Technical Support Document (TSD)

for the Transport Rule

Docket ID No. EPA-HQ-OAR-2009-0491

Alternative Significant Contribution Approaches Evaluated

U.S. Environmental Protection Agency

Office of Air and Radiation

July 2010

Alternative Significant Contribution Approaches Evaluated

The Transport Rule preamble discusses EPA’s proposed approach to
define emissions that constitute each upwind state’s significant
contribution to nonattainment and interference with maintenance
downwind.   This Technical Support Document (TSD) discusses alternative
approaches that EPA evaluated.  The TSD is organized as follows:

1.  Introduction.

2.  Air quality only approaches.

3.  Cost per ton approach with uniform cost in all states.

4.  Cost per air quality impact.

5.  “Binning” of states based on air quality impact.

1. Introduction

Section 110(a)(2)(D)(i)(I) in the Clean Air Act (CAA) requires states to
prohibit emissions that contribute significantly to nonattainment in, or
interfere with maintenance by, any other state with respect to any
primary or secondary NAAQS.  The proposed Transport Rule would define
these prohibited upwind state emissions with respect to the existing
8-hour ozone and fine particle (PM2.5) National Ambient Air Quality
Standards (NAAQS), i.e., the 1997 8-hour ozone and annual PM2.5 NAAQS
and the 2006 24-hour PM2.5 NAAQS.  Section IV.C and IV.D in the
Transport Rule preamble discusses EPA’s proposed approach to define
emissions that constitute significant contribution and interference with
maintenance in detail.

 EPA developed and evaluated a number of different approaches attempting
to quantify the emissions that are defined as each state’s significant
contribution and interference with maintenance, consistent with the
judicial opinions interpreting section 110(a)(2)(D)(i)(I) of the Clean
Air Act (see discussion of judicial opinions in preamble section IV.G). 
The Agency also held several “listening sessions” with interested
parties and stakeholder groups and received written comments from
several stakeholder groups.  EPA considered these ideas and suggestions
as it developed alternatives and further explored the most promising
approaches.  

This TSD describes and discusses the alternative approaches EPA
evaluated for defining significant contribution and interference with
maintenance.  EPA is not proposing any of the alternative approaches

The alternative approaches described in this TSD were listed in section
IV.H of the preamble.  EPA evaluated several “general” classes of
approaches.  These include approaches based on: (1) only air quality;
(2) only cost, with a uniform cost in all states; (3) cost per air
quality impact (e.g., dollars per g/m3 or dollars per ppb); and (4)
binning of states based on air quality impact.  Often, multiple
variations on each of the “general” approach classes were evaluated.
 Each of the general approaches is described in a separate section,
below. 

Sections 2, 3, 4, and 5, below, describe the approaches.

  

2. Air quality only approaches.

A number of air quality only approaches were considered.  Generally,
these approaches utilize air quality contribution modeling, then
determine reductions in emissions based solely on the results of this
modeling. These approaches include the “fixed air quality threshold
limit” approach and the “cumulative air quality threshold”
approach.  

a. Fixed air quality threshold limit approach: Under this approach, EPA
would define a threshold limit value for an air quality contribution
from an upwind state to a downwind monitor.  All upwind states would
have to reduce emissions such that their contribution to all downwind
monitors is at (or below) the level of the threshold.  The amount of air
quality contribution above the threshold limit would determine the
emission reduction required.  The amount of air quality impact above the
threshold limit to the maximally-impacted downwind
nonattainment/maintenance monitor would determine the amount of
significant contribution for the upwind state.  The contributions from
an upwind state to its maximal downwind nonattainment and maintenance
monitor can be found in Tables IV.C-13, 16, and 19 of the preamble.  

Generally, the largest contributors to a downwind monitor are from
nearby states (and the contributions can be quite large). Since there
are a number of nonattainment and maintenance monitors at locations
throughout the eastern United States, most upwind states have a
substantial air quality contribution to at least one monitoring site in
a downwind state (as an example, Table 2-1 summarizes the air quality
contributions for annual PM2.5).  As a result, assessed across all
states and monitors, the magnitude of the emissions reductions necessary
to reduce these contributions to a lower threshold level (e.g., 0.15
µg/m3) for nearly all of the upwind states would be very large and, by
consequence, lead to a cumulative reduction of air quality impact well
beyond what would be required for each individual downwind monitor to
reach attainment/maintenance.  The magnitude of emissions reductions
necessary to reduce the emissions to a higher threshold level (e.g., 1.0
µg/m3) results in fewer states making reductions, with the cumulative
air quality improvements possibly insufficient for each individual
downwind monitor to reach attainment/maintenance.  The emissions
reductions for some of the states might still be very large.

g/m3), or as specified in the Transport Rule proposed approach (i.e.,
defined as 1% of the NAAQS or 0.15 g/m3 for the annual PM2.5
standard) result in emissions reductions from nearly all upwind states
well beyond what would be needed for all of the downwind areas to attain
the standards and possibly exceeding what might be achievable through
available control measures.  For example, if a state were required to
reduce its NOx emissions from all sources by 92.5% (as described in the
example above), this would require elimination of nearly all NOx
emissions from the power generation sector and impose dramatic
reductions across many other sectors (including mobile).  

Table 2-1.  Maximum Downwind Air Quality Contributions to Nonattainment
or Maintenance for Annual PM2.5 (g/m3) and the Percent Reductions in
All Anthropogenic SO2 and NOx Emissions that would be Needed to Reduce
the Maximum Contributions to Threshold Levels of 0.15 g/m3 and 0.5
g/m3, Respectively.

Upwind State	Maximum Downwind Contribution to Nonattainment or
Maintenance for

Annual PM2.5 (g/m3)	 	Percent Reduction if Threshold is 0.15 g/m3
 	Percent Reduction if Threshold is 0.50 g/m3 

Alabama 	0.46	 	67%

0%

Arkansas 	0.09	 	0%

0%

Connecticut 	0.09	 	0%

0%

Delaware 	0.20	 	25%

0%

Florida 	0.29	 	48%

0%

Georgia 	0.63	 	76%

21%

Illinois 	1.01	 	85%

50%

Indiana 	2.09	 	93%

76%

Iowa 	0.31	 	52%

0%

Kansas 	0.09	 	0%

0%

Kentucky 	1.68	 	91%

70%

Louisiana 	0.34	 	56%

0%

Maine 	0.02	 	0%

0%

Maryland 	0.62	 	76%

19%

Massachusetts 	0.13	 	0%

0%

Michigan 	0.72	 	79%

31%

Minnesota 	0.19	 	21%

0%

Mississippi 	0.07	 	0%

0%

Missouri 	1.38	 	89%

64%

Nebraska 	0.08	 	0%

0%

New Hampshire 	0.02	 	0%

0%

New Jersey 	0.68	 	78%

26%

New York 	0.49	 	69%

0%

North Carolina 	0.19	 	21%

0%

North Dakota 	0.05	 	0%

0%

Ohio 	2.03	 	93%

75%

Oklahoma 	0.08	 	0%

0%

Pennsylvania 	1.60	 	91%

69%

Rhode Island 	0.01	 	0%

0%

South Carolina 	0.26	 	42%

0%

South Dakota 	0.02	 	0%

0%

Tennessee 	0.68	 	78%

26%

Texas 	0.13	 	0%

0%

Vermont 	0	 	0%

0%

Virginia 	0.37	 	59%

0%

West Virginia 	1.17	 	87%

57%

Wisconsin 	0.46	 	67%

0%



 Furthermore, the emissions reductions necessary to reduce all states’
contributions below a 0.15 µg/m3 threshold would be well beyond what
would be needed for all of the downwind areas to attain the standards. 
Likewise, if the threshold levels were increased to substantially higher
levels (e.g., 0.5 g/m3, as seen in Table 2-1), while the consequence
would be that fewer states would be required to make reductions (e.g.,
12 states compared with 23 states at thresholds of 0.5 and 0.15,
respectively), the emissions reductions required of those states would
still be significant since several states have large air quality
contributions to their maximally-impacted downwind monitors (Table 2-1).
 

One of the technical challenges of this approach is that there is not a
linkage between the levels of emissions reductions for each upwind state
and the air quality levels at each of the downwind monitors.  This
approach does not consider the cumulative air quality impact of
emissions reductions in multiple upwind states.  Consequently, depending
on the threshold chosen (e.g., 0.15 µg/m3), there could be substantial
over-control compared with cumulative air quality improvements needed
for all monitoring locations to meet the NAAQS standards.  

b. Cumulative air quality threshold approach:  This air quality-only
approach recognizes the multi-state nature of the sources contributing
to downwind air quality.  It uses an air quality threshold to identify
states making a significant contribution; all upwind states contributing
above the threshold would be required to reduce their individual air
quality contribution by an amount that is proportional to their
contribution.  

For example, consider a downwind monitoring location that is 1.2 µg/m3
above the NAAQS standard and is receiving 4.8 µg/m3 from all upwind
states combined (states A, B, C, and D) which are contributing 2.4, 1.2,
0.8, and 0.4 µg/m3, respectively.  The cumulative air quality approach
would guarantee that a downwind monitoring location would receive a
defined air quality improvement or, alternatively, would not receive
more than a defined contribution from all upwind states.

A requirement could be made that the location receive 1.2 µg/m3 of
relief from the upwind states, or alternatively, that the upwind states
could contribute no more than 3.6 µg/m3 to that location, with
reductions made in proportion to the upwind state’s individual
contributions.  Since the amount of relief is simply the remaining
contribution subtracted from the original total contribution, these two
cumulative approach variations produce identical results.

For both cumulative approaches, the result is that all states would be
required to reduce their emissions by the same percentage (which would
reduce their impact in direct proportion to their original
contribution).  Using the example contributions, with a 25% overall
reduction in the upwind contribution (1.2 µg/m3 out of 4.8 µg/m3), the
new contribution of states A, B, C, and D would be 75% of their original
contributions (or 25% less than their original contributions) or 1.8,
0.9, 0.6, and 0.3 µg/m3, respectively.

However, since all contributing states would be required to do the same
percent reduction of existing emissions, states that had previously
implemented stringent control programs might not be able to achieve the
required reductions using existing control technologies, while others
that had previously done little (and presumably have larger absolute
contributions) would achieve their required reductions using
significantly less than optimal control technologies.

There were two additional difficulties with the cumulative air quality
approaches.  First, it was not clear how to apportion the responsibility
for air quality reductions required from upwind states under section
110(a)(2)(D)(i)(I) relative to the responsibility for the state
containing the monitor.  Second, while it is possible to determine an
emission reduction percentage if there is a single downwind monitor,
most upwind states contribute to multiple downwind monitors (in multiple
states) and would have a different reduction percentage for each one. 
The technical difficulty lies in determining what percent emission
reduction is appropriate for each upwind state.  For example, if the
“maximum” reductions were applied to each upwind state, the
cumulative air quality result would be more than was needed for each
area to meet the NAAQS since the maximum contribution for each upwind
state is to a different monitor.

3. Cost per ton approach with uniform cost in all states.

Variations of a uniform cost per ton approach were evaluated.  This type
of approach has been successfully implemented before, with excellent
environmental results (i.e., the NOx Budget Trading Program).  

For the uniform cost per ton approaches evaluated, if a state’s air
quality impact to a nonattainment/maintenance monitor in a downwind
state was larger than a specified air quality threshold, EPA evaluated
that state’s contribution further to determine what, if any, of that
state’s contribution was “significant”.  The “significant”
portion of a state’s contribution is defined as the portion that could
be eliminated through the application of “cost-effective” controls. 
  “Cost-effective” would be defined using the cost to reduce a ton
of emissions of pollutant, and would be set at a particular dollar per
ton value.  Sources in covered states would then need to apply all
available controls (calculated using the IPM model) to that particular
value.  Following the model of the successful NOx Budget Program
approach, a single cost per ton value for all states in the region would
be identified through analysis of the cost effectiveness of other recent
control actions.  For this approach, the results of the emissions
reductions from each prospective cost per ton value in the cost analysis
would be assessed using air quality modeling.  

Compared with the way that a uniform cost methodology was applied in the
past, which focused on the cost-effectiveness of controls compared with
other rules and programs, EPA believes that a better approach to define
significant contribution and interference with maintenance is to jointly
consider both cost and air quality factors.

   

g/m3 or ppb removed).

Two “cost per air quality impact” approaches were evaluated. These
approaches attempt to balance the cost and air quality impacts of
emission reductions from upwind states affecting a particular downwind
monitor such that each dollar spent to reduce upwind emissions results
in the same downwind impact.  This means that more reductions would be
required in locations where the reductions will have a greater air
quality impact on the downwind monitor and fewer reductions would be
required where they will have less impact.  When air quality problems
are regionally widespread (i.e., there are a lot of monitors, each
requiring substantial air quality improvements to attain the air quality
targets), most upwind areas end up being close to some downwind area
with an air quality problem.  Even though the upwind state would not be
required to reduce a great deal for monitoring sites that are farther
away, it would still have larger reduction requirements for monitors
that are closer.  Therefore, with a larger number of geographically
distributed monitors with air quality problems, the results of such an
approach lead to reduction requirements similar to those of a uniform
cost approach.   

g/m3 reduced).  This was done by dividing the ratio of the marginal
cost per ton of emissions reductions by the air quality impact per ton
of emissions for each state.  All states would be required to achieve
all emissions reductions necessary for the cumulative air quality impact
to equal the cumulative air quality improvement needed for the monitor
to reach attainment/maintenance.  In other words, the total air quality
impact reduction required equaled the sum (across all states) of each
state’s air quality impact per ton of emissions multiplied by the
number of tons it reduced.

As an illustrative example of the first cost per air quality impact
approach, we could consider a particular monitor that is being impacted
by five states (A, B, C, D, and E).  The hypothetical contributions of
sulfate from each state are 2.4, 1.2, 0.8, 0.4, and 0.3 µg/m3,
respectively.  The hypothetical emissions of SO2 from each state are
2,000; 100,000; 8,000; 10,000; and 150,000 tons, respectively.  On a
state-by-state basis, if the sulfate contribution is divided by the
number of tons of emissions, the number of µg/m3 per ton of emissions
equals 0.0012, 0.000012, 0.0001, 0.00004, and 0.000002 for states A, B,
C, D, and E.  The cost-effectiveness of emissions reductions from each
state (on a cost per ton per µg/m3 per ton basis) can be evaluated. 
For this particular monitor, state A has the largest impact per ton of
emissions.  It is most cost-effective to have state A implement the
costliest controls, while the other states implement less expensive
controls.  

For this example, let’s assume that the emissions reductions for all
states lead to a cumulative air quality benefit such that the downwind
monitor achieves the air quality improvement required.  Let’s further
assume that this occurs when state A is implementing all SO2 emissions
reductions up to $10,000 a ton. If state A had to pay $10,000 to reduce
a ton of emissions, the cost per µg/m3 of that ton would be $ 8,333,333
per µg/m3.  This is calculated by dividing $10,000 by the number of
µg/m3 per ton of emissions (i.e., 0.0012 in this example).  The cost
per µg/m3 of this particular ton of reductions from state A is high,
but removing this ton has a large impact on the air quality at the
monitor. For the other states, the same cost per µg/m3 value
($8,333,333 per µg/m3) is reached at much lower marginal cost of
reduction levels: $100, $833.33, $333.33, and $16.67 per ton for states
B, C, D, and E, respectively.  These relatively low marginal cost levels
mean that these states make few emissions reductions, since they are
only making all emission reductions available at or below the marginal
cost levels applicable to their state.  The reason is that, even though
the tons are relatively cheap to reduce, they are not very effective at
reducing pollution levels downwind (the air quality impact per ton
values are small), because the monitor being assessed is farther away.

So far, this procedure has assumed that we identified the “right”
cost per impact level (i.e., $10,000 per ton for state A, leading to a
cost per air quality impact value of $8,333,333 per µg/m3).  How was
this cost per air quality impact value identified?  The procedure
requires an iterative approach, where a particular cost per air quality
impact value is identified, and the result translated to a
state-specific marginal cost of reduction level which, in turn, leads to
a particular emission reduction for each state.  The emissions
reductions are then translated to air quality improvements by
multiplying the emissions reductions by the air quality improvement per
ton of reductions.  The total air quality benefit at the downwind
monitor is the sum of the number of tons each state reduced multiplied
by its air quality impact per ton.  There is a particular cost per air
quality impact value where the cumulative benefit of reduction from all
upwind states will equal the total benefit required for the monitor to
reach the attainment/maintenance target level.     

For a particular monitor, the result of this approach is to ensure that
all of the emissions reductions required for bringing the monitor into
attainment/maintenance are below a particular cost per impact value. 
Thus, this approach should theoretically yield a relatively
cost-effective strategy for attainment/maintenance for a particular
monitor.  

Assessed across all states that impact a particular monitor, the costs
and emission reduction requirements varied widely between states (i.e.,
the marginal cost levels would be very different from each other for
different states).  For a particular state impacting a number of
different monitors, the marginal cost levels and emission reduction
requirements varied widely from monitor to monitor.  One of the
conclusions of the analysis for this approach was that it appeared that,
as part of the cost-effective method for many locations to reach
attainment, the state containing the monitor needed to implement all
available (e.g., up through expensive) control technologies.  This was
because the air quality impact per ton of emissions for the state on its
own local monitor was high.  In contrast, states located far from the
monitor had much smaller impacts per ton of emissions.  Consequently,
only extremely low-cost controls would be required for those upwind
states.

On a single monitor basis, this approach is effective and logical. 
However, it relies on an extremely high level of accuracy in both the
emissions modeling (estimating the magnitude and location of the
emissions reductions) and the air quality modeling (estimating the
downwind effects of the emissions reductions).  EPA concluded that
finer-scale emissions data from all sectors (not just the EGU sector) up
to very high marginal costs and fine-scale air quality modeling could be
needed to resolve differences in cost per air quality impact on the
downwind monitor from all emissions sources in the surrounding area. 
EPA concluded that these data and modeling techniques do not exist
and/or are too computationally demanding to be operationally
implemented.  

Such an approach is particularly challenging in larger states.  The air
quality impact of a power plant on a downwind site can be dependent on
where it is geographically situated.  Accounting for any such impacts
further increases the complexities described above.

When this approach variation was set aside due to the air quality and
emissions modeling technical challenges, significant additional
challenges had been identified and remained unresolved.  One of the
remaining challenges was determining a method to identify a single
reduction requirement for each upwind state from the myriad disparate
reduction requirements from individual monitors that it was affecting in
downwind states, without leading to over or under control when the
cumulative air quality benefits of reductions from all other upwind
states were assessed across the region.   EPA also concluded that it is
not sufficient to select a particular cost per air quality impact value
(i.e., a particular dollar per µg/m3 value), since this value is
determined based on a particular air quality impact value ascribed to a
particular downwind monitor.  The unique critical dollar per µg/m3
values associated with each individual monitor can be used to
back-calculate unique critical marginal costs for each state, for each
monitor.  Consequently, as with the “fixed air quality threshold
limit” air quality only approach, each state would have many different
possible emission reduction levels (one associated with each downwind
monitor).  EPA concluded that it would be difficult to determine each
state’s reduction requirement, since the cumulative air quality
improvement at each downwind monitor was not considered in this step. 
In the illustrative example, state A’s critical marginal cost was
$10,000 per ton.  For a different monitor, located far away from state
A, state A’s critical marginal cost would likely be lower (perhaps
$200 per ton), while state B’s critical marginal cost could be $15,000
per ton (state B is located close to the monitor).  What reductions
should state A and B make?  If both are required to do the maximum, the
cumulative air quality benefit would go well beyond what would have been
needed at either downwind monitor.

The second variation of the “cost per air quality impact” approach
assessed the cumulative cost per total air quality impact, rather than
the incremental benefit at particular marginal cost levels. The total
cost per total air quality benefit was calculated for each state,
monitor linkage, and several modeled emission levels and marginal cost
levels. The total cost was calculated as the sum of the emissions
reductions multiplied by the marginal cost per ton of each emission
reduction.  For example, if the cost was evaluated at a maximum marginal
cost level of $2,400 per ton for SO2 emissions, the incremental costs of
all marginal cost levels below $2,400 per ton were summed to get the
total cost (cost per ton multiplied by the tons of emissions reduced at
each marginal cost level).  The total air quality benefit was calculated
as the g/m3 or ppb concentration reduction at the maximum marginal
cost.  The total cost per total air quality benefit was found by
dividing the total cost by the total air quality benefit.  

g/m3 value of all linkages at $2,400/ton, then the upwind state would
be required to reduce all emissions up to the maximum marginal cost
level ($2,400/ton).  However, if the lowest cost per air quality linkage
at $2,400/ton was higher than the median of all linkages, then the
analysis determined the next highest marginal cost level at which the
lowest cost per air quality linkage was at or below the median.  This
might result in a different marginal cost requirement for some states. 
As a result, some states would be required to reduce emissions up to the
maximum marginal cost level of $2,400/ton and some states where the cost
per air quality benefit is high would be required to reduce emissions up
to a lower marginal cost level.   

There were several technical challenges with this approach, most related
to setting particular levels for the different parameters.  For example,
it was not clear how to specify the critical cost per air quality value
(i.e., is the median of all the linkages the appropriate cost per air
quality impact value?  Should all linkages count in calculating the
median if there are multiple receptors in a downwind area?).  As a
second example, it was not clear what “maximally-modeled” marginal
cost emission level to use (i.e., $2,400/ton was the example stated
above, but this could have been any number of other marginal cost
levels).  The final cost per air quality answer depends strongly on the
choice of the maximum marginal cost level.  Also, the cost per air
quality benefit varies considerably for SO2 and NOx.  Therefore, the
cost per air quality could be calculated independently for SO2 and NOx. 
Combined with separate calculations for annual and daily PM2.5, and
decisions that need to be made at several points in the calculations,
the analysis becomes quite complicated.  Ultimately, we were not able to
generate adequate justifications for the decision points that needed to
be made to support the methodology.

5. “Binning” of states based on air quality impact (i.e., subject
states with higher contribution to greater $ per ton reduction
requirement).

	

8

9

:

;

K

L

f

:

;

K



"

)

+

.

J

L

O

U

X

b

m

„

î

h

  h

h

kd,

Å欀遤

hÂ

h

- h

h

ôêãÙÒÙÆÙ¿Ù¿ÒÙÒÙÒÙÒÙÒÙãÙ¸±¸ª£¸ã£¸ªœ’
†ª’ªãªãª±ª±ªÙ

 h

h

õîõîõçîõîàîõîÙËÙ¿Ùõ¸±¸õ±õç±çõîõªÙËÙ¿
Ùõ¸ª±ª£œ•ÙËÙ¿ÙªŽõªõ¸

 h

 h

	g/m3) and resulting “bins”, or groups of states (Group A, B, C,
and D), based on three threshold levels (1, 3, and 5%), whether a state
is contributing less than 0.15, between 0.15 and 0.45, between 0.45 and
0.75, or greater than 0.75 g/m3, respectively.

Upwind State	Maximum Downwind Contribution to Nonattainment or
Maintenance for

Annual PM2.5 (g/m3)	 	Bin

Alabama 	0.46	 	Group C

Arkansas 	0.09	 	Group A

Connecticut 	0.09	 	Group A

Delaware 	0.20	 	Group B

Florida 	0.29	 	Group B

Georgia 	0.63	 	Group C

Illinois 	1.01	 	Group D

Indiana 	2.09	 	Group D

Iowa 	0.31	 	Group B

Kansas 	0.09	 	Group A

Kentucky 	1.68	 	Group D

Louisiana 	0.34	 	Group B

Maine 	0.02	 	Group A

Maryland 	0.62	 	Group C

Massachusetts 	0.13	 	Group A

Michigan 	0.72	 	Group C

Minnesota 	0.19	 	Group B

Mississippi 	0.07	 	Group A

Missouri 	1.38	 	Group D

Nebraska 	0.08	 	Group A

New Hampshire 	0.02	 	Group A

New Jersey 	0.68	 	Group C

New York 	0.49	 	Group C

North Carolina 	0.19	 	Group B

North Dakota 	0.05	 	Group A

Ohio 	2.03	 	Group D

Oklahoma 	0.08	 	Group A

Pennsylvania 	1.60	 	Group D

Rhode Island 	0.01	 	Group A

South Carolina 	0.26	 	Group B

South Dakota 	0.02	 	Group A

Tennessee 	0.68	 	Group C

Texas 	0.13	 	Group A

Vermont 	0	 	Group A

Virginia 	0.37	 	Group B

West Virginia 	1.17	 	Group D

Wisconsin 	0.46	 	Group C



As a particular variation of the approach evaluated by EPA, two
threshold levels (rather than three) were considered (e.g., 1% and 4% of
the respective NAAQS).  States were included in one of three groups
based on each state’s air quality contribution to their
maximally-impacted downwind monitor:  states below both thresholds,
which would not be included in the program; states above the 1%
threshold, but below the second threshold (4%), which would be subject
to some “basic” reductions; and a third “advanced” group for
states above both threshold levels (above 4%), which would be subject to
large emission reduction requirements.

For the magnitude of the reductions from different groups of states,
different levels of “cost-effective” controls would be applied.  As
an illustrative example, the “advanced” group might need to make all
reductions possible for less than $2,000 per ton, while the less
stringent “basic” group might need to make all reductions possible
for less than $500 per ton.  

EPA believes there are a number of questions that need to be asked with
the “binning” of states based on air quality approach.  First, it
would be necessary to pick the number of different threshold levels
(and, thus, the number of groups of states as well as their
composition).  For example, why two threshold levels, and not three or
more?  Second, it would be necessary to pick the cost threshold for each
different group.  

 The D.C. Circuit opinion granting petitions for review of the CAIR
indirectly discussed the use of fixed air quality thresholds.  It noted
that “[w]e observe initially that state SO2 budgets are unrelated to
the criterion (the “air quality factor”) by which EPA included
states in CAIR’s SO2 program. Significant contributors, for purposes
of inclusion only, are those states EPA projects will contribute at
least 0.2 μg/m3 of PM2.5 to a nonattainment area in another state.
While we would have expected EPA to require states to eliminate
contributions above this threshold, EPA claims to have used the measure
of significance we mentioned above: emissions that sources within a
state can eliminate by applying “highly cost-effective controls.””
North Carolina v. EPA, 531 F.3d 896, 916-17 (D.C. Cir. 2008).  

 The air quality benefit for each receptor for a particular cost level
was calculated using the Air Quality Assessment Tool.

 This particular set of bins was specifically recommended by a
stakeholder group in a letter to EPA (“LADCO Recommendations to EPA on
a CAIR Replacement Rule”).

  PAGE  1 

