrated to near dryness and reconstituted with an addition of 200 µL of acetonitrile. Next added was 20 µL of a 10 ng/µL solution containing d10-acenaphthene and d10-phenanthrene and 20 µL of a 5 ng/µL solution containing d3-clothianidin as internal standards before transferring the extract to an autosampler vial for analysis by GC-MS/MS and LC-MS/MS.

Surface-water grab samples for analysis by GC-MS and LC-MS/MS were extracted using the SPE method described in Sanders and others (2018). All surface-water samples were filtered through a 0.7-µm glass fiber filter before extraction, so all pesticide concentrations from surface-water samples were from the dissolved phase.

Chemcatcher Results by Disk Type

Disk-supply issues forced a change in disk configurations mid-study, so making comparisons using data from all five events was difficult. However, even with the discontinuation of the Empore SDB-RPS and the late addition of the AttractSPE SDB-RPS and AttractSPE HLB during the study, direct comparisons of disk configurations could be made between two subgroups of data. To make comparisons among disk configurations, the dataset was split into two subgroups in which only concurrently deployed disk configurations were compared to each other.

The first of these subgroups includes results from events 2 and 3 (hereinafter referred to as group A). Group A contained 11 direct comparisons of Resprep C18 disks deployed with no DLM versus Atlantic HLB disks deployed with PES DLMs. Of the 10 pesticides detected in group A, hexazinone was the only pesticide detected in the DLM covered HLB disks (fig. 10). The C18 disks without DLMs accumulated all 10 of the pesticides detected in group A.

The second subgroup includes results from events 4 and 5 (hereafter referred to as group B). Group B contained 12 direct comparisons of disks deployed without DLMs using hydrophilic-lipophilic balance (HLB) without diffusion-limiting membrane (DLM; Atlantic), styrene divinylbenzene-reverse phase sulfonate (SDB-RPS) without DLM (Affinisep), and octadecylsilane (C18) without DLM (Resprep) disks with a total of 9 pesticide compounds detected (fig. 11). Detections of the nine pesticide compounds in each of the three group B configurations were similar with detection counts of azoxystrobin, boscalid, cyhalothrin, dithiopyr, fipronil desulfinyl, and hexazinone equal among the three disk types. Bifenthrin, chlorantraniliprole, and methoxyfenozide were detected with a maximum deviation of two detection counts among the three disk types.

For compounds that were detected at least once in group B, there were 40 combinations of pesticides, collection dates, and sites available for comparison. In 36 of those combinations, the 3-disk configurations had uniform presence/absence performance. Overall, the AttractSPE HLB had the most detections, the AttractSPE SDB-RPS had fewer detections for one pesticide, and the Resprep C18 had fewer detections for three pesticides.

The improvement in HLB performance in group B (AttractSPE HLB disks deployed without a DLM) versus group A (Atlantic HLB deployed with a PES DLM) indicates that the HLB sorbent type can accumulate many of the compounds in this study’s analyte list, but the use of a PES DLM substantially limited the accumulation of target compounds for the sampling durations in this study. The comparisons made in groups A and B show that the factor that most affected disk configuration performance for the compounds detected in this study was the use of a DLM. Deploying the Chemcatcher with a DLM between the receiving phase and the PTFE retaining ring for the 14-day deployments used in this study restricted the uptake of pesticide compounds by the receiving phase (disk). The compound sorption to the disk was slowed appreciably by the DLM (extraction of a subset of DLMs revealed no appreciable sorption to the DLM itself). When deployed without DLMs (the Atlantic HLB disks disintegrated when deployed without a DLM), the three disk types capable of being used without a DLM (Resprep C18, AttractSPE HLB, and AttractSPE SDB-RPS) all detected the presence/absence of compounds analyzed for in this study similarly. When comparing pesticide concentrations in disks deployed in the same cage only, the Resprep C18 accumulated the greatest pesticide mass 50 percent of the time, the AttractSPE SDB-RPS accumulated the greatest pesticide mass 30 percent of the time, and the AttractSPE HLB accumulated the greatest pesticide mass 20 percent of the time (Uychutin and others, 2021).

One observation based on the dataset is that the accumulation of more hydrophobic pesticide compounds in the receiving phase is restricted by the application of a DLM (fig. 12). This observation supports the work of Charriau and others (2016) whose meta-analysis of Chemcatcher studies showed that the use of PES membranes in Chemcatcher can substantially limit the accumulation of more hydrophobic compounds in the receiving phase. When deploying Chemcatcher samplers with the goal of detecting presence/absence of pesticides that range from hydrophilic to hydrophobic, the results indicate that it is best to deploy disks with no DLM when calculating time-weighted averages is not required and the disk is sufficiently robust to allow for deployment without a DLM. Additional direct comparisons of Chemcatcher deployed concurrently using the same disks with and without DLMs would be necessary to further explore the correlation.

Silicone Band Passive Sampling Devices

During the study, 21 silicone bands were deployed, and 10 pesticides were detected in the silicone bands. Of the 21 silicone bands deployed, 14 bands accumulated at least 1 pesticide compound. Ten silicone bands were deployed between April and May 2019 in cages with hydrophilic-lipophilic balance (HLB) with diffusion-limiting membrane (DLM; Atlantic) and octadecylsilane (C18) without DLM (Resprep) disks (fig. 13). Nine silicone bands were deployed between October and December 2019 in cages with Hydrophilic-lipophilic balance (HLB) with diffusion-limiting membrane (DLM; Atlantic) and Octadecylsilane (C18) without DLM (Resprep) disks (fig. 14).

In the April–May 2019 comparisons, the silicone bands generally detected more hydrophobic pesticide compounds than Resprep C18 without DLM disks and Atlantic HLB with DLM disks. The azole fungicides propiconazole and difenoconazole were only detected in the silicone bands. The more hydrophilic compounds hexazinone, azoxystrobin, boscalid, and methoxyfenozide were detected in at least one of the Chemcatcher disks but not in silicone bands.

In the October–December 2019 comparisons, the silicone bands accumulated fewer pesticides than the Chemcatcher disks. Of the three compounds that silicone bands did accumulate in the October–December 2019 comparisons (azoxystrobin, dithiopyr, and pendimethalin), detection frequencies were lower than Chemcatcher disks for two of the compounds (azoxystrobin and dithiopyr). One silicone band did accumulate the single detection of pendimethalin in the study. Silicone bands did not consistently detect (or not detect) the same compounds as other passive samplers, which complicated efforts to directly compare results from silicone bands and other passive samplers. Additional testing is needed to make any conclusions about the performance of silicone bands when compared to Chemcatcher configurations tested in this study. The placement of silicone bands inside the cages (above the Chemcatcher disk holders) might have restricted flow across the silicone bands. Additional testing with the silicone bands deployed outside of the cages could be used to determine if their ability to accumulate compounds was affected by their placement inside the cages. The silicone bands are inexpensive, easy to deploy, and able to detect some compounds analyzed for in this study.

Bifenthrin

Bifenthrin is an insecticide that is registered for use on a variety of crops in agricultural applications as well as a variety of non-agricultural uses such as structural pest control or outdoor residential use (U.S. Environmental Protection Agency, 2010a). Bifenthrin was applied to Christmas trees in the Nork Fork Cosumnes River watershed in 2018 (table 1; California Department of Pesticide Regulation, 2020a). Bifenthrin also was applied for structural pest control and landscape maintenance in all counties containing study region sites. The actual location of non-agricultural pesticide application is not registered beyond the county level in California, so we could not confirm whether any county-level use occurred in study region watersheds. Bifenthrin also is an ingredient in products available for household use that would not be registered.

Bifenthrin was detected at least once at five out of six sites in the study region, including 11333500 North Fork Cosumnes River near El Dorado, California (table 1), where bifenthrin has a registered application. Detection frequencies were similar for wet- and dry-sampling events (fig. 15; 28 percent and 27 percent, respectively), but the 3 (May 2019) wet-sampling events had the highest number of detections, with samples from 3 out of 6 sites containing bifenthrin (11307550 North Fork Calaveras River near Mokelumne Hill, California; 392349121045601 Oregon Creek above Covered Bridge near North San Juan, California; and 384736120373501 South Fork American River above Forebay Road Bridge near Pollock Pines, California.)

Cyhalothrin

Cyhalothrin also is an insecticide that is registered for use on a variety of crops in agricultural applications and a variety of non-agricultural uses such as structural pest control and outdoor residential use (U.S. Environmental Protection Agency, 2010b). Registered agricultural use of cyhalothrin did not occur at the section level inside watersheds in the study region in 2018 (California Department of Pesticide Regulation, 2020a). Cyhalothrin was applied for structural pest control in all counties containing study region sites with minor applications for landscape maintenance and public health in some counties in 2018 (California Department of Pesticide Regulation, 2020a), but we could not confirm whether any of those applications occurred within study region watersheds. A guide released by the California Department of Pesticide Regulation listed cyhalothrin as one of several pesticides routinely found by law enforcement personnel at illegal cannabis grow sites (California Department of Pesticide Regulation, 2021). Aerial imagery indicates that cannabis grow sites are present in several of the study region watersheds, making cannabis grow sites a potential unregistered source of cyhalothrin. Cyhalothrin also is available for household use, which would not be registered.

Cyhalothrin was detected at least once at 4 out of 6 sites in the study region with similar detection frequencies for wet- and dry-sampling events (22 percent and 28 percent, respectively). The samples from wet-sampling event 5 (November 2019) had the highest number of cyhalothrin detections with samples from four out of six sites containing cyhalothrin.

Unlike bifenthrin and fipronil, cyhalothrin was detected in a surface-water grab sample and can be directly compared to EPA Office of Pesticide Programs aquatic life benchmarks. The presence of cyhalothrin in four of the study region streams is potentially concerning to resource managers because cyhalothrin has low aquatic life benchmarks for acute toxicity to invertebrates and fish (3.5 ng/L for invertebrates and 39 ng/L for fish; U.S. Environmental Protection Agency, 2020). One surface-water grab sample, collected from the 11333500 North Fork Cosumnes River near El Dorado, California, on November 26, 2019, contained cyhalothrin at a concentration of 48.3 ng/L (table 5). That concentration exceeds the aquatic life benchmarks for acute toxicity to invertebrates and fish and is in the 86th percentile of all cyhalothrin concentrations in whole water or dissolved surface-water samples from California where cyhalothrin was present (California Department of Pesticide Regulation, 2020b).

Fipronil

Fipronil is an insecticide from the phenylpyrazole class of pesticides. Its uses include forestry; a variety of crops in agricultural applications; outdoor residential; and a variety of indoor uses including industrial, residential, and pets (U.S. Environmental Protection Agency, 2011). Registered agricultural uses of fipronil did not occur inside the study region watersheds in 2018. Like the two other pesticides of special concern (bifenthrin and cyhalothrin), fipronil was applied for non-agricultural uses inside counties containing study region sites in 2018. Fipronil was mostly applied for structural pest control with minor uses registered for landscape maintenance and regulatory pest control at the county level, but we could not confirm whether any of those applications occurred within study region watersheds. Fipronil also is an ingredient in products available for household use, which would not be registered.

Fipronil and one of its degradates, fipronil desulfinyl, were detected at least once at four out of six sites in the study region. Two additional degradates (fipronil sulfide and fipronil sulfone) were detected at one site. Concentrations of fipronil degradates were not below the 20 ng/L threshold designated for this section, but research at the time of this study did indicate that degradates of fipronil are concerning at concentrations of as much as 4 orders of magnitude lower than the EPA OPP benchmarks (Miller and others, 2020). Fipronil and fipronil desulfinyl were detected more frequently in samples from wet-sampling events than from dry-sampling events. Fipronil sulfide and fipronil sulfone were only detected in samples from wet-sampling events. Samples collected during event 3 (May 2019) had the highest detection frequencies for fipronil or its degradates, with detections at four out of six sites.