United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/600/R-00/034
August 2000
Statistical Estimation and
Visualization of
Ground-Water
Contamination Data
-------
-------
EPA/600/R-0.0/034
August 2000
Statistical Estimation and Visualization of
Ground-Water Contamination Data
Rachel K. Boeckenhauer
Dennis D. Cox
Katherine B. Ensor
Philip B. Bedient
Anthony W. Holder
Rice University
Houston, Texas 77005-1892
Cooperative Agreement No. CR-821906
Project Officer
Mary E. Gonsoulin
Subsurface Protection and Remediation Division
National Risk Management Research Laboratory
Ada, Oklahoma 74820
National Risk Management Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
-------
Notice
The U.S. Environmental Protection Agency through its Office of Research and Develop-
ment partially funded and collaborated in the research described here under Cooperative
Agreement No. CR-821906 to Rice University. It has been subjected to the Agency's peer
and administrative review and has been approved for publication as an EPA document.
Mention of trade names or commercial products does not constitute endorsement or recom-
mendation for use.
All research projects making conclusions or recommendations based on environmen-
tally related measurements and funded by the Environmental Protection Agency are required
to participate in the Agency Quality Assurance Program. This project was conducted under
an approved Quality Assurance Project Plan. The procedures specified in this plan were
used without exception. Information on the plan and documentation of the quality assur-
ance activities and results are available from the Principal Investigator.
-------
Foreword
The U.S. Environmental Protection Agency is charged by Congress with protecting the
Nation's land, air, and water resources. Under a mandate of national environmental laws, the
Agency strives to formulate and implement actions leading to a compatible balance between
human activities and the ability of natural systems to support and nurture life. To meet these
mandates, EPA's research program is providing data and technical support for solving
environmental problems today and building a science knowledge base necessary to manage
our ecological resources wisely, understand how pollutants affect our health, and prevent or
reduce environmental risks in the future.
The National Risk Management Research Laboratory is the Agency's center for inves-
tigation of technological and management approaches for reducing risks from threats to
human health and the environment. The focus of the Laboratory's research program is on
methods for the prevention and control of pollution to air, land, water, and subsurface
resources; protection of water quality in public water systems; remediation of contaminated
sites and ground water; and prevention and control of indoor air pollution. The goal of this
research effort is to catalyze the development and implementation of innovative, cost-
effective environmental technologies; develop scientific and engineering information needed
by EPA to support regulatory and policy decisions; and provide technical support and
information transfer to ensure effective implementation of environmental regulations and
strategies.
This work presents methods of visualizing and animating statistical estimates of ground
water and/or soil contamination over a region from observations of the contaminant for that
region. The primary statistical methods used to produce the regional estimates are non-
parametric regression and geostatistical modeling (kriging). Nonparametric regression can
be used as a more "rough and ready" method to produce surface estimates with little outside
intervention, whereas geostatistical modeling produces prediction errors. Finally, a method
is proposed for estimating the total amount of contaminant present in a region. This report is
published and made available by EPA's Office of Research and Development to assist the
user community.
Clinton W. Hall, Director
Subsurface Protection and Remediation Division
National Risk Management Research Laboratory
-------
Abstract
This work presents methods of visualizing and animating statistical estimates of ground
water and/or soil contamination over a region from observations of the contaminant for that
region. The primary statistical methods used to produce the regional estimates are non-
parametric regression and geostatistical modeling (kriging). Nonparametric regression can
be used as a more "rough and ready" method to produce surface estimates with little outside
intervention, whereas geostatistical modeling produces prediction errors.
Animation of changes in the estimated level of contaminant or chemical as observa-
tions are removed illustrate the effect of each individual measurement on the overall estimate
and the error or variance of this estimate. Such methods are applied to the Eglin Air Force
Base (AFB) Florida site. The benefit of animating surface estimates in data which is taken
over time is clearly seen by an example from a site near Phoenix, AZ, where aberrations in the
data for one or several years were readily apparent by viewing a smoothed animation.
Finally, a method is proposed for estimating the total amount of contaminant present in
a region. The proposed method models the data as a realization of a lognormal stochastic
process and then capitalizes on conditional simulation to generate realizations of the
modeled process from which the distribution of the total contaminant (or integral of the
process) is estimated.
This report was submitted in fulfillment of cooperative agreement No. CR-821906 by
Rice University under the sponsorship of the United States Environmental Protection
Agency. This report covers a period from 10/01/93 to 03/31/97, and work was completed as
of 11/03/98.
IV
-------
Contents
1. Introduction 1
2. Sites and Data 3
2.1EglinAFB 3
2.2 Arizona 4
3. Statistical Methods 5
3.1 Nonparametric Regression 5
3.2KrigingandVariograms 5
3.3 Estimating an Integral via Sample-Mean Monte Carlo 6
4. EglinAFB: Visualization and Exploratory Analysis 9
4.1 Visualization of Estimated Plumes ._. , V. 9
4.1.1 Two-Dimensional Data 9
4.1.2Three-DimensionalData 10
4.2 Error Visualization 11
5. Estimation of Non-Linear Functionals of Random Processes for Environmental Problems 15
5.1 Description of the Problem 15
5.2 Using Monte Carlo to Estimate the Distribution of a Stochastic Integral 16
5.3 Application 17
5.3.1 Discussion of Data 17
5.3.2Results 29
6. Phoenix, AZ: Visualization with a Time Component 31
6.1 Exploratory Visualization 31
6.2 Animation 31
6.2.1 Trichloroethylene andDichloroethylene (TCE andDCE) 31
6.2.2 Sulfate Ions 38
6.3 Further Analytical Efforts 41
7. Summary and Conclusions 43
Appendix A. Cross-Validation 45
A.I Two-dimensional Data 45
A.2 Three-dimensional Data 45
Appendix B. Discussion of Spatial Estimation for Arizona 46
Bibliography 49
-------
List of Figures
2.1 Eglin AFB ground-water data points. Coordinates: depth-axis: 4400 to 5100;
width-axis: 4900 to 5600 and vertical-axis: Omg/kg to 9. Img/kg 3
22 Eglin AFB soil data points 4
4.1 Legends for two-dimensional perspective plot contours 10
42 Estimate of contaminant plume for Eglin AFB ground-water data 11
43 Estimate of soil contamination 7.0/?below the water table 12
4.4 Estimate of soil contamination (BTEX) at Eglin AFB 12
45 Estimates of ground-water contamination and absolute errors : 13
4.6 Estimate of ground-water contamination: contours represent magnitude of error 13
4.7 Smoothed absolute error estimate for Eglin AFB soil data 14
5.1 QQ-plot of logs of Eglin AFB ground-water data 18
5.2 Logs of Eglin AFB ground-water data points 18
53 Empirical semivariogram of logged Eglin AFB ground-water data 19
5.4 Empirical semivariograms with rotation angles from 0° on the left to
180° on the right and ratios from 1:25 at the bottom to 2 at the top 19
55 Empirical semivariograms with rotation angles from 45° on the left to
135° on the right and ratios from 1:45 at the bottom to 1:55 at the top 20
5.6 Semivariogram of logged Eglin AFB ground-water data with transformed locations 22
5.7 Kriged surface estimate of logged Eglin AFB ground-water data using
rational quadratic variogram details of the integral estimation 22
5.8 Standard errors for kriged surface estimate of logged Eglin AFB
ground-water data using a rational quadratic variogram 23
55 Surface of a grid simulation 25
5.10 Semivariogram calculated from a grid simulation 25
5.11 Histogram of integral estimates from 1000 samples of size 500: one realization 26
5.12 QQ-plot of integral estimates from 1000 samples of size 500: one realization 26
5.13 Histogram of integral estimates from 1000 samples of size 500: different realizations 27
5.14 Histogram of lower 97.5% integral estimates from 1000 samples of size 500: different realizations 27
5.15 Histogram of logged integral estimates from 1000 samples of size 500: different realizations 28
6.1 TCE prediction surface for 1991. Orientation: depth-axis: 892,600 to 896,800;
width-axis: 478,000 to 484,000; vertical-axis: 0 to 2.25 32
62 TCE standard errors of prediction for 1991 32
63 DCE prediction surface for 1991 33
6.4 DCE standard errors of prediction for 1991 33
65 TCE prediction surface for 1992 34
6.6 TCE standard errors of prediction for 1992 34
6.7 DCE prediction surface for 1992 35
6.8 DCE standard errors of prediction for 1992 35
vi
-------
6.9 Legend for prediction surfaces (measurements in mg/l).
6.10 Legend for standard errors
6.11 TCEandDCE surfaces for 1986
6.12 TCEandDCE surfaces for 1988
6.13 TCEandDCE surfaces for 1991
6.14 TCEandDCE surfaces for 1993
6.15 SO^" surface for 1986
6.16 SO^" surface for 1988
6.17 SOi" surface for 1991
A. 1 Bandwidth selection for ground-water data.
B. 1 Classical empirical variogram for 1990 TCE..
B2 Meany values for 1990 TCE
36
36
37
37
38
39
39
40
40
45
46
47
VII
-------
Tables
Table 5.1 Statistics of Integral Estimates for 1000 Realizations
Table 5.2 Means of Integral Estimates for 1000 Realizations ....
29
29
viii
-------
Acknowledgments
This project was initiated by Dr. M. E. Gonsoulin and Dr. C. Enfield, U. S. Environmental
Protection Agency. We thank Dr. S. Schmelling of the U.S. Environmental Protection
Agency, Dr. R. Michler of the University of North Texas, and Dr. D. McLaughlin of the
Department of Civil and Environmental Engineering Massachusetts Institute of Technol-
ogy, for their reviews and valuable comments. Also, we thank Ms. J. Elliott of the U. S.
Environmental Protection Agency and Ms. M. Williams of Orbiting Astronomical Observa-
tory Corp. for providing graphical support and developing the final format of the report.
ix
-------
-------
Chapter 1
Understanding the plume of any contaminant is a multi-faceted problem. Often data is limited and modeling exercises are
tedious. To aid the environmental researcher, we present several techniques for capitalizing on the measurements of the level
of contaminant in soil samples from a region. Our techniques rely on visual displays of statistical estimates of the contaminant
plume. We explore two- and three-dimensional estimation and visualization techniques and ways to examine related
contaminants. Furthermore, we propose a method for quantification of the total amount of contaminant within a region. Both
of these methods are investigated in the context of a site specific example, but the tools generalize to other similar problems.
As a simple method for displaying surface estimates from field data, exploratory visualization of the Eglin AFB in Section 4.1
was performed using nonparametric regression to produce the surface estimates. This methodology requires estimation of only
one parameter, the bandwidth, as opposed to the several parameters required by the more complicated art of variogram fitting
used in geostatistical modeling. Surface estimates were obtained for both two- and three-dimensional data for the Eglin AFB
site, using the program Geomview on a Silicon Graphics machine to facilitate display and animation. Geomview allows the
viewer to rotate images in real time, which aids greatly in examining the surface (i.e., looking for peaks and valleys, etc.). For
data in two dimensions, the third dimension (i.e., the z direction) may be used to plot the surface estimate as a perspective plot.
Color contours on the surface can be set to levels of interest to highlight areas where the contamination is above a fixed level.
This tool can be useful in cases where environmental regulations require contamination to be below some specific level, to
identify regions of high contaminant concentration, and to follow the movement of a contaminant over time. In an analogous
fashion, for three-dimensional visualization, shell contours are plotted at certain levels to illustrate regions of higher
concentrations.
Estimates of prediction errors in both the two-dimensional and three-dimensional setting provide an understanding of the
differing levels of uncertainty of the estimate of the level of the contaminant over the region. In the case of Eglin AFB ground
water, we also use visual tools in conjunction with cross-validation to ascertain the effect each of the data values has on the
estimate of the level of contaminant for the region. An animation of estimates produced excluding individual data points,
alternated with the overall surface estimate, lends insight to the question of where to obtain new samples. This sort of display
also helps us to determine the level of error in our estimates of the contaminant plume.
A second site providing a different type of complexity was examined from a statistical estimation and visualization perspective.
This second site, in Arizona, yields observations of several contaminants collected over a period of several years. Animations
of estimates capitalizing on the temporal component clearly illustrate major trends and aberrations in the data, which can then
be investigated more closely. Also, there were several different contaminant substances measured at the Arizona site, such as
TCE, DCE, and SO 4" - The behavior of these contaminants is expected to be interrelated. We present suggestions on how to
best visualize simultaneously two or more related substances in order to highlight possible relationships among the series. The
Arizona site also includes a common problem in that the region where measurements were taken increases over time. This
problem is addressed here, along with some possible solutions.
Another primary focus of this research is to answer the question of how much total contamination is present at a site and how
to best estimate this quantity given soil core samples from the site. It is important when producing such estimates to also
understand the level of uncertainty in the estimate, in other words to obtain the standard error of the estimate. A byproduct of
carefully implemented geostatistical methods such as kriging is standard errors for the estimated mean level over the region.
Estimation of total contaminant involves estimation of the integral of the modeled process over a region. We pursue estimation
of the total contaminant for Eglin AFB. The level of ground water BTEX was modeled as a realization of a lognormal
stochastic process, and estimates of the distribution of the integral were produced by Monte-Carlo simulation of the process
conditional on the observed data.
The data used in this research, from Eglin AFB and a site near Phoenix, AZ, are presented in Chapter 2. Chapter 3 contains
discussion of the statistical methods used, including estimation by nonparametric regression and kriging. Chapter 4 contains
results for Eglin AFB, including both exploratory data analysis and estimation of total contaminant Chapter 6 contains
discussion of the visualization and animation for the Arizona site, including discussions of temporal data and visualization of
two related substances. Finally, Chapter 7 contains conclusions and suggestions for future research.
-------
-------
Sites and Data
2.1£gIinAFB
In Chapters 4 and 5, we consider data from Eglin AFB, an example of a shallow aquifer in sandy soil. A leak of 30,000 -
40,000 gallons of JP-4 jet fuel was detected at Eglin AFB in Florida by Air Force personnel in 1984 (Boeckenhauer,
-------
Figure 2.2 Eglin AFB soil data points.
2.2 Arizona
In Chapter 6, we consider data from a contaminated site near Phoenix, AZ. Contaminants measured include trichloroethylene
(TCE), dichloroethylene (DCE), and sulfate (SO^"), all measured in \ig/L. Other measured contaminants contained somewhat
sparse data and were not used at this time. The DCE "measurements" are actually sums of measured values of 1,1-DCE and
1,2-DCE, so the measurement locations here are only used if measurements of both of these are available. These data were
gathered from 1985 to 1993.
-------
Chapter 3
Statistical Methods
3.1 Nonparametric Regression
For the exploratory analysis of Eglin AFB in Section 4.1, we used nonparametric regression to produce a surface estimate of
the plume. The model used for the contaminant plume is u(x) = f(x) + e(x) where:
x = a point hi the region of interest,
z/(x) = the observed level of contaminant at x,
Xx) = the true level of contaminant at x, and
e(x) = random noise in the measurement process.
The model assumes:
e The observation locations x[,...,xn are randomly chosen.
• The unknown function f(-) is twice continuously differentiable. Note that f(') is not a random process.
• The random noise E(X( ),...,e(xn) is independent, but not necessarily identically distributed.
An estimate of f(x) over the region of interest can be obtained via nonparametric regression methods (Scott, 1992) and is given
by:
(3.1)
where the weights are defined as:
•£*(*-*/)
(3.2)
and z/pz^Xj). The function Kh (•) is referred to as the scaled kernel function with bandwidth, or smoothing parameter, h. Note
that the bandwidth determines the smoothness of the surface estimate; a larger bandwidth yields a smoother estimate. For some
standardized (or unsealed) kernel function ^j(«), we define K^-) as K/t) - K(t/h)/h, so as h increases, the value off/h=(x- xj/h
decreases.
3.2 Krigmg and Variograms
Another method which we will use to estimate contaminant levels over the region is the geostatistical spatial prediction method
known as kriging. A complete and thorough exposition of geostatistical methods is given in (Cressie, 1993). A very brief
overview is provided here for purposes of introduction and definition of notation. The general idea of kriging is to first use the
observed levels of a contaminant to produce a model of the spatial covariance structure of the process. This spatial covariance
model is then used to obtain the "optimal" predictor p(Z; s0) of 2[s0), the value of the random process at s0. This predictor is
£[2X$0 )|Z], which is precisely the same asjy° (Z; s0) in (5.2) in the case where g(') is simply g(Z) - Z.
Assume that the data z = (Xs, ), ...Xs,,)) are a sample from a realization of the stochastic process (2(s): s e A}. In order to do
inference from the data, we need to make some assumptions. A common practice is to assume second-order stationarity. That
is, assume that
:{1 Vse^4 " (3.3)
(3.4)
(3.5)
or that Fs(z) = Pr(Z(s)(Z(si),Z(s.)) depends only on the distance \st - s.|, then the process is said to be isotropic.
-------
If second-order stationarity holds, a convenient way to model the covariance structure of the process is through use of a
variogram function
vc,r(z(Si)-Z(sj))=2j(Si-Sj) Vs,,sje4 (3.6)
The function 7(s. - sp is referred to as the semivariogram. Note that it is easy to show that 7(h) = 7(-h) and 7(0) = 0. If we have:
A-»0
then c0\s what is known as a nugget effect. This may be due either to some microscale variation or to measurement error. The
term nugget effect comes from spatial prediction's origins in mining, and refers to a variation caused by small nuggets of ore.
Regardless, in all real data, there is some measurement error and so we would be remiss to model our process without a nugget
effect. On the other end, as |h|-»°°, the semivariogram converges to the process variance. This follows easily by noting that we
assume that the covariance between two values of the process diminishes to zero as the distance between them increases.
In the case that the process is isotropic; i.e., the spatial covariance between values of the process depends on/yon the distance
between the observations, there are a number of standard variogram models available (Cressie, 1993). However, if the process
is anisotropic, it is sometimes possible to transform the locations so that an isotropic variogram model remains appropriate.
Specifically, such a transformation is possible in cases of geometric anisotropy; i.e., where rotating and scaling the locations
produces an isotropic process. For example, it is typically the case with ground-water data that the correlation is higher for
points a distance h apart if they lie in the direction of ground-water flow rather than perpendicular to it. In this case, the
variogram is of the form:
2Y(h)=2To(|Bh|)
h e A c 91
(3.8)
where B is a rf'x^/matrix and y0 is an isotropic variogram. We will be using this type of transformation on the Eglin AFB
ground-water contamination data where d= 2. In the case d- 2, the matrix B is given by:
"cos2 (9) + r*sin2 (0) (1 - r)* sin(9)* cos(9)"
(1 - r)* sin(9)* cos(9) sin2 (9) + r* cos2 (9)
(3.9)
meaning that d is the angle clockwise from North at which the scale is multiplied by r. The other axis is then the one which is
perpendicular to this, and the scale in this direction is not altered. For example, then, if we were working with ground-water
data where the flow was along the northwest-southeast direction, we might use 9=135 deg and some /•>!.
Assuming the modeled spatial covariance structure of the random process, we can now obtain optimal predictions. If our
optimization criteria is minimization of the squared-error loss then the optimal predictor of the random process at any point s0
is given by the expectation of the random process conditional on the observed values of the process; in other words, the best
predictor of 2(s0) is/?/^s0) = £(2(sa )\Z), where ^denotes the vector of data as in (Boeckenhauer, 1996). In the case that Z(?)
is a Gaussian process, this predictor is linear. Here we will be using the form of spatial prediction known as ordinary kriging
(Cressie, 1993) which requires the two assumptions:
1. There is a constant mean, i.e.
for s&A and |ie9l unknown.
2. The predictor is linear in the observations, i.e.
(3.10)
(3.11)
where the observations are at locations t15 ...,tn and ^T" A,,- = 1 . (Recall that this predictor is an estimate of £[%_s0)\2(t]),...,Z(tr)~\.)
Requiring ]£" X,- = 1 guarantees uniform unbiasedness, i.e. £[p(Z,s^) = (I = £(^sa)). For further information on optimal
prediction using the kriging equations, see (Cressie, 1993).
33 Estimating an Integral via Sample-Mean Monte Carlo
The question of the total amount of contaminant within a given region is equivalent to estimating the integral over the region
of the estimated spatial process for this contaminant. Monte Carlo methods provide an estimate of the integral (see also
(Rubinstein,1981) and (Hammersley and Handscomb, 1964)) by viewing the integral as an expectation and simulating the
sample mean as an estimate of this expectation. For example, suppose we wish to estimate the integral over a region^ of some
function ^(s):
-------
j(s)ds (3.12)
The basis of this method is to represent the integral O as the expected value of a random variable. For example, suppose that
S is a random variable which has density /Is) on A. We may then rewrite the integral 3> in (3.12) as:
g(s)1
L/s(S)j
provided that fs(s) > 0 when q(s) & 0.
In particular, suppose S is uniform on A. That is, S has density,
(3.13)
(3-14)
where
1 if s 8 A
0 otherwise
(3-15)
is the indicator function on A and ]A\ is the norm of A (e.g., the area or volume of A if A is two- or three-dimensional,
respectively). We may then simplify the integral in (3.13)
<& = J q(s) ds = E
= E
LV|A|J
= |A|E[q(S)]
(3.16)
To use this method to estimate <&, then, we will
1. Generate a "large" (say/>) number of locations, Sj,...,s uniformly over the region A.
2. Evaluate q(s) at each of these locations, yielding ^(s,) #(sp).
3. Compute the sample mean of these evaluations to yield an estimate of £[g(S)], i.e. ~2Ltj=\<2\sj).
4. Use the sample mean in (3.) to estimate the integral <3>, or
(3"1?)
The error inherent in this method relates to the randomness of the sampled sites and the number of sites which we sample.
Specifically, the variance of the integral estimate for the function q is given by:
(3.18)
\ P 7=1
This variance may be estimated by:
(3.19)
where s2 is the sample variance of q(s^,...,q(s ).
In order to estimate the total level of contaminant say for the Eglin AFB, we model the spatial process of interest. Using our
model as the truth, we can generate a/?-dimensional Multivariate Normal random vector with the appropriate mean structure
-------
and spatial covariance structure as given by our estimated model (see, for example, Johnson, 1987; Stewart, 1973). This
simulated random vector is then used in the above algorithm to ascertain the total amount of contaminant present, and the
standard error of this estimate.
-------
>r4
Eglin AFB: Visualisation and Exploratory Analysis
Our exploratory analysis of the observations of BTEX from Eglin AFB provides an understanding of the location and shape of
the contaminant plume. For the two-dimensional data, perspective plots of the surface estimate with color contours visually
display both the level of contamination and the rate of change over the region. The color contours can be set to specific
concentrations of interest, such as regulatory levels. In the three-dimensional case, nested contour visualization is used to
provide immediate characterization of the plume. Again, the contours could be keyed to concentrations of interest. Such
exploratory analyses are greatly enhanced by on-line manipulations of the visual tools provided. For example, it is possible to
rotate the surfaces to search for high levels of contamination which may be visible only from certain vantage points.
Furthermore, color facilitates identification of trouble spots. However, even the gray-scale static versions of the plots presented
here are useful for providing visual understanding of the plume.
The surface estimates for the exploratory visualization were produced using nonparametric regression, as discussed in
Section 3.1. In order to obtain an accurate surface estimate of the plume, it is necessary to choose appropriate bandwidths to
produce the nonparametric regression estimate. (Recall that the larger the bandwidth, the smoother the final surface estimate
will be.) Appendix A contains the details of bandwidth selection, via cross-validation, for both the two-dimensional and
three-dimensional Eglin AFB data.
Properties of spatial estimates and/or nonparametric regression estimates rely on asymptotic theory which, due to the small
number of observations available, is certainly not the situation with the data at hand. Therefore, we explore the robustness of
our estimate by examining the change hi the estimate as sample points are removed from the estimation process. The uses of
this examination are two-fold: (1) a better understanding of the the magnitude of the error of our estimates is obtained, and (2)
areas where additional observations are needed are highlighted. In other words, intuitively, we would take additional
observations in the region where the estimated level of contamination changed the most as data was removed. If removing an
observation has little effect on the estimate resulting in a small error, additional data would not be needed in that region. This
result was visualized in two ways:
1. by viewing an animation of the estimated surface alternated with surfaces estimated by the removal of one of the
sample points, and
2. by visualizations of the absolute differences (errors) between the surface estimated without point / and the
measured value at point i.
The animation in case (1) is very useful when viewed on the SGI, but does not appear here. Case (2) appears in Section 4.2.
By viewing a smoothed version of the absolute error of our plume estimate, both in the two-dimensional and three-dimensional
cases, it is clear where additional observations are needed. One suggestion for future sampling sites would be to sample in the
region where both the estimated level of contamination and the error associated with the estimate are high. Of course, any
measure of error or the amount of information contained in the data could be displayed in a similar fashion.
4.1 Visualization of Estimated Plumes
The program Geomviewfcn: an SGI was used to display plume estimates of the BTEX concentration. Geomview was written at
the NSF Geometry Center, University of Minnesota, and is available through anonymous ftp from ftp.geom.umn.edu. The
program ashreg, a modification of ashn (Scott, 1992), was used to produce plume estimates from the three-dimensional data.
A biweight kernel was used in all cases.
4.1.1 Two-Dimensional Data
Figure 4.2 contains a plot of the estimated plume for the Eglin AFB ground-water data. Figure 4.1 shows the legends for these
contour levels and those for the later figures, signifying estimated concentrations of:
(a) 5000
(b) 2000
(c) 1000
500
100
(d)
(e)
(g)
(h)
(0
10
0
U <
U <
U <
U <
U <
U <
U <
U =
No data
10000
5000
2000
1000
500
100
10
0
(4.1)
-------
Figure 4.1 Legends for two-dimensional perspective plot contours.
where all of the concentrations are in lig/L. The level labeled No data includes areas where there is no estimate as we are too
far from any of the measured points for the given bandwidth. (Note that, for the black and white figures, some of the lower
levels are shown in the same shade, but we are primarily interested in areas where there are high concentrations, and these are
distinctly different.) The same legend applies to the plot of the estimated plume for the Eglin AFB soil data at a depth of
approximately 1.0 Jf given in Figure 4.3, except that here the concentrations are:
(c)
(d)
00
(/)
to)
(h)
(0
50 <
10 <
5 <
1 <
0 <
u <
u <
u <
u <
u <
u =
No data
100
50
10
5
1
0
with all measurements in mg/kg. (Note that at this particular depth, the estimated concentration does not exceed
For both of the two-dimensional plots, the actual data locations are marked on the grid at the top of the plot. The large point
which is connected to the perspective plot by a vertical line is the mode of the estimate. An arrow pointing north indicates the
orientation of the plot.
4.1.2 Three-Dimensional Data
Figure 4.4 contains a plot of the estimated plume for the Eglin AFB soil data. Two different levels of contamination are
represented, with the lower (outer) shell being sliced so that we can see the higher (inner) one. The outer and inner shells
represent concentration levels of approximately 0.73 and l.^mg/kg, respectively. Also, the location of the water table and the
source are clearly marked.
10
-------
In an on-line version of the 3-D visualization of the plume, nested contours can be displayed using the transparency feature of
the SGI and Geomview. This allows the use of solid color nested contours, where the inner shells can be seen through the outer
ones. The use of transparency and the ability to rotate the graph greatly enhance the informative 7. Oy? below the Water Table
value of the plot to the user.
4.2 Error Visualization
We simultaneously display a perspective plot of the two-dimensional ground-water BTEX plume estimate and the absolute
errors of this estimate in Figure 4.5. Note that the heights of these two plots are on the same scale. The contours for the errors,
using the legend in Figure 4.1, are the same as given in 4.1 except for levels (a) and (b):
(a) 3000 < u < 5000
(b) 2000 < u < 3000
with errors in absolute \ig/L. In Figure 4.6, we show a perspective plot of the ground-water data where the contours are
determined by the smoothed absolute errors. By combining the plume estimate and the absolute errors into a single plot, we
more readily identify regions of high concentration and regions with a large amount of uncertainty in the estimated value.
To visualize the errors for the three-dimensional soil data, we simply took the absolute errors as calculated for cross-validation
in A.2 (i.e. |u, — u|, . . . , |un — un||) and plotted a smoothed contour shells (see Figure 4.7). Here the outer and inner shells
represent contaminant concentrations of approximately 0.48 and ^.ldmg/kg, respectively.
Figure 4.2 Estimate of contaminant plume for Eglin AFB ground-water data.
11
-------
Figure 43 Estimate of soil contamination 7.0y?below the water table.
Figure 4.4 Estimate of soil contamination (BTEX) at Eglin AFB.
12
-------
Figure 4.5: Estimates of ground-water contamination and absolute errors.
Figure 4.6 Estimate of ground-water contamination: contours represent magnitude of error.
13
-------
Figure 4.7 Smoothed absolute error estimate for Eglin AFB soil data.
14
-------
Chapter 5
Estimation of Non-Linear Functional of
Random Processes for Environmental Problems
In analyzing environmental sampling data, it is often of interest to estimate some function of the data. For example, one may
be interested in the maximum concentration attained within some region, the location of this maximum, the region for which
the concentration exceeds some set value, or the total amount of contaminant present in a region. For example, government
regulations on levels of ambient ozone typically involve exceedence of some threshhold deemed unsafe for human beings at any
location within a region (Cox, et al.,1995). In this case, one would wish to estimate the maximum concentration attained in the
region of interest. In the case where one is attempting to clean up ground water or soil contamination, it is of interest to know the
total amount of contaminant present in a region. The estimation of total contaminant involves estimating an integral over a
region and is what will be addressed in this section.
In this chapter, we will model the Eglin AFB ground- water BTEX observations as a realization of a stochastic process using the
methods described in §3 .2. The goal, then, will be to estimate the distribution of the integral of the process over some set region.
Now in the case where this process is Gaussian, estimation of this distribution is a solved problem. However, it is commonly the
case that environmental data are lognormal
-------
provided w/(Z(s)| x) > 0, so we cannot simply use the conditional mean of the ), s e A denotes this realization. We then estimate the integral of the realization in the manner discussed in §3.3:
(5.6)
P j=i
The variance of the integral estimate for the realization, conditional on the data and the realization, may be estimated by
(5.7)
as in (3.19). To estimate the variance of the integral of the process, notice that
Var [6|x] = Var
= Var
(5.8)
(Recall that x is the vector of data.) An unbiased estimate of #1 Vca^b X(*)j data^ is obtained from the sample mean of the
variances of the O , i— 1>— , m. Also, Van O|x may be estimated by the sample variance of the integral estimates from the
realizations. Sampling from a distribution which is not uniform may decrease the variance of the estimate in some cases,
however, it is the large variation between the integral estimates from different realizations which should be a cause of concern.
This latter error can only be reduced by either (1) improving the model, or (2) gathering additional data.
The above steps are repeated to obtain a large number of realizations thereby reducing error introduced by Monte Carlo
estimation. For each simulation /; /= 1,..., m, we:
1 . Generate/' locations s^ ..... s^ uniformly in A.
2. For these locations, calculate the covariance matrix and mean vector, conditional on the data, using the
geostatistical model.
16
-------
3. Generate multivariate normal dataz^s. t ),...,Zj(s.) using this mean vector and covariance matrix, and exponentiate
them (using the same base as for the logs we took of the data to produce the model).
4. Produce an estimate of the integral
«<•«)
using the sample-mean Monte Carlo method, i.e.
6, = „>
p j=i
We may then get an estimate of the mean of the integral of the process
4. . We may also estimate the variance of the integral
Far[0|;c] =Kar[«Jc]-
m ;=
(5.13)
as discussed above.
5.3 Application
5.3.1 Discussion of Data
A QQ-plot of the log of the ground-water BTEX data from Eglin AFB, Figure 5.1, indicates that lognormality is a reasonable
assumption in this instance. For further reference, we also plot the log of the data at the sampled spatial locations (Figure 5.2).
We now discuss the covariance modeling and integral estimation for the Eglin AFB site using the ground-water data introduced
in Section 2.1. Recall that these data appear to be lognormal. In Section 5.3.1, we will estimate a variogram from the log 10 data
to model the covariance structure, as discussed in Section 3.2. Section 5.3.1 contains details regarding integral estimation for
this data, and thus estimation of the total amount of contaminant.
Variogram Estimation
Before performing spatial estimation on any data, we must first model the covariance structure of the data. The empirical
semivariogram for the logged Eglin AFB ground-water data is shown in Figure 5.3. While this semivariogram does, in general,
seem to increase with distance as expected, it nonetheless is rather undesirable. In particular, there is high variability, which (a)
will make it difficult to estimate well with a vaiogram model and (b) will cause any estimates made from such a model to have
large error variance.
It should be noted that the above variogram was done assuming that the data was isotropic, that is, that the covariance of the
process at two locations depends only on the distance between these locations and not the direction. In fact, this does not really
appear to be the case here, as Figure 5.2 seems to indicate that the plume extends along the northwest-southeast direction, due
to the fact that the "large" observations seem to fall about this line. So it is possible here that we can get a better variogram, and
thus a better fit, by first transforming the coordinate axes through a rotation and then scaling one of the axes. This allows us to
take (this particular type of) anisotropic data and model it using a standard one-dimensional, isotropic variogram model. Using
the Splus spatial module function anisotropy.plot, we are able to try various rotations and scalings of the coordinate axes in an
attempt to produce a more appropriate variogram. Examples of these empirical variograms, along with loess smooth lines, are
shown in Figures 5.4 and 5.5. Note that a number of these are obviously bad choices, as they do not even have y increase with
distance. We are looking for something with the points "well-clustered" about an increasing line. We concluded the best
17
-------
ffl
u.
ff
I
-2
Quantiles of Standard Normal
Figure 5.1 QQ-plot of logs of Eglin AFB ground-water data.
Figure 5.2 Logs of Eglin AFB ground-water data points.
18
-------
in
d
p
d
100
200
distance
300
Figure 5.3 Empirical semivariogram of logged Eglin AFB ground-water data.
oo
O -2.0
•1.5
-1.0
-0.5
100 200 300 400 500 100 200 300 400 50C 100 200 300 400 100 150 800 250 300 100 200 300 400 500
ca
CO
en
2.0-
1.5
1.0-
0.5-
100 200 300 400 BO 100 150 200 250 300 350 100 150 200 250 300 350 100 150 200 250 300 100 200 300 400
100 200 300 400 100 150 200 250 300 350 100 150 200 250 300 100 150 200 250 300 100 200 300 400
SO 100 150 200 250 300 3560 100 150 200 250 300 350 100 150 200 250 100 150 200 250 30K> 100 150 200 250 300 351
distance
Figure 5.4 Empirical semivariograms with rotation angles from 0° on the left to 180° on the right and ratios from 1:25 at the bottom to
2 at the top.
19
-------
2.0-
1.5-
1.0-
0.5-
o o
o o
135
CO
o>
100 150 200 250 300 35 100 150 200 250 300 35 100 150 200 250 300
2.0-
1.5-
1.0-
0.5-
100 150 200 250 300 350 100 150 200 250 300 100 150 200 250 300
O O
100 150 200 250 300 350 100 150 200 250 300 100 150 200 250 300
distance
Figure S3 Empirical semivariograms with rotation angles from 45° on the left to 135° on the right and ratios from 1:45 at the bottom
to 1:55 at the top.
20
-------
option was a rotation angle of 45° and an axis ratio of 1.45. That is, for the f' data point at location (x[i\, j{/]), we multiply the
location by the matrix B where
B =
( cos2 (45) +1.45* sin2 (45) (1 -1.45)* sin(45)* cos(45)^j
[(1 -1.45)* sin(45)* cos(45) sin2 (45) +1.45* cos2 (45) J
( 1.225 -0.225N
I -0.225 1.225
(5.14)
as in (3.9). This empirical semivariogram, along with two different semivariogram models, is shown in Figure 5.6. The
spherical semivariogram, with the formula
fO A = 0
y(/z) = \c0+ cs(l.5*(h/as)-.5*(h/as)3) 0 < h < as
[c0 + cs h>as
(Cressie, 1993) was produced with nugget cg— 0, (partial) sill cs— 1.51, and range. ^ = 325. This model appears as the dashed
line in Figure 5.6. The rational quadratic semivariogram has formula
c0 +
0
c*h2
h > 0
(5.15)
(Cressie, 1993) and was calculated with nugget ^=0.19, cr— 1/14580, and ar— 24750. This variogram model appears as the
solid line in the figure. In both cases, h refers to the distance (in the transformed space) between the data points. The two
variogram models appear very similar for most distances in Figure 5.6, but are different in the tails. While either one would
likely work reasonably well, we chose to use the rational quadratic model as it produces a plausible variogram which has a
positive nugget effect (and a nugget effect of zero would assume that there was no measurement error, which is certainly an
unreasonable assumption). The kriged surface estimate produced using this variogram model is shown in Figure 5.7 and the
standard error surface is depicted in Figure 5.8. The kriging here was done using the spatial module of S-PLUS (MathSoft,
1995), which estimated the constant mean to be p. = 1.55344 on the loglo scale.
Details of the Integral Estimation
Programs to estimate the integral for the Eglin AFB ground-water data were written in the C programming language and appear
in (Boeckenhauer, 1996). The method is essentially that described in Section 5.2. That is, for each of M realizations of the
process, we simulate values at N locations within the region, where N and M are contained in the file cons tants . h. The main
program, contained in runsim. c, first uses the data locations to produce Su , the covariance matrix for these locations. As
discussed above, the locations are first corrected for geometric anisotropy by multiplying each pah- of locations tj = (x,y) by the
matrix B hi (5.14). Then the distance (in transformed space) between each pair of points is calculated, followed by the
covariance between the locations using the rational quadratic model selected above. Specifically, notice that since we are
assuming the data are of the form
where e. and £,. are independent for fcj and
) = Z(t,.) + e, *• = !,... ,22
is independent of &. for every / andyj we have
(5.16)
Cbv(z(t,.),z(ty.)) if
if
(5.17)
21
-------
« .
Ol
°. .
100
200
distance
300
Figure 5.6 Semivariogram of logged Eglin AFB ground-water data with transformed locations.
predicted'
Figure 5.7 Kriged surface estimate of logged Eglin AFB ground-water data using rational quadratic variogram details of the integral
estimation.
22
-------
standard error
0.6
Figure 5.8 Standard errors for kriged surface estimate of logged Eglin AFB ground-water data using a rational quadratic variogram.
Recall that we are using the rational quadratic model. So letting h = t,. - ty (i.e., h is the distance between locations i and j
after transforming locations), we have:
and
(i.e., the nugget effect). Thus the /;_/* element of Zu is given by:
crar
=
c0 +crar- 2, =c0+cra for i = j
(5.18)
(5.19)
(5.20)
23
-------
Recall that
-------
predicted
Figure 5.9 Surface of a grid simulation.
03
I 2
CO
O)
q
O
200
Figure 5.10 Semivariogram calculated from a grid simulation.
400
distance
600
25
-------
o
II
n
o
11
o J
4*10*9
6*10A9
8*10A9
int
Figure 5.11 Histogram of integral estimates from 1000 samples of size 500: one realization.
10*10
1.2*10*10
o
i
a 2 -
-z
Quantiles of Standard Normal
Figure 5.12 QQ-plot of integral estimates from 1000 samples of size 500: one realization.
26
-------
I
2*10*11 4*10*11 6*10*11 8*10*11 10*12 1.2*10*12
estimates
Figure 5.13 Histogram of integral estimates from 1000 samples of size 500: different realizations.
S -,
8-
8 -
8-
8-
10*10 2*10*10 3*10*10 4*10*10 5*10*10 6*10*10
estimates
Figure 5.14 Histogram of lower 97:5% integral estimates from 1000 samples of size 500: different realizations.
27
-------
o
a
8-
8-
10 11
Iog10(est)
12
Figure 5.15 Histogram of logged integral estimates from 1000 samples of size 500: different realizations.
The function sim then calls the function multnorm, located in multnorm. c, which uses p. and X to generate multivariate
normal random variables from ^((l, S). We first generate an A^vector Y of standard normal random variates, i.e. Y~ JV(QJ.}
where 0 is a/^-vector of zeroes and I is the/rx/7 identity matrix. This is done using the function gauss (Reilly, 1995) m
boxmul. c. The lower triangular Cholesky decomposition matrix L is calculated by using the LAPACK function dpptrf
(Anderson, et al.,1992). (Recall that L is the unique lower triangular matrix such that Z = LLr.). Finally, we calculate
X — LY+{1. Since a^tf/calculates L in packed format (Anderson, et al., 1992), it was necessary to write code to do the matrix
multiplication. This was done using as few operations as possible.
Now predictions of the process at the simulated locations are obtained by letting Z(s.) = 10x(s.) = 10XJ . The integral of the
realization is then estimated by
N j=
(5.25)
as discussed in §3.3. Note that in this case A is rectangular, so \A\ is easy to calculate. The constant k— 3 .048 refers to the number
of dm/ft, since the (x,y) locations are in ft and the concentration predictions >2(s.) are in \ig/L. Since this region is only
2-dimensional, the integral estimate here is in iig/dm, and would need to be multiplied by a depth in dm to give an estimate of
the total amount of contaminant. Also, the variance of the integral estimate is calculated in the manner discussed in §3.3.
However, instead of calculating var(6,) = (k4\A\2/N2) varj it is calculated as
var(6,.) = (l/N2} varf f W,] = (l/N^ (5.26)
v'=1 )
where W/ , = k2\A\l(f^ u' and s^, is the sample variance of W/{ , ... ,^/JV • This amounts to the same thing but is easier to
program, as we can then use the integral estimate in calculating the sample variance.
This entire procedure is performed M times and the resulting integral estimates from the realizations are then used to estimate the
conditional distribution of the integral of the process, given the data. A point estimate of the integral may be found by, for
example, taking the mean of the integral estimates for the realizations. Approximate 100(1 — oc)% prediction intervals may be
found by using the o/2 and 1 — oc/2 quantiles of the simulated integral estimates.
28
-------
5.3.2 Results
As a quick assessment that our computer code is simulating the desired process, we generate, from the assumed model, a single
realization over a 20x20 grid which encompasses the region, where the jr-values range from 4900 to 5600 and the ^-values
range from 4400 to 5100. A perspective plot of this single realization is given in Figure 5.9. The realization behaves as one
might expect given the conditional mean shown in Figure 5.7. Also, the empirical variogram of this realization compares well
to the rational quadratic variogram model used to generate the data (see Figure 5.10).
Recall that the estimate of the integral for a realization, as discussed in §3.3, is given by
(5.27)
where the g(&^,j— 1,..., yVare the values of realization at locations s,,..., s^,. Noting that this is in fact a sample mean of the
N?(sy)>y= 1.—, N, we have by the central limit theorem that, for a single realization, the estimate of the integral should have a
normal distribution as N—> <*>. We would like to check the integral estimates produced from samples of a single realization to
see that this is in fact the case.
Ideally, we would like to be able to take a large number of samples from a single realization of size P, where the samples are
sufficiently sized for the asymptotics to "kick in" (say N— 500 or 1000). However, since we are simulating the realization, we
must simulate all of the values of the realization which we wish to sample together initially. If there are /"such values of the
realization, this means that we must not only calculate a PxJ1 covariance matrix, but must then calculate the Cholesky
decomposition of this matrix. As /'gets large, near singularities (i.e., singularities within machine precision) in the conditional
covariance matrix X cause the Cholesky decomposition routine to fail. The largest value of P for which the code could be
successfully run with any consistency was 2000, which does not allow for a great number of independent samples of size 500,
to say the least.
In lieu of independent samples we rely on dependent samples; i.e., samples which share some of the same realization values.
A single realization of size 2000 was simulated and subsamples of size 500 were obtained from the 2000. The subsampling was
performed with replacement. So each sample is then a sample from this realization taken at 500 independent uniform locations
over the region. From each of these samples, an integral estimate was calculated using the integral estimate (5.25). Figure 5.11
shows a histogram of the integral estimates taken from 1000 such samples. The histogram appears to be approximately normal.
Figure 5.12 contains a qq-plot of the 1,000 integral estimates. The qq-plot is very close to a straight line, although a bit off in the
tails. Again, this indicates that these integral estimates are approximately normally distributed. Also, the integral estimate
produced by the entire realization was approximately 7.548 x 109, which is also very close to the center of the histogram and
the median of the integral estimates from the samples. Thus, the distribution of integral estimates for samples from a single
realization appears to be fairly normal, as it should be.
Finally, to actually estimate the distribution of the integral, we generated samples of 500 from each of 1000 different
realizations. Figure 5.13 shows a histogram of the 1000 different integral estimates. This histogram is obviously very skewed,
to the extent that we cannot see any of the detail in the lower part of the histogram, where most of the values reside. Figure 5.14
shows a histogram of the lower 97.5% of these estimates and Figure 5.15 shows a histogram of the log,0 estimates. These allow
us to see more detail in the lower part of the histogram. Note that the units for the estimates are \ig/dm, and they would have to
be multiplied by a measurement of depth in dm to provide an estimate of total contaminant for a /#ra?-dimensional region.
Tables 5.1 and 5.2 contain summary statistics about the integral estimates for the realizations. We can see the effect of the
skewness in these summary statistics. For example, the value of the sample mean is more than twice the value of the sample
median. Furthermore, the values of the a% trimmed means rapidly approach the value of the median as a increases.
Table 5.1 Statistics of Integral Estimates for 1000 Realizations
minimum
5.604 x 10s
0.025 quantile
1.034X 109
median
4.527 x 109
0.975 quantile
6.354 xlO10
maximum
1.252 xlO12
Table 5.2 Means of Integral Estimates for 1000 Realizations
mean 5% trimmed mean 10% trimmed mean 20% trimmed mean
1.204 x 10'° 7.776 x 109 6.849 x 109 5.946 x 10'
Recall that we said in Section 5.1 that the conditional mean (i.e., the mean calculated above) was an optimal estimate of the
integral under squared error loss. Now either the median or any of the means could be used as a point estimate of the integral
of the process. In particular, in the case of absolute error loss
29
-------
L(g(X\p(x;g(X}}} = \g(X) - p(x;g(X))\
(5.28)
the median is actually optimal. However, due to the high skewness and the large difference between the median and the mean,
we would do well to exercise caution in our choice. More investigation is necessary to determine which of these is more
representative of the truth, or if some other statistic would be better, or if what is "better" depends on the particular application.
It is important to reiterate here that even though the distribution of integral estimates from samples from one realization is quite
normal, the distribution of the integral of the process is not even close to normal. However, this is not surprising as these are two
entirely different distributions. The first is the distribution of estimates for a realization with variation coming only from the
"sampling error", i.e. the error induced by estimating a value of the integral for a region with only a finite number of points. The
second involves this sampling error, along with the actual variation of the process.
As the distribution of the integral is not normal, we obviously cannot use typical normal prediction intervals. We may, however,
use quantiles of the distribution to get estimated prediction intervals for the integral of the process. For example, for a 95%
prediction interval, we may use the values in Table 5.1 to get an interval of (1.034 x 109; 6.354 x 10'°). Note, however, that
this interval still involves the aforementioned sampling error.
The variances for the integral estimates for the realizations, <£>,., were estimated as discussed in Section 5.2. The estimated
standard deviations ranged from 5.645 x 107 to 4.602 x 10". However, for the lower 97.5% of the integral estimates, the largest
estimated standard deviation was 1.708 x 10'°, nearly 30 times smaller than the overall maximum. This would lead us to
suspect that the highest integral estimates for realizations come about due to one or two very high simulated values in that
realization, thus increasing the variance of the sample from the realization tremendously. Further investigation is necessary to
determine if this is in fact the case.
The variance of the integral is estimated here as discussed in Section 5.2. That is
(5.29)
where Var\ OjxJ may be estimated by the sample variance of the integral estimates from the realizations and an unbiased
estimate of jEp^rHOjX^jJdata is obtained by taking the sample mean of the variances of the ;. calculated from the 1000
realizations as discussed above. In this example, this yields
I = 2.423 x 1021 - 2.796 x 1020 = 2.143 x 1021 (5.30)
So the estimated standard deviation of the integral is JFar[3>|x] = 4.629 X1024. It is of concern here that this value is
actually /fa/gevthan the estimated mean of the integral of 1.204 x 1010. This is caused by the large variation between the integral
estimates for the different realizations, and the high skewness of the distribution.
30
-------
Chapter 6
Phoenix, AZs Visualizatioii with a Time Component
In this chapter, we use methods of kriging and related spatial estimation (Cressie,1993) to study various concentration plumes
at a site near Phoenix, Arizona. Since we have data from several years for this site, we are able to incorporate time into our
analysis. In all cases, the plots were intended to represent a yearly average, so <7//of the data from each year was used. In cases
where there was more than one measurement for a particular location in a year, the average of these values was used. In
Section 6. 1 , we do an exploratory visualization of contaminant levels, along with accuracy assessments in the form of prediction
standard errors. In Section 6.2, we present portions of animations of contaminants together, including a method for animation of
two possibly related contaminants.
6.1 Exploratory Visualization
Figures 6.1, 6.3, 6.5, and 6.7 show "prediction" surfaces for TCE and DCE concentrations for two selected years. The
predictions were done using the log of the data, so these surfaces are on a log scale. Figure's 6.2, 6.4, 6.6, and 6.8 show the
corresponding standard error surfaces. The xan&y coordinates are the same for all of the plots in this section and the next, with
the z coordinates varying slightly. The legends for the ^coordinates are given in Figures 6.9 and 6.10 for the prediction surfaces
and error surfaces, respectively.
Figures 6.9 and 6.10 indicate the color codes for different levels of concentrations and standard errors. (Note that the legend for
the prediction surfaces in Figure 6.9 refers to the level of the original data, rather than the log of the data.) Predicted values tend
to be most accurate in the neighborhood of wells where the data were taken, which explains the downward spikes in the standard
error surfaces. (That is, each small downward spike represents an observation well location.)
One sees, when comparing TCE concentration maps (Figures 6.1 and 6.5 for the years 1991 and 1992, respectively) with the
corresponding DCE concentrations (Figures 6.3 and 6.7), that both substances are highest in the northeast portion of the region,
and the TCE plume drops off more quickly than the DCE plume. This suggests scavenging of TCE to create DCE over time as
the plume is transported from northeast to southwest. Thus, a map of the ratio of concentration of DCE to TCE should show a
large increase moving down the plume, but the sum of the two concentrations might be relatively constant, assuming the region
is relatively "closed." Of course, conclusions drawn from such displays would have to be tempered by the accuracy of the
estimated quantities.
6.2 Animation
6.2.1 Trichloroethylene and Dichloroethylene (TCE and DCE)
Since it is suspected that the TCE and DCE plumes are somehow interrelated, it is desirable to plot the two together in such a
way as to make the relationship (if it exists) more visible. With this in mind, animation of the TCE and DCE data is
accomplished by combining the estimated concentration plots for each on the same plot. This allows the two substances to be
easily animated together in time, which in turn should facilitate our attempts to see relationships between the levels of the two
substances through space and time. The animation has been performed using the measured data from the years 1985 to 1993, as
discussed in Section 2.2.
Spatial estimation, via kriging, was then performed on the logs (base 10) of the yearly data. (See Appendix B for a brief
discussion of the spatial modeling and an explanation of the problems there encountered.) For the animation, interpolations were
done in between the years (5 time slices between each pair of years), allowing a smoother progression which aids greatly in
seeing general trends in the data.
Examples of the plots used in the animation are shown in Figures 6. 1 1 , 6. 12, 6. 13, and 6.14, representing the years 1986, 1988,
1991, and 1993, respectively. The TCE surface is the upper one in all of the plots and the DCE the lower. The size of the region
is approximately 18000y?in the E-W direction and 8300y?in the N-S direction. All plots are shown on the Iog10 scale; e.g., the
purple regions which are labeled as 1 to 2 on the scale are regions where the estimated contaminant level is between 10 and
These plots all confirm our assumption that the values of the contaminants are generally highest in the Northeast region, where
the source is located. These plots help to point out some general features in the data, as well as introducing some new questions.
In particular, Figure 6.13 would seem to indicate that there is a ridge of high values of contaminant along a line from
East-Northeast to West-Southwest, along with some low values next to this region of high values. Investigation into the site
properties yielded the information that the ground-water flow is generally from E-NE to W-SW, so it would appear that the
contaminant levels are highest along a direct line from the source in the direction of ground- water flow, perhaps indicating very
slow dispersion of the contaminants through other means. The 1986 plot, Figure 6.1 1, is representative of the earlier years in that
data were only taken in an area relatively close to the source. In later years, gradually points were added farther to the west of
the source. (There were none added to the east, presumably since it was unexpected that the contaminant would spread greatly
in a direction counter to the direction of ground-water flow.) By 1993, both contaminants appear to have dissipated greatly, and
many of the mild variations seen are likely due to measurement error. Figure 6.12, from 1988, shows lower values of
31
-------
Figure 6.1 TCE prediction surface for 1991. Orientation: depth-axis: 892,600 to 896,800; width-axis: 478,000 to 484,000;
vertical-axis: 0 to 2.25.
Figure 6.2 TCE standard errors of prediction for 1991.
32
-------
Figure 63 DCE prediction surface for 1991.
Figure 6.4 DCE standard errors of prediction for 1991.
33
-------
Figure 6.5 TCE prediction surface for 1992.
Figure 6.6 TCE standard errors of prediction for 1992.
34
-------
Figure 6.7 DCE prediction surface for 1992.
Figure 6.8 DCE standard errors of prediction for 1992.
35
-------
[100, 1000)
[10, 100)
[1, 10)
[0,1)
Figure 6.9 Legend for prediction surfaces (measurements in mg/E).
[1-75,2)
[1.5, 1.75)
[1.25, 1.5)
[0, 1.25)
Figure 6.10 Legend for standard errors.
36
-------
Figure 6.11 TCE and DCE surfaces for 1986.
0
Figure 6.12 TCE and DCE surfaces for 1988.
37
-------
this occurred. It has been supposed that perhaps an excessive amount of rainfall in this year would have diluted the
concentrations of contaminant in the ground water, but it should be noted that this is definitely only suspicion and further
investigation is necessary. In any event, however, animation of the spatial estimates through time allowed us to easily pick up on
this seeming aberration.
6.2.2 Sulfate Ions
Measurements of sulfate ions
) are of interest due to a supposed connection between levels of SO^" , TCE, and DCE.
That is, anaerobic bacteria which consume SO^" also consume TCE, converting it into DCE. It is believed that regions of low
sulfate, dubbed "sulfate holes," indicate the presence of such bacteria. If this is true, we also would expect to see TCE drop in
these regions. There are much fewer sulfate data than there are TCE and DCE data, to the extent that there are insufficient data
in several of the years to do a reasonable surface estimate. For this reason, SO^" surfaces were produced only for the years 1 985
through 1989, inclusive, and 1991. As there was some concern that the large flat sections of the surfaces in the portions of the
region with no data might be somewhat misleading, a new visualization technique was tried with the SO^" data. For each year,
the surface was only plotted in the area where there were data, with a bounding box to indicate the region and keep all years on
the same scale. The success of this approach is perhaps mixed. It does indeed make it very clear in what regions we do not have
any good estimates due to lack of data. However, it also makes it nearly impossible to interpolate between years to produce a
smooth animation, and in fact increases the "jumpy" effect seen when viewing the surfaces in chronological order. (This is not
such a problem with the sulfate data, as the time gap makes them not entirely suitable for animation, anyway.) Again, the plots
are all on a log]0 scale with the same color contours as for TCE and DCE.
Figures 6.15, 6.16, and 6.17 contain examples of these SO^" surfaces. These surfaces have several interesting features. In all
three, we see a small peak in sulfate levels near the source of TCE and DCE contamination. Also, in all three, we see one or more
"sulfate holes" near to this peak. This indicates that perhaps TCE is being converted to DCE near to the source, which would
lead to TCE values dropping off more rapidly than DCE as distance from the source increases. Efforts to see if such a
relationship exists will be discussed further in Section 6.3. Figure 6.16 reveals another interesting and somewhat odd feature of
0
Figure 6.13 TCE and DCE surfaces for 1991.
38
-------
0
Figure 6.14 TCE and DCE surfaces for 1993.
2—
Figure 6.15 SOl" surface for 1986.
39
-------
Figure 6.16
' surface for 1988.
Figure 6.17 SO$ surface for 1991.
40
-------
the sulfate data. The overall sulfate levels for the years 1987, 1988, and 1989 over the region are significantly higher than for
the other years studied. We have not yet been able to identify a reason for this, but it certainly seems to warrant further
investigation.
6.3 Further Analytical Efforts
As mentioned previously, the region for which there were data available increased as time progressed. Specifically, in the
earlier years (e.g., 1985), there were data points only relatively close to the source. Later, more data points were added to the
south and especially to the west. This can create some problems with analysis, so it was decided following a conversation with
Dr. Joe Hughes that we should try to do local analyses using some smaller region around the source. It is hoped that this will
help us to better understand the chemical and transport processes at work. We expect that
• TCE should decrease in regions of high sulfate due to the likely presence of sulfate- consuming bacteria in these regions,
and that
• TCE and SO^" should decrease quicker with distance from the source than DCE.
Two different types of local analyses were attempted. The first involves simply using only a small rectangular region around
the source; e.g., the Northeast comer of the total region. The other involves using a relatively narrow region from the source
and extending in the direction of flow. For this second method, we then want to do spatial estimates of contaminant (i.e., log,0
of TCE, DCE, and SO^" ) vs. the following:
1. distance from the source,
2. time since Jan. 1, 1985 and distance from the source, and
3. time since Jan. 1, 1985 and distance from the source at time 0.
As the region chosen in this case is narrow, the distance from the source is approximately equal to the distance from the source
along the line of flow. This method seems reasonable because the primary method of transport of contaminant in this system is
ground-water flow, with dispersion being decidedly less. So if the speed of ground-water flow is sft/day, then the contaminant
present a distance jVfrom the source in the direction of flow on day /should have been at the source on day 0. This is precisely
the motivation for the third case listed above, for which we calculate d'- d- st, the estimated distance from the source at time 0.
Unfortunately, there is one problem with this latter method which has not yet been resolved: it is not entirely clear where the
source of contaminant is. The general vicinity is certainly known, but to pinpoint an "exact" location allowing us to place a
narrow strip about the source has proved to be a difficult problem. It is possible that the source is in fact a large area and would
not be well approximated by a point source. From recent investigations, what appears likely is that there are at least 2 point
sources approximately 500 - lOOOy?apart. (We have hypothesized this after noticing that in the general vicinity of the source,
there are two clusters of measurement locations which both contain very high levels of TCE. As it is common to place large
numbers of wells near a known source, the presence of two sources known to previous investigators seems a logical hypothesis.)
This presents some unforeseen problems and is what has led us to consider the first method mentioned; i.e., doing more
standard estimation on a small region about the source.
41
-------
-------
Chapter 7
Summary and Conclusions
Site characterization and estimation of contaminant plumes is a complex problem which requires the compilation by the
environmental researcher of many sources of information. Observations on the contaminant level over the region are expensive
and sometimes difficult to obtain. In this research effort, we suggested several methods of examining such valuable data to
further the researcher's understanding of the environmental problem under study.
Based on observational data, we explored analytical methods for estimating the level and extent of the contaminant plume.
Nonparametric regression methods proved useful for quick summaries of the contaminant plume, whereas the more difficult to
implement geostatistical methods were required for quantitative measures of the contaminant plume, such as the total amount of
contaminant present.
In addition to exploring the analytical issues associated with estimation of the contaminant plume and functional of this plume,
we investigated how best to display this information through visualization methods. Two and three-dimensional perspective
plots with color contours proved useful in our investigation. To associate the error in the estimated plume with the estimated
level of the plume, we suggest associating the height of the perspective plot with the estimated level of contaminant and the color
contours with the estimated amount of observed error in the estimate.
Our investigations also found that animation of the estimated level of contaminants or estimated errors was a useful exploratory
tool. For the Eglin data, surface estimates produced with all but one point are animated alternately with the surface estimate
using all the data points. This allows us to readily see the effect each data point has on the surface estimate. For points whose
absence produces a large change in the surface estimate, it may be desirable to take additional samples near this point to help
stabilize the estimate in this area; estimates of the prediction errors at these points are also useful for this reason. Animations
through time, with smoothing, were used for the Arizona data allowing quick identification of atypical behavior in time. Also for
the Arizona site, we investigated methods of simultaneously animating two related substances. Simultaneous animations of TCE
and DCE helped identify the relationship between these two substances. Furthermore, we examined the issue of a growing
region or plume. Both of these issue are the focus of further research.
We proposed in Section 5 a method for estimating the integral of a random process in the case where the process is lognormal
by modeling the process through geostatistical methods and simulating the process conditional on the data. This is useful for
estimating the total amount of contaminant present in a region. When implemented on the Eglin ground water observations for
BTEX, this method produced reasonable point estimates but large confidence intervals. Large confidence intervals are to be
expected from such a small number of observations, however we are hopeful that further research into improved statistical
methods can yield tighter confidence intervals for small sample sizes.
Future Research
This work has surfaced a number of topics which would be appropriate for future research. It is still an open question how to best
view two possibly related substances to see how they are related. For the TCE and DCE data from Arizona, it was thought that
examining the ratio of the two substances might be useful, but this did not seem to reveal very much. An examination of
functions of two such substances so they may be viewed as a single surface seems like a promising idea, however.
For the sulfate data from Arizona, we attempted to deal with the issue of an increasing design region over time, due to more
information, and interest, on the part of those taking observations. Attempts to plot various portions of the estimated surface,
dependent on the region, seem to be of dubious value. The idea of estimating a small region near the source, or perhaps
estimating along the line of ground-water flow from the source, as discussed in Section 6.3, is a promising idea. The lack of a
well-defined source in this instance made such an examination difficult, but such a method could certainly still be examined for
this and other data.
The method for estimating the integral discussed in Section 5 is promising, but reductions of variance and better prediction
intervals are areas which need to be addressed. Specifically, if one looks at the values obtained for the integral estimates and
their standard deviations, one notices that in fact the standard deviations are extremefy\n$\. In particular, recall that if we use
the mean of the integral estimates for the realizations as a point estimate for the integral of the process , we get an estimate of
total contaminant of 1.204 x 10'°. However, the estimated variance of this value is then the mean of the variances for the
realizations, giving a variance of 2.796 x 1020 , and a standard deviation of 1.672 x 1010 . That is, the estimated standard
deviation of total concentration is actually higher than the estimated concentration. It is possible that by sampling locations
using some distribution other than a uniform, i.e. importance sampling (Rubinstein, 1981), may yield integral estimates with a
smaller variance. For example, we may wish to sample from a smaller, possibly non-rectangular region where the data are more
dense. Or we may wish to sample with higher probability along the direction of geometric anisotropy than in the perpendicular
direction. At any rate, it is desirable to investigate ways of reducing the variance of the integral estimates, and this is a topic of
current research for some of the authors of this report.
For the prediction intervals, we used what are referred to as equal /^//intervals. That is, the interval is a two-tailed interval with
equal probability in either of the tails. This is not always the best type of interval to use, and particularly may not be in the case
43
-------
of such an asymmetric distribution. It would be worthwhile to investigate other types of intervals, particularly those known as
highest'posterior density'(HPD) regions (Casella andBerger, 1990). In this case, the 1- a interval is chosen so as to be as short
as possible. Specifically, if the posterior density is denoted by 7t, the 1- a HPD region is given by {jc: K(X) > c} where c is such
that 1 — OC = J. ^^ n(x) ax Also, the prediction intervals discussed here are actually too large, as they contain additional
variation due to Monte Carlo sampling error. To get better prediction intervals, this factor needs to be corrected for.
Finally, it was mentioned at the beginning of Section 5 that there are actually severa/nonlinear functional of random processes
which are of interest to estimate. The integral of the process is the only one of these which we have investigated in detail to this
point. Other functions which are of interest are:
• the maximum concentration attained within a region,
• the location where this maximum concentration occurs, and
• the region for which the concentration exceeds some set value.
These other three are quantities which are of interest for various types of environmental contamination and these warrant further
investigation. Further, in ozone modeling, it is common to use a square root transform rather than a log transform as in Carroll
et al.,1997. Thus it is also of interest to estimate the total contaminant in the case where the process is transformed by a square
root rather than a log.
44
-------
Appendix A
Cross-Validation
In all cases, it is assumed that bandwidths in the x andj/ directions should be the same, i.e. h - (hfh^ where hj is the bandwidth
in the longitude and latitude directions and k2 is the bandwidth in the vertical direction. We will use h to denote either h{ or (hf
h2 ) depending on whether the estimate is in 2-D or 3-D, respectively. The bandwidths were chosen by minimizing over h
i=\
where // is the number of data points, u. is the /* observation, and u! h is the nonparametric regression estimate, based on
bandwidth k, of the value at the /* data point obtained when this point is removed. (Note: «/>A = /(^/J from 4.1.)
A.1 Two-dimensional Data
For the ground- water data, a single cross-validation was performed on bandwidths varying from 100 to 250/?in increments of
ten. As seen in Figure A.I, the bandwidth selected here is 120y?. See 4.1 for a plot of the estimated plume.
Similarly, a cross-validation was performed for twordimensional soil data from Eglin at approximate depths of 7.0 and 7.6/f
below the water table. In both of these cases, the minimization of SSHJi) was performed by a grid search from 100 to 250y?in
increments often. The bandwidth chosen for the depth of 7.0/?is 170/?, while 230/? was chosen for the depth of 7.6/f. See 4.1
for a plot of the estimated plume at a depth of 7.07?.
In addition to an estimate of the contaminant plume, we also visualize a smoothed estimate of the absolute error of this estimate
as each data point is removed. A cross-validation of the absolute errors for the ground-water data resulted in a bandwidth of
130y?, which is close to the 120;? found for the plume estimate. The plots for visualizing the error are given in Section 4.1.
A.2 Three-dimensional Data
For the three-dimensional soil data, we needed to perform cross-validation in both the x and y directions and the z direction.
(Here the x direction is longitude, the ^direction is latitude, and the z direction is distance above or below the water table.) First
we performed cross-validation in the x and y directions, allowing the bandwidth in the z direction to vary along with the
bandwidth in the x andy directions. This yielded an x and ^bandwidth of 24ff. Then we examined the SSEJi) using x and^
bandwidths varying from 20 to 4Q/? (which includes the minimum of 24/5), and six different z bandwidths ranging from
approximately 1 .5 to approximately 3 . This caused us to choose an xy bandwidth of approximately 24/? and a z bandwidth of
approximately 1.8/?.
150
250
Figure A.1 Bandwidth selection for ground-water data.
45
-------
Appendix B
Discussion of Spatial Estimation for Arizona
There were a number of difficulties encountered when doing the spatial estimates of TCE, DCE, and SO^" for the Arizona site.
The biggest problem was finding suitable variogram models. It was decided that it would be most reasonable to find only one
variogram model for each substance, which would be used for all years of data. Many of the classical empirical variograms
with standard default binwidths, etc., produced totally unreasonable variograms (e.g., variograms which were flat or indicated
stronger correlation between points which were a long ways apart than for points which were close together). For example,
consider the classical empirical variogram for 1990 TCE, shown in Figure B.I. This plot is highly variable and shows a general
decreasing trend, whereas a variogram should be generally increasing. Cressie's robust variogram estimator did not usually
solve these problems. It was conjectured that the high variability could largely be due to low numbers of pairs of data points at
many of the higher distances. To combat this, we binned the data point pairs into groups with equal numbers of pairs, rather than
equal width bins as hi the classical estimator. We tried taking both means and medians within these groups, analogous to the
classical and robust estimators. A plot of the mean case for the 1990 TCE data is shown in Figure B.2. Apart from the last bin,
where the data point pairs used are so far apart hi distance as to be suspicious anyway, this empirical variogram estimator looks
much better than the classical one. Specifically, it is much less variable and has a decidedly increasing trend. Using this type of
empirical variogram estimator, then, spherical variogram models were fit first for each year (and each substance). Then, from
these, a variogram model for each substance was chosen which would be "best" hi most years, and hopefully reasonable in all.
This is intended to allow us to get satisfactory variogram estimates even for those years which do not contain much data and to
provide a more unified approach.
CO
en
co -
CM -
O -
1000
2000 3000
distance
4000
5000
Figure B.I Classical empirical variogram for 1990 TCE.
46
-------
co
CO
O)
-------
-------
Bibliography
E. Anderson, Z. Bai, C. Bischof, J. Demmel, J. Dongarra, J. Du Croz, A. Greenbaum, S. Hammarling, A. McKenney, S.
Ostrouchov, and D. Sorensen. LAPACK Users' Guide. SIAM, Philadelphia, 1992.
R.K. Boeckenhauer. Estimating nonlinear functional of a random field. PhD. dissertation. Department of Science, Rice
University, 2000.
R.K. Boeckenhauer, K.B. Ensor, D.W. Scott, and P.B. Bedient Visualization of ground water contamination data. Proceedings
of the 27th Interface Between Statistics and Computing Science, Editors Michael M. Meyer and James L. Rosenberger.
Interface Foundation of North America, Fairfax, VA, pp 76-84, 1996.
R.J. Carroll, R. Chen, E.I. George, T.H. Li, H.J. Newton, H. Schmiediche, and N. Wang. Ozone exposure and population
density in Harris County, Texas. Journal of the American Statistical Association, 92:392-415, 1997.
G. Casella and R.L. Berger. Statistical Inference. Duxbury Press, Belmont, CA, 1990.
D.D. Cox, L.H. Cox, and K.B. Ensor. Spatial Sampling for the Environment. Environmental and Ecological Statistics, Vol. 4,
pp 219-233, 1996.
N.A.C. Cressie. Statistics for Spatial Data. John Wiley & Sons, New York, revised edition, 1993.
C.V. Deutsch and A.G. Journel. GSLIB: Geostatistical Software Library and User's Guide. Oxford University Press, 1993.
E.J. Englund and N. Heravi. Power curves for spatial sampling. Draft: 4/95.
J.M. Hammersley and D.C. Handscomb. Monte Carlo Methods. Methuen & Co. Ltd., London, 1964.
M.E. Johnson. Multivariate Statistical Simulation. John Wiley & Sons, New York, 1987.
E.L. Lehmann. Theory of Point Estimation. John Wiley & Sons, New York, 1983.
K.V. Mardia, J.T. Kent, and J.M. Bibby. Multivariate Analysis. Academic Press, Inc., San Diego, 1979.
MathSoft Inc. S-plus, 1988, 1995. Version 3.3 Release 1 for Sun SPARC, SunOS 4.1.x.
P.L. Reilly. Pmmlcg. IsoQuantic Technologies, LLC, 1994, 1995.
R.Y. Rubinstein. Simulation and the Monte Carlo Method. John Wiley & Sons, New York, 1981.
D.W. Scott. Multivariate Density Estimation. John Wiley & Sons, New York, 1992.
G.W. Stewart. Introduction to Matrix Computations. Academic Press, San Diego, 1973.
H.G. Sweed, P.B. Bedient, and S.R. Hutchins. Surface application system for in situ ground-water bioremediation: Site
characterization and modeling. GroundWater, 34(2):211-222, March 1996.
49
-------
-------
-------
United States
Environmental Protection Agency
National Risk Management
Research Laboratory, G-72
Cincinnati, OH 45268
Please make all necessary changes on the below label,
detach or copy, and return to the address in the upper
left-hand corner.
If you do not wish to receive these reports CHECK HEREtH ;
detach, or copy this cover, and return to the address in the
upper left-hand corner.
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use
S300
EPA/600/R-00/034
-------
United States Prevention, Pesticides EPA738-R-00-018
Environmental Protection And Toxic Substances October 2000
Agency (7508C)
** EPA 'nterim Reregistration
-------
-------
United States
Environmental Protection
Agency
Prevention, Pesticides
and Toxic Substances
(7508C)
EPA738-F-00-16
October 2000
Propetamphos Facts
EPA has assessed the risks of propetamphos and reached an Interim Reregistration Eligibility
Decision (IRED) for this organophosphate (OP) pesticide. Provided that the risk mitigation measures
outlined in this document are adopted, propetamphos fits into its own "risk cup"; that is, its aggregate
risks are within acceptable levels. Propetamphos is also eligible for reregistration, pending a full
reassessment of the cumulative risk from all OPs.
Propetamphos is an insecticide used indoors
for the control of insects, such as ants, cockroaches,
fleas and termites. Propetamphos residues in food
and drinking water do not pose risk concerns.
Additionally, risks are low to workers who mix, load,
and apply propetamphos at commercial and
residential use sites. There are also no environmental
risk concerns. However, there are post-application
risk concerns for adults, and especially children
entering areas treated with propetamphos. With
mitigation canceling all residential use, propetamphos
fits into its own "risk cup". With other mitigation
measures, propetamphos' worker risks also will be
below levels of concern for reregistration.
EPA is reviewing the OP pesticides to
determine whether they meet current health and safety
standards. OPs need decisions about their eligibility
for reregistration under FIFRA. Additional OPs with
residues in food, drinking water, and other non-
occupational exposures also must be reassessed to
make sure they meet the new Food Quality
Protection Act (FQPA) safety standard.
EPA's next step under the Food Quality Protection Act (FQPA) safety standard is to complete
a cumulative risk assessment and risk management decision encompassing all the OP pesticides, which
share a common mechanism of toxicity. The interim decision on propetamphos cannot be considered
final until this cumulative assessment is complete. Further risk mitigation may be necessary at that time.
The OP Pilot Public Participation Process
The organophosphates are a group of
related pesticides that affect the functioning of the
nervous system. They are among EPA's highest
priority for review under the Food Quality Protection
Act.
EPA is encouraging the public to
participate in the review of the OP pesticides.
Through a six-phased pilot public participation
process, the Agency is releasing for review and
comment its preliminary and revised scientific risk
assessments for individual OPs. (Please contact
the OP Docket, telephone 703-305-5805, or see
EPA's web site, www.epa.gov/pesticides/op .)
EPA is exchanging information with
stakeholders and the public about the OPs, their
uses, and risks through Technical Briefings,
stakeholder meetings, and other fora. USDA is
coordinating input from growers and other OP
pesticide users.
Based on current information from
interested stakeholders and the public, EPA is
making interim risk management decisions for
individual OP pesticides, and will make final
decisions through a cumulative OP assessment.
-------
The propetamphos BRED was made through the OP pilot public participation process, which
increases transparency and maximizes stakeholder involvement in EPA's development of risk
assessments and risk management decisions. EPA worked extensively with affected parties to reach
the decisions presented in this IRED document, which concludes the OP pilot process for
propetamphos.
Uses
Propetamphos is an OP insecticide used indoors for the control of insects, primarily ants,
cockroaches, fleas, and termites. Propetamphos may be applied at indoor residential, medical,
commercial, and industrial buildings and equipment, such as homes, apartments, stores, schools,
hospitals, offices and factories. It may also be used in food service establishments where there
is no contact with food, and where no processing, packing, or warehousing of food occurs.
Total annual usage is low, and estimated at 90,000 pounds active ingredient. The typical rate of
dilution varies from 0.5% to 1.0% active ingredient solution. Propetamphos is applied as a
water dilution through a compressed air sprayer, often with a low pressure hand wand.
Health Effects
Propetamphos can cause cholinesterase inhibition in humans; that is, it can overstimulate the
nervous system causing nausea, dizziness, confusion, and at very high exposures (e.g., accidents or
major spills), respiratory paralysis and death.
Risks
Dietary exposures from food are not of concern for the entire U.S. population, including infants
and children, provided food is removed or covered prior to an area being treated. Because
propetamphos is only used indoors, exposure from drinking water sources is not expected.
Risks are low, but still of concern for workers who mix, load, and apply propetamphos at
commercial and residential use sites.
Risks are of concern for adults, and especially children, from combined dermal, inhalation, and
(for children only) oral routes of post-application exposure from re-entering areas treated with
propetamphos.
Because propetamphos is used indoors, exposure to the environment is not expected, and
therefore, ecological risks are not of concern to the Agency.
In order to support an ERED for propetamphos, the following risk mitigation measures are
necessary:
-------
To mitigate dietary (food) risks:
for use in food service establishments, all food must be either covered or removed prior
to the area being treated.
To mitigate worker risks:
reduce the maximum rate of dilution from 1.0% to 0.5% active ingredient solution;
• applicators must wear personal protective equipment consisting of a long-sleeve shirt,
long pants, shoes and socks, and chemical-resistant gloves; and
• only protected handlers may be in the area during applications.
To mitigate non-occupational risks to persons re-entering treated areas (post-application risks):
cancel all residential uses;
prohibit use in structures children and the elderly occupy, such as or including homes,
schools, day-cares, hospitals, nursing homes (except for areas of food service when
food is covered or removed prior to treatment);
• cancel all spot, broadcast, and termiticide treatment; and
• restrict the method of application to crevice treatment only, as defined in OPPTS
860.1460 Food Handling.
Next Steps
Numerous opportunities for public comment were offered as this decision was being
developed. The Propetamphos IRED, therefore, is issued in final (see www.epa.gov/REDs/ or
www.epa.gov/pesticides/op ) without a formal public comment period. The docket remains
open, however, and any comments submitted in the future will be placed in this public docket.
To effect risk mitigation as quickly as possible, time frames for making the changes described
in the Propetamphos BRED are shorter than those in a usual RED. All labels need to be
amended to includefthe above mitigation and submitted to the Agency within 90 days after
issuance of this IRED.
For propetamphos, tolerances for residues in food commodities will remain in effect and
unchanged until a full reassessment of the cumulative risk assessment for all OP pesticides is
completed. Upon completion of the cumulative risk assessment, EPA will issue its final
tolerance reassessment decision for propetamphos and may request further risk mitigation
measures. For all OPs, raising and/or establishing tolerances will be considered once a
cumulative assessment is completed.
-------
-------
I
\
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
OFFICE OF
PREVENTION, PESTICIDES
AND TOXIC SUBSTANCES
nn v ' •'
u u s ,.>,_ s
CERTIFIED MAIL
Dear Registrant:
This is to inform you that the Environmental Protection Agency (hereafter referred to as EPA or the
Agency) has completed its review of the available data and public comments received related to the
preliminary and revised risk assessments for the organophosphate (OP) pesticide propetamphos. The
public comment period on the revised risk assessment phase of the reregistration process is closed.
Based on comments received during the public comment period and additional data received from the
registrant, the Agency revised the human health and environmental effects risk assessments and made
them available to the public on December 1, 1999. This action brought an end to Phase 4 of the OP
Public Participation Pilot Process developed by the Tolerance Reassessment Advisory Committee, and
initiated Phase 5 of that process. During Phase 5, all interested parties were invited to participate and
provide comments and suggestions on ways the Agency might mitigate the estimated risks presented in
the revised risk assessments. This public participation and comment period commenced on December
1, 1999, and closed on February 1, 2000.
Based on its review, EPA has identified risk mitigation measures that the Agency believes are necessary
to address the human health risks associated with the current use of propetamphos. The EPA is now
publishing its interim decision on the reregistration eligibility of and risk management decision for the
current uses of propetamphos and its associated human health and environmental risks. The
reregistration eligibility and tolerance reassessment decisions for propetamphos will be finalized once
the cumulative assessment for all of the OP pesticides is complete: The enclosed "Interim
Reregistration Eligibility Decision for Propetamphos," which was approved September 29,2000,
contains the Agency's decision on the individual chemical propetamphos.
A Notice of Availability for this Interim Reregistration Eligibility Decision (IRED) for propetamphos is
being published in the Federal Register. To obtain a copy of this IRED document, please contact the
OPP Public Regulatory Docket (7502C), US EPA, Aerial Rios Building, 1200 Pennsylvania Avenue
NW, Washington, DC 20460, telephone (703) 305-5805. Electronic copies of the IRED and all
supporting documents are available on the Internet. See http:www.epa.gov/pesticides/op.
The IRED is based on the updated technical information found in the propetamphos public docket. The
docket not only includes background information and comments on the Agency's preliminary risk
assessments, it also now includes the Agency's revised risk assessments: Updated Revised
-------
Preliminary Risk Assessment: Propetamphos, June 7, 1999; Updated Occupational and
Residential Dermal Exposure Assessment addendum, September 27, 2000; EFED Integrated
Science Chapter for Propetamphos, December 2, 1997; soft. Propetamphos Errata Sheet For
EFED Chapter, January 12,1999; and a document summarizing the Agency's Response to
Comments. The Response to Comments document addresses corrections to the preliminary risk
assessments submitted by chemical registrants, as well as responds to comments submitted by the
general public and stakeholders during the comment period on the risk assessment. The docket will
also include comments on the revised risk assessment, and any risk mitigation proposals submitted
during Phase 5. For propetamphos, a proposal was submitted by Wellmark International, the technical
registrant. Mitigation suggestions were also submitted by the National Pest Management Association
(NPMA).
This document and the process used to develop it are the result of a pilot process to facilitate greater
public involvement and participation in the reregistration and/or tolerance reassessment decisions for
OP pesticides. As part of the Agency's effort to involve the public in the implementation of the Food
Quality Protection Act of 1996 (FQPA), the Agency is undertaking a special effort to maintain open
public dockets on the OP pesticides and to engage the public in the reregistration and tolerance
reassessment processes for these chemicals. This open process follows the guidance developed by the
Tolerance Reassessment Advisory Committee (TRAC), a large multi-stakeholder body that advised the
Agency on implementing the new provisions of the FQPA. The reregistration and tolerance
reassessment reviews for the OP pesticides are following this new process.
Please note that the propetamphos risk assessment and the attached IRED concern only this particular
OP pesticide. This IRED presents the Agency's conclusions on the dietary risks posed by exposure to
propetamphos alone. The Agency has also concluded its assessment of the ecological and worker
risks associated with the use of propetamphos. Because the FQPA directs the Agency to consider
available information on the basis of cumulative risk from substances sharing a common mechanism of
toxicity, such as the toxicity expressed by the OPs through a common biochemical interaction with
cholinesterase enzyme, the Agency will evaluate the cumulative risk posed by the entire OP class of
chemicals after completing the risk assessments for the individual OPs. The Agency is working towards
completion of a methodology to assess cumulative risk and the individual risk assessments for each OP
are likely to be necessary elements of any cumulative assessment. The Agency has decided to move
forward with individual assessments and to identify mitigation measures necessary to address those
human health and environmental risks associated with the current uses of propetamphos. The Agency
will issue the final tolerance reassessment decision for propetamphos and finalize decisions on the
reregistration eligibility once the cumulative assessment for all of the OPs is complete.
This document contains a generic and a product-specific Data Call-In(s) (DCI) that outline(s) further
data requirements for this chemical. Note that a complete DCI, with all pertinent instructions, is being
sent to registrants under separate cover. Additionally, for product-specific DCIs, the first set of
required responses to is due 90 days from the receipt of the DCI letter. The second set of required
responses is due eight months from the date of the DCI.
In this BRED, the Agency has determined that propetamphos will be eligible for reregistration provided
-------
current uses of propetamphos may pose unreasonable adverse effects to human health and the
environment, and that such effects can be mitigated with the risk mitigation measures identified in
this IRED. Accordingly, the Agency recommends that registrants implement these risk mitigation
measures immediately. Section IV of this IRED describes labeling amendments for end-use
products and data requirements necessary to implement these mitigation measures. Instructions
for registrants on submitting revised labeling and the tune frame established to do so can be found
in Section V of this document.
Should a registrant fail to implement any of the risk mitigation measures outlined in this
document, the Agency will continue to have concerns about the risks posed by propetamphos.
Where the Agency has identified any unreasonable adverse effect to human health and the
environment, the Agency may at any time initiate appropriate regulatory action. At that tune, any
affected person(s) may challenge the Agency's action.
If you have questions on this document or the label changes necessary for reregistration, please
contact the Special Review and Reregistration Division Chemical Review Manager, Gary Mullins
at (703) 308-8044. For questions about product reregistration and/or the Product DCI that
accompanies this document, please contact Karen Jones at (703) 308-8047.
Lois A. Rossi, Director
Special Review and
Reregistration Division
Attachment
-------
-------
INTERIM REREGISTRATION
ELIGIBILITY DECISION
for
PROPETAMPHOS
Case No. 2550
-------
-------
TABLE OF CONTENTS
Executive Summary ii
I. Introduction 1
n. Chemical Overview 3
A. Regulatory History 3
B. Chemical Identification 3
C. Use Profile 4
D. Estimated Usage of Pesticide 5
HI. Summary of Propetamphos Risk Assessment . 7
A. Human Health Risk Assessment 7
1. Dietary Risk from Food 7
a. Toxicity 7
b. FQPA Safety Factor 8
c. Population Adjusted Dose (PAD) 9
d. Hazard Determination 9
e. Cancer Determination 9
f. Acute Dietary (Food) Risk 10
g. Chronic Dietary (Food) Risk 10
2. Dietary Risk from Drinking Water 11
3. Occupational and Residential Risk 11
a. Toxicity 11
b. Hazard Determination 12
c. Exposure 13
d. Occupational and Residential Risk Summary 15
4. Aggregate Risk 18
5. Human Incident Reports 19
B. Environmental Risk Assessment 20
IV. Interim Risk Management and Reregistration Decision 21
A. Determination of Interim Reregistration Eligibility 21
B. Summary of Phase 5 Comments and Responses 22
C. FQPA Assessment 23
1. "Risk Cup" Determination 23
2. Tolerance Summary 23
3. Endocrine Disrupter Effects 24
-------
D. Regulatory Rationale 25
1. Human Health Mitigation Measures 25
a. Dietary (Food and Drinking Water) Risk 25
b. Occupational Risk 25
c. Residential (Post-Application) Risk '..... 26
2. Environmental Risk Mitigation Measures 27
E. Label Amendments 27
V. What Registrants Need to Do 29
A. Manufacturing-Use Products 29
1. Additional Generic Data Requirements 29
2. Labeling for Manufacturing-Use Products 30
B. End-Use Products 30
1. Product-Specific Data Requirements 30
2. Labeling for End-Use Product 31
C. Existing Stocks 31
D. Labeling Changes Summary Table 32
VI. Related Documents and How to Access Them 37
A: Use Patterns Eligible For Reregistration 41
B: Table Of Generic Data Requirements And Studies Used To Make The Interim
Reregistration Decision 43
C: Technical Support Documents 47
D: Citations Considered To Be Part Of The Database Supporting the Interim
Reregistration Eligibility Decision (Bibliography) 49
E: Generic Data Call-In 59
F: Product Specific Data Call-In 63
G: List of Registrants Sentthis Data Call-In 71
H: List of Related Documents and Electronically Available Forms 73
-------
Propetamphos Reregistration Team
Office of Pesticide Programs:
Health Effects Risk Assessment
Steven A. Knizner
Julianna Cruz
Jerome Blondell
Environmental Fate Risk Assessment
William Evans
Sid Abel
Pat Jennings
James Goodyear
Use and Usage Analysis
Virginia W. Dietrich
Steven Nako
EvanVillianatos
Registration Support
Marilyn Mautz
Risk Management
GaryMullins
Susan Jennings
Michael Goodis
-------
-------
Glossary of Terms and Abbreviations
ai Active Ingredient
aPAD Acute Population Adjusted Dose
AR Anticipated Residue
ARI Aggregate Risk Index
C/CPAS Certified/Commercial Pesticide Applicator Survey
CFR Code of Federal Regulations
ChEI Cholinesterase inhibition
cPAD Chronic Population Adjusted Dose
CSF Confidential Statement of Formula
DCI Data Call-In
DEEM Dietary Exposure Evaluation Model
EC Emulsifiable Concentrate Formulation
EDSP Endocrine Disrupter Screening Program
EDSTAC Endocrine Disrupter Screening and Testing Advisory Committee
EPA Environmental Protection Agency
EP End-Use Product
ExpoSAC Exposure Science Advisory Committee
FIFRA Federal Insecticide, Fungicide, and Rodenticide Act
FFDCA Federal Food, Drug, and Cosmetic Act
FHEs Food Handling Establishments
FSEs Food Service Establishments
FQPA Food Quality Protection Act
FR Federal Register
GLN Guideline Number
GC Gas Chromatography
GC/MSD Gas Chromatography/Mass Spectrometry Detection
HED Health Effects Division
IDS The OPP Incident Data System
IPM Integrated Pest Management
IRED Interim Reregistration Eligibility Decision
LC50 Median Lethal Concentration. A statistically derived concentration of a substance that
can be expected to cause death in 50% of test animals. It is usually expressed as the
weight of substance per weight or volume of water, air or feed, e.g., mg/1, mg/kg or ppm.
Median Lethal Dose. A statistically derived single dose that can be expected to cause
death in 50% of the test animals when administered by the route indicated (oral, dermal,
inhalation). It is expressed as a weight of substance per unit weight of animal, e.g.,
mg/kg.
LOD Limit of Detection
LOAEL Lowest Observed Adverse Effect Level
MCCEM Multi-Chamber Concentration and Exposure Model
mg/kg/day Milligram Per Kilogram Per Day
MOE Margin of Exposure
MRID Master Record Identification (number). EPA's system of recording and tracking studies
submitted.
MUP Manufacturing-Use Product
LD
50
-------
Glossary of Terms and Abbreviations
NA Not Applicable
NHGPUS National Home and Garden Pesticide Use Survey
NOAEL No Observed Adverse Effect Level
NPMA National Pest Management Association
NPTN National Pesticide Telecommunications Network
OPIDN Organophosphate Induced Delayed Neurotoxicity
OP Organophosphate
OPP EPA Office of Pesticide Programs
OPPTS EPA Office of Prevention, Pesticides and Toxic Substances
PAD Population Adjusted Dose
PAM Pesticide Analytical Manuel
PCC Pest Control Centers
PCO Pest Control Operator
PHED Pesticide Handler's Exposure Data
PPE Personal Protective Equipment
ppm Parts Per Million
QUA Quantitative Usage Assessment
RBC Red Blood Cell
RED Reregistration Eligibility Decision
RfD Reference Dose
SAP Science Advisory Panel
SF Safety Factor
SOP Standard Operating Procedure
TGAI Technical Grade Active Ingredient
TRAC Tolerance Reassessment Advisory Committee
USDA United States Department of Agriculture
UF Uncertainty Factor
UV Ultraviolet
WPS Worker Protection Standard
-------
Executive Summary
Propetamphos is an organophosphate (OP) insecticide registered by Wellmark International for
the control of insects indoors. Target pests include ants, cockroaches, fleas, and termites in buildings and
structures. Propetamphos may be applied at indoor residential and medical sites, such as homes, apartment
buildings, stores, schools or hospitals. It may also be used in food service establishments, commercial, and
industrial buildings. Based upon available pesticide usage information between the years 1990 and 1997,
average annual domestic use at approximately 90,000 Ibs of active ingredient per year.
EPA has completed its review of public comments and has revised the risk assessments and
developed interim risk management decisions for propetamphos. The decisions outlined in this document
do not include the final tolerance reassessment decision for propetamphos. For propetamphos, the only
tolerance for residues in food commodities will remain unchanged. The final tolerance reassessment
decision for this chemical will be issued once the cumulative assessment for all the OPs is complete. The
Agency may need to pursue further risk management measures for propetamphos once the cumulative
assessment is finalized.
The revised risk assessments are based on review of the required target data base supporting the
use patterns of currently registered products and new information received. The Agency invited
stakeholders to provide proposals, ideas or suggestions on appropriate interim mitigation measures before
the Agency issued its risk mitigation decision on propetamphos. After considering the revised risks, as well
as mitigation proposed by Wellmark International, the technical registrant of propetamphos, mitigation
suggestions by the National Pest Management Association, and comments from other interested parties,
EPA developed its interim risk management decision for uses of propetamphos that pose risks of concern.
This decision is discussed fully in this document. Results of the risk assessments, and necessary label
amendments to mitigate those risks, are presented in this interim reregistration eligibility decision (IKED).
Overall Risk Summary
EPA's human health risk assessment for propetamphos indicates some risk concerns. Dietary
(food and drinking water) risk is not expected for all populations and is not of concern to the Agency.
Additionally, risks are low to workers who mix, load, and apply propetamphos at commercial and
residential use sites. However, there are post-application risk concerns for adults, and especially children
entering areas treated with propetamphos. Also, there are no environmental risk concerns.
To mitigate risks of concern posed by the uses of propetamphos, EPA considered the mitigation
proposal submitted by the technical registrant, as well as comments and mitigation suggestions from other
interested parties, and has decided on a number of label amendments to address the residential risk
concerns. Results of the risk assessments, and the necessary label amendments to mitigate those risks, are
presented in this IRED.
-------
Dietary (Food and Drinking Water)
There are no acute dietary (food) risks associated with propetamphos, and chronic (food) dietary
exposure for propetamphos residues is not expected. Because propetamphos is only used indoors,
exposure from drinking water sources are not expected and no drinking water assessment was conducted.
Provided that the label is amended to require that food is covered or removed prior to treatment, no further
mitigation measures are necessary at this time for dietary (food and drinking water) exposure to
propetamphos.
Occupational
Based on a proposed maximum dilution rate of 0.5 % solution of active ingredient, and the addition
of minimum personal protective equipment (PPE) consisting of single-layer clothing and chemical-resistant
gloves, both dermal and inhalation risks to applicators are low and not of concern to the Agency.
Residential
Risks resulting from use of propetamphos in the residential setting are of concern. Combined risks
(oral, inhalation, and dermal routes of exposure) for residential broadcast (flea) treatment using
propetamphos are high for adults, and especially high for children. Combined risks (dermal and oral (hand-
to-mouth)) for residential spot treatment, and crack and crevice applications using propetamphos are high
for children, but dermal risks are low for adults. Because of these risk concerns, the registrant has agreed
to voluntarily cancel all residential uses of propetamphos.
Chronic residential inhalation exposure to propetamphos is possible because of the termiticide use
of this pesticide, however, dermal or incidental oral exposure is not anticipated based on the use pattern
(gallery treatment). Based on a conservative exposure assessment, chronic inhalation risks are high for
adults and children, and are of concern to the Agency. In response, the registrant has informed the Agency
that it does not support the continued registration of termiticide use for propetamphos and has voluntarily
canceled this use.
Ecological Risk
Ecological risks associated with propetamphos use are not of concern to the Agency. Because all
currently registered uses of propetamphos are limited to indoor use, exposure to nontarget terrestrial and
aquatic plants and animals are not expected.
For the uses of propetamphos, the Agency has determined that, with the adoption of all of the label
amendments noted in this document, these uses may continue until the outcome of the cumulative
assessment of all OPs has been decided.
The Agency is issuing this IRED for propetamphos, as announced in a Notice of Availability
published in the Federal Register. This IRED includes guidance and time frames for complying with any
necessary label changes for products containing propetamphos. There is no comment period for this
document, and the time frames for compliance with the necessary changes outlined in this document are
shorter than those given in previous REDs. As part of the process discussed by the Tolerance
Reassessment Advisory Committee, which sought to open up the process to interested parties, the
-------
Agency's risk assessments for propetamphos have already been subject to numerous public comment
periods, and a further comment period for propetamphos was deemed unnecessary. Phase 6 of the pilot
process does not include a public comment period; however, for some chemicals, the Agency may provide
for another comment period, depending on the content of the risk management decision. With regard to
complying with the requirements in this document, the Agency has shortened this time period so that the
risks identified herein are mitigated as quickly as possible. Neither the tolerance reassessment nor the
reregistration eligibility decision for propetamphos can be considered final, however, until the cumulative
risk assessment for all OP pesticides is complete.
111
-------
IV
-------
I.
Introduction
The Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) was amended in 1988 to
accelerate the reregistration of products with active ingredients registered prior to November 1,1984. The
amended act calls for the development and submission of data to support the reregistration of an active
ingredient, as well as a review of all submitted data by the U.S. Environmental Protection Agency (referred
to as EPA or "the Agency"). Reregistration involves a thorough review of the scientific database underlying
a pesticide's registration. The purpose of the Agency's review is to reassess the potential hazards arising
from the currently registered uses of the pesticide; to determine the need for additional data on health and
environmental effects; and to determine whether the pesticide meets the "no unreasonable adverse effects"
criteria of FIFRA.
On August 3,1996, the Food Quality Protection Act of 1996 (FQPA) was signed into law. This
Act amends FIFRA to require tolerance reassessment of all existing tolerances. The Agency decided that,
for those chemicals that have tolerances and are undergoing reregistration, the tolerance reassessment will
be initiated through this reregistration process. It also requires that by 2006, EPA must review all
tolerances in effect on the day before the date of the enactment of the FQPA, which was August 3,1996.
FQPA also amends the Federal Food, Drug, and Cosmetic Act (FFDCA) to require a safety finding in
tolerance reassessment based on factors including an assessment of cumulative effects of chemicals with
a common mechanism of toxicity. Propetamphos belongs to a group of pesticides called OPs, which share
a common mechanism of toxicity by affecting the nervous system by inhibiting cholinesterase. Although
FQPA significantly affects the Agency's reregistration process, it does not amend any of the existing
reregistration deadlines. Therefore, the Agency is continuing its reregistration program while it resolves the
remaining issues associated with the implementation of FQPA.
This document presents the Agency's revised human health and ecological risk assessments; and
the interim decision on the reregistration eligibility of propetamphos. It is intended to be only the first step
in the reregistration process for propetamphos. The Agency will eventually proceed with its assessment
of the cumulative risk of the OP pesticides and issue a final reregistration eligibility decision for
propetamphos.
The implementation of FQPA has required the Agency to revisit some of its existing policies relating
to the determination and regulation of dietary risk, and has also raised a number of new issues for which
policies need to be created. These issues were refined and developed through collaboration between the
Agency and the Tolerance Reassessment Advisory Committee (TRAC), which was composed of
representatives from industry, environmental groups, and other interested parties. The TRAC identified
the following science policy issues it believed were key to the implementation of FQPA and tolerance
reassessment:
• Applying the FQPA 10-Fold Safety Factor
• Whether and How to Use "Monte Carlo" Analyses in Dietary Exposure Assessments
• How to Interpret "No Detectable Residues" in Dietary Exposure Assessments
• Refining Dietary (Food) Exposure Estimates
1
-------
• Refining Dietary (Drinking Water) Exposure Estimates
• Assessing Residential Exposure
• Aggregating Exposure from all Non-Occupational Sources
• How to Conduct a Cumulative Risk Assessment for Organophosphate or Other Pesticides with
a Common Mechanism of Toxicity
• Selection of Appropriate Toxicity Endpoints for Risk Assessments of Organophosphates
• Whether and How to Use Data Derived from Human Studies
The process developed by the TRAC calls for EPA to provide one or more documents for public
comment on each of the policy issues described above. Each of these issues is evolving and in a different
stage of refinement. Some issue papers have already been published for comment in the Federal Register
and others will be published shortly.
In addition to the policy issues that resulted from the TRAC process, the Agency issued on
September 29, 2000 a Pesticide Registration Notice (PR 2000-9) that presents EPA's approach for
managing risks from OP pesticides to occupational users. The Worker PR Notice describes the Agency's
baseline approach to managing risks to handlers and workers of OP pesticides. Generally, basic protective
measures such as closed mixing and loading systems, enclosed cab equipment, or protective clothing, as
well as increased restricted entry intervals will be necessary for most uses where current risk assessments
indicate a risk and such protective measures are feasible. The policy also states that the Agency will assess
each pesticide individually, and based upon the risk assessment, determine the need for specific measures
tailored to the potential risks of the chemical. The measures included in this IRED are consistent with that
draft Pesticide Registration Notice.
This document consists of six sections. Section I contains the regulatory framework for
reregistration, as well as descriptions of the process developed by TRAC for public comment on science
policy issues for the OP pesticides and the worker risk management PR notice. Section H provides a
profile of the use and usage of the chemical. Section DI gives an overview of the revised human health and
environmental effects risk assessments resulting from public comments and other information. Section IV
presents the Agency's interim decision on reregistration eligibility and risk management decisions. Section
V summarizes the label changes necessary to implement the risk mitigation measures outlined in Section IV.
Section VI provides information on how to access related documents. Finally, the Appendices A list the
use patterns eligible for reregistration; B, the necessary studies for reregistration; and C, the bibliography
listing citations of all studies considered relevant to the IRED document. The revised risk assessments are
not included in this document, but are available on the Agency's web page www.epa.gov/oppsrrdl/op, and
in the Public Docket.
-------
II.
Chemical Overview
A. Regulatory History
Propetamphos technical was first registered to Sandoz Crop Protection (Company No. 11273)
by the Agency in December 1980. In March 1981, the first end-use product was registered as a non-
food/non-feed use for indoor structural pest control. In 1983, a food/feed use in food/feed handling
establishments was registered. This permitted propetamphos to be used in food processing facilities (mills,
dairies, etc.), meat and poultry plants, food processing facilities (packing, canning, bottling, etc.), food
and/or feed warehouses, and food service establishments. The regulations to permit residues in food/feed
resulting from application in food handling establishment were announced in the Federal Register Notice
of November 23, 1983 (48 FR 52902). The registrations were transferred to Zoecon Industries
(Company No. 2724) in 1984. On June 23, 1997, the company name was subsequently changed to
Wellmark International (retaining the same company number of 2724).
In 1998, all propetamphos labels were amended to delete the food and feed handling establishment
uses, except food service establishment uses where food is prepared and served (e.g., restaurants).
B. Chemical Identification
CH O
CH,
H C
3
I O
HNL
'CH.
Propetamphos is a yellowish oily liquid with a boiling point of 87-89°C. Propetamphos is
practically insoluble in water (110 mg/L at 20° C), but is completely miscible in most organic solvents
including acetone, chloroform, diethyl ether, ethanol, hexane, and xylene. The vapor pressure of
propetamphos is 2.6 x 10'7 mm Hg at 25°C.
Chemical Name:
([(e)-]-methylethyl 3-
[[(ethylamino)methoxyphosphinothioyl]oxy]-2-butenoate)
Common Name:
Chemical family:
CAS registry number:
OPP chemical code:
Propetamphos
Organophosphate
31218-83-4
113601
3
-------
Empirical formula:
Molecular weight:
Trade and other names:
Basic manufacturer:
C10H20N04PS
281.3 g
Catalyst™, Safrotin ™ , ZoeconT
Wellmark International
C. Use Profile
Type of Pesticide
Propetamphos is an insecticide used for indoor structural pest control. The following is a summary
of propetamphos use sites:
Indoor Food/Non-Food: There are no food uses of propetamphos, however, propetamphos may be used
in food service establishments. Application is limited to spot and crack and crevice treatments. Food
service establishments may include restaurants, cafeterias, taverns, delicatessens, mess halls, school and
institutionaldining areas, hospitals, mobile canteens, vending machines, groceries and markets. Indoor non-
residential non-food areas (may include eating establishments, office buildings, commercial and industrial
premises and equipment) where there is no contact with food, and where no food processing, packing, and
no food and/or feed warehousing occurs.
Residential: Propetamphos is used inside residential homes on carpets (limited to broadcast applications
for fleas) and other surfaces, on hard surfaces (e.g., floors, counters, walls), spot applications (areas up
to 2' X 2'), crack and crevice (primarily for cockroach control), and galleries for termites (e.g., crawl
spaces, foundations).
Public Health According to the National Center of Infectious Diseases of the Centers for Disease Control
and Prevention, "propetamphos is not used regularly as an insecticide in public health programs in the
United States." Propetamphos is not on the Agency's proposed listing of Public Health Pesticides.
Other Non-Food: Propetamphos is used in pet living/sleeping quarters, and in institutional/medical and
veterinary facilities.
Target Pests
Propetamphos is used to control silverfish, cockroaches, earwigs, beetles, fleas, ants, termites,
ticks, other indoor insects, and spiders.
Formulation Types
There are three current registered products that contain propetamphos: one manufacturing-use
product (MUP) (EPA Reg. No. 2724-313) containing 90% active ingredient (ai), and two end-use
products (EPs). One EP consists of a 46.5% ai emulsifiable concentrate (Zoecon 8718 EW, EPA Reg.
No. 2724-449) formulation, and the other is an 18.9% ai soluble concentrate (Zoecon 9001 EW, EPA
-------
Reg. No. 2724-450) formulation. Only Zoecon 9001EW is currently manufactured and used in the United
States, whereas Zoecon 8718 EW is manufactured for export only and has never been sold in the United
States. The registrant has voluntarily canceled the Zoecon 8718 EW product registration. There are no
section 24(c) special local need registered propetamphos products or uses.
Method and Rates of Application
Propetamphos is applied as a water dilution through a compressed air sprayer, often with a low
pressure hand wand. Termite applications use a crack and crevice or injection tube nozzle. For general
surface application, propetamphos is applied at a rate of 0.5% ai in a fine spray. Approximately 1 gallon
of finished spray is used per 1500 square feet for broadcast application. For spot, and crack and crevice
applications, propetamphos is applied as a 0.5 to 1.0 % ai solution. For spray applications, propetamphos
is applied as a 1.0% ai spray. Gallery (termite) applications are applied at a 1.0% ai spray using low
pressure equipment. For all applications, additional treatment may be repeated as needed, but not more
than once every 7 days, and not to exceed 2 treatments in a 30-day period.
Use Classification
The 46.5 % ai emulsifiable concentrate formulation (Zoecon 8718 EW, EPA Reg. No. 2724-449)
is classified as a restricted-use product, due to acute oral and dermal toxicity. The 18.9 % ai soluble
concentrate product (Zoecon 9001 EW, EPA Reg. No. 2724-450) is not classified as a restricted use
product.
D. Estimated Usage of Pesticide
This section summarizes the best estimates available for the pesticide uses of propetamphos, based
on available pesticide usage information between the years 1990 and 1997. Totalannual usage has been
estimated at 90,000 Ibs ai/year. About 70% of this total annual propetamphos usage is applied to
residential areas (by Pest Control Operators (PCOs)), while the remaining 30% is applied to various
commercial sites. About 90% of application is carried out by PCOs, while most of the remaining 10% of
applications are by not-for-hire applicators, such as maintenance workers.
An estimated 1.2% of all residences, and 3.3% of all food handling establishments are treated with
propetamphos each year (food service establishments are a subset of food handling establishments, and
annual treatment based on this use alone would be less than 3.3%). Estimates of propetamphos use (Ibs
ai) are based on the 1993 Certified/Commercial Pesticide Applicator Survey (C/CPAS), 1992 National
Home and Garden Pesticide Use Survey (NHGPUS), and other proprietary data sources. The quantitative
usage assessment for propetamphos is provided in Table 1.
-------
a
42
C3
•a
r-
o\
Os
1-H
Os
as
e
•o
cu
en
03
55
0
i
"cu
P
2
*«
«2
•4-)
CU
cn
en
CU
cn
CU
SJ
C3
P
CU
"S
•£
c
O
«
I
g-.
o
S3
i
e
O
T;
«
"3.
cu
"*•
»lf
eS-
o •«
ll
&M4 "^
1
1-8
13 15
O O
"3
"3
U
•a
13
i"
F^
«
2
t/j
a,
1
cu
£
s. -5
•if
&
CO*
^ 1
trt «5
S"
>> 3>
"3 f
sS ^
•J <
E
>» a
"3 E
ai
>> Si
« 2
™ «>
^s
*flj C
31
1" 1
A
-.-S
'S
P
s
S
Use Site
OS
o
o
t
Os
VI
o
o
m
VO
2?
cn
CN
.1 •
• — ; eft
CO
C
.9
*~^
c
O c/}
••-;
U
Os
is
c
U
•a
"1
g
|£
imercial
dings Total
g?
u oa
0
o
0
VO
oo
o
CN
CN
^
O
cn
cn
c
_o
H tK
2 IT
Os CO
VO
c
_o
10
•2«
PQ o"
vq w
i Handling
blishments
O co
Os
0
O
o
VO
oo
o
10
3?
CN
00
0
g .
3 «
c
o
== *
"H
c
.2
3 "S-
^ CO
VO
"c3
er Commerci
dings
J3 H3
0 ^
o
OS
£
O co
cjH o
fa -S S cn .^
•S "Si W CS CO
w u =s in
-------
III. Summary of Propetamphos Risk Assessment
Following is a summary of EPA's revised human health and ecological risk findings and conclusions
for the OP pesticide propetamphos, as fully presented in the documents: Updated Revised Preliminary
Risk Assessment: Propetamphos, June 7, 1999; Updated Occupational and Residential Dermal
Exposure Assessment addendum, September 27, 2000; EFED Integrated Science Chapter for
Propetamphos, December 2,1997; andPropetamphos Errata Sheet For EFED Chapter, January 12,
1999. The purpose of this summary is to assist the reader by identifying the key features and findings of
these risk assessments, and to better understand the conclusions reached in the assessments.
The risk assessment summaries presented here form the basis of the Agency's risk management
decision for propetamphos only; the Agency must complete a cumulative assessment of the risks of all the
OP pesticides before any final decisions can be made.
A. Human Health Risk Assessment
EPA issued its preliminary risk assessments for propetamphos on December 15, 1998. In
response to comments and studies submitted during Phase 3, the risk assessments were updated and
refined and were included in the revised risk assessment and addendum, dated June 7, 1999 and
September 27,2000, respectively. This risk assessment serves as the basis for this IRED. Major revisions
to the human health risk assessment are listed below:
1. Dietary Risk from Food
a. Toxicity
The Agency has reviewed all submitted toxicity studies and has determined that the toxicity
database for propetamphos is complete, and that it supports an interim reregistration eligibility determination
for all currently registered uses. Further details on the toxicity of propetamphos can be found in the June
7,1999 Human Health Risk Assessment and the September 27,2000 addendum. A brief overview of the
studies used for the dietary risk assessment is outlined in Table 2.
The toxicity data base provides evidence that cholinesterase inhibition is the most sensitive
toxicological observation in laboratory animals. Propetamphos, like other OPs, has anticholinesterase and
neurotoxic effects in all species tested, including dogs, rabbits, rats, and mice. Signs of neurotoxicity, such
as muscle tremors, fasciculations and cholinesterase inhibition (ChEI) have been observed in acute,
subchronic, chronic and developmental/reproductive toxicity studies. Propetamphos did not, however,
induce organophosphate induced delayed neurotoxicity in hens when orally dosed as part of a delayed
neurotoxicity study. Propetamphos is acutely toxic via the oral route of exposure and is classified as a
toxicity category II, based on an oral rat study (MRID 41607417) with a Lethal Dose (LD50) = 116-1
mg/kg in males and Lethal Dose (LD50) = 96.4 mg/kg in females.
-------
The subchronic and chronic toxicity studies demonstrate that propetamphos inhibits cholinesterase
activity in plasma, red blood cells (RBC), and/or brain in rats, dogs, and mice. Clinical signs associated
with cholinesterase activity inhibition were observed and included ataxia, tremors, salivation, constricted
pupils, and dyspnea. Propetamphos was not toxic to the visual system of dogs in a chronic toxicity study.
There is no evidence of increased susceptibility for infants and children, based on adequate
developmental toxicity studies in rats and rabbits and an adequate two-generation reproduction study in
rats. Following in utero exposures, no developmental toxicity was seen in rats. In the rabbit study,
developmental toxicity occurred only at a dose that also caused maternal toxicity. In the two-generation
rat reproductive toxicity study, offspring toxicity was only seen in the presence of maternal systemic toxicity.
The Agency has concluded that there are no metabolites of toxicological concern and that the
residues to be regulated in food commodities will consist of propetamphos per se.
b. FQPA Safety Factor
The FQPA Safety Factor Committee determined that the lOx FQPA safety factor should be
removed (equivalent to Ix), based on the following factors:
• In prenatal developmental toxicity studies following in utero exposure in rats and rabbits, there was
no evidence of developmental effects being produced in fetuses at lower doses as compared to
maternal animals nor was there evidence of an increase in severity of effects at or below maternally
toxic doses.
• In the pre/post natal two-generation reproduction study in rats, there was no evidence of enhanced
susceptibility in pups when compared to adults (i.e., effects noted in offspring occurred at
maternally toxic doses or higher).
• There was no evidence of abnormalities in the development of the fetal nervous system in the
pre/post natal studies.
• There was no concern for positive neurological effects from the available neurotoxicity studies or
for histopathology in the central nervous system from the other toxicological studies (e.g.,
subchronic rat, chronic dog, chronic rat and mouse).
• The toxicology data base is complete, and there are no data gaps according to the Subdivision F
Guideline requirements.
• Adequate actual data, surrogate data, and/or modeling outputs are available to satisfactorily assess
dietary and residential exposure.
-------
c. Population Adjusted Dose (PAD)
The PAD is a term that characterizes the dietary risk of a chemical, and reflects the Reference Dose
(RfD), either acute or chronic, that has been adjusted to account for the FQPA safety factor (i.e., RfD •*-
FQP A safety factor). The RfD is the level of daily exposure to a pesticide residue which is believed to have
no significant deleterious effects. In the case of propetamphos, the FQPA safety factor is 1; therefore, the
acute and chronic RfDs are equal to the acute and chronic PADs, respectively. A risk estimate that is less
than 100% of the acute or chronic PAD does not exceed the Agency's risk concern.
d.
Hazard Determination
Cholinesterase inhibition was the toxicity endpoint chosen for the acute and chronic dietary
endpoints. For risk assessments describing acute oral exposures, the dose selected was the no observed
adverse effect level (NOAEL) of 0.05 mg/kg/day based on brain cholinesterase inhibition at a lowest
observed adverse effect level (LOAEL) of 0.1 mg/kg/day observed in the 4-week oral toxicity study in
mice. An uncertainty factor of 100 (lOx for inter-species extrapolation and lOx for intra-species variation)
and an FQPA safety factor of Ix was applied to the NOAEL, therefore, the acute PAD is 0.0005
mg/kg/day.
For the chronic dietary risk assessment, the dose selected for risk assessment was the NOAEL
of 0.05 mg/kg/day based on plasma cholinesterase inhibition at a LOAEL of 0.1 mg/kg/day observed in
the 1-year chronic toxicity and carcinogenicity study in mice. An uncertainty factor of 100 (lOx for inter-
species extrapolation and lOx for intra-species variation) and an FQPA safety factor of Ix was applied
to the NOAEL, therefore, the chronic PAD is 0.0005 mg/kg/day. This toxicity and endpoint selection
information is summarized in Table 2.
Table 2. Toxicology Endpoints for Dietary Risk
Exposure
Scenario
Acute
Dietary
Chronic
Dietary
Dose
(mg/kg/day)
NOAEL = 0.05 mg/kg/day
(4-week oral mouse study)
NOAEL = 0.05 mg/kg/day (mouse chronic
feeding/ carcinogenicity study)
•- " •••• •:-•£..-: -.
Endpoint ;
Brain cholinesterase
inhibition (ChEI)
Brain, RBC, and plasma
ChEI
UF
100
100
FQPA
SF
1
1
PAD
(mg/kg/day)
0.0005
0.0005
e.
Cancer Determination
The Agency has classified propetamphos as "not likely to be a human carcinogen." This
classification is based on the lack of evidence of carcinogenicity in male and female rats and in male and
female mice when tested at dose levels that caused cholinesterase inhibition and, therefore, were judged
to be adequate to assess the carcinogenic potential of propetamphos. Additionally, propetamphos was
non-mutagenic both in vivo and in vitro.
-------
f. Acute Dietary (Food) Risk
Acute dietary risk considers all food that is eaten in one day (in this instance, by the individual who
consumed the most) and maximum, or high-end residue values in the food. It is the Agency's policy that
acute dietary exposure analysis does not take into account food handling establishments. Residues resulting
fiompesticide use in food handling establishments (or food service establishments-a subset of food handling
establishments) are not likely to result in incidental contamination of all foods at tolerance levels on a
uniform and consistent basis, and not all foods consumed by an individual in a day are likely to have come
from a food handling establishment. Therefore, an acute dietary (food) exposure and risk assessment is
not needed for pesticides having only food handling establishment tolerances, such as propetamphos.
g-
Chronic Dietary (Food) Risk
Because a tolerance is required for pesticides used for treatments of food service establishments,
the Agency assesses chronic dietary (food) exposure, due to concerns of inadvertent residues on food in
food service establishments when sprayed applications are made. Chronic dietary (food) exposure is
calculated using the average consumption value for food and average residue values on those foods over
a 70-year lifetime. Chronic dietary exposure is then compared with the chronic PAD (cPAD). ThecPAD
is the dose at which an individual could be exposed over the course of a lifetime and no adverse health
effects would be expected. The chronic dietary risk estimate is expressed as percent of the cPAD. A risk
estimate that is less than 100% of the cPAD does not exceed the Agency's level of risk concern.
For propetamphos, a Tier IE chronic dietary exposure assessment was conducted based upon
anticipated residues and the estimate of 11% of food handling establishments being treated with
propetamphos. Magnitude of the residue data showed that propetamphos residues were non-detectable
(<0.01 ppm) in/on foods that were held in closed containers. Therefore, anticipated residues of 0.005 ppm
('/2 Limit Of Detection (LOD)) were used in the Tier HI chronic dietary assessment.
Also, this chronic dietary assessment was conducted prior to refinements to the quantitative usage
assessment (QUA) in Table 1. Since the time of this analysis, the percent of food handling establishments
treated with propetamphos has been lowered from 11% to 3.3%. Incorporating this refined usage
informationinto the analysis will lower the chronic dietary risks. Presently, the chronic dietary risks are low,
thus, further refinements to the chronic dietary analysis to reflect this usage information were not conducted.
The Tier HI chronic analysis, based on non-detectable residues on foods held in covered containers
during pesticide application, indicates that chronic dietary (food) exposure and risk estimates for
propetamphos are below the Agency's level of concern. Refer to Table 3 for the propetamphos chronic
dietary risk estimates.
10
-------
Table 3. Chronic Dietary Risk of Prooetamohos8 For Covered Food
Population Subgroups
U.S. Population
Non-nursing infants (< 1 year old)
Children (1-6 years old)
Exposure (mg/kg/day)
0.000030
0.000104
0.000061
--. i Chronic Risk (% cPAD)
6%
21%
12%
a Expressed in terms of propetamphos per se.
As indicated above, the chronic dietary assessment is based on no detectable residues. It is the
Agency's policy to use 1A LOD, which is 0.005 ppm for propetamphos, to estimate dietary risk when no
residues are detected. Realistically, provided foods are covered or removed prior to treatment of the area,
actual chronic (food) dietary risk for treatment in food service establishments may be as low as zero.
2. Dietary Risk from Drinking Water
Propetamphos is presently not registered for use on food/feed crops, potable water, or aquatic
food, and is not expected to be released to water. Therefore, exposure from drinking water sources is not
expected and no drinking water risk assessment was conducted.
3. Occupational and Residential Risk
Occupational workers can be exposed to propetamphos through mixing, loading, and applying, or
re-entering treated sites. Residents or homeowners can be exposed to propetamphos through entering or
performing other activities in treated areas. Occupational handlers of propetamphos include pest control
operators (PCOs) who mix, load, and apply pesticides. Risk for all of these potentially exposed
populations is measured by a Margin of Exposure (MOE), which determines how close the occupational
or residential exposure comes to a NOAEL, or,if necessary, by the Aggregate Risk Index (ART), which
is a way to aggregate MOEs that have dissimilar target MOEs. For propetamphos, dermal and oral MOEs
greater than 100, inhalation MOEs greater than 300, and ARIs that are greater than 1.0 are not of concern
to the Agency.
a. Toxicity
In summary, propetamphos is acutely toxic via the oral and dermal routes of exposure, has low
inhalation toxicity, is not a skin or eye irritant, and is not a dermal sensitizer. Propetamphos, technical, is
placed in toxicity category H for acute oral and dermal toxicity, category IH for acute inhalation, and
category IV for acute eye and skin irritation. A summary of the acute toxicity profile of propetamphos is
provided in Table 4.
11
-------
Table 4. Acute Toxicitv Profile of Pro
Study Type
Acute Oral-Rat
Acute Dermal-Rabbit
Acute Inhalation-Rat
Primary Eye Irritation
Primary Skin Irritation
Dermal Sensitization
MRIDNo.
41607417
41607418
41529301
41607419
41607420
41607412
petamphos
Results
LD50= 116.1 mg/kg, males
LD50= 96.4 mg/kg, females
LD50= 486.4 mg/kg, both sexes combined
LC50= 1.5 mg/L, males
LC50= 0.69 mg/L, females
Negative for eye irritation
Negative for dermal irritation
Negative for dermal sensitization
Toxicity Category
II
II
m
IV
IV
N/A
b.
Hazard Determination
For the short- and intermediate-term (< 30 days) dermal risk assessment, the dose selected was
the NOAEL of 1.25 mg/kg/day, based on brain cholinesterase inhibition at a LOAEL of 2.5 mg/kg/day
observed in the 21-day dermal toxicity study in rats. Due to concerns of rapid detoxification of some OPs
when rabbits are used for dermal toxicity tests, and thereby sometimes underestimating risk, the registrant
conducted a 21-day dermal toxicity study in rats. The Agency has recently received the 21-day dermal
toxicity study in rats, and has conducted a preliminary review. The Agency is currently conducting a final
review of the study and is confident that the NOAEL is 1.25 mg/kg/day and will be selected by the
Agency's Hazard Identification and Assessment Review Committee. An MOE of greater than 100 (lOx
for inter-species extrapolation and lOx for intra-species variation) does not exceed the Agency's level of
concern for these risk assessments. Because a dermal study was used to determine the toxicity endpoint,
a dermal absorption factor is not necessary.
For the intermediate- (> 30 days) and long-term dermal risk assessment, the dose selected was
the NOAEL of 0.08 mg/kg/day, based on RBC cholinesterase inhibition at a LOAEL of 0.17 mg/kg/day
observed in the 6-month subchronic toxicity study in dogs. An MOE of greater than 100 (1 Ox for inter-
species extrapolation and lOx for intra-species variation) does not exceed the Agency's level of concern
for these risk assessments. However, based on current use patterns, it is expected that applicators will not
be continuously exposed to propetamphos for greater than 30 days. Therefore, the dermal risk assessment
is based on the short- and intermediate-term (< 30 days) toxicity endpoint discussed above and listed in
Table 5.
For inhalation exposure (of any duration), the dose selected for risk assessment was the LOAEL
of 4.7 mg/kg/day based on plasma cholinesterase inhibition at this dose in a 14-day rat inhalation toxicity
study. Because a NOAEL was not established in this study, an extra uncertainty factor of 3x was applied.
Therefore, a MOE of greater than 300 (lOx for inter-species extrapolation, lOx for intra-species variation,
and 3x for use of LOAEL) does not exceed the Agency's level of concern for these risk assessments. A
summary of the toxicological endpoints, and other factors used in the occupational and residential risk
assessments for propetamphos are listed below in Table 5.
12
-------
For the oral ingestion (children) route of exposure, the toxicological endpoint was based on a 4--
week oral mouse study. This study is further described in Section HI. A.l.d Hazard Determination for
human dietary risk (see Table 2).
TableS. Summary of Toxicological Endpoints for Occupational and Residential Risks
Assessment
Short- and Intermediate
term dermal (<30 days)
[ntermediate-term dermal
(>30 days)
Long-term dermal
(> 180 days)
Oral ingestion (children)
Inhalation
(Any time period)
Dose
(mg/kg/day)
NOAEL=1.25
•VT/~\AT?T n fiQ
NOAEL = 0.05
LOAEL=4.7
"Endpoint
Brain cholinesterase
inhibition (ChEI)
RBC ChEI at 4 weeks.
This is supported by a NOAEL of
0.05 mg/kg/day for brain ChEI in a
4-week mouse study
Brain ChEI
Plasma ChEI in both sexes. No
NOAEL established.
Study
21 -day
dermal rat
6-month
oral dog
study
4 week oral
mice
14- day
inhalation
rat
Absorption
factor
N/A
100
N/A
100
Target
MOE
100
100
100
300
c. Exposure
Occupational Exposure
Chemical-specific exposure data for handlers were not available for propetamphos, so risks to
pesticide handlers were assessed from data derived from the Pesticide Handlers Exposure Database
(PHED), using standard assumptions based on the exposure scenarios and types of equipment supported
by current labeling. The basic premise of PHED is that the chemical formulation (i.e., soluble concentrate)
and method of application are the major determinants of pesticide exposure, rather than chemical specific
properties. PHED is a database containing exposure data for surrogate chemicals used in a number of
different formulations and application scenarios. The occupational exposure assessment was conducted
for a worker who not only mixes, but loads and applies this insecticide in one day (with the assumption that
one worker may perform all three tasks in a day and, therefore, will have additive exposures from all three
tasks). The quality of the data and exposure factors represent the best sources of data currently available
to the Agency for completing these kinds of assessments. The exposure factors (e.g., body weight, amount
ai treated per day, protection factors, etc.) are all standard values that have been used by the Agency over
several years. For more information about PHED and the data used for each scenario, see the Updated
Revised Preliminary Risk Assessment: Propetamphos, June 7, 1999 and the Updated Occupational
and Residential Dermal Exposure Assessment addendum, September 27, 2000, which is available in
the public docket and on the Internet.
Anticipated use patterns and application methods, range of application rates, and typical rate of
coverage were derived from current labeling. Application rates specified on propetamphos labels range
from 0.5 to 1.0% concentration of active ingredient per gallon of finished solution. One gallon of finished
13
-------
spray (at a diluted solution of 0.5%) will typically cover 1500 square feet for broadcast application. There
are no restrictions on the label stipulating how much product may be used in any given day.
Occupational handler exposure assessments are conducted by the Agency using different levels of
personal protection. The Agency typically evaluates all exposures with minimal protection and then adds
additional protective measures using a tiered approach to obtain an appropriate MOE (i.e., going from
minimalto maximum levels of protection). The lowest tier is represented by the baseline exposure scenario,
followed by, if needed (i.e., MOEs are less than 100 for dermal exposure and MOEs are less than 300 for
inhalation exposure), increasing levels of risk mitigation to include personal protective equipment (PPE).
Currently, there is no requirement for PPE on the propetamphos labels. The levels of protection that
formed the basis for calculations of occupational exposure from propetamphos activities include:
Baseline:
Minimum PPE:
Maximum PPE:
Long-sleeved shirt and long pants, shoes and socks.
Baseline + chemical resistant gloves.
Coveralls over long-sleeved shirt and long pants, shoes and socks, and
chemical-resistant gloves.
Residential Exposure
Residential exposure is assessed by determining how a person could come into contact with a
pesticide in and around a home. There are no registered homeowner uses for propetamphos at the present
time. However, post-application exposure is possible as a result of PCO indoor broadcast (flea control)
or spot, and crack and crevice (e.g., cockroach, ant, cricket control) applications. Since propetamphos
is used strictly indoors, and only applied by PCOs, residential exposure to propetamphos takes place when
people come into contact with post-application residues either by touching, breathing, or ingesting them.
Therefore, residential post-application exposure scenarios were considered for the broadcast, spot, and
crack and crevice use scenarios.
^ Where available, chemical-specific post-application exposure data have been used for these
scenarios. When no chemical-specific data is available, the post-application exposure assessment is based
on the newly proposed Standard Operating Procedures (SOPs) for Residential Exposure Assessments and
recommended approaches by the Agency's Health Effects Division (HED), Exposure Science Advisory
Committee (ExpoSAC). The newly proposed SOPs for Residential Exposure Assessments alter the
residentialpost-application scenario assumptions. Compared with the previous SOPs, the newly proposed
SOPs are expected to better represent residential exposure, but are still considered to be high-end,
screening level assumptions.
For the post-application scenario resulting from the indoor broadcast use (carpet treatment for flea
control), residential exposures were estimated using a chemical-specific (Jazzercise) post-application study.
Because there are no chemical-specific studies measuring post-application exposures resulting from the
spot, and crack and crevice use of propetamphos, the proposed Residential SOPs were used to assess
exposure.
14
-------
To assess chronic inhalation exposure resulting from the termiticide use, the Agency utilized the
Multi-Chamber Concentration and Exposure Model (MCCEM), as outlined in the SOPs for Residential
Exposure Assessments. The MCCEM is a model that is capable of calculating indoor air concentrations
and the corresponding exposure assessments for chronic scenarios. The MCCEM contains a database
of various default house data that are needed to complete each calculation, such as air exchange rates,
geographically based inter-room air flows, and house/room volumes.
d. Occupational and Residential Risk Summary
Occupational Risk
An occupational exposure assessment was conducted for a worker who mixes, loads, and applies
propetamphos (one worker is considered to perform all three tasks). The Aggregate Risk Index (ART)
is a way to aggregate MOEs that have dissimilar target MOEs. Because the target MOE for dermal
exposure is 100 and the target MOE for inhalation exposure is 300, an ARI method to combine the MOEs
is necessary. ARIs that are greater than 1.0 are not of concern to the Agency. As indicated in Table 6,
the ARIs are greater than 1.0 for all occupational use scenarios and are, therefore, not of concern.
Table 6. Occupational Mixer/Loader/Applicator Risk Assessment
Use Scenario -
Low Pressure
Handwand,
Broadcast or
Crack and
Crevice
Gallery Injection
Treatment for
Termites
5 homes/day, 0.5% ai
1 0 apartments/
day, 0.5% ai
5 homes/day, 1.0% ai
1 0 apartments/
day, 1.0% ai
Igal, l%ai
2 gal, l%ai
3 gal, l%ai
Dermal MOEs a
Minimum
PPE
625
310
310
160
3000
1500
1000
Maximum
v PPE
740
370
370
180
4500
2200
1500
Inhalation MOEs b
; No Respirator
>8400
>8400
>8400
8400
>6.3E5
>6.3E5
6.3E5
v? ARIs0
Minimum
PPE
>5.1
>2.8
>2.8
1.5
>30
>15
10
Maximum
PPE
>5.8
>3.3
>3.3
1.7
>45 .
>22
15
a Dermal NO AEL = 1.25 mg/kg/day, (21 -day dermal rat study).
b Inhalation NOAEL = 0.027 mg/L = 4.7 mg/kg/d (14 day inhalation toxicity study in rats).
c ARI<1 is of concern to the Agency.
Residential Risk
Most residential exposures to propetamphos are from entering or performing some activity on
treated areas. Post-application exposure was assessed on the same day the pesticide was applied, since
it was assumed that homeowners could contact treated areas immediately after application.
Similarly with the occupational risk assessment, because the target MOEs for propetamphos are
100 for dermal and oral exposure, and 300 for inhalation, an ARI method to combine the MOEs for
residential risk is necessary. ARIs that are greater than 1.0 are not of concern to the Agency.
15
-------
Broadcast Application
As indicated in Table 7, the dermal MOEs for both adults and children are significantly below the
target MOE of 100. Incidental oral exposures (hand-to-mouth) for children is also below the target MOE
of 100. However, inhalation MOEs are above the target MOE of 300 for adults and children. Therefore,
the combined (ARI) exposure from broadcast carpet treatment is less than 1.0 and of concern to the
Agency. Because a chemical-specific exposure study is available (Jazzersize study using 0.5% Safrotin
solution), the Agency has a high level of confidence in these exposure and risk estimates. A summary of
these risk estimates are provided in Table 7.
Table 7. Summary of Dermal, Inhalation, and Oral MOEs for Broadcast Carpet Treatment
Population
Adults
Children
Dermal MOE"
10
2
Inhalation MOE?
3900
1400
OralMOF
N/A
0.4
'^^.-v^.ART1 ;.;...
0.1
0.003
Dermal NOAEL = 1.25 mg/kg/day, (21 -day dermal rat study).
•"Inhalation NOAEL = 0.027 mg/L = 4.7 mg/kg/d (14 day inhalation toxicity study in rats).
cAcute Oral NOAEL = 0.05 mg/kg/d (4 week oral toxicity study in mice).
d ARI<1 is of concern to the Agency.
Spot, and Crack and Crevice Application
Chemical specific data were not available depicting exposures resulting from the spot, and crack
and crevice application. The residential post-application exposure assessment for the crack and
crevice/spot treatment application of propetamphos was conducted using the proposed revisions to the
Residential Exposure Assessment SOPs.
The following considerations and assumptions were used to estimate post-application exposure and
risk from spot, and crack and crevice applications, based on the proposed reduced maximum application
rate and current label instructions for spot, and crack and crevice (i.e., spot applications to baseboards):
• a proposed maximum rate of dilution of 0.5% ai solution
one quart of diluted material would be used to treat a 2,500 ft2 home
based on chemical-specific data, only 0.5% of the residue on carpet is dislodgeable using the hand
roller method
• only 1% of the residue is dislodgeable on hard surfaces
post-application exposure was assessed on the same day the pesticide was applied, since it is
assumed that homeowners could contact the treated surfaces immediately after application.
• the duration of exposure is assumed to be 8 hours per day for carpet and 4 hours for hard surfaces
• the mean dermal transfer coefficient was assumed to be 16,700 cnf/hr for adults and 6,000 cnf/hr
for children
• for children incidental hand-to-mouth exposures, the surface area of the hand put into the mouth
was assumed to be 20 cm2 with 20 events/hr, and this activity lasts 2 hours
At the proposed maximum dilution rate of 0.5% ai solution, the dermal MOEs for adults are above
the target MOE of 100. Dermal and oral (hand-to-mouth) MOEs for children are below the target MOE
16
-------
of 100. Because the dermal and oral target MOEs are the same (100), the MOEs for both routes of
exposure can be combined to assess risks to children. Therefore, dermal risks to adults are not of concern,
and risks to children are of concern to the Agency. Table 8 summarizes the risk results from spot, and
crack and crevice applications of propetamphos.
Table 8. Residential Post-Application Risks from Crack and Crevice/Spot Treatment Use
•'-: ~' : ./Scenario;:;--".;., ":f .'-.-?':'••:'
Exposure from residue deposition on carpet
Exposure from residue deposition on hard
surfaces
Population -.:
Children
Adult
Children
Adult
Dermal MOi?
80
140
80
140
OralMOI?
50
Combined MOE
31
NA
23
18
NA
"Dermal MOE based on NOAEL =1.25 mg/kg/day (21-day rat dermal toxicity study)
b Oral MOEs based on NOAEL = 0.05 mg/kg/day (4-week oral mouse study)
Termiticide Application
Chronic residential inhalation exposure to propetamphos is possible because of the termiticide use
of this pesticide. Dermal or incidental oral exposure is not anticipated based on the use pattern (gallery
treatment). The exposure assessment for the gallery treatment is based on the Multi-Chamber
Concentration and Exposure Model (MCCEM), as outlined in the SOPs for Residential Exposure
Assessments.
The termiticide assessment represents a conservative Tier I estimate of exposure. It is assumed
that 100% gallery treatment (i.e., applied inside the home) technique is a source for oflfgassing for long-term
inhalation exposure. Based on this conservative (Tier I) exposure assessment, chronic inhalation MOEs
for adults and children were 150 and 48 respectively. Because the chronic inhalation MOEs were below
the target MOE of 300, the inhalation exposure from termiticide use of propetamphos is of concern to the
Agency. This risk information is summarized in Table 9.
Table 9. Residential Chronic Post-Application Risks from Termiticide Use
Scenario
Chronic exposure from termiticide use
-Population
Adult
Children
Inhalation MOI?
150
48
a MOEs based on LOAEL = 47 mg/kg/day (14-day inhalation rat study)
Because the Agency does not have chemical-specific termiticide use data for propetamphos, the
actual use pattern of propetamphos (gallery injections with sealing of holes in dry wall) may well result in
less than 100% of the total amount applied being available as a source. This model is intended to be a
conservative screening scenario because it assumes 21 hours of residential exposure in a generic house with
a moderate air exchange rate. The application of a 1% solution was also assumed. Because of these
factors, the risk estimates provided in Table 9 are considered to be an overestimate and actual risk resulting
from termiticide applications are expected to be much lower.
17
-------
4. Aggregate Risk
An aggregate risk assessment looks at the combined risk from dietary exposure (food and drinking
water routes) and residential exposure (dermal and inhalation exposure, and incidental hand-to-mouth oral
exposure for children). For propetamphos, all individual and combined MOEs must be greater than the
target MOE (i.e., 100 for dermal and oral, and 300 for inhalation), and the ARI must be greater than 1 to
be not of concern to the Agency. Results of the aggregate risk assessment are summarized here, and are
further discussed in Propetamphos Updated Revised Preliminary Risk Assessment, June 7,1999 and
Updated Occupational and Residential Dermal Exposure Assessment addendum, September 27,2000.
Acute Aggregate Risk
Acute aggregate exposure assessments take into account acute dietary food and drinking water
exposures. An acute aggregate risk assessment is not needed because only food handling establishment
tolerances are established for propetamphos. Residues resulting from pesticide use in food handling
establishments are not likely to result in incidental contamination of all foods at tolerance levels on a uniform
and consistent basis, and not all foods consumed by an individual in a day are likely to have come from a
food handling establishment. Also, based on the nature of propetamphos uses (in buildings and structures),
residues are not expected in drinking water, therefore, an acute aggregate assessment of risk is not
necessary.
Short-Term Aggregate Risk
Short-term aggregate risk takes into account short-term residential exposures (dermal and
inhalation for adults), and dermal, inhalation and oral [incidental hand-to-mouth] for children, combined with
chronic dietary (food) exposure. Because propetamphos is not expected in drinking water, the dietary
component of the aggregate risk assessment is based on food exposure only. As indicated in Table 3, there
are no chronic dietary food concerns (provided that foods are removed or covered during applications).
For broadcast carpet treatments, the ARIs for adults (combined MOEs for dermal and inhalation)
for residential post-application exposure are less than 1.0 and, therefore, are of concern (see Table 7).
The ARIs for children (combined MOEs for dermal, inhalation and oral [incidental hand-to-mouth]
exposure) for residential post-application exposure are also less than 1.0 and are of concern (see Table
7). The ARIs are 0.1 and 0.003 for adults and children, respectively. Therefore, an aggregate risk
assessment with dietary exposure was not be conducted.
For spot, and crack and crevice treatments, dermal MOEs for residential post-application exposure
to adults were above the target MOE. Therefore, an aggregate assessment with chronic dietary (food)
exposure was conducted and the resultant aggregate risks are not of concern. For children, combined
dermal and oral (hand-to-mouth) MOEs for all scenarios are below the target MOE and are of concern
(see Table 8). Therefore, an aggregate risk assessment with dietary exposure to children was not be
conducted.
18
-------
Chronic Aggregate Risk
Aggregate chronic risk estimates consider chronic dietary (food) and chronic residential
(termiticide) exposure scenarios. Provided foods are covered or removed prior to application of
propetamphos in food service establishments, chronic dietary (food) risk estimates for propetamphos are
not of concern to the Agency.
For chronic inhalation exposure resulting from the termiticide use, post-application inhalation MOEs
for children (48) and adults (150) are well below the inhalation target MOE of 300. Therefore, the
aggregate chronic risk estimate was not conducted and is of concern to the Agency. However, as indicated
in the previous section, this chronic inhalation risk assessment represents a conservative Tier I estimate of
exposure, and actual risks are expected to be lower.
5. Human Incident Reports
OPs as a group, have a well documented and disproportionately higher rate of poisonings than
other pesticides. The incident reports associated with propetamphos are disproportionately higher than
other pesticides used interiorly in both the number of indoor incidents reported, and in the number of
incidents involving PCOs. Incident reports from the following sources were reviewed for their potential
relationship to propetamphos exposure:
The OPP Incident Data System (IDS)
Poison Control Centers (PCCs)
The National Pesticide Telecommunications Network (NFTN)
California Pesticide Illness Surveillance Program (1982-1995)
Based on data from the NPTN, reported poisoning incidences involving propetamphos have
steadily declined between the years 1984 and 1998. Incidents where propetamphos was the only source
of exposure or where it was the only cholinesterase inhibitor and the symptoms were consistent with
cholinesterase inhibition were included. It is not clear at this point whether that decline is due to a lower
odor formulation or due to the reduction in usage of propetamphos products, or a change in use pattern.
However, three recent cases reported in California and submitted to EPA's Incident Data System (one in
July of 1999, two in March of 2000) suggest that offensive odor continues to be a serious problem for
propetamphos products. The specific cause of many of the reported effects from these incidents and others
could be odor, due to constituents other than the active ingredient.
In 235 out of 301 detailed descriptions of cases submitted to the California Pesticide Illness
Surveillance Program (1982-1995), propetamphos was used alone and was judged to be responsible for
the health effects. Only cases with a definite, probable or possible relationship were reviewed.
Propetamphos ranked 7th as a cause of systemic poisoning in California and 36th as a cause of
hospitalization. Non-occupational exposure and residue from structural applications was associated with
the overwhelming majority (88%) of the poisonings. Symptoms of these illnesses included difficulty
breathing, chest tightness, shortness of breath, mental confusion, nausea, dizziness, headaches, vomiting,
and eye irritation. Also common were cluster poisonings where large groups of office workers were
19
-------
exposed, poisonings due to workers returning to offices that did not receive proper ventilation, and
incidences where there was improper dilution by the applicator. Additionally, cluster poisonings have been
reported where there was no evidence of either poor ventilation or label violations. The total number of
poisoning cases related to structural pest control applications appears excessive when compared to the
extent of use. The main concern with propetamphos appears to be inappropriate use or misuse by PCOs
indoors. Most of the more serious poisonings appear to involve misuse, especially improper dilution,
application in enclosed spaces with bystanders present, inadequate ventilation of structures before
occupants are readmitted, and site inappropriate applications. A number of illnesses occurred despite the
apparent adherence to label directions. In some of these cases, it appears symptoms are brought on by
the offensive odor of the compound. This was supported by the finding that only one case out of 235
needed hospitalization. It should be recognized that individuals developing symptoms brought on by odor
effects are poisonings by definition. Cholinesterase depression, though a useful indicator for exposure, does
not have to be present to prove that poisoning has occurred. If odors are offensive enough to cause illness
and to seek medical attention, then the circumstances that lead to such morbidity should be examined so
that risk reduction measures can be identified and implemented
Poison Control Center data were obtained and reviewed for all pesticides for the years 1993-96.
This review reported on 199 exposures to propetamphos alone. Thirteen of the OP insecticides used in
residential settings were ranked on a variety of hazard measures. Propetamphos ranked in the top three
highest, and higher than any other OP except phosmet. Propetamphos ranked first for proportion of
exposures and symptomatic cases that were due to environmental residue. As with the California data,
Poison Control Center data suggests that propetamphos ranks high due to problems associated with
exposure to residues as a result of inappropriate use by PCOs.
In summary, propetamphos continues to rank high in the total number of poisoning cases related
to problems likely to be associated with exposure to residues and inappropriate use by PCOs, and appears
excessive when compared to the extent of use. In a nationwide survey of residential and commercial PCO
use, which estimated the total number of pounds of active ingredient of propetamphos applied indoors
compared to a total of 9,232,000 pounds active ingredient for all pesticides used indoors, propetamphos
accounted for only one percent of indoor use but accounted for 10 percent of the systemic poisonings.
B. Environmental Risk Assessment
Because all currently registered uses of propetamphos are limited to indoor use, exposure to
nontarget terrestrial and aquatic plants and animals is not expected. Therefore, no ecological risk
assessment was conducted for propetamphos.
20
-------
IV. Interim Risk Management and Reregistration Decision
A. Determination of Interim Reregistration Eligibility
Section 4(g)(2)(A) of FIFRA calls for the Agency to determine, after submissions of relevant data
concerning an active ingredient, whether products containing the active ingredients are eligible for
reregistration. The Agency has previously identified and required the submission of the generic (i.e., an
active ingredient specific) data required to support reregistration of products containing propetamphos
active ingredients. Appendix A identifies the use patterns eligible for reregistration that the Agency has
reviewed as part of its determination of reregistration eligibility of propetamphos.
The Agency has completed its assessment of the occupational and ecological risks associated with
the use of pesticides containing the active ingredient propetamphos, as well as a propetamphos-specific
dietary risk assessment that has not considered the cumulative effects of OPs as a class. Based on a review
of these generic data and public comments on the Agency's preliminary risk assessments for the active
ingredient propetamphos, EPA has sufficient information on the human health and ecological effects of
propetamphos to make interim decisions as part of the tolerance reassessment process under FFDCA and
reregistration under FIFRA, as amended by FQPA. The Agency has determined that propetamphos is
eligible for reregistration provided that: (i) current data gaps and additional data needs are addressed; (ii)
the risk mitigation measures outlined in this document are adopted, and label amendments are made to
reflect these measures; and (iii) the cumulative risk assessment for the OPs support a final reregistration
eligibility decision. Label changes are described in Section IV. Appendix B identifies the generic data
requirements that the Agency reviewed as part of its interim determination of reregistration eligibility of
propetamphos, and lists the submitted studies that the Agency found acceptable.
Although the Agency has not yet completed its cumulative risk assessment for the OPs, the Agency
is issuing this interim assessment now in order to identify risk reduction measures that are necessary to
support the continued use of propetamphos.
Based on its current evaluation of propetamphos alone, the Agency has determined that
propetamphos products, unless labeled and used as specified in this document, would present risks
inconsistent with FIFRA. Accordingly, should a registrant fail to implement any of ttie risk mitigation
measures identified in this document, the Agency may take regulatory action to address the risks concerns
from use of propetamphos.
At the time that a cumulative assessment is conducted, the Agency will address any outstanding risk
concerns. For propetamphos, if all changes outlined in this document are incorporated into the labels, then
all risks will be mitigated. But, because this is an IRED, the Agency will take further actions to finalize the
reregistration eligibility decision for propetamphos after assessing the cumulative risk of the OP class. Such
an incremental approach to the reregistration process is consistent with the Agency's goal of improving the
transparency of the reregistration and tolerance reassessment processes. By evaluating each OP in turn
and identifying appropriate risk reduction measures, the Agency is addressing the risks from the OPs in as
timely a manner as possible.
21
-------
Because the Agency has not yet completed the cumulative risk assessment for the OPs, this IRED
does not specifically address the reassessment of the existing propetamphos food residue tolerances as
called for by the Food Quality Protection Act (FQPA). When the Agency has completed the cumulative
assessment, propetamphos tolerances will be reassessed. At that time, the Agency will reassess
propetamphos along with the other OP pesticides to complete the FQPA requirements and make a final
reregistration eligibility determination. By publishing this interim decision on reregistration eligibility and
requesting mitigation measures now for the individual chemical propetamphos, the Agency is not deterring
or postponing FQPA requirements, rather, EPA is taking steps to assure that uses which exceed FIFRA's
unreasonable risk standard do not remain on the label indefinitely, pending completion of assessment
required under the FQPA. This decision does not preclude the Agency from making further FQPA
determinations and tolerance-related rulemakings that may be needed on this pesticide or any other in the
future.
s
If the Agency determines, before finalization of the RED, that any of the determinations described
in this IRED are no longer appropriate, the Agency will pursue appropriate action, including but not limited
to, reconsideration of any portion of this IRED.
B. Summary of Phase 5 Comments and Responses
When making its interim reregistration decision, the Agency took into account all comments
received during Phase 5 of the OP Pilot Process. As stated previously, a mitigation proposal was received
from the registrant, Wellmark International, a summary of which is outlined below. Several other comments
on mitigation were also received from the National Pest Management Association (NPMA), as well as
approximately thirty comments from commercial pest companies and other interested stakeholders. A
general summary of the majority of the comments received indicate a concern that propetamphos continue
to be available as one more additional tool in Integrated Pest Management (IPM) programs, where a
variety of chemicals are rotated to reduce potential resistance to any one type of chemical. Additionally,
most comments made the statement that propetamphos is particularly effective in the control of heavy pest
infestations, when other chemicals are not as efficacious.
Wellmark International's submission on proposed mitigation measures included the following:
cancel the restricted-use product Zoecon 8718 EW (EPA Reg. No. 2724-449)
amend the Catalyst end-use product label (EPA Reg. No. 2724-450) to state that foods must be
covered or removed during application in food handling establishments
• specify for Pest Control Operator (PCO) use only
• add personal protective equipment requirements
• conduct a 21-day dermal toxicity study in rats to refine the dermal NOAEL
The registrant also provided comments on data from the National Pesticide Telecommunications
Network (NPTN), suggesting that the decline in the number of reports from 1984-91 (35 calls per year)
to the later time period, 1995-98 (7 calls per year) is due to the introduction of a low odor formulation.
The original formula, Safrotin EC (EPA Reg. No. 2724-314), had volatile sulfides, which the registrant
contends were largely responsible for the adverse effects reported (i.e., nausea, headaches and eye
22
-------
effects). A new formulation replaced this product in 1995. However, the Agency believes the comparison
made between 1984-91NPTN data and 1995-98 data may not be appropriate. This information suggests
that there has been a recent decline in the number of propetamphos incidents, but may only represent a
decline in the number of propetamphos incidents reported, which may be the result of a change in reporting.
It is not clear at this point whether the decline in number of propetamphos incidents reported is due to a
lower odor formulation, a reduction in usage of propetamphos products, or a change in use pattern.
C. FQPA Assessment
1. "Risk Cup" Determination
As part of the FQPA tolerance reassessment process, EPA assessed the risks associated with this
OP. The assessment was for this individual OP, and does not attempt to fully reassess these tolerances as
required under FQPA. FQPA requires the Agency to evaluate food tolerances on the basis of cumulative
risk from substances sharing a common mechanism of toxicity, such as the toxicity expressed by the OPs
through a common biochemical interaction with the cholinesterase enzyme. The Agency will evaluate the
cumulative risk posed by the entire class of OPs once the methodology is developed and the policy
concerning cumulative assessments is resolved.
EPAhas determined that risk from exposure to propetamphos is within its own "risk cup." In other
words, if propetamphos did not share a common mechanism of toxicity with other chemicals, EPA would
be able to conclude today that the tolerances for propetamphos meet the FQPA safety standards. In
reaching this determination EPA has considered the available information on the special sensitivity of infants
and children, as well as the chronic and acute food exposure. An aggregate assessment was conducted
for exposures through food, residential uses, and drinking water. Results of this aggregate assessment
indicate that the human health risks from these combined exposures are considered to be within acceptable
levels; that is, combined risks from all exposures to propetamphos "fit" within the individual risk cup.
Therefore, for propetamphos, the tolerances remain in effect and unchanged until a full reassessment of the
cumulative risk from all OPs is completed.
2, Tolerance Summary
Propetamphos is not registered for use on plants (either food or feed crops). The only food or
feed-related use is the spot, and crack and crevice treatment of food service establishments. Tolerances
for propetamphos residues in food commodities exposed to the insecticide during treatment of food or feed
handling establishments are established at 0.1 ppm and are expressed in terms of propetamphos per se,
([(e)-]-methylethyl 3-[[(ethylamino) methoxyphosphinothioyl]oxy]-2-butenoate), [40 CFR §180.541].
The qualitative nature of the residue in food commodities is adequately understood based upon
metabolism studies examining the degradation of [C14] propetamphos in tomato juice, butter, bread, and
hamburger meat. Adequate analytical methodology is available for enforcing tolerances and collecting data
on propetamphos residues in food commodities. A gas chromatography/flame photometric detection
enforcement method for determining propetamphos on fruit, meats, milk, and vegetables is listed in the
23
-------
Pesticide Analytical Manual (PAM), Vol. n, as method I. The registrant also submitted a gas
chromatography/mass spectrometry detections method (GC/MSD) for tolerance enforcement. This
method has been successfully validated by the Agency. The validated limit of quantitation (LOQ) is 0.1
ppm and the limit of detection (LOD) is 0.01 ppm.
Reregistrationrequirements for magnitude of the residue in food handling establishments are fulfilled.
Adequate data (obtained using the GC/MSD analytical method) are available depicting residues of
propetamphos in representative food commodities (apples, beer, bologna, bread, butter, flour, hamburger,
lettuce, macaroni, milk, Rice Krispies®, and sugar) exposed, in open and closed containers on tables, to
propetamphos treatments reflecting the registered use pattern for food handling areas.
Tolerance Listed Under 40 CFR §180.541:
Registration requirements for data depicting residues of propetamphos in/on food commodities
following applications representative of the use in food handling establishments are fulfilled, and sufficient
data are available to ascertain the adequacy of the established tolerance for residues in/on food
commodities. The available data indicate that the current 0.1 ppm tolerance for residues of propetamphos
in food commodities is appropriate, based on the validated LOQ of the analytical method.
3. Endocrine Disrupter Effects
EPA is required under the FFDCA, as amended by FQPA, to develop a screening program to
determine whether certain substances (including all pesticide active and other ingredients) "may have an
effect in humans that is similar to an effect produced by a naturally occurring estrogen, or other such
endocrine effects as the Administrator may designate." Following the recommendations of its Endocrine
Disrupter Screening and Testing Advisory Committee (EDSTAC), EPA determined that there were
scientific bases for including, as part of the program, the androgen and thyroid hormone systems, in addition
to the estrogen hormone system. EPA also adopted EDSTAC's recommendation that the program include
evaluations of potential effects in wildlife. For pesticide chemicals, EPA will use FTFRA and, to the extent
that effects in wildlife may help determine whether a substance may have an effect in humans, FFDCA
authority to require the wildlife evaluations. As the science develops and resources allow, screening of
additional hormone systems may be added to the Endocrine Disrupter Screening Program (EDSP).
When the appropriate screening and/or testing protocols being considered under the Agency's
EDSP have been developed, propetamphos may be subjected to additional screening and/or testing to
better characterize effects related to endocrine disruption.
24
-------
D. Regulatory Rationale
The following is a summary of the rationale for managing risks associated with the use of
propetamphos. Where labeling revisions are warranted, specific language is set forth in the summary tables
of Section V of this document. The Agency has determined that the mitigation measures discussed below,
combined with additional amendments to the label, will reduce risks to workers, homeowners and children
to an acceptable level, and that unreasonable adverse effects are unlikely to result from such use. Provided
the following risk mitigation measures are incorporated into amended labels for propetamphos, the Agency
finds that all remaining registered uses of propetamphos are eligible for interim reregjstration, pending a
cumulative assessment of the OPs.
1. Human Health Mitigation Measures
a. Dietary (Food and Drinking Water) Risk
Acute Dietary (Food)
Acute dietary exposure and risk assessment is not necessary for propetamphos, a pesticide having
only food handling establishment tolerances. Therefore, there are no acute dietary (food) mitigation
measures necessary for propetamphos.
Chronic Dietary (Food)
The chronic dietary risk of propetamphos from food residues does not exceed the Agency's level
of concern, provided that language stating food be removed or covered prior to pesticide application is
added to the product labels.
Drinking Water
Because propetamphos is not expected to be released to water, exposure to drinking water is not
expected. Therefore, there are no drinking water mitigation measures necessary for propetamphos.
b. Occupational Risk
As indicated in Table 6, the ARIs are greater than 1.0 for all occupational use scenarios and are,
therefore, not of concern. These risk estimates are based on a reduced dilution rate of 0.5% ai solution
(from 1.0% ai), and applicators wearing personal protective equipment (PPE) consisting of a long-sleeve
shirt, long pants, shoes and socks, and gloves. Because PPE statements are not on the current
propetamphos label, the Agency has included as a mitigation measure that product labels be amended to
state that applicators must wear PPE consisting of a long-sleeve shirt, long pants, shoes and socks, and
chemical-resistant gloves. Additionally, to further mitigate these risks, the following measures are
necessary:
• Reduce the maximum rate of dilution to 0.5% ai solution.
Require that only protected handlers may be in the area during applications.
25
-------
The dermal exposure component of the occupational risk assessment is based on the recently
received 21-day dermal toxicity study in rats. Based on a preliminary review of the study, the Agency has
determined that the NOAEL =1.25 mg/kg/day based on brain cholinesterase inhibition at a LOAEL of 2.5
mg/kg/day. The Agency is currently conducting a final review of the study and is confident of its
determination and that it will be selected by the Agency's Hazard Identification and Assessment Review
Committee.
c. Residential (Post-Application) Risk
Broadcast Applications
As indicated in Table 7, the dermal MOEs for both adults and children are significantly below the
target MOE of 100. Incidental oral exposures (hand-to-mouth) for children is also below the target MOE
of 100. However, inhalation MOEs are above the target MOE of 300 for adults and children. Therefore,
the combined (ART) exposure from broadcast carpet treatment is less than 1.0 for all populations and of
concern to the Agency. Because these risk estimates are based on a chemical-specific exposure study,
the Agency has a high level of confidence in these exposure and risk estimates. Because of these risk
concerns, broadcast carpet treatment with propetamphos products shall be prohibited and removed from
the label.
Spot, and Crack and Crevice Applications
As indicated in fable 8, for crack and crevice/spot treatment, dermal MOEs for residential post-
application exposure to adults were above the target MOE. Therefore, an aggregate assessment with
chronic dietary (food) exposure was conducted and the resultant aggregate risks are not of concern. For
children, combined dermal and oral (hand-to-mouth) MOEs for all scenarios are below the target MOE
and are of concern (see Table 8). To mitigate these risks to children and other potentially sensitive
populations, the following measures are necessary:
• Cancel all residential uses.
Prohibit use in structures children and the elderly occupy, such as or including homes, schools, day-
cares, hospitals, nursing homes, with the exception of areas of food service within those structures,
when food is covered or removed prior to treatment.
Additionally, provided that a crack and crevice treatment meets the following application
restrictions (as defined in OPPTS 860.1460 Food Handling): "crack and crevice treatment is
application of small amounts of pesticides into crack and crevices in which pests hide or through
which they may enter a building. Openings of this type commonly occur at expansion joints,
between different elements of construction, and between equipment and floors. These openings may
lead to voids such as hollow walls, equipment legs and bases, conduits, motor housings, and junction
or switch boxes.", dermal and inhalation exposure and risk to persons re-entering the treated area is
expected to be negligible. To further mitigate these risks from non-residential uses, the following measures
are also necessary:
26
-------
Cancel all spot treatment applications and restrict its use to crack and crevice treatment only, as
defined in OPPTS 860.1460 Food Handling.
The product may only be used for crack and crevice treatment in food service establishments (e.g.,
restaurants, taverns, delicatessens, mess halls, mobile canteens, around vending machines, grocery
stores and.markets-where there is no contact with food) including schools, hospitals and nursing
homes in food service areas only; indoor non-food areas (e.g., office buildings, commercial, and
industrial premises and equipment); and non-food areas of eating establishments where there is no
contact with food, and where no food processing, packing, and no food and/or feed warehousing
occurs.
Termiticide Applications
Chronic residential inhalation exposure to propetamphos is possible because of the termiticide use
of this pesticide. Dermal or incidental oral exposure is not anticipated based on the use pattern (gallery
treatment). Based on the exposure assessment, chronic inhalation MOEs for adults and children are 150
and 48, respectively. This risk information is summarized in Table 9. Because the chronic inhalation MOEs
are below the target MOE of 300, the inhalation exposure from termiticide use of propetamphos is of
concern to the Agency. However, as discussed previously, this chronic inhalation risk assessment
represents a conservative Tier I estimate of exposure and actual risks are expected to be lower.
Consequently, the registrant has informed the Agency that it does not support the continued termiticide use
and has requested voluntarily cancellation of the termiticide use for propetamphos.
2. Environmental Risk Mitigation Measures
Because all currently registered uses of propetamphos are limited to indoor use, exposure to
nontarget terrestrial and aquatic plants and animals is not expected. Therefore, no ecological risk mitigation
measures are necessary for propetamphos.
E.
Label Amendments
Provided the following risk mitigation measures are incorporated in their entirety into labels for
propetamphos-containing products, the Agency finds that all remaining registered uses of propetamphos
would be eligible for reregistration, pending a cumulative assessment of the OPs. The regulatory rationale
for each of the mitigation measures outlined below is discussed in the previous section of this IRED. Also,
in order to remain eligible for reregistration, other use and safety information need to be placed on the
labeling of all end-use products containing propetamphos. For specific labeling statements, refer to Section
V of this document.
• Cancel all residential uses.
• Prohibit use in structures children and the elderly occupy, such as or including homes, schools, day-
cares, hospitals, nursing homes, with the exception of areas of food service within those structures,
when food is covered or removed prior to treatment.
27
-------
Cancel all spot, broadcast, and termiticide treatments.
The product may only be used for crack and crevice treatment in food service establishments (e.g.,
restaurants, taverns, delicatessens, mess halls, mobile canteens, around vending machines, grocery
stores and markets-where there is no contact with food) including schools, hospitals and nursing
homes in food service areas only; indoor non-food areas (e.g., office buildings, commercial, and
industrial premises and equipment); and non-food areas of eating establishments where there is no
contact with food, and where no food processing, packing, and no food and/or feed warehousing
occurs.
Amend the label to include the following crack and crevice treatment definition as defined in
OPPTS 860.1460 Food Handling: "crack and crevice treatment is application of small
amounts of pesticides into crack and crevices in which pests hide or through -which they may
enter a building. Openings of this type commonly occur at expansion joints, between
different elements of construction, and between equipment and floors. These openings may
lead to voids such as hollow walls, equipment legs and bases, conduits, motor housings, and
junction or switch boxes. "
Reduce the maximum rate of dilution from 1.0% ai to 0.5 % ai solution.
For food service establishment use, all food must be either covered or removed prior to application
of the product.
Applicators must wear personal protective equipment consisting of a long-sleeve shirt, long pants,
shoes and socks, and chemical-resistant gloves.
For use by Pest Control Operators (PCOs) only.
Only protected handlers may be in the area during applications.
28
-------
V. What Registrants Need to Do
A. Manufacturing-Use Products
1. Additional Generic Data Requirements
The generic data base supporting the reregistration of propetamphos for the above eligible uses has
been reviewed and determined to be substantially complete. The following confirmatory data in Table 10
are required:
Table 10. Confirmatory Data Requirements
s Guideline Test Name ,;" .
Dissociation Constant in Water
Partition coefficient (w-octanol/water), shake flask method
Stability to normal and elevated temperatures, metals, and metal ions
UV/Visible Absorption
i New Guideline No.
OPPTS 830.7370
OPPTS 830.7550-70
OPPTS 830.6313
OPPTS 830.7050
Old Guideline No.
63-10
63-11
63-13
none
Chemistry Studies
Pertinent product chemistry datarequirements remain unfulfilled for the Wellmark International 90%
T/TGAI concerning stability, pH, UV/visible absorption, and octanol/water partition coefficient (OPPTS
830.6313, 830.7370, 830.7050, and 830.7550-70). The registrant must submit the data required in the
attached data summary tables for the 90% T/TGAI, and either certify that the suppliers of beginning
materials and the manufacturing process for the propetamphos technical grade active ingredient (TGAI)
have not changed since the last comprehensive product chemistry review or submit a complete updated
product chemistry data package.
Neurotoxicity Studies
A Data Call-in (DCI) Notice has been sent to registrants of OP pesticides currently registered
under FIFRA (August 6, 1999 64FR42945-42947, August 18 64FR44922-44923). DCI requirements
included acute, subchronic, and developmental neurotoxicity studies. The Agency has received acceptable
acute (MRID 43403901) and subchronic (MRID 43403902 and 43995601) neurotoxieity studies,
therefore, the DCI referenced above only refers to the developmental neurotoxieity study for
propetamphos. After further consideration of the risk mitigation measures discussed in Section IV of this
IRED and other factors discussed below, the requirement for the developmental neurotoxicity study is
waived, provided the registrant complies with the necessary label amendments and annual Emit of 25,000
pounds of propetamphos active ingredient sold or distributed. If the registrant sell or distributes more than
25,000 pounds of propetamphos active ingredient within any calendar year, the registrant will be required
to submit to the Agency the developmental neurotoxicity study. The following factors were considered for
waiving these studies:
Based on the risk assessments and limited use pattern of propetamphos, there are no dietary (food
and water), occupational, or ecological risk concerns.
29
-------
• There is no evidence of neuropathology in the acute and subchronic studies; chronic dog study; and
no organophosphate induced delayed neurotoxicity (OPIDN) in the hen study. Also, there is no
evidence of increased susceptibility, based on adequate developmental toxicity and reproduction
studies. Therefore, the FQPA Safety Factor for propetamphos was removed (equivalent to Ix).
• The use of propetamphos will be restricted to (non-residential) crack and crevice only treatment
in food service establishments; indoor non-food areas; and non-food areas of eating establishments
where there is no contact with food, and where no food processing, packing, and no food and/or
feed warehousing occurs. All residential uses will be canceled, thereby significantly reducing
potential exposure to children.
• Provided propetamphos is restricted to PCO use for crack and crevice only treatment (excluding
baseboard and spot treatment applications), and because the low vapor pressure of propetamphos
(2.6 x 10"7 mm Hg at 25°Q will significantly limit the volatization of the compound, exposure to
persons re-entering treated areas is not expected to occur.
• Provided all foods are covered or removed prior to treatment of food service establishments, there
is no expectation of detectable residues on food.
• To assure that potential exposure to propetamphos does not increase significantly beyond current
levels, the amount of propetamphos active ingredient shall be limited to 25,000 pounds.
2. Labeling for Manufacturing-Use Products
To remain in compliance with FTFRA, manufacturing-use product (MUP) labeling should be
revised to comply with all current EPA regulations, PR Notices, and applicable policies. The MP labeling
should bear the labeling contained in Table 11 at the end of this section.
B. End-Use Products
1. Product-Specific Data Requirements
Section 4(g)(2)(B) of FIFRA calls for the Agency to obtain any needed product-specific data
regarding the pesticide after a determination of eligibility has been made. Registrants must review previous
data submissions to ensure that they meet current EPA acceptance criteria and if not, commit to conduct
new studies. If a registrant believes that previously submitted data meet current testing standards, then the
study MRED numbers should be cited according to the instructions in the Requirement Status and
Registrants Response Form provided for each product. A product-specific DCI, outlining specific data
requirements, accompanies this IRED.
30
-------
2. Labeling for End-Use Product
Labeling changes are necessary to implement measures outlined in Section IV. Specific language
to incorporate these changes is specified in the Table 11 at the end of this section.
C. Existing Stocks
Registrants may generally distribute and sell propetamphos products bearing old labels/labeling for
12 months from the date of the issuance of the RED document. Persons other than the registrant may
generally distribute or sell such products for 24 months from the date of the issuance of this interim RED.
However, existing stocks time frames will be established case-by-case, depending on the number of
products involved, the number of label changes, and other factors. Refer to "Existing Stocks of Pesticide
Products; Statement of Policy"; Federal Register. Volume 56, No. 123, June 26,1991.
The Agency has determined that registrants may distribute and sell propetamphos products bearing
old labels/labeling for 8 months from the date of issuance of this IRED. Persons other than the registrant
may distribute or sell such products for 18 months from the date of the issuance of this IRED. Registrants
and persons other than the registrant remain obligated to meet pre-existing label requirements and existing
stocks requirements applicable to products they sell or distribute.
31
-------
03
H
en
CJ
feC
^
"o3
Table 1 1 : Summary of Lab eling Changes for Prop etamphos
"3
3
c
0
•M
c
1
-2
P-4
Amended Labeling Language
s
o
••a
D.
b
o
P
Manufacturing-Use Products
V
z
»
^~}
J_
£
IS
_o
o
5
ie product may be used to formulate products for specific
(s) not listed on the MP label if the formulator, user group, or
wer has complied with U.S. EPA submission requirements
arding support of such use(s)."
ie product may be used to formulate products for any
litional use(s) not listed on the MP label if the formulator, user
up, or grower has complied with U.S. EPA submission
uirements regarding support of such use(s)."
^~| u o bo rj ^J o o"
s 3 b£l 1-1 s cdb!}^
0 o o
•S 3 D,
•21 §•
•T3 t- to
*Q ^* W
-------
--.
Table 11: Summary of Labeling Changes for Propetamphos
u
^
nJ
>•
3
u
1
S
w
Amended Labeling Language
JS
o
1
O)
Q
D
3
Tt
K
2 tn
£ "3
J £
5 Jj
'£ o
c -a
£ P
§ £
5 °
/3 „
?*•» C
G w
I i
11
B K
£ 2
•'This chemical is toxic to fish, aquatic invertebrates and other
wildlife, and poses a risk to reproduction of birds. Do not
discharge effluent containing this product into lakes, streams,
ponds, estuaries, oceans or other waters unless in accordance
with the requirements of a National Pollutant Discharge
Elimination System (NPDES) permit and the permitting authority
has been notified in writing prior to discharge. Do not discharge
effluent containing this product to sewer systems without
previously notifying the local sewage treatment plant authority.
For guidance contact your state Water Board or Regional Office
of the EPA."
•u
*2 3
£ S §
-s s-B
^03
ts Q ca
H « £
u e
J- .S (D
1 'Personal Protective Equipment (PPE)
Mixers, loaders, applicators, and other handlers must wear:
Long-sleeve shirt, long pants
• Shoes plus socks
Chemical-resistant gloves" (registrant inserts correct
chemical-resistant material)
Note: PPE that is established on the basis of Acute Toxicity oft
end-use product must be compared to the active ingredient PPE
this document. The more protective PPE must be placed in the
product labeling. For guidance on which PPE is considered mor
protective, see PR Notice 93-7.
•o
CL>
.52
1
1
-2
fi ED
11
u -o
•- 2
P £
(S ~
ao cr
^ !*j
o °
O Q
•- §
u ""
£ £•
"c u
« -s
e M
ca ,S
3 o £3
^* *^ S
o
"?
1
u.
OJ
S
CO
-------
|
Summary of Labeling Changes for Propetai
f— 4
O
3
a
H
•g
tJ
s
o
'S
i
t»
*2
P-
Amended Labeling Language
s
0
•c
.2-
u
S
Q
c em
1 •§
o o
1-2
N 0
ca a> ^^
EC M ^2 X
... en C x
v v~ A) ^^
j3 ca Si „
— P E C3
**-« .S ^ d,
•^ C «S _Q
TO O K] "~
0 P OT w
111 I
S -S5 •§
a a 1 s_
Recommendations
wash hands before eating, drinking, chewin
o, or using the toilet.
remove clothing/PPE immediately if pesticic
wash thoroughly and put on clean clothing,
remove PPE immediately after handling this
:side of gloves before removing. As soon a;
>h thoroughly and change into clean clothinj
>, 30 ^g 23«
oi 3 pa "5 ,ri "3 ® ^
e3 ^*° 2E"1 oa>^
CO w 2 w . w "S --^
w oc o3^2 ow^
p ^"3 r^c ^?S°
en
C
.2
Cd
TD
U
E
H
o
u
(S
C3
CO
fe
to
D
CO
^
<§
co
e
0
22
5
•d
-------
a
e
o
0 T3
rt
2 S
11
II is
2 § 2
3 c t>
o '
o
s E
~ a>
CA
1
o
u
c a
£ s
C "O
I g
' W3
>:= £
,-e 2
o tS
c£ M '> c1
g.ss>:§
| I|I
« "1 a {C
•S
"flj »*
,1 i
&
W M
'S t>5 3
111
U g 00
•Hi
?-S 2
•o
o
c oo
0^3
111
SI*
«J C 0>
^o g -a
>. __ "~'
co
JJ S
o J3
2=0-0
•S J3 J3 G
T3
0>
" £
: &
4-» i-l
•2 -Ss
oo S c
.5 ^ "
-a .S
•S C •> S
S S
l
S M
T3
C
CO
' -a
I
o
c
•§
o
T3
C
&S c
§••£.•£ g
O c« G
o o o o
O JS C C
em
c
O
_o
*C
&.
•a
u
o
u c
-O J3
s
< S
a -3
0 s »
•a M o
C IK Cu
« O j-
111
211
<2 O. co
T-l CX (D
^ « o
M co >.
3 « S
"Go
•e s-S
^ s c
G « =8
° a •*
—* i3 O
"2^2
43^0
III
!•§ s
o. « 3
en ^ .«
J3 c3 w
r"Q_S
3 S
>>g .
*^ S M
W ,Q> ^
sfc 8
<*- T3 C
0 S -a
w c
g> § §
•il g
s.-°.l
O S .o.
£?£,!
2 § g
I'll
«J CX
V- X J5
S " -a
a
u _
^3 §
1§
|l
§•&
u C
_«" o
11
£3
&-^s
J3 C i3
•*-• O co
•s I §
i i-
5 o O
^•g
2 *
_C en
M 00
« C
J3 "5
u 3
s.§
•S S
11
•S.f
« 3
u -a
II
a s
§>s-.
•a-° s
u T3 X
a, c o
o "»-g
S
C/3
u. a> *5 c
?s g €
"o
o.
D.
o
-------
V)
mary of Labeling Changes for Propetampho
E
3
..
*••<
.2
3
c2
"3
a
c
o
B
o
g
U
_«
a.
Amended Labeling Language
e
o
•c
a.
1
O
Q
&>
CO
t-l
,2 "o
^0 =
c ™
O 2
'S -2
J-i ^
o a
£ ^
.s g
E E
O
i ?
S S 'io
co T3 c
.£2 § 2
•£ E 0
o o 't-
0 '•" M
E 1 (S
flj • n "*"* "
CA W C3 ^ ^ E
c D •*-• ~™ 5t r*" *C
.« jy o o cL s
J3 *tS QJ w « >* ci<
•g-2 >2 c° g^>
II fe II-. J | ii
E" "= U-C~ a EE
gu °S*:0 >•« ^'M
Go "c7 '> 'c '^ '•= P ~S a
-C" O Ptacj |^ fc^ «>.2
S| fc. If I ^ ^ 11
!l i ni l i 11
.§ | s « s -g B | a^
£» S. -s >••« -g 5 s^
aj -^- /™-i 4-t -t— i tu ^Z y bri •-• ™ S«fc U S?
•oo o-o.Sc3'a t -^S
»C "O ^ •!-• ^5 •—• ^ C Qi c? ^
.So S si « -° S^ss: Gc
— J- PH ^W^ ^SkC- O1^
w a* *" • «O* ^* 5 S *••" *^ tt O "O c
||| | 1 s.j 1 J.|| li
^•sta o 5^"E -a Itv^ 5-5
=0
QJ
3
B
'E
o
^
to
E
0
'o
^Q
P*
E
O
u
"o.
o.
^
2
u
(U
o
(L
•Z
£
2^
^
c
"t
j:
c
a>
c
CA
C
O
O
1
,
X
£
u
5
l_l
K
«•>
en
B
O
tis represent the exact language that must app<
ke to amend their labels or product registrati
o
a
3
3
o-
S
CO
3
s
M
.£ a
an J3
- w
3
a
3.
a
3
a
3
U
q
3
J
bJ
™
-------
VI. Related Documents and How to Access Them
This IRED document is supported by documents that are presently maintained in the OPP docket.
The OPP docket is located in Room 119, Crystal Mall #2,1921 Jefferson Davis Highway, Arlington, VA.
It is open Monday through Friday, excluding legal holidays from 8:30 am to 4 pm.
The docket initially contained preliminary risk assessments and related documents as of January
15,1999. Sixty days later the first public comment period closed. The EPA then considered comments,
revised the risk assessment, and added the formal "Response to Comments" document and the revised risk
assessment to the docket on December 1, 1999.
All documents, in hard copy form, may be viewed in the OPP docket room or downloaded or
viewed via the Internet at the following site: "http://www.epa.gov/pesticides/op."
If any of the conditions of this interim decision are not satisfied, including but not limited to the
submission of an unacceptable study, missing established deadlines, or failing to amend product labels, the
Agency may take other regulatory actions. If the Agency later determines (based upon consideration of
the cumulative assessment) that any of the determinations described in this DRED are no longer appropriate,
the Agency will pursue appropriate action, including but not limited to, reconsideration of any portion of
this IRED.
37
-------
38
-------
VII. APPENDICES
39
-------
40
-------
e
o
•fl
C3
S
fi
OS
PL!
"S
c
1
ft
M-
di
—
JS
s
.- ta S
•— j~ C
s (2 "
O W3
!Z c
.9
C "™"
1 &
s <
1 *
3 ^,
_C _2; (D
S .s 3
S W .
rt ' p,
•^
c ni
.2 £
^ bb
|<
F"^,
g;^
c* 'S
^ 0
.2 s
jrt "S
o -^
-— • OH
&
**•
CO r
P
.W
ervice Est
oo..
T3
1
co~ CD
C R 0
Limit re-treatment intervals to not more than 2
treatments per 30 days.
For indoor, non-residential crack and crevice
treatments only for the following use areas:
food service establishments (e.g. restaurants, taver
delicatessens, mess halls, mobile canteens, around
vending machines, grocery stores and markets whei
there is no contact with food, and when food is
removed or covered prior to treatment), including
schools, hospitals and nursing homes in food servi
areas only;
VI
C >>
a a
•S "°
0 .S
£ i
\-. O OJ
o \=. -o
Z M 3
For indoor, non-residential crack and crevice
treatments only for the following use areas:
indoor non-food areas (e.g., office buildings;
commercial; and industrial buildings and warehous
and institutions, except those where children and rt
elderly occupy, such as and including schools, day
cares, hospitals, and nursing homes); and
non-food areas of eating establishments where then
no contact with food, and where no food processin
packing, and no food and/or feed warehousing occ
CO
J-g
0 .S
O r-
IZ o
i! »
flj 03 ^^
o s ~a
E S; o
^c^ «
c
_o
Jjj
o
CO
'S
^
0
o1
il
CO f^
1— 1 t — I
v- a>
'?- ^ 1 i a
.§ | S H t|
^ 2 "§ N Si--
o ."ti rt o c 'p*
"c ^ "S e 'S &
•S ^, s '1 ^ i
2 S S3 .2, "° «
O & Q..S 0 to
-------
42
-------
Appendix B. Table Of Generic Data Requirements And Studies Used To Make The Interim
Reregistration Decision
GUIDE TO APPENDIX B
Appendix B contains listing of data requirements which support the reregistration for active
ingredients within case #2550 (propetamphos) covered by this Interim RED. It contains generic data
requirements that apply to propetamphos in all products, including data requirements for which a "typical
formulation" is the test substance.
The data table is organized in the following formats:
1. Data Requirement (Column 1). The data requirements are listed in the order in which they
appear in 40 CFR part 158. the reference numbers accompanying each test refer to the test
protocols set in the Pesticide Assessment Guidance, which are available from the National
technical Information Service, 5285 Port Royal Road, Springfield, VA 22161 (703) 487-
4650.
2. Use Pattern (Column 2). This column indicates the use patterns for which the data
requirements apply. The following letter designations are used for the given use patterns.
A. Terrestrial food
B. Terrestrial feed
C. Terrestrial non-food
D. Aquatic food
E. Aquatic non-food outdoor
F. Aquatic non-food industrial
G. Aquatic non-food residential
H. Greenhouse food
I. Greenhouse non-food
J. Forestry
K. Residential
L. Indoor food
M. Indoor non-food
N. Indoor medical
O. Indoor residential
3. Bibliographic Citation (Column 3). If the Agency has acceptable data in its files, this column
list the identify number of each study. This normally is the Master Record Identification
(MIRD) number, but may be a "GS" number if no MRED number has been assigned. Refer
to the Bibliography appendix for a complete citation of the study.
43 .
-------
APPENDIX B
(OLD/NEW GUIDELINE) REQUIREMENTS
OLD
NEW
STUDY
USE PATTERN
CITATION(S)
Product Chemistry
51-1
61-2A
61-2B
62-1
62-2
62-3
63-2
63-3
63-4
63-5
63-6
63-7
63-8
63-9
63-13
63-17
63-20
71-1
71-2A
72-1A
72-1 C
72-2A
81-1
81-2
81-3
830.1550
830.1600
830.1670
830.1700
830.1750
830.1800
830.6302
830.6303
830.6304
830.7200
830.7220
830.7300
830.7840
830.7860
830.7950
830.7370
830.7550
830.6320
Chemical Identity
Start. Mat. & Mnfg. Process
Formation of Impurities
Preliminary Analysis
Certification of limits
Analytical Method
Color
Physical State
Odor
Melting Point
Boiling Point
Density
Solubility
Vapor Pressure
Stability
Storage stability
Corrosion Characteristics
IX^LOGICAL EFFECTS
830.2100
850.2200
850.1075
850.1075
850.1010
870.1100
870.1200
870.1300
Acute Avian Oral -Quail/Duck
Avian Dietary - Quail
Fish Toxicity-Bluegill
Fish Toxicity Rainbow Trout
Invertebrate Toxicity
„.,, TOXICOLOGY , «,
Acute Oral Toxicity - Rat
Acute Dermal Toxicity -Rabbit/Rat
Acute Inhalation Toxicity-Rat
ALL
ALL
ALL
ALL
ALL
ALL
ALL
ALL
ALL
ALL
ALL
ALL
ALL
ALL
ALL
ALL
ALL
41607414
41607414
41607414
42355803
42355802
42355803,
42355804
41607411
41607411
41607411
41607411
41607411
41607411
41607408
41607416
42254701
41997304,
41607402
41997304
" x „.„•*>" » * ' "* '
ALL
ALL
ALL
ALL
ALL
^t^^, ^.f ,^ ^ ^ ^ ^^
ALL
ALL
ALL
00097891,
41607401
42144701,
42144702
41607409
41607415
41607401,
41607404
,:l •""•-';_-
41607417
41607418
45198401
41529301
44
-------
OLD
81-4
81-5
81-6
81-7
82-1B
82-2
83-1A
83-2A
83-4
84-2A
84-2B
85-1
81-8
85-4-SS
160-5
171-2
171-4E
NEW
870.2400
870.2500
870.2600
870.6100
870.3150
870.3200
870.4100
870.4200
870.3800
870.5140
870.5375
870.7485
870.6200
^Jone
>Jone
None
860.1380
STUDY
Primary Eye Irritation -Rabbit
Primary Dermal Irritation-Rabbit
Dermal Sensitization-Guinea Pig
Acute Delayed Neurotoxicity - Hen
90-Day Feeding - Non-rodent
21 -Day Dermal -Rabbit/Rat
Chronic Feeding Toxicity-Rodent
Oncogenicity - Rat
2-Generation Reproduction - Rat
Gene Mutation (Ames Test)
Structural Chromosomal Aberration
General Metabolism
Acute Neurotoxicity Study
6-Mo Ocular Toxicity Study
Chemical identity
Chemical identity
Storage Stability
USEPATTERN
ALL
ALL
ALL
ALL
ALL
CLMNO
CLMNO
CLMNO
ALL
ALL
ALL
ALL
ALL
ALL
ALL
ALL
ALL
CITATION(S)
41607419
41607420
41607412,
42194401
42194401,
92150013
00039596
00052920,
00052921
42399001
42399001
43039801
41607405
41607406
42978201
43403901
43049501
41607414
41607414
43193303
45
-------
46
-------
Appendix C: Technical Support Documents
Additional documentation in support of this Interim RED is maintained in the OPP docket, located in Room
119, Crystal Mall #2,1921 Jefferson Davis Highway, Arlington, VA. It is open Monday through Friday,
excluding legal holidays, from 8:30 am to 4 pm.
The docket initially contained the preliminary risk assessments and rekted documents as of September 23,
1998. Sixty days later the first public comment period closed. The Agency considered comments on the
revised risk assessments and added the formal "Response to Comments" document and the revised risk
assessment to the docket on September 24, 1999.
All documents, in hard copy form, may be viewed in the OPP docket room or downloaded or viewed via the
Internet at the following site:
www.epa.gov/pesticides/op
47
-------
48
-------
Appendix D. Citations Considered To Be Part Of The Database Supporting the Interim
Reregistration Eligibility Decision (Bibliography)
GUIDE TO APPENDIX D
1. CONTENTS OF BIBLIOGRAPHY. This bibliography contains citations of all studies considered
relevant by EPA in arriving at the positions and conclusions stated elsewhere in the Reregistration
Eligibility Document. Primary sources for studies in this bibliography have been the body of data
submitted to EPA and its predecessor agencies in support of past regulatory decisions. Selections
from other sources including the published literature, in those instances where they have been
considered, are included.
2. UNITS OF ENTRY. The unit of entry in this bibliography is called a "study". In the case of
published materials, this corresponds closely to an article. In the case of unpublished materials
submitted to the Agency, the Agency has sought to identify documents at a level parallel to the
published article from within the typically larger volumes in which they were submitted. The resulting
"studies" generally have a distinct title (or at least a single subject), can stand alone for purposes of
review and can be described with a conventional bibliographic citation. The Agency has also
attempted to unite basic documents and commentaries upon them, treating them as a single study.
3. IDENTIFICATION OF ENTRIES. The entries in this bibliography are sorted numerically by
Master Record Identifier, or "MRED number". This number is unique to the citation, and should be
used whenever a specific reference is required. It is not related to the six-digit "Accession Number"
which has been used to identify volumes of submitted studies (see paragraph 4(d)(4) below for
further explanation). In a few cases, entries added to the bibliography late in the review may be
preceded by a nine character temporary identifier. These entries are listed after all MRID entries.
This temporary identifying number is also to be used whenever specific reference is needed.
4. FORM OF ENTRY. In addition to the Master Record Identifier (MRID), each entry consists of a
citation containing standard elements followed, in the case of material submitted to EPA, by a
description of the earliest known submission. Bibliographic conventions used reflect the standard of
the American National Standards Institute (ANSI), expanded to provide for certain special needs.
a. Author. Whenever the author could confidently be identified, the Agency has
chosen to show a personal author. When no individual was identified, the Agency
has shown an identifiable laboratory or testing facility as the author. When no author or laboratory
could be identified, the Agency has shown the first submitter as the author.
b. Document date. The date of the study is taken directly from the document. When the date is
followed by a question mark, the bibliographer has deduced the date from the evidence contained in
the document. When the date appears as (19??), the Agency was unable to determine or estimate
the date of the document.
49
-------
c. Title. In some cases, it has been necessary for the Agency bibliographers to create or enhance a
document title. Any such editorial insertions are contained between square brackets.
d. Trailing parentheses. For studies submitted to the Agency in the past, the trailing parentheses
include (in addition to any self-explanatory text) the following elements describing the earliest known
submission:
1) Submission date. The date of the earliest known submission appears immediately
following the word "received".
2) Administrative number. The next element immediately following the word "under" is the
registration number, experimental use permit number, petition number, or other administrative
number associated with the earliest known submission.
3) Submitter. The third element is the submitter. When authorship is defaulted to the
submitter, this element is omitted
4) Volume Identification (Accession Numbers). The final element in the trailing parentheses
identifies the EPA accession number of the volume in which the original submissions of the
study appears. The six-digit accession number follows the symbol "CDL," which stands for
"Company Data Library." This accession number is in turn followed by an alphabetic suffix
which shows the relative position of the study within the volume.
50
-------
Appendix D
PROPETAMPHOS BIBLIOGRAPHY
MRID Number
00039595 Hamburger, F.; Carpy, S.; Klotzsche, C.; et al. (1979) San 52. 139 I: 6-Month Feeding
Study in Dogs: Report No. TOX 21/79. (Unpublished study received May 8, 1980 under
11273-EX-19; prepared by Sandoz, Ltd., submitted by Sandoz, Inc. Crop Protection, San
Diego, Calif.; CDL:242461-B)
00039596 Klotzsche, C.; Carpy, S.; Luginbuehl, H. (1978) Propetamphos (San 52.1391): 13-Week
Feeding Study in Rats: Report No. 24/77. (Unpublished study including report 47/78,
received May 8,1980 under 11273-EX-19; prepared by Sandoz, Ltd. and Univ. of Bern,
Institute for Animal Pathology, submitted by Sandoz, Inc. Crop Protection, San Diego,
Calif.; CDL:242462-A)
00052919 Goldenthal, E.I.; Wazeter, F.X., Geil, R.G.; et al. (1976) Three Week Dermal Study in
Rabbits: IRDC No. 163-373. (Unpublished study received May 5, 1976 under 876-252;
prepared by International Research and Development Corp., submitted by Velsicol Chemical
Corp., Chicago, 111.; CDL:228723-G)
00052922 Goldenthal, E.I.; Wazeter, F.X.; Geil, R.G.; et al. (1975) Fourteen Day Inhalation Toxiciry
Study in Rats: IRDC No. 163-334. (Unpublished study received May 5,1976 under
876-252; prepared by International Research and Development Corp., submitted by Vesical
Chemical Corp., Chicago, 111.; CDL:228723-J)
00063021 StoU, R.E. (1980) Interim 18 Month Report on Lifetime Oral (Diet) Carcinogenicity/Toxicity
Study in the Mouse on San 52-139: Sandoz Project T-1220; WIL # 79218. (Unpublished
study received Nov 24,1980 under 11273-22; prepared in cooperation with WIL Research
Laboratories, Inc., submitted by Sandoz, Inc. Crop Protection, San Diego, Calif.;
CDL:243800-B)
00085152 Hamburger, F.; Klotzsche, C. (1978), Safrotin (R) 50 EC: Primary Skin Irritation in Rabbits:
Agro Dok CBK 3155/78. (Unpublished study received Nov. 1, 1978 under 11273-21;
prepared by Sandoz, Ltd., Switzerland, submitted by Sandoz, Inc. Crop Protection, San
Diego, Calif.; CDL: 235623-H)
51
-------
00085153 Klotzsche, C., Hamburger, F.; (1978), Safrotin (R) 50 EC: Primary Skin Irritation in
Rabbits: Agro Dok CBK 3154/78. (Unpublished study received Nov. 1, 1978 under
11273-21; prepared by Sandoz, Ltd., Switzerland, submitted by Sandoz, Inc. Crop
Protection, San Diego, Calif.; CDL: 235623-1)
00085154 Hamburger, F.; Klotzsche, C. (1978), Safrotin (R) 50 EC: Diluted for Use: Primary Skin
Irritation in Rabbits: Agro Dok CBK 3153/78. (Unpublished study received Nov. 1,1978
under 11273-21; prepared by Sandoz, Ltd., Switzerland, submitted by Sandoz, Inc.-Crop
Protection, San Diego, Calif.; CDL: 235623-J)
00085155 Hamburger, F.; Klotzsche, C. (1978), Propetamphos: Primary Skin Irritation in Rabbits:
Agro Dok CBK 3152/78. (Unpublished study received Nov. 1,1978 under 11273-21;
prepared by Sandoz, Ltd., Switzerland, submitted by Sandoz, Inc.-Crop Protection, San
Diego, Calif.; CDL: 235623-K)
00085156 Leuschner, F.; Leuschner, A.; Klie, R.; et al. (1978) Two-weeks toxicity of Safrotin in
Sprague-Dawley Rats when Administered by Inhalation. (Unpublished study received Nov 1,
1978 under 11273-21; prepared by Laboratorium fur Pharmakologie und Toxikologie, West
Germany, submitted by Sandoz, Inc. Crop Protection, San Diego, Calif.; CDL:235623-L)
00085157 Hartman, H. A.; Hrab, R.; Buechle, P.;et al. (1978) San 52-139: Investigation of Teratogenic
Potential in the Rabbit: Exp. #T-1183. (Unpublished study, including letter dated Oct. 17,
1978 from H. A. Hartman and R. Hrab to R. J. Van Ryzin, received Nov. 1,1978 under
11273-21; submitted by Sandoz, Inc. Crop Protection, San Diego, Calif. CDL: 235623-
M)
00085158 Sandoz, Incorporated-Crop Protection (1978) Fish & Wildlife: Safrotin 4 Emulsifiable
Concentrate Insecticide]. Summary of studies 235623-O and 235623-P. (Unpublished
study received Nov 1,1978 under 11273-21; CDL:235623-N)
00085159 Morrissey, A.E. (1978) The Acute Toxicity of Propetamphos (92% Pure) to the Water
Flea-Daphnia magna-Straus: UCES Proj. No. 11506-16-01. (Unpublished study, including
letter dated Sep 27,1978 from R.E. Stoll to R.J. Van Ryzin, received Nov 1,1978 under
11273-21; prepared by Union Carbide Environmental Services, submitted by Sandoz,
Inc.-Crop Protection, San Diego, Calif.; CDL:235623-P)
00097891 Fink, R.; Beavers, J.B.; Brown, R. (1978) Final Report: Acute Oral LD50 Mallard Duck:
Project No. 131-105; Sandoz Project T-l 177. (Unpublished study received Nov 1, 1978
under 11273-21; prepared by Wildlife International, Ltd. and Washington College, submitted
by Sandoz, Inc.-Crop Protection, San Diego, Calif; CDL:235623-Q)
52
-------
00102928
00117996
00142110
00164890
41529301
41581201
41607401
41607403
41607404
41607405
StoU, R.; Adamik, E.; LeQuire, M.; et al. (1982) Final Report on: Lifetime Oral (Diet)
Carcinogenicity/Toxicity Study in the Mouse on SAN 52-139: WIL-79218; T-1220.
(Unpublished study received May 12, 1982 under 11273-22; prepared in cooperation with
WIL Research Laboratories, Inc. and Toxpath Services, Inc., submitted by Sandoz,
Inc.-Crop Protection, San Diego, CA; CDL:247482-A; 247483; 247484; 247485-
247486; 247487;247488)
Bagdon, R.; Heilman, J.; Krause, R.; et al. (1978) 8 Weeks Preliminary Toxicity (Dose
Range Finding) Study of Propetamphos in Mice: Project T-1217. (Unpublished study
received Nov 5,1982 under 11273-22; submitted by Sandoz, Inc., Crop Protection San
Diego, CA; CDL:248795-D)
Eschbach, B.; Klotzsche, C. (1984) Propetamphos: Teratogenicity Study in Rats: Agro Dok
cbk I. 6058/84. Unpublished study prepared by Sandoz Ltd. 110 p.
Luginbuehl, H. (1980) Propetamphos: Chronic Feeding Study in Rats: Project No.: 279.
Unpublished study prepared by Sandoz Ltd., Basle. 2424 p.
Carpy, S. (1984) Propetamphos Technical Grade: 4-Hour Acute Inhalation LC50
Determination in Rats: Lab Project Number: AGRO DOK CBK 1.5909/8. Unpublished
study prepared by Sandoz Ltd. 43 p.
Huang, F. (1988) Propetamphos Dislodgeability Study Report: Lab Project Number:
88-820-0400. Unpublished study prepared by Mid west Regional Chemistry
Laboratory/Environmental Science & Engineering, Inc. 71 p.
Burgess, D. (1990) Acute Flow-through Toxiciry of Propetamphos Technical (...) to Daphnia
magna: Final Report: Lab Project Number: 38677: 1422. Unpublished study prepared by
Analytical Bio-chemistry Laboratories, Inc. 25 p.
Marshall, R. (1990) Study to Evaluate the Potential of Propetamphos to Induce Sister
Chromatid Exchanges (SCE) in Cultured Chinese Hamster Ovary (CHO) Cells- Lab Project
Number: SAD 2/SCE; 2CSRESAD.002. Unpublished study prepared by Microtest
Research Ltd. 35 p.
Bussard, J. (1990) Method Validation for the Analysis of Propetamphos in Aquatic Test
Water: Final Report: Lab Project Number: 38674. Unpublished study prepared by
Analytical Bio-chemistry Laboratories, Inc. 15 p.
Clare, C. (1989) Study to Determine the Ability of Propetamphos to Induce Mutations to
6-Thioguanine Resisitance in Mouse Lymphoma L517Y Cells using a Fluctuation Assay: Lab
Project Number: SAD 2/ML; 2MLRESAD.002. Unpublished study prepared by Microtest
Research Ltd. 29 p.
53
-------
41607406 Marshall, R. (1989) Study to Evaluate the Chromosome Damaging Potential of
Propetamphos by its Effects on the Bone Marrow Cells Treated Rats: Lab Project Number:
SAD 2/RBM; RBMRESAD.002. Unpublished study prepared by Microtest Research Ltd.
35 p Battelle, Columbus Laboratories. 20 p.
41607407 Kazee, B. (1990) N-octinol/water Partition Coefficient of Propetamphos: Lab Project
Number: 1456. Unpublished study prepared by Battelle. 14 p.
41607408 Kazee, B. (1990) Solubility of Propetamphos: Lab Project Number: 1457. Unpublished
study prepared by Battelle. 13 p.
41607410 Kazee, B. (1990) Dissociation Constant of Propetamphos: Final Report: Lab Project
Number: SC900059: 1454. Unpublished study prepared by Battelle. 10 p.
41607411 Schweitzer, M. (1990) Physical Characterization of Propetamphos: Final Report: Lab Project
Number: SC900060; 1455. Unpublished study prepared by Battelle. 12 p.
41607412 Wilkinson, G.; Singer, A. (1990) Delayed Contact Skin Hypersensitivity Study of
SAN139I90 TC (Propetamphos Technical) in the Guinea Pig: Lab Project Number:
SC900075; 1444. Unpublished study prepared by Battelle, Columbus Laboratories.
22 p.
41607416 Dublaski, A. (1990) Determination of the Vapor Pressure of Propetamphos: Lab Project
Number: BE-P-106-90-A04-01; 1458. Unpublished study prepared by Battelle-Institut E.
V. 17 p.
41607417 Wilkinson, G.; Singer, A. (1990) Acute Oral Toxicity Study of SAN 139190 TC
(Propetamphos Technical) in the Rat: Lab Project Number: SC900071; 1440. Unpublished
study prepared by Battelle, Columbus Laboratories. 27 p.
41607418 Wilkinson, G.; Singer, A. (1990) Acute Dermal Toxicity Study of SAN 139190 TC
(Propetamphos Technical) in the Rabbit: Lab Project Number: SC900072; 1441.
Unpublished study prepared by Battelle, Columbus Laboratories. 29 p.
41607419 Wilkinson, G.; Singer, A. (1990) Primary Eye Irritation Study of SAN 139190 TC
(Propetamphos Technical) in the Rabbit: Lab Project Number: SC900073; 1442.
Unpublished study prepared by Battelle, Columbus Laboratories. 20 p.
41607420 Wilkinson, G.; Singer, A. (1990) Primary Dermal Irritation Study of SAN139I90 TC
(Propetamphos Technical) in the Rabbit: Lab Project Number: SC900074; 1443.
Unpublished study prepared by Battelle Columbus Laboratories. 15 p.
54
-------
41841402 Allen, T.; Corney, S.; Janiak, T.; et al. (1991) 52-Week Oral Toxicity (Feeding) Study with
SAN 52.1391 Technical Grade in the Dog: Lab Project Number: 226912. Unpublished
study prepared by Research and Consulting Co., AG.; in cooperation with RCC (U.K.) Ltd.
and EPS (U.K.) Ltd. 522 p.
41997101 Schweitzer, M.; Summer, S. (1991) Method Development and Validation of Propetamphos
Residue Analysis in Food Commodities: Final Report. Lab Project Number: SC900078.
Battelle. 154p
41997102 Neslund, C. (1991) Confirmation of the Tolerance Enforcement Method for Propetamphos
Residue on Food Commodities: Final Report: Lab Project Number: 2985: 1618.
Unpublished study prepared by Lancaster Labs., Inc. 178 p.
41997103 Rudolph, R. (1991) Propetamphos Residue in Representative Food Commodities Resulting
from Exposure to Safrotin 1% Aerosol: Lab Project Number: 1452:
R256SAN139I1AE-RES
42144701 Fink, R.; McCormack, R. (1979) LC50 Determination of Propetamphos in the Mallard
Duck: pFinal Reporto: Sandoz Project No. T-1389: WI Study No. 131-112; Report
T-l-10/12/79. Unpublished study prepared by Wildlife International Ltd. 39 p.
42144702 Fink, R.; McCormack, R. (1979) LC50 Determination of Propetamphos in the Bobwhite
Quail: Final Report: Sandoz Project No. T-1390; WI Study No. 131-111;
42275801 Ferdinandi, E. (1991) Metabolism, Mass Balance of Radioactivity and Plasma
Pharmacokinetics of [carbon 14]-Propetamphos in Male and Female Sprague-Dawley Rats
Following its Oral Administration: Lab Project Number 38804: 1532. Unpublished study
prepared by Bio-Research Labs., Ltd. 427 p T-2-10/12/79. Unpublished study prepared
by Wildlife International, Ltd. 24 p.
42144701 Fink, R.; McCormack, R. (1979) LC50 Determination of Propetamphos in the Mallard
Duck: Final Report: Sandoz Project No. T-1389: WI Study No. 131-112; Report
T-l-10/12/79. Unpublished study prepared by WUdlife rntemational Ltd. 39 p.
42144702 Fink, R; McCormack, R. (1979) LC50 Determination of Propetamphos in the Bobwhite
Quail: Final Report: Sandoz Project No. T-1390; WI Study No. 131-111;
42254701 Clark, A. (1992) Stability for Propetamphos: Lab Project Number: 6449-F: 1715.
Unpublished study prepared by Midwest Research Institute. 24 p.
42275801 Ferdinandi, E. (1991) Metabolism, Mass Balance of Radioactivity and Plasma
Pharmacokinetics of [carbon 14]-Propetamphos in Male and Female Sprague-Dawley Rats
55
-------
Following its Oral Administration: Lab Project Number: 38804: 1532. Unpublished study
prepared by Bio-Research Labs., Ltd. 427 p T-2-10/12/79. Unpublished study prepared
by Wildlife International, Ltd. 24 p.
42355801 Burleson, I; lhada, S. (1992) Propetamphos: Manufacturing Procedure and Beginning
Materials: Lab Project Number: 9005 SAN 139191 TC. Unpublished study prepared by
Nippon Kayaku Co. Ltd. and Zoecon Corp. 91 p
42355802 Reuter, K.; Burleson, J.; Kayaku, N. (1992) Discussion of Impurities of Propetamphos
Technical: Lab Project Number: REF 4500/KRE/RC: 9005 SAN 139191 TC. Unpublished
study prepared by Sandoz Ltd., Nippon Kayaku Co. Ltd., and Zoecon Corp. 23 p.
42355803 Ko, J.; Nguyen, J.; Lewis, S.; et al. (1992) Analysis and Certification of Ingredients and
Impurities in Five Separate Batches of Propetamphos Technical Material: Lab Project
Number: 1864. Unpublished study prepared by Zoecon Corp. 72
42355804 Ko, J.; Nguyen, J.; Lewis, S.; et al. (1992) Precision and Accuracy for Current Analytical
Procedures, CAP 315, CAP 341, CAP342, and CAP 344. Used to Analyze Components
of Propetamphos Technical Material: Lab Project Number: 1863: 9005 SAN 139191 TC.
Unpublished study prepared by Zoecon Corp. 64 p.
42399001 Fresh, R. (1992) Dietary Analysis Data, MRID 164890: Propetamphos: Combined Chronic
Toxicity/Carcinogenicity Study to the Rat (Supplement): Lab Project Number: I. 5214/81.
Unpublished study prepared by Zoecon Corp. 16 p.
43039801 Eschbach, B.; Aemi, R.; Hopley, J.; et al. (1991) Propetamphos: Two Generation
Reproduction Study in Rats: Final Report: Lab Project Number: 442 R: 1309: BS2238.
Unpublished study prepared by Sandoz Agro Ltd. 943 p..
43049501 Garg, R.; Weber, K.; Allen, T. (1993) Data to Support the Ocular Toxicity Requirement of
Propetamphos Technical in Dogs: Lab Project Number: RCC 226912: ZOECON 1627.
Unpublished study prepared by RCC, Research and Consulting Co. AG; RCC,
Umweltchemie AG; RCC (UK) Ltd.; and EPS (UK) Ltd. 452 p.
43193301 Lephart, J. (1994) Response to United States Environmental Protection Agency Letter,
October 28, 1993, Regarding Propetamphos Residue Studies (Part 1: Method
Development): Supplement: Lab Project Number: 1538. Unpublished study prepared by
Sandoz Agro, Inc. 13 p.
43193302 Lephart, J. (1994) Response to United States Environmental Protection Agency Letter,
October 28, 1993, Regarding Propetamphos Residue Studies (Part 2: Method
56
-------
Confirmation): Supplement: Lab Project Number: 1618. Unpublished study prepared by
Sandoz Agro, Inc. 13 p.
43193303 Lephart, J. (1994) Response to United States Environmental Protection Agency Letter,
October 28,1993, Regarding Propetamphos Residue Studies (Part 3: Residue Quantitation):
Supplement: Lab Project Number: 1452. Unpublished study prepared by Sandoz Agro, Inc.
13 p.
43403901 Minnema, D. (1994) Acute Neurotoxicity Study of Propetamphos (Technical) in Rats: Final
Report: Lab Project Number: HWA 777-140: 9005. Unpublished study prepared by
HazletonWashington, Inc. 397 p.
43403902 Minnema, D. (1994) Subchronic Neurotoxicity Study of Dietary Propetamphos (Technical)
in Rats: Final Report: Lab Project Number: HWA 777-141: 9005. Unpublished study
prepared by Hazleton Washington, Inc. 488 p.
43890201 Cannon, J. (1995) Evaluation of Propetamphos Degradation in Four Food Matrices: Lab
Project Number: 3861: 0924-167: 2207. Unpublished study prepared by Midwest Research
Institute. 71 p.
43995601 Minnema, D. (1996) Subchronic Neurotoxicity Study of Dietary Propetamphos (Technical)
in Rats: Supplemental Information: Lab Project Number: HWA 777-141. Unpublished study
prepared by Hazleton Washington, Inc. 10 p.
44150501 Andrews, K. (1996) Method Development and Validation of Propetamphos Residue
Analysis in Food Commodities: Lab Project Number: SC900078: 1538. Unpublished study
prepared by Battelle. 9 p.
45198401 Findley, J.; Wacek, B. (2000) Wellmark International, 21-Day Dermal Toxicity Study of
Propetamphos (Technical) In Rats; Lab Project No. IITRI: 1313 SN2. Prepared by HT
Research Institute (HTRI). 352 p.
92150004 Garg, R. (1990) Zoecon Corporation Phase 3 Summary of MRID 00029976. Acute
Delayed Neurotoxicity Study in the Chicken on SAN 52-139; Sandoz Project T-1447,
Bio/Dynamics 6182-79. Prepared by Bio/dynamics Inc. 13 p.
92150013 Ben-Dyke, R. (1990) Zoecon Corporation Phase 3 Reformat of MRID 00029976. Acute
Delayed Neurotoxicity Study in the Chicken on SAN 52-139, Sandoz Project T-1447,
Bio/Dynamics 6182-79. Prepared by Bio/Dynamics Inc. 183 p.
92150015 Hartman, H.; Hrab, R.; Buechle, P.; et al. (1990) Zoecon Corporation Phase 3 Reformat of
MRID 00085157. San 52-139: Investigation of Teratogenic Potential in the Rabbit: Sandoz
Project T-l 183. Prepared by Sandoz, Inc. Teratology Laboratory. 115 p.
57
-------
58
-------
Appendix E. Generic Data Call-In
See attached table for a list of generic data requirements. Note that a complete Data Call-in (DCI), with all
pertinent instructions, is being sent to registrants under separate cover.
59
-------
r*
4-
C
t_
a
d
fr
2
n.
M-t
o
u
E-<
rt|
&
Q
4 ^
' H
n
5 ^ p, *
o m Q
1*H O 4)
0 0 rl
*O 1 1 -rt
> ? gg 8-
dp o o sa
ft MM
Pi ,—|
Gt * (0
P
£ 1 1
u
a
1
§
•H
$IH
Lj ^ O
W
^ S
i i
States Environment
Washington, D
DATA CALL-IN
•d
/it
(D
-H
S
B
1
Q
•H
JS
AJ
C
o
•o
01
AJ
a
01
IT
K
§
-rl
JJ
m
o
U-t
C
•H
JS
JJ
1
TJ
rue t ion o
M
•jj
0
c
•H
'O
V
X
u
m
jj
o
.c
rH
rH
M-l
S
15
U
•O
a
S
a
D
0)
rH
a*
4
•rl
JJ 63
C Q
•a s
0, U
• i
K
O JJ
S £
01 £
rH Q
o.
rH
.. (8
INSTRUCTIONS
Ise addition
rf iJ
H
Q
U4
O
1 ° '
1 *
? s
" w
Q V™ /
n
CO
o
rC H
(0 ro
4J H
Q) H
Si
15 C
2 15 a
0
*O 4fc .C
15 "irt «
„. O O 5
3
C 01 m
(5 Oi J3 AJ
DOC
Q. M 10 11
tD Al E D
M 41 U 4) -
C AJ -H C
15 01 3 (8
Q UJ AJ 01 A)
-H m ai ta
-H C * -iH
A) A) 0 Dl
O (5 T3 4)
3 01 m 01 B!
8O C A) T) i
JJ 0) -H C -
01 01 C S
>, 01 rl U 01 C
E rl -H 30
Dl 3 E Al S,
• 15 D1 M 15 01
43 O 0 AJ 0)
r-H r< 1 4^ D1 H
01 tu AJ 0) AJ
-H 01 D{ TO
-H C * *H
AJ AJ O Dl
O 15 -O 01
3 01 HI 01 U
'O AJ rH
0 O C AJ T3 «
S A) 01 -ri C -
Qj E Al 15 01
41 0) C 01
>. 01 r< 01 0 C
E r) -ri 30
Di 3 e AI a
- (5 D1 fi 15 O
U 0) O AJ 01
r- H r. «w w K
U TJ
H O 'O
H jj d)
U (0 i-H
C O i)
P -H -rl
9 T3 JJ T3
. C C C *
>i*rf 4) 0 •
J 0)
m a s o to
H m R 3 c
•U 0 4J 0
(B CD *W flj ft
to jj jj a
C T3 CO CJ
2 15 0*
• s s^ -
01 -ri Al 01 Al
H 3 Al E c
Dl (T 15 01 15
15 01 r< r<
rl 0) -H AJ
H J3 3 O
15 AJ D"-ri
• Al 01 Dl
J3 15 C CU 01
vo Q 0 « K
. I am claiming a Generic
ta Exemption because I
tain the active ingredient
om the source EPA regis-
ation number listed below.
a s> vi n
Q O M-l Al
i i
0 C
O -H 3
i) O CD rH
•SSB?
*«SI^
H O -O U -H
C p « ^
• 15 S H (8
in o Qi u 4J
u
-§§
83
0. «
< i)
Oi Q
ta -H
Di
- 0»
•» o:
1
1
AJ
s
o->
sS
c c
rH O
1-5
O Pi
cT
2 >,
u c
"1
» (5
01 £
3 01
irm and all attachments are tr
lleading statement may be puni
:epresentative
u ra u.
•W -iH
e -o
a 01
•H H N
-C O -ri
tements made on t
knowingly false
le law.
Company's Author
« >..Q V
AJ C 15 O
m 15 o
J> •" rH rH
c 5 5 a -H
O -P (B H
•H 4J
•U (5 fl) rl T3
O JJ T3 -O OS
H 4) C
*W >nrH 3 0
H IW S V,
ti 3 § 5 3
01 rl i! 0 18
J 1) U .0 C
U (0 D)
rl -H
03 H H 0 W
-D
i
S
£
§
£
Al
U
15
AJ
c
S
16
E
U
U-l
0
o
H
-------
o
O
H
r- t-
o m ro
iH O 4>
o o ^
'D I 1 -H
4) O O O,
> t- r- X
0 O O W
i-i n n
& • nJ
B J 1
0 E ft
>i U
U CQ
d ^S
S S
tn PI
< TO
H
C OS
O
•H TO
4-1 v
O o 2
ft ™ TO
H
>H rj CP
fd . W
^ p PS
Ig 1
O jj
•H O P
> -H H
si fl
(U •* TO
4-) EH
ftf Z
CO §
0) H
4-) p
^^
g
M-l
0)
•H
**
§
•o
01
01
3
O1
S
-H
JJ
o
c
-H
5
rH
a
3
Q
IB
Q
§
id instruct!
HI
JZ
U
re
jj
jj
re
01
rH
rH
3
S
rt
o
13
flj
4)
^
Q)
ID
(0
4)
On
i
•H
JJ (D
C m
•H (Q
}-t 4>
p, y
4)
0
0, "H
jj "o
41 JJ
to 4)
4) .C
rH (Q
CU
INSTRUCTIONS:
Use additional
H
u
Q
tw
O
|u
•O OH
18 s
" W
Q vD
•
H
to o
o <^>
rd n
E *~~i
ft5
a! 8
a §
1^ 1 §
•0 * .C
m rH E
:* o S Jj
in -g j5,
g cQ 1 1
«
JJ
2
H
(Q 4)
•H 0)
« S
Qj Q,
(0
o;S
i
S 1
co ft
4)
O
JJ C
a re
4) JJ
B tQ
,Q
• 3
f- CQ
fi
4) S
tQ 0)
AJ
VO Oi
r,
to
Dl 0 M
g it
iH
a,o!OHOUOJ
0)
rH
-H
H
*u
JJ
CQ
•
4 . Guideline
Requirement
Number
'. • :• ••x''!:.:
tQ tQ IQ tQ
O.;;0 00
PC EE
CN fSl
JJ. O ft
0 -rt . JT
n -u o
O -H V ffl
C ••:.«
0 • & fl>
•rl" . , >-»
(0 '4) jJ • Iff
•H i> -H fll
8flJ i-l -H
JE -H >
0) ^*v St
0 4J S • >
-r) O JJ -X.
GO CQ 13
in
o
o -H ro r*-.
H t-4 H -
II I O
ro ro ro ro
4>
(0
Q
rH
H
JJ
* C
4) 4)
§jj e
8 cL
TJ -H
c
(D -
41
- C
41 -H
B N
8-°
U 01
«rH
- m
(0 -H
rH JJ JJ
rH Q (0
10 JJ
Dl C
'DC W <
re >i o
e rH
01 >,
0 C • B
JJ -H & IB
c 3 re ft
0) O r-H £
S C 0
0) ^ 0) U
JJ rH
IB >, J3 U
jj c re o
Q re u
•rl 0)
0) JJ rH rH
C £ rt ft jJ
O JJ .C ft -H
•H jj re H
10. Certificat
I certify that
I acknowledge
"or both under
Signature and
H
|
2
o>
c
S
m
JJ
u
IB
JJ
U
C
i*
S
U-t
0
I
TH
-------
62
-------
Appendix F: Product Specific Data Call-In
See attached table for a list of product-specific data requirements. Note that a complete Data Call-in (DCI),
with all pertinent instructions, is being sent to registrant under separate cover.
63
-------
H
m
o
,_)
0)
tn
rt
CM
^H
O
O
&
Q
r^N-
otn
OO
-p i i
Q) OO
O OO
u rucM
E z
o s:
11 Q
0
a
(U
"5i
o
•r)
JJ 0
O VD
0) <* W
4J 0 CO
O OQ 15
M O
CM • CU
CJ CO
nS • w
C D fe
(1) ~ H
° aJ ^
•H CJ S
W m Ej
tti rt!
to !S Q
(U
aJ
nJ
CO
T)
Q)
oJ
•H
B
E
CO
p
c
o
s
1
t_
c
Q
ro
Q
QJ
.e
^"
g.
w
1
CO
c
o
4J
U
3
t_
tt
tc
CU
JC
u
ro
+•*
ro
JC
4-*
_x
H—
CU
t.
CD
U
•s
o
t.
o
CA
CD
01
a.
j±
C
4-> CO
C U
'L! cu
°- o
o
M-
0) 4?
U) O
CD (U
CU J=
— • M
a.
• • CD
u> c.
I-B
J-^ •—
§3 ^
in cu
~ =
CJ
H
« H
*** QJ
o ^^
& Cj '
i— t"i
T2 ^
a p
£8
O fit
•
ro
tn
o
I
0)
a
§ o
=tt O
0, jn
3 CN
CXI
0
O
CO 0
CO 0
w °
S !H PC!
2! S O
2 rt; Q ><
"H 0 EH
» 0 H -
g W >-i
« pq pc! EH
c J H H
g-CU CO CJ
|-§ o o
*"
CD
4->
CD
O
O
M-
8
CL
V)
4-»
U
"S
c_
o_
^
CO
4-*
ro
u
cu
g
u
CO
c a. j: 4-;
ro ID u c
(X 4-» E CO
i % « .2 ^
CO H- 4-* CU 4-»
r- CO O£ CO
4-» 4-< O ~ O)
o ro TJ Q)
3 U) CO QJ OS
O O C 4J TJ S
D. *"* § '-M ro QJ
cu a3 c co
>. OJ U CU CO C
•E *- — 3 O
D) 3 E 4-> CL
• ro cr t. n co
ro cu o 4-< QJ
N- »-i U H- W OS
O "O
.*- oj "D
I- 4-» O
QJ (Q _s
C U 4-i
cu .— •—
CD "U 4^ "O
>*.p- cu ro "•
co co E to co
4-* O 4-* O
ro w H- co o.
« 4-» 4-> CO
C TJ C/) QJ
O QJ ft) rv
4-> E JC W
CU U 4-* 03
cu t- ro c *.
CU IP- 4-f OJ 4-»
L. 3 4J E C
ro cu L. t-
^- f 3 CO
ro *-• a-—
« 4-» QJ D)
JD ra c a: o
^ O O S gv
O C 3
•p- QJ I O
C ""* QJ 5)^3
QJ QJ L. QJ
CD Cfl O) t- TJ
3 c: cu
ro ro — *
-P- .P- QJ
E C 4-" O t-
.5 0 0 t- QJ
ro -p- ro 3 -Q
— ' 4-» O g
CJ Q. QJ W 3
§ *-•
Is
o •—
o- ro
• c
4-1 E
QJ c;
— ' 0
a. co
o t!
1"".
- c
QJ *p~
ro
t_ >x
u
u cu
11
O JC
3 0)
c_ ••-
•u c
o) a.
,00
4-* X
QJ E
1 c
S §
ro 4-> >
ro •—
— ' M ro
;? s
C -p- W
« TJ QJ
ro t—
E cu CL
C ' QJ
O CO O£
»'E "S
'*- L. H
•5 ° 'L
o o
CM -C
"Sx f
'm V-
CO C • C
4J •»- 3E CO
C 3 JO Q.
O H O O
+•> — '
CD X JO H-
4J C CO O
CO CO U
— O
O 4J — > —
JC CD Q. 4->
C V JC Q. •-
O *"» CD h-
4-t CO O t_ *O
CO JC 0) O C
U *-* "T3 "O CD
'•^ X-2 § 0)
•p- M- 3 £.
4-> — O JC 3
1- •*-» C 4-* 4->
CJ O O JD C
O CD O)
• L. ••-
oo "- •— o tn
1
3
o
Q
1.
^
U
CD
C
o
u
X
c
CD
U
O
i
ro
z
o
-------
r>-i"-
om
oo
o oo
t- CMCM
<" 0
sc
L. m
O E
Ll_ o
>1 M
0 CO
g §
CD &
<; co
w
F4 rt
-H CO
4J O •
Q) *# fe
4-) O 2
Orvi ^c
i \| U>|
P[ • QJQ
Ul_l
nn
rH O
_J ful
fO ' pM
i-) Q CA
^— f—t ^^
C
£ pj S
S fi iff
O 4J
H tT> CO
"s! c E
? "'H t^
W CD EH
ri co
m js
a) co
5 I
co S
(U H
-H O
fl W
£3 S
E
o
H-
U)
£
4-*
g
1
01
01
g
^
0
c
0)
JZ
V)
1
01
g
(J
o
L.
C
g
u
TJ
a
;
J
^
Hi
U
D
U
3
U
o
0
(D
OJ
L
.
:
t_
4-* CD
C 01
- W
L U
CJ
I- C
o
H-
J «-
K *"*
j 4_»
U) JC
-- w
• CD
in c
s o
-* 4-»
3 TJ
3 T3
£ CD
0 Q>
z w
fe
b^
CJ g
i t Cj
~HQ
° O PH
° W 1
|"|
! B s
£ O *
ra # P
0 CM H
to
f|
g
i
03 z
O S
ro
4-) •
D< S3
a o
(Q (j ^
Z dl |J)
"E (u
% 0 ^
Q Li) W|
% in cu
S OJ H
rJ
O
O
CO 0
CO O
01 H °
O £*j Q
-^ |2 Q brf
3 cu ^ ><
1 0 E-"
o w »
i H >n
c J En H
JS-CM CQ o
CO 2
|§ 0 0
*~
4-*
C
(0
4-1
VI OJ
o> c
iS Q.
. 0
O* (X
Is
• L,
CO u.
4- 1
01 CO
V 4-1
1^ tn
3 £
• CD
*O EX
to
OJ -M
t_ t. —
O) Q l>l
o S.
CL. K.
*"
xo:oi— ouo— i
i—
t
m
u
C 4J
— C
"D w
••- L_ i_
3 •— O
-* o: z
• •* ,-• . . '. * :< •''*' *•••*•
m m 03 tn to to ra to ra ra m
oo o o o oooooo
eg € g S g g g g .;g g
00 -" ."
08 4J Q. S t- E .01 4J
!• ffl t- D • L. jv -^
jj • . >-. s _« .0. *• .0 ^ « —
2 •- 't *• •*; '•'*•• ^ "- ;™ a)
9 4J O O O M- '-fij E ro 4J
S CO • :O Bl '•" -S . 10
ae ' u .c o • c c o>,_4J ™
TJO 3 O O C •- t? C Z
I ,»•-.«»- "D •— « H- (0 o 4J 10 TJ O
_ 4-> O *-> M 4J Ul «— •*- C S E '— '
B 4-« .CL t- , CL Ot CL Q) 01 C- ••- ••- CU -CO
8 0 ••" CL -^ 0 — 0 W 3 E M- 0 :5
X^ 3 , t- t. O L. O 3 ZL '^ •»- L. £_ •-•ri
° • "o tn ° o) ^" ;° ° '* *•
TS t_ :0) . OJ OJ >p- t— QJ C O _T! "D
o Q- d o a Q o. o LJJ c_> n. o
a. ' "•• ' • • . •' •
o o o o o: o o o o) n ^
in o oj in c^ o tn o o o o
in \o u> vx) -U) c~- i> oo n n n
rH H H H H :H H rH U) VO <>D
•'• .. • '• •• • • -«. »...,.
; : o o o o o oooooo
fi ro n ro ro rorororororo
oo oo oo oo oo ao oo oc> oo ao oo
01
CO
^
*~
4-*
Q) QJ
4-* E
^ ^
tco
L.
0 g.
^ "~
to •>
- c
4-> **-
ra
u
U 0)
OJ — '
-Q
*- (D
0) _C
4-* 'c
0, S.
L.
tO Q)
CO
i io"
CJ E
.C 4-*
U C
CD <1)
4J CD aj
CO 4J >
CO •«-
—•01 CO
CO 4-1
TJ C S
C "- w
CD C-
E CU CL
O Ol - f^
"*" 'I TJ
CO QJ
^ O ••-
OJ o
C 01 ^
O — • 4J
CO 3
ro _>. to
"•Sj §
C X CD D.
E C ""* O
S ^ o> cj
ro >. S H-
4J C CO 0
01 CD U
OJ 4-* —J — *
- -C CD Q. 4J
o 4-> j: CL •—
•- 4J CD I—
0 CO 0) t- TJ
j f D> Ol C
- — O .c 3
J 4-* C 4-> 4«*
CJ t- ^ O CD
QJ O .Q EZ
• U CO Dl
O 1- •.-
«-•—•-• o
-------
ro
,. .
O
f>j
0)
tn
rf
o<
X
CM
0
O
EH
h
«!
tf
P
N-N.
28
•g II
o oo
L- CM CM
! *
E to
£ §
t^ M
O CO
d g
Q) O
tn CM
rtj CO
0 §
o
-H CO
4-J O *
O VO &H
Q) ^ 2
4J 0 S
O O3 P^
M EH
O4 • CO
O H
tH O
(ti • W
OJ P K
d
(L) - P
£ d 3
O 4J
s-i tn co
">-H EH
w m EH
fti CO
m !s
O CO
4-) EH
(rf pZ
4J M
co S
•ct 3
Jj P
•H O
d M
ID rt
•
5
CO
^
3
3
jj
u
I
g
i
o
5
01
c
CO
•g
ro
(/)
g
«»—
u
t_
c
^
OJ
1
CO
•5
jx
•=;
H-
01
L.
CO
o
"8
01
Ol
CO
Ol
a.
4
"~
•M CO
C CO
•r- CO
t- at
01 oi
t- C.
o
«*-
Ig
to cu
CO 0
CU .C
— • CO
a.
z o
o —
I-* +•*
CJ "O
a: co
H~
co at
g
{^i
o g
fa
„ H P
3 OS
o W i
v ra g
^ EH g
"g B i
« p
i § p
0 Oi H
ro
g
g
t
m g
O S
Pi S
C P-J
ro
OJ
Q) O
D4 a
i o
CD (^ •
z Oi tn
•g Q)
co p^
* 0
cu m rf;
g m Oi
S M W
B
CVJ
0
o
CO O
CO 0
W fc^j f^J
(0 ?H P5
ills
1 0 EH
0 U W -
sy w >>H
«? W Cd EH
c i-3 EH H
g- Oi CO O
O. 5*1 ,-~N .'^
& ri; O O
cj CO p? S
^
c
CD
C_
CO CU
'ro c
cu o
a: a.
• cu
0 a:
,
10
• L.
CO u-
0)
4-* C
1— W
N- C/3
V) 0)
• ro
O O.
O 4-*
D) O ^^
0 0.
I_ Q — —
°" ^ «-
XOiO*—OCJO.J
o
+J
1—
t
4-»
M
in
gw
••- c
"O 5
'5-g fc
rj- "g
««»
^
m ro ro m co m ro m m ro ro
ooooo °°i§§§
£ £ c S £ SSEScc
CO CO CO 00 CO CO CO rCO : 00 ;CO : CO
£u CM Oi CM CM CM CM CM CM CM CM
wwwww wwwwww
ssssi i si sis
OOOOO OOOOOO
3§§3§ 333333
H"j ^O ^"3 ^D ^O ^^ ^3 ^"D ^-5 *D ^3
MHHHH HHHHHH
*TJ |T! JT! ^rl ?T| ^ri ?T! ^r! h j( j ji.} |j ]
r*5 >*h fi^ rh r ri rij rft f^j r^j f'j r^
|V| pt| p][^ pT| jjn yV[ pt| ^^ pn fT[ pLt
JiJfiltilpLjJxJ pt) p£] ^ pp {.xl PL]
88888 888888
§§§§*! §§§§§§
r* ir> o ••'•*- «M ; . sf • .'- ui
S"1 ° *s
ro a
§*" <0 4J i=
.C ^ 0 g
0 V > -C
c ^o u
•*•* ° v. >. B -S ^S
O O ••— **•" t0 €0 5
jQ T3 "~ >•* ~*^* (0 &- fiC
CO 41 O in *•* jf^ J^1 J3
£. a ^ it- £OO g
^O O O' O *!}* 'tn VO .C> • O\ ;O rH
:o in o o :i-t 1-1 H rt 1-1 :IN tN
;O O -rrl: ro ro ro ro TO ro co ro
:i>..c^ t> P*:iD \o u> «J VD v0 .•;*- O O (D
'o *x c L- W
*- 'ig c >. a> c
"~ «
flj , ,^> fl) . Qt d)
3 ? 3 g 3 •£
5 S 5 ^ S CO
O O O O O O
o O o o o o
H tN ro ^ in ^D
H H H O5 O3 CS1
oooooo
D1^ 'C** C^* f^1 C** ^^
00 CO CO 00 00 CO
fl)
CD
Q
Si
D)
D.
to
•£
g
g
•4-)
i
o •
1?
o
o
*j Ol
O)
CO CO
CO CL
gg
•M W
(0 —
H! o
L. (D
O U
O •••*'
M-
0) ..-
(0 t—
u o>
K
O 4-**
-M X
_* °*
CO
— "3
" w
-------
N.N.
otn
oo
o oo
1 ~~
« 0
g
1_ CD
o z
u. o
>i M
O CO
£H (25
(U O
Cn ft
< CO
a 5
-H CO
4_) O *
Q) ^ a
4-) 0 2
8 ^ P
ft • CO
OLJ
^1
•H C5
(tf • W
4J P K
Q) «. P
C d S5
C O 2
O 4J
k Cn CO
-H « p
^ "H 2
W CD E-l
(d co
ra |s
4J EH
4-> W
CO S
•d 3
Q) H
-H O<
S M
D «
o
H-
CO
!c
v*
g
1
cn
01
§•
u
c
o
4J
I_
O
c
"
.c
1
cn
CD
cn
C
o
*j
E
1
1
is
CD
CU
.c
**
>»
careful 1
CD
U
L.
CU
cn
CD
Q.
jj
C
L.
•M CO
C cn
•- cn
r Q
Q. u
cu
I- C
o
H"
<-• cn
» (l)
CD OJ
CU f
— ' cn
o.
n c
il
I s
~ =
o S
H^5
f£i
— H P
° W ^
§-W §
ffr'Z
a Q
2 o*
ra Pi P
0 ft H
to
g
g
g
g
i
O 2
D, 1
g
4J
Pi g
I o
^ ft Cn
"g OJ
% o
cn LD ft
o CN W
4
CM
O
O
CO O
co o
cn W °
2 a P
!l^
<° O H -
t— 1 t^ tl
^ ft CO U
ill o o
Q CO S S
*-
4->
c
ip
£_
4->
••- to
D> C
Q) O
C£ Q.
. S
O* &f
i
•t~ (U
. c_
CO U.
QJ
CO CO
0) -M
• 3
r^-
cu IE
cn cu
• CD
*O D_
to
cn
cu •>->
i_ t. _.
D) O "^
O Q.
r> r*r
3.0:01— ouo—i
CU
1—
1
in
8*.
•— c
— • 0)
0 E
^3 Q)
• •-£_£_
3 .c- 0)
"11
stf- O£ X
• «
to m :T*
i i
co oo
ft ft
H W
O O
?! 1
:" •-
O
"\ ^
^. ! ** u"»
W) * ' ••
f M- ~-
% <- <-
O1 ^x ^x
"o
:^*3 ' '
CJ «
- .• cu
-4) >^ 4J
*J . / O "
a co . 13
S m ." "o
*5 S >t '>T
t. S tu c
g 1 t r -^
- 2 s § s
1 I- CD — ' C.
' 1- . CD CD
to {0 co o
O «j- _ 1 o
£ :S
^- . i. ' •
Ul 0.
." V H H
H • H
.1 i
:••.. i:in •. .in
: ; X
cu
(0
M — '
~~
-------
(N
O
H
01
nJ
0
a
QJ
*§
C
Ow
-H
O 0
0) VD
aJ «4<
O o
M tN
•
rH 0
rt
rt Q
(U
is
H 4->
•H tn
> a
H!E!
03
m n5
(U !S
rt
u
CO
•a
0)
•H
,§
:*
PL,
O
O
Ct!
Q
CO
i
M
g
3
F5
H
§
§
[TIONS
1
b
M
P
B
i
w
1
1
o
en
o
%
PJ
S
jj
§.
M
P4
0
in
in
aJ
i
•g
ro
0)
9
u
U) ~ ' ' ~~~~~~ —— -^— — ^— —
a 4J
4J C O — •
... a 4^ co • ">
o e 4J — ' o "D ""
£ c-ctotj— cu co
L. L. ooj4JcuC4J ro
O "- M- E C 4J .C 4J _Q •
.t f 1 K c, 2 J -| 2 1 - H
S ° "° *~ "*~ -2 fu 3 o> « 4Jjo
wS'^o'co in »0 eu x I* 43 cu oo
coo •- T-.-eucji.fi _ 13
S •" ra "g £ * « fc c -° £ ~ - t
C >- = o » t- ••-••.• SL c "TOO
>. -u n „ *£ £ t. o w in o -M -j i£ "in
G) O 4-' CO O C~ i *— ••— C ro u «^ Si
jco< co v o 1-1 w *~ * o -0 ro 2 £" ro
-Q«-^crociE oi>^3"po) *rr«
•.3 «^ «£ u. t-^ £_ C"U"PCI(UC OO
SWO~ iii«£ 'StoiSro'0^ 'ito*
3O4J UJ -3 O UJ — » O •» -rrtO> I— C)
SCS S -0 £ CM ^ - 1 ^ «C
CU*»- O M-Q.ICOCDO COCD
"Dt~"D °" 0)CNJ°OC<: T3W
CUCOD) Q.Ct£ 4Jt^— *4J'«-O "~-5
J3 43 c • o t. co — • c 13 M-CO
coc--.i-efl C34J3eueuoj 'S; ni
• P-CO CJ^ •l-4^*'-M-4^t-13 COO)
D)L.tU I- OE COCD CO -•
0 4J > T3 Z t- tO — t- — C C- CO •;-
I-CO— D.O g ^>
L. »» "cj O ••"• UJ "D O "D C, O C ~^ *
H- 4^4-> t ' •. CO 4J (0—*
CWC Q.-'ZW tOO)OO4-»OC Jl»2
"" '5» .2i z c £ **" .5 "S 'E — • o J5
cu cu "o m cu 4^ in o 3 ^r co eu j cu
.2 *" £ LU •5o5T-4JCOO)—3 CO
4Jt,C3)O.C^ CU4J«CO 4J3 ••>£
oeuco « c. co co cu eo jr — . in
coj:— t- r> oiino.— eueuto »J JJ
L-^CU O.UJ •r-o5To)**4J— 'C CO T-JD
— c > TJ o • as ^^'^Pg 1i cJS
CUID'r- OC- M-C4^E^OE C C — ^
Jzij ooeu o o •» E *• tS .S o> toco
*"**oco *c"D4^ c'-^T^.r-D-Coeu — • 4-*c
0) OOT3 OD.I tOD. O- ••-
coeueu COOM- .-.,-cMeueuL.t.x CO-M t-
cop>.c M-OO 4^c_>oocooa.cu c cue
•o 5 4J "coeu'c'^ 5«v'M^:3Eeiit- §6 co
L.OM- — cooeu •^-eueoM-— euj=o ao. cu-cu
3CDO t-3C— M-13— 3C04J4JM- "0*3 D) •£ 3
CUCU CO.CI-"- 4JL.X CXM-T3 C-O"
COL.13 0)COI— I-O— •— 'CUCOOeU 4JO) t-4JM-
t.cD 1-0)0 CUM-IO— •_£« 0)^ O^O
JS S _. I~u5>~t2 CT^xna.a'I'u 4J— ci-o
3L.CD IIIC/5 CO«1-COXC_'^Q. ••-«_ CD ••-
Ecuu cocooeuroeo 34J •CZ^,4J
OC m COT— •--OC4JQ.4-* 130) '^13
""S" " "•H^°-re-£sl • | o So§
cu^-S — •! — co'euccc o — • "a — '22
1310 D.CO 4J COt-OtOCOM-^t-^ T33 ZO)
'>co" £ '43 c S CM* 4J e >. c« 4J o a.M-13
S-S °S § .Sfe 3 .2 -"c'S-H S i ID a« °g
O.4JO "D*a — • c>M-134J'Deyca tn — * CD co
oi— eu-r- o oco ooo)£.ac0'co 013 —jrcc.
.*3... CUCO U »-inOl-3OCU4JL. 4^0) 4J O3
4J-D4J M-CU 4JT-C«-D.1334JeUCeUO) CO COC-— O
ooo t- c •— T- o T3 ro w cu -M c co3 — • o *> o
3L.3 13 ••- CO»-CUt-O4JCO--CD'«- "- O '^O
OO OO "D Q.»Oin4J Q.>0)3> *U^ '— C4J
C.COC- M-OO) E «— T— CU fO>*t-.c!^CO C X^£OO
Q.Q. M- O O»4: M-— 't-J3O)4-'0)13 COO1 ^CO(OC
M- — ' C 13 C OCO O CO CU » i C •*- I- «•- •*J'9 ^Cl)
01 *-« Q) CO O O CU in •*> O JZ *O 0) <^* 3 3 • *O 0) D) O 4J L« CO
3 .. 3 'tlM-eU X*~M 0 0)4JC04JO)O— > 3 ••- •— ^ 0)130
Tuil 4JO M- 4J 1 ••- — '^COOt-— — ..— CT3 OM-C13
"So"S CU4JO tL 4J..S Q.'DCU "ococ— ceub ^ceu o— eu eu
Qi^ECU t-COO C^^*^*«— M-^»CO'»-CO"-— ^— ^ COL. (0 C_L.
£ 1.3130) O)T- i- .QO. — C C-N^Q.0) COOCO3
n • i cj)trc c- "DICO o t-coui oc.eu>^o
Q_OO SO"-0)T3'^'L.^"PM-CO"-CO C03_tO O4JM-CL
ui3.— 111 co 4J^4JU ccoo eti—xix— IT cootoox
• «OX CDC9-J4^ 3CL. M- CD Q. O •— 0) "£:_*•••--. ' 4J COCO"-CO
4-1 C- 4-T O T3O3 0) XZC3t.tUCOC013CI) (0 — — L. 4J — •
t> CL C O—5-— "OCIl CtT— L.COO — — — -C013 L.COCO
3|| L.4JEL.CO—CCOCU3CO 4J33— —.— ..— ..-£
•D-a — • o) O-— •— jaeueooL-trcococtTEX COCOCOL.
oeuo. to c co »L.C cucocceU"-0)4JO --'i -ecu
L.COU1 *^* •— OOM-»C-4J— JOM-4JO)t-0)'»-"-4J— rf OCO ^ L.L.CDO^
a. co i— L. x 4JQ.o^oj:to— o c J3 u ro ro o C3o a ci m E o
j; 4J 0 -^ EIO — D) 4J 0) 4JC4J4JQ.COCD13"- U 4J4JL. 13
CUO*« CLCO — • CO O)OClT334JCOC"-— 'OIDCC OM- ••- tOCOfl)4JCU
U)L.r^ 03 — 0 COOCMCOOEOTI— C4JOOCI)COCI)t-M- E EE13O4J
33.. L.T3L. o •— — — >o co co c n — eu ••- co o o ••- — — Q. eu u 3co
1 Q^ CO 13 OCO M— E C ^3 4^ ^^ O 4V L, 3C Q) .Q Jv 4^ 4"* ^2 ^s
c cu ^ ^ D 13 a> J= tocDgcocDC—ci-ccococoooooococo cua>coL.o)
— ^ J3 0) O134JJ: U 4J L.13L..— — 4J — Q. 3333330) t. 4J4J D.L.
L- 4J CO J3 O031— L.XOOO1CUO) CO :£C04J1JT>-I3T3T3i— • <0 0)
3 4-i ro •• M- O O «-" L. fl)4JM-j:C0^113CO— 0) 4JCOOOOOO13C ^ M-M-XeUCO
O4JCUOCil-'C— ' ,, ^ 4JM-0)13lOOj:0)— OE4JQ.Q.Q.Q.Q.CU W _ ^ _ , ** JJ
M-D14J'5 .^C"- Q3 W 4JQ) O*"-M OCO — >M-M-M-M-M-M-M-0) QC CUCUfOM-C
c *^ c t— fli 4J o c GJ oej cu •— o to 4J *^ cu c L. uj "D *^ ' "^ *^ *^
CO C '^* '*" CO •-- 0) ^* E '^ O CO E O- fl) O M— 0) "O ^ ~O ~O 1J ^ ^3 3 3 3 C "O "O
" £.2 S fffc g-S S S 3-i 3^-g °-£ ° 3-g co'43 3 3 3 3 3 3 3 "~ § U ^ i 3 3
tLOiM-.^wt-^Q: I, £ irocotocoeu tu to4Jt-crtriT(riTtTcr4J H- 4J4J trtr
£a.— i3ta 4J u a)L.».-c coM-.cOL.coa)0)0)0)Q)a)0)0)O OOM-CUQ)
*j cu (J i i i Q cea.t3cD<3<-'4^»-O)aua:c£tga:ce:c£a:z CD zzoua:
X4JO)CDCU
-------
Cx,
U-4
o
OJ
Q)
Cn
X
cu
0
o
EH
b
<
P
O
fl
Q)
to
c
O
„,_{
4-)
U 0
(1) VD
4_) ^<
O 0
Oi
rH CJ
rt
4J •
C P
<1J
Environn
ihington,
m
CQ (tf
QJ !2
(d
CQ
•d
Q)
-H
D
CO
c
H
1
§
H
H^
M
§
1
i
EFINITIO
Q
1
Q
1
CO
c^
Jjj
|
Pn
CQ
5
4J
CD
0
H
04
O
in
tn
ex
• *
0)
S3
*§
§
(U
CQ
U
1)
1
^J
CQ
0)
JJ
_j
a
o
£
t_
0
CD •
0) ^
•— (0
u
01 •
-C -4-
+•» in
o a:
CD u-
E p
E
O..S
~?
QJ «+-
4-* ••—
O
-5 S.
3 to
a>
3
M- OP
0 •-
V) CU
•M L.
3 — •
CO CO
0) •—
<_ u
a>
cu a.
4-»
«*-
4J O
£ c
4-> O
CU CD
4-* ••"
CO 4-*
•— E
O •—
*J L.
C O
CO H-
X CO
1 fe
C 4->
O •—
to i-
CD U
O
I- CU
o-£
§ °
O ^
-e
H- «—
(M
TJ in
cu «-
t_
'i-s
cu
t- o
•- c
a. *"
21 cu
' o C
a. u
UJ CU
CL
0) CO
4-> CO
1_
o o
°l
*j "o.
— D.
2 co
•s
CO 13
(U
c •—
C *J
O C-
.^ tt)
U
^
o
a. co
W D
cu o
j: -u
4->
"S.2
4-*
cn o
— t-
4-» 4->
8 8
(^
ro
(A
4J
I
^
I
u
4-*
re
!
c.
1
§
o
4-
H-
UJ
CO
— o
* CO —
4J . ro -j
«n O C JO -Q
3-0 =3
o E cu "~ cu a.
L. ^s 4-» CO
3 4-J (O CO G)
CO C tO t- O' J=
C (0 ••- 4-> i • 4-*
CU t- CO «3 X TJ
4-» tO JD O) I 4V O
(0 — GJ QJ CO M-
t_ (0 t- O CO CU
4-> O O C 4J
C CU I M- CO O 4-» •
(D JZ X E — CO
t- I— J3 TJ C GJECU
4_i i GJ O TJ J3 O
CO CU CO D •«-
O) tO CO Q. O CU O
QJ QJ O O C CPQJCO
i. O L. G) J3-J3 (-
— to o. o) a.
JC 4-» «£ O) 4-»
O O C t_ CO CO X
to to o o cu _e 3 t-
0) t- J= E 0
CL * TJ 4-* cu 4^
tT— ' £_ L. O ._G)t_
(U O *r- QJ 4-* OOO
QJ 4-» O* CO TJ CO 4J (0
Z C CU *^ CU GJ O '
OOL.O) 4-* o. co
^ O GJ 4-i I- TJ
o t- .^- GJ CL o
•M 4-* E .CO
• to * _a 4v — • cu
CO GJ 4-> 4-> 3 O
GJ D-JZ O CO O C. TJ
CO D} 3 4-» 4-» C
3 TJ ••- TJ CU C CO
JC 4-> L. O O *•*
4-> Q. CU Q. -M C «-
— • QJ J= CO (0 4^ ^~
CO O 4-* CU 3 4^ CO i
CU O TJ E CO GJ in
GJ 'o -g *to %-*
•t- — * t_ 4-» CO t- QJ tO
— ' C GJ tO 4-> G)
JO CU CO O QJ C Q. C
3 E QJ Q. — ^ QJ CJ —
tE t- «^- JC O -^
0 X 4- Z O GJ
O X C CD TJ
C 0 o jz 4-J to jr o
GJ 4-» TJ O) « JZ
D) O CU Z C TJ 4-»
CO GJ • X t- QJ O C
I- TJ O — • CO JC
— ' — CJ CO CJ > CO TJ 4-»
O TJ tO O JCCUCGJCO't-
(_ to -i— 4->C~OECZ
4-* — ' J2 M- -r- »P- O
CO) M- C — ' 4-» tO — 4-»
O CO — — TJ O O G)
tU~"xO GJGJOCOI-CO
(04-»COC »--« COTJW
L. ••- U 0 CD TJGJ C
J3 Z ••- — 4-» M- — ' Q. -J O
QJ H- CO C O CJ QJ O
4-< Q) H- CO •»- »— — ' JO
I_ O QJ «- (O tO H- — ' tO QJ
> CO **- j5 CO t-
CTJO3 TJCOOi-JC4J
••-L. CO CO O 4J CO
O C €0 * H- — 3
t-OO*^. TJGJ ZE
O O — CO TJ C CO CJ
H-(D4-'Z 5 tO 3 — ' QJ tO
S .£.£'? 5 ftl51 S^
CO ECU GJCGJCOTJ3
TJTJt-L. > -^ QJ - t_4-»
QJ CJ QJ TJ t_ 4J O tO
>. CO -M — * TJCt3)CO
O3GJCO — COOX
CO TJ"- GJ H- •» t_ (0 O
0 C 0 JU X 4-> CO
...CUGJGJ 4-> t_ CO C O
»*- JZ JC Q. O CO O "-•--••-
H- Z 4-» tO 4-» CJ 4-> 0) M-
CJ L. CO D"TJ M-
§JZ M- TJ QJ t- QJ QJ QJ
O O GJ 4-> O O£ CO
— i- C JO 3 GJ
3 (0 CO O" O 4-» CU
tO O C tO QJ *^- L- JCH-
— O O t 'GJCZO
CO CO " CO 4) O —
4->GJ4-*CD £_ TJ C — 4-»
§i_ CO 4-> (0 O Q) •*- (J O
tO TJ tO 4-» M- 4~* fO (0 D
TJ OJ 0 4-» O TJ
CUC04-» TJC03C—C
L.4-'CCO CO TJ — *»- O
••-OI04-* GJ CtDH-O
33>— OJCOEGJ
D"TJ GJ M- COO GJ
t- t- QJ C E CO QJ C — h-
CL ^ QJ t— " CO
— ' J2 O CO 4-1 •
— ' CO QJ M- QJ C TJ O CO
CD<>-JZi. t- CO (0 QJ 3 QJ
JC 4-> O CU X 0 Q.TJ JZ
TJ Q.— ' OOO
CJ 4-» C CO tO CO — • t- tO
> CO — GJ 4-*CCGJO.O
•— JC CO CO 0 CO O > t_
(D4-»4-»3 3 -p-CJJC^
Z C TJ4J4JTJOO
D> — 3 - O ••- tD COO
COCCDCUXl-H-3(UaJO
(0 — E C C- CL 0) — * J3
JC 4-» CO C CD O> »
tOTJ4->tOCUCU>OC4J
XGJCCCO>ja 014-^— CO
O 4-» (0 tO QJ — "^ Z QJ
C O O -M J-; XTJ O Q.
GJJCCL"-GJ CO CO O GJJZ
D1D1O4- C t_ .^- (0 L- tO JZ
tO 3 — • — (0 L. O — 4J
O QJ C C D. — 3 CO — '
GJ t-> O)GJ Et_M-CT'4^ CO
JZ JC CU — JZ 0 OH-GJ (OGJ
*- (M ro «* o
tn
-------
70
-------
Appendix G: List of Registrants Sent this Data Call-In
71
-------
o
o
s
m
tfi
01
1
o
o
o
a
CO
a
!
n
-H
tn
0)
.u
ra
-H
i
n
m
o
4J
0)
o
nJ
o
-H
o
a
, (U
:s
^* r~i
5 VD
'g ro
1
&
-------
Appendix H: List of Related Documents and Electronically Available Forms
Pesticide Registration Forms are available at the following EPA internet site:
http://www.epa.gov/opprd001/forms/.
Pesticide Registration Forms (These forms are in PDF format and require the Acrobat reader)
Instructions
1. Print out and complete the forms. (Note: Form numbers that are bolded can be filled out on
your computer then printed.)
2. The completed form(s) should be submitted in hardcopy in accord with the existing policy.
3. Mail the forms, along with any additional documents necessary to comply with EPA
regulations covering your request, to the address below for the Document Processing Desk.
DO NOT fax or e-mail any form containing 'Confidential Business Information1 or 'Sensitive
Information.'
If you have any problems accessing these forms, please contact Nicole Williams at (703)
308-5551 or by e-mail at williams.nicole@epamail.epa.gov.
The following Agency Pesticide Registration Forms are currently available via the internet:
at the following locations:
8570-1
8570-4
8570-5
8570-17
8570-25
8570-27
8570-28
8570-30
Application for Pesticide Registration/ Amendment
Confidential Statement of Formula
Notice of Supplemental Registration of Distribution of a
Registered Pesticide Product
Application for an Experimental Use Permit
Application for/Notification of State Registration of a
Pesticide To Meet a Special Local Need
Formulator's Exemption Statement
Certification of Compliance with Data Gap Procedures
Pesticide Registration Maintenance Fee Filing
httn://www.ena.20v/ODrtrd001/forms/8570-l.pdf.
httn://www.eDa.eov/oDDrd001/forms/8570-4.Ddf.
http://www.eDa.eov/ODDrd001/forms/8570-5.Ddf.
http://www.eDa.sov/ODnrd001/forms/8570-17.Ddf,
http://www.eDa.eov/opprd001/forms/8570-25.Ddf.
http://www.epa.eov/oDnrd001/forms/8570-27.pdf.
http://www.eDa.BOV/ooprd001/forms/8570-28.odf.
httpV/www.eoa.eov/opprdOOl /forms/8570-3 O.odf.
73
-------
8570-32
8570-36
8570-37
Certification of Attempt to Enter into an Agreement with
other Registrants for Development of Data
Certification with Respect to Citations of Data (in PR Noti
98-5)
Data Matrix (in PR Notice 98-5)
Summary of the Physical/Chemical Properties (in PR Notics
98-1) q
Self-Certification Statement for the Physical/Chemical
Properties (in PR Notice 98-1)
httD://www.eDa.gov/ODDrd001/fr>nris/8570-3?
ehttD://www.epa.gov/QDODmsd1/PR Notices/or98-i
pdf.
http://www.epa.gov/opppmsdl/PR_Notices/pr98-'
pdf. • ^
http://www.epa.gov/opppmsd1/PR Notices/pr9S-l
pdf.
http://www.epa.gov/nppDmsd1/PB Notices/prQR-
pdf.
Pesticide Registration Kit
Dear Registrant:
www.epa.gov/pesticides/registratinTikif/
For your convenience, we have assembled an online registration kit which contains the Mowing
pertinent forms and information needed to register a pesticide product with the U.S. Environmental Protection
Agency s Office of Pesticide Programs (OPP):
1. The Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) and the Federal Food
oTl 99^ C°SmetiC ACt (FFDCA> M Amended by the Food Quality Protection Act (FQPA)
2. Pesticide Registration (PR) Notices
"? Label Improvement Program-Storage and Disposal Statements
u-r-1 Clarification of Label Improvement Program
c. 86-5 Standard Format for Data Submitted under FIFRA
87-1 Label Improvement Program for Pesticides Applied through Irrigation Systems
(Chermgation)
87-6 Inert Ingredients in Pesticide Products Policy Statement
S"i ^^Ingredients in Pesticide Products; Revised Policy Statement
h oo 7 ™™cations> Non-notifications, and Minor Formulation Amendments
"• r*."1 self Certification of Product Chemistry Data with Attachments (This document
is in PDF format and requires the Acrobat reader.)
a.
b.
e.
f
g-
Other PR Notices can be found at http://www.epa.gov/npppmsdl/PR
3 .
Pesticide Product Registration Application Forms (These forms are in PDF format and will
require the Acrobat reader.)
?JA ?0im S0< SZ2"1' APPlication for Pesticide Registration/Amendment
EPA Form No. 8570-4, Confidential Statement of Formula
c. EPA Form No. 8570-27, Formulator's Exemption Statement
74
-------
d. EPA Form No. 8570-34, Certification with Respect to Citations of Data
e. EPA Form No. 8570-3 5, Data Matrix
4. General Pesticide Information (Some of these forms are in PDF format and will require the
Acrobat reader.)
a. Registration Division Personnel Contact List
2. Biopesticides and Pollution Prevention Division (BPPD) Contacts
41. Antimicrobials Division Organizational Structure/Contact List
d. 53 F.R. 15952, Pesticide Registration Procedures; Pesticide Data Requirements
(PDF format)
e. 40 CFR Part 156, Labeling Requirements for Pesticides and Devices (PDF format)
f. 40 CFR Part 158, Data Requirements for Registration (PDF format)
g.. 50 F.R. 48833, Disclosure of Reviews of Pesticide Data (November 27,1985)
Before submitting your application for registration, you may wish to consult some additional sources
of information. These include:
1. The Office of Pesticide Programs' Web Site
2. The booklet "General Information on Applying for Registration of Pesticides in the United
States," PB92-221811, available through the National Technical Information Service (NTIS)
at the following address:
National Technical Information Service (NTIS)
5285 Port Royal Road
Springfield, VA 22161
The telephone number for NTIS is (703) 605-6000. Please note that EPA is currently in the
process of updating this booklet to reflect the changes in the registration program resulting
from the passage of the FQPA and the reorganization of the Office of Pesticide Programs.
We anticipate that this publication will become available during the Fall of 1998.
3. The National Pesticide Information Retrieval System (NPIRS) of Purdue University's Center
for Environmental and Regulatory Information Systems. This service does charge a fee for
subscriptions and custom searches. You can contact NPIRS by telephone at (765)
494-6614 or through their Web site.
4. The National Pesticide Telecommunications Network (NPTN) can provide information on
active ingredients, uses, toxicology, and chemistry of pesticides. You can contact NPTN by
telephone at (800) 858-7378 or through their Web site: ace.orst.edu/info/nptn.
The Agency will return a notice of receipt of an application for registration or amended
registration, experimental use permit, or amendment to a petition if the applicant or petitioner
75
-------
encloses with his submission a stamped, self-addressed postcard. The postcard must contain
the following entries to be completed by OPP:
Date of receipt
EPA identifying number
Product Manager assignment
Other identifying information may be included by the applicant to link the acknowledgment of
receipt to the specific application submitted. EPA will stamp the date of receipt and provide
the EPA identifying File Symbol or petition number for the new submission. The identifying
number should be used whenever you contact the Agency concerning an application for
registration, experimental use permit, or tolerance petition.
To assist us in ensuring that all data you have submitted for the chemical are properly coded
and assigned to your company, please include a list of all synonyms, common and trade
names, company experimental codes, and other names which identify the chemical (including
"blind" codes used when a sample was submitted for testing by commercial or academic
facilities). Please provide a CAS number if one has been assigned.
Documents Associated with this RED
The following documents are part of the Administrative Record for this RED document and may
included in the EPA's Office of Pesticide Programs Public Docket. Copies of these documents are not
available electronically, but may be obtained by contacting the person listed on the respective Chemical
Status Sheet.
a. Health and Environmental Effects Science Chapters.
b. Detailed Label Usage Information System (LUIS) Report.
76
------- |