6.2.1. Air

There is no information on releases of endosulfan to the atmosphere from manufacturing and processing facilities because these releases are not required to be reported (EPA 2005).

As a result of its use as an insecticide on fruit trees, vegetables, and other crops, endosulfan is released directly to the atmosphere during application. The compound is applied principally by aerial spray, ground spray, and airblast. The direct release to the atmosphere is commonly a result of spray drift, which immediately contaminates the air surrounding the application area. Volatilization and wind suspension occurring from post-application periods can also be a source of endosulfan in air. Atmospheric endosulfan derived from these sources has the potential to contribute to regional and long-range transport (EPA 2010a).

6.2.2. Water

There is no information on releases of endosulfan to water from manufacturing and processing facilities because these releases are not required to be reported (EPA 2005).

Endosulfan is most commonly released to water by atmospheric deposition, spray drift, runoff, and erosion. Direct release to water bodies is restricted and application restrictions require a buffer distance of 300 feet from surface waters for aerial application and 100 feet for ground application. In California, the buffer is 300 feet for both types of applications (EPA 2010a). Endosulfan is not expected to leach through soil to groundwater based on its low water solubility and its tendency to absorb to soil (EPA 2010a; HSDB 2010).

6.2.3. Soil

There is no information on releases of endosulfan to the soil from manufacturing and processing facilities because these releases are not required to be reported (EPA 2005).

The main routes of release of endosulfan to soils are direct application to crops and atmospheric deposition from spray drift, volatilized material, or from long-range atmospheric transport (EPA 2010a).

6.3.1. Transport and Partitioning

Endosulfan isomers and endosulfan sulfate are found throughout the environment in various media due to its widespread use, physical properties, and relative persistence. Figure 6-2 contains a conceptual model of how endosulfan moves between environmental compartments starting from field application to potential ecological receptors (EPA 2010a). However, transport and partitioning of endosulfans can be complex and depends greatly on environmental conditions (Weber et al. 2010).

Endosulfan is applied directly to soil through application to crops by aerial, hand, groundboom, airblast, rights-of-way, or backpack sprays, or by chemigation (EPA 2002). Measured K_oc values (average for four soils) for α- and β-isomers were 10,600 and 13,500 mL/g, respectively, indicating limited mobility in soil (EPA 2010a). Therefore, leaching to groundwater is not expected to be a concern. This is supported by available groundwater monitoring data, which indicate very low rates of detection in extensive groundwater monitoring programs (USGS 2012b). Model field studies also support this conclusion. Endosulfan did not leach from sandy loam soil following incorporation of 6.7 kg/hectare of the compound (Stewart and Cairns 1974). After sampling periods of 503–828 days, 90% of the residues were found in the top 0–15 cm of soil, 9% at 15–30 cm, and 1% at 30–45 cm. In model soil evaporation beds constructed to test the feasibility of treating pesticide wastes, endosulfan exhibited no movement in loamy sand soil beds up to 54 weeks after the start of the tests (Hodapp and Winterlin 1989). Endosulfan is metabolized in soil to endosulfan sulfate, which is also expected to be immobile based on its estimated K_oc of 9,800 (EPA 2010a).

Although it is not applied directly or in the vicinity of water bodies, endosulfan is transported to water via spray drift or atmospheric deposition, or through soil runoff and erosion (EPA 2002, 2010a). Endosulfan has been regularly detected in surface water samples taken from South Florida canals that drain agricultural areas (Harman-Fetcho et al. 2005; Pfeuffer 2011; Scott et al. 2002). Endosulfan transported to water is expected to eventually partition to sediments (EPA 2010a; Weber et al. 2010). However, estimated log air/water partition coefficients for α- and β-endosulfan and endosulfan sulfate ranging from −3.56 to −4.78 indicate that volatilization from water to air is expected to occur, and these chemicals can be considered semi-volatile (EPA 2010a). In a field dissipation study, volatilization was considered to be the dominant route of dissipation for endosulfan and endosulfan sulfate in cotton fields of sub-tropical India. High temperatures and low rainfall were likely influential factors for this behavior (Kathpal et al. 1997). Volatilization of α-endosulfan accounted for 34.5% of total losses from freshly tilled soil during another field study, while volatilization losses of β-endosulfan were much less (14.5%), indicating that the β-isomer is less volatile than the α-isomer (Rice et al. 2002). Volatilization from plant surfaces is also expected to occur, and may be more significant than from soil surfaces. In air sampling studies done in a wind tunnel, 12% of the initial endosulfan application volatilized from a silty sand soil after 24 hours, as compared to 60% from plant surfaces in 24 hours (Rudel 1997). When pure β-endosulfan was allowed to equilibrate in the apparatus, the ratio of the β-isomer to the α-isomer in the gas phase became 8:92 at 20 °C, suggesting a β- to α- conversion (Rice et al. 1997). This conversion would also contribute to total volatilization losses of endosulfans from treated fields.

Figure 6-2

Conceptual Model of the Potential Effects of Endosulfan Application on Ecological Receptors. Note: Endosulfan is expected to absorb strongly to soils and is not likely to leach to groundwater.

Because of its semi-volatile nature and relative stability in the atmosphere, endosulfan is susceptible to long-range transport in the environment (Weber et al. 2010). These transport and deposition processes can be localized, regional, or long-range. Atmospheric deposition rates of endosulfans (α-, β-, and sulfate) were estimated in the agricultural intensive Choptank River watershed of the Chesapeake Bay region. Total wet deposition (± combined absolute and relative error) for α-endosulfan was estimated at 0.96 ±0.1 kg/year. Estimated depositions for β-endosulfan and endosulfan sulfate were 2.7±0.3 and 0.5±0.06 kg/year, respectively. Deposition processes can be regional in nature. For example, probabilistic source contribution function (PSCF) modeling performed by Hafner and Hites (2003) suggests that atmospheric endosulfans in the Great Lakes Region are generated in the lower Michigan peninsula, and New York State, and to a lesser extent Pennsylvania.

Regional transport of endosulfans to remote, non-agricultural areas has been documented. Endosulfan residues have been detected in air, snow, rain, lichen, surface water, and sediment samples from Yosemite National Park, California. West to east atmospheric movement of these residues from the agricultural San Joaquin Valley is the likely source. Estimated winter and summer deposition rates of endosulfans (α-, β-, and sulfate) were calculated from data collected during the spring and summer of 2009. Winter deposition was estimated at 1.11 μg/m² with 72% of the total deposition occurring during this season. Summer deposition was estimated at 0.44 μg/m² making the total annual deposition 1.55 μg/m². The dominance of winter deposition is due to the fact most of the precipitation in this region occurs as snow (Mast et al. 2012a, 2012b). Long-range transport to remote regions is evident in the numerous studies that have monitored endosulfan in Arctic environmental media (Hageman et al. 2006a; Hung et al. 2005, 2010; Stern et al. 2005; Weber et al. 2010). Like other organochlorine pesticides in Arctic areas, α-endosulfan has been observed to undergo a “spring maximum event”, where air concentration peaks in the spring (April/May) and again in the fall (October/November). This is different from seasonal trends observed in temperate regions where air concentrations peak in the summer months (Hung et al. 2005).

Organochlorine pesticides like endosulfan undergo a phenomenon known as “cold mountain trapping” where cold temperatures (leading to condensation) and high precipitation rates of high-elevation areas in temperate regions cause increased deposition (mostly through snow) (Daly et al. 2005). This phenomenon for endosulfan has been noted in the results of monitoring activities in National Parks in western United States (Hageman et al. 2006a), in Yosemite National Park (Mast et al. 2012a, 2012b), and in National Parks of western Canada (Daly et al. 2007) (see Section 6.4).

Bioaccumulation and biomagnification potential of endosulfan in organisms varies, but generally suggests that it has the potential to bioaccumulate in organisms and biomagnify in the food webs. This potential has been extensively investigated in aquatic organisms. EPA’s Ecological Fate and Risk Assessment for Endosulfan (2010a) extensively summarized experimental bioconcentration factors (BCFs) and bioaccumulation factors (BAFs) for aquatic organisms available in the scientific literature. BCF values ranged from 17.1 to 11,583 in fish species (Hansen and Cripe 1991; Jonsson and Toledo 1993; Schimmel et al. 1977; Toledo and Jonsson 1992; Rajendran and Venugopalan 1991). However, values from Hansen and Cripe (1991) and Schimmel et al. (1977) that met quality screening criteria were 1,146 (in sheepshead minnow for α- and β-isomers) and 2,755 (in striped mullet for α- and β-isomers and endosulfan sulfate). Experimental BCF values in aquatic invertebrates (shrimp, mussel, oyster, clam, and crayfish) ranged from 12 to 600 (EPA 2010a; Ernst 1977; Rajendran and Venugopalan 1991; Roberts 1972; Schimmel et al. 1977). Mesocosm and microcosm bioaccumulation studies reported BAF values similar in range to BCF values ranging from 115 to 1,262 in Bluegill fish, water flea, green algae, oyster, and macrophytes (DeLorenzo et al. 2002; EPA 2010a; Pennington et al. 2004; and unpublished industry studies). BAF is similar to BCF except that BAF takes into account multiple routes of exposure and not just intake from gills. Weber et al. (2010) summarized BAFs based on monitoring data reported for Arctic aquatic fish and mammals. Wet weight BAF values (for α- and β-isomers and endosulfan sulfate) were 1,690 (Arctic char), 7,280 (salmon), and 3,260 (Arctic cod). Lipid weight BAFs (for α- and β-isomers and endosulfan sulfate) were 1.45x10⁵ (whole salmon), 3.13x10⁵ (whole Arctic cod), 6.76x10⁵ (female beluga blubber), 5.65x10⁵ (male beluga blubber), 2.21x10⁵ (female ringed seal blubber), 2.41x10⁵ (male and female ringed seal blubber), and 2.98x10⁴–3.52x10⁴ (Arctic char muscle). Biomagnification factors (BMFs) were 2.2 for cod to beluga, 1.5 for salmon to beluga, and 0.77 for cod to seals. BMF values represent the ratio of the level of chemical in the predictor versus the concentration in its diet and BMF values greater than 1 indicate a potential to biomagnify up the food web. However, these BMF values may be overestimated since they do not account for metabolism.

Assessing the bioconcentration and biomagnification potential in terrestrial organisms is difficult. Kelly et al. (2007) estimated BMFs for air-breathing organisms ranging from 4.9 to 23 for β-endosulfan. However, this model did account for metabolism. Armitage and Gobas (2007) suggest that application of soil-earthworm-shrew food-chain model illustrates that chemicals with log K_oa ≥ 5.25 and logK_ow values ranging from 1.75 to 12 may biomagnify given that they are not rapidly metabolized (half-life ~2.5 days). Estimated logK_oa values are 6.41 for both isomers and 8.45 for endosulfan sulfate. A measured log K_oa value for both isomers is reported at 8.64 (EPA 2010a). Analyses of two cases of intentional endosulfan poisoning reported terminal half-lives in blood serum of 15.2 hours (from 35% endosulfan formulation) and 8.8 hours (endosulfan content unknown) (Eyer et al. 2004). These half-lives suggest that metabolism of endosulfan may significantly attenuate biomagnification in the terrestrial food chain, based on the Armitage and Gobas model. Although there is a lack of comprehensive, standard biological monitoring data in humans, a variety of endosulfan metabolites (-sulfate, -ether, -lactone, -diol) have been detected in adipose tissue, placenta, cord blood, and breast milk of women at nanoscale concentrations. It is unclear the role of continuous, low-dose exposures and metabolism play in these concentrations (Cerrillo et al. 2005). At the very least, these data indicate that human metabolism of endosulfan occurs and may be complex.

6.3.2. Transformation and Degradation

6.4.1. Air

Endosulfan has been included in air monitoring programs and in many air monitoring studies conducted within the last 10 years. Results from these studies establish endosulfan as an air contaminant not only in agricultural areas, but in rural, mountainous, and Arctic regions.

Many studies have established α-endosulfan as one of the most prevalent organochlorine pesticides in the Arctic region (Weber et al. 2010). Results from the Arctic Monitoring and Assessment Program (AMAP) analyses of pesticide concentrations sampled from various sites in the Arctic region at various times between 1993 and 2006, indicate endosulfan undergoes long-range atmospheric transport. Sampling stations were located in Canada, Finland, Iceland, Svalbard/Norway, Russia, United States, and Greenland. Endosulfan was measured in Alert, Canada at average air concentrations ranging from 3.3 pg/m³ in 1993 to 6.5 pg/m³ in 2003. Concentrations sampled at Nuuk, Greenland between 2004 and 2005 averaged 4.8 pg/m³. Similar average concentrations were reported at the Russian Arctic stations. The Yukon region in the Canadian Arctic reported the highest average concentrations of endosulfan at 8.3 pg/m³ at Tagish in 1994 and Little Fox Lake in 2002–2003 (Hung et al. 2010). Like other organochlorine pesticides, α-endosulfan has been observed to undergo a “spring maximum event” at several Arctic sampling stations, where air concentration peaks in the spring (April/May) and again in the fall (October/November). This is different from seasonal trends observed in temperate regions where air concentrations peak in the summer months (Hung et al. 2005).

Shen et al. (2005) mapped air concentrations of α- and β-endosulfan collected from XAD passive air samplers from 2000 to 2001 located across North America, with 31 stations located in Canada, five in the United States, and four in Central America. α-Endosulfan was detected at higher concentrations than β-endosulfan, ranging from 3.1 to 685 pg/m³, detectable at 39 stations, compared to 0.03–119 pg/m³, detectable at 30 stations, respectively. The highest concentrations were detected in areas of agriculture.

Endosulfans were detected in outdoor air sampled from three mountainous national park locations in Canada. Passive air samples were taken from sites within Mount Revelstoke National Park in British Columbia, Yoho National Park in British Columbia, and Observation Peak in Banff National Park, Alberta. All three areas lay west of land used for agriculture, but the land in the immediate vicinity of the sampling sites had limited to no agricultural use (Daly et al. 2007).

In 2001, ground level and mid-troposphere (~4,400 m) air samples were collected from the Fraser Valley, British Columbia and analyzed for endosulfan (α- and β-isomers). Ground level samples were taken from rural and urban areas, and mid-troposphere samples were obtained during flight times in aircraft.

Concentrations of the two isomers in rural areas ranged from ~18 to ~82 pg/m³, with the exception of a period where concentrations exceeded 250 pg/m³ for several days. This was attributed to high local use of endosulfan products. Concentrations in urban areas were less variable and ranged from ~4 to ~62 pg/m³. Endosulfans were consistently below the detection limits in the high altitude samples, and therefore, possible trans-pacific movement of atmospheric endosulfans could not be discussed. The authors suggested that the low detection may be attributed to the low temperatures (0 °C) in the troposphere, which would result in partitioning to particulate matter (Harner et al. 2005).

Total endosulfans (α-, β-, and sulfate) were detected in passive air samplers located within the Global Atmospheric Passive Sampling (GAPS) network, which included polar sites (n=4), background sites (n=16), urban sites (n=6), and rural/agricultural sites (n=12) located worldwide. Sampling took place between December 2004 and March 2005. Endosulfan concentrations were highest compared to other organochlorine pesticides tested. Concentrations were highly variable, ranging from tens to hundreds of pg/m³, the geometric mean was 58 pg/m³. The α-isomer was the most abundant in the samples, accounting for ~90% of total endosulfans. Results for β-endosulfan and endosulfan sulfate were often below the detection limit. Concentrations were highest in tropical regions where regional applications and greater soil to air exchanges due to temperature may have occurred (Pozo et al. 2006).

Concentrations of endosulfans (α- and β-isomers and endosulfan sulfate) in air were measured by Hoh and Hites (2004) at four sampling sites located in four states in order to model potential sources. Samples were collected from Sleeping Bear Dunes National Lake Shore in Michigan, Bloomington, Indiana, Rohwer, Arkansas, and Cococrie, Louisiana between 2002 and 2003. Concentrations ranged from 0.56 to 1,200 (mean of 142), from 2.7 to 2,000 (mean of 260), from 4.7 to 390 (mean of 100), and from 3.6 to 480 (mean of 100) pg/m³ for Michigan, Indiana, Arkansas, and Louisiana, respectively. Both Michigan and Indiana sites experienced spikes in endosulfan concentrations during July and August, presumably from local application. Applying the Potential Source Contribution Function (PCSF) model to the results, a large potential source region of endosulfan running from Kentucky, Tennessee, Alabama, and Florida was identified. With the exception of Alabama, these results correlated well with known regional use patterns. Modeling results did not indicate any potential sources west of the sampling sites.

Sun et al. (2006) measured gas phase and particulate phase α- and β-endosulfan in air from seven sampling sites in the Great Lakes region between 1996 and 2003. Samples were taken from sites within the Integrated Atmospheric Deposition Network (IADN), with five of the seven sites located in rural areas. Mean gas phase concentrations from these sites are included in Table 6-1 and mean particulate phase concentrations are included in Table 6-2. Calculated half-lives for gas phase α-endosulfan at these locations ranged from 8.2 to 19 years for four of seven sites (results were not statistically significant for other sites). Particulate phase α-endosulfan half-lives ranged from 3.4 to 8.1 years for five of the seven sites (data were not available for the other sites). Calculated half-lives for gas phase β-endosulfan ranged from 3.2 to 9.7 years and particulate phase half-lives ranged from 3.0 to 8.0 years for five of the seven sites (data were not available for the other sites). These half-lives are based on a regression of the temporal trends of the endosulfan levels at these locations and are not to be confused with the estimated atmospheric photooxidation half-lives of α- and β-endosulfan. High outlier concentrations were observed during summer months, and were likely due to agricultural use. Results indicated that endosulfan concentrations (both isomers) increased from the western Great Lakes Region to the eastern region, possibly due to regional use patterns. Particulate α-endosulfan concentrations declined compared to concentrations recorded in previous years. It was not clear whether usage patterns in the region contributed to this declining trend.

Table 6-1

Average Concentrations of Gas-Phase α- and β-Endosulfan in Air from the Great Lakes Region.

Table 6-2

Average Concentrations of α- and β-Endosulfan in Air from the Great Lakes Region.

α-Endosulfan, β-endosulfan, and endosulfan sulfate were detected in residential indoor air sampled (n=52) as a part of the Arizona Border Study (NHEXAS-AZ), which collected samples from sites in the Yuma, Nogales/Naco, and Douglas areas. These three testing sites had varied geographies and land use. The Yuma area is highly agricultural with a history of pesticide use. The Douglas area is mountainous with a history of mining and smelting. The Nogales/Naco area is a border town with industrial activity prevalent across the border in Mexico. α-Endosulfan averaged 190 ng per 4 standard semipermeable membrane devices (SMPDs), with a range of 10–1,600 ng per 4 SMPD and an 85% detection rate. β-Endosulfan averaged 87 ng per 4 SMPDs, with a range of 3.7–490 ng per 4 SMPD and an 89% detection rate. Endosulfan sulfate averaged 48 ng per 4 SMPDs, with a range of 19–100 ng per 4 SMPD, but a significantly lower detection rate of 5% (Gale et al. 2009).

6.4.2. Water

Endosulfan is monitored extensively in surface water and groundwater through various state, regional, and national programs. Studies analyzing rainwater, snow, and runoff from across the United States have also been published.

The USGS National Water Quality Assessment (NAWQA) program began in 1991 and obtains water quality data from 51 basins nationwide. These basins include approximately 7,300 surface water sites and 9,800 groundwater wells. Consistent with evidence that endosulfan will adsorb to soil, the available NAWQA groundwater samples obtained between 2002 and 2011 revealed an extremely low detection rate (0.12%) for α-endosulfan and endosulfan sulfate and no detection of β-endosulfan. Only 10 samples collected during this period reported measured or estimated concentrations at levels above detection limits (see Table 6-3). A similarly low rate of detection in surface water and bed sediment was observed for samples obtained between 2006 and 2011. Results from samples containing concentrations above detection limits are summarized in Table 6-4. It is important to note that endosulfan is expected to hydrolyze in aquatic environments to endosulfan diol, which is not analyzed in these samples.

In 2010, EPA published an ecological fate and risk assessment report that summarized the extensive water monitoring data available from NAWQA, EPA’s STORET database, California’s Department of Pesticide Regulation (CDPR) Surface Water Protection Program Database, and the South Florida Water Management District (SFWMD) DBHydro database, among others. Combining the data from these sources, EPA illustrated regional and national trends of endosulfan presence in U.S. waters over the period of almost 20 years (EPA 2010a).

In south Florida, endosulfan levels in agricultural runoff were analyzed by sampling surface water concentrations in the extensive canals that drain urban and agricultural areas. These canals are managed by the SFWMD. Samples were collected between 1993 and 1997. Average endosulfan concentrations ranged from 9 to 99 ng/L during this period. Using more sensitive analytical methods for samples obtained between 1996 and 1997, endosulfans were detected at a rate of 100% with a peak concentration of 477 ng/L (Scott et al. 2002).

Further analysis of the SFWMD data from 1992 through 2007 by Pfeuffer (2011) revealed several trends concerning endosulfan concentrations in south Florida canals. Surface water concentrations in selected basins for this period are summarized in Table 6-5. Endosulfans were detected in sediment samples taken from South Miami-Dade County (n=142). Endosulfan sulfate had the highest concentration among the three endosulfans in this basin, with an average of 16 μg/kg (25 detections), and a maximum concentration of 120 μg/kg. α-Endosulfan had an average concentration of 6 μg/kg (17 detections) and a maximum of 30 μg/kg. β-Endosulfan had an average concentration of 5 μg/kg (24 detections) and a maximum of 24 μg/kg. Endosulfan has been identified as a chemical of concern in the South Miami-Dade County agricultural area, but the frequency and magnitude of endosulfan detections have decreased since the 1994–1995 growing seasons.

Endosulfans (α- and β-isomers and endosulfan sulfate) were detected in water sampled from 13 sites located in the Biscayne Bay canals of southern Florida (n = 88) obtained between November 2002 and March 2004. These sites were located near ecologically sensitive areas of Everglades National Park and Biscayne National Park. Many of the canals in these areas drain agricultural or mixed agricultural and urban areas. Concentrations of α-endosulfan ranged from 0.21 to 54 ng/L with an 81% detection rate. β-Endosulfan concentrations ranged from 0.20 to 16 ng/L with a 75% rate of detection. Endosulfan sulfate concentrations ranged from 0.22 to 28 ng/L with a 91% rate of detection. Endosulfan concentrations were highest towards the end of the growing season (Harman-Fetcho et al. 2005).

Table 6-3

α-Endosulfan and Endosulfan Sulfate Detected in Groundwater Sampled for the USGS National Water Quality Assessment Between 2002 and 2011^,.

Table 6-4

α-Endosulfan and Endosulfan Sulfate Detected in Surface Water and Bed Sediment Sampled for the USGS National Water Quality Assessment Between 2006 and 2011^,.

Table 6-5

Endosulfan Concentrations (μg/L) in Surface Water Measured from the South Florida Water Management District (SFWMD).

A combination of discrete water and passive water samples were collected from six stream sites in the Potomac River basin in 2007 and tested for chemical pollutants. β-Endosulfan and endosulfan sulfate were detected in 20 and 40% of passive water samples collected, respectively. No α- and β-endosulfan or endosulfan sulfate was detected in discrete water samples (Kolpin et al. 2013).

Levels of endosulfans were measured in rain water and air from the Choptank River watershed on the Delmarva Peninsula of the Chesapeake Bay. This watershed is located in an agricultural area and is vulnerable to pesticide runoff and atmospheric deposition. Samples were collected from 8 stations in 2000. In rainwater, α-endosulfan had a 13% rate of detection (n=71), an average concentration of 5.1 ng/L, and a range of 1.3–31 ng/L. β-Endosulfan had a rate of detection of 28%, an average concentration of 7.2 ng/L, and a range of 0.27–81 ng/L. Endosulfan sulfate had a rate of detection of 8.5%, an average concentration of 4.1 ng/L, and a range of 0.98–14 ng/L. Endosulfans were generally only detected in rain from June to early August. Total wet deposition rates were estimated as 0.96±0.1 kg/year for α-endosulfan, 2.7±0.3 kg/year for β-endosulfan, and 0.5±0.06 kg/year for endosulfan sulfate. The authors noted that these estimates are probably low compared to actual rates in areas of high pesticide usage (Kuang et al. 2003).

Snow and rain samples collected from 12 sites within Yosemite National Park, California during the spring and summer of 2008 and 2009 and were analyzed for endosulfan. α-Endosulfan and β-endosulfan were detected in 100% of the snow samples with concentrations ranging from 0.15 to 2.1 ng/L. Endosulfan sulfate was detected in 85% of the samples with concentrations ranging from 0.06 to 1.5 ng/L.

They were also among the most frequently detected current-use pesticides in the rainwater samples. Examination of the 2009 rain samples revealed a strong positive correlation between increasing concentrations of endosulfans during the summer months and increasing applications rates in the San Joaquin Valley during the same time. Winter and summer deposition rates were estimated for endosulfans (α- and β-isomers and endosulfan sulfate) and are included in Section 6.3.1, Transport and Partitioning. α-Endosulfan was the only pesticide present above the method quantitation limit in the surface water samples. The authors concluded that estimated water concentrations were sub-parts per trillion and orders of magnitude lower than aquatic benchmarks. Concentrations may be higher during the snowmelt periods of April and May (Mast et al. 2012a, 2012b).

Endosulfan levels in western National Parks were also analyzed by Hageman et al. (2006a, 2006b) as a part of a research project initiated by the U.S. National Parks Service (NPS). Snow pack samples were collected from alpine, sub-Arctic, and Arctic ecosystems from seven National Parks in the spring of 2003. These parks include Sequoia, Rocky, Rainier, Glacier, Denali, Noatak, and Gates of the Arctic. Concentrations of total endosulfans ranged from <0.0040–1.5 ng/L. Calculated deposition rates ranged from <0.19–1,400 ng/m². In a follow-up study using data collected from 2003 to 2005, Hageman et al. (2010) estimated the percent concentration due to regional transport for total endosulfans as approximately >90%.

6.4.3. Sediment and Soil

Endosulfans are monitored extensively through national programs such as the USGS NAWQA and National Oceanic and Atmospheric Administration’s (NOAA) National Status and Trends (NS&T) Program. The National Sediment Quality Survey (NSQS) includes sediment monitoring data from January 1990 to December 1998. Data from this source has been summarized in EPA’s ecological fate and risk assessment report for endosulfan, which was published in 2010. Sediment concentrations of endosulfan reported by NAWQA between 2006 and 2011 showed only one sample above the detection limit in Kewaunee, Wisconsin (see Table 6-5).

The NOAA NS&T Program reported a similarly low rate of detection (5.3%) in sediment samples obtained between 2005 and 2009 (NOAA 2012). Approximately 77 samples reported endosulfan (α- and β-isomers and endosulfan sulfate) above detection limits. These concentrations ranged from 0.08 to 12.59 ng/g dry weight. Details are provided in Table 6-6.

Lake sediment samples were collected from 19 lakes within Yosemite National Park, California during the summer of 2008 and 2009 and analyzed for endosulfan. Endosulfan sulfate was the most dominant endosulfan form detected. Total endosulfan concentrations ranged from 1.0 to 5.7 ng/g dry weight. Concentrations in lichen, lake sediments, and surface water (using SPMDs) displayed a positive correlation between increasing concentrations with rising elevation (Mast et al. 2012a, 2012b).

Table 6-6

Sediment Concentrations (ng/g Dry) Obtained by the National Oceanic and Atmospheric Administration (NOAA) National Status and Trends (NS&T) Program Between 2005 and 2009^a.

Bed sediment samples were collected from six stream sites in the Potomac River basin in 2007 and tested for chemical pollutants. α-Endosulfan was detected in 14% of samples collected at a maximum concentration of 0.01 μg/kg. No β-endosulfan or endosulfan sulfate was detected (Kolpin et al. 2013).

Endosulfan was detected in sediments obtained from 2 major rivers, 11 creeks or sloughs, 8 irrigation canals, and 2 tailwater ponds located in agricultural areas of California’s Central Valley. Samples were obtained during “peak use” periods between July and November 2002, and “winter” samples obtained in March 2003. Peak endosulfan concentrations were limited to ponds adjacent to lettuce fields, but a concentration of 17 ng/g was reported in Del Puerto Creek. The authors noted that endosulfan concentrations were generally below acute toxicity thresholds, but may have contributed to toxicity in a few tailwater ponds and irrigation canals with concentrations greater than several hundred ng/g (Weston et al. 2004).

Residential soil samples collected from 11 homes in Atlanta, Georgia in 2006 did not show β-endosulfan in either yard or foundation samples. The method detection limit for this study was 0.60 ng/g (Riederer et al. 2010).

6.4.4. Other Environmental Media

Endosulfan residues have been detected in a variety of the consumer products, as well as aquatic and terrestrial organisms.

The U.S. Department of Agriculture (USDA) monitors levels of endosulfans and endosulfan sulfate in commodity food items for its Pesticide Data Program. Endosulfan (α- and β-isomers) and endosulfan sulfate were detected in samples collected in 2010 from apples, asparagus, cantaloupe, cilantro, cucumbers, hot peppers, lettuce, mangoes, pears, and sweet bell peppers. The results are summarized in Table 6-7. Cucumbers typically had the highest rates of detection (25.3–38.1%), while asparagus and mangoes had very low detection rates (<1% for all three forms of endosulfan). Endosulfan sulfate residues were detected in 44.5% of cantaloupe sampled with a reported concentrations ranging from 0.005 to 0.064 ppm (USDA 2012). Levels of endosulfan and endosulfan sulfate in domestic foodstuffs were determined as part of FDA’s Total Diet Studies series (FDA 2005). The results of this monitoring study are summarized in Table 6-8. The highest mean concentrations were reported for endosulfan sulfate in items such as olive oil (0.01363 ppm), fresh/frozen summer squash (0.02050 ppm), peeled cucumber (0.01099 ppm), and fresh/frozen spinach (0.03654 ppm). Generally, concentrations were higher in fresh/frozen fruits and vegetables versus processed food products. USDA’s program analyzes a greater number of samples for each food item compared to FDA’s Total Diet Studies. It is important to note that the residue samples in USDA’s Pesticide Data Program tended to be higher when compared to results from FDA.

Table 6-7

U.S. Department of Agriculture (USDA) Pesticide Data Program: Distribution of Endosulfan Residues in Fruits and Vegetables (2010).

Table 6-8

Endosulfan Levels in Food Products Sampled for the 2003–2005 Market Basket Survey.

Studies of carrot and tomato crops sprayed with endosulfan 2–8 days prior to harvest showed that more pesticide remains in the pulp than in the juices of these vegetables. Washing and peeling the vegetables lowered the endosulfan concentration considerably (Burchat et al. 1998).

The NOAA’s Mussel Watch Program monitors contaminant levels in mussels and oysters in over 280 U.S. coastal sites. In EPA’s ecological fate and risk assessment report (EPA 2010a), monitoring levels were summarized from samples analyzed between 1994 and 2009 (see Table 6-9). Endosulfans (α- and β-isomers and endosulfan sulfate) were detected in 64% of samples with an average concentration of 2.0 μg/kg dry weight and a 90^th percentile concentration of 4.9 μg/kg dry weight.

Two recent studies monitored endosulfan levels in freshwater fish. In a study monitoring chemical contaminants in bass and carp species in southeastern U.S. rivers, endosulfans were detected at low concentrations (<1.2 ng/g wet weight) and mean concentrations were not calculated due to the large number of samples below the limit of detection. Samples were taken for the Mobile River Basin, Apalachicola-Chattahoochee-Flint River Basin, Savannah River Basin, and Pee Dee River Basin in 2004 (Hinck et al. 2008). Bass, carp, and catfish sampled from the Colorado River basin and tributaries had reported total endosulfan concentrations of <0.02 μg/g, but the largest concentrations (>0.07 μg/g) were found in carp and bass from the Gila River near Arlington, Arizona in August 2003 (Hinck et al. 2007).

USDA’s Pesticide Data Program continued its study of pesticide residues in domestic and imported catfish intended for human consumption. The catfish were mostly farm-raised. In 2010, 384 samples were analyzed for residues of α-endosulfan, β-endosulfan, and endosulfan sulfate. α-Endosulfan was only detected at the limit of detection (0.001 ppm) and β-endosulfan was not detected in any samples. Endosulfan sulfate was detected in 30 samples with a rate of detection of 7.8% and at concentrations ranging from the limit of detection (0.001 ppm) to 0.028 ppm (USDA 2012).

Endosulfan was also detected in the muscle of Pacific cod and Pacific halibut collected from coastal waters of Aleutian Islands, Alaska. Sampling areas were grouped according to their level of military activity. In the contemporary military group, the geometric mean concentrations of endosulfan in Pacific cod (n=18) and Pacific halibut (n=26) were 2.5 and 3.3 ng/g wet weight, respectively. In the historical military group, the geometric mean concentration in Pacific cod (n=13) was 5.3 ng/g wet weight and 4.6 ng/g wet weight in Pacific halibut (n=23). The reference group concentrations were 3.2 ng/g wet for Pacific cod (n=16) and 3.0 ng/g wet weight for Pacific halibut (n=13). Endosulfan was not detected in any of the fish samples of rock greenling (Miles et al. 2009).

Table 6-9

Concentration of Total Endosulfans (μg/kg Dry Weight) in Bivalves from the National Oceanic and Atmospheric Administration (NOAA) Mussel Watch Database.

Endosulfan was measured in the major food items of Koeye River grizzly bears in British Columbia, Canada in the fall of 2004. Measured mean concentrations in spawned pink salmon, spawned sockeye salmon, crab, mussels, fall terrestrial vegetation, and spring terrestrial vegetation were 4.21, 2.18, 0.0410, 2.23, 0.926, and 0.833 ng/g lipid weight, respectively (Christensen et al. 2013).

China is the world’s largest producer and exporter of fishery products. Seafood products including 6 species of shrimp, 2 species of crab, and 14 species of shellfish were collected in June and October 2005 and analyzed for various chemical contaminants. Samples were collected from the Guangdong Province, which borders the South China Sea. α-Endosulfan residues had a frequency rate of 3.8%, arithmetic mean of 0.04 ng/g wet weight and a range of 0.02–1.25 ng/g wet weight. β-Endosulfan residues had a frequency rate of 1.4% and a range of 0.02–0.29 ng/g wet weight. Endosulfan sulfate residues had a frequency rate of 1.4% and a range of 0.01–0.35 ng/g wet weight. Residues were found mostly in shellfish species Perna uiridis, Sinonovacula constricta, and Crassostrea gigas (Guo et al. 2007).

Endosulfans were detected in wine corks produced from the bark of the cork oak tree (Quercus suber), which are grown widely in regions of western Mediterranean. Wine corks were collected in spring 1999 from wines produced in Greece, wines produced in Cleveland, Ohio, and from a winery in Bloomington, Indiana (these corks were not used as bottle stoppers). The authors did not specify the origin of the corks used in these wine samples. Total endosulfan levels ranged from 3.8 to 29 ng/g lipid. Concentrations were generally higher in the samples obtained from Bloomington, Indiana. Variation of concentrations was high, and the authors suggested that this was due to differences in pesticide usage where cork oak trees are harvested. The production steps, which include the produce, the cork retailers, and the winery, may also contribute to these variations. Wine to cork exchanges may also have contributed to variations (Strandberg and Hites 2001).

Lichen samples collected from Yosemite National Park, California during the spring and summer of 2008 and 2009 were analyzed for endosulfans (α- and β-isomers and endosulfan sulfate). They were detected in 100% of the samples ranging from 2.0 to 24 ng/g dry weight. These concentrations also showed a positive correlation with increasing elevation, suggesting the occurrence of “cold-mountain trapping” (Mast et al. 2012a, 2012b).

Sparling et al. (2001) claimed that heavy pesticide use in the San Joaquin Valley is contributing to population decline of certain amphibians in the Sierra Mountains, which lies downwind from this agricultural area. Residue levels in Pacific tree frog (Hyla regilla) tadpoles and adults were analyzed from samples taken from coastal sites (used as controls), Lassen Volcanic National Park, Lake Tahoe, Yosemite National Park, and Sequoia National Park. Generally, endosulfan residues were often zero in the coastal sites and in Lassen Volcanic National Park, which are located west and north, respectively, of the San Joaquin Valley. The maximum concentrations (21.9 ppb) occurred in Sequoia or Yosemite National Park. Endosulfans along with DDx compounds also had the highest frequency of detection (~70–80%) throughout the Lake Tahoe, Sequoia, and Yosemite locations. Although endosulfan was not solely implicated, the authors concluded that there was a mixed but increasing occurrence of endosulfan residues along the west to east gradient, which is consistent with the declining amphibian populations in this area.

Endosulfans were detected in blubber of beluga whales (Delphinapterus leucas) collected from 15 sites in the Canadian Arctic at various times between 1993 and 2001. Geometric means for endosulfan (isomers not specified) ranged from 9.7 to 76.3 ng/g wet weight from samples taken from males and females of Resolute Bay (1996, 1999), Grise Fjord (2000), Igloolik (1995, 1997), Coral Harbor (2000), and Arviat (1999). Endosulfan sulfate concentrations (geometric means) ranged from 7.0 to 70.6 ng/g wet weight from sites in Chesterfield Inlet (1997, 1999), Sanikiluaq (1994, 2000), Cape Dorset (1999, 2000), Kimmirut (1994, 1996), Iqaluit (1992, 1996), and Pangnirtung (1996, 1997). Endosulfan along with α-hexachlorocyclohexane, comprised the larger proportion of organochlorine residues in the northern Hudson Bay areas. Concentration variations between the sites were statistically significant but concentrations between male and female samples were not (Stern et al. 2005).

Endosulfan was also detected in other higher trophic aquatic organisms. A study analyzed pesticide concentrations in Bonnethead sharks (Sphyrna tiburo) in the Florida estuaries of Apalachicola Bay (9 females, 13 males), Tampa Bay (17 females, 15 males), Charlotte Harbor (5 females, 5 males), and Florida Bay (18 females, 13 males). β-Endosulfan and endosulfan sulfate were not detected in liver samples from any of the sampling locations. β-Endosulfan geometric mean concentrations ranged from 12.77 to 15.13 ng/g in muscle and 1.55–6.60 ng/mL in serum. Endosulfan sulfate was only detected in liver samples from Charlotte Bay (geometric mean 1.00±037 ng/g) and serum samples from Florida Bay (geometric mean 1.93±0.93 ng/mL) (Gelsleichter et al. 2005). Skin and blubber samples (post-mortem) from two blue whales (Balaenoptera musculus) stranded off the coast of Baja, California were analyzed for chlorinated hydrocarbons. The two whales were juvenile males, both approximately 18 m long. Endosulfan sulfate was detected at approximately 70 ng/g of lipid in the blubber of whale 1 and at approximately 230 and 380 ng/g of lipid in the skin of whales 1 and 2, respectively (Valdez-Marquez et al. 2004).

6.8.1. Identification of Data Needs

Physical and Chemical Properties. The physical/chemical properties of endosulfan and endosulfan sulfate are sufficiently well characterized to enable assessment of the environmental fate of the compound (HSDB 2009, 2010; NIOSH 2011; O’Neil et al. 2006; Tomlin 2003).

Production, Import/Export, Use, Release, and Disposal. According to the Emergency Planning and Community Right-to-Know Act of 1986, 42 U.S.C. Section 11023, industries are required to submit substance release and off-site transfer information to the EPA. The TRI, which contains this information for 2009, became available in February of 2011. This database is updated yearly and should provide a list of industrial production facilities and emissions.

Although all U.S. producers of endosulfan must be registered under FIFRA, data concerning quantities of endosulfan produced domestically are limited. Annual production can be estimated by analogy to its annual use, for which quantitative data are available. The available use trends over the last two decades indicates use of endosulfan is in decline, based on a reported annual use of 1.38 million pounds between 1990 and 1999 to only 380,000 pounds per year between 2006 and 2008 (EPA 2002, 2010a). As of March 2012, there were only four active registrants of endosulfan in the United States. Over the 4-year phase-out schedule, these registrants will be cancelling their endosulfan products according to a voluntary agreement (EPA 2012e). Despite this cancellation, production data and continued recordkeeping of endosulfan use would be valuable during the years of cancellation and the years following, so as to understand the impact and its effects on exposure to the public and to workers.

Releases and disposal of endosulfan to the environment are well defined based on the regulatory restrictions and the available monitoring data (CDPR 2011; EPA 2001, 2002, 2010a, 2012c, 2012e; NOAA 2012; USGS 2012c).

Environmental Fate. Overall, the environmental fate mechanisms associated with endosulfan are well documented. Endosulfan partitions to air and is subject to long-range transport (EPA 2010a; Hafner and Hites 2003; Hageman et al. 2006a; Hung et al. 2005, 2010; Kathpal et al. 1997; Mast et al. 2012a, 2012b; Rice et al. 2002; Rudel 1997; Stern et al. 2005; Weber et al. 2010). It is immobile in soils (EPA 2010a; Stewart and Cairns 1974). It is transformed in surface waters and soils via hydrolysis (Greve and Wit 1971; HSDB 2010; Kaur et al. 1998) and biodegradation (Cotham and Bidleman 1989; HSDB 2010).

Bioavailability from Environmental Media. Endosulfan can be absorbed following inhalation of contaminated workplace air and ingestion of insecticide-contaminated food (Ely et al. 1967). Dermal contact with or ingestion of endosulfan that is tightly bound to soil particles is an exposure route of concern at hazardous waste sites. No quantitative information is available on the absorption of endosulfan in either adults or children following ingestion or dermal contact with contaminated soils. Therefore, additional information is needed on the uptake of endosulfan from contaminated soil following ingestion or dermal contact. This information would be useful in determining the bioavailability of soil-bound endosulfan.

Food Chain Bioaccumulation. Endosulfan is bioconcentrated by aquatic organisms and supporting data are well established (DeLorenzo et al. 2002; EPA 2010a; Ernst 1977; Pennington et al. 2004; Rajendran and Venugopalan 1991; Roberts 1972; Schimmel et al. 1977; Weber et al. 2010). BMFs from fish to aquatic mammals suggest the potential for biomagnification in aquatic food chains. However, these estimates do not take metabolism into account (Kelly et al. 2007; Weber et al. 2010). Estimating bioconcentration and biomagnification in terrestrial organisms is also difficult due to the lack of biomonitoring data, especially in humans. Recent BMF estimates indicate potential for biomagnification in humans, but understanding of continuous, low-dose metabolism in terrestrial organisms is also lacking (Kelly et al. 2007). Acute poisoning cases suggest that metabolism in humans may be rapid (Eyer et al. 2004). Methods for estimating biomagnification in terrestrial organisms based on octanol/air and octagonal/water partition coefficients require an understanding of its metabolic potential (Armitage and Gobas 2007; EPA 2010a). Further investigations into the bioaccumulation, biomagnification, and metabolism of endosulfan in humans and terrestrial organisms are needed.

Exposure Levels in Environmental Media. Reliable monitoring data for the levels of endosulfan in contaminated media at hazardous waste sites are needed so that the information obtained on levels of endosulfan in the environment can be used in combination with the known body burden of endosulfan to assess the potential risk of adverse health effects in populations living in the vicinity of hazardous waste sites.

Monitoring data for endosulfan in environmental media are current and extensive. It has been detected in ambient air of temperate and Arctic regions (Daly et al. 2007; Gale et al. 2009; Harner et al. 2005; Hoh and Hites 2004; Hung et al. 2005, 2010; Pozo et al. 2006; Shen et al. 2005; Sun et al. 2006; Weber et al. 2010). Endosulfan is a monitored in groundwater and surface water samples through the USGS NAWQA program, EPA’s STORET database, CDPR Surface Water Protection Database, USDA Pesticide Data Program, and SFWMD DB Hydro database, etc. Altogether, these sources contain over 20 years’ worth of monitoring data for endosulfan in the United States (EPA 2010a). Additional studies have also detected endosulfan in surface water (Harman-Fetcho et al. 2005; Pfeuffer 2011; Scott et al. 2002). Studies have also detected endosulfan in rainwater and snow samples (Hageman et al. 2010; Kuang et al. 2003; Mast et al. 2012a, 2012b). It has been detected in sediment and soil (NAWQA; NOAA NS&T) (Mast et al. 2012a, 2012b; Riederer et al. 2010; Weston et al. 2004). Endosulfan levels in bivalves are monitored as part of the NOAA’s Mussel Watch Program and have been summarized by EPA (2010a). The FDA’s Total Diet Studies (2005) and USDA’s Pesticide Data Program (USDA 2012) detected endosulfan residues in a variety of food products, mostly in fresh and frozen fruits and vegetables. Residues have been detected in fresh and seawater fish and seafood, including catfish caught and raised for human consumption (Guo et al. 2007; Hinck et al. 2007, 2008; Miles et al. 2009; USDA 2012). They have been detected in lichen samples from Yosemite National Park, California (Mast et al. 2012a, 2012b) and wine corks made from cork oak tree (Q. suber) (Strandberg and Hites 2001). Endosulfans have been detected in the Pacific tree frog (H. regilla) of the Sierra Mountains and California, downwind from the agricultural San Joaquin Valley (Sparling et al. 2001). Endosulfans have been detected in high-trophic aquatic organisms such as beluga whales, blue whales, and Bonnethead sharks (S. tiburo) (Gelsleichter et al. 2005; Stern et al. 2005; Valdez-Marquez et al. 2004).

Exposure Levels in Humans. Comprehensive biomonitoring studies for endosulfan are not available. Exposure levels in the general population have been extensively evaluated using available monitoring data. Dietary intake is expected to be the major source of endosulfan exposure to the general public. Dietary exposures were extensively evaluated for various population subgroups by both EPA and CDPR using different data sets and analytical methods. Although the estimated chronic dietary exposures from these studies were significantly different, both values were below protective benchmarks. The major difference between these exposure assessment deals with the incorporation of drinking water into dietary intake. CDPR assumed that drinking water was not a significant source of exposure based on the monitoring data for California. However, EPA incorporated drinking water into their assessment and their model results estimated drinking water exposure to be the same as food intake (Silva and Carr 2010). Comprehensive biomonitoring studies using samples collected from the U.S. population that can be used to better assess endosulfan exposure in the general population would be valuable. CDPR evaluated short-term, bystander inhalation exposure and estimated long- and short-term exposures for child and adult swimmers (Beauvais et al. 2010a). Lee et al. (2002) estimated inhalation HQs using CDPR data from 1990 to 2000.

This information is necessary for assessing the need to conduct health studies on these populations.

Exposures of Children. Dietary exposure assessments were conducted by EPA and CDPR, which included estimates for child subgroups. These assessments used monitoring data in food and drinking water rather than biomonitoring data from urine or blood samples. Results from these assessments indicated that childhood exposures were generally higher compared to exposure estimates for the general adult population (Silva and Carr 2010). Several studies measured endosulfan levels in breast milk, placenta, umbilical cord blood, and maternal adipose tissue, and some explored correlations between elevated endosulfan levels and reproductive malformations in male infants (Cerrillo et al. 2005; Damgaard et al. 2006; Fernandez et al. 2007; Freire et al. 2011; Shen et al. 2007). These studies were conducted in Europe and may not be representative of U.S. exposures. Comprehensive biomonitoring studies using samples collected from the U.S. population would be valuable in assessing endosulfan exposure in children. More data would also be helpful to properly assess endosulfan exposure to children who live, play, or attend school near farmlands that are treated with endosulfan. CDPR estimated inhalation exposure levels under a bystander scenario, but this analysis incorporated several assumptions (Beauvais et al. 2010a). The possibility that farming parents’ work clothes and shoes may carry endosulfan residues into the home also should be studied.

Child health data needs relating to susceptibility are discussed in Section 3.12.2, Identification of Data Needs: Children’s Susceptibility.

Exposure Registries. Although endosulfan is monitored extensively in food, surface water, groundwater, air, and media, it is not included in comprehensive biomonitoring studies that analyze tissue, blood, or urine. Endosulfan is not included in the CDC’s National Health and Nutrition Examination Survey (NHANES), which monitors chemical concentrations in urine and blood collected from the U.S. population. These data would give the best basis for estimating exposures to the general population, children, and workers. Endosulfan will be gradually phased-out through 2016 and exposures are expected to decrease. As a result, large-scale biomonitoring and environmental monitoring programs will likely be decreased or discontinued. Continuing studies would be valuable in understanding the immediate and long-term impacts on the general population.

6.8.2. Ongoing Studies

No ongoing studies sponsored by NIH or EPA were identified for endosulfan.