UNITED STATES ENVIRONMENTAL PROTECTION AGENCY

	WASHINGTON, D.C. 20460

							                                               	OFFICE OF

				PREVENTION, PESTICIDES

	AND TOXIC SUBSTANCES

MEMORANDUM

Date:		April 18, 2007

Subject:	Agency Responses to Phase 3 Public Comments Related to Methyl
Bromide’s Soil and Other Uses without Food Tolerances 

From:	Steven Weiss, Chemical Review Manager

Special Review Branch, 

Special Review and Reregistration Division

To:		Methyl Bromide Docket (EPA-HQ-OPP-2005-0123)

The Environmental Protection Agency (EPA) is developing a Reregistration
Eligibility Decision for the uses of methyl bromide (MeBr) through a
6-phase public participation process that the Agency uses to involve the
public in developing pesticide reregistration and tolerance reassessment
decisions.  

In June of 2005, the Agency released revised risk assessments for all
MeBr uses and other documents for Phase 3.  After Phase 3 was completed,
EPA moved forward with the reregistration process on commodity uses
having food tolerances.  However, the reregistration process for the
MeBr soil uses was delayed due to review of a chloropicrin human study
by the Human Study Review Board (HSRB).   The Agency determined that it
was important to keep the soil fumigants on the same reregistration
schedule in order to ensure that the risk assessment approaches are
consistent and to ensure that economic outcomes can adequately predicted
in reaching risk management decisions for the soil fumigants. 
Therefore, Phase 5 for MeBr was delayed until the chloropicrin risk
assessment was revised based on the comments from the HRSB.

On March 21, 2006 the Agency provided responses to Phase 3 comments
related to commodity uses that have food tolerances Document (see
EPA-HQ-OPP-2005-0123-0134 in the docket).  On August 9, 2006 the Agency
completed the Report of Food Quality Protection Act (FQPA) Tolerance
Reassessment and Risk Management Decision (TRED) for Methyl Bromide and
Reregistration Eligibility Decision (RED) for Methyl Bromide’s
Commodity Uses.  On August 9, 2006 the Agency also opened a 60-day
comment period on the TRED/RED document.   The Agency is currently in
the process of reviewing comments received during the 60-day TRED/RED
comment period.

The Agency is now beginning Phase 5 for soil and other MeBr uses that do
not have food tolerances and has prepared the attached responses to
Phase 3 comments that were not fully addressed in the March 21, 2006
document.   The March 21 document can be found in the docket at
regulations.gov.

 

Table 1. Summary of Phase 3 Public Comments 

1	OPP-2005-0123-0052 	Comments Submitted by Barbara Sachau 

2	OPP-2005-0123-0053 	Comments from Various Private Citizens (Campaign
Letters). A Total of 837 Letters were Received.  

3	OPP-2005-0123-0054 	Comments from the Al-FLEX Exterminators, Inc. 

4	OPP-2005-0123-0056 	Comments from the Association of Floral Importers 


5	OPP-2005-0123-0057 	Comments submitted by the Peruvian Asparagus
Industry  

6	OPP-2005-0123-0058 	Comments submitted by Lee Murphy, President/CEO
California Cut Flower Commission  

7	OPP-2005-0123-0076 	Comments from the Lassen Canyon Nursery, Inc. 

8	OPP-2005-0123-0077 	Comments from the International Paper Company

9	OPP-2005-0123-0083 	Comments from the Smurfit-Stone Rock Creek Nursery

10	OPP-2005-0123-0084 	Comments from the La Department of Agriculture
and Forestry – Oberlin Nursery

11	OPP-2005-0123-0085 	Comments from the Associated Oregon Hazelnut
Industry

12	OPP-2005-0123-0086     	Comments from the Association of Floral
Importers of Florida

13	OPP-2005-0123-0087     	Comments from the Georgia Forestry
Commission’s Flint River Nursery

14	OPP-2005-0123-0088     	Comments from the Southern Forest Nursery
Management Cooperative

15	OPP-2005-0123-0089    	Comments from the Louisiana Department of
Agriculture & Forestry

16	OPP-2005-0123-0090    	Comments from the Texas Forest Service

17	OPP-2005-0123-0091    	Comments from the South California Forestry
Commission

18	OPP-2005-0123-0092   	Comment from the Tennessee Department of
Agriculture  

19	OPP-2005-0123-0093   	Comments from the Sunrise Growers Inc.   

20	OPP-2005-0123-0094   	Comments from the Sunrise Growers, Inc. (repeat
of previous comment)

21	OPP-2005-0123-0095   	Comments from the Peruvian Asparagus Importers
Association  

22	OPP-2005-0123-0096   	Comments from the FJ Pugliese Company Inc.     


23	OPP-2005-0123-0097   	Comments submitted by Don Stringfield   

24	OPP-2005-0123-0101   	Comments from the Warren Co. Ag. Extension
Service 

25	OPP-2005-0123-0102   	Comment submitted by Plum Creek    

26	OPP-2005-0123-0103   	Comments from the California Carrot Board  

27	OPP-2005-0123-0104   	Comments from the California Pepper Commission 


28	OPP-2005-0123-0105   	Comments from the Fumigation Service & Supply,
Inc. 

29	OPP-2005-0123-0106   	Comments from the California Tomato Commission 
 

30	OPP-2005-0123-0107   	Comments from the California Tomato Growers
Association   

31	OPP-2005-0123-0108   	Comments from the California Pistachio Industry


32	OPP-2005-0123-0109   	Comments from the University of Hawaii  

33	OPP-2005-0123-0110   	Comments from the Royal Fumigation Inc.        

34	OPP-2005-0123-0111	Comments from California Minor Crops Council   

35	OPP-2005-0123-0112 	Comments from the Pesticide Action Network (these
comments were re-submitted, see below)

36	OPP-2005-0123-0113 	Comments from the Pesticide Action Network (these
comments were re-submitted, see below)

37	OPP-2005-0123-0114 	Comments from the Western Industries-North, Inc. 
     

38	OPP-2005-0123-0115 	Comments from the National Pest Management
Association Inc.  

39	OPP-2005-0123-0116 	Comments from the Golf Course Superintendents
Association of America   

40	OPP-2005-0123-0116 	Comments from the Golf Course Superintendents
Association of America 

41	OPP-2005-0123-0117 	Comments from Fumigation Service & Supply,
Inc. 

42	OPP-2005-0123-0118 	Comments from Fred Leitz, Leitz Farms LLC 

43	OPP-2005-0123-0119 	Comments from the Plant Protection and
Quarantine Program/US Department of AHHIS

44	OPP-2005-0123-0121   	Comments submitted by Bayer Corporate and
Business Services, LLC  

45	OPP-2005-0123-0122  	Comments submitted by the Natural Resources
Defense Council  

46	OPP-2005-0123-0123   	Comments on Methyl Bromide from Private Citizen
Carol Fisher

47	OPP-2005-0123-0124   	Comments from John Kepner, Beyound Pesticides
(re-submitted version of OPP-2005-0112)

48	OPP-2005-0123-0125      	Comment from Private Citizen RL McKay

49	OPP-2005-0123-0126   	Comments from Chava McKeel of GCSA

50	OPP-2005-0123-0127   	Comments of the MBIP of the ACC

51	OPP-2005-0123-0128	Comments from Pesticide Action Network 

52	OPP-2005-0123-0129	Comment submitted by Almond Hullers & Processors
Association



Table 2.  Responses to Phase 3 Public Comments 

Comment#  (based on numbering from 3/21/06 document)	OPP-2005-0123
Document ID	Public Comment	Agency Response

1 

	0056, 0076, 0077,

0083, 0084, 0085,

0086, 0087, 0088,

0089, 0090, 0091, 0092, 0093, 0102, 0110, 0111, 0118	The Agency makes a
number of assumptions on the rates of application, methods of use and
other use practices for methyl bromide. These assumptions do not reflect
typical use practices.	The Agency assessed risk based on a range of
application rates and other parameters such as field/chamber size,
weather conditions, etc.  Data submitted by stakeholders on typical
rates, methods and other practices throughout the risk assessment
process has been reviewed and incorporated into the revised risk
assessment as appropriate.  The Agency has also worked with USDA to
refine the risk assessment.

Information received pertaining to use patterns and the importance of
soil fumigants in the production of many crops is very valuable to the
Agency.  The Agency plans to incorporate the information received in the
docket and directly from stakeholders, as appropriate, in its
assessments of the benefits of fumigants and in its analysis of the
impacts of potential regulatory changes.



41

	124	Chloropicrin is combined with methyl bromide as a warning agent and
also as an active ingredient. Chloropicrin is used with methyl bromide
in various products at ratios varying from approximately 1:400 to 1:1.
In California, 4.93 million pounds of chloropicrin were reported used in
1997, compared to 7.38 million pounds of methyl bromide (Pesticide Use
Report, CDPR, 2003). This is particularly relevant in applications to
strawberries, for which reported methyl bromide use in 2003 was 3.67
million pounds, and reported chloropicrin use was 3.28 million pounds
presumably applied together). US EPA includes a brief discussion of the
public health data regarding chloropicrin/methyl bromide formulations in
its risk assessment. However, there is not enough information to
adequately convey the impact of combining these two active ingredients
on human health. We have provided general comments regarding US EPA’s
methods to evaluate cumulative risks of the soil fumigants (see Section
1.5). It is essential that US EPA quantify the health risks of the
exposure to the combination of methyl bromide and chloropicrin,
including a complete toxicological evaluation of combined toxicity,
exposure, and interaction between chloropicrin and methyl bromide to
address the risk of using these formulations. When US EPA completes its
assessment for chloropicrin as an active ingredient, we believe that the
findings will demonstrate that chloropicrin is much more acutely toxic
than methyl bromide (up to about 50 times more potent as an irritant).
Therefore, we believe that the effects of chloropicrin, and not methyl
bromide will dominate the acute toxicity hazard from the use of
chloropicrin/methyl bromide mixtures. In addition, because the
volatility and evaporation rate of chloropicrin is lower than that of
methyl bromide, it is likely that chloropicrin persists longer in the
environment. Therefore, measured levels of methyl bromide in ambient air
would not accurately predict chloropicrin levels based on the initial
mixture ratio. It appears that the longer- duration inhalation exposures
from use of the combined products could be essentially chloropicrin
exposure. We have already alluded to the polymorphism of
glutathione-S-transferase (GST) enzymes in humans and the role of GST in
mediating methyl bromide toxicity. Because GST also activates
chloropicrin66 there is clearly a need for US EPA to research the
association between the mechanism of action and the cumulative toxicity
of these two agents. In looking toward the future for assuring worker
and public health and safety related to fumigants, we remain concerned
about the absence of regulatory controls specific to chloropicrin, an
agent with high acute toxicity, and chloropicrin used in combination
with methyl bromide and other fumigants. Some registered methyl bromide
products currently contain up to 50% chloropicrin, thereby increasing
our concern regarding this agent. Even if methyl bromide use continues,
or if its use declines or is phased out altogether, it is likely that
some growers will choose to use other fumigant that also utilize
chloropicrin as an active ingredient and warning agent. Since products
containing chloropicrin are being re-evaluated for registration purposes
at this time, it seems appropriate and timely that US EPA initiate rule
making for chloropicrin, as it is used in combination with other soil
fumigants, in particular methyl bromide and Telone. We believe that when
assessed together, the combination of the toxicity levels of concern and
the exposure risks for residents and workers will require even more
stringent mitigation and control measures than the use of either of
these chemicals alone. It is quite possible that the data will
demonstrate that there is no acceptable means to handle mixtures of
methyl bromide and chloropicrin.	See response to comment#1 regarding
pounds of fumigant used.

Along with the revised risk assessment for MeBr, the Agency has also
provided in the docket revised risk assessments for Chloropicrin and the
other soil fumigants.  Furthermore, EPA is soliciting comments on risk
mitigation options for all of the fumigants (see “Risk Mitigation
Options to Address Bystander and Occupational Exposures from Soil
Fumigant Applications”).  

The Agency recognizes the role of GST on the metabolism of methyl
bromide & chloropicrin.  Though GST metabolism is common for both
compounds it does not, however, lead to a common toxic effect/profile.
For methyl bromide the endpoints of concern are developmental toxicity
for acute assessments and nasal basal cell hyperplasia for the chronic
assessments.  In contrast, the acute risk assessment for chloropicrin is
based on eye irritation in humans while the chronic assessments are
based on rhinitis and exudate, hyaline epithelial inclusions, and
olfactory epithelial atrophy.  As these two compounds elicit different
toxic effects, the Agency would not conduct a cumulative risk
assessment.  Moreover, according to Agency policy a clearly defined mode
of action - not a single event - is required to consider compounds for
cumulative assessments.

58

.

	127	Bystander Exposure and Risk From Pre-Plant Agricultural Use —
PERFUM Model -- Estimation of Flux Rates — Pages 32-45 of Phase 3 Risk
Assessment (included as Comment Iva on page 5).  The comment is
summarized below.

The MBIP supports EPA’s use of the PERFUM Model to estimate
non-occupational exposures. Using this approach leads to more accurate
exposure estimates than applying the ISCST3 model. However, there is a
serious technical flaw in the way EPA has calculated the flux rates that
were used in the PERFUM modeling. Compared to the other soil fumigants,
methyl bromide has an extensive array of field studies from which flux
rates can be estimated. In the PRA, EPA relies on work done by the
California Department of Pesticide Regulation (CDPR) to generate
composite flux rate profiles (i.e., flux rate versus time since
application) for each type of field application (see Johnson, 19992;
Johnson and Segawa, 2000 Johnson, 2004 This work was not intended to be
used for developing flux rate profiles and certain aspects of CDPR’s
procedure for calculating the profiles are problematic and lead to the
overestimation of risk.

Under CDPR’s approach, a mean “emission ratio” is estimated for
each application type based on all of the field study data. The emission
ratio is defined as the percentage of methyl bromide mass that
volatilizes in the first 24 hours after the application. This part of
the procedure is scientifically sound and acceptable. CDPR then develops
a composite decline profile for methyl bromide that is meant to reflect
the proportion of mass that volatilizes over each hour since the start
of the application. There is a significant technical flaw in the manner
in which EPA is using this aspect of CDPR’s analysis, the result of
which is an overestimation of buffer zones.

The technical flaw arises from the decline curves developed by CDPR.
Those curves were based on the average downwind concentration as opposed
to the flux rate. While there is a correlation between these two
parameters, it is not perfect. For example, a higher flux rate is needed
during the daytime to achieve the same concentrations near the field as
during the nighttime, due to the more conducive dispersion conditions
during the daytime.

The problem with using the concentration to develop a decline curve for
the flux rate is illustrated by the following example. Using data from a
study at Seal Beach, California, CDPR developed a decline curve for the
bedded tarp application using the shank-ahead method The application
occurred on June 24, 1999. The decline curve is reported in Johnson,
1999 (see Figure A9). Figure 1 below reproduces the decline curve in the
original CDPR report and plots the concentrations and flux rates using
the mid-point of the measurement interval. The concentration values
(triangle symbol) fall very close to the decline curve. This is expected
since the concentration values were used to estimate the decline curve.
The flux rate is in different units. Therefore, it was plotted on a
separate axis and the peak value at hour 3 was matched to approximately
fall on the decline curve. The second flux rate point falls
substantially below the decline curve. The result is that the flux rate
estimated with the CDPR decline curve is much higher during the
nighttime period than actually occurred during the study.

To illustrate the effect of the poorly specified flux rate decline
curve, the MBIP modeled the actual flux rate data from the Seal Beach
shank-ahead study to compare the results with the composite curve used
by EPA. Table 1 summarizes the flux rates for the Seal Beach study and
the composite flux rate used by EPA, both scaled to an application rate
of 430 lbs/acre. Using the actual data, the flux rates are generally
higher in the daytime following the application. However, during the
nighttime period, the smoothed flux rates are several-fold higher than
the actual flux rates. This is problematic and scientifically unsound
because the atmosphere is less conducive to dispersion at night, and
placing more mass in the nighttime period than actually occurred results
in an overestimation of the buffer zones as shown in Figure 2.

Figure 2 presents the 95 percentile buffer distances for both the whole
field and maximum concentration approaches using both the datasets
summarized in Table 1. The buffer zones using the actual flux rates are
about 20% lower than with the smoothed flux rates. The reason, as noted
above, is that the smoothed flux rates artificially place more mass in
the nighttime period than actually occurred in the volatility study. In
the PRA, EPA should correct this issue by either: (1) using the actual
flux rates in the modeling as was done for all of the other fumigants,
or (2) developing a smoothing method based on the flux rates, and not
raw concentration values.	The Agency solicited the assistance of the
California Department of Pesticide Regulation (DPR) in addressing this
comment.

MBIP suggests that use of the concentrations to develop a flux curve
leads to overestimation of buffer zones because the concentration-based
flux curve overestimates flux during night time hours, when stability
conditions would exaggerate downwind concentrations.  This is based on a
shank-ahead study (Johnson and Wofford 1999).  It is true that if flux
estimates during the night are too high, then downwind concentration
estimates will also be accordingly high. The selection of the particular
study is not believed to be in error because of the following:

1) The particular study chosen as an example was atypical of methyl
bromide studies because the start of the application was delayed until
14:00.   With reference to onset of nighttime conditions and in relation
to other studies, nighttime conditions, when measured as time after
application, occurred sooner for this particular study.  Of 15 studies
listed in Johnson (2004), 12 studies started at 10AM or earlier, 1 study
started at 11:30AM and 2 studies started at 2PM.  Most methyl bromide
applications start earlier in the day, rather than later in the day, but
there is no regulatory requirement for restricting start times.

2) The reasons for using concentrations to pin down the flux curve,
instead of measured fluxes included:

a) Many studies only provided concentrations (no direct measurements or
flux estimates were possible) and in order to consistently use all of
the studies, the concentrations were averaged to gain a profile of flux.

b) The focus at the time this methodology was developed was to estimate
how long buffer zones needed to be maintained.

c) The emission ratios, which MBIP supported, were based on flux
analyses and were used in the process of defining these curves.

3) The phrase “actual flux rate data from the Seal Beach shank-ahead
study” (pg7, MBIP) must be understood in terms of sampling and
analysis procedures.  Generally, off-site air monitors are run for
periods of time from 6-12 hours.  It is difficult to sample at a greater
frequency (i.e. shorter time periods). Thus, the estimated flux
utilizing the back-calculation procedure (Johnson et al 1999) provides a
gross estimate of flux for that time period.  Other factors which can
influence flux, but which would be reflected in these time periods are
(1) temperature affect on flux (2) wind affect on flux (3) spatial
variability of flux (4) finer temporal variability in flux.  If flux
were measured at hourly or half-hour intervals, one would expect a
considerably more jagged line.  However, it is believed that the overall
shape of jagged line would follow the composite flux profile line,
notwithstanding differences in individual studies between the composite
flux profile and the individual study profile.

4) Analysis of emission ratios indicates wide variability between
studies.  Coefficients of variation in emission ratios for broadcast no
tarp, broadcast tarp and bed tarp emission ratios were 47%, 52% and 38%,
respectively (Barry 1999).  With this level of variability, it would be
expected that any particular study, when compared to the composite flux
profile, would show over- and/or underestimation of actual flux.

5) The MBIP Figure 1 is somewhat misleading because in order to
accommodate the different units, MBIP fixed the first measured flux to
the peak of the fitted concentration curve.  This resulted in lowering
the subsequent measured flux values in relation to the composite flux
line.  To gain a more accurate picture of the situation, Figure 1 here
plots the MBIP modeled fluxes (composite vs particular study, normalized
to 430 lbs/acre) from their Table 1. It is clear that with respect to
this particular study, the composite flux at first underestimates
‘actual’ flux, then overestimates ‘actual’ flux.  

References

Barry, Terri. 1999. Memorandum to Randy Segawa on Methyl bromide
emission ratio groupings dated Dec 2 1999.

Johnson, Bruce and Pam Wofford. 1999. Memorandum to John Sanders on
Monitoring results from a comparison test of bedded tarped application
equipment dated September 24, 1999.

Johnson, Bruce, Terrell Barry and Pamela Wofford. 1999. Workbook for
Gaussian modeling analysis of air concentration measurements. EH99-03.
State of California, Environmental Protection Agency, Department of
Pesticide Regulation, Environmental Monitoring and Pest Management
Branch, Environmental Hazards Assessment Program 

Johnson, Bruce. 2004. Memorandum to Kean Goh on Recent Help on Air
Modeling Provided to the United States Environmental Protection Agency
dated Dec 17, 2004. Sacramento, California 95814-3510

Lewis, Susan A., David McAllister, Vincent J. Piccirillo, Paul M. Price
and Richard Reiss. 2005. Comments of the Methyl Bromide Industry Panel
of the American Chemistry Council on U.S. EPA’s methyl bromide risk
assessments for fumigant pesticide. Docket No. OPP-2005-0123 Notice of
Availability 70 Fed. Reg. 40336 (July 13, 2005). Methyl Bromide Industry
Panel c/o The American Chemistry Council, 1300 Wilson Boulevard,
Arlington VA 22209.





59

	127

	Bystander Exposure and Risk From Pre-Plant Agricultural Use — PERFUM

Model — Whole Field v. Maximum Concentration Approach -- Pages 36-46

The Human Health Risk Assessment presents results from the PERFUM model
based on twenty-five combinations of flux and meteorological data for
field (pre-plant agricultural) applications. The PERFUM model generates
two statistical distributions of buffer zones in its output — one
based on the maximum concentration approach and the other based on the
whole field approach. Both are presented in the PRA. For the maximum
concentration approach, the direction from the field with the highest
concentration is selected for each simulated day and the values for all
of the simulated days are assembled into a distribution. Essentially,
this represents a variability distribution around the maximally exposed
individual (MEI), except that in this case, it is not necessarily known
that there is an individual at the maximally exposed location. In fact,
given the sparse population near many applications, the odds are against
someone being right at the maximum concentration location (and for all
24- hours after the application as EPA assumed for methyl bromide).

By contrast, the whole field distribution expands upon the maximum
concentration distribution by including locations in every direction
around the field, not just the direction that produces the maximum
concentration. Therefore, the whole field distribution is essentially a
population distribution at the buffer zone. It represents the
variability in exposure at the perimeter of the buffer zone.

The debate between using the whole field versus the maximum
concentration

approach largely falls towards what is meant by probabilistic risk
assessment. EPA recently explained what “probabilistic analysis”
means in its Risk Assessment Principals & Practices. There it stated
that probabilistic analysis is “a means for describing the uncertainty
in risk estimates by characterizing the uncertainty and population
variability in the individual steps by probability distributions.”
Only the whole field approach represents a “population variability.”
The maximum concentration approach only represents the variability in
exposure of a single individual, the ME In fact, the whole purpose of
probabilistic risk assessment was to move away from just focusing on the
MEl. In other OPP assessments such as for dietary exposure, the
percentiles used for regulation define population variability.
Furthermore, in the OP cumulative assessment of all exposure pathways,
the percentiles also referred to a population variability. For all of
these reasons, EPA should utilize the whole field approach in assessing
risks of exposure to methyl bromide.

The MBIP believes that exposure should be assessed on the basis of a
population, not a single individual. Although specification of the
exposed population is not entirely straightforward given the nature of
field fumigation, the PERFUM model defines the population in the most
constrained way possible. The population only includes the locations
around the buffer zone perimeter. As one moves away from the perimeter,
the concentrations fall and defining the population as alternatively
being some distance from the field beyond the buffer zone would have
resulted in smaller values. Going the other way, it is clearly
inappropriate to include those locations inside the buffer zone, as
these locations will be protected from exposures while the buffer zone
is in place. Therefore, the PERFUM model (using the whole field
approach) defines the population in the most conservative manner
possible.	The explanation of whole field and maximum concentration
outputs have been updated in the risk assessment.  Both have been
presented for characterization purposes.  No decisions on buffer zones
or other risk mitigation measures have been made yet.  This and other
comments received will be considered by the Agency.  Furthermore, EPA is
soliciting comments on risk mitigation options for all of the fumigants
(see “Risk Mitigation Options to Address Bystander and Occupational
Exposures from Soil Fumigant Applications”).  



64

	127

	Overview of Errors and Flaws In EPA’s Analysis of Pre-plant and Field
Fumigation Workers Available data on worker exposure and the proper use
of methyl bromide toxicity data indicate that current risks are
significantly lower than the estimates given in the PRA. As a threshold
issue, the MOE calculations presented in Table 6 of the Agency’s
“Overview” document and Table 20 in the Revised Human Health Risk
Assessment (June 13, 2005) appear to be incorrect. It appears that the
Agency applied a Human Equivalent Concentration (HEC) of 10 ppm to
calculate the MOE for both the acute and the short- and
intermediate-term scenarios. However, this HEC value does not match the
HEC concentrations for these occupational scenarios that are presented
in Table 2 in the “Overview” document (Table 4 in the Revised Human
Health Risk Assessment). For acute occupational exposure, the HEC is
reported as 30 ppm (not 10 ppm) and the required MOE is 30. In essence,
this is equivalent to an acceptable acute human exposure concentration
of 1 ppm. It also appears that the acute and the short- and
intermediate-term exposure concentrations for “All Workers at Field
Application” are in error.

As stated in the introduction of this section, the Agency’s analysis
is based solely on data from workers in the State of California. Based
on that data, the Agency concludes that all of the job classifications
evaluated in the assessment fail the 1 ppm risk criterion. However, in
California, soil applications are currently being conducted to meet a
210 ppb 24- hour TWA standard. This standard has been set by the
California Department of Pesticide Regulation (CDPR) and is applicable
to all workers. The standard has the effect of placing a cap on the
exposures that occur in any 8-hour period. The highest amount of
exposure that can occur in this shorter period is three times the
24-hour standard or 630 ppb (0.63 ppm). Methyl bromide applications in
California currently meet this standard. Therefore, EPA should expect
that data from California would support a finding that current exposures
do not exceed 0.63 ppm, and as a result, that current exposures also
will not exceed the higher health criterion of 1 ppm.

Despite the fact that the soil fumigation industry is meeting the more
stringent standard, EPA has determined in the PRA that the available
data indicate that all seven jobs classifications associated with soil
fumigation pose unacceptable risks. The Agency is led to this conclusion
because it has failed to account for current practices that control
exposure. The root of this error is the Agency’s use of monitoring
data that dates back to 1981. Prior to 1992, the standards for exposure
were 15 ppm and as a result, levels of exposure above I ppm were
permitted in the 1980s.

Application practices have changed since the 1 980s and current
exposures are lower. Modifications currently in place in California that
reduce current exposures include:

• Reduction of the application rates;

• Changes in applications methods (improved injection designs, rapid
closure of injection shank furrows, and separation of tarp cutting and
tarp removal processes) that minimize the amount of methyl bromide that
is released;

• Use of high barrier tarps;

• Modification of tractors; and

• Use of ventilation fans in the tractor cabs.

Because the worker monitoring data from the 1990s is more recent and
because of the documented changes in engineering practices data from the
1 980s, the studies performed in the 1980s should not be used in the
PRA. A second problem with the Agency’s analysis is that it has
included all of the data from studies of engineering controls where the
above technologies have been demonstrated. These data included the
“control” data where the technology was not applied. If EPA had
limited measurements used in these studies to the data that reflect the
current controls, the results would indicate that workers are not
exposed to unsafe levels.

While the decision to include monitoring data that do not reflect
current practices is the major reason for EPA’s finding of
unacceptable risk, the PRA also contains a large number of factual
errors in the compilation of monitoring data from the California
studies. These errors include:

• Errors in the citation of studies;

• Failure to include all of the data in a given study;

• Duplicative citations of the same data;

• Inclusion of studies for pests that are not supported by current
labels (ground squirrels);

• Misapplying data to the wrong job classifications.

These errors are outlined below. In preparing these comments, the MBIP
has sought to correct the data and to provide a more accurate
characterization of the currently available studies. A corrected data
set is presented in Attachment 4.

Aside from these errors, EPA has failed to interpret properly the
available monitoring data in several respects. These errors include:

1. EPA does not give equal weight to data from studies from the early
1980s that are not representative of current exposures and data from
more recent studies.

2. EPA incorrectly gives equal weight to samples taken with and without
controls.

3. EPA incorrectly gives equal weight to short term measurements (30 mm.
and less in duration) and measurements over an entire working day.

4. EPA assessment fails to consider the impact of reduced application
rates and other mitigation measures already in use in California on
current exposures.

These points are discussed in detail below.

Finally, this assessment has focused on the impacts of a proper
interpretation of the toxicity and exposure for acute risks,
specifically whether workers in 2005 will be exposed to levels above 1
ppm 8 h TWA on any given day. The comments have focused on the acute
risks because these are the risks that the PRA identified as being of
the greatest concern. However, the proper evaluation of the exposure
studies described below will also result in lower estimates of
short/intermediate-term and long-term exposures. These lower estimates
also result in acceptable risks for these toxicological endpoints
without the need for respirators.	The error pertaining to the proper HEC
has been corrected.  An 8 hour HEC of 30 ppm with an associated
uncertainty factor of 30 was used as the basis for all worker risk
calculations in the updated Phase 5 assessment.  The use of the HEC = 10
ppm was clearly in error.  [See Table 3, page 19 in revised Phase 5 risk
assessment for further information.]

With regard to the specifics of the exposure monitoring data which have
been used by the Agency, MBIP is correct on several points pertaining to
how the data were used in the previous (Phase 3) risk assessment.  The
Agency has attempted to correct these errors.  In fact, in table 6 of
the MBIP comments (pg 35) revised estimates of exposure were presented
as the “corrected version”.  The revised Agency estimates are
similar to the values presented by MBIP.  In some cases, exact agreement
is noted while in other cases values vary but only based on data from a
few worker monitoring events.

The Agency acknowledges that application practices have evolved since
the 1980s and that exposure levels would likely be reduced if such
practices are employed.  However, as MBIP commented, most of the
evolution in application practices occurred in California and are a
direct result of the regulatory process over that timeframe through
CDPR.  While a significant amount of methyl bromide use occurs in
California, it is not clear if in other areas of the country similar
practices have evolved.  As such, it is difficult to delineate which
exposure monitoring data may be or may not be applicable to current
practices on a national level.  For this reason, the Agency has retained
and used all available monitoring data.  It should be noted that in many
cases the monitoring data inherently reflect the use of various
engineering controls such as ventilation fans for tractor drivers or the
use of updated devices such as Nobel plows.

MBIP comments mentioned several specific issues that should be
considered by the Agency.  These are addressed below:

Rate reduction:  It is intuitive that rate reductions would reduce
exposure levels but these values can be impacted by many other factors
such as soil type, shape and orientation of the injection shank, soil
moisture, or attention paid by the tractor driver with regard to end of
row turning activities.  As such, no attempt has been made to
extrapolate using the available data to possibly lower application
rates.  If MBIP monitored exposures at lower application rates, such
data would be considered for exploring the possible impact of this
phenomenon.

Application Methods:  As indicated above, it is likely that applicators
in California could have lower exposure due to more refined application
methods but it is not clear how such changes could be considered on a
national level. 

High Barrier Tarps:  In some cases, high barrier tarps were used during
the monitoring events captured in the available worker exposure data. 
For tarp cutters and tarp removers, where the use of high barrier films
is believed to have the most potential impact because of the possibility
of trapping more residues than a typical polyethylene, a majority of the
available monitoring data were generated with high barrier films.

Tractor Ventilation:  Some of the monitoring data reflect the use of
tractor ventilation while others do not.  All have been included in this
analysis.  Based on the responses above, it is expected that on a
national level that application devices are in use that both include or
are not outfitted with tractor cab ventilation devices.

Study Citation & Summary Errors:  These errors have been corrected based
on the information provided on pages 29 through 35 of the MBIP comments,
including Table 6 of the MBIP comments.

Deleted Exposure Scenarios:  The Agency concurs on the comment regarding
the ground squirrel use.  These data have been removed from the
analysis.

Lack of Controls:  The Agency guidelines for exposure data typically
require thorough, well documented quality control regimens.  Given that
most data were either generated by CDPR or the MBIP (where much of this
information is available) it was decided to use the data for risk
assessment purposes.  Use of data where only control measurements were
collected would limit the amount of monitoring events that could be
considered for the analysis which could possibly add as much uncertainty
to the overall result as the lack of control information. 

Short-term Measurements:  In many circumstances, the measurements used
in this analysis were collected over shorter durations of exposure that
do not represent 8 hour time weighted averages.  It is expected,
however, that the resulting monitoring values represent an exposure rate
for the monitored period.  It follows that if an individual worked for 8
hours doing the same specific task that the exposure rate would not
change regardless of the duration worked.  Given the size and scale of
methyl bromide use in agriculture where 8 hour work days for specific
tasks could be expected, it is reasonable to use monitored exposure
rates (i.e., the measured concentrations) to calculate risks using 8
hour HECs.

Overall Summary:  There are many possible issues associated with the
available methyl bromide occupational monitoring data as described
above.  In order to provide a broad a characterization of the
occupational risks associated with methyl bromide all of the appropriate
information was used.  Further refinements would likely require that
additional information be generated based on modern cultural practices
(e.g., currently available high barrier films and typical application
rates/timing).



65

	127

	Errors and Data Related Issues in the PRA Analysis

The MBIP found numerous errors in the data underlying the Agency’s
analysis. These errors were not evident in the prior rounds of the error
correction phase because the underlying data have only now been made
available. Thus, this is the first opportunity that the MBIP has had to
discover them.

As more fully detailed below, the errors in the data are numerous. For
ease of reference, the errors are organized based on the study names as
listed in the bibliography to the PRA (Appendix C). In the tables
provided in Appendix T, 16 studies are listed as contributing data to
the PRA (Field 3, 4, 5, 7, 16, 18, 20, 21, 21, 23, 38, 39, 49, and 50).
Three studies have duplicate listings (Field 22 is also Field 40, Field
38 is also Field 41, and Field 49 is also Field 50). Thus there are
actually only 13 named studies, not 16. In addition, as the following
indicates, the names on the actual documents do not necessarily match up
with the names in the bibliography.

• Field 4. Siemer and Associates, Inc. (1992c) AMBI: Shallow
Shank-Tarped Bed Fumigation Assessment (Interim Report No. SM924096C,
M).

This “study” consists of two reports. The EPA analysis omits data
from the second of the two reports. In this analysis these data are
included. This adds 19 additional measurements to the database. As
discussed below, the missing data also provide information on how
commonly used controls reduce worker exposures in the fumigation of
beds.

In addition, the data from Siemer and Associates, Inc. (1992c) on
Shovelman (Table 4 of the study) were not used in the EPA analysis. The
MBIP used these data in the assessment summarized below. Data on the job
that EPA labeled as “Plastic Truck Driver” are included in the
“Second Driver” job. However, the “Second Driver” appears to be
intended to cover workers in vehicles that follow the tractor pulling
the methyl bromide injection equipment. These vehicles can include
tractors laying tarps and pulling cultipackers, but they do not include
the “Plastic Truck Drivers”. These drivers bring the tarps to the
work site but do not enter the fields. Therefore these workers should
not be included in the “Second Driver” job. The data in this study
is also reported under Field 39.

• Field 5. Maddy, K.; Lowe, J.; Fredrickson, S. (1984) Employee
Exposure To Methyl Bromide Used For Ground Squirrel Control: HS-1238.
Unpublished study prepared by California Dept. of Food & Agriculture,
l3p.

This study reports on exposures that occur from the control of ground
squirrels. The registrants no longer support this use. In addition, the
method of application is so different from soil fumigation (squirting
methyl bromide down a burrow by hand) that the data are not a relevant
measure of agricultural worker exposures from soil fumigants. The data
from this study should not be included in the PRA.

• Field 18. Maddy, K.; Gibbons, D.; Lowe, S.; et al (1984) A Study Of
the Inhalation Exposure of Workers to Methyl Bromide during Preplant
Soil Fumigation (Shallow Injection) in 1980 and 1981: HS-900.
Unpublished study prepared by California Dept. of Food and Agriculture,
Div. of Pest Management, Environmental Protection and Worker Safety,
Worker Health and Safety Unit. 37p.

The name of the studies that provide the data listed under Field 18a-e
in the Tables in Appendix I is incorrectly listed in Appendix C. The
data in Field 18 are taken from three sources. Two, and possibly three
of these sources are from studies other than the one cited for Field 18.
The first set of data listed as 18a comes from a 1982 Study by Maddy (A
Study of the Inhalation Exposure of Workers to Methyl Bromide During
Preplant Soil Fumigants (Shallow injection) in 1980 and 1981.) The
second set of data, listed as I 8b comes from an unreferenced table
described as “preliminary results — for review and discussion
only.” The table does not include any supporting information on the
origin of the data or the date that it was collected. The third set of
data, listed as 1 8c-e, is a series of internal studies from TriCal Inc.
collected by Donald Richmond of the State of California. This set of
data is presented later as part of the data entered as Field 49.

The data from the first two sources (labeled 1 8a and 1 8b) are used to
evaluate the Driver, Co-pilot, and Shovelman job classifications. The
data from all three sources (labeled 1 8a, 1 8b, 1 8c, 1 8d, and 1 8e)
are used in the determination of “All Workers at Field Application.”
In the corrected analysis presented below, only the data from the first
source is used. The data from the second source, the 1984 table, should
not be used since the table lacks any description on where the data
originates, how it was collected, and how the values were calculated.
Finally these data are described as preliminary. The data from the third
source appear to be a subset of the data presented in Field 49 and are
entered under that Field number.

• Field 19. Gillis M. Becker (1994 c) Report on New High Barrier Film
Evaluation Studies, TC233 1, TC233 2 and TC233 3, TRICAL Inc.

This study is not listed in Appendix C but is listed in Appendix T.

• Field 20. TriCal, Inc. (1993a) Tarp Removal Worker Exposure Vol. 1-2
(TC211).

EPA used a measurement of Tarp Puller exposure that reflects a technical
problem in the field and is not representative of typical tarp remover
exposures. This data point should be removed from EPA analysis.

• Field 22 (Also listed as Field 40). Siemer and Associates, Inc.
(1993b) Deep Shank, Nontarped Fumigation-Mitigation of Methyl Bromide
Worker

Exposure and Offsite Drift (Interim Report No. SM934104, 1-2, Report No.

SM934104.2-1, Vol. 1-2).

Data from this study are listed under Fields 22 and 40 in the RED. Thus
the data are double-counted in the EPA analysis. In the MBIP’s
corrected analysis, these data are entered only once under the heading
of Field 22.

• Field 39. Siemer and Associates, Inc. (1992b) Nontarp Deep Injection
for Measurement of Methyl Bromide Exposure to the Applicator, Applicator
Assistant, and Cultipacker Tractor Driver (Interim Report No.
SM924096B).

This Field is mislabeled in Appendix C. The data for Field 39 actually
comes from the study, Siemer and Associates, Inc. (1992c) AMBI: Shallow
Shank-Tarped Bed Fumigation Assessment (Interim Report No. SM924096C,
M). The data from Siemer and Associates, Inc (1992c) is already entered
into the PRA analysis under Field 4.

In the corrected analysis presented below, the data from Siemer and
Associated, Inc (1 992c) is removed from the Field 39 entry and is
replaced with the data from Siemer and Associates, Inc. (1992b).

• Field 49 (Also listed as Field 50). TriCal, Inc. (1987)
Response/California Notice 87-5: Risk Assessment/Methyl Bromide. DPN
123-099. Record No. 64748 to 64753.

Field 49 is incorrectly listed in Appendix C. It is actually an
interoffice memo dated 1981 (Great Lakes Chemical Corporation, (1981)
Interoffice Memorandum From D. L. McAllister to R. J. McKeand, Subject
Methyl Bromide — Applicator Exposures) and is composed of a series of
data sheets on worker monitoring.

A number of errors were identified in the transcription of the data. The
EPA analysis failed to use data labeled “Swamper.” This is an
alternative name for co-pilot. These data have been included in the
corrected data as part of the co-pilot data. The data on tarp cutters
was also omitted. These data are also included in the MBIFs analysis.

Field 50 is also incorrectly listed in Appendix C and differs from the
TriCal Inc. (1987) reference. The data actually come from a document
defined as “Results of Worker Monitoring of Deep Tarpless Application
at Various Sites in CA, by Rick Stange Sponsored by TriCal,
04/03/1986.”

A measurement point from a grab sample taken when an applicator fixed a
connection is listed as an applicator exposure. This measurement is too
short a duration (1 minute) to be indicative of the 8 hr TWA exposures.
This data point should not be used

EPA’s analysis.

Finally it is unclear why a “Shovelman” exposure is reported when
the study is described as Tarpless.

• Additional Errors -- Miscellaneous

In Appendix T, EPA Lists the same data from Field 4 under both the
“Second Tractor Driver” and “Irrigation Worker” job
classifications. This is incorrect. In the MBIP’s reanalysis, the data
have been listed under “Irrigation Worker.”	See response to comment
64 above related to correcting errors in the exposure monitoring data. 
Essentially, the Agency acknowledges most of the errors cited pertaining
to specific studies on pages 29 to 35 of the MBIP comments document. 
The Agency has attempted to correct these errors in the revised Phase 5
risk assessment document.  In fact, in table 6 of the MBIP comments (pg
35) revised estimates of exposure were presented as a “corrected
version”.  The revised Agency estimates are similar to the values
presented by MBIP.  In some cases, exact agreement is noted while in
others values vary but only based on data from a few worker monitoring
events.

66



Temporal Trends in Worker Exposures Due to Improved Technology

The available monitoring data on worker exposure levels present a strong
case that technological changes in California have greatly reduced
exposure to methyl bromide over the period of 1981 to 1994. The
following figure presents the maximum values for each of the seven job
classifications as given by each of the available monitoring studies.
Where a study demonstrates the use of a technology that is now in common
use in California, only the data from the samples that reflect the use
of that technology are presented. Thus, in Field 4 only data from the
fields that used the rapid closure technology (closing shoes followed by
a roller that compacts the soil prior to tarp placement) are used’. In
Field 23 only data from the fields treated with a noble plow are used.
As the figure shows there is a strong trend of reduced exposure over
time. Data from the early 1990s indicates that when controls are used,
most of the jobs are below 1 ppm. This finding is expected because CDPR
lowered the permissible exposure to Methyl Bromide to a 24-hour TWA of
210 ppb in the early 1990’s.	See responses to comments 64 and 65
above.  While use of the available exposure monitoring data could
possibly not be representative of current cultural practices in
California, it is difficult to justify not using the available data on a
national level.  This is because the same regulatory process which has
occurred in California has not transpired in other regions of the
country so it is possible that current application practices in other
regions have not similarly evolved.

67

	127

	Changes in Engineering and Practice

In the 1980s a number of engineering practices in soil fumigation were
modified to reduce exposures. Some of these practices were captured in s
guidance on the use of methyl bromide

When methyl bromide is injected into soil the injection is performed by
a series of injections at the bottom of the shanks of plows pulled
behind a tractor. The tractor pulls the application equipment across a
field. Once the equipment reached the end of the field, the plows and
injection equipment must be raised to allow the tractor to swing around
and begin plowing the next section of the field. In the early 1980s the
practice was to shut off the supply of methyl bromide into the injection
equipment prior to raising the plows. This was done using a “shut-off
valve.” This prevented the spraying of methyl bromide into the air
when the injection shanks were in a raised position. However, a
substantial amount of methyl bromide remained in the tubes leading down
to the shanks. This residual methyl bromide in the injection equipment
would escape during the time the tractor repositioned itself and began
plowing the next portion of the field. These releases occurred each time
the tractor reached the end of the field and resulted in increased
exposures to all employees at the site.

In the mid 1980s, the injection equipment was modified to allow
compressed air to flush the methyl bromide from the shanks prior to
lifting the shanks from the soil. This flushing was performed after the
cutoff valve was closed and before the shanks were raised from the soil.
As a result, methyl bromide is not released to the air.

A second modification that also occurred during this period was the
requirement that engine fans of the tractor blow back towards the
injection equipment. Prior to this modification certain types of
tractors used to pull the injection equipment (Caterpillar tractors) had
fans that pulled air from the rear of the tractor over the engine and
blew the air out of the front of the vehicle. This practice “pulled”
air over the injection equipment and carried the methyl bromide towards
the driver and co-pilot. In the mid 1980s, this mechanism of exposure
was identified and the tractors were modified to blow air from the front
of the tractor back across the engine to the rear of the tractor. This
caused any methyl bromide released from the injection equipment to be
blown away from the driver and co-pilot and resulted in reduced
exposures to these jobs.

In addition to these changes, in the early 1 990s CDPR required the use
of tarps with low permeability to Methyl Bromide’ Specifically these
regulations require that tarps used for methyl bromide soil fumigations
in California shall have a permeability factor between 5 and 8
milliliters per hour per square meter per 1000 ppm of methyl bromide at
30 degrees Celsius and be approved by the Department. Tarps used prior
to this time had much higher permeability and as a result general air
levels at treated fields were higher in the 1980s than after the 1990s.
See responses to comments 64 through 66 above.  While use of the
available exposure monitoring data could possibly not be representative
of current cultural practices in California, it is difficult to justify
not using the available data on a national level.  This is because the
same regulatory process which has occurred in California has not
transpired in other regions of the country so it is possible that
current application practices in other regions have not similarly
evolved.

68

	127

	Decreased Application Rates will Reduce Occupational Exposures

The existing monitoring data are all more than 12 years old. Typical
application rates of Methyl Bromide are now significantly lower then the
levels in the monitoring studies performed in the 1980s and 1990s. As
discussed in Section II above, typical rate for soil fumigation are now
below 200 lb/acre. As the data in Attachment 4 shows, almost all of the
monitoring data utilized by EPA in the PRA reflects usage rates from
greater than 200 lbs/acre and includes data from fields where the
application rates were as high as 400 lbs/acre.

The role of application rates as a factor in methyl bromide exposure has
been discussed by a number of authors (Maddy et al. Abdalla et al.’
and higher levels of exposure are believed to occur when application
rates are high. Quantitative evidence of the role of application rate
and air concentrations can be seen in the monitoring data.

In Field 23 (Siemer and Associates, Inc., 1992a), First Tractor Drivers,
Co-Pilots, and Shovelman were monitored. The purpose of the study was to
demonstrate the reduction of exposure from the use of the Nobel plow.
However, as the following three figures show the exposures to all three
jobs declines as the application rates decline. This decline occurs for
both types of plow. These data show that as the application rates
approach 200 lbs/acre, the measured exposures to methyl bromide drop to
less than 0.5 ppm for all three jobs.

Similar patterns can be seen in other studies. The finding that
concentrations drop with application rates indicates that the current
use rates of methyl bromide result in lower exposures than the studies
from the 1980s and 1990s indicate. Since the studies from the 1990s
indicate that uses in the range of 200-400 lbs can meet the current 1
ppm criterion, future levels of exposure are likely to be well below 1
ppm.	It is intuitive that rate reductions would reduce exposure levels
but these values can be impacted by many other factors such as soil
type, shape and orientation of the injection shank, soil moisture, or
attention paid by the tractor driver with regard to end of row turning
activities.  As such, no attempt has been made to extrapolate using the
available data to possibly lower application rates.  If MBIP monitored
exposures at lower application rates, such data would be considered for
exploring the possible impact of this phenomenon.





69

	127

	Use of Short Term Data to Predict 8-hour TWA

In the PRA, EPA has assumed that the air concentrations reported by the
authors of the various studies can be assumed to be predictors of 8 hr
TWA. This is incorrect. The range of sampling times for these
measurements has ranged from less than 30 minutes up to six hours with
the majority of the samples less than three hours in duration. Since
short-term sampling is taken at those times where the potential for
exposure is the highest, the values reported may be higher than the
8-hour TWA exposure received by the worker. This occurs for two reasons.
First, if the sample was taken over a period of relatively high exposure
(correction of a malfunction of a piece of equipment) then the
measurement could over estimate the average exposure during the time of
injection.

Second, any measure of exposure taken during the time of actual
injection of methyl bromide will overestimate the 8-hr TWA, since methyl
bromide is not injected for 8 hours a day. Soil fumigation requires that
large pieces of equipment be transported to the site of the soil
fumigation. Once there, equipment preparation and set up requires
approximately 1 - 1.5 hours. Application of methyl bromide must cease
during lunch. At the end of the day shut down and removal processes
require approximately 1 hour. Thus in a nine hour working day only 5.5
hours of injection occurs (personal communication Tom Dufala, 2005).

Evidence that applications occur for less than 8 h per day can be seen
in the sampling times from a number of the recent surveys where workers
were monitored for an entire injection period. The durations in these
surveys are typically in the range of 4-5 hours. Additional evidence for
a 5.5 h injection time is seen in a 1998 study that investigated the
temporal patterns of worker exposure in the methyl bromide soil
fumigation industry (MBIP, 1998’s). In this study, data on the number
of acres treated by each of 47 workers in each month of a 1 2-month
period (May 1997- April 1998) was analyzed. The time necessary to treat
an acre and the assumption the injection time will be limited to 5.5 hrs
per day were used to solve for how many days in the month the worker was
injecting methyl bromide. In every case, the total number of acres
treated in a given month by a worker did not exceed the number that
could be treated in a 5.5 hr workday (given a 5-6 day workweek).

For both of these reasons the finding that a specific sample is greater
than 1 ppm cannot be taken as evidence that the worker has had an 8 hour
TWA exposure of greater than 1 ppm.	In many circumstances, the
measurements used in this analysis were collected over shorter durations
of exposure that do not represent 8 hour time weighted averages.  It is
expected, however, that the resulting monitoring values represent an
exposure rate for that particular monitored period.  It follows that if
an individual worked for 8 hours engaged in doing the same specific task
that the exposure rate would not change regardless of the duration
worked.  Given the size and scale of methyl bromide use in agriculture
where 8 hour work days for specific tasks could be expected, it is
reasonable to use monitored exposure rates (i.e., the measured
concentrations) to calculate risks using 8 hour HECs.

70

	127

	Comments on Specific Job Classifications in the PRA

EPA establishes seven job classifications in its assessment of soil
fumigation exposures. The following section presents the specific
comments on these seven job classifications. As the comments indicate,
two of the jobs classes (second tractor, and irrigation worker) should
be dropped. As the comments indicate MIBP believes that the available
data indicate that the current exposures for the remaining five jobs are
not likely to exceed 1 ppm.

• First Tractor Driver

There is a large amount of data for the First Tractor Driver or
Applicator. As data on historical trends indicates, the exposures to
methyl bromide for this job have been reduced to levels below 1 ppm in
the early 1990s. This reduction was achieved by improved injection
methods, use of high barrier films, modifications in the engine fans,
and the introduction of ventilation fans in the driver’s cabin.
Moreover, current levels are expected to be lower than the levels in the
studies from the early 1 990s since application rates have declined over
the last 10 years.

• Co-pilot

The co-pilot’s exposure tends to be more variable than the driver’s
reflecting the wider range of tasks performed by the co-pilot. However,
the co-pilots exposures have been reduced over time by the use of better
injection methodologies, improved barrier films, and modifications in
the engine fans of the tractors. As in the case of the tractor driver
and applicator, current levels are expected to be lower than the levels
in the studies from the early 1990s since application rates have
declined over the last 10 years.

• Shovelman

Exposures to these workers are controlled by improved injection
techniques and by use of high barrier films. Current levels are expected
to be lower than the levels in the studies from the early 1 990s since
application rates have declined over the last 10 years.

• Tarp Cutter and Tarp Remover

Tarp cutters and tarp removers have a potential for exposure that is
potentially increased by the use of high barrier films (tmpping more
methyl bromide). However as the study listed as Field 20 indicates,
these exposures can be greatly minimized if the cutting of the tarps is
performed on one day and the tarps are removed on a subsequent day.
Cutting the tarps allows the residual methyl bromide to dissipate prior
to removal of the tarps. Current levels are expected to be lower than
the levels in the studies from the early 1990s since application rates
have declined over the last 10 years.

• Second Tractor Driver

MBIP recommends that EPA drop this classification for the following
reasons. First, the Pipe Layer and Pipe Tractor workers are listed under
the “Irrigation Workers” job category and should not be considered
under this category. Second, the classification includes measurements at
are no longer common practice. The reference to “Tarp Layer” occurs
in the study listed as “Field 3”. In this study, a second tractor
applied the tarp following the tractor that injected the Methyl Bromide.
This practice is not common today. The reference to “Cultipacker
Driver” occurs in the study listed as “Field 22”. In this study, a
second tractor pulls a cultipacker that closes the soil furrows
following injection of Methyl Bromide. Again this practice is not common
today. Under current practices, a single tractor pulls equipment that
both injects the Methyl Bromide, and lays the tarp in one operation.

Third, the remaining workers, “Drip Tape”, and “Plastic Truck
Driver” represent two very different jobs. The “Drip Tape Layer”
is the driver of a tractor lays a drip line, or drip tape, in the field
prior to the injection of the Methyl Bromide and tarp placement. The
“Plastic Truck Driver” delivers the rolls of plastic tarp to the
tractor but does not enter the treated field. The differences between
these two tasks makes the grouping useless in characterizing
occupational exposures. Fourth, there is only a single measurement for
each of the categories. This is too limited a data set to reach any
conclusions.

Finally, the various jobs occur at locations that are farther from the
soil injection process than the locations of the First Tractor Driver,
Co-pilot, and Shovelman and are likely to have exposures that are lower
than those jobs. Mitigations that protect these three exposure
categories will also protect the more distant workers.

• Irrigation Worker

The irrigation workers are the pipe layers and the pipe tractor drivers.
Data on these workers comes from two studies of raised bed tarped
applications (Fields 4 and 22). These workers were involved in placing
water on top of the tarp as a means of forcing the tarp to remain on the
ground. In these instances, the tarp remains on the field as mulch. This
practice is not commonly performed today. Because of this change in
practices there is no need to evaluate this job classification.	The job
tasks that are included in the risk assessment are reflective of the
tasks which were monitored in the occupational exposure data.  No
information has been submitted to the Agency which suggests that any of
these tasks are not completed in current cultural practices.  The Agency
acknowledges that certain tasks may be completed less frequently than
once occurred (e.g., second tractor driver as cited by MBIP).  However,
it is clear that such activities still can occur even though they would
be infrequent.  In fact, in describing these tasks MBIP uses the term
“is not common today” and not something like this does not occur in
modern agriculture.  Based on the available information, the Agency will
continue to include risks associated with these tasks but certainly
recognizes that they may occur less frequently than in the past.  

In its revised risk analysis, the Agency has also corrected the
mischaracterizations of job tasks as cited by MBIP.

71

	127

	Overall Conclusions on Occupational Exposures for Pre-plant Soil
Fumigation

In short, the available data on methyl bromide exposures resulting from
soil fumigation indicate that the current exposures are significantly
lower than the estimates given in the PRA. The MBIP urges EPA to take
the foregoing comments into account and reconsider its conclusions on
potential occupational risks from soil applications.	The Agency does not
concur with this conclusion since changes driven by regulation have
occurred in California but it is not clear that a similar evolution has
occurred for other regions of the country.  Again, additional exposure
monitoring data would be considered that are more reflective of current
practices on a national level should they become available.

86	128	There is insufficient detail to follow or comment on the indirect
back-calculation method for determining flux. Detail lacking includes
the field monitoring studies used, the criteria for combining or
eliminating study results, and the comparison of study results. The flux
calculation and hence the exposure modeling calculation are therefore
not transparent.	This has been corrected in the updated Phase 5 risk
assessment document.

87	128	The model (including its calibration through the back-calculation
of flux values) was not corroborated with known major poisoning
incidents. Demonstrating that the calibrated model correctly predicts
past incidents would increase confidence that the model results will
correctly validate risk mitigation strategies and prevent future
incidents.	The Agency has attempted in the revised Phase 5 risk
assessment to better corroborate PERFUM outputs with monitoring data and
incident events.  As is explained in the risk assessment one might not
necessarily expect to be able to corroborate modeling outputs with
specific incident events because key modeling inputs for that particular
situation would not be available.  For example, emissions measurements
for the specific treated field and the weather conditions at the time of
the incident would not be available which would be necessary for a
correct modeling prediction of the results of an incident.  Also,
comparison of the modeling results to incidents is also difficult
because by definition the model is determining the distance at which a
target concentration is achieved at the NOAEL HEC adjusted by an
uncertainty factor (i.e., 30).  This means that the target concentration
at the Agency level of concern would be 30 times less than the HEC at
the NOAEL.  At such low levels, one would not expect to see incidents. 
Conversely, there could potentially be under-reporting issues associated
with methyl bromide incidents because it is not a sensory irritant like
other fumigant chemicals which means that even though individuals could
be exposed at high levels of methyl bromide it could possibly be hard
for them to attribute their symptoms, if any, to a chemical
over-exposure.



92	129	… one problem with the current modeling is that it does not
take the seasonality of applications by crop into account . For example
more than 75% of the pre-plant MeBr applications in almonds occur
Nov-Jan .

In 2002 5%, in 2003 9%, and in 2004 13% of the applications took place
between May 1 and October 31, and most of those were in October. Given
that rain fall, air temperature and winds are very different during the
winter months from the summer months in the Central Valley, accounting
for seasonality of applications by crop is critical to understand the
actual bystander exposure risks posed.	This comment is correct in that
seasonal periods (i.e., several months in a row) were not quantitatively
evaluated in the risk assessment.  The results which are summarized in
the document are for 5 years of sequential weather data which is
consistent with Agency air modeling policy.  If so desired, monthly
outputs which are included in all PERFUM analyses, could be consulted to
evaluate how differences in monthly weather patterns could impact
results.  PERFUM output files can be provided for any emissions/weather
station combination considered in the assessment.

93	129	Nor does the model account for the days when methyl bromide
cannot be applied due to , the weather conditions. Any MeBr application
requires a permit from the county ag. commissioner's office as well as a
24 h pre-notification of the planned application. If there are inversion
weather conditions the ag . commissioner will not allow the fumigation.
Similarly if the soil is too moist, which leads to poor penetration, the
ag. commissioners will not permit the application . Thus, there is not
an equal probability at methyl bromide will be applied any of the 365
days of the year by almond growers as currently assumed in the model .
It is unclear whether EPA made any effort to exclude weather conditions
where MeBr would not be applied in their calculations and thus in the
data presented as bystander exposure.	The comment is correct in that no
effort was made to account for “days when methyl bromide cannot be
applied due to weather conditions.”  The goal in the assessment was to
evaluate all potential operational conditions where methyl bromide could
possibly be used.  It should also be noted that outside of California,
there are not the same kinds of use restrictions in other regions of the
country.  Also, it is not clear what specific weather conditions and
field conditions would prohibit an application.  Many applicators can
recognize possible inversion conditions but these are extremely site
specific.  For modeling purposes in order to reduce peak emission
periods, and not just overall emissions, a myriad of possible factors
would have to be controlled for including field conditions such as
moisture levels and soil type.

PERFUM provides results at varying percentiles of exposure.  The Agency
has been considering how to account for possible inversion and other
prohibitive site or weather conditions in its risk management decision
making process where the frequency of prohibitive conditions could
impact the selection of the percentile of exposure for regulatory
purposes.

95	129	The pre-plant applications of methyl bromide in almonds are
exclusively done by custom applications as required by the State of
California . Thus, it is likely that a single applicator will be exposed
to methyl bromide several days in a row . However, the state has also
recently introduced limitations to the number of hours and acres that a
single applicator may treat in a day to reduce worker exposure. The
typical application will involve the applicator, who also does the
mixing and loading, along with a supervisor.

The presence of a second trained person is required by California for
monitoring. Most commonly a third person will be disking the soil as a
seal several rows behind the applicator. Tarps are not used in the
almond industry. EPA does not discuss the single tree-hole treatment.
These applications are typically done by the growers or employees. The
state requires special training of anyone working with methyl bromide as
well as the presence of two persons during the application. The methyl
bromide is injected into the prepped hole (loose soil) ideally at 2-3 ft
depth using a injector attached to a pressurized canister. As mentioned
above the typical rate is 1 lb/hole. This is method is most commonly
used to treat replants of young trees in an existing orchard that have
died due to soil pest issues.	The shank injection method for orchard
replant is discussed in the current risk assessment and it is believed
that the results for pre-plant soil analyses can be used to determine
appropriate buffer zone estimates for that type of use.  Single tree
hole applications were also discussed in the revised risk assessment. 
Typically tree hole treatments with handheld equipment are infrequent. 
If single (or a few) trees are being replaced, a handheld injector is
typically used.  No emissions data are available for this pattern,
however, it is believed that this technique presents minimal risks for
bystander populations because of the low amount of material used.  

萏֠萑褐葞֠葠褐摧䒌4

0

1

t

u

À

Ä

È

É

×

ñ

õ

û

(

1

kd

kd£

愀Ȥv欀।

愀Ȥv欀뱤

h

@

A

µ

¶

á

ã

[

A

B

¶

hB

hB

nt.  It is difficult to ascertain the possible level of risk associated
with single tree hole applications because of the handheld equipment
used or the proximity to the application.  In fact, there are likely to
be other specialized scenarios that haven't been addressed as well in
the assessment similar to the single tree hole uses due to a lack of
monitoring data.  For this reason, additional exposure monitoring data
may be required to address the potential for risks from the single tree
hole scenario, as well as others that are identified, in the risk
management process in order to ensure that methyl bromide users are
adequately protected.

96	129	EPA discusses the bystander exposure both in terms of the maximum
buffer distances from the field where the MOE is achieved as well as
whole field buffer distances.

Unfortunately it is still very difficult to understand what the whole
field buffer distances mean both mathematically as well as in reality.
If EPA would like to use whole field buffer calculations in the risk
assessments and risk mitigation, then a greater effort is needed to
educate readers on what the whole field buffers represent.	See response
to comment #59





Page   PAGE  3 

