Potential Toxicity of Pesticides in Midwestern
Small fractions, estimated at <1 to 2% of the pesticides applied to
Midwestern cropland are lost from fields and enter nearby streams
during rainfall events. In many cases aquatic organisms are exposed
to mixtures of chemicals, which may lead to greater non-target risk
than that predicted based on traditional risk assessments for single
chemicals. Relatively little research has been directed at determining
the risk of environmental mixtures of pesticides to non-target aquatic
One approach involves computing a toxicity index using actual measurements
of pesticide residues in stream water and published estimates of acute
toxicity. We evaluated the potential toxicity of environmental mixtures
of 5 classes of pesticides using concentrations from water samples
collected from sites (fig 1) on 76 Midwestern streams during late
spring or early summer runoff events in 1998.
The 5 classes of pesticides examined are: (1) ALS-inhibiting herbicides
(including sulfonylureas, sulfonamides, and imidazolinones), (2) triazine
herbicides, (3) chloroacetamide herbicides, (4) organophosphate insecticides,
and (5) carbamate pesticides. Acute toxicity data are primarily from
the EPAS ECOTOX database, but other sources were also used.
Published EC50 values for two aquatic plants (duckweed, Lemna gibba;
and green algae, Selenastrum capricornutum), and LC50 values
for two aquatic vertebrates (bluegill sunfish, Lepomis macrochirus;
and chorus frogs, Pseudacris triseriata) are used as toxicity
metrics in this study (fig 2). Potential chronic effects of low-level
pesticide exposures or the non-cancer or mutagenic properties of pesticide
exposures are not address in this study.
Toxicity index (TI) values are calculated as the concentration of
a compound in the sample divided by the EC50 or LC50 of that compound
for an aquatic organism. All non-detects were treated as zeros for
this calculation. Individual TI values were summed for all pesticides
in a pesticide class. When an EC50 or LC50 value was not available
(for example there was no reported EC50 value for imazaquin on duckweed),
the mean value for the pesticide class was used. If less than 50%
of the pesticides in a class had acute toxicity estimates, then a
mean value was not calculated and a TI was not calculated for the
pesticide with missing data.
The TI values are summed within a pesticide class and for all classes
to determine additive pesticide class and total pesticide toxicity
indices. There is some debate over the validity of summing TI values
for classes of compounds with different modes of action. Studies of
mixtures of chemicals have generally concluded that an additive model
is appropriate for estimating the toxicity of mixtures of chemicals
with the same mode of action. However, for mixtures of chemicals with
differing modes of action, additive models may overestimate the mixture
toxicity and independent action models may be more accurate. Some
recent research has also documented synergistic (more than additive)
toxicity of mixtures of pesticides from different classes. Acute toxicity
estimates for herbicide transformation products were not found in
any of the listed sources. There are limited studies that show that
herbicide transformation products may or may not be as toxic as their
corresponding parent compounds. For this study, primary herbicide
transformation products where estimated to be one-half as toxic as
the parent herbicides.
TI values greater that 1.0 indicate probable toxicity of a class of
pesticides to the subject aquatic organism. TI values greater than
0.5 indicate potential toxicity, while TI values greater than 0.1
indicate limited toxicity.
The percentage of samples with TI values greater than 0.1, 0.5 and
1.0 for each class of pesticide and for total pesticides were calculated
and are shown in figure 3. None of the samples had probable, potential,
or limited toxicity from any of the five pesticide classes to bluegill
Less than 10 percent of the samples showed probable or potential toxicity
from ALS inhibitors to three of the four organisms (fig. 3A). There
was no toxicity data for frogs.
Less than 10 percent of the samples showed probable or potential toxicity
from triazine herbicides to any of the four organisms (fig. 3B). A
few samples had potential or limited toxicity from triazines to frogs.
For this calculation, the bullfrog (R. Catesbeiana) with a
reported LC50 of 410 ug/L was used as the test organism. LC50ís for
triazines on chorus frogs were not found. Reported LC50 values of
atrazine on other frog species ranged from 220 to 127,000 μg/L.
More than 15% of samples showed potential toxicity from chloroacetamides
to duckweed and 10% had TI values of 1.0 or more (fig. 3C). There
was no chloroacetamide toxicity data for frogs.
Only a few samples had limited or potential toxicity from organophosphate
insecticides to frogs (fig. 3D), all as a result of the occurrence
of chlorpyrifos. Chlorpyrifos toxicity to frogs is highly species
dependent, but data for chorus frogs was not be found. For this calculation
the American toad (Bufo americanus) with an LC50 of 1 ug/L
was used. Other reported LC50 values for chlorpyrifos on frogs ranged
from 10 to 3,000 ug/L. EC50 values for most organophosphates to duckweed
or green algae were not be found.
None of the samples had probable, potential, or limited toxicity from
carbamates to the four organisms (fig. 3E). There were no carbamate
toxicity data for frogs and limited data for duckweed and green algae.
The TI values for the 5 classes of pesticides were summed to estimate
a total pesticide TI values (fig. 3F). Duckweed was the most susceptible
organisms investigated; 17% of the samples had total pesticide TI
values for duckweed greater than 1.0 and 27% had TI values greater
than 0.5. For green algae, 8% of the samples had total pesticide TI
values greater than 1.0, and 15% had TI values greater than 0.5. Only
1% of samples had total pesti-cide TI values for frogs greater than
0.5, but there was limited LC50 data for frogs.
The effects of herbicide transformation product occurrence on the
potential toxicity of stream water to aquatic organisms have not been
well studied. TI values for triazine and chloroacetamide herbicides
and for total pesticides were recalculated using the assumption that
herbicide degradates were one-half as toxic as their parent compounds
to the four aquatic organisms.
The addition of the triazine degradates had only a minor effect on
the percentage of samples with TI values for duckweed or green algae
that were greater than 0.1, 0.5, or 1.0 (figs 3B and 3G).
However, accounting for the chloroacetamide degradates added substantially
to the percentages of samples with TI values for duckweed and green
algae that were greater than 0.1 0.5 or 1.0 (figs. 3C and 3H).
The effect on the total pesticide TI values was also substantial.
Twenty-six percent of the samples had total pesticide plus degradate
TI values for duckweed greater than 1.0, and 50% had TI values greater
than 0.5 (fig. 3I). Ten percent of the samples had total pesticide
plus degradate TI values for green algae greater than 1.0, and 23%
had TI values greater than 0.5.
Three conclusions can be drawn from the results of this investigation.
First, water in Midwestern streams during spring and early summer
runoff events can contain pesticides in sufficient quantities to
be toxic to non-target aquatic organisms. In this study, the concentrations
of some pesticides did exceed concentrations thought to affect aquatic
plants. Still, this data set may underestimate potential effects
of pesticides on aquatic systems in smaller streams because peak
concentrations are generally inversely related to stream size. Second,
accounting for herbicide transformation products can substantially
increase the estimated toxicity of stream water to aquatic plants.
More information is needed on both the occurrence of herbicide and
insecticide transformation products in streams and their toxicity
to aquatic organisms. Finally, the quality of this analysis is limited
by the lack of acute toxicity data for many of the pesticide-organism
This report is based on a poster presented at the 2001 Annual SETAC
meeting and an article published in the journal "Water &
Battaglin, W.A. and Fairchild, J. 2001. Potential toxicity of pesticides
measured in Midwestern streams to aquatic organisms. abstract. Society
of Environmental Toxicology and Chemistry 22nd Annual Meeting Abstract
Book, p. 204.
Battaglin, W.A. and Fairchild, J. 2002. Potential toxicity of pesticides
measured in Midwestern streams to aquatic organisms. Water Science
and Technology 45 (9):95-103.