UNITED STATES ENVIRONMENTAL PROTECTION AGENCY

WASHINGTON D.C., 20460

OFFICE  OF 

PREVENTION, PESTICIDES AND 

TOXIC SUBSTANCES

MEMORANDUM					DATE:  October 31, 2007

				DP Barcode:  D346213

				PC Code:  079401 

SUBJECT:	Addendum to the Ecological Risk Assessment of Endosulfan 

FROM:	Keith Sappington, Ph.D., Senior Biologist

		Faruque Khan, Senior Agronomist

		Endosulfan Chemical Team

		Thomas Steeger, Ph.D., Senior Biologist

		Reviewer

		Environmental Fate and Effects Division (7507P)

		

THROUGH:	Mah Shamim, Ph.D., Branch Chief

		Environmental Fate and Effects Division (7507P)

TO:		Tracy Perry, Risk Manager Reviewer

		Robert McNally, Risk Manager

		Special Review and Reregistration Division (7508P)

In 2000, the Environmental Fate and Effects Division (EFED) completed
its draft environmental fate and ecological risk assessment chapter
(ERA) in support of the reregistration eligibility decision (RED) for
endosulfan. The ERA was later revised to reflect public comments and the
final version of the science chapter was completed in February 2002
(USEPA, 2002; DP Barcode D238673).    SEQ CHAPTER \h \r 1 At that time,
the chapter incorporated nonconservative assumptions in its
deterministic aquatic risk assessment and included a probabilistic
exposure assessment for aquatic impacts as well.  In addition, because
the risk assessment did not take into account the toxicity of the
primary transformation product, i.e., endosulfan sulfate, the overall
terrestrial and aquatic risk of endosulfan may have been underestimated.
 

Subsequent to the publication of the final 2002 ecological risk
assessment, new information related to endosulfan toxicity,
bioaccumulation, monitoring and transport, and ecological incidence have
been obtained by OPP.  The purpose of this addendum is to conduct a
preliminary assessment of this new information in order to: (1) address
the extent to which the previous ecological risk assessment for
endosulfan might change in relation to this new information, and (2)
indicate future avenues where additional data analysis and risk
characterization endosulfan are needed.  Areas of focus include new
information on endosulfan bioaccumulation, ecological effects, exposure
modeling, and monitoring and long-range transport.  The findings from
each of these focus areas are summarized below.  More detailed
information in support of these findings is contained in Appendix 1 and
its attachments (under separate cover).

EXPOSURE CHARACTERIZATION

  	Persistence

Endosulfan is expected to be more persistent than how it was
characterized in the 2002 ERA.  This conclusion is based on the newly
submitted toxicity data on endosulfan sulfate, the major degradate of
endosulfan, which shows that endosulfan sulfate is similar in toxicity
to the parent compound.  Therefore, EFED has used the total residue
method to estimate exposure concentrations for endosulfan and endosulfan
sulfate.  This means that the Agency has based its exposure estimates on
the half life of total endosulfan residue, a comparatively longer half
life than the parent compound.  The total residue aerobic soil half-life
is 1336 days compared to the parent only half-life of 57 and 208 days
for α  and β endosulfan, respectively.

	Exposure Modeling

Revised exposure modeling was conducted based on total endosulfan
residues (α+β+sulfate) for two crop scenarios with the highest
predicted estimated exposure concentrations (EEC): Florida tomatoes and
California strawberries.  Predicted surface water EECs increased by
20-40% compared to EECs used in the 2002 ERA, which reflects the
addition of endosulfan sulfate in the exposure modeling. Predicted peak,
21-d mean, and 60-d mean surface water EECs are: 23, 9.3, and 6.8 ppb
(Florida tomato scenario).  Sediment pore water EECs of about 4 ppb were
predicted using this same crop scenario and did not vary across these
averaging periods.  The EECs predicted from the California strawberry
scenario are about half those of Florida tomato.  However, both
California strawberry and Florida tomato EECs may be underestimated
because the use of plastic mulch practices, which are expected to
increase endosulfan runoff, could not be modeled accurately.  

	Monitoring and Long-Range Transport

Numerous literature sources have documented the regional or long-range
transport of endosulfan through various environmental media such as air,
precipitation, water, and sediment. Monitoring studies have detected
endosulfan or endosulfan sulfate in environmental media with regular
frequency in both agriculturally influenced and remote regions.  Recent
studies suggest that desorbed residues of endosulfan volatilize and
continue to recycle in the global system through a process of migration
and redeposition (via wet and dry depositions) as well as air-water
exchange in the northern Hemisphere.  Findings from selected monitoring
studies of endosulfan are summarized below within various environmental
media or regions.

Surface Water

The presence of endosulfan in many ecosystems throughout the U.S. has
led to concern regarding continuing point and non-point sources of
endosulfan in vulnerable areas that may result in acute and chronic
effects on aquatic communities.  These data suggest that in the vicinity
of row crops where endosulfan is reportedly applied, endosulfan residues
have been routinely detected in both the water column and benthic
sediments.  Additionally, the data indicate that residues exceed OPP
acute and chronic risk levels of concern (e.g., south Florida and
California) and that total endosulfan residues have moved to areas
distant from where it was initially applied. Furthermore, in 1998,
Calleguas Creek in California and the Yakima River in Washington State
were classified as impaired water bodies under the Section 303d of the
Clean Water Act due to presence of endosulfan as well as other chemicals

Sediments

etween 1980 and 1999, 199 detections were reported for α-endosulfan
(ranging from 0 to 11000 μg·Kg-1).  For β-endosulfan, there were 667
detections reported (ranging from 0 to 67500 μg·Kg-1).  For endosulfan
sulfate, there were195 detections reported (ranging from 0.2 to 900
μg·Kg-1).

Air

Abundant regional concentration data are available for the Great Lakes
Region from the Integrated Atmospheric Deposition Network (IADN), a
joint US EPA / Environment Canada-monitoring project (Sun et al., 2003,
3006).  These data provide compelling evidence for medium-range airborne
transport of endosulfan and endosulfan sulfate, as endosulfan
concentrations in vapor phase show a clear increasing trend from the
west to east across the region. 

Remote Regions 

in south-central Canada show the presence of α-and β-endosulfan (Muir
et al., 2004). Mean concentration levels of α-endosulfan ranged from
28.5 – 1.3 pg L-1, and those of β-endosulfan ranged from 10.3 – 0.0
pg L-1 in lakes Opeongo, Nipigon, Britt Brook, and Virgin Pond. No
agricultural area was within 31 miles (50 Km) of any of these lakes,
suggesting that the presence of endosulfan resulted from atmospheric
transportation and deposition. Monitoring and modeling results suggest
that under the conditions prevailing in south-central Canada, endosulfan
can potentially undergo regional-scale atmospheric transport and reach
lakes outside endosulfan use areas. 

Precipitation and Snowpack

ion of the Great Lakes region, α- and β-endosulfan concentrations were
regularly determined by IADN at various stations during the period of
1987–1997. Concentration levels of α-endosulfan ranged from 0.13 –
1.95 ng·L-1 and those of β-endosulfan ranged from 0.19 – 6.09
ng·L-1 in Lake Superior and Lake Erie. Higher values were reported from
Lake Michigan ranging from 0.54 – 8.22 ng·L-1 for α- and from 1.06
– 12.13 ng·L-1 for β-endosulfan. Endosulfan and endosulfan sulfate
were detected in seasonal snow pack samples at six national parks in the
Western United States (Hagman et al., 2006). Concentrations of total
endosulfan ranged from 1.5 ng·L-1 to 0.0040 ng·L-1 in Sequoia, Mount
Rainier, Denali, Noatak-Gates, Glacier and Rocky Mountain National
Parks.

Mountainous Areas

The effect of "global distillation" which is believed to account for
transport of persistent organic pollutants (POPs) whereby a compound
could volatilize from warmer regions, undergo long-range atmospheric
transport and subsequently recondense to an accumulation of these
substances in the temperate, higher mountainous and Arctic regions.

 Levels of α-endosulfan found in rain were in a range of < 0.0035 ng/L
to 6.5 ng L-1 while β-endosulfan was determined at concentrations of <
0.012 ng L-1 up to 1.4 ng L-1 McConnell et al. (1998).

Arctic Region

Endosulfan has routinely been detected in arctic regions.  As excerpted
from GFEA (2007): 

“ Endosulfan was reported as a widely distributed pesticide in the
atmosphere of Northern polar regions [37]. Unlike for most other
organochlorine pesticides average concentrations of endosulfan in the
Arctic have not changed significantly during the last five years [38].
Concentrations of endosulfan (isomers unspecified) from Arctic air
monitoring stations increased from early to mid-1993and remained at that
level through the end of 1997 at 0.0042-0.0047 ng/m3.”

Bioaccumulation

Three lines of evidence were used to evaluate and characterize the
bioaccumulation potential of endosulfan and its metabolites: (1)
empirical bioconcentration and bioaccumulation experiments, (2)
bioaccumulation modeling in aquatic food webs, and (3) bioaccumulation
modeling in terrestrial food webs. Based on these lines of evidence, the
following conclusions are drawn from this review.

Empirical data on endosulfan bioconcentration and bioaccumulation are
compromised substantially by data quality concerns (i.e., none of the
studies appear to meet OPP data quality guidelines) and therefore, BCF
and BAF values from these studies should be interpreted with caution. 

Given these caveats, measured BCF values in fish based on the highest
quality studies range from about 1000 to 3000 (L/kg wet wt.).  This
range in BCF values is also supported by aquatic bioaccumulation
modeling which predict mean and 90th percentile BCFs in the range of
1000 to 2000 L/kg wet wt.  Considering all studies, BCF values exceeded
5000 for one of seven fish species (yellow tetra), which approximated
5700 (based on the BCF ratio method) and 11,600 (based on the kinetic
BCF method).  Substantial uncertainties in the BCF values from the
yellow tetra study are noted, however, as discussed herein.  

sulfan (α and β isomers) and its primary degradate (endosulfan
sulfate) appears rapid in fish (half life 2-6 days) based on limited
data.  However, the ability of fish and other aquatic organisms to
depurate endosulfan in the environment may be ameliorated in areas of
broad-scale endosulfan contamination (e.g., certain
agriculturally-influenced ecosystems and even selected remote ecosystems
receiving endosulfan loadings through long-range transport).

 

Based on one study with daphnids and preliminary aquatic bioaccumulation
modeling results, biomagnification of endosulfan appears unlikely in
typical aquatic food webs.  This finding is consistent with other
bioaccumulation modeling studies published in the literature for
endosulfan and related compounds.

Based largely on indirect evidence from terrestrial bioaccumulation
modeling, the biomagnification of endosulfan by terrestrial organisms
appears to be a concern with predicted biomagnification factors ranging
from 2.5 to 28 for herbivorous and carnivorous wildlife at the top of
arctic food webs.  Additional information on endosulfan in terrestrial
food webs should be evaluated to more fully characterize the
biomagnification of endosulfan by terrestrial organisms.  

EFFECTS CHARACTERIZATION

α and β) and the endosulfan sulfate degradate.  The new endosulfan
sulfate toxicity data consisted of:

acute toxicity with two species of estuarine/marine fish 

acute and chronic toxicity to one estuarine/marine invertebrate

sediment toxicity with one freshwater and one estuarine/marine
invertebrate

acute oral and dietary toxicity with two avian species. 

Based on comparisons made within the same species, the toxicity of
endosulfan sulfate and the parent isomers (α, β) generally appears
similar based on acute studies with bluebill sunfish and mallard duck,
but somewhat less toxic with bobwhite quail.  This information was used
to support the assumption that endosulfan sulfate is similar in toxicity
to the parent compounds.  New information on the toxicity of endosulfan
sulfate did not alter the toxicity values used in determining risk
quotients for aquatic organisms via water column exposure.

h–g

␱币炄怈悄懺̤摧┞

% h–g

 

	̤̀␱愁̤摧┞

㄀Ĥ摧┞

ༀ킄ሂĀ㄀Ĥ葞ː摧┞ᜀ 

£

␃؃Ĥ萏ː萑ː␱币킄怂킄愂̤摧┞

̤̀☊଀͆㄀Ĥ␷㠁Ĥ⑈愁̤摧┞

̤̀萑Ũ␱态梄愁̤摧┞

̤̀萑ː␱态킄愂̤摧┞

̤̀☊଀͆㄀Ĥ␷㠁Ĥ⑈愁̤摧┞	̤̀␱愁̤摧┞ሀosures
with an estuarine/marine amphipod.

RISK CONCLUSIONS 

it is of similar toxicity to the parent isomers (α and β).   Based on
the tomato crop scenario that yielded the largest EECs, acute and
chronic RQ values range from about 30 to 60 for freshwater fish,
respectively (compared to 23 and 44 from the 2002 ERA) to about 230 and
680 for estuarine/marine fish, respectively (compared to 190 and 490
from the  2002 ERA). Acute and chronic RQ values for invertebrates range
from approximately 4 to 130 for freshwater (compared to 3.3 and 93 from
the 2002 ERA) and from 50 to 190 for estuarine/marine (compared to 42 to
130 from the 2002 ERA), respectively.  Findings from ecological
incidents support the findings of endosulfan risk to fish and
invertebrates.

Risks to freshwater and estuarine/marine invertebrates resulting from
sediment exposure to endosulfan are evident from the integration of
exposure and effect characterization.  The RQ values for
sediment-dwelling invertebrates range form 0.9 to 2.8 depending on
species. These RQ values are based on predicted total endosulfan
residues compared to endosulfan sulfate toxicity values, which assumes
similar toxicity of endosulfan sulfate and total endosulfan residues. 

Based on preliminary results from an aquatic food web bioaccumulation
model, risks to piscivorous wildlife appear relatively modest with mean
predicted acute RQ values exceeding the Agency acute LOC of 0.1 for one
of eight species modeled (0.15 for river otter) and 90th percentile
estimates exceeding the LOC for three of eight species modeled (0.18,
0.39, 0.20 for mink, river otter and belted kingfisher, respectively). 
Predicted chronic RQ values did not exceed the Agency LOC for any of the
eight species modeled.

Risks to nontarget terrestrial wildlife did not change from the 2002 ERA
as a result of this addendum, because currently available terrestrial
exposure models could not address total residue exposure.  Based on the
2002 ERA, RQs for birds and mammals exceed the Agency’s acute and
chronic risk LOCs and range up to a maximum of 2.7 for birds and 40 for
mammals.

㄀Ĥ摧┞

̤̀萑ː␱态킄愂̤摧┞	̤̀␱愁̤摧┞ᤀ Monitoring data
and incident reports confirm that endosulfan is moving through aquatic
and terrestrial food chains and that its use has resulted in adverse
effects on the environment adjacent to and distant from it registered
use sites.

 Meakin, S. What´s New with POPs Research in the Arctic. Northern
Perspectives 26 (1), 6-7 (2000)

Page   PAGE  1  of   NUMPAGES  7 

