UniM States
Enviroiirwiilal Protection
Agency
AERMOD/BLP Development and Testing
-------
-------
EPA-454/P-15-004
July 2015
AERMOD/BLP Development and Testing
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Air Quality Assessment Division
Air Quality Modeling Group
Research Triangle Park, North Carolina
-------
Preface
This document provides the modification and evaluation procedures used to incorporate the
Buoyant Line and Point Source (BLP) model into AERMOD.
-------
Acknowledgements
The incorporation of BLP into AERMOD was performed by Jim Paumier of AMEC
Foster/Wheeler Environment & Infrastructure Inc., Part 4, Task Order No. EP-G15-D-00208,
contract No. GS-35F-0851R with Chris Owen as Part 4 Task Order Project Officer.
in
-------
Contents
Preface ii
Acknowledgements iii
Figures v
Tables vi
1. Introduction 1
2. Source Code Modifications 1
3. Testing 2
3.1 Initial Tests 2
3.2 Further Testing of AERMOD 17
4. Differences between AERMOD/BLP and BLP 24
5. Additional Changes to AERMOD/BLP 24
5.1 Background Concentration 24
5.2 Event Processing 24
5.3 BLP Source Group 24
6. Recommendations 24
APPENDIX 26
1. AERMOD Keywords for Buoyant Line Sources 26
IV
-------
Figures
Figure 1. Buoyant Line Source and Receptor Locations.
-------
Tables
Table 1. Beginning and Ending Coordinates of the Buoyant Lines 2
Table 2. Bouyant Line Source Average Parameters 2
TableS. Top 50 1-hour Concentration Estimates (|ig/m3): 1-day Scenario 4
Table 4. Top 50 3-hour Concentration Estimates (|ig/m3): 1-day Scenario 5
Table 5. Top 50 24-hour Concentration Estimates (|ig/m3): 1-day Scenario 6
Table 6. Top 50 1-hour Concentration Estimates (|ig/m3): 4-day Scenario 7
Table?. Top 50 3-hour Concentration Estimates (|ig/m3): 4-day Scenario 8
Table 8. Top 50 24-hour Concentration Estimates (|ig/m3): 4-day Scenario 9
Table 9. Top 50 1-hour Concentration Estimates (|ig/m3): 1-month Scenario 10
Table 10. Top 50 3-hour Concentration Estimates (|ig/m3): 1-month Scenario 11
Table 11. Top 50 24-hour Concentration Estimates (|ig/m3): 1-month Scenario 12
Table 12. Top 50 1 -hour Concentration Estimates (|ig/m3): 1-year Scenario 13
Table 13. Top 50 3-hour Concentration Estimates (|ig/m3): 1-year Scenario 15
Table 14. Top 50 24-hour Concentration Estimates (|ig/m3): 1-year Scenario 16
Table 15. Additional Tests - 1-year Scenarios 17
Table 16. Top 5 Concentration Estimates (|ig/m3), Test A (Gridded Network) 18
Table 17. Top 5 Concentration Estimates (|ig/m3), Test B (Includes Two Point Sources) 18
Table 18. Concentration Estimates (|ig/m3), Test C 18
Table 19. Concentration Estimates (|ig/m3), Test C 19
Table 20. Concentration Estimates (|ig/m3) for TestD 19
Table 21. Concentration Estimates (|ig/m3) for Test E 20
Table 22. Concentration Estimates (|ig/m3) for Test F (24-hr PM-2.5) 21
Table 23. Concentration Estimates (|ig/m3) for Test G (1-hr SO2) 22
Table 24. Concentration Estimates (|ig/m3) for Test H - 23
Table 25. Concentration Estimates (|ig/m3) for Test I - 23
Table 26. Concentration Estimates (|ig/m3) for Test J - 23
VI
-------
1. Introduction
The AERMOD dispersion model was modified to incorporate the buoyant line algorithms of the
Buoyant Line and Point source (BLP) dispersion model. For purposes of this discussion, we will
refer to this draft of AERMOD as AERMOD/BLP.
2. Source Code Modifications
The buoyant line algorithms in BLP (version 99176) were first incorporated into AERMOD
version 14134 and later into version 15181. For this phase of development, the buoyant line
algorithms were ported unchanged from BLP to AERMOD. To keep the routines easily
identifiable in AERMOD/BLP, "BL_" was added to the beginning of the subroutine and function
names. Within those subroutines and functions, variable names were added as needed or
changed to match the AERMOD variable names, such as the reference height wind speed. BLP
was written decades ago and used GOTO statements and other obsolescent structures to control
conditional programming. Those structures were replaced with IF..THEN..ELSE..ENDIF and
SELECT.. CASE structures. Other parts of the original BLP code were updated to reflect more
recent versions of the Fortran language such as named DO loop structures, double precision, and
allocatable arrays. Some of the GOTO statements were retained since deconstructing those
structures would require a much greater effort. A new module, buoyant_line, was added to
centralize the global variables associated with BLP. Variables and constants in the module and
routines were defined as double precision.
BLP version 99176 is very limited in the number of sources (up to 10 lines in a single buoyant
line source and 51 point sources) and receptors (up to 100) that can be processed in a single
model run. As part of the porting process, array size limitations imposed by BLP were changed
to allocatable arrays in AERMOD/BLP.
Additional keywords and parameters were added to process buoyant lines. These additions can
be found in the appendix attached to this memo.
The following files were modified in the process of incorporating BLP into AERMOD:
aermod.f, calcl.f, soset.f, reset.f, setup.f, modules.f, metext.f, and evcalc.f (minimal changes; not
all necessary changes were made for event processing). For testing AERMOD/BLP, 64-bit
executables were generated.
-------
3. Testing
i.l Initial Tests
Modifications to BLP version 99176were required to compare concentration estimates between
BLP and AERMOD/BLP. These included reading and processing command line arguments
(complying with the Intel Fortran requirements) and increasing the number of receptors. Since
BLP was originally limited to 100 receptors, the BLP code was modified to allow up to 1000
receptors.
During the initial development and debugging process, one-day and four-day meteorological data
sets were used to evaluate whether or not BLP and AERMOD/BLP were giving the same results.
A buoyant line source consisting of three lines parallel to each other was defined, with the lines
oriented northwest-to-southeast at a 45-degree angle as shown in Figure 1. The endpoints of the
three lines that make up the line source are in Table 1.
Table 1. Beginning and Ending Coordinates of the Buoyant Lines
Source ID
BLINE1
BLINE2
BLINE3
Xbegin
300
300
300
Ybegin
800
1000
1200
Xend
900
900
900
Yend
200
400
600
BLP requires the average value for several parameters in its calculations. These values are
defined in the BLP User's Manual. The average parameters associated with the buoyant line
source used in this testing are taken from one of the example cases packaged with BLP version
99176 and are presented in Table 2.
Table 2. Bouyant Line Source Average Parameters
Bldg Length
(m)
848.0
Bldg Hgt
(m)
20.0
Bldg Width
(m)
100.0
Line Width
(m)
2.0
Bldg Separation
(m)
37.5
Buoyancy Param.
(m4/s3)
300.0
The release height for the lines was 20.0 meters (m) and the elevation for the lines was 0.0 m.
A 3 l-by-31 discrete receptor grid, with the origin at (0,0), and 50-meter node spacing was
defined for the analyses, for a total of 961 receptors. The receptors are shown in Figure 1 as
light grey dots.
-------
For these initial tests, each buoyant line was modeled with an emission rate of 1.0
grams/second/square meter (g/s/m2). The averaging times modeled were 1-, 3-, and 24-hr. Some
test scenarios were modeled with 8-hr averaging periods, but the results are not reported here.
Flat terrain was assumed for all model runs.
1 3ftfll
1100
HIV1
fton
criri
3on
V
%
V
Buoyant Line Source & Receptors
N
%
1 ^
•
V
V
S1
' ! : : :
\
•
s
V
S'
N
S1
\
?!'::•
: ! r : :
— • — BLinel
-^-BLine 3
• Receptors
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500
Figure 1. Buoyant Line Source and Receptor Locations
To run the AERMOD meteorology (or as near as possible) in BLP, which requires
PCRAMMET-like format, AERMOD/BLP was modified to output the necessary parameters for
BLP to read. The parameters are: wind speed (meters/second, m/s), flow vector (although it is
referred to as wind direction in the BLP manuals) (degrees), ambient temperature (kelvin),
Pasquill-Gifford (P-G) stability class, and rural and urban mixing heights (m). The wind
direction, an input to AERMOD, was converted to flow vector for BLP. Since the P-G stability
class is not an input to AERMOD, the subroutine LTOPG found in AERMOD was used to
generate and output the appropriate stability class for the hour. The rural and urban mixing
heights were set to the same value and are the larger of the convective and mechanical mixing
heights that are input to AERMOD. For the one-day and four-day periods, all meteorological
parameters were present (i.e., no missing data).
-------
The results from BLP and AERMOD/BLP were compared and significant differences were noted
with the one- and four-day runs. The code was reviewed and a few bugs were corrected, but
differences remained. The difference was traced to the use of double precision in
AERMOD/BLP (real variables are declared double precision). BLP was recompiled with the
real-size:64 option (the default is real-size:32) and the results from both models (AERMOD/BLP
and BLP) for the two short meteorological data sets were nearly identical. Table 1 shows the top
50 1-hour concentration impacts for the one-day scenario with the discrete receptor grid. Table 2
shows the 3-hour impacts, and Table 3 shows the 24-hour impacts. The 'NaN' in Table 3 for
BLP indicates 'not a number'; the cause of this is unclear or why it would appear in the top 50.
Table 3. Top 50 1-hour Concentration Estimates (|ig/m3): 1-day Scenario
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
BLP
1493.2
843.5
517.9
211.7
193.0
175.2
108.1
97.9
97.5
96.6
95.0
94.1
91.7
91.0
90.3
90.3
90.0
89.8
89.6
89.1
88.4
88.3
88.2
88.1
87.9
AERMOD
1345.38
843.51
517.93
211.68
181.35
175.19
108.07
97.87
97.50
96.57
95.03
93.87
91.72
91.02
90.85
90.31
90.31
89.98
89.07
88.42
88.31
88.18
88.14
87.94
87.68
Rank
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
BLP
87.7
87.4
87.0
86.9
86.4
86.2
86.1
85.9
85.5
85.3
85.2
84.7
84.5
84.4
84.2
83.8
83.6
83.1
83.0
82.2
82.2
81.7
81.3
81.3
81.1
AERMOD
87.39
87.01
86.94
86.40
86.20
86.10
85.90
85.45
85.31
85.18
84.70
84.48
84.41
84.21
83.78
83.62
83.14
83.03
82.24
82.15
81.66
81.34
81.30
81.14
81.07
-------
Table 4. Top 50 3-hour Concentration Estimates (|ig/m ): 1-day Scenario
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
BLP
498.0
281.3
203.0
93.3
89.6
71.5
55.2
54.4
54.0
53.3
52.6
51.3
51.1
50.9
46.2
45.5
44.3
43.0
42.8
42.2
40.2
40.1
39.8
39.8
39.5
AERMOD
448.63
281.34
199.91
75.40
75.33
71.53
54.35
53.97
52.64
51.33
51.31
50.93
46.15
45.53
45.01
44.31
43.21
42.97
42.82
42.20
40.52
40.22
40.07
39.79
39.52
Rank
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
BLP
39.4
39.0
37.6
37.6
37.3
37.2
37.0
36.4
36.3
36.3
35.4
35.1
35.0
33.7
33.6
33.2
33.0
33.0
32.7
32.2
31.9
31.8
31.7
31.2
31.2
AERMOD
39.41
39.01
38.06
37.59
37.26
37.20
37.18
37.04
36.30
36.30
35.75
35.05
33.66
33.56
33.00
32.98
32.71
32.71
32.17
32.16
31.75
31.59
31.50
31.24
31.23
-------
Table 5. Top 50 24-hour Concentration Estimates (|ig/m ): 1-day Scenario
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
BLP
110.0
54.5
26.9
21.7
20.5
19.6
18.0
17.0
15.7
14.1
13.8
12.8
12.7
12.7
12.3
12.3
11.9
11.7
11.7
11.7
11.6
11.4
11.3
11.1
11.1
AERMOD
102.23
52.89
25.14
22.32
20.76
20.64
18.60
17.07
16.66
15.79
14.87
14.53
13.51
13.20
13.19
12.81
12.59
12.59
12.56
12.40
11.96
11.95
11.80
11.73
11.61
Rank
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
BLP
11.1
11.1
11.1
10.9
10.6
10.6
10.5
10.2
10.2
10.2
10.2
10.1
10.1
10.1
10.0
10.0
10.0
9.9
9.9
9.8
NaN
9.8
9.7
9.7
9.6
AERMOD
11.58
11.49
11.38
11.38
11.28
11.27
11.23
11.19
11.14
11.00
10.97
10.96
10.91
10.68
10.67
10.60
10.59
10.58
10.49
10.42
10.40
10.30
10.18
10.16
10.02
-------
For the 4-day scenario with the discrete receptor network, Table 4 shows the top 50 1-hour
concentration impacts for the one-day scenario with a discrete receptor grid. Table 5 shows the
3-hour impacts, and Table 6 shows the 24-hour impacts.
Table 6. Top 50 1-hour Concentration Estimates (|ig/m3): 4-day Scenario
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
BLP
1493.2
843.5
517.9
265.5
244.3
211.7
206.5
193.0
191.2
189.2
188.7
184.3
180.2
178.8
178.4
176.2
175.2
171.5
170.5
169.9
168.3
167.8
166.3
163.9
163.5
AERMOD
1345.38
843.51
517.93
265.46
226.88
211.68
191.17
189.72
189.16
188.66
184.32
181.35
180.25
178.78
178.39
176.16
175.19
171.54
170.48
169.94
168.32
167.82
166.27
163.86
163.47
Rank
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
BLP
157.9
157.6
156.0
155.5
154.8
154.6
154.2
152.7
150.7
149.1
148.7
146.7
146.5
145.8
145.2
142.0
141.9
140.3
138.1
137.7
135.5
135.2
134.6
133.8
133.8
AERMOD
157.92
157.55
156.02
155.47
154.76
154.63
154.19
152.70
150.71
149.13
148.66
146.67
146.53
145.80
145.19
142.44
141.87
140.26
138.07
137.66
135.50
135.17
134.64
133.81
133.76
-------
Table 7. Top 50 3-hour Concentration Estimates (|ig/m ): 4-day Scenario
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
BLP
498.0
281.3
202.5
145.9
145.4
129.2
100.1
97.7
97.2
95.1
91.9
89.4
88.8
88.2
80.3
78.8
77.5
77.3
77.2
75.4
75.1
75.0
73.8
73.8
73.5
AERMOD
448.63
281.34
199.91
145.88
145.43
129.17
100.05
97.69
91.92
89.95
89.37
88.21
80.40
79.88
78.80
77.46
77.28
77.18
76.10
75.40
75.33
75.03
73.83
73.75
73.52
Rank
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
BLP
72.2
71.5
69.4
68.8
68.2
68.0
67.0
66.6
63.7
63.3
63.1
63.1
62.8
62.5
62.3
62.2
62.1
62.0
62.0
61.9
61.7
61.5
61.4
61.1
61.1
AERMOD
72.23
71.53
69.39
68.81
68.18
67.02
66.63
64.45
63.72
63.26
63.09
63.05
62.78
62.50
62.32
62.17
62.13
62.04
62.01
61.86
61.73
61.51
61.39
61.14
61.07
-------
Table 8. Top 50 24-hour Concentration Estimates (|ig/m ): 4-day Scenario
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
BLP
109.6
52.9
38.6
33.0
32.9
32.6
31.6
30.6
30.3
30.2
29.8
29.4
29.4
28.5
28.1
28.0
28.0
27.9
27.7
27.7
27.7
26.7
26.7
26.1
26.0
AERMOD
102.23
52.89
38.60
32.99
32.87
32.63
31.39
30.60
30.40
30.30
30.16
29.65
29.48
29.37
28.19
28.10
27.92
27.88
27.82
27.77
27.68
26.95
26.06
26.01
25.83
Rank
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
BLP
25.8
25.3
25.2
25.2
24.9
24.7
24.6
24.5
24.5
24.3
23.8
23.4
23.1
23.0
22.7
22.7
22.4
22.4
22.4
22.3
22.1
22.0
21.9
21.7
21.7
AERMOD
25.34
25.32
25.14
25.00
24.92
24.83
24.62
24.61
24.50
24.26
23.87
23.60
23.09
23.04
22.73
22.66
22.56
22.43
22.32
22.28
22.05
22.03
22.00
21.70
21.68
-------
A 1-month scenario (January) was run using meteorological data processed for the Lovett
evaluation. To get AERMOD to accept the data, the version number in the header of the .SFC
file was changed to 14134. For the discrete receptor network, Table 7 shows the top 50 1-hour
concentration impacts for the one-day scenario with a discrete receptor grid. Table 8 shows the
3-hour impacts, and Table 9 shows the 24-hour impacts. For the 1-hour estimates, differences
are apparent immediately after the highest value, but are more in line with each other further
down.
Table 9. Top 50 1-hour Concentration Estimates (|ig/m3): 1-month Scenario
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
BLP
1208.3
841.6
787.3
763.2
755.3
676.9
674.6
654.5
604.5
548.5
545.8
521.7
508.1
500.0
499.9
497.1
494.1
494.1
485.1
484.7
475.8
475.1
471.7
465.5
461.5
AERMOD
1208.26
1098.06
1038.66
841.62
787.32
763.23
755.30
676.90
674.59
654.49
604.45
548.53
545.84
521.73
499.99
499.93
497.10
494.11
494.07
471.74
461.54
454.19
454.09
453.48
451.39
AERMOD
(UREF=50)
1208.26
1098.06
1038.66
841.62
787.32
763.23
755.30
676.90
674.59
654.49
604.45
548.53
545.84
521.73
508.09
499.99
499.93
497.10
494.11
494.07
485.13
484.74
475.76
475.14
471.74
Rank
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
BLP
454.2
454.1
453.5
451.4
451.0
447.5
447.4
445.0
443.0
441.7
432.8
429.4
425.4
421.4
419.0
411.4
410.7
407.2
404.8
403.9
403.0
401.0
400.6
399.2
399.0
AERMOD
447.49
447.39
443.00
429.36
425.40
421.43
418.98
411.38
404.82
403.88
402.95
400.58
399.16
399.02
398.14
397.58
397.13
386.52
384.66
383.36
383.24
378.09
374.51
371.74
371.59
AERMOD
(UREF=50)
465.48
461.54
454.19
454.09
453.48
451.39
450.96
447.49
447.39
444.99
443.00
441.69
432.77
429.36
425.40
421.43
418.98
411.38
410.67
407.23
404.82
403.88
402.95
401.00
400.58
10
-------
Table 10. Top 50 3-hour Concentration Estimates (|ig/m3): 1-month Scenario
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
BLP
440.4
420.8
402.8
390.3
378.3
370.5
350.1
331.4
327.1
326.7
323.3
280.4
275.0
247.8
237.4
234.2
233.6
229.0
219.6
218.8
217.9
213.0
199.0
196.1
195.9
AERMOD
440.37
420.79
402.75
390.34
378.27
370.46
367.97
366.02
350.14
331.44
327.06
326.73
323.27
280.36
275.02
247.83
237.36
234.17
229.00
219.57
218.82
217.91
212.95
196.06
195.88
AERMOD
(UREF=50)
440.37
420.79
402.75
390.34
378.27
370.46
367.97
366.02
350.14
331.44
327.06
326.73
323.27
280.36
275.02
247.83
237.36
234.17
233.61
229.00
219.57
218.82
217.91
212.95
199.03
Rank
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
BLP
195.8
194.9
193.4
192.8
192.8
191.1
189.9
189.9
188.5
186.7
186.0
184.2
183.7
183.3
180.3
179.9
179.6
179.2
178.0
177.5
174.0
171.4
171.2
171.0
170.3
AERMOD
195.75
194.91
193.38
192.80
192.77
191.09
189.92
188.48
186.66
185.98
184.15
183.70
183.34
180.25
179.92
179.24
178.04
177.52
174.03
171.62
171.37
170.34
170.33
169.57
168.73
AERMOD
(UREF=50
196.06
195.88
195.75
194.91
193.38
192.80
192.77
191.09
189.92
189.90
188.48
186.66
185.98
184.15
183.70
183.34
180.25
179.92
179.56
179.24
178.04
177.52
174.03
171.37
171.22
11
-------
Table 11. Top 50 24-hour Concentration Estimates (|ig/m3): 1-month Scenario
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
BLP
84.7
79.9
76.3
73.6
70.8
70.5
70.1
68.9
68.3
67.2
67.1
66.4
65.9
65.9
65.8
65.8
64.7
64.7
64.2
63.7
63.6
63.2
62.9
62.9
62.5
AERMOD
84.53
76.75
72.75
70.79
63.65
61.60
61.02
61.02
60.73
60.60
60.39
59.07
58.64
58.53
58.12
57.45
57.44
56.31
56.13
55.63
55.03
54.95
54.56
54.05
53.26
AERMOD
(UREF=50)
84.67
79.74
76.30
73.63
70.90
70.81
70.54
70.08
68.93
68.33
68.01
67.19
66.39
65.92
65.86
65.78
65.75
64.71
64.68
64.18
63.65
63.54
63.16
62.94
62.48
Rank
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
BLP
62.3
61.9
61.7
61.6
61.5
61.3
61.3
61.2
60.9
60.8
60.3
60.1
60.0
59.9
58.9
58.7
58.3
57.0
56.9
56.7
56.7
56.5
56.4
56.3
56.1
AERMOD
52.66
52.62
52.62
52.56
52.26
52.25
52.14
51.53
51.49
51.48
51.30
50.99
50.91
50.60
50.39
50.30
50.24
50.22
50.16
50.01
49.90
49.54
49.25
49.20
48.42
AERMOD
(UREF=50
62.36
62.33
61.86
61.70
61.64
61.34
61.29
61.15
60.91
60.76
60.25
60.14
60.06
59.97
59.91
58.91
58.71
58.64
58.31
57.04
56.86
56.67
56.66
56.47
56.41
It is important to note that the reference measurement height for winds in the Lovett-based
meteorology can vary from 10 m to 50m and even as high as 100 m, whereas the 'measurement
height' in BLP is a user-defined static value. For these tests it was set at 50 m. To eliminate this
height as a source of possible differences, the reference height in AERMOD/BLP (UREFHT)
was set to 50 m for all hours. As seen in Table 11, it makes some difference, ranging from 1-2%
to as much as 15%.
Once most of the 'bugs' had been worked out using the short meteorological data sets, testing was
extended to using a full year of meteorological data. The Lovett evaluation database profile and
surface files were used: LOVETT.PFL and LOVETT.SFC. As with the one- and four-day cases,
AERMOD/BLP was run to output the year of meteorological data for BLP.
BLP requires a complete period of meteorological record but AERMOD will skip the hour if
critical parameters are missing. There were approximately 250 hours skipped in the output
meteorological file generated by AERMOD/BLP for the year of meteorological data. Periods
12
-------
ranged from one hour to several consecutive hours. The resulting output file for BLP was edited
manually to fill those hours. Interpolation, persistence, and professional judgment were used to
fill the parameters, depending on the variable and, in part, on the length of the missing data
period.
A single case was developed to test AERMOD/BLP initially and identify any further bugs with
the three lines and receptors shown in the figure above. Tables 12-14 show the top 50 1-hr, 3-
hr, and 24-hr impacts, respectively.
Table 12. Top 50 1-hour Concentration Estimates (|ig/m3): 1-year Scenario
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
BLP
2528.8
2525.6
2364.9
2220.2
1825.4
1821.3
1602.4
1501.9
1488.1
1332.6
1330.8
1291.5
1208.3
1171.8
1167.5
1153.9
1133.0
1131.1
1082.2
1079.9
1078.1
1066.2
1046.9
1013.9
1013.0
AERMOD
146052.19
2528.82
2525.60
2434.01
2364.93
2220.21
2208.78
1825.36
1821.31
1761.65
1602.43
1548.45
1501.89
1488.07
1332.55
1330.77
1291.44
1208.26
1171.84
1167.48
1153.89
1132.99
1131.08
1098.06
1079.88
BLP
(real-
size^)
146002.9
6802.1
3010.3
3006.0
2713.6
2517.8
2472.9
2434.0
2208.8
1827.7
1825.4
1821.3
1761.6
1626.7
1548.4
1501.9
1488.1
1332.6
1330.8
1291.4
1282.3
1275.1
1264.0
1242.3
1240.1
Rank
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
BLP
966.4
940.5
919.8
895.3
793.3
766.7
751.6
730.2
718.8
696.3
691.3
674.2
671.0
668.4
663.0
662.8
660.9
660.1
650.8
650.1
648.0
644.3
637.3
637.3
636.4
AERMOD
1078.10
1071.29
1066.18
1065.83
1054.35
1046.85
1038.66
1013.93
1012.96
966.38
940.46
939.13
919.75
895.25
874.41
841.62
831.39
793.28
791.68
790.26
787.32
766.66
763.23
761.52
755.30
BLP
(real-
size^)
1208.3
1185.6
1183.6
1174.5
1149.7
1113.7
1112.6
1098.1
1071.3
1066.2
1065.8
1054.3
1038.7
1012.2
999.8
999.8
982.7
966.3
819.5
1.3
NaN
919.8
874.4
847.5
838.8
With the full year of data the 1-hour concentration estimates were found to be vastly different:
146,052 |ig/m3 for AERMOD/BLP and 2,528 |ig/m3 for BLP. The remaining impacts were two
orders of magnitude smaller, with the concentration estimates between AERMOD/BLP and BLP
13
-------
somewhat similar, although there are ranks where the AERMOD/BLP is 30-40 percent (%)
greater than BLP. The large hourly value also affects the maximum 3-hour and 24-hour results.
AERMOD/BLP also was run with the wind reference height set a constant 50 meters (as is the
case in BLP). There is little difference between the two AERMOD/BLP model runs, with most
differences 2% or less. These results are not shown here.
On further investigation, the highest AERMOD/BLP impact was at receptor at x=450, y=1050.
That receptor is on one of the lines (#3, the uppermost line in Figure 1) and the wind direction
associated with the date of the highest value is 315° or long the line. This may partly explain the
extreme value (a receptor right on the line). However, several other receptors appear to fall
directly on the lines and there are nearly 40 cases where the winds direction is 315° as well, yet
the impacts are much smaller. Why this single hour stands out has not been investigated.
Why this single hour at one receptor yielded such a large impact while similar receptors (i.e., on
a line) and wind directions did not has not been investigated in detail. However, BLP was
recompiled with real-case:32 (32-bit precision) and, unexpectedly, the 32-bit version of BLP
yielded an estimate of the same magnitude as AERMOD/BLP (Table 12). The remaining BLP
maximum values are all greater than the corresponding AERMOD/BLP values to the 44th
maximum value. The 45th (1.3) and 46th (NaN) values cannot be explained at this time, and is
similar to the NaN seen in Table 5. Nor is it clear why the 47th-50th values are about the same
magnitude as the 44th value. This issue may require further exploration, although since it is an
issue with BLP, it may not be important for AERMOD/BLP.
14
-------
Table 13. Top 50 3-hour Concentration Estimates (|ig/m ): 1-year Scenario
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
BLP
917.8
848.8
763.7
652.8
614.4
545.4
543.1
507.0
472.2
455.0
440.4
435.5
434.8
427.5
425.2
400.9
389.9
384.6
380.1
369.3
369.0
362.8
352.6
350.4
342.6
AERMOD
48710.30
917.78
849.38
815.06
807.64
781.35
763.73
652.83
632.31
614.38
578.72
543.10
507.04
504.81
472.22
454.98
440.37
440.29
435.54
434.75
425.18
420.79
402.75
400.92
390.61
Rank
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
BLP
340.7
315.2
313.9
289.9
282.1
281.3
280.4
266.3
264.2
264.2
252.6
251.8
248.5
245.1
243.4
242.6
236.6
235.5
231.6
229.9
227.1
222.7
222.4
217.0
216.2
AERMOD
390.34
389.85
384.60
380.06
378.27
372.41
370.46
369.49
369.30
368.99
367.97
366.02
362.82
357.25
355.91
350.39
350.14
347.01
342.61
340.66
334.85
333.33
331.44
328.37
327.06
15
-------
Table 14. Top 50 24-hour Concentration Estimates (|ig/m ): 1-year Scenario
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
BLP
138.7
131.0
126.2
103.6
101.3
100.8
93.3
90.8
90.1
88.6
88.3
83.2
82.9
82.3
79.1
78.2
76.6
74.9
74.4
74.4
74.3
71.6
71.4
70.9
69.7
AERMOD
6094.37
142.76
131.84*
131.19
129.66
125.95
122.70
121.06
119.73
109.95*
109.74
106.66
104.74
103.60*
98.31
97.62
95.75
95.56
93.79
93.34
92.73
89.54
89.09
88.60
86.93
Rank
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
BLP
69.5
69.3
64.7
64.2
63.2
63.2
63.1
62.3
61.5
61.3
61.2
59.8
59.1
58.4
56.3
56.3
55.7
54.1
53.9
53.5
52.2
52.2
51.9
51.8
51.6
AERMOD
84.53
84.14
83.88
82.44
82.14
80.33
80.10
78.89
78.86
78.22*
78.06
77.90
77.52
77.27
76.75
76.68
76.45
75.68
74.64
74.14
73.92
73.78
73.54
72.90
72.75
missing data processing
At this juncture, it is important to note that the original BLP model removes receptors that are
within the rectangle defined by the maximum extents that enclose the multiple lines that define a
buoyant line source (the three lines in this case). In Figure 1 this would result in about 275
receptors being removed. To be able to compare AERMOD/BLP to BLP, that part of the BLP
code (only about 3 lines) was commented out, otherwise it would be impossible to compare and
debug AERMOD/BLP with BLP. This removal of receptors likely would result in the
elimination of the large AERMOD/BLP impact since it is exactly on the line. Recall that it is up
to the user to eliminate receptors inside a facility boundary in AERMOD; as such, the receptors
within the rectangle defined by the buoyant line source would likely be removed during this
process and this large discrepancy never an issue.
16
-------
3.2 Further Testing of AERMOD
Several additional one-year test cases were created. The buoyant line source always consisted of
three lines. In those test cases that indicate two point sources, the sources were located slightly
south of the center of the buoyant line source. These additional cases are described here.
Table 15. Additional Tests - 1-year Scenarios
Test
A
B
C
D
E
F
G
H
I
J
Scenario
Buoyant
Buoyant
line source and gridded receptor network
line source and two point sources
Buoyant line source, using an hourly emissions file: two scenarios - hourly emissions for a
single line and hourly emissions for all three lines
Buoyant
Buoyant
the lines
Buoyant
Buoyant
Buoyant
Buoyant
Buoyant
line source and two point sources, using HROFDY emission factors for the lines
line source and two point sources, using three different emission factors for each of
(SEASON, HROFDY, MONTH)
line source and two point sources modeling 24-hr PM2.5 and using MAXDCONT
line source and two point sources modeling 1-hr SO2 and using MAXDCONT
line source and two point sources with the lines oriented ENE to WSW
line source and two point sources with the lines oriented vertically
line source and two point sources with the lines oriented horizontally
Since most of the testing was to exercise AERMOD options, and similar options are not
available in BLP, model-to-model comparisons are not possible, except for Tests A and B. For a
few of the tests, variations on input parameters were modeled to see if the results were as
expected. For example, the emission rate in the hourly emission file was doubled or tripled to
see if the resulting impacts doubled or tripled. All results appear to be in line with what is
expected.
With the exception of the highest value, the remaining impacts in Tables 16 and 17 are
comparable between AERMOD/BLP and BLP for all three averaging periods.
17
-------
Table 16. Top 5 Concentration Estimates (|ig/m3), Test A (Gridded Network)
Rank
1
2
3
4
5
1-hour
BLP
3010.3
3006.0
2713.6
2517.8
1827.7
AERMOD
146052.19
2528.82
2525.60
2434.01
2364.93
3-hour
BLP
1083.3
1010.3
865.8
652.8
619.4
AERMOD
48710.30
917.78
849.38
815.06
807.64
24-hour
BLP
157.2
154.2
148.7
105.2
103.6
AERMOD
6094.37
142.76
131.84
131.19
129.66
Table 17. Top 5 Concentration Estimates (|ig/m3), Test B (Includes Two Point Sources)
Rank
1
2
3
4
5
1-hour
BLP
2528.8
2525.6
2364.9
2220.2
1825.4
AERMOD
146052.19
2528.82
2525.60
2434.01
2364.93
3-hour
BLP
917.8
848.8
763.7
652.8
614.4
AERMOD
48710.30
917.78
849.39
815.06
807.67
24-hour
BLP
138.7
131.0
126.2
103.6
101.3
AERMOD
6094.38
142.81
132.32
131.27
129.76
The remaining tests were only performed with AERMOD/BLP since BLP does not have all the
processing options that AERMOD employs. Tables 18-25 correspond to Test C-J. The
maximum 5 values are shown for the 1-hour, 3-hour, and 24-hour impacts. For the annual
average, the high-first-high through high-fifth-high are shown.
Table 18 represents hourly emissions for all three of the buoyant lines, whereas Table 19 show
the results when Line 3 is the only contributor to the estimates (lines 1 and 2 are not included in
the hourly emission file and the emission rate is set to 0.0 g/s in the control file). The only effect
is seen mostly for the annual averaging period and somewhat in the 24-hr averaging period.
Table 18. Concentration Estimates (|ig/m3), Test C
(Hourly Emissions File with All Three Lines Contributing)
Rank
1
2
3
4
5
1-hour
146052.19
2528.82
2525.60
2434.01
2364.93
3-hour
48710.30
917.78
849.38
815.06
807.64
24-hour
6094.37
142.76
131.84
131.19
129.66
Annual
47.58
27.84
26.99
24.88
22.45
18
-------
Table 19. Concentration Estimates (|ig/m3), Test C
(Hourly Emissions File with Only Line 3 Contributing)
Rank
1
2
3
4
5
1-hour
146052.19
2528.82
2525.60
2434.01
2364.93
3-hour
48710.30
917.78
848.13
815.06
800.67
24-hour
6094.37
131.83
130.61
130.43
129.34
Annual
42.39
15.47
15.24
14.42
13.50
Table 20 shows the impacts when the HROFDY keyword is used to control the emissions. The
following factors were used.
EMISFACT BLINE1 HROFDY 5*0.0 0.0 18*0.0
EMISFACT BLINE2 HROFDY 5*0.0 0.0 18*0.0
EMISFACT BLINE3 HROFDY 5*0.0 0.5 18*0.0
The only contributor to the concentration estimates is BLINE3 and only for 0600. This hour
corresponds to the hour when the extremely large impacts were seen on October 3. Since the
emission factor is 0.5, the resulting hourly rate should be half of the largest value seen in the
tables above. This is in fact the case: 146052/2 = 73026. One would expect slight differences in
the 3-hour and 24-hour values since there are no contributions from the other hours to these
averaging periods.
Table 20. Concentration Estimates (|ig/m3) for Test D
Rank
1
2
3
4
5
1-hour
73026.09
523.42
468.59
465.44
465.26
3-hour
24342.03
297.76
294.39
294.32
292.62
24-hour
3042.76
72.29
72.22
71.55
71.44
Annual
9.41
5.26
5.16
5.08
5.00
Table 21 shows the impacts when the three different emission factor keywords are used to
control the emissions. The following factors were used. In this scenario, the SEASON emission
factor was used for line 3 (BLINE3) and was set to 0.0 for the fall (October-December). The
highest impacts seen in the tables above (e.g., Table 18) are the result of the extreme
concentration estimate on October 3 at 0600. With the emission factor set to 0.0, this hour's
impact does not factor in to the results in Table 21.
EMISFACT BLINE3 SEASON 1.0 1.0 1.0 0.0
EMISFACT BLINE2 HROFDY 5*0.0 1.0 18*0.0
EMISFACT BLINE1 MONTH 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0
19
-------
Table 21. Concentration Estimates (|jg/m3) for Test E
Rank
1
2
3
4
5
1-hour
2528.82
2525.60
2434.01
2220.21
1761.65
3-hour
917.78
848.14
815.06
763.73
632.31
24-hour
3042.76
72.29
72.22
71.55
71.44
Annual
9.41
5.26
5.16
5.08
5.00
Table 22 shows the results for the 24-hr PM-2.5 impacts with the three buoyant lines and the two
point sources modeled and the MAXDCONT output option specified for source groups ALL,
STACKS (both stacks), and BLINES (three lines). The emission rate was 1.0 g/s/m2 for all three
lines, i.e. no emission factors were applied.
20
-------
Table 22. Concentration Estimates (u,g/m3) for Test F (24-hr PM-2.5)
Group
ALL
STACKS
LINES
Rank
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
Average
Concentration
6094.38
126.54
110.07
104.95
103.66
98.67
97.67
93.06
90.05
87.16
72.29
72.22
71.55
71.43
71.06
70.69
70.64
70.47
69.51
69.51
6094.37
125.95
109.74
104.74
103.60
98.31
97.62
92.73
89.54
86.93
Contribution
ALL
6094.38
126.54
79.67
79.85
6094.38
126.54
Contribution
STACKS
0.01
0.58
72.29
72.22
0.01
0.58
Contribution
BLINES
6094.37
125.95
7.37
7.63
6094.37
125.95
21
-------
Table 23 shows the results for the 1-hr SO2 impacts with the three buoyant lines and the two
point sources modeled. The MAXDCONT output option was specified for source groups ALL,
STACKS, and BLINES. The emission rate was 1.0 g/s/m2 for all three lines, i.e. no emission
factors were applied.
Table 23. Concentration Estimates (u.g/m3) for Test G (1-hr SCh)
Group
ALL
STACKS
LINES
Rank
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
Average
Concentration
146052.19
1825.36
1488.07
1171.84
841.63
831.39
790.26
787.33
763.23
761.52
471.12
466.96
464.74
459.90
456.18
446.70
437.29
436.62
427.32
413.75
146052.19
1825.36
1488.07
1171.84
841.62
831.39
790.26
787.32
763.23
761.52
Contribution
ALL
146052.19
1825.36
475.09
469.95
146052.19
1825.36
Contribution
STACKS
0.00
0.00
471.12
466.96
0.00
0.00
Contribution
BLINES
146052.19
1825.36
3.97
2.98
146052.19
1825.36
The next three tests demonstrate the impacts resulting from different orientations of the buoyant
line source: rotated 90 degrees, vertical, and horizontal. In these tests, the receptor and the
buoyant line that caused the extremely large impact seen in the tables above do not coincide, thus
the smaller impacts. The purpose of these tests is to show consistency between the results, i.e.
there is no apparent anomalous behavior.
22
-------
Table 24 shows the impacts when the buoyant lines are oriented from the WSW to the ENE, i.e.
rotated 90 degrees from the orientation shown in Figure 1.
Table 24. Concentration Estimates (|ig/m3) for Test H -
Buoyant Lines Oriented ENE to WSW
Rank
1
2
3
4
5
1-hour
468.59
465.44
465.26
459.41
452.37
3-hour
297.76
294.39
294.32
292.62
288.80
24-hour
73.06
72.85
72.62
72.41
71.67
Annual
20.67
19.33
18.47
17.83
17.38
Table 24 shows the impacts when the buoyant lines are oriented from the WSW to the ENE, i.e.
rotated 90 degrees from the orientation shown in Figure 1.
Table 25. Concentration Estimates (|ig/m3) for Test I -
Buoyant Lines Vertically
Rank
1
2
3
4
5
1-hour
608.86
608.85
566.36
566.36
566.36
3-hour
325.94
314.33
314.33
314.33
304.85
24-hour
89.16
87.73
85.22
81.96
78.68
Annual
30.85
28.59
26.04
23.81
20.50
Table 24 shows the impacts when the buoyant lines are oriented from the WSW to the ENE, i.e.
rotated 90 degrees from the orientation shown in Figure 1.
Table 26. Concentration Estimates (|ig/m3) for Test J -
Buoyant Lines Horizontally
Rank
1
2
3
4
5
1-hour
469.16
465.80
465.70
459.92
452.98
3-hour
297.88
294.54
294.52
292.72
288.97
24-hour
81.28
81.05
80.44
80.31
79.98
Annual
25.27
25.09
24.64
24.39
24.26
23
-------
4. Differences between AERMOD/BLP and BLP
Some differences between the two models (AERMOD/BLP and BLP) are to be expected due in
part, to:
1) The AERMOD/BLP meteorology has a variable reference height as developed by
AERMET. However, in BLP the measurement height (ZMEAS in the METIN namelist)
is a constant 50 meters. Some tests were run to see if this made a difference. Although it
changed the AERMOD/BLP results, the changes were not significant.
2) Filling of missing meteorological data for BLP.
3) Use of a compiler option to enforce double precision versus explicit declaration in the
code - the effect of this option may contribute to differences.
Further testing may yield insights of other reasons differences are seen between AERMOD/BLP
and BLP.
5. Additional Changes to AERMOD/BLP
5.1 Background Concentration
Investigation of the use of the background keyword was initiated. The results were not as
expected, but there was no opportunity to determine why the results were not as expected.
5.2 Event Processing
A couple lines of code were added in the file evcalc.f for event processing but did not go any
further. The buoyant_line module was added as was code to save the ambient temperature (note
that when processing point sources, AERMOD changes the value of the ambient temperature
stored in the variable TA), and a comment about the P-G stability class.
5.3 BLP Source Group
The BLPGROUP keyword was not added since it would not provide any additional capabilities
at this stage of incorporating a single buoyant line source (composed of multiple lines). This
keyword will be useful as a future enhancement to allow more than one buoyant line source to be
processed in a single model run.
6. Recommendations
24
-------
A primary recommendation for AERMOD/BLP is to develop the meteorology and source
information associated with the 1979 SFe tracer study at the Reynolds Metals Patterson Plant.
Since all the relevant data are hardcopy (paper), time will be needed to enter the data
electronically.
25
-------
APPENDIX
1. AERMOD Keywords for Buoyant Line Sources
One LOCATION record for each line that comprises the buoyant line source.
SO LOCATION sourcelD BUOYLINE Xb Yb Xe Ye (Zs)
Xb, Yb = coordinate of the beginning of the line source (m)
Xe, Ye = coordinate of the end of the line source (m)
Zs = line elevation (m) - optional; defaults to 0.0 if omitted
One SRCPARAM record for each line that comprises the buoyant line source.
SO SRCPARAM sourcelD blemis relhgt
Blemis = emission rate (g/s/m2)
Relhgt = release height (m)
The BLPINPUT keyword defines the average values for the buoyant line (as a whole and not as
the individual lines that comprise the buoyant line source). The order shown is the same as the
input in BLP (in the namelist RISE), with the variable name used in BLP shown in parentheses.
This keyword can only appear once.
SO BLPINPUT blavgblen blavgbhgt blavgbwid blavglwid blavgbsep blavgfprm
blavgblen (L) = average building length (m)
blavgbhgt (HB) = average building height (m)
blavgbwid (WB) = average building width (m)
blavglwid (WM) = average line source width (m) (of the individual lines)
blavgbsep (DX) = average building separation (m) (between the individual lines)
blavgfprm (FPR1ME) = average buoyancy parameter (m4/s3)
26
-------
-------
United States Office of Air Quality Planning and Standards Publication No. EPA-454/P-15-004
Environmental Protection Air Quality Assessment Division July 2015
Agency Research Triangle Park, NC
------- |