MEMORANDUM

National Volatile Organic Compound Emission

Standards for Aerosol Coatings

Review of the Technical Basis for use of the One-dimensional MIR Scale
in the National Volatile Organic Compound Emission Standards for Aerosol
Coatings

Prepared by:

Deborah Luecken

Office of Research and Development

National Exposure Research Laboratory

Atmospheric Modeling /Human Exposure and Atmospheric Sciences Division

Prepared for:

J. Kaye Whitfield

Office of Air Quality Planning and Standards

Sector Policies and Programs Division

Research Triangle Park, North Carolina 27711

March 15, 2007

UNITED STATES ENVIRONMENT ADVANCE \l 0 AL PROTECTION AGENCY

NATIONAL EXPOSURE RESEARCH LABORATORY

Research Triangle Park, NC 27711

										Office of

									  Research and Development

						March 15, 2007

MEMORANDUM

SUBJECT:	Review of the Technical Basis for use of the One-dimensional
MIR Scale in the National Volatile Organic Compound Emission Standards
for Aerosol Coatings

FROM:	Deborah Luecken

Atmospheric Modeling /Human Exposure and Atmospheric Sciences Division
National Exposure Research Laboratory, ORD 

TO:		J. Kaye Whitfield

		Project Lead,   

		Office of Air Quality Planning and Standards

		Sector Policies and Programs Division	

		Natural Resources and Commerce Group		

CC: 		J. David Mobley

		Associate Director, Atmospheric Modeling Division

		National Exposure Research Laboratory, ORD 

		Larry Cupitt

		Associate Director, Human Exposure Research

		National Exposure Research Laboratory, ORD

		

I have reviewed the literature on reactivity to determine if it supports
the use of a reactivity-based control for aerosol coatings, and in
particular, the use of a single, one-dimensional, box model-based
Maximum Incremental Reactivity (MIR) scale for national applicability. 
While I believe that we should continue to do more research work in this
area to ensure that the best possible metric/scale is being employed,
there’s evidence that, given what we already know, EPA can already
make a compelling case for considering reactivity as a basis for VOC
control. While no single scale will be a perfect predictor of ozone in
every location at every time, a single one-dimensional, box model-based
MIR scale, is remarkably consistent with the more complex,
three-dimensional, Air Quality Model-based scales.  The
three-dimensional scales are attractive for their robustness and general
applicability, but at the current time, the one-dimensional MIR scale
has several advantages.  We have a history of implementation and
scientific review, as well as existing product formulations using the
MIR, based on the 2001 CARB rule.  We have peer-reviewed values for over
800 compounds based on the box model MIR, but for only about 30 explicit
compounds from the three-dimensional models.   Because it is much easier
to implement, the box model-based MIR scale gives us a way to easily
incorporate updates to the chemical mechanism and information on new
chemical species that are not part of the existing 800 compounds.  

In the following pages, I list four of the technical questions that we
need to address in order to proceed with use of one-dimensional MIR
values in a national regulation, and summarize how the scientific
literature addresses these issues.  Based on this review, I believe it
is reasonable to proceed with the regulation, with the caveat that we
allow the flexibility to modify the scale as the science improves.  I
include a more detailed discussion, plots, and references in an attached
report to this memo. 

Question 1:  Does the mechanism used to derive MIRs accurately
characterize ozone formation from individual VOCs?

In general, the SAPRC99 chemical mechanism which has been used to
predict the reactivity of ozone formation for VOCs, and derive the MIRs,
has been shown to be a valid predictor of ozone formation under
controlled evaluation studies.  Ozone formation in the atmosphere is a
complex process and there is still some uncertainty in the mechanism
predictions.  While the mechanisms may underpredict ozone at low VOC/NOx
ratios, which are of most interest to this regulation, use of SAPRC99
minimizes this bias, and the bias errs on the side of being slightly
conservative.  There is some uncertainty in the detailed chemical
mechanisms used to calculate reactivities of individual VOCs, but for
the compounds used most often in aerosol coatings, this uncertainty is
estimated to be largely limited to categories 1 or 2.  I believe this is
equal to (probably less than) the uncertainties involved in overall
calculations of ozone changes due to mass reductions.

Question 2:  Is the MIR a reasonable way to define reactivity?

All of the different metrics that have been proposed, approximately 4
box model-based metrics and 8+ three-dimensional metrics, give generally
consistent predictions in the relative ranking of reactivities.  None of
the detailed studies performed to date have shown serious deficiencies
in the long-used MIR metric.  The aldehydes are sometimes outliers, but
these are not important components of aerosol coatings.   

The MIR describes ozone formation in areas where ozone is most sensitive
to VOC emissions, in upwind areas with high emissions.  It is
complementary to NOx control programs which address ozone formation in
NOx-limited areas.  The box model-based MIR scale gives the largest
difference between the high reacting compounds and the lower reacting
compounds than any of the other scales.  On the negative side, this
allows a larger amount of overall mass to possibly be emitted in a
reactivity-limited solvent, as compared with an equal mass-restricted
solvent, which would benefit ozone in the immediate area, but may
increase ozone slightly downwind (see Question 4 for further
discussion).  On the positive side, it provides more incentive for
manufacturers to use low-reactivity chemicals in aerosol coatings –
the major goal of this policy.   

In addition, the MIR has a long history of use and analysis, and is the
basis of some existing paint formulations under California law.  Given
the uncertainties inherent in any air quality modeling, is a reasonable
way to characterize reactivity.   The three-dimensional AQM metrics are
arguably more robust, but there are no detailed reactivity scales yet
available for all of the possible chemicals involved in aerosol coatings
using AQMs.

Question 3:  Does a single MIR scale account for potential spatial
variability across the United States?

Several three-dimensional modeling studies have shown that spatial
variations in reactivity are generally found to be small, and averaging
time used in calculation of the scale (whether 1 hour or 8 hours) makes
little difference in the values.  When viewed on a relative basis, as is
applicable for solvent mixtures, the one-dimensional, city-specific MIR
values are approximately the same for the 39 different cities studied. 
This gives us confidence that one metric and scale can be used with
equal applicability throughout the country.  

Question 4:  Is the large effective range of the MIR scale a potential
problem when replacing high reactivity compounds with lower reactivity
chemicals?

A major benefit as well as a concern with the use of MIR is its large
effective range.  The box model-based MIR scale gives the largest
difference between the high reacting compounds and the lower reacting
compounds.  This means that a substitution using MIR would result in a
larger permissible amount of mass than using another metric, i.e. the
MOIR.  Compared with an equal mass-restricted solvent, would benefit
ozone in the immediate area but may have some disbenefits downwind.  

Current studies indicate that some increases in ozone due to increased
mass of low-reacting compounds are possible, but the studies have only
looked at very extreme cases (i.e. substitution of all VOCs with
equal-ozone amounts of ethane, tracking ozone formation downwind from
substitutions only in an urban area). While ozone could increase due to
these upwind substitutions, the increases tended to be much smaller (by
a factor of 12-20) than the magnitude of concurrent ozone decreases. 
The substitutions had a larger effect on reducing the higher ozone
concentrations than they did on increasing downwind concentrations. Even
in these extreme cases, the benefits for ozone (reduction in ozone peak)
were significant.  

Inherent in this rule is a reduction in ozone equivalent to a reduction
in mass, therefore the substitutions will be done on a less-than-equal
ozone basis, relative to the pre-rule mixture.  Realistic changes in
formulation, especially if limited to aerosol coatings are unlikely to
result in a noticeable increase in ozone downwind, given that downwind
areas are usually NOx-limited, so small amounts of additional VOCs
won’t influence ozone formation much. 

More detailed and realistic studies are needed to determine whether any
potential increase in mass due to reactivity-based VOC limits are truly
a problem.

