-
An Ecological Risk Assessment for Chlorpyrifosin an
Agriculturally Dominated Tributaryof the San Joaquin River
Nicholas N. Poletika,1* Kent B. Woodburn,2 and Kevin S.
Henry2
Single-species toxicity testing of ambient water samples and
national-scale probabilistic riskassessment have implicated the
organophosphorous (OP) insecticide chlorpyrifos (O,O-diethyl
O-(3,5,6-trichloro-2-pyridyl)-phosphorothioate) as a potential
chemical stressor ofaquatic organisms residing in the lower San
Joaquin River basin. This site-specific aquaticecological risk
assessment was conducted to determine the probability of adverse
effectsoccurring from exposure to chlorpyrifos in an agriculturally
dominated tributary of the SanJoaquin River and to assess the
ecological significance of such effects. Assessment endpointswere
fish population persistence and invertebrate community
productivity. Daily chemicalmeasurements collected over a period of
one year were analyzed temporally for frequency,duration, and
spacing between events for acute and chronic exposure episodes.
Effectsthresholds for fish and freshwater lotic invertebrates were
determined from single-specieslaboratory toxicity tests. Potential
risk was characterized by the degree of overlap ofdistributions of
exposure events and effects, with consideration given to additive
toxicity ofother OP insecticides, recovery periods, and duration of
chronic exposure ( 21 d).Ecological significance was determined by
analysis of fish assemblage dietary andreproductive habits in
relation to the surrogate invertebrate taxa judged at risk. Results
ofanalysis indicated no direct effects on fish, and indirect
effects on fish through elimination ofinvertebrate food items were
considered unlikely. Biological survey information will benecessary
to address uncertainty in this risk conclusion, especially as it
relates to the benthicinvertebrate community. Results of this
site-specific risk analysis suggest that fish populationpersistence
and invertebrate community productivity were not adversely affected
bymeasured chlorpyrifos residues during a year-long monitoring
period.
KEY WORDS: Chlorpyrifos; diazinon; aquatic risk assessment;
ecological significance; agriculturallydominated tributary
1. INTRODUCTION
1.1. Background
Chlorpyrifos is a widely used agricultural OPinsecticide.
Although only one freshwater aquatic
incident report has been compiled by the U.S.Environmental
Protection Agency (U.S. EPA) inthe 30 years this product has been
marketed foragricultural use,(1) toxicity testing in California(2)
andpreliminary risk assessments for U.S. EPA
pesticidereregistration(3) have suggested the potential foradverse
ecological effects of chlorpyrifos in aquaticsystems. A recent
probabilistic risk assessment forchlorpyrifos in North American
aquatic ecosystems
1 Dow AgroSciences, Indianapolis, IN.2 The Dow Chemical
Company.* Address correspondence to Nicholas N. Poletika, Dow
Agro-
Sciences LLC, 9330 Zionsville Road, Indianapolis, IN 46268.
Risk Analysis, Vol. 22, No. 2, 2002
291 0272-4332/02/0400-0291$22.00/1 2002 Society for Risk
Analysis
-
has addressed this potential problem on a nationalscale.(3) The
conclusions of the assessment were that,overall, chemical
monitoring data in freshwatersystems do not suggest ecologically
significant risks,except in a few locations. The authors
recommended,however, conducting site-specific risk assessments
forthese few locations to refine the risk characterizationand to
reduce the uncertainty associated with thenational assessment. One
area identified as possiblyexperiencing risk was a region in the
San JoaquinValley of California with heavy chlorpyrifos use
thatcontributes agricultural drainage to the perennialreach of the
San Joaquin River. The tributaries inthis region thus appeared to
be good candidates for asite-specific ecological risk assessment
for the stres-sor chlorpyrifos.
1.2. Objectives
The primary objectives of this risk assessmentwere (1) to
estimate the probability of chlorpyrifosuse to cause adverse
effects in various taxa ofaquatic organisms in an agriculturally
dominatedtributary of the San Joaquin River, and (2) todetermine
the ecological relevance of any predictedrisk in the local aquatic
community. Ideally, detailedinformation on exposure and effects
would becoupled with site-specific biological and habitatsurvey
information in performing this type ofassessment. Intensive
chemical monitoring data areavailable for a site where significant
chlorpyrifos useoccurs within the San Joaquin Valley,(4) and
thenational-scale assessment provides an excellent setof effect
endpoints, derived from laboratory toxicitytests and
microcosm/mesocosm studies, which aregenerally applicable to
theoretical assemblages ofvertebrate and invertebrate freshwater
organisms.(3)
Unfortunately, little biological survey work has beenconducted
in the agricultural regions of California,(5)
nor has biocriteria development progressed to thepoint where
established indices of biological integ-rity are available to
interpret the effects of varioustypes of human activity on
California surfacewaters.(6)
2. PROBLEM FORMULATION
2.1. Site Description
Detailed chemical monitoring, stream flow, andweather data are
available for the lower reach ofOrestimba Creek,(4) an
agriculturally dominated
natural drainage in Stanislaus County, California(Fig. 1).
Orestimba Creek originates in the CoastRange of mountains in
western Stanislaus County,passes through irrigated farmland in the
San JoaquinValley at elevations of 66 to 20 m above sea level,and
terminates at its confluence with the SanJoaquin River. The most
important crops in thestudy area receiving insecticide applications
arealfalfa, walnuts, almonds, and dry beans.
Recent assessments of fish communities in theNational Water
Quality Assessment (NAWQA)Program San Joaquin-Tulare Basin study
unit wereperformed by the U.S. Geological Survey (U.S. GS)in
conjunction with habitat surveys and waterchemistry
determinations.(5) The San Joaquin MainStem group was characterized
by high specificconductance (high salinity), decreased fish
cover,and high percentage of agricultural land. Largepercentages of
introduced fish species tolerant ofaltered environmental conditions
were found atthese sites: fathead minnow (Cyprinidae:
Pimiphalespromelas), red shiner (Cyprinidae: Cyprinellalutrensis),
threadfin shad (Clupeidae: Dorosoma
N
0 10 Kilometers
0 100 Kilometers
StanislausCounty
SanJoaquinR
.
San J
oaqu
inR.
Orestimba C
r.
AgriculturallyDominated Reach
Fig. 1. Location of study area.
292 Poletika, Woodburn, and Henry
-
petenense), and inland silverside (Atherinidae: Men-idia
beryllina). Minor components of the SanJoaquin Main Stem group
include, in order ofdecreasing importance, other introduced
species,largemouth bass and sunfish, catfish, native species,and
smallmouth bass. Incidence of external abnor-malities (parasites
and lesions) in San Joaquin MainStem resident fish individuals
averaged 17%, indi-cating impaired conditions may exist.(5)
The benthic invertebrate species assemblagespecific to Orestimba
Creek is unknown. However, arecent paper by Brown and May(7)
examined benthicinvertebrates in lotic systems in the lower
Sacra-mento and San Joaquin river drainage basins. Theyfound that
tributaries have statistically greater spe-cies richness than the
main stem rivers and drainagewatersheds (Orestimba Creek and
others, N 7), adifference likely due to land-use practices.
Theauthors report that the predominate benthic inverte-brate taxa
in the drainage watersheds are Odonata,Oligochaeta, Ephemeroptera,
and Diptera, and thatthe drainage watersheds have substrates of a
finerparticle size.
2.2. Conceptual Model
In view of the site description given above, wedecided to
evaluate the risk of chlorpyrifos use inthe agriculturally
dominated reach of OrestimbaCreek (Fig. 1) for ecologically
significant adverseeffects on the principal components of the
SanJoaquin Main Stem fish assemblage. It would bepreferable to
focus the assessment on native speciesof concern across the entire
Sacramento-San Joa-quin River system rather than on these
introducedspecies tolerant of altered environmental condi-tions.
However, there is no site-specific informationavailable to
determine whether particular nativespecies inhabited the lower
reach of OrestimbaCreek before commercial agriculture was
intro-duced or whether the current habitat would supportnative fish
of concern should restoration beattempted.
Important invertebrate components of fishdiets were identified
from the literature andassumed to be present in the creek for
riskcharacterization using surrogate species in thetoxicity
database. Thus, both direct effects on fishand invertebrates and
indirect effects on fishthrough impact on their diet were
evaluated. Basedon the toxicity profile of chlorpyrifos(8) and
thereview by Giesy et al.,(3) effects on aquatic plants
and microorganisms were judged unlikely to occur.The toxicity
data also indicate that, as a taxonomicgroup, aquatic invertebrates
are more sensitive tothe broad-spectrum insecticide chlorpyrifos
thanare fish; therefore, for a given exposure concentra-tion, more
severe direct effects were expected withinvertebrate organisms.
2.3. Assessment Endpoints and Measures of Effect
The ecological entities of value in this assess-ment were fish
and benthic invertebrates (plank-tonic invertebrates were assumed
to be minorcomponents of this lotic system). Ecologically rele-vant
characteristics of these valued entities requir-ing protection were
fish population persistence andinvertebrate community productivity.
Indicators ofchlorpyrifos effects included the concentration
caus-ing death in 50% of a population of organisms(LC50), and, for
chronic effects, the no-observed-effect concentration (NOEC). The
effect endpointsdetermined in Giesy et al.,(3) as discussed below
inSection 4.3, were used as measures of adverseimpact.
2.4. Analysis Plan
The analysis plan for the assessment consistedof the following
elements: stressor characteriza-tion, exposure analysis, effects
analysis, riskcharacterization, and analysis of uncertainty
andvariability. Stressor characterization includeddescriptions of
chlorpyrifos physicochemical prop-erties, environmental fate, and
mechanism ofaction. Reviews of chlorpyrifos properties andfate(9)
and mechanism of toxicity(3) were brieflysummarized.
Exposure analysis focused principally on chlor-pyrifos use in
the watershed and the temporalpatterns of chemical pulses
(frequency, duration,and intervals between pulses). Previously
reportedOP insecticide dissolved concentrations present indaily
time-proportional composite samples collectedfor the period of May
1996 through April 1997(4)
were examined with the data analysis computerprogram RADAR (Risk
Assessment Tool to Evalu-ate Duration and Recovery) (Waterborne
Environ-mental, Inc., Leesburg, VA). Using the dailycomposite
concentration data, exposure events wereidentified by RADAR in
terms of exceedance of athreshold concentration value appropriate
for acuteor chronic effects.
Risk Assessment for Chlorpyrifos 293
-
Effects analysis evaluated effects on individualspecies,
populations, and communities (fish andinvertebrate). Several
measures of effects associatedwith acute and chronic exposure were
compared tothe pattern of exposure events identified byRADAR to
determine which endpoints wereappropriate.
Risk characterization related the previous ana-lysis to the
ecologically relevant assessment end-points of fish population
persistence andinvertebrate community productivity. Where
appro-priate, additive toxicity from co-occurring OP insec-ticides
was also considered. Surrogate speciespresent in the laboratory
toxicity database wereselected to represent fish species identified
byNAWQA as components of the San Joaquin MainStem group fish
assemblage. Invertebrate food itemsconsumed by fish were identified
from publisheddescriptions of fish feeding habits.
Invertebratesdocumented to be important in fish diets wererelated
to rank-order distributions of chemicalsensitivity to determine
whether a tested taxon (orclosely related surrogate) was likely to
be affected atthe observed exposure levels.
3. ANALYSIS
3.1. Stressor Characteristics
3.1.1. Physicochemical Properties andEnvironmental Fate
Chlorpyrifos is moderately lipophilic, whichlimits its water
solubility (1.4 mg L1) and explainsits tendency to partition (log
Kow of 4.75.3) intosoil, sediment, and organic material.(9) The
dissipa-tion of chlorpyrifos in water is relatively rapid,
withhalf-lives on the order of days, while dissipation ratesin soil
and sediment are only moderate (half-lives ofweeks). When used at
agricultural application rates,residues of chlorpyrifos do not
accumulate over timein soil/sediment systems. The principal
degradate ofchlorpyrifos is 3,5,6-trichloro-2-pyridinol (TCP).TCP
is considerably more polar than the parentchlorpyrifos molecule and
is therefore more hydro-philic.
3.1.2. Mechanism of Action
Toxicity of chlorpyrifos results from metabolicactivation to
form the chlorpyrifos oxon, whichinactivates the neurotransmitter
acteylcholinest-erase at neural junctions. Inactivation occurs
through reversible phosphorylation of the enzymeactive site.
Enzyme inactivation exerts toxicity byoverstimulation of the
peripheral nervous system.Acute toxicity of OP insecticides
co-occurring in thewater column appears to be adequately explained
bya simple additive model.(10) The TCP metabolitedoes not contain
the phosphorous-dialkyl ethermoiety and cannot phosphorylate the
target enzymeactive site. Consequently, TCP does not exhibit
OPinsecticide activity, and TCP was not considered achemical
stressor in this risk assessment. Thisjudgment is supported by the
toxicity data availablefor TCP, where the lowest reported LC50/EC50
valueis 1,800,000 ng L1.
3.2. Exposure Data
3.2.1. Insecticide Use
Previous monitoring(2,11,12) in the perennialreach of the San
Joaquin River identified three OPinsecticides as important
contributors of residues tosurface water, especially during the
dormant tree app-lication period (DecemberFebruary):
chlorpyrifos,diazinon
(O,O-diethyl-O-(2-isopropyl-4-methyl-6-pyrimidinyl)
phosphorothioate), and methidathion(O,O-dimethylphosphorodithioate,
S-ester
with4-(mercaptomethyl)-2-methoxyD2-1,3,4-thiadiazolin-5-one).
Monitoring conducted by Poletika et al.(4) fromMay 1, 1996 to April
30, 1997 generated daily data setsfor each of these compounds,
although a staggeredmethod validation process limited analysis for
resi-dues of diazinon and methidathion to days21364 and 56364,
respectively.
3.2.2. Daily Chemical Monitoring in Water:May 1996April 1997
ISCO Model 2700 autosamplers (ISCO, Inc.,Lincoln, NE) were used
to collect hourly samplescomposited over 24 hours. Data used in the
riskassessment came from a sampling location at RiverRoad, near the
confluence with the San JoaquinRiver. Methods of sample handling,
extraction,analysis, and quality control were
previouslyreported.(4) Method detection limits were 10, 10,and 24
ng L1, for chlorpyrifos, diazinon, andmethidathion, respectively.
Individual chemical datawere combined and plotted on the same graph
toshow the overall detection patterns for the threemonitored OP
insecticides (Fig. 2). The patterns arecharacterized by numerous,
moderate chlorpyrifos
294 Poletika, Woodburn, and Henry
-
peaks, fewer, larger diazinon peaks, and few, smallmethidathion
peaks.
3.2.3. Exposure in Sediment
No sediment data were collected during thisstudy period.
However, bed sediment samples col-lected near River Road in October
1992 andanalyzed by the U.S. GS contained
-
The 10th centile single-point estimate of spe-cies acute
sensitivity was selected as a benchmarkto evaluate exposure
profiles that could impact themost sensitive organisms in the tail
of the distribu-tion and direct further analysis of indirect
effectson taxa higher in the food chain. Previous prob-abilistic
aquatic ecological risk assessments haveused the 10th centile
benchmark in a similarmanner.(3,15)
Bluegill sunfish (Lepomis macrochirus) was themost sensitive
fish species, with a 48-h LC50 value of4800 ng L1. The 48-hour
period was viewed as atypical maximum exposure window, similar to
thepattern found in midwestern U.S. stream chemicalmonitoring
data.(3) We assumed initially that the48-hour exposure timeframe
would also be applic-able to Orestimba Creek and later confirmed
thevalidity of the assumption through temporal analysisof exposure
data.
The freshwater organism acute toxicity databaseused in this
site-specific risk assessment was subdi-vided in the following
manner. First, due to therelative insensitivity of fish, and to a
greater extent,plants, only aquatic invertebrates were
considered.Second, each taxon in the database was evaluatedfor
habitat requirements with respect to lentic andlotic environments.
Grouping was based on descrip-tions in the literature(1618) and
professional judg-ment. Tables I and II list the freshwater
invertebratespecies sensitivity distributions resulting from
thissubdivision. Note that where an animal inhabits bothstill and
moving water, the species appears in bothtables. Table II is
referenced as a listing of surrogatespecies for invertebrates
likely to be found inOrestimba Creek.
Fig. 3 presents the rank-ordered (i1 . . . in)cumulative
freshwater invertebrate species sensitiv-ity distributions for the
lentic and lotic groups,plotted using the formula i/n +1 100.(19)
Also givenin Fig. 3 are the best-fit linear regressions and
10thcentile sensitivity estimates calculated from theregression
equations. Compared to the all-species(vertebrates and
invertebrates) 10th centile sensitiv-ity of 102 ng L1, the
equivalent 10th centile valuewas lower for the lentic group of
invertebrates, 35 ngL1, and higher for the lotic invertebrate
group, 177ng L1. As shown in Table I, the sensitivity of thelentic
group was greatly influenced by the inclusionof mosquito larvae and
cladocerans, taxa unlikely tobe found in Orestimba Creek. The most
sensitivelotic species include amphipods, a mysid, midges,and
mayflies (Table II).
Giesy et al.(3) used reported sediment acutetoxicity data for a
midge (Chironomus tentans) andan amphipod (Hyalella azteca) and
equilibriumpartitioning methods to determine
whole-sedimentchlorpyrifos concentration ranges having potentialfor
adverse effects. The analysis suggests that effectsare not probable
at dry weight sediment concentra-tions 500 lg kg1.Applying the
local estimated Kd value of 379lg mL1 to these whole-sediment
concentrationranges for potential adverse effects, the
equivalentpore-water concentrations are 1; 320 ng L1,
respectively.
Fewer data are available to describe chlorpy-rifos chronic
toxicity endpoints. Chronic toxicity ininvertebrates has been
observed as reproductiveinhibition in two sensitive species, the
freshwaterdaphnid, Daphnia magna, and the saltwater
mysid,Mysidopsis bahia.(8) Life-cycle studies (exposureperiods of
21 to 35 d) report ranges of lowest-observed effect concentration
(LOEC) of 100 to300 ng L1 for D. magna and 4 to 10 ng L1 forM.
bahia. Reported LOECs for saltwater andfreshwater fish evaluated in
early life stage studies(flow-through exposures of 28 to 32 d)
range from480 ng L1 with the Atlantic silverside, Menidiamenidia,
to 3,200 ng L1 with the fatheadminnow, Pimephales promelas.(8)
These findings,combined with results from two full
life-cyclestudies of Pimephales promelas and other chronictoxicity
tests on various saltwater species, indicatethat growth generally
is the most sensitive meas-ure of chronic toxicity for vertebrates
exposed tochlorpyrifos.(8) However, survival may occasionallybe as
important or a more sensitive chronicendpoint.
3.3.2. Microcosm Studies
Microcosm and mesocosm studies providevaluable effects
information from controlled experi-mental ecosystems containing
several trophic levels.Both direct and indirect effects can be
observed, andif a suitable dosing regime is employed, generationof
chronic effects is also possible. Giesy et al.(3)
summarized several microcosm studies conducted inthe United
States and the Netherlands and conclud-ed that consistent findings
from these studiessupported a no-observable-adverse-effect
concen-tration (NOAEC) of 100 ng L1. Recovery ofaffected sensitive
invertebrate populations usually
296 Poletika, Woodburn, and Henry
-
Table I. Acute Toxicity Values Estimated for 48-Hour Exposure to
Chlorpyrifos for Invertebrates Requiring Lentic Habitat
Organism Taxonomic Name TaxonHabitat
Requirement Rank48-h Acute
Value (ng L1)
Mosquito various species I Lentic 1 0.7Mosquito Aedes aegypti I
Lentic (temporary pools
and ponds)2 4.8
Mosquito Culex pipiens I Lentic 3 28Cladoceran Ceriodaphnia
dubia C Lentic 4 113Cladoceran Ceriodaphnia sp. C Lentic 5 148Midge
Chironomus tentans I Lentic 6 157Mysid Neomysis mercedis I
Estuarine, brackish lakes,
freshwater lakes and streams(lotic and lentic)
7 198
Cladoceran Daphnia pulex C Lentic 8 210Mosquito Culex
quinquefasciatus I Lentic 9 212Cladoceran Daphnia sp. C Lentic 10
255Amphipod Gammarus lacustris C Lentic and lotic 11 289Midge
(various species) I Lentic, lotic, estuarine, some
marine12 297
Mosquito Culicoides pipiensquinquefasciatus
I Probably lentic 13 354
Cladoceran Daphnia longispina C Lentic 14 424Mosquito Aedes
vexans I Lentic 15 424Fairy Shrimp [Anostraca] C Lentic 16
438Amphipod Gammarus fasciatus C Lentic, lotic, estuarine 17
453Cladoceran Simocephalus vetulus C Lentic 18 707Mosquito
Culicoides variipennis I Lentic 19 771Pygmy Backswimmer Neoplea
striola I Lentic 20 849Caddisfly Leptoceridae sp. I Lotic and
lentic 21 900Mosquito Aedes cantans I Lentic (temporary pools
and ponds)22 900
Amphipod Hyalella sp. C Primarily lentic, occasionallylotic
23 919
Cladoceran (not specified) C Lentic 24 1,000Midge Tanypus
grodhaus I Lentic 25 1,061Diptera Paratanytarsus sp. I Lentic and
lotic 26 1,131Crawling water beetle Peltodytes sp. I Lentic 27
1,131Diving Beetle Laccophilus fasciatus I Lentic and lotic 28
1,485Diptera Chaoborus sp. I Primarily lentic, occasionally
lotic29 1,543
Cladoceran Daphnia magna C Lentic 30 1,700Pygmy Backswimmer Plea
sp. I Lentic 31 2,400Midge Chricotopus sp. I Unknown
distribution,
collected from S. Californiadrainage ditches
32 2,475
Tadpole shrimp Triops longicaudatus C Lentic 33 2,828Water
Boatman
(not Water Strider)Corixa punctata I Uncertain, generally lentic
34 2,828
Planaria Dugesia dorotocephala O Lentic and lotic 35
3,742Diptera Chaborus punctipennis I Lentic 36 3,818Dragonfly
Crocothemis erthryaea I Uncertain, collected from
Sudanese irrigation canals37 4,101
Mayfly Caenis horaria I Lotic and lentic 38 4,243Ostracod (not
specified) C Lentic and lotic 39 6,300Damselfly Enallagma/Ishnura
spp. I Primarily lentic, also lotic
depositional40 8,061
Crayfish Orconectes immunis C Lentic and lotic 41 8,485Diptera
Chaoborus obscuripes I Lentic 42 9,334Cladoceran Daphnia sp. C
Lentic 43 11,314Ostracod Cyprinotus incongruens O Lentic and lotic
44 14,142
Risk Assessment for Chlorpyrifos 297
-
occurred at concentrations < 500 ng L1. Becausemost
experimental ecosystems of this type tend to belentic systems, the
100 ng L1 NOAEC in partreflects the more sensitive species
associated withlentic habitats (Fig. 3, Table I) and is
thereforeprobably too conservative for moving water bodies.The
lotic species sensitivity distribution (Fig. 3,Table II) suggests
that a higher value such as the10th centile concentration of 177 ng
L1 is moreappropriate for characterizing acute effects in
Ores-timba Creek.
3.3.3. Effects Thresholds
3.3.3.1. Chlorpyrifos Alone. For temporal eventanalysis, the
acute effect threshold used was 177 ngL1, the 10th centile
sensitivity of the lotic speciesgroup. Applying a mean
acute-to-chronic ratio of 8(3)
to the acute value produced an estimated chroniceffect threshold
of 22 ng L1. Note that thesethresholds are most accurate for event
periods of 2 dfor acute events and 21 d or longer for
chronicevents. The acute duration is based on the normal-ization of
acute toxicity tests to a constant exposureperiod of 48 hours. A
21-d minimum for chroniceffects is associated with the exposure
periodrequired in laboratory toxicity studies to observetypical
chronic endpoints for both invertebrate andvertebrate aquatic
species.
3.3.3.2. Chlorpyrifos and Diazinon. The expo-sure data in Fig. 2
indicate that diazinon may havecontributed additional toxicity to
increase theseverity of adverse effects in Orestimba
Creek.Methidathions contribution appears to be negli-gible.
Accordingly, we utilized the diazinon com-bined acute toxicity
database(20) to form a subgroup-ing of invertebrates requiring a
lotic habitat,employing the same approach as that used
forchlorpyrifos. These diazinon data are presented inTable III. We
then estimated the 10th centile speciessensitivity value for
diazinon by linear regression onthe cumulative rank-order
distribution of acuteEC50/LC50 concentrations (r
2 0.878). The resulting10th centile value for lotic
invertebrates was 1,142ng L1. Diazinon acute values in the
distributionwere not normalized to 48 hours as was done
forchlorpyrifos. Instead, Giddings et al.(20) combined alldata for
exposure periods of 4896 hours.
Assuming that the additive effect modelexplains acute
toxicity,(10) we assessed exposure forthe combined residues of
chlorpyrifos and diazinonby summing the daily measured water
concentra-tions of chlorpyrifos and the toxic equivalent
con-centrations of diazinon to estimate a totalchlorpyrifos
equivalent concentration. Total chlor-pyrifos equivalent
concentrations were computed asfollows. First, each daily diazinon
concentration wasconverted to a chlorpyrifos equivalent by taking
the
Table I. Continued
Organism Taxonomic Name TaxonHabitat
Requirement Rank48-h Acute
Value (ng L1)
Backswimmer Notonecta undulata I Lentic and lotic 45
24,890Crayfish Procambarus clarki C Lentic and lotic 46
41,713Oligochaete Limnodrilus hoffmeisteri O Lentic and lotic 47
50,912Diving Beetle Hydrophylus spp. I Lentic and lotic 48
70,711Ostracod Chlamydotheca arcuata C Lentic 49 141,421Snail Anius
vortex O Lentic 50 210,190Snail Lymnaea stagnalis O Lentic 51
210,190Leech Nephelopsis obscura O Lentic 52 586,570Midge
Chironomus decorus I Primarily lentic, also lotic
depositional53 1,039,447
Snail Aplexa hypnorum O Lentic 54 1,139,856Snail Lanistes
carinatus O Unknown, probably lentic,
African genus55 1,916,259
Snail Bromphalaria alexandra O Unknown, probably lenticEgyptian
genus
56 2,070,391
Snail Helisoma trivolvis O Lentic and lotic 57 2,449,490Rotifer
Brachionus calyciflorus O Lentic 58 8,449,928
Note: Crustacea (C), Insect (I), or Other (O).
298 Poletika, Woodburn, and Henry
-
product of the diazinon concentration and the ratioof the
chlorpyrifos and diazinon 10th centile acutelotic species
sensitivities, 177/1,142. Second, each ofthe resulting daily
chlorpyrifos equivalent concen-trations was summed with the
corresponding repor-ted daily chlorpyrifos concentration.
Chronic toxicity from the presence of diazinonresidues does not
appear to be important, based on a
comparison of chronic endpoints derived in speciestested for
both chlorpyrifos and diazinon toxi-city.(8,20) Typically, the
no-observed-effect levels fordiazinon chronic effects were two to
three orders ofmagnitude higher than those determined for
chlor-pyrifos.
Table II. Acute Toxicity Values Estimated for 48-Hour Exposure
to Chlorpyrifos for Invertebrates Requiring Lotic Habitat
Organism Taxonomic Name TaxonHabitat
Requirement Rank48-h Acute
Value (ng L1)
Amphipod Gammarus pulex C Generally lotic,
occasionallylentic
1 99
Mysid Neomysis mercedis I Estuarine, brackish lakes,freshwater
lakes and streams(lotic and lentic)
2 198
Amphipod Gammarus pseudolimnaeus C Lotic 3 248Amphipod Gammarus
fasciatus C Lentic, lotic, estuarine 4 285Amphipod Gammarus
lacustris C Lentic and lotic 5 289Midge (various species) I Lentic,
lotic, estuarine, some
marine6 297
Mayfly Cloeon dipterum I Lotic 7 360Mayfly Ephemerella sp. I
Lotic 8 383Stonefly Pteronarcella badia I Lotic 9 537Caddisfly
Leptoceridae sp I Lotic and lentic 10 922Diptera Paratanytarsus sp.
I Lentic and lotic 11 1,131Diving Beetle Laccophilus fasciatus I
Lentic and lotic 12 1,485Stonefly Claassenia sabulosa I Lotic 13
2,162Midge Chricotopus sp. I Unknown distribution,
collected fromS. California drainage ditches
14 2,475
Planaria Dugesia dorotocephala O Lentic and lotic 15 3,742Isopod
Asellus aquaticus O Lotic 16 3,818Dragonfly Crocothemis erthryaea I
Uncertain, collected from
Sudanese irrigation canals17 4,101
Mayfly Caenis horaria I Lotic and lentic 18 4,243Midge
Dicrotendipes californicus I Lotic 19 4,950Ostracod (not specified)
C Lentic and lotic 20 6,300Crayfish Orconectes immunis C Lentic and
lotic 21 8,485Ostracod Cyprinotus incongruens O Lentic and lotic 22
14,142Blackfly Simulium vitattum I Lotic 23 19,092Mayfly
[Heptageniidae] I Primarily lotic, occasionally
lentic24 20,506
Caddisfly Hydropschy/Cheumatopsyche sp. I Lotic 25
21,637Stonefly Pteronarcys californica I Lotic 26 22,361Stonefly
Claassenin sp. I Lotic 27 24,495Backswimmer Notonecta undulata I
Lentic and lotic 28 24,890Isopod Proasellus coxalis O Probably
lotic 29 28,284Crayfish Procambarus clarki C Lentic and lotic 30
41,497Oligochaete Limnodrilus hoffmeisteri O Lentic and lotic 31
50,912Water Scavenging
BeetleHydrophylus spp. I Lentic and lotic 32 70,711
Snail Bithynia tentaculata O Primarily lotic, also lentic 33
210,190Snail Helisoma trivolvis O Lentic and lotic 34 2,449,490
Note: Crustacea (C), Insect (I), or Other (O).
Risk Assessment for Chlorpyrifos 299
-
4. RISK CHARACTERIZATION
4.1. Temporal Analysis of Chlorpyrifos WaterExposure
Patterns
Fig. 4 presents the RADAR acute event analy-sis for chlorpyrifos
alone and chlorpyrifos + diazinonfor events at or above the
chlorpyrifos lotic thresh-old value of 177 ng L1. There were eight
acutechlorpyrifos events with an average duration ofthree days, and
the percent of time above thethreshold was 7%. Addition of the
diazinon chlor-pyrifos-equivalent concentrations to the data
analy-sis increased the number of acute events to 11, with
an average duration of two days. The percent of timeabove the
threshold increased slightly to 9%. Theseacute exposure patterns
were characterized by along recovery duration of 175 days,
occurring in theperiod between the end of summer applications
andbefore initiation of winter treatments.
Chlorpyrifos exposure events at or above thechronic effect
threshold of 22 ng L1 are depicted inFig. 5. A total of 25 exposure
periods were found toexceed the chronic effect level, averaging
five daysin duration (range of one to 32 days). Only one ofthese
episodes, the one beginning on Day 315 (3/10/97), persisted beyond
the 21-day exposure periodnecessary to elicit chronic effects in
laboratory
Fig. 3. Distribution of 48-hour normalizedchlorpyrifos
EC50/LC50s for all freshwaterinvertebrate species tested. Data
pointswith error bars are geometric means andgeometric standard
errors. The linearregression line is plotted with the 95thcentile
prediction interval. Table inset: 10thcentile species sensitivity
concentrationwith upper and lower concentration boundsof regression
probability point estimate.NS Normal Score. Upper: species
asso-ciated with lentic habitat. Lower: speciesassociated with
lotic habitat.
300 Poletika, Woodburn, and Henry
-
studies with freshwater invertebrates. The amount oftime the
exposure concentrations were above thelotic chronic threshold value
was 36%. The long fall/winter recovery period observed in the acute
eventanalysis also appeared in the pattern of chronicevents (155
d). As Fig. 5 shows, interpretation of thechronic exposure pattern
is complicated by thepresence of acute events for sensitive
invertebrates.
4.2. Probability of Exposures Exceeding EffectsThresholds and
Taxa at Risk
4.2.1. Acute Effects
The 11 chlorpyrifos + diazinon acute exposureevents (Fig. 4) can
be rank-ordered by arithmeticmean concentration and plotted as a
cumulativeprobability distribution (Fig. 6). Using linear
regres-sion, estimates of the typical (50th centile) andtypical
worst-case (90th centile) acute event weregenerated: 356 and 766 ng
L1, respectively. Com-parison of these values to the lotic species
chlorpy-rifos sensitivity ranking in Table II suggests thatsome
species of amphipods, mysids, and midges maybe impacted by the
typical event, and some mayfliesand stoneflies could be affected by
the typical worst-case event. Regression on a similar distribution
forchlorpyrifos-alone acute-event concentrations (r2 0.912, data
not shown) produced 50th and 89thcentiles of 330 and 624 ng L1,
respectively (N 8,so the 90th centile cannot be computed). The
significance of this analysis is that diazinon contri-buted to
the upper end of the event distributionmore than chlorpyrifos
did.
The 50th and 89th centile chlorpyrifos-aloneexposure events can
be related to reference water-column concentrations predicting risk
in bed sedi-ment habitat. (No sediment acute toxicity data
orsite-specific Kd information are available for diazi-non, so the
analysis is restricted to chlorpyrifosonly.) In this approach, we
utilized the single-valueestimate of Kd to predict bioavailable,
toxic concen-trations in bed sediment pore water,
assumingequilibrium conditions exist in each time period ofinterest
in the water column (water + suspendedsediment) and in the sediment
bed (pore water +sediment). Thus, dissolved concentrations in
bothcompartments are assumed to be equal or nearly so.The 50th
centile event (330 ng L1 ) falls at thelower end of the range in
possible effect concentra-tions in bed sediment pore water (264 to
1,320 ngL1 ), while the 89th centile event (624 ng L1 ) fallsin the
middle of the range. Effects on sediment-exposed life stages are
slightly possible to possiblebut not probable.
Returning to the situation in the water column,data from Table
III indicate that there are nosensitive taxa in the group of lotic
invertebratestested for diazinon acute effects. However, amphi-pods
are sensitive to relatively low concentrations ofchlorpyrifos
(Table II). Stoneflies appear to be
Table III. Acute Toxicity Values for Exposure to Diazinon for
Invertebrates Requiring Lotic Habitat
Organism Taxonomic Name Taxon Habitat Requirement RankAcute
Value (ng L1 )
Amphipod Gammarus pseudolimnaeus C Lotic 1 2,000Mysid Neomysis
mercedis I Estuarine, brackish lakes, freshwater
lakes and streams (lotic and lentic)2 4,150
Mayfly Cloeon dipterum I Lotic 3 7,800Crayfish Orconectes
propinquus C Lentic and lotic, primarily Laurentian
Great Lakes and drainages4 15,000
Stonefly Acroneuria ruralis I Lotic and lentic 5 16,000Amphipod
Asellus communis C Probably lotic 6 21,000Amphipod Hyallela azteca
C Primarily lentic, occasionally lotic 7 22,000Mayfly Baetis
intermedius I Lotic and lentic 8 24,000Stonefly Pteronarcys
californica I Lotic 9 25,000Mayfly Paraleptophlebia pallipes I
Lotic 10 44,000Snail Helisoma trivolvis O Lentic and lotic 11
528,000Asian leech Hirudo nipponia O Uncertain, probably lakes
and/or slow
streams12 1,900,000
Cyclops Cyclops sp. O Primarily lentic, occasionally lotic 13
2,510,000Oligochaete Tubifex sp. O Lentic and lotic 14
3,160,000
Note: Crustacea (C), Insect (I), or Other (O).
Risk Assessment for Chlorpyrifos 301
-
tolerant to diazinon, and most tested stonefly generaare also
tolerant to chlorpyrifos. Stoneflies, there-fore, can be removed
from the group that could beaffected by the typical worst-case
event dominatedby diazinon residues. Equal numbers of mayflygenera
were sensitive and tolerant to chlorpyrifos,while all tested genera
were tolerant to diazinon.More of the chlorpyrifos-tested midge
genera weretolerant of the observed acute events than
weresensitive, but there is no information on midgesensitivity to
diazinon. Only the tested amphipodgenera were uniformly sensitive
to the observedtypical and typical worst-case acute events.
Tosummarize: (1) amphipods and mysids appear to
be at risk, primarily from the presence of chlorpy-rifos, (2)
mayflies and stoneflies are less likely to beaffected by either
compound due to a range ofgenera sensitivities, and (3) for the
same reason,midges are also less likely to suffer acute effectsfrom
chlorpyrifos alone (no data for diazinon).
Table IV lists the spawning and dietary habitatsof the fish
known to dominate the main stem regionof the San Joaquin River
watershed. When thisinformation is related to the groups of
potential preyspecies at risk from typical and typical
worst-caseacute effects, it is possible to determine
whetherindirect effects on fish through dietary impacts arelikely
to have occurred.
Fig. 4. Temporal analysis of acute toxicevents occurring above a
threshold of177 ng L1 (horizontal line). Upper:chlorpyrifos alone.
Lower: chlorpyrifosequivalent concentrations ([chlorpyrifos]+
177/1,142[diazinon]).
302 Poletika, Woodburn, and Henry
-
During the spring to fall period, flow of waterinto the lower
reach of Orestimba Creek came fromirrigation tailwater and
spillwater. Flow volumesvaried daily,(4) and only during winter
flooding in themain stem river was the water flow
consistentlyslow-moving in the lower reach of the creek.
Habitatrequirements for spawning and for juveniles suggestthat few
immature individuals would inhabit thecreek (Table IV). Adult
fathead minnow, red shiner,and inland silverside are more dependent
on aqua-tic insects and other small macroinvertebrates,and
therefore would be more likely to enter thelower reach of Orestimba
Creek from downstream
breeding areas in search of food. Adult threadfinshad would be
less attracted to the creek, where onlysmall populations of
microscopic food sources wouldbe available for filter feeding.
In summary, the potential impacts on amphi-pods and mysids
during acute chlorpyrifos + diazi-non events would appear to be
relativelyunimportant to the feeding requirements of adultfish
(Table IV), the life stage most likely to inhabitthe lower reach of
the creek. However, if there werealso reductions in populations of
mayflies, stoneflies,and midges, due to the presence of sensitive
species,indirect impacts on fish could have occurred.
Fig. 5. Temporal analysis of chronic toxicevents occurring above
a threshold of22 ng L1 (lower horizontal line; upperhorizontal line
represents the acutethreshold of 177 ng L1).
Fig. 6. Chlorpyrifos + diazinon acute toxicevents from the lower
panel of Fig. 4plotted as a cumulative probability distri-bution of
event arithmetic average con-centrations. NS Normal Score.
Risk Assessment for Chlorpyrifos 303
-
4.2.2. Chronic Effects
The arithmetic average concentration of the32-day chlorpyrifos
chronic event was 155 ng L1,considerably lower than the reported
freshwater fish(fathead minnow) LOEC of 3,200 ng L1 and nearthe
midpoint of the 100300 ng L1 range ofDaphnia magna LOECs. Assuming
that lenticcladocerans such as Daphnia magna, which areacutely
sensitive to chlorpyrifos, possess more sen-sitive chronic
endpoints than do animals that areless acutely sensitive, then
chronic effects on loticinvertebrates are unexpected from this
32-dayevent.
5. ECOLOGICAL RELEVANCE
5.1. Population and Community Effects
Interpretation of available data discussedabove suggests that
populations of sensitive inver-tebrates could be impacted, if
present. Affectedpopulations, however, do not appear to be
critic-ally important components of fish diets. Alternat-
ive food sources such as algae, other plantmaterial, other
insects and small invertebrates,insensitive crustaceans, and
mollusks would beunaffected by exposure to chlorpyrifos +
diazinon.Food of terrestrial origin, including live inverte-brates
and invertebrate carcasses, commonlymakes up a significant
percentage of energy flowin low-order streams with adequate
riparian hab-itat, so this would also mitigate against an
adversechemical effect on fish diets (good riparian habitatis
reported for Orestimba Creek(5)). Recovery oftemporarily affected
invertebrate populationswould take place via reproduction by
unaffectedindividuals and those protected by refugia, repop-ulation
from downstream drift, or immigration byterrestrial life stages. If
recovery did not occur,functional replacement by more tolerant
inverteb-rate species would likely fill the vacated
ecologicalniches.(21,22) In view of all these considerations,
weconclude that no significant alterations in ecosys-tem function
are expected from the occurrence ofthese chemical stressors in the
temporal patternobserved during the year-long chemical
monitor-ing.
Table IV. Habitat, Spawning Periods, and Diet and for Fish
Inhabiting the Main Stem Region of the San Joaquin Watershed
Fish Spawning Habitat Spawning Period Foraging Habitat Diet
References
Fathead minnow Shallow waterwith vegetation
As early as MarchLate spring tomid-summer
Juveniles:shallow weedy areas
Young: filamentousalgae, diatoms,detritus, smallinvertebrates
Algae,other plant material,aquatic insects
38, 39, 40, 41, 42
Red shiner Aquatic vegetation,gravel, sand, mud,mostly in calm
water
More than onceper season Late springto early fall,mid-summer
peak
Shallow waters Juveniles: smallcrustaceans, aquaticinsects,
larvae, algae,plant leaves, detritusInsects, other
smallinvertebrates
43, 44, 45
Threadfin shad Submergedvegetation inshallow sluggishwater
Spring, when waterwarms to 21 C, maycontinue atintervals
throughoutthe warmer monthsof the year
Shallow and openwater
Filter feeder(young and adults):microscopic plants andanimals
suspended inwater column
44, 46, 47, 48
Inland silverside Shallow weedywaters
April throughSeptember, withpeaks in May andAugust
Shallow sluggishweedy areas;ditches, reservoirs,irrigation
systems;avoid fast currents
Juveniles:zooplankton, includinglarge cladocerans,larvae of
chironomidand phantom midgesSmall crustaceans,mollusks, insects
38, 49
304 Poletika, Woodburn, and Henry
-
5.2. Ecological Significance of AgriculturallyDominated
Streams
Presence of the San Joaquin Main Stem groupof fish species
indicates the existence of alteredenvironmental conditions, many of
which resultdirectly from human activities in the lower Ores-timba
Creek watershed. In addition to the occur-rence of pesticide
residues in surface water,additional habitat alterations resulting
from agri-culture include increased salinity and nutrients,eroded
soil contributions to suspended and bedsediment, channel dredging,
flow inconsistency,and, possibly, presence of animal waste.(5)
Thebasic stream type has changed from intermittent toperennial.
Ecological effects from the presence ofOP insecticides over time
must be interpreted inthe context of the type of habitat and
aquaticcommunities present in a creek dominated byhuman activities.
Cumulative ecological risk assess-ment, a process that considers
aggregate riskcaused by multiple stressors,(23) could be useful
togain better understanding of this situation andpoint to actions
necessary to implement any desiredhabitat restoration.
6. UNCERTAINITY AND VARIABILITYANALYSIS
6.1. Exposure Characterization
The exposure data are limited to a one-yearsampling period.
Farming practices were typical forthe year, with the exception of
fewer winter insec-ticide applications to dormant trees.
Increaseddormant applications could result in shortening ofthe long
recovery period observed in the monitoringdata set.
Cholinesterase-inhibiting insecticides otherthan chorpyrifos,
diazinon, and methidathion werenot monitored. Historically, use of
other products ofthis type has been minimal.(24)
Sources of variation in the exposure datainclude sampling error
and analytical error. Theuse of autosamplers, calibrated daily, to
collecthourly water samples composited each 24 hoursfor analysis
provided samples representative of thetemporal changes in
concentration from day today. Analysis of additional samples
collected usingthe U.S. GS Equal Width Increment protocol(25)
indicated that the autosampler single-point intakesalso were
representative of the spatial concentra-tion profile across the
stream channel in this
well-mixed stream.(26) Sampling error appears tobe minimal.
Variability in the analytical data set was quan-tified by
quality control samples. Samples fortified ata level of 25 ng L1
yielded a mean standarddeviation of 23 2.9 ng L1 for chlorpyrifos
(N 218) and 24 1.9 ng L1 for diazinon (N 188).(26)Compared to the
variation in effects data discussedin the following section,
analytical variability wasnegligible in the key threshold
concentrations usedin the RADAR analysis (22177 ng L1).
Conse-quently, only the reported point estimates of con-centration
were utilized in the assessment.
The sediment concentration estimates should beconsidered more a
screening level exposure assess-ment than a quantitative analysis
due to uncertaintyin the cross-compartment equilibrium
assumption.Paired sampling of the water column and bedsediment,
along with abundance measurements ofbenthos, is necessary to test
the assumption andgeneral approach.
6.2. Effects Data
The use of the constructed surrogate assem-blage for the
invertebrate community inhabitingOrestimba Creek generates
uncertainty related tothe presence or absence of entire groups
(example:amphipods and caddisflies). Another uncertaintyarises from
the relative sensitivities of genera withina group. A good example
is mayfly sensitivity tochlorpyrifos: two genera were tolerant and
two weresusceptible (Table II). It is not known whethermayflies
inhabit Orestimba Creek, or which generaor species are
represented.
There are unequal numbers of species tested forchlorpyrifos and
diazinon toxicity, and the speciestested are not the same. This may
bias the additivityanalysis, because 10th centile points on the
acutespecies sensitivity distributions were selected toestimate
chlorpyrifos equivalent concentrations.
Sediment toxicity data is quite limited relativeto laboratory
results reported from simple watersystems. Interpretation of
sediment toxicity valuesgenerally relies on the equilibrium
partition method,which appears to be valid for
chlorpyrifos.(27)
Application of the additive toxicity model is notpossible in the
absence of diazinon sediment toxicitytest data. In 1992, the site
was also characterized byrelatively large concentrations of total
DDT resi-dues in all of the sampled matrices (range of 24 ngL1 to
4,350 ng g1).(13) The contribution of legacy
Risk Assessment for Chlorpyrifos 305
-
DDT toxicity to benthic macroinvertebrates needsto be factored
into the assessment.
Variability in the effects database reflectsinherent biological
variation in the responses oftest organisms and interlaboratory
variability rela-ted to conditions of cultured organisms,
methods,and, possibly, poorly controlled parameters duringthe
testing period. Fig. 3 shows the combinedeffects of all these
contributing sources of vari-ation and quantifies the 95th centile
predictionerror for a point estimate of probability in
thedistribution of acute species sensitivity (4th to 19thcentile
for the 10th centile point estimate). Theprediction interval
accounts for the error of theregression coefficient, the error of
the mean, andthe variation of individual data points around
theestimated mean.
6.3. Risk Characterization
Comparison of the complex temporal pattern ofobserved exposures
in Orestimba Creek to thesimple, short-duration laboratory toxicity
tests gen-erates uncertainty in interpreting overlaps betweenthe
exposure and effects distributions. A betterunderstanding of the
dynamics of exposure, uptake,elimination, and acetylcholinesterase
regenerationwould allow application of modeling to improve therisk
characterization.
Repeated exposures of fish and invertebrateswere assumed to be
independent, with full recoveryfrom cholinesterase inhibition
occurring during eachrecovery period. The actual kinetics of
depurationand enzyme regeneration may be more complex,thus
violating this assumption. Depuration rates forchlorpyrifos and
diazinon are quite rapid, and fishmay tolerate substantial levels
of cholinesterasedepression without experiencing toxicity.(8)
Recov-ery of invertebrate organisms from repeated chlor-pyrifos
exposures has been demonstrated by Naddyet al.(28) In pulsed
exposure work with Daphniamagna, a sensitive freshwater lentic
invertebrate,these authors found recovery following pulsedexposures
to chlorpyrifos, if the critical body burdenfor chlorpyrifos with
the daphnid was not achieved;the required timeframe between acute
exposures forrecovery appeared to be three days or more.
The 10th centile single-point estimate of speciesacute
sensitivity benchmark assumes the LC50/EC50values are realistic
predictors of population-leveleffects in the stream environment,
and those specieslocated outside the sensitive tail of the
distribution
would not be at risk for sufficient mortality to affectthe
assessment endpoints. Research indicates thatthis is a conservative
assumption for some sensitiveaquatic species, as studies on the
effects of insecti-cides on zooplankton and miticides on
terrestrialarthropods have shown that laboratory LC50 valuesmay
significantly overestimate field effects at thepopulation
level.(2931) If one therefore interprets theerror in the point
estimate of the 10th centile as arange of possible concentrations
impacting thesensitive grouping of species in the distribution
tail(error bounded by prediction interval, Fig. 3), thenRADAR
analysis can be performed to determinethe sensitivity of the
exposure characterization tothis error.
RADAR runs for chlorpyrifos-equivalent con-centrations using the
thresholds 63, 177, and 426ng L1 (Fig. 3) produced the following
patterns ofexposure (reported by increasing concentration
ofexposure threshold): 18, 11, and 7 events; 18, 9,and 4% of time
above threshold; and 3, 2, and 1days average duration. The sets of
event concen-trations, when ranked and considered as cumula-tive
distributions of acute exposure events,generate 90th centile
typical worst-case concentra-tions of 416, 766, and 964 ng L1,
respectively, foreach threshold value (the last number is
extrapo-lated on the regression line outside the datarange). Only
one additional taxon, ranked 10 inTable II, is added to the list of
organismspotentially impacted by the 964 ng L1 typicalworst-case
exposure event. We conclude that thepoint estimate of the 10th
centile is adequate todescribe the sensitive species grouping for
riskcharacterization. Moreover, the prediction intervalaround the
10th centile regression estimate in-cludes both the 4th and 19th
centiles; this bracketsthe 5th centile level of species sensitivity
com-monly used in regulatory schemes as a brightlinefor ecosystem
protection.(32,33)
Seasonality of acute-event occurrence was notrelated to
invertebrate reproduction strategy. Moresite-specific information
on invertebrate communitycomposition and knowledge of life
histories wouldimprove understanding of the significance of
expo-sure and recovery patterns.
Additive exposures based on a common point inthe species
sensitivity distribution can predict gen-eral levels of effect but
do not well characterizetoxicity to specific taxa, unless each
component inthe mixture is similar in its activity. A more
reliablepredictor for specific taxa would employ a ratio of
306 Poletika, Woodburn, and Henry
-
toxicity values derived from tests on the sameorganism.
6.4. Ecological Relevance
The principal uncertainty associated with theanalysis of
ecological relevance of effects is the lackof knowledge of the
specific ecological entitiesrequiring protection in a system
dominated byagricultural activities. In particular, more
informa-tion is necessary to determine whether
sensitiveinvertebrate populations important in fish diets
arepresent.
7. CONCLUSION
Analysis of a detailed chemical exposure dataset in combination
with a relatively robust toxicitydatabase of surrogate species for
chlorpyrifos anddiazinon suggests that no direct adverse effects
onfish should have occurred during the year of mon-itoring. Certain
genera of sensitive aquatic inverte-brates, if present, may have
experienced populationreductions during acute-exposure events and
subse-quent functional replacement by more tolerantspecies if
recovery did not take place. These reduc-tions/replacements would
not have had appreciableeffect on the diets of common fish species
currentlyinhabiting the main stem San Joaquin River and
itsagriculturally dominated tributaries. Neither of theassessment
endpoints, fish population persistence orinvertebrate community
productivity, appeared tobe adversely affected by the presence of
thesechemical residues.
8. RESEARCH RECOMMENDATIONS
This risk assessment can be refined with addi-tional information
on the site-specific nature of theaquatic community. Surveys of
benthic macroinver-tebrates and fish in the lower reach of
OrestimbaCreek and in an appropriate reference site canincrease
understanding of the types of organismsexpected in this type of
system in the absence of OPinsecticides. However, reference sites
may be diffi-cult to identify, given the widespread use of
chem-ical pest control in commercial agriculture. Surveysof
physical habitat would also contribute to under-standing the
relative roles different types of stressorsplay in determining
species composition. Once thisbiological information is obtained,
toxicity testing of
local species could provide additional data forinterpretation of
site-specific effects.
ACKNOWLEDGMENTS
The authors thank the three anonymous re-viewers for the many
constructive suggestions forimproving the manuscript, particularly
in the area ofuncertainty and variability analysis.
REFERENCES
1. U.S. EPA, Reregistration Eligibility Science Chapter
forChlorpyrifos, Fate and Environmental Risk AssessmentChapter,
June 2000 Revision, U.S. Environmental ProtectionAgency, Office of
Pesticide Programs, Washington, DC(2000).
2. C. Foe, Insecticide Concentrations and Invertebrate
BioassayMortality in Agricultural Return Water from the San
JoaquinBasin, Staff Report to the Central Valley Regional
WaterQuality Control Board, Sacramento, CA (1995).
3. J. P. Giesy, K. R. Solomon, J. R. Coats, K. Dixon, J.
Gid-dings, E. E. Kenaga, Ecological Risk Assessment ofChlorpyrifos
in North American Aquatic Environments,Rev. Environ. Contam.
Toxicol. 160, 1129 (1999).
4. N. N. Poletika, P. L. Havens, C. K. Robb, R. D.
Smith,Organophoshorous Insecticide Concentration Patterns in
anAgriculturally Dominated Tributary of the San JoaquinRiver, Ch.
19 in Agrochemical Fate and Transport: Per-spectives and Scale of
Study, ed. by T. Steinheimer, L. Ross,T. Spittler, Amer. Chem. Soc.
Symposium Series No. 751(1999).
5. N. M. Dubrovsky, C. R. Kratzer, L. R. Brown, J. M. Gron-berg,
K. R. Burow, Water Quality in the San Joaquin-TulareBasins,
California, 19921995, U.S. Geological SurveyCircular 1159, Denver
CO (1998).
6. J. M. Harrington, P. Ode, A. Montalvo, D. Post, C. Sheehy,M.
Dawson, Russian River Index of Biological Integrity(RRIBI) for
First to Third Order Tributary Streams, WaterQuality Inventory
Series: Biological and Physical/HabitatAssessment of California
Water Bodies, California Depart-ment of Fish and Game, Rancho
Cordova, CA (1999).
7. U.S. EPA, Framework for Ecological Risk
Assess-ment,EPA/630/R-92/001, Risk Assessment Forum,
U.S.Environmental Protection Agency, Washington, DC (1992).
8. U.S. EPA, Proposed Guidelines for Ecological RiskAssessment,
Federal Register 61, 4755247631, U.S.Environmental Protection
Agency, Washington DC (1996).
9. U.S. EPA, Guidelines for Ecological Risk
Assess-ment,EPA/630/R-95/002F, Risk Assessment Forum,
U.S.Environmental Protection Agency, Washington, DC(1998).
10. L. R. Brown, Aquatic Biology of the San
Joaquin-TulareBasins, California: Analysis of Available Data
Through1992, U.S. Geological Survey Water-Supply Paper 2471,Denver,
CO (1996).
11. L. R. Brown, J. T. May, Macroinvertebrate Assemblages
onWoody Debris and their Relations with EnvironmentalVariables in
the Lower Sacramento and San Joaquin RiverDrainages, California,
Environ. Monitor. Assess. 64,322329 (2000).
12. M. G. Barron, K. B. Woodburn, Ecotoxicology of
Chlor-pyrifos, Rev. Environ. Contam. Toxicol. 144, 193 (1995).
Risk Assessment for Chlorpyrifos 307
-
13. K. D. Racke, Environmental Fate of Chlorpyrifos,
Rev.Environ. Contam. Toxicol. 131, 1154 (1993).
14. H. C. Bailey, J. L. Miller, M. J. Miller, L. C. Wiborg,
L.Deanovic, T. Shed, Joint Acute Toxicity of Diazinon
andChlorpyrifos to Ceriodaphnia dubia, Environ. Toxicol.Chem. 16,
23042308 (1997).
15. L. J. Ross, R. Stein, J. Hsu, J. White, K. Hefner,
Distribu-tion and Mass Loading of Insecticides in the SanJoaquin
River, California. Winter 199192 and 199293,Report EH 96-02,
California Environmental ProtectionAgency, Department of Pesticide
Regulation, Sacramento,CA (1996).
16. J. L. Domalgalski, N. M. Dubrovsky, C. R. Kratzer,
Pesti-cides in the San Joaquin River, California: Inputs
fromDormant Sprayed Orchards, J. Environ. Qual. 26,
454465(1997).
17. W. E. Pereira, J. L. Domagalski, F. D. Hostettler, L.
R.Brown, J. B. Rapp, Occurrence and Accumulation of Pes-ticides and
Organic Contaminants in River Sediment, Waterand Clam Tissues from
the San Joaquin River and Tribu-taries, California, Environ.
Toxicol. Chem. 15, 172180(1996).
18. P. L. Havens, N. N. Poletika, C. K. Robb, Evaluation ofWater
Quality Best Management Practices to Reduce RunoffTransport of
Chlorpyrifos, Internal Report of Dow Agro-Sciences, GH-C 4940
(1999).
19. K. R. Solomon, D. B. Baker, P. Richards, K. R. Dixon, S.
J.Klaine, T. W. LaPoint, R. J. Kendall, J. M. Giddings, J. P.Giesy,
L. W. Hall, Jr., C. P. Weisskopf, M. Williams, Eco-logical Risk
Assessment of Atrazine in North AmericanSurface Waters, Environ.
Toxicol. Chem. 15, 3176 (1996).
20. R. W. Merritt, K. W. Cummins, An Introduction to theAquatic
Insects of North America, third edition, Kendall/Hunt Publishing
Company, Dubuque, IA (1996).
21. R. W. Pennak, Fresh-Water Invertebrates of the United
States,third edition, John Wiley & Sons, New York (1989).
22. J. H. Thorp, A. P. Covich, Ecology and Classification
ofNorth American Freshwater Invertebrates, Academic Press,Inc., San
Diego, CA (1991).
23. B. R. Parkhurst, W. Warren-Hicks, T. Etchison, J. B.
But-cher, R. D. Cardwell, J. Volison, Methodology for
AquaticEcological Risk Assessment, RP91-AER, Final Report Pre-pared
for the Water Environment Research Foundation,Alexandria, VA
(1995).
24. J. M. Giddings, L. W. Hall, Jr., K. R. Solomon,
EcologicalRisks of Diazinon from Agricultural Use in the
Sacramento-San Joaquin River Basins, California, Risk Anal.
20,545570 (2000).
25. P. D. Kingsbury, D. P. Kreutzweiser, Permethrin Treat-ments
in Canadian Forests. Part I: Impact on Stream Fish,Pestic. Sci. 19,
3548 (1987).
26. D. P. Kreutzweiser, P. D. Kingsbury, Permethrin Treatmentsin
Canadian Forests. Part II: Impact on Stream Invertebrates,Pestic.
Sci. 19, 4960 (1987).
27. U.S. EPA, Waquoit Bay Watershed Ecological Risk Assess-ment
Planning and Problem Formulation, Draft, June 13,1996, Risk
Assessment Forum, EPA/630/R-96/004A, NSCEP,Cincinnati, OH
(1996).
28. Stanislaus County Agricultural Commissioner and Sealer
ofWeights and Measures, Pesticide Use Data Reports for 19921997.
Modesto, CA.
29. Federal Inter-Agency Sedimentation Project, A Study
ofMethods Used in Measurement and Analysis of SedimentLoads in
Streams, Rev. November 1972, St. Anthony FallsHydraulic Laboratory,
Minneapolis, MN (1972).
30. N. N. Poletika, C. K. Robb, A Monitoring Study to
Charac-terize Chlorpyrifos Concentration Patterns and
EcologicalRisk in an Agriculturally Dominated Tributary of the
SanJoaquin River, Internal Report of Dow AgroSciences, GH-C4854
(1998).
31. G. T. Ankley, D. J. Call, J. S. Cox, M. D. Kahl, R. A. Hoke,
P.A. Kosian, Organic Carbon Partitioning as a Basis forPredicting
the Toxicity of Chlorpyrifos in Sediments, Envi-ron. Toxicol. Chem.
13:621626 (1994).
32. K. Day, N. K. Kaushik, An Assessment of the ChronicToxicity
of the Synthetic Pyrethroid, Fenvalerate, to Daphniagaleata
mendotae, Using Life Tables, Environ. Pollut. 44,1326 (1987).
33. R. E. Daniels, J. D. Allan, Life Table Evaluation of
ChronicExposure to a Pesticide, Can. J. Fish. Aquat. Sci. 38,
485494(1981).
34. W. K. Walthall, J. D. Stark, A Comparison of AcuteMortality
and Population Growth Rate as Endpoints ofToxicological Effect,
Ecotoxicol. Environ. Saf. 37, 4557(1997).
35. C. E. Stephan, D. I. Mount, D. J. Hansen, J. H. Gentile, G.
A.Chapman, W. A. Brungs, Guidelines for Deriving NumericalNational
Water Quality Criteria for the Protection of AquaticOrganisms and
Their Uses, NITS PB85-227049, USEPAResearch Laboratory, Duluth, MN,
Narragansett, RI, andCorvallis, OR (1985).
36. Health Council of the Netherlands, Ecotoxicological
RiskAssessment and Policy-Making in the Netherlands,
Network6(3)/7(1), 811 (1993).
37. R. B. Naddy, K. A. Johnson, S. J. Klaine, Response ofDaphnia
magna to Pulsed Exposures of Chlorpyrifos, En-viron. Toxicol. Chem.
19, 423431 (2000).
38. P. B. Moyle, Inland Fishes of California, University of
Cali-fornia Press, Berkeley, CA (1976).
39. V. C. Wynne-Edwards, The Breeding Habits of the Black-Headed
Minnow (Pimephales promelas Raf.), Trans. Am.Fish. Soc. 62, 382383
(1933).
40. H. C. Markus, Life History of the Black-Headed
Minnow(Pimephales promelas), Copeia 3, 116122 (1934).
41. W. C. Starrett, Food Relationships of the Minnows of theDes
Moines River, Iowa, Ecology 31, 216233 (1950).
42. E. E. Coyle, The Algal Food of Pimephales promelas(fathead
minnow), Ohio J. Sci. 30, 2335 (1930).
43. R. H. Hardwood, Diurnal Feeding Rhythm of Notropislutrensis
Baird and Giard, Texas J. Sci. 24, 9799 (1972).
44. J. C. S. Wang, Fishes of the Sacramento-San Joaquin
Estuaryand Adjacent Waters, California: A Guide to the Early
LifeHistories, Technical Report 9, Interagency Ecological
StudiesProgram for the San Joaquin Estuary (1986).
45. W. L. Pflieger, The Fishes of Missouri, Missouri
Departmentof Conservation (1975).
46. C. Taber, The Distribution and Identification of
LarvalFishes in the Buncombe Creek Arm of Lake Taxoma
withObservations on Spawning Habits and RelativeAbundance, Ph.D.
Dissertation, University of Oklahoma(1969).
47. R. V. Miller, Food of the Threadfin Shad, Dorosomapetenense,
in Lake Chicot, Arkansas, Trans. Am. Fish. Soc.96, 243246
(1967).
48. H. E. Bryant, A. Hauser, Growth of Threadfin Shad in
BullShoals Reservoir, Proc. 22nd Ann. Conf. S.E. Assoc. Gameand
Fish Comm., 275283 (1968).
49. C. M. Breder, Field Book of Marine Fishes of the
AtlanticCoast from Labrador to Texas, G.P. Putnams Sons, NewYork
(1948).
308 Poletika, Woodburn, and Henry