Article pubs.acs.org/est
Occurrence of Fipronil and Its Biologically Active Derivatives in Urban Residential Runoff J. Gan,*,† S. Bondarenko,† L. Oki,‡ D. Haver,§ and J. X. Li⊥ †
Department of Environmental Sciences, University of California, Riverside, California 92521, United States Department of Plant Sciences, University of California, Davis, California 95616, United States § University of California Cooperative Extension, Orange County, Irvine, California 92618, United States ⊥ Department of Statistics, University of California, Riverside, California 92521, United States ‡
S Supporting Information *
ABSTRACT: Insecticides are commonly used around homes for controlling insects such as ants, termites, and spiders. Such uses have been linked to pesticide contamination and toxicity in urban aquatic ecosystems. Fipronil is a relatively new and popular urban-use insecticide that has acute toxicity to arthropods at low-ppb levels. In this study, we collected runoff water from 6 large communities, each consisting of 152 to 460 single-family homes, in Sacramento County and Orange County, California, and evaluated the occurrence of fipronil and its biologically active derivatives over 26 months under dry weather conditions. Statistical modeling showed that the levels of fipronil and derivatives in the runoff water were both spatially and temporally correlated. More than 10-fold differences were observed between the Sacramento and Orange County sites, with the much higher levels for Orange County (southern California) coinciding with heavier use. The median concentrations of combined fipronil and derivatives for the Orange County sites were 204−440 ng L−1, with the 90th percentile levels ranging from 340 to 1170 ng L−1. These levels frequently exceeded the LC50 values for arthropods such as mysid shrimp and grass shrimp. The highest levels occurred from April to October, while decreases were seen from October to December and from January to March, likely reflecting seasonal use patterns and the effect of rain-induced washoff. Fipronil and fipronil sulfone (oxidation derivative) each accounted for about 35% of the total concentrations, with desulfinyl fipronil (a photolytic product) contributing about 25%. Results of this study clearly established residential drainage as a direct source for pesticide contamination in urban waterways, and for the first time, identified fipronil as a new and widespread contaminant with potential ecotoxicological significance.
■
INTRODUCTION Urbanization places an increasing level of environmental stress on surface aquatic ecosystems such as streams, lakes, and estuaries. One form of ecological stress is the contamination of surface water by pesticides, leading to acute and chronic toxicities to the indigenous aquatic organisms.1−7 For instance, in California, a large number of urban streams and estuaries are designated as “impaired” water-bodies due to toxicities associated with insecticides.8 Nationwide, an early study by the U.S. Geological Survey showed sustained levels of multiple insecticides in 75% of urban surface water streams in the United States.9 The primary source of pollution in urban watersheds is generally believed to originate from the off-site movement of pesticides used around residential homes, which is facilitated by irrigation and rain-induced surface runoff over impervious urban hard surfaces.10−12 Insecticides have long been used for the control of common household and garden pests such as termites, ants, spiders, and mosquitoes. Over the last few decades, due to the development of pest resistance or concerns over adverse environmental or human health effects, the types of urban-use insecticides have undergone several iterations of changes. Chlorinated insecti© 2012 American Chemical Society
cides such as DDT and chlordane were banned in most developed countries in the 1970s and were replaced primarily by organophosphates (OPs) and carbamates. In late 1990s, owing to concerns for cholinesterase toxicities to humans and aquatic organisms, the use of many OPs and carbamates became restricted, whereas the use of pyrethroids has since increased. Along with the increased use of pyrethroids, another new compound, fipronil, has also become an insecticide of choice for termite and ant control. In California, the reported urban use of fipronil by licensed applicators has rapidly increased over the past decade.13 A number of recent studies showed ubiquitous contamination of bed sediment and water column by pyrethroids in urban areas in California.1−6 However, so far no study has considered the occurrence of the other primary replacement fipronil. Before its registration for urban use, fipronil was used in rice production in several southern states in the United Received: Revised: Accepted: Published: 1489
August 19, 2011 January 3, 2012 January 6, 2012 January 6, 2012 dx.doi.org/10.1021/es202904x | Environ. Sci. Technol. 2012, 46, 1489−1495
Environmental Science & Technology
Article
States.14−17 Fipronil residue discharged from rice paddies was linked to mortality of crawfish.11,16 Subsequent studies consistently showed that fipronil has acute toxicity to aquatic arthropods, chironomids, and other aquatic invertebrates at very low concentrations.18−25 For instance, the LC50 of fipronil was 0.32 μg L−1 for grass shrimp (Palaemonetes pugio),20,23 0.14 μg L−1 for mysid shrimp (Neomysis americana),26 and 0.18− 0.31 μg L−1 to a black fly (Simulium vittutum).22 At a concentration of 0.42 μg L−1 fipronil also eliminated the reproduction of Amphiascus tenuiremis, an estuarine copepod.18 In addition, studies show that fipronil derivatives, including fipronil sulfone, fipronil sulfide, and fipronil desulfinyl (Figure S1 in the Supporting Information), are readily formed from fipronil in the environment.27−33 These derivatives have biological activity similar to or even greater than fipronil itself to nontarget organisms.14,18,24 The primary objective of this study was to evaluate the occurrence of fipronil and its biologically active derivatives in runoff from residential homes, and the temporal and spatial patterns. Irrigation runoff water was collected at 6 sites in California weekly or biweekly for 26 months, with each site (i.e., neighborhood) consisting of 152−460 single-family homes. This study represented the first of its kind in scope and magnitude, and was unique in that samples were collected directly from residential areas. Study findings may be used not only for determining sources of pesticide contamination in urban watersheds, but also for devising strategies for regulation and pollution mitigation.
from Fisher Scientific (West Chester, PA). Individual stock solutions of fipronil and derivatives were prepared in acetone and stored in amber bottles at 4 °C before use. Sample Collection. Water samples were collected at the storm drain outfalls weekly or biweekly from July 2006 through December 2008 during the dry season that typically extended from April through October and around and through storm events from November to March. Storm runoff samples were taken under different schemes, and the uncertainties due to the strong dependence of contaminant distribution on the exact sample collection time precluded a simultaneous consideration in this paper. Water was collected manually by dipping a 1-L amber glass bottle into the center of the flow. Selected water quality parameters, including pH, temperature, and dissolved oxygen, were measured on site when a sample was taken. The water samples were placed in coolers filled with ice bags, and shipped to the laboratory within 24 h from the time of sampling. The water samples were analyzed within 24 h from when they were received or otherwise preserved with methylene chloride and stored at 4 °C. All samples were extracted within 48 h upon arrival. Sample Analysis. The whole water sample was fortified with 50 g of sodium chloride,100 μL of a surrogate mixture of mirex (20 ng L−1) and decachlorobiphenyl (ng L−1), and transferred to a 2-L glass separatory funnel. The sample was then extracted with 50 mL of methylene chloride for 1 min, followed by the collection of the solvent phase into a 500-mL round-bottom flask. The sample container was rinsed with 20 mL of acetone and the rinse solution was combined with the solvent extract. The same extraction step was repeated two additional times with fresh solvent, and the extracts were combined. The extract was dried by passing through 50 g of anhydrous sodium sulfate and condensed to 5 mL on a vacuum rotary evaporator, followed by drying to 1.0 mL under nitrogen. The final extract was spiked with 13C6-cis-permethrin and analyzed on a Varian 3800 GC equipped with a Varian 1200 triple quadrupole mass spectrometer (GC/MS-MS) (Varian, Sunnyvale, CA). A VF-5MS column (30 m × 0.25 mm × 0.25 μm) (Varian) or a DB-5 MS column (30 m × 0.25 mm × 0.25 μm) (J&W Scientific, Folsom, CA) was used for separation. The pulsed splitless mode at 45 psi was used for injection with the purge valve closed for 1.0 min. The inlet temperature was 260 °C. The initial column temperature was set at 80 °C for 1 min, ramped to 160 at 25 °C min−1, further ramped to 300 at 7 °C min−1, and held at 300 °C for 7 min. High-purity helium was used as the carrier gas at 1 mL min−1. The mass spectrometer was operated in the electron ionization (EI) mode with selected reaction monitoring (SRM). The transfer line, ionization source, and manifold temperatures were 300, 170, and 40 °C, respectively. Argon (99.999%) was used as a collision gas. The EI-MS/MS library was created for target analytes by injecting individual compounds. Detailed GC-MS/MS conditions are given in the Supporting Information (Table S1). Quality Control, Data Analysis, and Statistical Modeling. The method detection limit (MDL) was determined as 3 times the standard deviation of the signal in the fortified extract with a known amount of the compound before analysis. The MDLs of fipronil and derivatives ranged from 1.2 to 2.0 ng L−1 (Table S1). One laboratory blank was included for each set of analyses. For the total of 572 field samples analyzed, 151 laboratory blanks, 47 duplicate field samples, 5 laboratory spikes, and 3 matrix spike duplicate samples were also run. Fipronil and derivatives were not detected in the blanks. The
■
MATERIALS AND METHODS Study Sites. Two residential communities in Sacramento, CA and four communities in Orange County, CA were selected based on their representativeness in terms of lot sizes, age of the homes, household income, and accessibility to storm drain outfalls. These neighborhoods did not include any other land use type (e.g., commercial, industrial) within the area served by the same storm drain system. For confidentiality reasons, the two sites in Sacramento were labeled as N1 and N2, while the four sites in Orange County were labeled as S1−S4. Each residential community contained 152−460 single-family parcels, each of which typically consisted of a driveway, a front lawn, and a backyard. Pesticide applications in these areas are made by either professional applicators or homeowners, but fipronil products may be used only by licensed applicators in California. The information on the use of fipronil at individual properties was unavailable as such information is considered proprietary by pest control companies. Surface runoff water in these neighborhoods moved along street curbs into storm drains and was discharged under gravity at the point of storm drain outfall into a nearby tributary of a creek or river. Under dry weather conditions, discharge at the storm drain outfall may be assumed to originate entirely from runoff as the result of landscape irrigation or other outdoor human activities. Chemicals. Standards of fipronil (98.0%), mirex (100 μg mL−1 in methanol, as a surrogate), and decachlorobiphenyl (2000 μg mL−1, as a surrogate) were purchased from Chem Service (West Chester, PA). Standards of fipronil sulfone (99.7%), fipronil sulfide (98.8%), and fipronil desulfinyl (97.8%) were obtained from the U.S. Environmental Protection Agency’s National Pesticide Standard Repository in Fort Meade, MD. Phenoxy-13C-cis-permethrin (99%, as the internal standard) was purchased from Cambridge Isotope Laboratories (Andover, MA). All solvents used were GC grade and were 1490
dx.doi.org/10.1021/es202904x | Environ. Sci. Technol. 2012, 46, 1489−1495
Environmental Science & Technology
Article
Table 1. Descriptive Statistics of Occurrence (ng L−1) of Fipronil and Derivatives in Runoff Water from Residential Parcels Located in Sacramento County (N1, N2) and Orange County (S1−S4), California, under Dry Weather Conditions during 2006−2008
average recovery was 95.2 ± 5.8% to 111 ± 19.5% for all analytes in laboratory spike samples and matrix spike samples. To estimate descriptive statistics for the pesticide data set containing less than 50% nondetects, the nonparametric Kaplan−Meier method was used.34 In the nonparametric Kaplan−Meier method, nondetects were not substituted. The large data sets afforded a closer examination of fipronil occurrence as a function of location and time.35,36 The concentration values were analyzed using a spatial−temporal linear mixed model using SAS:37
ln(Cijm) = δi + θm + εijm
site
(1)
where i = 1, 2, or 3 represents year 2006, 2007, or 2008, j = 1, 2, ..., 12, represents distinct month effects, and m = 1, 2, ..., 6, represents location effects. Four residual error structures were explored: (1) an autoregressive, order 1 (AR1) error structure I assuming that the temporal correlation parameters (ρm) and variance components (σm2 ) changed across all sites; (2) an AR1 error structure II assuming that ρm and σm2 were different between the north and south sites; (3) a diagonal independent distributed error structure (ρm = 0∀m) assuming no temporal dependence and the diagonal variance components changed across sites; and (4) an error structure assuming an independent and identically distributed error structure exhibiting no temporal dependence. For statistical modeling, nondetects were treated as a half of the respective detection limits.35
■
RESULTS AND DISCUSSION Frequency, Concentrations, and Patterns of Fipronil in Runoff. The water flow at some sampling sites was intermittent during the summer, affecting the actual number of samples taken at each site. Overall, from 69 to 98 water samples were collected from each site spanning over 26 months. The sampling frequency, time span, number of sampling sites, and their locations together provided a comprehensive sampling scheme both spatially and temporally, allowing the construction of a snapshot for the occurrence of fipronil and its derivatives at the source of contamination for these geographic areas. At a limit of detection of 1.5 ng L−1, fipronil was detected at 85% (N1) and 66% (N2) at the two northern California sites. In comparison, fipronil was found in nearly all samples that were taken from the four southern California neighborhoods (S1−S4) (Table 1). The concentrations of fipronil varied over a wide range at each individual site, as shown in Figure 1A for S2, with maximum concentrations in the low ppb (μg L−1) range for most sites (Table 1). As a Gaussian distribution was not always observed, the median concentration, along with other statistics describing the occurrence of fipronil, was calculated for each site (Table 1). For N1 and N2, the median concentrations were
