IMPACT ASSESSMENT ON USE OF
ROUGH TERRAIN DISPERSION MODEL (RTDM)
November 1985
Source Receptor Analysis Branch
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
Environmental Protection Agency
-------
Table of Contents
Page
1.0 Introduction 1
2.0 Methodology 3
2.1 Identification of Major Sources 3
2.2 Identification of Major Sources Likely to Change Emissions. 5
2.3 Calculation of Emission Changes for Major Sources 6
2.4 Description of Sources not Covered 6
3.0 Results 8
3.1 Major Sources Located m Complex Terrain 8
3.2 Major Sources Likely to Change Emissions 8
3.3 Emission Changes for Major Sources 15
3.4 Emission Changes for Small Sources 19
4.0 Summary and Conclusions 23
REFERENCES 24
Appendix A A-l
l
-------
IMPACT ASSESSMENT ON USE OF
ROUGH TERRAIN DISPERSION MODEL (RTDM)
1. Introduction
The Guideline on Air Quality Models RevisedO) does not recommend any
refined models for sources located in complex terrain. Complex terrain is
defined here as terrain above physical stack height. Instead, the Guideline
recommends that specific refined models be developed, tested and applied on a
case-by-case basis. In the absence of such a case-specific refined model, the
Guideline allows the use of three screening techniques to establish emission
limits. These models are Valley(2), Complex l(3) and SHORTZ/LONGZW . As
screening techniques, each of these models yields a conservative estimate of
maximum concentration, occurring on high terrain.
Since the publication of the original Guideline on Air Quality Models(5),
few sources have gone through the effort and expense to develop, test and
apply site-specific complex terrain models. Instead they rely on the above
mentioned screening techniques to estimate their impact on ambient pollutant
concentrations. (Prior to 1978 many sources located in complex terrain apparently
had their emission limits based on rollback or on models not designed to account
for the effects of terrain.) Industrial interests have, since 1978, expressed
concern that the estimates obtained from these screening techniques are
overly conservative, i.e., they provide an unnecessarily large margin of
safety for maintaining the NAAQS (or PSD increments). The Rough Terrain
Dispersion Model (RTDM)(6) has been advanced by the Utility Air Regulatory
Group (UARG) as a more appropriate screening technique which yields more
realistic, but still conservative, estimates of concentrations in complex
terrain. Based on the technical evaluation of this model(?)»(8) EPA agrees
-------
that RTDM is acceptable and plans to propose it as a third-level screening
technique for rural point sources.
Since RTDM provides lower concentration estimates than the other screening
techniques, the question of whether approval of RTDM would result in signi-
ficant increases in SO2 emissions must be considered. The purpose of this
report is to address the potential for such increases. Section 2 of this
report describes the data bases and analysis methodology used to determine
the answer to this question. Section 3 presents the results of the analysis
and Section 4 the conclusions drawn.
2
-------
2. Methodology
A four step process was used to address the potential for increased SO2
emissions as a consequence of approval of RTDM. The steps were as follows:
1. Identify large (5000 T/Y) sources that could use RTDM to establish
emi ssion 1imits;
2. Identify those large sources likely to change emissions;
3. Calculate emission changes for the large sources;
4. Qualitatively describe smaller sources not covered in Steps 1-3.
Detailed procedures for carrying out these four steps are described in
tne following subsections.
2.1 Identification of Major Sources
Sources larger than 5000 T/Y, located in complex terrain and with
the potential to use RTDM, were identified. Included were sources with emission
limits previously set by Valley or Complex I, sources with emissions previously
set by other techniques, and sources with tall stacks subject to EPa's new
stack height regulation.
The 5000 T/Y cutoff was chosen for three reasons: (1) larger
sources have the greatest potential to adversely impact the environment, (2)
a similar cutoff has been used in the various impact analyses to support the
GEP stack height regulations, and (3) resource limitations would preclude
developing an inventory for all of the SO2 sources in complex terrain in
the U.S. The smaller SO^ sources were qualitatively identified and the
possible consequences of using RTDM on these sources is discussed according
to procedures identified in Section 2.4.
Exiting SO2 data bases were generally of limited use for this
analysis since they did not identify the sources which have, or could
3
-------
potentially have, emission limits based on a complex terrain model. Never-
theless, in connection with the development of the GEP stack height regulations
at least two specialized emission inventories were developed. rCf(9) developed
an emissions inventory of power plants potentially subject to the stack height
regulation. Although this inventory identified sources whose emission limits
could potentially be set by a complex terrain model, both at existing and at
GEP stack height, the method used to do this was extremely crude; many plants
may have been incorrectly identified as being or not being located in complex
terrain. Also, this data base was limited to power plants and it is known
that a significant amount of industrial source emissions occur in complex
terrain settings. Another study by GCAHO) identified, in a more useful
manner, sources located in complex terrain. However, comparison of the
sources with basic emission inventory data revealed that a number of large
plants were missing and/or the emissions estimates were questionable.
Because of the difficulties and uncertainties in working with
existing data bases, it was decided to develop a specialized data base for
this study based on the firsthand knowledge of EPA staff. A memorandum was
sent to the ten EPA Regional Offices; each was asked to identify sources with
emissions greater than 5000 T/Y that could potentially have emission limits
based on a complex terrain model. The Regions were also asked to separate
out the sources according to (1) those (including PSD sources) that currently
have emission limits based on Valley or Complex I, (2) those (in complex
terrain) that have their emission limits based on techniques other than
Valley or Complex I, or currently have no emission limits, and (3) those
that would have their emission limits based on a complex terrain model in
order to comply with the GEP stack height regulations.
For existing and PSD sources with SO2 emissions sources greater
4
-------
than 5000 T/Y, the Regional Offices provided the source name, location
(State), type of source, and SO2 emissions in tons per year (i.e., the
emission limit (if one exists), the expected emission limit, or the actual
emission rate).
Sources were identified according to the following categories:
A. existing sources with emission limits based on Valley/Complex 1;
B. existing sources, located in complex terrain, with emission
limits based on techniques other than Valley or Complex I (the other model/
technique was identified);
C. any other existing sources, located in complex terrain, which
did not have a current SIP limit or were out of compliance with the current
SO2 1imi t;
D. States that may want to relax the statewide SO2 limit if
RTDM were available as an approved screening technique;
E. Existing sources located in complex terrain which, because they
are subject to the stack height regulation, could have their emission limits
set based on a complex terrain model or fluid model;
F. PSD sources with emission limits and incremental consumption
based on Valley/Complex I; and
G. sources which could not or did not choose to because the
emission limitations based on Valley or Complex I would be too stringent.
2.2 Identification of Major Sources Likely to Change Emissions
Out of the group of sources located in complex terrain, the ones
that were likely to model with RTDM in order to gain an emission relaxation
were further identified. The Regional Offices were asked to identify, to
the extent possible, those sources that would want to use RTDM to relax
5
-------
emission limits and the extent to which the States' policy/regulations or
other factors would allow such a relaxation. The Regional Offices were
also asked if any of their States might want to relax statewide emission
limits, given that RTDM could justify a relaxation (Category B, above).
Finally the Regional Offices were asked to identify potential PSD sources
which, because of the stringent emission limitation imposed by Valley or
Complex I, chose not to build at the present time, but might now wish to do
so because of the availability of RTDM (Category G, above).
2.3 Calculation of Emission Changes for Major Sources
Next the potential increase in SO2 emissions from the sources
likely to use RTDM was estimated. Emissions were considered for the eastern and
western halves of the nation and also were aggregated to the national level.
Modeling of each source would be prohibitive from a resource
standpoint. Instead, the results of comparative model output data from RTDM
and Complex I for several "example" sources were used to infer the emission
relaxations for all the sources. A multiplicative factor or range of multi-
plicative factors were used to estimate the potential emissions increase from
the identified sources. These multiplicative factors were derived from an
EPA m-house study, the results of which are described in Section 3.3 below,
and from the results of comparative model evaluation studies between RTDM and
Complex l(7)»(8).
2.4 Description of Sources not Covered
The Regional Office responses serve to identify only large sources
(> 50U0 T/Y) with the potential to use RTDM. A number of smaller sources,
especially PSD sources, may have had their emission limits based on Valley
6
-------
or Complex I. Although the number of such sources is unknown, their total
amount of emissions is probably not large compared to those of the 500U T/Y
sources.
A recent report by Radian(H) quantified, by EPA Regions, the
aggregate SO2 emissions from PSD sources that have been permitted between
1977 and 1983. While these aggregates include sources located in flat terrain
as well as complex terrain, they did serve to identify an upper bound on
emissions that could be relaxed on the basis of the availability of RTDM.
No data base exists to identify other small sources (that are not PSD sources)
whose emission limits were based on Complex I or Valley. The potential impacts
of these smaller sources is qualitatively discussed in Section 3.
7
-------
3,0 Results
This section describes the results and analyses for each of the four
steps described in Section Z.
3.1 Major Sources Located in Complex Terrain
Table 1 summarizes data from the Regional Offices concerning major
sources located in complex terrain, that emit more than 5000 T/Y of SO2.
Data for Categories A and F are based on allowable emissions. Category C
data are is based on actual emissions. Categories B and E contain a mixture,
but for the most part represent actual emissions; the Region VI and VIII
smelters arid some sources in Region II are allowable emissions. The emissions
tallied in the Table are for sources in the contiguous U. S.
Subject to these caveats, Table 1 shows that approximately 4.6
million T/Y of SO2 are emitted from large (> 5000 T/Y) sources located in
or near complex terrain. This may be compared to a 5983 estimate of 22,9
million T/Y from all SO2 sources in the country (including small sources,
area sources and sources outside of the 48 contiguous states)(11). If it is
assumed that Regions I-V comprise the eastern U.S. and Regions VI-X the
West, Table 1 shows that 3.4 million T/Y of emissions come from sources
located in complex terrain in the East and 1.2 million T/Y in the West.
3.2 Major Sources Likely to Change Emissions
Category A of Table 1 shows that there are only 14,000 T/Y of
permitted emissions from existing sources, other than PSD sources, with
emission limits based on Valley or Complex I. These sources are comprised
of one power plant in Connecticut and two power plants in New York State.
The Connecticut source could potentially increase its emissions (currently
based on 0.5$S oil) up to the Statewide limit associated with US oil, if
8
-------
Table 1. SO? Emissions (KT/Y) from Point Sources'1) larger than 6001) T/Y with Stack Heights Less Than Stirroundi
Terrain Heignt.
REGION
Source Category
I
II
III
IV
V
VI
VII
VIII
IX
X
TOTAL
A. Existing sources with
emission limits based on
Valley or Complex l(2)(5)
5
9
0
0
0
0
0
0
0
0
14
B. Existing sources with
emission limits based on
techniques other than Valley
or Complex I
(B.l. Power plants)
(B.2. Primary smelters)
(B.3. Industrial sources
other than primary smelters)
0
216
(175)
(41)
1727
(1555)
(172)
1179
(1101)
(78)
15b
(55)
(101)
261
(152)
(103)
(6)
50
(50)
80
(80)
608
(608)
106
(74)
(32)
4383
(3112)
(841)
(430)
C. Existing sources with no
SIP emission limit(3)
0
0
0
11
99
0
0
0
0
0
110
D. States which may want to
relax their emission limit
based on RTDM
0
0
0
0
0
0
0
0
0
0
0
E. Existing sources which
could have GEP emission limits
based on a complex terrain
model(4)(6)
0
0
677
794
28
103
0
80
44
74
1800
F. PSD sources with emission
limits based on Valley or
Complex I(2)
0
19
0
0
0
0
0
54
14
20
107
G. Future PSD sources
0
0
0
0
0
0
0
0
0
0
0
Tota
less Column £
4614
Footnotes on following page
-------
Table 1. Footnotes
(1) Does not include emissions from sources located outside the continental United States
(2) Allowable emissions
(3) Actual emissions, in most cases
(4) Emissions from Regions III, IV, VI, VIII, and X are also, to a large extent, included
in Category B. Emissions from Region V are included in Category C
(5) Other than PSD souces, which are included in Category F
(6) For Region II a data base is not yet developed. However, few sources in Region II have
stack heights above GEP.
-------
RTDM would allow. Region II, in their response, indicated that the New York
State sulfur policy would cap the emissions from the two power plants at or
below the levels listed in Table I. Thus for Category A there is only one
source that could potentially relax its emission limit based on the availability
of RTDM.
Category B of Table 1 provides estimates of current emissions from
large sources located in complex terrain with emission limits based on tech-
niques other than Valley or Complex I. Approximately 4.4 million T/Y, or
approximately 95% of the total emissions in Table 1, are in Category B. To
determine whether RTDM would allow a relaxation of emission limits for these
sources, some additional discussion on the nature of these sources and the
bases for their emission limits is useful.
The numbers in parentheses divide the emissions in Category B into
those coming from power plants, primary smelters and other industrial facilities.
As can be seen, the bulk of the emissions are associated with power plants in
the East and smelters in the Uest. The smelters have emission limits based
on multipoint rollback (MPR); application of any of the currently available
screening models, including RTDM, to those sources would very likely result
in a more stringent emission limitation than MPR allows.
The bulk (in terms of emissions) of the Category B power plants
and industrial sources in Region III have emission limits based on straight
rollback. While application of straight rollback results in perhaps more
stringent emission limits than MPR, it is doubtful that these limits would
be more stringent than those imposed by any screening technique, including
RTDM. Rollback relies on the assumption that an existing monitor is located
at the point of maximum concentration; this is seldom the case especially
11
-------
where the monitor must exist at a high, remote terrain point. From
a temporal standpoint, rollback also assumes that the data base used in the
rollback calculation was collected under worst case meteorological conditions,
which is not the case as a rule. Thus it is not believed that RTDM would
allow emissions relaxations from sources whose emission limits are based on
straight rollback.
The bulk of the emission limits for the Region IV sources in
Category B are based on flat terrain models such as CRSTER or MPTER.
Apparently these emission limits were either established before the applica-
tion of complex terrain models was required, or it was assumed that the
Hat terrain impacts were greater £ba*) those on high terrain, in either
ewent the application of RTDM, or -anj< other complex ter-a~n nodel , ta these
scurces woulc either result in reafflmetion that the limit imposEc by the
flat terrain model is appropriate or4 if complex terrain impacts were higner,
establish a more stringent limit.
For the remaining sources in Regions III and IV and for sources in
other Regions listed under Category B, the basis for emission limits varied.
Included were (l) site-specific nonguideline complex terrain models, (2)
rollback, and (3) non-modeling techniques such as sulfur in fuel, BACT/NSPS,
etc. Two sources in Region III have emission limits based on the SHORTZ
screening technique. For sources whose limits are based on site-specific
nonguideline models, it is assumed that the models have been adequately
justified; thus calculated impacts are considered to be lower than those
obtained from conservative screening techniques, including RTDM. Of sources
whose emission limits are not based on modeling, it is assumed that the
availability of RTDM would not change this basis.
12
-------
Of the two sources in Region III whose emission limits are based
on application of SHORTZ, Region III indicated that for one of the sources
(emissions of 45,000 T/Y), application of RTDM would probably cause a tighten-
ing of the emission limit. The reason is that SHORTZ estimated that the design
concentration, used to set the emission limit, occurs in flat terrain and not
on the nearby terrain feature, RTDM could estimate a higher concentration
on the terrain feature. The other source, a power plant emitting 20,000 T/Y,
presumably could relax its limits to the extent that RTDM would estimate lower
concentrations than SHORTZ.
In summary, for the 4.3 million T/Y in Category b, it is believed
that the availability of RTDM would not result in a relaxation of emission
limits, with the possible exception of one relatively small source ir Region III.
In fact, if the Category B sources were modeled according to current modeling
policy, or with RTDM, the need for a significant reduction in allowable
emissions would undoubtedly be evident.
Category C contains 110,000 T/Y of actual emissions from sources
for which emission limits have not yet been set. The majority (99,000 T/Y)
of these sources, which includes a mixture of paper mills and power plants,
are in the State of Wisconsin. Preliminary modeling indicates that only
18,200 T/Y of actual emissions in Wisconsin are likely to be affected by
complex terrain modeling. Given that no emission limits currently exist and
that the amount of emissions effected is a small percentage of the total,
Category C sources will not be included with those who could potentially
increase emissions based solely on the availability of RTDM,
13
-------
For Category D, none of the Regions believe that their States
would want to relax Statewide emission limits (sulfur in fuel regulations)
based on the availability of RTDM. The one exception is Puerto Rico, which
is not included in Table 1.
Skipping to Category F for the moment, there are 107,000 T/Y of
allowable emissions from 5000 T/Y sources permitted under the PSD program.
Of this total, 73,000 T/Y of emissions are capped by BACT/NSPS or by other
non-air quality related criteria. This leaves 34,000 T/Y from sources
with emission limits based or Valley or Complex I that could potentially
be relaxed based on RTDM.
Category £ includes sources wfncft could have emission J mm ts
imposed because of GEP regulations. It should be noted that the new emission
limits would only be tighter (or perhaps the same as current limits) regard-
less of which complex terrain model is used. The only difference between
the application of RTDM or Valley/Complex I is that Valley/Complex I would
tend to tighten the emission limits further than RTDM. Also, in almost all
of the cases for Category E, the current emission limits are based on
rollback, MPR or a flat terrain model. Thus, following the reasoning above
concerning Category B sources, the GEP emission limits set by a complex
terrain model for all of these sources will be tighter than current limits.
Thus, it is assumed that none of the Category E emissions could increase
because of the availatnlity of J?TDM,
Another point to note about Category E is that the bulk of the emissions
are also included in Category B; thus the total rational emissions in complex
terrain at the bottom of Table 1 does not include Category £. The only sources
included in Category E that are not included in Category B are (1) several power
plants in Region IV, which become "complex terrain sources" because their GEP
H
-------
stack height is lower than terrain and,(2) one smelter in Region IX which is
currently not operating. A final point regarding Category E is that the
emission estimates in Table 1 are subject to a high degree of uncertainty.
States are just now beginning tc inventory sources subject to GEP regulations;
estimates in the Table reflect very rough first approximations of the sources
i nvolved.
The somewhat unanticipated result for Category G sources was that
none of the Regions believe that there are new sources wishing to build but
are unable do so because of the stringent emission limitations that would be
set by the current screening techniques. Historically there have been several
cases, especially in the West, where proposed major sources negotiated with EPA
on siting. A major issue in these negotiations was the emission limitations
that would be imposed by use of Valley or Complex I. Apparently these sources
are no longer interested in siting, perhaps due to other factors, e.g. economic,
reduced demand for power, etc. In any event, based on the Regions' response,
the availability of RTDM would not by itself cause an increase in SO2 emissions
due to reconsideration/resubmittals of PSD permit applications.
Giver the above rationale for Categories A through F, Table 2
sunmarizes the emissions that could potentially be increased because of the
availability of RTDM. Currently only 59,000 T/Y of actual or permitted
emissions from large sources could potentially be increased. Of this total
only one source, a PSD source in Region IX that is not yet constructed, has
its emission limit based on Complex 1.
3.3 Emission Changes for Major Sources
Appendix A summarizes the results of an internal EPA study where
concentration estimates from RTDM and Complex 1 were compared. The study
15
-------
Table 2. Emissions Potentially Subject to Relaxations
Source Category
Emi ssions(KT/Y)
Basis of Current Limits
A
5(1)
Va 11 ey
B
20(2)
SHORTZ
C
*(3)
None
D
0
Sulfur in Fuel
E
0
Various
F
20(1)
Va 11 ey
14C)
Complex I
G
0
—
Total
59
(1) Allowable emissions
(2) Actual emissions
(3) Si rice no emission limits exist, the availability of RTDH would
not, by itself, provide the opportunity for relaxation.
16
-------
considered three hypothetical plants, i.e. three stack heights, located in a
moderately complex terrain setting. The results suggest that there may be
significant differences in the concentrations estimated by the two models
and that these differences may vary widely from plant to plant and with the
specific terrain height. When the plume is below the dividing streamline,
RTDM seems to estimate concentrations that range from a factor of 10 lower
than Complex I to a factor of two higher. A factor of five lower might be
a rough typical ratio. Under stable conditions when the plume is above the
dividing streamline, but the below the height of terrain, RTDM estimates
concentrations up to factor of 20 lower than Complex I. When the plume is
near or above the height of terrain, RTDM estimates seem to be about a factor
of three lower than Complex I.
In addition to the study described in Appendix A, concentration estimates
with Complex I and with RTDM from a total of 25 receptor points (corresponding
to monitoring sites) for two actual sources are available (?)(8). From these data
the high, second-high estimates from the two models were compared for both the
3-hour and 24-hour averaging times, for both paired in space and unpaired data
and for each source. The comparison showed a fairly consistent set of results,
regardless of the averaging time or degree of pairing. RTDM estimates were lower
than Complex I estimates by a factor of 2.5 to 3.7. It should be noted that
these data bases contained much fewer receptor sites than did the study in
Appendix A (11 for one source and 14 for the other compared to 180 for the
Appendix A study). Thus it is not surprising that a narrower range of
results was found for the studies involving two actual sources.
For purposes of calculating potential emission increases, the factor
of five is being used as a conservative estimate. As can be seen from the
discussion that follows, it does not make a big difference in emissions
17
-------
increases from an overall standpoint whether one uses this factor of five or
some other more or less conservative factor. The magnitude of the potential
emissions increases still is small compared to nationwide SOg emissions. It
should also be noted that a ratio larger than a factor or five is probably
not realistic since a number of other considerations would likely limit the
emissions increases, e.g. flat terrain modeling, control equipment currently
in place, State policies, etc.
Given the factor of five, the only major plant identified in Section 3.2
where the factor can be directly applied is the PSD source in Region IX. The
factor of five sugyests that an emission increase of 56,000 T/Y from this
source could occur. For the sources in Region VIII based on the Valley model,
there are no sensitivity studies to suggest what the emission relaxation
factor would be if RTDM were used. In most cases Valley estimates would likely
be somewhat higher than Complex I since Valley assumes worst case meteorological
conditions rather than using actual conditions. Thus, the Valley/RTDM ratio
would be higher than the Complex 1/RTDM ratio, and perhaps a factor of 6 would
be more appropriate for the Region VIII sources. Thus the emission relaxation
might be as much as 100,000 T/Y from these sources.
Some sketchy information from the performance evaluation for the Westvaco
pi ant(8) suggests that SHORTZ estimates might be about 1.5 times higher
than RTDM estimates. Thus the Region III source could potentially increase
its emissions by 10,000 T/Y.
As mentioned earlier, the power plant in Region I could potentially
increase its emissions by a factor of 2, up to the limit associated with the
Statewide \% S oil limit. It is apparent that RTDM would probably allow
such a relaxation. Thus an emission increase of 5000 T/Y could be allowed.
18
-------
Table 3 summarizes the potential emissions increases from major
sources that could result if RTDM were made available to revise emission
limits. Given the assumptions described previously, the estimates are that
there could be a 171,000 T/Y increase in SO2 emissions from these sources.
Because of the uncertainties associated with the assumptions and with the data
base, these estimates could be in error by a factor of 3 or more.
3.4 Emission Changes for Small Sources
As mentioned in Section 2, information provided by the Regional
Office was limited to identifying sources larger than 50U0 T/Y. Several
Regional Offices indicated in their responses that there are a number of
smaller PSD sources whose emission limits are based on Valley or Complex I.
No mention was made of other small sources with emission limits based on
Valley/Complex I. It is not believed that there are very many such sources
since most would have had their limits established in the initial SIP's in
the early 1970's, before complex terrain models were generally required.
In fact, based on the pattern of Table 1, it is expected that most such
sources would fall into Category B with emission limits based on rollback,
flat terrain models and non-modeling criteria. Thus it is not expected
that the availability of RTDM would allow any significant relaxation from
the smaller, non-PSD sources.
Based on data contained in the report by Radian(12)} Table 4 lists,
by Region, emissions from PSD sources permitted between 1977 and 1983.
This Table shows that there are approximately 940,000 T/Y of SO2 emissions
from such sources. Of this total, 483,000 T/Y are in Regions V, VI and VII.
It probably can be safely assumed that almost all of the sources in these
three Regions are located in flat terrain, i.e. their emission limits were
imposed by a flat terrain model (or by BACT/NSPS). Of the remaining
19
-------
Table 3. Potential Emission Increases Allowed by RTDM
Source Category
Emissions{KT/Y)
Potential Emission
Increase (KT/Y)
A
5(1)
5
B
20(2)
10
C
*(3)
*(3)
D
0
0
E
0
0
F
20(1)
100
14(1)
56
G
0
Total
59
171
(1) Allowables emissions
(2) Actuals emissions
(3) Since no emission limits exist, the availability of RTDM
would not, by itself, provide the opportunity for relaxation.
20
-------
457,000 T/Y, many of the sources associated with these emissions probably
have emissions capped or at least the relaxations would be limited by BACT,
NSPS, current control equipment, State policies, or by a flat terrain model.
Given these factors, together with the pattern of emission limitations for
the larger sources provided in Table 1, it is doubtful that there would be
any significant emission increases from these PSD sources.
21
-------
Table 4. Total Emissions Associated with PSD Permit Applications
(Tons/Year)*
EPA Region
SO2 Gross
Emissions
I
14,855
II
46,658
III
22,888
IV
301,721
V
89,619
VI
402,610
VII
18,344
VIII
2,572
IX
5,434
X
35,819
Total 940,519
*These emissions include those that were subject to PSD regulations,
as well as those that were not subject to PSD regulations.
22
-------
4.0 Summary and Conclusions
The analysis shows that the availability of RTDM could result in a
national emissions ?rrcrease of SO2 of approximate?/ 171tOUU T/Y. Of this
total 156,000 T/Y are associated with sources in the western U.S. and
15,000 T/Y from sources in the East. These increases are compared to a
current estimate of 4.6 million T/Y gross emissions from sources located in
complex terrain and a 1983 estimate of 22.9 million T/Y from all SO2 sources
in the country.
The 156,000 T/Y potential increase m the western U.S. are associated
with three power plants. One plant has not been built yet; thus the
potential change in emissions would involve a reapplication for a new PSD
permit. Some question could be raised on the possible visibility impacts
associated with the increased SO2 emissions from the ether two existing
plants. However, these plants are quite distant from any Class 1 areas,
120 and 165 miles, and it is doubtful that any perceptable visibility
impacts would occur at such distances.
Estimates of emissions increase have a sizeable uncertainty associated
with them, perhaps as much as a factor of three or more. This uncertainty
is associated with uncertainties in the emissions inventory, the highly
variable differences between RTDM estimates and those obtained from current
screening techniques, and the assumption that emissions increases from small
sources could be neglected. Nevertheless it appears that the emissions
increases would be small compared to nationwide estimates of total SO2
emissions.
These results are somewhat surprising since there is a wide perception
that complex terrain models have had a major restrictive impact on emission
limitations. However, the results are reasonable because most of the major
23
-------
existing sources in complex terrain do not have their current emission
limitation based on a complex terrain model. Instead these emission limita-
tions are most often based on rollback, flat terrain models or non modeling
criteria and as NSPS, BACT or State policies. Many of the emission limits
based on rollback or flat terrain were probably set before there was a
requirement to apply complex terrain models.
24
-------
REFERENCES
1. Environmental Protection Agency. Guideline on Air Quality Models (Revised),
Proposed Revisions, U. S. Environmental Protection Agency, Research Triangle
Park, NC 2771 1 , November 1984.
2. Burt, E. W. Valley Model User's Guide, EPA 450/2-77-018, U. S. Environmental
Protection Agency, Research Triangle Park, NC 27711, September 1977.
3. Environmental Protection Agency. User's Network for Applied Modeling
of Air Pollution (UNAMAP), Version 5, (Computer Programs on Tape),
National Technical Information Service, Springfiela, VA 22161, August
1983.
4. Bjorklund, J. R. and J. F. Bowers. Users' Instructions for the SHORTZ
and LONGZ Computer Programs, Volumes I and II, EPA 903/9-82-004A and B,
U. S. Environmental Protection Agency, Region III, Philadelphia, PA 19107,
March 1982.
5. Environmental Protection Agency. Guideline on Air Quality Models, EPA 450/
2-78-027, U. S. Environmental Protection Agency, Research Triangle Park,
NC 2771 1 , April 1978.
6. Environmental Research and Technology. User's Guide to the Rough Terrain
Dispersion Model, (RTDM) (Rev. 3.10), ERT Document No. M2209-585, ERT, Inc.,
696 Virginia Road, Concord, MA 01742, 1984.
7. Environmental Research and Technology. Performance Evaluation of Complex I
and RTDM using 1979-1980 Data from the TVA Widows Creek Monitoring Network,
ERT Document PD 23-400, ERT, Inc., 696 Virginia Road, Concord, MA 01742,
June 1985.
8. Londergan, R. J. Performance Evaluation of the Rough Terrain Diffusion
Model at Cinder Cone Butte and Westvaco Luke Mill, Report for TRC Project
No. 3017-R61, TRC Environmental Consultants, Inc., 800 Connecticut Blvd,
East Hartford, CT 06108, June 1985.
9. ICF, Inc. Final analysis of the Proposed Stack Height Regulations, Final
Report to U. S. Environmental Protection Agency, ICF, Inc., 1850 K Street
NW, Washington, DC 20006, June 1985.
10. GCA Corporation. Evaluation of Sources Affected by Revisions to the Stack
Height Regulations, Final Report to the U. S. Environmental Protection
Agency, Contract No. 68-02-3510, Work Assignment No. 51, GCA-TR-84-129-6,
GCA Corporation, 213 Burlington Road, Bedford, MA 01730, June 1985.
11. Environmental Protection Agency. National Air Quality and Emissions Trend
Report, 1983, EPA 450/4-84-029, U. S. Environmental Protection Agency,
Research Triangle Park, NC 27711, April 1985.
12. Radian Corporation. Analysis of New Source Review (NSR) Permitting
Experience - Part 2, EPA 450/2-85-005, Radian Corporation, 3200 East Chapel
Hill Road, Research Triangle Park, NC 27709, September 1985.
25
-------
Appendix A
A Comparison Between RTDM and COMPLEX I
A.l Introduction
A comparison has been made between RTDM and Complex I model estimates
using input data for sources having stack heights of 35m, 100m, and 200m,
respectively. Stack characteristics were selected to be reasonably typical
of small, medium, and large sources, respectively. Five rings of 36 receptors
each at 0.8, 2.0, 4.0, 7.0, and 15.0 km were used. Synthesized, but realistic,
terrain was input as receptor elevations which, for some receptors, extended
well above stack height. The meteorological data set for Pittsburgh for 1964
was used. One-hour and 24-hour averaging times were considered. Model
output of the highest concentrations at each receptor were used in the
comparison.
The options set in RTDM for this study are those proposed in the draft
Guideline on Air Quality Models:
° wind speed scaled to stack height for plume rise calculations, and to
plume height for dilution calculations;
° default RTDM wind speed profile exponents;
° Briggs' rural dispersion coefficients;
° buoyancy-induced dispersion, with coefficient of 3.162;
0 unlimited mixing for stable conditions;
° transitional plume rise;
° plume path coefficients of 0.5 for all stabilities;
° stack-tip downwash;
° partial reflection from the ground; and
0 22.5° sectors for all stabilities.
The following optional features are disabled:
-------
° partial plume penetration of mixing lid; and
° wind direction shear.
A.2 Discussion
The terrain used in this analysis extends to about 250 meters above the
stack base. This places the highest terrain 50 meters above the top of the
high stack, 150 meters above the top of the medium stack and 215 meters above
the top of the low stack.
Under stable conditions, RTDM calculates a critical streamline height
above which the plume remains above the terrain by an amount determined by
by using a half-height plume path correction. Below the critical stream-
line height, the plume is assumed to flow horizontally until it impacts the
hillside. Complex I always assumes a horizontal plume under stable conditions
and uses a half-height plume path correction during unstable and neutral con-
ditions. Note that the half-height plume path correction in calculated
differently in Complex I than in RTDM, with potentially different results
for otherwise identical cases.
Another difference is the treatment of plume impaction on terrain.
Complex I deflects the plume up the hillside after impact occurs, decreasing
the plume centerline concentration linearly as a function of height until
the concentration reaches zero at 400 meters above the original plume center-
line. RTDM allows the impacting plume to intersect the ground without
deflection or change in shape. This difference does not affect the maximum
concentration, but does affect the spacing of receptors needed to obtain
that peak value. Specifically, the size of the stable plume is relatively
small - a few tens of meters in size even up to several kilometers downwind.
If a receptor is not placed at an elevation near the centerline of the
2
-------
impacting plume, the peak concentration will be missed. With Complex I,
the receptor location is not so critical, since the plume is deflected up
the slope toward the next higher receptor.
Two other differences in the concentration calculations made by the
two models for the impact cases should be noted. The first is that RTDM
uses Briggs' formulation for the vertical dispersion curves, while Complex I
uses the Turner formulation. Both are based on the Pasquil1-Gifford curves,
and should not give greatly different results for E and F stability. The
second difference is that Complex I assumes the plume is completely "reflected"
from the ground. RTDM, on the other hand, uses a partial reflection that
depends on the slope of the terrain.
As a result of the stack heights, plume rise, and critical streamline
heights for stable cases, the following statements can be made regarding
each stack used in this study.
Low (35 meters) stack: The plume is frequently under the dividing
streamline height for stable conditions, resulting in frequent plume
impaction cases with RTDM as well as with Complex I.
Medium (100 meters) stack: The stable plume is rarely under the
dividing streamline height with RTDM, but nearly always below the height of
the highest terrain. (In this case the dividing streamline height is always
within 80 meters of the highest terrain elevation.) Thus the plume, as
modeled in RTDM, will rise over the terrain features following a half-height
plume path correction. The plume, as modeled in Complex I, will frequently
impact on the terrain.
High (200 meter) stack: The plume remains above the terrain in both
models. Differences in concentration estimates are due primarily to the
3
-------
effects of RTDM following the half-height correction, and Complex I using the
horizontal plume for stable cases and the half-height correction (different
from that used in RTDM) for neutral and unstable cases.
The results of the model runs comparing RTDM and Complex I are summarized
in Table 1. No comparisons were made for the 3-hour averaging time; however,
the results for the 1-hour averaging time are believed to adequately represent
the 3-hour results. The entries in Table 1 represent factors by which RTDM
predicts lower concentrations than Complex I. Thus, 3.7 means that RTDM
predicts a factor of 3.6 lower than Complex I for that case, while 0.6 means
that RTDM predicts higher, by a factor of 1.0/0.6, or 1.7.
For the plant whose plume is frequently below the dividing streamline,
represented by the 35 meter stack, this study indicates RTDM could predict
concentrations ranging from more than a factor of 10 lower to nearly a
factor of 2 higher than those from Complex I. It is probable that the
large range of factors found in this study will be reflected in a similarly
large range of factors at a variety of real sources in a variety of terrain
settings. However, based on this study alone, a factor of 5 lower should
represent a reasonable rough estimate of the typical effects of such a
modeling change on calculated emissions if the modeling change is applied
to a substantial number of sources in a variety of high terrain situations.
For plants located where there is just enough terrain to result in a
calculated impaction with Complex I, but not with RTDM, represented in this
study by the 100 meter stack, this study indicates that RTDM could range
from predicting a factor of 20 lower than Complex I to a factor of 1.3
higher. The highest concentrations are predicted to be factors of 11 to 17
lower by RTDM, with factors of roughly 11 and 13 being most conspicuous.
4
-------
For this case concentration estimates using RTDM instead of Complex I may
reasonably be expected to be roughly on the order of a factor of 12 lower,
with wide variations at individual sources.
Plants with terrain above stack height, but below plume height, are
represented by the 200 meter stack case. Here, the concentrations are
less affected by terrain, with the result that the factors in Table 1 are
less variable. Table 1 shows, for this case, RTDM predicting from a factor
of 4.3 lower than Complex I to a factor of 1.6 higher. Unlike the other
two cases, the highest concentration is caused by unstable conditions, rather
than stable. Also, unlike the other two cases, the factors for the 1-hour
averages differ substantially from those for the 24-hour averages. Thus,
while RTDM appears to predict within a factor of 1.5 either way from Complex I
for 1-hour averages, it predicts a factor of 3 lower for the 24-hour averages
for this case.
A.3 Conclusions
When terrain is sufficiently high to permit the plume to remain below
the dividing streamline height, this study indicates that RTDM tends to
predict concentrations from a factor of 10 lower than Complex I to
a factor of 2 higher. A rough estimate of a typical ratio is five.
When terrain is just high enough to exceed plume height under stable
conditions, but not high enough to place the plume below the dividing stream-
line height, RTDM could easily predict concentrations as low as a factor of
20 lower than Complex I, to a factor of two higher. A rough estimate of a
typical ratio is a factor of 12.
When terrain is higher than stack height, but below plume height, RTDM
tends to predict about the same high and high-second high values as Complex I
5
-------
for 1-hour averages. For 24-hour averages, RTDM tends to predict on the
order of a factor of three lower.
Three limitations to this study should be noted:
° Because of the manner in which plume impaction cases are handled
by the models, RTDM would require a much closer spacing of
receptors than Complex I in order to be assured of calculating
the highest concentrations.
° Large variations in factors can be expected from source to source,
which cannot be accurately addressed by a study involving only three
source configurations.
0 In cases where RTDM predicts concentrations substantially lower
than Complex I, RTDM may also predict lower than a model applicable
to flat terrain. Since use of a flat terrain model is always required,
the net factor for the plant may actually be much less than the
factors developed in this report.
6
-------
Table 1. Summary of Comparison Between RTDM and Complex I
(Numbers indicate factors by which RTDM predicts lower than Complex I)
Stack Averaging Highest
Height Times Concentration
(meters) (hours) (Factors)
Highest
Concentration
by Direction
(Range of 36
Factors)
35
100
200
1
24
1
24
1
24
3.6
5.3
13.
11.
1.4
2.7
7.1
11.
20.
17.
1.5
4.3
0.6*
0.6
0.8
0.8
0.6
0.8
~Factor less than 1.0 indicate that RTDM predicts higher than Complex I,
7
------- |