Requirements for Preparation, Adoption, and Submittal of Implementation Plans

ENVIRONMENTAL PROTECTION AGENCY

40 CFR Parts 51 and 52

[AH-FRL-5107-1; Docket No. A-92-65]

Requirements for Preparation, Adoption, and Submittal of

Implementation Plans

AGENCY: Environmental Protection Agency (EPA).

ACTION: Notice of proposed rulemaking.

-----------------------------------------------------------------------

SUMMARY: EPA is issuing this proposal to augment the final rule that

was published on July 20, 1993. Today's notice proposes to make several

additions and changes as supplement C to the ``Guideline on Air Quality

Models (Revised)''. Supplement C does the following: incorporates

improved algorithms for treatment of area sources and dry deposition in

the Industrial Source Complex (ISC2) model, adopts a solar radiation/

delta-T (SRDT) method for estimating atmospheric stability categories,

adopts a new screening approach for assessing annual NO2 impacts,

and adds SLAB and HGSYSTEM as alternative models. The Guideline sets

forth air quality models and guidance for estimating the air quality

impacts of sources and for specifying emission limits for them. The

purpose of the proposed changes is to enhance the guidance in response

to a substantial number of public comments urging the Agency to do so.

For the purposes of this document, EPA is soliciting public comments

only on the four proposed changes associated with supplement C and will

not respond to any comments that are outside the scope of this

document. This limiting of EPA's responses to comments within the scope

of this document allows the Agency to focus on the issues, data, and

information relevant to this rulemaking.

DATES: The period for comment on these proposed changes closes January

12, 1995.

ADDRESSES: Comments: Written comments should be submitted (in duplicate

if possible) to: Air Docket (6102), Room M-1500, Waterside Mall,

Attention: Docket A-92-65, U.S. Environmental Protection Agency, 401 M

Street, S.W., Washington, D.C. 20460.

Copies of supplement C (draft) to the ``Guideline on Air Quality

Models (Revised)'' may be obtained by writing or calling Joseph A.

Tikvart, Source Receptor Analysis Branch, MD-14, U.S. Environmental

Protection Agency, Research Triangle Park, NC 27711, phone (919) 541-

5561. Supplement C (draft) is also available to registered users of the

Support Center for Regulatory Air Models Bulletin Board System (SCRAM

BBS) by downloading the appropriate file. To register or access this

electronic bulletin board, users with a personal computer should dial

(919) 541-5742.

Docket: Copies of reports referenced herein (unless otherwise

noted) and public comments made on this Notice of Proposed Rulemaking

(NPR) are maintained in Docket A-92-65. The docket is available for

public inspection and copying between 8:00 a.m. and 4:00 p.m., Monday

through Friday, at the address above.

FOR FURTHER INFORMATION CONTACT: Joseph A. Tikvart, Chief, Source

Receptor Analysis Branch, Office of Air Quality Planning and Standards,

U.S. Environmental Protection Agency, Research Triangle Park, NC 27711;

telephone (919) 541-5561 or C. Thomas Coulter, telephone (919) 541-

0832.

SUPPLEMENTARY INFORMATION:

Background\1\

The purpose of the Guideline\2\ is to promote consistency in the

use of modeling within the air management process. The Guideline

provides model users with a common basis for estimating pollution

concentrations, assessing control strategies and specifying emission

limits; these activities are regulated at 40 CFR 51.46, 51.63, 51.112,

51.117, 51.150, 51,160, 51.166, and 51.21. The Guideline was originally

published in April 1978. It was incorporated by reference in the

regulations for the Prevention of Significant Deterioration of Air

Quality in June 1978 (43 FR 26380). The Guideline was subsequently

revised in 1986 (51 FR 32176), and later updated with the addition of

supplement A in 1987 (53 FR 393). The last such revision was supplement

B, issued on July 20, 1993 (58 FR 38816). The revisions in supplement B

included techniques and guidance for situations where specific

procedures had not previously been available, and also improved several

previously adopted techniques.

---------------------------------------------------------------------------

\1\In reviewing this preamble, note the distinction between the

terms ``supplement'' and ``appendix''. Supplements A, B and C

contain the replacement pages to effect Guideline revisions;

appendix A to the Guideline is the repository for preferred models,

while appendix B is the repository for alternate models justified

for use on a case-by-case basis.

\2\``Guideline on Air Quality Models (Revised)'' (1986) [EPA-

450/2-78-027R], with supplement A (1987) and supplement B (1993),

hereinafter, the ``Guideline''. The Guideline is published as

appendix W of 40 CFR Part 51. The text of appendix W will be

appropriately modified to effect the revisions proposed for

supplement C.

---------------------------------------------------------------------------

During the public comment period for supplement B, EPA received

requests to consider several additional new modeling techniques and

suggestions for enhanced technical guidance.\3\ However, because there

was not sufficient time for the public to review the new techniques and

technical guidance before promulgation of supplement B, the new models

and enhanced technical guidance could not be included in the supplement

B rulemaking. Thus, in this subsequent regulatory proposal, EPA is

proposing to revise the Guideline and is seeking public comment on the

four items described below. Once promulgated, these four items will be

included in supplement C to the Guideline. A copy of supplement C

(draft) is available for public review (Docket Item III-B-1).

---------------------------------------------------------------------------

\3\The official public hearing for EPA's proposal to adopt

supplement B was the Fifth Conference on Air Quality Modeling, March

1991 (56 FR 7694). Full transcripts filed in Docket No. A-88-04; IV-

F-1 (see ADDRESSES). See also ``Summary of Public Comments and EPA

Responses on the Fifth Conference on Air Quality Modeling: March

1991'', February 1993. (Docket No. A-88-04; V-C-1)

---------------------------------------------------------------------------

Proposed Action

Appendix W of 40 CFR part 51 will be appropriately amended to

effect the following revisions, proposed as supplement C to the

Guideline. EPA solicits comment on each of the following revisions.

1. Enhancements\4\ to the Industrial Source Complex Model (ISC2)

---------------------------------------------------------------------------

\4\For clarification, these enhancements are discussed

separately. EPA intends to integrate these enhancements into one

model for actual use.

---------------------------------------------------------------------------

A. Area Source Algorithm

Today's action proposes to replace the area source algorithm in the

Industrial Source Complex model (ISC2) with a new one based on a double

integration of the Gaussian plume kernel for area sources.

(1) Short-term algorithm: ISCST2. A previous EPA study\5\ indicated

that the currently implemented ISCST2 area source algorithm, based on a

finite line segment approximation, estimates concentration

distributions with limited accuracy, especially for receptors located

close to the area source. An independent but later evaluation confirmed

these findings.6,7 These studies suggested that the integrated

line source algorithm for modeling impacts from area sources provides a

better treatment of near-source geometry than that currently

recommended in ISCST2, and a reasonable far-field behavior. Based on

these performance evaluations and limited field data, the integrated

line source algorithm is a candidate to substitute for the current

ISCST2 area source algorithm. Responding to public comments received at

the time supplement B was proposed, steps were taken to develop and

test this algorithm. In the new algorithm,\8\ the ground-level

concentration at a receptor downwind of all or a portion of the area

source is given by a double integral in the upwind and crosswind

directions. The integral in the lateral direction is solved

analytically. The integral in the longitudinal direction (i.e., the

summation of the contributions from the line sources in the upwind

direction) is approximated with a Romberg integration technique.\9\ The

new algorithm, essentially equivalent to PAL\10\ and the convergent

mode of the FDM\11\ integrated line source algorithm, has been shown to

perform very well in terms of efficiency and of the reasonableness of

the results.\12\

---------------------------------------------------------------------------

\5\Environmental Protection Agency, 1989. Review and Evaluation

of Area Source Dispersion Algorithms for Emission Sources at

Superfund Sites. EPA Publication No. EPA-450/4-89-020. U.S.

Environmental Protection Agency, Research Triangle Park, NC. (NTIS

No. PB 90-142753)

\6\American Petroleum Institute, 1992. Evaluation of Area and

Volume Source Dispersion Models for Petroleum and Chemical Industry

Facilities, Phase I (Final Report). API Publication No. 4539.

(Docket No. A-92-65; II-A-1)

\7\American Petroleum Institute, 1992. Area and Volume Source

Air Quality Model Performance Evaluation, Phase II (Final Report).

API Publication No. 4540. (Docket No. A-92-65; II-A-2)

\8\``User Instructions for a New Area Source Algorithm'' (August

1993), uploaded to the SCRAM BBS. (Docket No. A-92-65; II-A-3)

\9\W.B., B. Flannery, S. Teukolsky, and W. Vetterling, 1986.

Numerical Recipes. Cambridge University Press, New York; 797 pp.

\10\Petersen, W.B., 1978. User's Guide for PAL--A Gaussian-Plume

Algorithm for Point, Area, and Line Sources. EPA Publication No.

EPA-600/4-78-013. U.S. Environmental Protection Agency, Research

Triangle Park, NC. (NTIS No. PB 281306)

\11\Environmental Protection Agency, 1991. User's Guide for the

Fugitive Dust Model (FDM) (Revised). EPA Publication No. EPA-910/9-

88-202R. U.S. Environmental Protection Agency, Region X. (NTIS No.

PB 90-502410)

\12\Environmental Protection Agency, 1992. Comparison of a

Revised Area Source Algorithm for the Industrial Source Complex

Short Term Model and Wind Tunnel Data. EPA Publication No. EPA-454/

R-92-014. U.S. Environmental Protection Agency, Research Triangle

Park, NC. (NTIS No. PB 93-226751)

---------------------------------------------------------------------------

Existing field studies of impacts within and nearby area sources

being scarce and limited in scope, EPA compared model predictions to

measured results using a wind tunnel simulation at the Fluid Modeling

Facility, Atmospheric Research and Exposure Assessment Laboratory.\13\

Both qualitative physical and quantitative statistical analyses were

performed. The analysis results\12\ show that the new algorithm

predicts the concentration distribution with relatively good accuracy

(i.e., \10%), especially for the ground-level

receptors located near the downwind edge of the area source, a

situation of concern to regulatory modeling applications. For receptors

near ground level and within or near the area source, the

normalized modeled concentrations generally matched the wind tunnel

measured concentrations to within 20%. EPA considers this

to be an acceptable correspondence.

---------------------------------------------------------------------------

\13\Snyder, W.H., 1991. DATA REPORT: Wind Tunnel Simulation of

Dispersion from Superfund Area Sources. Part: Neutral Flow. (Docket

No. A-92-65; II-a-4)

---------------------------------------------------------------------------

To examine the sensitivity of the design concentrations across a

range of source characteristics, scenarios considering source size,

elevation, and downwind distance were simulated.\14\ For each scenario,

the high-second high (HSH) 1-hour, 3-hour, 24-hour averages and high

annual averages were determined using a full year of meteorological

data; both rural and urban mode dispersion options were used.

Generally, the concentration ratio\15\ averaged 1.2 (1-

hour) to 1.0 (annual). However, for receptors located

within and nearby the area source, the ratio averaged 2 (1-

hour) to 3 (annual). Thus, for receptors inside the area

source, the ratio is higher than for receptors outside the source,

where the effect is a function of averaging time and proximity to the

source in question.

---------------------------------------------------------------------------

\14\Environmental Protection Agency, 1992. Sensitivity Analysis

of a Revised Area Source Algorithm for the Industrial Source Complex

Short Term Model. EPA Publication No. EPA-454/R-92-015. U.S.

Environmental Protection Agency, Research Triangle Park, NC. (NTIS

No. PB 93-226769)

\15\RATIO = XNEW/XOLD

---------------------------------------------------------------------------

The proposed algorithm is equivalent to that in PAL and FDM and is

more efficient than either of these algorithms. Based on comparisons

with wind tunnel data, the proposed algorithm provides a more realistic

characterization of the magnitude of impacts at receptors located

within and nearby the area than that currently in ISC2, and gives

comparable results to the FDM convergent algorithm when modeled based

on the same assumptions for release height, mixing height, and

dispersion parameters. Furthermore, these findings confirm that the

currently used area source algorithm in ISC2 is an approximation that

routinely under-estimates (and underrepresents) the actual ambient

impact, especially for receptor locations within and near an area

source.

(2) Long-term algorithm: ISCLT2. The studies previously cited in

footnotes 5, 6, and 7 have also indicated the deficiencies of the

virtual point source algorithm used in ISCLT2. While it is

computationally efficient, the virtual point source algorithm used in

the original ISCLT2 yields estimates of limited accuracy for receptors

located near the edges and corners of the area, a problem also seen

with the original ISCST2. The algorithm cannot predict the area source

impact for receptors located inside the source itself, and does not

adequately handle effects of complex source-receptor geometry.

Thus, a new area source algorithm for the ISCLT2, based on the

numerical integration algorithm described above, was developed and

evaluated.\16\ Detailed performance tests, statistical analyses and

sensitivity analyses were completed to assure the reliability and

reasonableness of the modeling results. Using idealized meteorological

conditions, the new algorithm yields very good comparison results when

compared with the newly developed ISCST2 area source algorithm. For

realistic meteorological data, the differences between ground level

concentration values simulated with the new ISCLT2 algorithm and with

the new ISCST2 counterpart are within about 10% for a typical source.

The differences between the long-term and short-term algorithms using

actual meteorological data are because ISCLT2 uses a meteorological

frequency distribution to represent the meteorological conditions, and

does not contain precise hour-to-hour information on specific

combinations of wind speed, wind direction, stability class and mixing

height that typically control the design values for the short-term

model. Furthermore, sensitivity analyses show that the current ISCLT2

area source algorithm, based on the virtual source approach, routinely

underestimates (and underrepresents) the actual maximum concentration

impacts by a factor of 2 to 4, especially when the receptors are

located inside or near the source.

---------------------------------------------------------------------------

\16\Environmental Protection Agency, 1992. Development and

Evaluation of a Revised Area Source Algorithm for the Industrial

Source Complex Long Term Model. EPA Publication No. EPA-454/R-92-

016. U.S. Environmental Protection Agency, Research Triangle Park,

NC. (NTIS No. PB 93-226777)

---------------------------------------------------------------------------

B. Dry Deposition Algorithm

Deposition phenomena can be conceptualized in a two by two matrix,

with a wet/dry dichotomy on one side and a particle/gas dichotomy on

the other. Each of the four cells can then be further subdivided into

simple and complex terrain components. Today's action proposes to

replace the plume depletion and dry deposition algorithm\17\ in the

Industrial Source Complex model (ISC2) with a new algorithm that

estimates the amount of material depleted from the plume as a

combination of processes involving atmospheric turbulence and

gravitational settling. This proposal embodies the simple terrain

component of one cell in the conceptual matrix: dry deposition applied

to particles. It is proposed that the new algorithm be implemented to

treat dry deposition in rolling terrain, which is not possible in the

current versions of ISC2. Future efforts may be directed at better

characterizing gaseous and wet deposition in simple and complex.

---------------------------------------------------------------------------

\17\``User Instructions for the Draft Deposition Models DEPST

and DEPLT'' (March 1994) have been uploaded to the SCRAM BBS.

(Docket No. A-92-65; II-A-5).

---------------------------------------------------------------------------

The dry deposition algorithm currently used in ISC2 is applicable

to large particles (i.e., those with diameters greater than

20m) for which deposition is dominated by

gravitational settling. In 1993, EPA initiated a study to evaluate the

performance of alternative deposition algorithms. A review of the

technical literature identified four core algorithms and six variants

suitable for testing, producing a field of ten algorithm candidates.

Estimates based on these algorithms were compared with observations

from several data bases. Objective statistical procedures\18\ were used

to measure model performance. The main feature of this approach is to

compute normalized statistical measures of the fractional bias between

observed and predicted values.

---------------------------------------------------------------------------

\18\Environmental Protection Agency, 1992. Protocol for

Determining the Best Performing Model. EPA Publication No. EPA-454/

R-92-025. U.S. Environmental Protection Agency, Research Triangle

Park, NC. (NTIS No. PB 93-226082)

---------------------------------------------------------------------------

Based on the evaluation,\19\ the performance among the three top-

ranked dry deposition algorithms was statistically indistinguishable.

The three top-ranked models were UAM 2, CARB 3 and ADOM 1. The UAM 2

and CARB 3 algorithms represent a hybrid variant of their respective

core algorithms with an added Leaf Area Index (LAI)\20\ adjustment.

ADOM 1, currently employed in the Acid Deposition and Oxidant Model, is

a core algorithm (does not include a LAI adjustment). The results of

the evaluation suggest that the reflection coefficient method used in

ISC2 does not perform well for particle sizes less than 20m in

diameter.

---------------------------------------------------------------------------

\19\Environmental Protection Agency, 1994. Development and

Testing of a Dry Deposition Algorithm (Revised). EPA Publication No.

EPA-454/R-94-015. U.S. Environmental Protection Agency, Research

Triangle Park, NC. (NTIS No. PB 94-183100)

Note: This report replaces one previously completed because an

error was discovered after the earlier report was issued. The

following memorandum details the nature of the error and documents

the validity of the newer report.

Memorandum from Jawad S. Touma et al. to Joseph A. Tikvart:

Comments on the report ``Development and Testing of a Dry Deposition

Algorithm (Revised)'', 6 May 1994 (3pp. w/5 attachments) (Dockets

No. A-92-65; II-E-1)

\20\The LAI is a ration of leaf surface area divided by ground

surface area and can be estimated from land use type and season.

---------------------------------------------------------------------------

The technical applicability of a LAI adjustment, as implemented for

particle deposition velocity, has not been extensively studied. Thus,

the robustness of using a LAI in routine model applications is

uncertain. Excluding algorithms with LAI adjustments, the ADOM 1 scheme

produces the best composite fractional bias measure (CPM) and was

significantly better than other models tested at the 95% confidence

level. ADOM 1 slightly underestimates observed deposition velocities, a

trait that is shared by all the algorithm candidates. Considering all

of these factors, ADOM 1 is recommended for estimating dry deposition

velocity in the ISC2 model.

The ADOM 1 dry deposition algorithm has been tested within the

framework of the ISC2 model and comparisons of deposition estimates

using the old and new deposition algorithms have been made for a range

of source types and particulate emission scenarios. Similar comparisons

have been made of particulate concentration estimates as affected by

the old and new deposition algorithms. A report\21\ documenting these

analyses and assessing the potential consequences of replacing the

current deposition algorithm in ISC2 with the proposed algorithm has

been prepared.

---------------------------------------------------------------------------

\21\Environmental Protection Agency, 1994. Comparison of ISC2

Deposition Estimates Based on Current and Proposed Deposition

Algorithms. EPA Publication No. EPA-454/R-94-018. U.S. Environmental

Protection Agency, Research Triangle Park, NC.

---------------------------------------------------------------------------

The results of the comparative analyses of the proposed dry

deposition algorithm vary with release type, particle size, and

averaging period. Consequently, care should be exercised in

interpreting the generalizations that follow regarding deposition and

concentration estimates.

The effects on the actual deposition predicted by ADOM 1 were

examined. For surface releases, the new algorithm gives higher annual

and 24-hour deposition estimates for all particle sizes. For 1-hour and

3-hour estimates for surface releases the results were mixed. For

elevated releases, deposition estimates given by the new algorithm are

higher for 0.1m and 1m particles, lower for 10 and

20m particles, and higher for 80m and 100m

particles. The results for elevated releases of 50m particles

depend on release height.

The effects on ambient concentrations predicted by ISC2 were also

examined. For both surface and elevated releases of small and

intermediate particle sizes (i.e., 0.1, 1.0, 10 and 20m), the

differences in concentration estimates between the old and new

algorithms are less than 10 percent. These differences are considered

insignificant. Results for the large particle sizes (i.e., 50, 80, and

100m) depend on release height. For surface releases, the

concentration estimates using the new algorithm are diminished. For

elevated releases, concentration estimates using the new algorithm are

increased.

EPA is also soliciting public comment on whether it would be

appropriate to require the proposed dry deposition algorithm to be used

for all ISC2 analyses involving particulate matter in any of the

programs for which Guideline usage is required under 40 CFR parts 51

and 52 (see Summary). Heretofore, use of the deposition algorithm has

been optional, depending on the relevance of particle deposition to a

particular application. However, with the more accurate deposition

algorithm proposed herein, its use may result in the systematic

prediction of more accurate ambient concentrations. Therefore, EPA is

soliciting comment on whether it would be appropriate to revise

Guideline section 8.2.7 (Gravitational Settling and Deposition) to

require use of the deposition algorithm, and if so, whether the

implementation guidance provided in the User's Instructions\17\ is

sufficient.

2. Enhancements to On-site Stability Classification

EPA is proposing to revise the on-site stability classification

with the adoption of a new technique, adapted from Bowen et al.\22\ and

herein referred to as the solar radiation/delta-T (SRDT) method. This

method uses total solar radiation during daytime and temperature

difference, delta-T (T), at night and is a replacement for the

one originally proposed (56 FR 5900). As proposed in supplement C, the

hierarchy of stability classification schemes in the Guideline will be

changed to reflect a preference for SRDT-derived stability categories.

Operation of the method is fully described in section 6.4.4.2 of ``On-

Site Meteorological Program Guidance for Regulatory Modeling

Applications'' (EPA-450/4-87-013), hereafter, ``on-site guidance''.

---------------------------------------------------------------------------

\22\Bowen, B.M., J.M. Dewart, and A. I. Chen, 1983. Stability

Class Determination: A Comparison for One Site. Proceedings, Sixth

Symposium on Turbulence and Diffusion, American Meteorological

Society, Boston, MA; pp. 211-214. (Docket No. A-92-65; II-A-6)

---------------------------------------------------------------------------

The new method has been completely reconfigured in terms of its

classification criteria, in response to the public comments provided at

the Fifth Conference on Air Quality Modeling (March 1991) regarding the

original proposal. The comments (Docket A-88-04, Category IV-D; see

footnote #4) were generally favorable to the concept of a SRDT method

for determining stability category. However, there were some

substantial criticisms of specific SRDT components. Most significant

were comments on:

(1) Accuracy of measurements associated with a 2-10m T;

(2) Limitations on temperature measurements made at 2m;

(3) Use of a 10-60m T in lieu of one measured from 2-10m;

(4) Lack of evaluation data bases;

(5) Use of net radiation measurements in lieu of solar radiation;

and

(6) Merits of measurements for stability determination.

Regarding the use of net radiation, it is not apparent that there

is sufficient experience with routine use of such measurements to

justify requiring their use, whereas there has been extensive

experience with T systems. Regarding the use of

measurements, experience has been that, unless such systems are tuned

for site-specific regimes, the -based methods do not represent

Pasquill-Gifford (P-G) stability classification well. Evaluation

results,\23\ based on on-site measurements from three widely separated

locations, indicate that the SRDT method seems to be less sensitive to

local measurement configurations and is expected to be geographically

robust. Furthermore, the new SRDT method has been configured so that

the system accuracy will not be limiting. Thus, the method will be less

sensitive to random temperature differences. The claim (commenter IV-D-

27 in Docket Item V-C-1; see footnote #3) that accurate measurement of

the 2m temperature may be adversely affected by surface conditions

under the tower has merit in certain circumstances. The new SRDT method

does not mandate that the location of the lower temperature sensor be

at 2m. EPA believes that proper siting of temperature probes in

accordance with Chapter 3 of the on-site guidance, coupled with sound

judgment, should obviate any such problem. Use of a 10-60m T,

an interval specified in the meteorological monitoring protocol used by

the Nuclear Regulatory Commission, is accommodated by the new SRDT

method. Finally, substantial effort was made in acquiring suitable on-

site data bases with which to evaluate the new SRDT method; the new

SRDT method has been more extensively evaluated.

---------------------------------------------------------------------------

\23\Environmental Protection Agency, 1993. An Evaluation of a

Solar Radiation/Delta-T (SRDT) Method for Estimating Pasquill-

Gifford (P-G) Stability Categories. EPA Publication No. EPA-454/R-

93-055. U.S. Environmental Protection Agency, Research Triangle

Park, NC. (NTIS No. PB 94-113958)

---------------------------------------------------------------------------

To make the stability classification comparisons for the SRDT

evaluation, a surrogate for the preferred Turner classification

scheme\24\ was devised. This surrogate method utilized ``off-site''

National Weather Service (NWS) observations in lieu of those otherwise

made ``on-site''. To ensure the integrity of this surrogate method, it

was necessary that candidate sites be sufficiently near a

representative NWS station from which cloud cover and ceiling height

observations could be obtained. Of ten on-site data bases considered

for supporting the evaluation, three were ultimately selected because

they had the requisite attributes. The data bases thus selected were:

Kincaid, IL (21 weeks in 1980), Longview, WA (CY 1991), and a site near

Bloomington, IN (7/91-7/92). Proximity of these sites to NWS stations

ranged from 17 to 45 miles.

---------------------------------------------------------------------------

\24\This method requires on-site measurements of wind speed

coupled with observations of cloud cover and ceiling height. Turner,

D.B., 1964. A Diffusion Model for an Urban Area. Journal of Applied

Meteorology, 3(1): 83-91.

---------------------------------------------------------------------------

For theoretical reasons, as well as for consistency with the

approach originally proposed, the SRDT method was initially evaluated

using T data from 2-10m; such data were available for all

three sites. At two of the sites, T data from 10-50m were also

available. These data were of interest in trying to accommodate

T measurements from alternative height intervals.

As substantial site-to-site variability was seen in initial

analyses using the 2-10m T data, it was decided to pool the

data from all three sites and then determine optimum SRDT ``cutpoints''

(i.e., meteorological criteria for discriminating stability category).

Thus, optimum cutpoints were derived in an empirical, iterative fashion

from a data base of 19,540 valid hours. Use of these optimum cutpoints

resulted in a SRDT system that estimated the same P-G stability as the

preferred Turner scheme for 62% of the hours; the categories were

within one class for 89% of the hours. A randomization procedure in

which the composite data were split into two complementary sets was

done to ascertain robustness (insensitivity to random variations in the

data) of the method. The optimum cutpoints from the composite data were

then applied to the three sites individually to document site-specific

residuals.

For the two sites with 10-50m T data, the SRDT system

using the optimum (for pooled data) cutpoints was applied in the same

way as with the 2-10m T data, with reasonably accurate and

consistent results. Stability categories were duplicated by the SRDT

method at least 56% of the hours, and were within one class for about

90% of the hours. Overall, the analyses show that the SRDT system works

adequately for either T interval: the system does not appear

to be unduly sensitive to the actual T height interval. Based

on these analyses, EPA does not feel it should be overly prescriptive

regarding the use of particular T intervals. Rather, in

guidance for implementation of the method, actual placement of

temperature probes is related to fundamental site-specific phenomena,

e.g., surface roughness. While the method was evaluated using only 2-

10m and 10-50m T data, it is considered to be robust enough to

accommodate other T height intervals as well, so long as

section 6.4.4.2 of the on-site guidance cited above is followed.

Finally, consequence analyses were performed using a Gaussian

dispersion model (i.e., ISC2) to document the effect of the SRDT method

on design concentration ratios.\25\ These analyses were performed for

the 2-10m T comparisons at all three sites and for the 10-50m

T comparisons at two sites. For all such analyses, scenarios

included single 35m, 100m and 200m stacks and 180 receptors configured

radially in 5 concentric rings. Averaging times included 1-hour, 3-

hour, 24-hour, and period. Modeled concentrations of interest were the

high, and high 2nd high value. Using stability categories derived from

the 2-10m T data for the three sites, the concentration ratios

averaged 1.06-1.24 across three source types, four averaging times and

two concentration types. Likewise, using those categories derived from

the 10-50m T data, the same concentration ratios also averaged

1.06-1.24.

---------------------------------------------------------------------------

\25\RATIO=XSRDT/XTurner

---------------------------------------------------------------------------

In the supplement B revisions to the Guideline, EPA referenced

``On-Site Meteorological Program Guidance for Regulatory Modeling

Applications'' in section 9.3.3. This document continues to serve as

the primary source of supplementary technical guidance on the

collection and use of on-site meteorological data. EPA is proposing an

addendum\26\ to accommodate the technical details of the SRDT system.

Once finalized, the hierarchy of stability classification schemes in

that document will also be changed to reflect the preference for those

derived via SRDT. The use of other techniques prior to a year following

promulgation will be exempt from this provision, after which they will

not be considered the primary method for estimating stability. Finally,

the module designed to implement the SRDT system in Version 1.3 of the

Meteorological Processor for Regulatory Models (MPRM), EPA-600/3-88-

043, will be activated and configured with the optimum cutpoints

derived in the evaluation.

---------------------------------------------------------------------------

\26\ADDENDUM: On-Site Meteorological Program Guidance for

Regulatory Modeling Applications. Draft for Public Comment

(September 1993). (Docket No. A-92-65; II-A-7)

---------------------------------------------------------------------------

3. Screening Approaches for Assessing Annual NO2 Impact

EPA is proposing a revision to simplify the screening approaches

for assessing annual NO2 concentration impact in Guideline section

6.2.3.

These revisions respond to public comments contending that the

initial screening level (which assumed total conversion of NO to

NO2) was overly conservative, and that the ozone limiting approach

described in the second and third screening levels was sometimes

inapplicable or impracticable. Thus, a second level screening approach

that embodies use of an empirically derived NO2/NOX ratio is

proposed. This method replaces the multi-tiered ozone limiting method

now recommended in the Guideline. As described in Chu and Meyer

(1991),\27\ the new approach reflects a review of 10 years of ambient

NO2 and NOX concentration data collected at a variety of

monitoring sites throughout the United States.

---------------------------------------------------------------------------

\27\Chu, S.-H. and E.L. Meyer, 1991. Use of Ambient Ratios to

Estimate Impact of NOX Sources on Annual NO2

Concentrations. Proceedings, 84th Annual Meeting & Exhibition of the

Air & Waste Management Association, Vancouver, B.C.; 16-21 June 1991

(16pp.). (Docket No. A-92-65; II-A-8)

---------------------------------------------------------------------------

The underlying basis for the ambient ratio method (ARM) is that,

for a well mixed plume, the photochemical conversion of NO to NO2

is essentially controlled by the characteristics of the ambient air.

This, in turn, is reflected in the annual NO2/NOX ratio

monitored downwind. Since the photochemistry involved in converting NO

to NO2 is implicitly accounted for by the annual NO2/NOX

ratio monitored downwind, no long-term complex photochemical

calculation is needed. Thus, it makes the modeling exercise much

simpler, yet still provides results consistent with available plume

observational studies.

The method is conservative since, in many cases, maximum estimated

ground level NOX concentration may occur prior to thorough mixing

of the plume. A second, less important, source of conservatism is that

the existing NO2 and NOX data may overestimate the actual

NO2 and NOX concentrations due to interference of PAN and

nitric acid in the measurement. However, since the same amount is added

to both the numerator and denominator of the NO2NOX ratio, it

only makes the conversion ratio slightly more conservative. As shown by

Chu and Meyer (1991), the ARM, while likely to be conservative, is

somewhat less so than existing screening methods (such as the total

conversion and the ozone limiting method) for estimating annual

NO2 concentrations and PSD NO2 increments for NOX

sources. Serving as a second level screening method, ARM has the

quality of simplicity, is easy to apply and is likely to be somewhat

conservative. It relies only on the standard regulatory Gaussian models

and data from nationwide NOX monitoring networks. EPA has

therefore selected this method to propose as a revision to the

Guideline in supplement C.

4. Modeling Techniques for Toxic Air Pollutants

In response to a request made by the American Petroleum Institute

(see footnote 3), two new models for treating toxic air pollutant

releases are being proposed for addition to appendix B of the

Guideline. These models, SLAB and HGSYSTEM, will then accompany

DEGADIS, another appendix B model for treating dense gas releases for

use on a case-by-case basis. (See footnote 2.)

Administrative Requirements

A. Executive Order 12866

Under Executive Order (E.O.) 12866 [58 FR 51735 (October 4, 1993)],

the Agency must determine whether the regulatory action is

``significant'' and therefore subject to review by the Office of

Management and Budget (OMB) and the requirements of the Executive

Order. The Order defines ``significant regulatory action'' as one that

is likely to result in a rule that may:

(1) Have an annual effect on the economy of $100 million or more

or adversely affect in a material way the economy, a sector of the

economy, productivity, competition, jobs, the environment, public

health or safety, or State, local, or tribal governments or

communities;

(2) Create a serious inconsistency or otherwise interfere with

an action taken or planned by another agency;

(3) Materially alter the budgetary impact of entitlements,

grants, user fees, or loan programs of the rights and obligations of

recipients thereof; or

(4) Raise novel legal or policy issues arising out of legal

mandates, the President's priorities, or the principles set forth in

the Order.

It has been determined that this rule is not a ``significant

regulatory action'' under the terms of E.O. 12866 and is therefore not

subject to OMB review.

B. Paperwork Reduction Act

This final rule does not contain any information collection

requirements subject to review by OMB under the Paperwork Reduction Act

of 1980, 44 U.S.C. 3501 et seq.

C. Regulatory Flexibility Act

The Regulatory Flexibility Act (5 U.S.C. 601 et seq.) requires EPA

to consider potential impacts of regulations on small ``entities''. The

action here proposed is a supplement to the notice of final rulemaking

that was published on July 20, 1993 (58 FR 38816). As described earlier

in this preamble, the revisions here proposed as supplement C to the

Guideline encompass the use of new model algorithms and techniques for

using those models. This rule merely updates existing technical

requirements for air quality modeling analyses mandated by various

Clean Air Act programs (e.g., prevention of significant deterioration,

new source review, SIP revisions) and imposes no new regulatory

burdens. As such, there will be no additional impact on small entities

regarding reporting, recordkeeping, compliance requirements, as stated

in the notice of final rulemaking (op. cit.). Furthermore, this

proposed rule does not duplicate, overlap, or conflict with other

federal rules. Thus, pursuant to the provisions of 5 U.S.C. 605(b), EPA

hereby certifies that the attached proposed rule will not have a

significant impact on a substantial number of such entities.

List of Subjects

40 CFR Part 51

Administrative practice and procedure, Air pollution control,

Intergovernmental relations, Reporting and recordkeeping requirements,

Ozone, Sulfur oxides, Nitrogen dioxide, Lead, Particulate matter,

Hydrocarbons, Carbon monoxide.

40 CFR Part 52

Air pollution control, Ozone, Sulfur oxides, Nitrogen dioxide,

Lead.

Authority: This notice of proposed rulemaking is issued under

the authority granted by sections 110(a)(2), 165(e), 172(a) & (c),

173, 301(a)(1) and 320 of the 1990 Clean Air Act Amendments, 42

U.S.C. 7410(a)(2), 7475(e), 7502(a) & (c), 7503, 7601(a)(1) and

7620, respectively.

Dated: November 7, 1994.

Carol M. Browner,

Administrator.

Parts 51 and 52, Chapter I, Title 40 of the Code of Federal

Regulations are amended as follows:

PART 51--REQUIREMENTS FOR PREPARATION, ADOPTION, AND SUBMITTAL OF

IMPLEMENTATION PLANS

1. The authority citation for part 51 continues to read as follows:

Authority: 42 U.S.C. 7410(a)(2), 7475(e), 7502(a) and (b), 7503,

7601(a)(1) and 7620.

Sec. 51.112 [Amended]

2. In Sec. 51.112, paragraphs (a)(1) and (a)(2) are amended by

revising ``and supplement B (1993)'' to read ``, supplement B (1993)

and supplement C (1994)''.

Sec. 51.160 [Amended]

3. In Sec. 51.160, paragraphs (f)(1) and (f)(2) are amended by

revising ``and supplement B (1993)'' to read ``, supplement B (1993)

and supplement C (1994)''.

Sec. 51.166 [Amended]

4. In Sec. 51.166, paragraph (l)(1) and (l)(2) are amended by

revising ``and supplement B (1993)'' to read ``, supplement B (1993)

and supplement C (1994)''.

5. Appendix W to Part 51, section 4.2.2, is amended by revising

footnote 1 in Table 4-1 to read as follows:

Appendix W to Part 51--Guideline on Air Quality Models (Revised)

* * * * *

4.2.2 * * *

Table 4-1. * * *

------------------------------------------------------------------------

Land

Use Model\1\

------------------------------------------------------------------------

*****

\1\The models as listed in this table reflect the applications for which

they were originally intended. Several of these models have been

adapted to contain options which allow them to be interchanged. For

example, ISCST2 could be substituted for ISCLT2. Similarly, for a

point source application, ISCST2 with urban option can be substituted

for RAM. Where a substitution is convenient to the user and equivalent

estimates are assured, it may be made.

* * * * *

Appendix W [Amended]

6. Appendix W to Part 51, section 6.2.3, is revised to read as

follows:

* * * * *

6.2.3 Models for Nitrogen Dioxide (Annual Average)

a. A tiered screening approach is recommended to obtain annual

average estimates of NO2 from point sources for New Source

Review analysis, including PSD, and for SIP planning purposes. This

multi-tiered approach is conceptually shown in Figure 6-1 below:

Figure 6-1--Multi-tiered Screening Approach for Estimating Annual

NO2 Concentrations From Point Sources

BILLING CODE 6560-50-M

TP28NO94.000

BILLING CODE 6560-50-C

b. For Tier 1 (the initial screen), use an appropriate Gaussian

model from appendix A to estimate the maximum annual average

concentration and assume a total conversion of NO to NO2. If

the concentration exceeds the NAAQS and/or PSD increments for

NO2, proceed to the 2nd level screen.

c. For Tier 2 (2nd level) screening analysis, multiply the Tier

1 estimate(s) by an empirically derived NO2/NOX value of

0.75 (annual national default).\36\ An annual NO2/NOX

ratio differing from 0.75 may be used if it can be shown that such a

ratio is based on data likely to be representative of the

location(s) where maximum annual impact from the individual source

under review occurs. In the case where several sources contribute to

consumption of a PSD increment, a locally derived annual NO2/

NOX ratio should also be shown to be representative of the

location where the maximum collective impact from the new plus

existing sources occurs.

d. In urban areas, a proportional model may be used as a

preliminary assessment to evaluate control strategies to meet the

NAAQS for multiple minor sources, i.e. minor point, area and mobile

sources of NOX; concentrations resulting from major point

sources should be estimated separately as discussed above, then

added to the impact of the minor sources. An acceptable screening

technique for urban complexes is to assume that all NOX is

emitted in the form of NO2 and to use a model from appendix A

for nonreactive pollutants to estimate NO2 concentrations. A

more accurate estimate can be obtained by: (1) calculating the

annual average concentrations of NOX with an urban model, and

(2) converting these estimates to NO2 concentrations using an

empirically derived annual NO2/NOX ratio. A value of 0.75

is recommended for this ratio. However, a spatially averaged annual

NO2/NOX ratio may be determined from an existing air

quality monitoring network and used in lieu of the 0.75 value if it

is determined to be representative of prevailing ratios in the urban

area by the reviewing agency. To ensure use of appropriate locally

derived annual NO2/NOX ratios, monitoring data under

consideration should be limited to those collected at monitors

meeting siting criteria defined in 40 CFR 58, appendix D as

representative of ``neighborhood'', ``urban'', or ``regional''

scales. Furthermore, the highest annual spatially averaged NO2/

NOX ratio from the most recent 3 years of complete data should

be used to foster conservatism in estimated impacts.

e. To demonstrate compliance with NO2 PSD increments in

urban areas, emissions from major and minor sources should be

included in the modeling analysis. Point and area source emissions

should be modeled as discussed above. If mobile source emissions do

not contribute to localized areas of high ambient NO2

concentrations, they should be modeled as area sources. When modeled

as area sources, mobile source emissions should be assumed uniform

over the entire highway link and allocated to each area source grid

square based on the portion of highway link within each grid square.

If localized areas of high concentrations are likely, then mobile

sources should be modeled as line sources with the preferred model

ISCLT2.

f. More refined techniques to handle special circumstances may

be considered on a case-by-case basis and agreement with the

reviewing authority should be obtained. Such techniques should

consider individual quantities of NO and NO2 emissions,

atmospheric transport and dispersion, and atmospheric transformation

of NO to NO2. Where they are available, site-specific data on

the conversion of NO to NO2 may be used. Photochemical

dispersion models, if used for other pollutants in the area, may

also be applied to the NOX problem.

* * * * *

7. Appendix W to Part 51, section 9.3.3.2, is revised to read as

follows:

* * * * *

9.3.3.2 Recommendations--Site-specific Data Collection.

a. The document ``On-Site Meteorological Program Guidance for

Regulatory Modeling Applications''\66\ provides recommendations on

the collection and use of on-site meteorological data.

Recommendations on characteristics, siting, and exposure of

meteorological instruments and on data recording, processing,

completeness requirements, reporting, and archiving are also

included. This publication should be used as a supplement to the

limited guidance on these subjects now found in the ``Ambient

Monitoring Guidelines for Prevention of Significant

Deterioration''.\63\ Detailed information on quality assurance is

provided in the ``Quality Assurance Handbook for Air Pollution

Measurement Systems: Volume IV''.\67\ As a minimum, site-specific

measurements of ambient air temperature, transport wind speed and

direction, and the parameters to determine Pasquill-Gifford (P-G)

stability categories should be available in meteorological data sets

to be used in modeling. Care should be taken to ensure that

meteorological instruments are located to provide representative

characterization of pollutant transport between sources and

receptors of interest. The Regional Office will determine the

appropriateness of the measurement locations.

b. All site-specific data should be reduced to hourly averages.

Table 9-3 lists the wind related parameters and the averaging time

requirements.

c. Solar Radiation Measurements. Total solar radiation should be

measured with a reliable pyranometer, sited and operated in

accordance with established on-site meteorological guidance.\66\

d. Temperature Measurements. Temperature measurements should be

made at standard shelter height (2m) in accordance with established

on-site meteorological guidance.\66\

e. Temperature Difference Measurements. Temperature difference

() measurements for use in estimating P-G

stability categories using the SRDT methodology (see Stability

Categories) should be obtained using two matched thermometers or a

reliable thermocouple system to achieve adequate accuracy.

f. Siting, probe placement, and operation of T systems

should be based on guidance found in Chapter 3 of reference 66, and

such guidance should be followed when obtaining vertical temperature

gradient data for use in plume rise estimates or in determining the

critical dividing streamline height.

g. Wind Measurements. The wind speed for determining plume rise

using the methods of Briggs56,57 should be measured at stack

top. For refined modeling applications in simple terrain situations,

if a source has a stack below 100m, select the stack top height as

the wind measurement height for characterization of plume dilution

and transport. In some cases, collection of stack top wind speed may

be impractical. For sources with stacks extending above 100m, a 100m

tower is suggested unless the stack top is significantly above 100m

(i.e., 200m). In cases with stack tops 200m,

the Regional Office should determine the appropriate measurement

height on a case-by-case basis. Remote sensing may be a feasible

alternative.

h. For refined modeling applications in complex terrain,

multiple level (typically three or more) measurements of wind speed

and direction, temperature and turbulence (wind fluctuation

statistics) are required. Such measurements should be obtained up to

the representative plume height(s) of interest (i.e., the plume

height(s) under those conditions important to the determination of

the design concentration). The representative plume height(s) of

interest should be determined using an appropriate complex terrain

screening procedure (e.g., CTSCREEN) and should be documented in the

monitoring/modeling protocol. The necessary meteorological

measurements should be obtained from an appropriately sited

meteorological tower augmented by SODAR if the representative plume

height(s) of interest exceed 100m. The meteorological tower need not

exceed the lesser of the representative plume height of interest

(the highest plume height if there is more than one plume height of

interest) or 100m.

i. Specifications for wind measuring instruments and systems are

contained in the ``On-Site Meteorological Program Guidance for

Regulatory Modeling Applications''.\66\

j. Stability Categories. The (P-G) stability categories, as

originally defined, couple near-surface measurements of wind speed

with subjectively determined insolation assessments based on hourly

cloud cover and ceiling height observations. The wind speed

measurements are made at or near 10m. The insolation rate is

typically assessed using observations of cloud cover and ceiling

height based on criteria outlined by Turner.\50\ In the absence of

site specific observations of cloud cover and ceiling height, it is

recommended that the P-G stability category be estimated using the

solar radiation/delta-T (SRDT) method described in section 6.4.4.2

of reference 66. This method requires measurements of total solar

radiation during the daytime and temperature difference ()

at night (see Temperature Difference Measurements), coupled with

average wind speed at 10m above ground level. This technique is

modified slightly from that published by Bowen et al. (1983),\136\

has been evaluated with three on-site data bases,\137\ and allows

practical and reasonable implementation of the preferred Turner

method.\55\

k. Two methods of stability classification which use wind

fluctuation statistics, the and

methods, are also described in detail in

reference 66 (note applicable tables in Chapter 6). As a primary

method, these two techniques may only be used for processing data

collected within 1 year following the promulgation date of

Supplement C, and then only when data are unavailable to implement

either the preferred Turner method\55\ or the SRDT method. After

promulgation of Supplement C, these turbulence methods should only

be used to provide back-up stability category estimates for missing

hours in the record according to an established data substitution

protocolg and after valid data retrieval requirements have been

met.

---------------------------------------------------------------------------

\2\Such protocols are usually part of the approved monitoring

program plan. Data substitution guidance is provided in section 5.3

of reference 66.

---------------------------------------------------------------------------

l. In the case of the method it should be

noted that wind meander may occasionally bias the determination of

and thus lead to an erroneous determination of

the P-G stability category. To minimize wind direction meander

contributions, may be determined for each of four

15-minute periods in an hour. However, 360 samples are needed during

each 15-minute period. If the method is being

used for stability determinations in these situations, take the

square root of one-quarter of the sum of the squares of the four 15

minute 's, as illustrated in the footnote to Table

9-3. While this approach is an acceptable alternative for

determining stability, as qualified above, 's

calculated in this manner are not likely to be suitable for input to

models that are designed to accept on-site hourly 's based

on 60-minute periods, e.g., CTDMPLUS. For additional information on

stability classification using wind fluctuation statistics, see

references 68-72.

m. In summary, when on-site data are being used, P-G stability

categories should be determined by (1) Turner's method\55\ using

site specific data which include cloud cover, ceiling height and

surface (10m) wind speeds, or (2) the radiation-based

technique (SRDT) described in reference 66.

n. The following techniques may only be applied to on-site data

bases collected within 1 year following the promulgation date of

Supplement C, and then only when data are unavailable to implement

the preferred Turner\55\ or SRDT method; or to provide back-up

stability category estimates for missing hours in the record

according to an established data substitution protocolg and

after valid data retrieval requirements have been met (choice is

based on data availability and site suitability):

(1) from site-specific measurements in accordance

with guidance;\66\

(2) from site-specific measurements in accordance

with guidance;\66\

(3) Turner's method\55\ using site-specific wind speed with cloud

cover and ceiling height from a nearby NWS site.

o. Meteorological Data Processors. The following meteorological

preprocessors are recommended by EPA: RAMMET, PCRAMMET, STAR,

PCSTAR, MPRM,\135\ and METPRO.\24\ RAMMET is the recommended

meteorological preprocessor for use in applications employing hourly

NWS data. The RAMMET format is the standard data input format used

in sequential Gaussian models recommended by EPA. PCRAMMET\138\ is

the PC equivalent of the mainframe version (RAMMET). STAR is the

recommended preprocessor for use in applications employing joint

frequency distributions (wind direction and wind speed by stability

class) based on NWS data. PCSTAR is the PC equivalent of the

mainframe version (STAR). MPRM is the recommended preprocessor for

use in applications employing on-site meteorological data. The

latest version (MPRM 1.3) has been configured to implement the SRDT

method for estimating P-G stability categories. MPRM is a general

purpose meteorological data preprocessor which supports regulatory

models requiring RAMMET formatted data and STAR formatted data. In

addition to on-site data, MPRM provides equivalent processing of NWS

data. METPRO is the required meteorological data preprocessor for

use with CTDMPLUS. All of the above mentioned data preprocessors are

available for downloading from the SCRAM BBS.\19\

* * * * *

8. Appendix W to Part 51, section 12.0, is amended by:

a. Redesignating footnote g and h as footnotes h and i;

b. Revising references 36 and 90; and

c. Adding references 136 through 138.

The revisions and additions read as follows:

* * * * *

12.0 * * *

* * * * *

36. Chu, S.-H. and E. L.Meyer, 1991. Use of Ambient Ratios to

Estimate Impact of NOx Sources on Annual NO2

Concentrations. Proceedings, 84th Annual Meeting & Exhibition of the

Air & Waste Management Association, Vancouver, B.C.; 16-21 June

1991. (16pp.) (Docket No. A-92-65, II-A-7)

* * * * *

90. Environmental Research and Technology, 1987. User's Guide to the

Rough Terrain Diffusion Model (RTDM), Rev. 3.20. ERT document No.

PD535-585. Environmental Research and Technology, Inc., Concord, MA.

(NTIS No. PB 88-171467)

* * * * *

136. Bowen, B.M., J.M. Dewart and A.I. Chen, 1983. Stability Class

Determination: A Comparison for One Site. Proceedings, Sixth

Symposium on Turbulence and Diffusion. American Meteorological

Society, Boston, MA; pp. 211-214. (Docket No. A-92-65, II-A-5)

137. Environmental Protection Agency, 1993. An Evaluation of a Solar

Radiation/Delta-T (SRDT) Method for Estimating Pasquill-Gifford (P-

G) Stability Categories. EPA Publication No. EPA-454/R-93-055. U.S.

Environmental Protection Agency, Research Triangle Park, NC. (NTIS

No. PB 94-113958)

138. Environmental Protection Agency, 1993. PCRAMMET User's Guide.

EPA Publication No. EPA-454/B-93-009. U.S. Environmental Protection

Agency, Research Triangle Park, NC.

* * * * *

Appendix W [Amended]

9. Appendix W to Part 51, section 13.0, is amended by redesignating

footnote i as footnote j.

Appendix W [Amended]

10. Appendix W to Part 51, Appendix A, is amended by:

a. Revising section A.5.d;

b. Revising section A.5.m;

c. Adding four references in alphabetical order in section A.5.n;

and

d. Adding a reference at the end of section A.REF.

The revisions and additions read as follows:

Appendix A to Appendix W of Part 51--Summaries of Alternative Air

Quality Models

* * * * *

A. 5 * * *

d. Type of Model

ISC2 is a Gaussian plume model. It has been revised to perform a

double integration of the Gaussian plume kernel for area sources.

* * * * *

m. Physical Removal

Dry deposition effects for particles are treated using a

resistance formulation in which the deposition velocity is the sum

of the resistances to pollutant transfer within the surface layer of

the atmosphere, plus a gravitational settling term (EPA, 1994),

based on the modified surface depletion scheme of Horst (1983).

n. Evaluation Studies

* * * * *

Environmental Protection Agency, 1992. Comparison of a Revised

Area Source Algorithm for the Industrial Source Complex Short Term

Model and Wind Tunnel Data. EPA Publication No. EPA-454/R-92-014.

U.S. Environmental Protection Agency, Research Triangle Park, NC.

(NTIS No. PB 93-226751)

Environmental Protection Agency, 1992. Sensitivity Analysis of a

Revised Area Source Algorithm for the Industrial Source Complex

Short Term Model. EPA Publication No. EPA-454/R-92-015. U.S.

Environmental Protection Agency, Research Triangle Park, NC. (NTIS

No. PB 93-226769)

Environmental Protection Agency, 1992. Development and

Evaluation of a Revised Area Source Algorithm for the Industrial

Source Complex Long Term Model. EPA Publication No. EPA-454/R-92-

016. U.S. Environmental Protection Agency, Research Triangle Park,

NC. (NTIS No. PB 93-226777)

Environmental Protection Agency, 1994. Development and Testing

of a Dry Deposition Algorithm (Revised). EPA Publication No. EPA-

454/R-94-015. U.S. Environmental Protection Agency, Research

Triangle Park, NC.

* * * * *

A. Ref Rerences

* * * * *

Horst, T. W., 1983. A Correction to the Gaussian Source-

depletion Model. In Precipitation Scavenging, Dry Deposition and

Resuspension. H. R. Pruppacher, R. G. Semonin, and W. G. N. Slinn,

eds., Elseview, NY.

11. Appendix W to Part 51, Appendix B, is amended by:

a. Adding two entries to the Table of Contents in numerical order;

and

b. Adding sections B.32 and B.33 immediately following section

B.31.

The additions read as follows:

Appendix B to Appendix W of Part 51--Summaries of Alternative Air

Quality Models

Table of Contents

* * * * *

B.32 HGSYSTEM

B.33 SLAB

* * * * *

B.32 HGSYSTEM: Dispersion Models for Ideal Gases and Hydrogen Fluoride

References

Witlox, H.W.M., 1991. HGSYSTEM: dispersion models for ideal

gases and hydrogen fluoride, tutorial and quick-reference guide.

Report TNER.91.007. Thornton Research Centre, Shell Research,

Chester, England. [EGG 1067-1150] (NTIS No. DE 93-000952)

Availability

The PC-DOS version of the HGSYSTEM software (HGSYSTEM: Version

NOV90, Programs for modeling the dispersion of ideal gas and

hydrogen fluoride releases. [EGG 1067-1153]), executable programs

and source code, can be installed from ten 5\1/4\'' diskettes. These

diskettes and all documentation are available as a package from

Energy, Science & Technology Center: (615) 576-1301.

Technical Contacts

Doug N. Blewitt, Amoco Corporation, Environmental Affairs & Safety

Department, Mail Code 4901, 200 East Randolph Drive, Chicago, IL

60601, (312) 856-4099

Howard J. Feldman, American Petroleum Institute, 1220 L Street,

Northwest, Washington, D.C. 20005, (202) 682-8340

Abstract

HGSYSTEM is a software package consisting of mathematical models

for simulating one or more of the consecutive phases between

spillage and far-field dispersion of a non-reactive ideal gas or

hydrogen fluoride (HF). The individual models can be described as

follows: (1) HFSPILL calculates the time-dependent spill rate of HF

liquid or HF vapor from a pressurized vessel; (2) EVAP calculates

the spreading and evaporation of a boiling liquid pool on water or

non-boiling liquid pool on land; (3) HFPLUME calculates the

depressurization to ambient pressure, the jet release and the near-

field dispersion from a pressurized release of HF; (4) PLUME

calculates the depressurization to ambient pressure, the jet release

and the near-field dispersion from a pressurized release of non-

reactive, ideal gases; (5) HEGADAS calculates the steady-state or

time-dependent ground-level heavy-gas dispersion resulting from

either a ground-level pool or a source in a vertical plane; and (6)

PGPLUME simulates passive-gas dispersion downwind of a transition

point based on a simple Pasquill/Gifford similarity model. The

models assume flat, unobstructed terrain. HGSYSTEM can be used to

model steady-state, finite-duration and time-dependent releases. The

models can be run in either the interactive or batch mode.

a. Recommendations for Regulatory Use

HGSYSTEM can be used as a refined model to estimate short-term

ambient concentrations. For toxic chemical releases (non-reactive

chemicals or hydrogen fluoride; 1-hour or less averaging times) the

expected area of exposure to concentrations above specified

threshold values can be determined. For flammable non-reactive gases

it can be used to determine the area in which the cloud may ignite.

b. Input Requirements

1. HFSPILL input data: reservoir data (temperature, pressure,

volume, HF mass, mass-fraction water), pipe-exit diameter and

ambient pressure.

2. EVAP input data: spill rate, liquid properties, and

evaporation rate (boiling pool) or ambient data (non-boiling pool).

3. HFPLUME and PLUME input data: reservoir characteristics,

pollutant parameters, pipe/release data, ambient conditions, surface

roughness and stability class.

4. HEGADAS input data: ambient conditions, pollutant parameters,

pool data or data at transition point, surface roughness, stability

class and averaging time.

5. PGPLUME input data: link data provided by HFPLUME and the

averaging time.

c. Output

1. The HGSYSTEM models contain three post-processor programs

which can be used to extract modeling results for graphical display

by external software packages. GET2COL can be used to extract data

from the model output files. HSPOST can be used to develop

isopleths, extract any 2 parameters for plotting and correct for

finite release duration. HTPOST can be used to produce time history

plots.

2. HFSPILL output data: reservoir mass, spill rate, and other

reservoir variables as a function of time. For HF liquid, HFSPILL

generates link data to HFPLUME for the initial phase of choked

liquid flow (flashing jet), and link data to EVAP for the subsequent

phase of unchoked liquid flow (evaporating liquid pool).

3. EVAP output data: pool dimensions, pool evaporation rate,

pool mass and other pool variables for steady state conditions or as

a function of time. EVAP generates link data to the dispersion model

HEGADAS (pool dimensions and pool evaporation rate).

4. HFPLUME and PLUME output data: plume variables

(concentration, width, centroid height, temperature, velocity, etc.)

as a function of downwind distance.

5. HEGADAS output data: concentration variables and temperature

as a function of downwind distance and (for transient case) time.

6. PGPLUME output data: concentration as a function of downwind

distance, cross-wind distance and height.

d. Type of Model

HGSYSTEM is made up of four types of dispersion models. HFPLUME

and PLUME simulate the near-field dispersion and PGPLUME simulates

the passive-gas dispersion downwind of a transition point. HEGADAS

simulates the ground-level heavy-gas dispersion.

e. Pollutant Types

HGSYSTEM may be used to model non-reactive chemicals or hydrogen

fluoride.

f. Source-Receptor Relationships

HGSYSTEM estimates the expected area of exposure to

concentrations above user-specified threshold values. By imposing

conservation of mass, momentum and energy the concentration,

density, speed and temperature are evaluated as a function of

downwind distance.

g. Plume Behavior

1. HFPLUME and PLUME: (1) are steady-state models assuming a

top-hat profile with cross-section averaged plume variables; and (2)

the momentum equation is taken into account for horizontal ambient

shear, gravity, ground collision, gravity-slumping pressure forces

and ground-surface drag.

2. HEGADAS: assumes the heavy cloud to move with the ambient

wind speed, and adopts a power-law fit of the ambient wind speed for

the velocity profile.

3. PGPLUME: simulates the passive-gas dispersion downwind of a

transition point from HFPLUME or PLUME for steady-state and finite

duration releases.

h. Horizontal Winds

A power law fit of the ambient wind speed is used.

i. Vertical Wind Speed

Not treated.

j. Horizontal Dispersion

1. HFPLUME and PLUME: Plume dilution is caused by air

entrainment resulting from high plume speeds, trailing vortices in

wake of falling plume (before touchdown), ambient turbulence and

density stratification. Plume dispersion is assumed to be steady and

momentum-dominated, and effects of downwind diffusion and wind

meander (averaging time) are not taken into account.

2. HEGADAS: This model adopts a concentration similarity profile

expressed in terms of an unknown center-line ground-level

concentration and unknown vertical/cross-wind dispersion parameters.

These quantities are determined from a number of basic equations

describing gas-mass conservation, air entrainment (empirical law

describing vertical top-entrainment in terms of global Richardson

number), cross-wind gravity spreading (initial gravity spreading

followed by gravity-current collapse) and cross-wind diffusion

(Briggs formula).

3. PGPLUME: It assumes a Gaussian concentration profile in which

the cross-wind and vertical dispersion coefficients are determined

by empirical expressions. All unknown parameters in this profile are

determined by imposing appropriate matching criteria at the

transition point.

k. Vertical Dispersion

See description above.

l. Chemical Transformation

Not treated.

m. Physical Removal

Not treated.

n. Evaluation Studies

1. PLUME has been validated against field data for releases of

liquified propane, and wind tunnel data for buoyant and vertically-

released dense plumes. HFPLUME and PLUME have been validated against

field data for releases of HF (Goldfish experiments) and propane

releases. In addition, the plume rise algorithms have been tested

against Hoot, Meroney, and Peterka, Ooms and Petersen databases.

HEGADAS has been validated against steady and transient releases of

liquid propane and LNG over water (Maplin Sands field data), steady

and finite-duration pressurized releases of HF (Goldfish

experiments; linked with HFPLUME), instantaneous release of Freon

(Thorney Island field data; linked with the box model HEGABOX) and

wind tunnel data for steady, isothermal dispersion.

2. The validation studies are contained in the following

references:

McFarlane, K., Prothero, A., Puttock, J.S., Roberts, P.T. and

Witlox, H.W.M., 1990. Development and validation of atmospheric

dispersion models for ideal gases and hydrogen fluoride, Part I:

Technical Reference Manual. Report TNER.90.015. Thornton Research

Centre, Shell Research, Chester, England. [EGG 1067-1151] (NTIS No.

DE 93-000953)

Witlox, H.W.M., McFarlane, K., Rees, F.J., and Puttock, J.S.,

1990. Development and validation of atmospheric dispersion models

for ideal gases and hydrogen fluoride, Part II: HGSYSTEM Program

User's Manual. Report TNER.90.016. Thornton Research Centre, Shell

Research, Chester, England. [EGG 1067-1152] (NTIS No. DE 93-000954)

B.33 SLAB

Reference

Ermak, D.L., 1990. User's Manual for SLAB: An Atmospheric

Dispersion Model for Denser-than-Air Releases (UCRL-MA-105607),

Lawrence Livermore National Laboratory.

Availability

1. The computer code is available on the Support Center for

Regulatory Air Models Bulletin Board System (Upload/Download Area;

see page B-1), and can also be obtained from:

Energy Science and Technology Center, P.O. Box 1020, Oak Ridge, TN

37830, (615) 576-2606

2. The User's Manual (NTIS No. DE 91-008443) can be obtained

from:

Computer Products, National Technical Information Service, U.S.

Department of Commerce, Springfield, VA 22161, (703) 487-4650

Abstract

The SLAB model is a computer model, PC-based, that simulates the

atmospheric dispersion of denser-than-air releases. The types of

releases treated by the model include a ground-level evaporating

pool, an elevated horizontal jet, a stack or elevated vertical jet

and an instantaneous volume source. All sources except the

evaporating pool may be characterized as aerosols. Only one type of

release can be processed in any individual simulation. Also, the

model simulates only one set of meteorological conditions; therefore

direct application of the model over time periods longer than one or

two hours is not recommended.

a. Recommendations for Use

The SLAB model should be used as a refined model to estimate

spatial and temporal distribution of short-term ambient

concentration (e.g., 1-hour or less averaging times) and the

expected area of exposure to concentrations above specified

threshold values for toxic chemical releases where the release is

suspected to be denser than the ambient air.

b. Input Requirements

1. The SLAB model is executed in the batch mode. Data are input

directly from an external input file. There are 29 input parameters

required to run each simulation. These parameters are divided into 5

categories by the user's guide: source type, source properties,

spill properties, field properties, and meteorological parameters.

The model is not designed to accept real-time meteorological data or

convert units of input values. Chemical property data are not

available within the model and must be input by the user. Some

chemical and physical property data are available in the user's

guide.

2. Source type is chosen as one of the following: evaporating

pool release, horizontal jet release, vertical jet or stack release,

or instantaneous or short duration evaporating pool release.

3. Source property data requirements are physical and chemical

properties (molecular weight, vapor heat capacity at constant

pressure; boiling point; latent heat of vaporization; liquid heat

capacity; liquid density; saturation pressure constants), and

initial liquid mass fraction in the release.

4. Spill properties include: source temperature, emission rate,

source dimensions, instantaneous source mass, release duration, and

elevation above ground level.

5. Required field properties are: desired concentration

averaging time, maximum downwind distance (to stop the calculation),

and four separate heights at which the concentration calculations

are to be made.

6. Meteorological parameter requirements are: ambient

measurement height, ambient wind speed at designated ambient

measurement height, ambient temperature, surface roughness, relative

humidity, atmospheric stability class, and inverse Monin-Obukhov

length (optional, only used as an input parameter when stability

class is unknown).

c. Output

No graphical output is generated by the current version of this

program. The output print file is automatically saved and must be

sent to the appropriate printer by the user after program execution.

Printed output includes in tabular form:

1. Listing of model input data;

2. Instantaneous spatially-averaged cloud parameters--time,

downwind distance, magnitude of peak concentration, cloud dimensions

(including length for puff-type simulations), volume (or mole) and

mass fractions, downwind velocity, vapor mass fraction, density,

temperature, cloud velocity, vapor fraction, water content, gravity

flow velocities, and entrainment velocities;

3. Time-averaged cloud parameters--parameters which may be used

externally to calculate time-averaged concentrations at any location

within the simulation domain (tabulated as functions of downwind

distance);

4. Time-averaged concentration values at plume centerline and at

five off-centerline distances (off-centerline distances are

multiples of the effective cloud half-width, which varies as a

function of downwind distance) at four user-specified heights and at

the height of the plume centerline.

d. Type of Model

As described by Ermak (1989), transport and dispersion are

calculated by solving the conservation equations for mass, species,

energy, and momentum, with the cloud being modeled as either a

steady-state plume, a transient puff, or a combination of both,

depending on the duration of the release. In the steady-state plume

mode, the crosswind-averaged conservation equations are solved and

all variables depend only on the downwind distance. In the transient

puff mode, the volume-averaged conservation equations are solved,

and all variables depend only on the downwind travel time of the

puff center of mass. Time is related to downwind distance by the

height-averaged ambient wind speed. The basic conservation equations

are solved via a numerical integration scheme in space and time.

e. Pollutant Types

Pollutants are assumed to be non-reactive and non-depositing

dense gases or liquid-vapor mixtures (aerosols). Surface heat

transfer and water vapor flux are also included in the model.

f. Source-Receptor Relationships

1. Only one source can be modeled at a time.

2. There is no limitation to the number of receptors; the

downwind receptor distances are internally calculated by the model.

The SLAB calculation is carried out up to the user-specified maximum

downwind distance.

3. The model contains submodels for the source characterization

of evaporating pools, elevated vertical or horizontal jets, and

instantaneous volume sources.

g. Plume Behavior

Plume trajectory and dispersion is based on crosswind-averaged

mass, species, energy, and momentum balance equations. Surrounding

terrain is assumed to be flat and of uniform surface roughness. No

obstacle or building effects are taken into account.

h. Horizontal Winds

A power law approximation of the logarithmic velocity profile

which accounts for stability and surface roughness is used.

i. Vertical Wind Speed

Not treated.

j. Vertical Dispersion

The crosswind dispersion parameters are calculated from formulas

reported by Morgan et al. (1983), which are based on experimental

data from several sources. The formulas account for entrainment due

to atmospheric turbulence, surface friction, thermal convection due

to ground heating, differential motion between the air and the

cloud, and damping due to stable density stratification within the

cloud.

k. Horizontal Dispersion

The horizontal dispersion parameters are calculated from

formulas similar to those described for vertical dispersion, also

from the work of Morgan, et al. (1983).

l. Chemical Transformation

The thermodynamics of the mixing of the dense gas or aerosol

with ambient air (including water vapor) are treated. The

relationship between the vapor and liquid fractions within the cloud

is treated using the local thermodynamic equilibrium approximation.

Reactions of released chemicals with water or ambient air are not

treated.

m. Physical Removal

Not treated.

n. Evaluation Studies

Blewitt, D. N., J. F. Yohn, and D. L. Ermak, 1987. An Evaluation

of SLAB and DEGADIS Heavy Gas Dispersion Models Using the HF Spill

Test Data, Proceedings, AIChE International Conference on Vapor

Cloud Modeling, Boston, MA, November, pp. 56-80.

Ermak, D. L., S.T. Chan, D. L. Morgan, and L. K. Morris, 1982. A

Comparison of Dense Gas Dispersion Model Simulations with Burro

Series LNG Spill Test Results, J. Haz. Matls., 6: 129-160.

Zapert, J. G., R. J. Londergan, and H. Thistle, 1991. Evaluation

of Dense Gas Simulation Models. EPA Publication No. EPA-450/4-90-

018. U.S. Environmental Protection Agency, Research Triangle Park,

NC.

PART 52--APPROVAL AND PROMULGATION OF IMPLEMENTATION PLANS

1. The authority citation for Part 52 continues to read as follows:

Authority: 42 U.S.C. 7401-7671q.

Sec. 52.21 [Amended]

2. In Sec. 52.21, paragraphs (l)(1) and (l)(2) are amended by

revising ``and supplement B (1993)'' to read ``, supplement B (1993)

and supplement C (1994)''.

[FR Doc. 94-28456 Filed 11-25-94; 8:45 am]

BILLING CODE 6560-50-P