EPA 910-R-15-001 c Alaska
United States Region 10 Idaho
Environmental Protection 1200 Sixth Avenue Oregon
Agency Seattle WA 98101 Washington
Office of Environmental Assessment October 2015
-------
-------
Combined WRF/MMIF/
AERCOARE/AERMOD Overwater
Modeling Approach for Offshore
Emission Sources
Volume 3 - Analysis of AERMOD
Performance Using Weather Research
and Forecasting Model Predicted and
Measured Meteorology in the Arctic
EPA Contract No. EP-W09-028
Work Assignment No. M12PG00033R
Prepared for:
U.S. Environmental Protection Agency
Region 10
1200 Sixth Avenue
Seattle, WA 98101
and
U.S. Department of the Interior
Bureau of Ocean Energy Management
45600 Woodland Road
Sterling, VA 20166
Prepared by:
Ramboll Environ US Corporation
773 San Marin Drive, Suite 2115
Novato, CA, 94998
and
Amec Foster Wheeler
Environmental & Infrastructure, Inc.
4021 Stirrup Creek Dr., Suite 100
Durham, NC 27703
October 2015
-------
-------
The Region 10 Project Officer for the Interagency Agreement No. M12PGT00033R and EPA
contract number EP-W-09-028 was Herman Wong with technical support provided by Robert
Kotchenruther, PhD and Robert Elleman, PhD. From BOEM, the Project Officer was Eric J.
Wolvovsky and the Technical Coordinator was Ronald Lai, PhD. The Project Lead for the prime
contractor Amec Foster Wheeler was James Paumier while Project Lead for subcontractor
RAM BOLL ENVIRON was Ken Richmond. Peer review of draft Volume 2 and/or draft Volume 3
was provided by Steven Hanna, PhD of Hanna Consultants, Robert Paine, CCM of AECOM and
Christopher Lindsey, Shell Exploration and Production. Their reviews and comments are greatly
appreciated by R10 and BOEM.
The collaboration study was funded in part by the U.S. Department of the Interior, Bureau of
Ocean Energy Management, Environmental Studies Program, Washington DC, and the U.S.
Environmental Protection Agency, Region 10, Seattle, WA.
-------
DISCLAIMER
The opinions, findings, conclusions, or recommendations expressed in this report are those of
the authors and do not necessarily reflect the view of the U.S. Environmental Protection Agency
or the U.S. Department of the Interior, Bureau of Ocean Energy Management, nor does the
mention of trade names or commercial products constitute endorsement or recommendation for
use by the Federal Government.
-------
PREFACE
The recommended American Meteorological Society/Environmental Protection Agency
Regulatory Model (AERMOD) dispersion program continues to be studied for assessing air
quality concentration impacts from emission sources located at overwater locations under an
Interagency Agreement (IA) Number M12PGT00033R dated 9 August 2012 between the U.S.
Environmental Protection Agency (EPA), Region 10 and the U.S. Department of the Interior
(DOI), Bureau of Safety and Environmental Enforcement (BSEE) on behalf of the Bureau of
Ocean Energy Management (BOEM). Specifically, the work scope under the IA calls for Region
10 and BOEM to (1) assess the use of AERMOD as a replacement for the Offshore and Coast
Dispersion (OCD) model in a near-source (< 1,000 meters source-receptor distance) ambient air
quality impact analysis for sea surface based emission sources and (2) evaluate the use of
Weather Research and Forecasting (WRF) model predicted meteorology with AERMOD in lieu
of overwater meteorological measurements from platforms and buoys.
Results of the Region 10/BOEM collaboration study are described in a three volume report.
Volume 1 describes all six tasks completed under the IA. However, only a summary of the work
completed under Task 2 and Task 3 appears in Volume 1. Volume 2 and Volume 3 provides a
detailed description of the work in Task 2 and Task 3, respectively. The six tasks are:
Task 1. Evaluation of two Outer Continental Shelf Weather Research and Forecasting Model
Simulations
Task 2. Evaluation of Weather Research and Forecasting Model Simulations for Five Tracer
Gas Studies with AERMOD
Task 3. Analysis of AERMOD Performance Using Weather Research and Forecasting Model
Predicted Meteorology and Measured Meteorology in the Arctic
Task 4. Comparison of Predicted and Measured Mixing Heights
Task 5. Development of AERSCREEN for Arctic Outer Continental Shelf Application
Task 6. Collaboration Study Seminar
Prior to the collaboration study, Region 10 on 1 April 2011 approved the use of the Coupled
Ocean-Atmosphere Response Experiment (COARE) air-sea flux algorithm with AERMOD to
preprocess overwater measured meteorological data from platforms and buoys. Initially, the
preprocessing of the overwater measurements was done manually with COARE. Subsequently,
Region 10 funded a study that was completed in September 2012 that coded the COARE air-
sea flux procedure into a meteorological data preprocessor program called AERMOD-COARE
(AERCOARE). The AERCOARE program was uploaded to the EPA Support Center for
Regulatory Atmospheric Modeling (SCRAM) website on 23 May 2013 as a beta option for case-
by-case approval by EPA regional offices.
-------
[Blank]
-------
TABLE OF CONTENTS
LIST OF FIGURES X
LIST OF TABLES XVII
LIST OF ABBREVIATIONS AND ACRONYMS XIX
1 INTRODUCTION 1
2 ARCTIC WRF SIMULATIONS 3
2.1 Polar WRF 3
2.2 WRF Simulation Methodology 3
2.3 Sea Surface Temperature Datasets 8
3 EVALUATION OF WRF PERFORMANCE OVER OPEN-WATER PERIODS 11
3.1 METSTAT Statistics 11
3.2 METSTAT Benchmarks 12
4 EVALUATION OF WRF SIMULATION RESULTS 15
4.1 Overland METSTAT Results 15
4.2 Overwater METSTAT Results 18
4.3 Meteorological Time Series at Site C1: Chukchi-Burger 22
4.3.1 Wind speed 22
4.3.2 Wind direction 22
4.3.3 Air temperature 22
4.3.4 Sea surface temperature 22
4.3.5 Air-sea temperature difference 23
4.3.6 Relative humidity 23
4.3.7 PEL height 23
4.4 SiteC2: Chukchi-Klondike 39
4.4.1 Wind speed 39
4.4.2 Wind direction 39
4.4.3 Air temperature 39
4.4.4 Sea surface temperature 39
4.4.5 Air-sea temperature difference 39
4.4.6 Relative humidity 40
-------
4.4.7 PEL height 40
4.4.8 Discussion 40
4.5 Site B2: Reindeer Island 55
4.5.1 Wind speed 55
4.5.2 Wind direction 55
4.5.3 Air temperature 55
4.5.4 Sea surface temperature 55
4.5.5 Air-sea temperature difference 55
4.5.6 Relative humidity 56
4.5.7 PEL height 56
4.5.8 Discussion 57
4.6 Site B3: Beaufort-Sivulliq 72
4.6.1 Wind speed 72
4.6.2 Wind direction 72
4.6.3 Air temperature 72
4.6.4 Sea surface temperature 72
4.6.5 Air-sea temperature difference 72
4.6.6 Relative humidity 73
4.6.7 PEL height 73
4.6.8 Discussion 74
4.7 Summary 96
5 AERMOD SIMULATION METHODOLOGY 99
5.1 AERMOD Meteorological Input Files 99
5.2 Overwater Measurement Datasets and AERCOARE Processing 101
5.3 PBL Height Diagnosis Methods 104
5.4 Hypothetical Sources 105
5.5 Receptor Grid 107
5.6 WRF Meteorology Extraction and Processing 108
5.7 AERMOD Evaluation Methodology 110
5.7.1 Simulation scenarios 110
5.7.2 Statistical measures and methods 113
-------
6 AERMOD SIMULATION RESULTS 115
6.1 1 -hour Averages 116
6.1.1 Robust high concentration 116
6.1.2 Fraction-factor-of-two 121
6.1.3 Discussion 121
6.2 3-hour Averages 129
6.2.1 Robust high concentration 129
6.2.2 Fraction-factor-of-two 134
6.2.3 Discussion 134
6.3 8-hour Averages 141
6.3.1 Robust high concentration 141
6.3.2 Fraction-factor-of-two 142
6.4 24-hour Averages 152
6.4.1 Robust high concentration 152
6.4.2 Fraction-factor-of-two 153
6.5 Period Averages 165
6.5.1 Robust high concentration 165
6.5.2 Fraction-factor-of-two 171
6.6 SiteB2 Results 171
6.7 Site B3 Results 180
6.8 Sited Results 185
6.9 SiteC2 Results 189
6.10 Far-Source Results 190
7 CONCLUSIONS 201
8 REFERENCES 207
APPENDIX A: TASK 3 PROTOCOL
APPENDIX B: AERMOD RESULTS STATISTICAL SCORES
APPENDIX C: REPORT DISK
-------
LIST OF FIGURES
Figure 1. WRF four-year Arctic dataset nested modeling domains: 36km, 12km, and 4km 5
Figure 2. 12 km and 4 km WRF domains 6
Figure 3. Land-based temperature METSTAT results 16
Figure 4. Land-based humidity METSTAT results 17
Figure 5. Land-based wind speed METSTAT results 17
Figure 6. Land-based wind direction METSTAT results 18
Figure 7. Overwater temperature METSTAT results 20
Figure 8. Overwater humidity METSTAT results 20
Figure 9. Overwater wind speed METSTAT results 21
Figure 10. Overwater wind direction METSTAT results 21
Figure 11. C1 2011 wind speed time series 25
Figure 12. C1 2012 wind speed time series 26
Figure 13. C1 2011 wind direction time series 27
Figure 14. C1 2012 wind direction time series 28
Figure 15. C1 2011 air temperature time series 29
Figure 16. C1 2012 air temperature time series 30
Figure 17. C1 2011 sea-surface temperature time series 31
Figure 18. C1 2012 sea-surface temperature time series 32
Figure 19. C1 2011 air-sea temperature difference time series 33
Figure 20. C1 2012 air-sea temperature difference time series 34
Figure 21. C1 2011 relative humidity time series 35
Figure 22. C1 2012 relative humidity time series 36
Figure 23. C1 2011 PEL height time series 37
Figure 24. C1 2012 PEL height time series 38
Figure 25. C2 2010 wind speed time series 41
Figure 26. C2 2012 wind speed time series 42
Figure 27. C2 2010 wind direction time series 43
Figure 28. C2 2012 wind direction time series 44
-------
Figure 29. C2 2010 air temperature time series 45
Figure 30. C2 2012 air temperature time series 46
Figure 31. C2 2010 sea surface temperature time series 47
Figure 32. C2 2012 sea surface temperature time series 48
Figure 33. C2 2010 air-sea temperature difference time series 49
Figure 34. C2 2012 air-sea temperature difference time series 50
Figure 35. C2 2010 relative humidity time series 51
Figure 36. C2 2012 relative humidity time series 52
Figures?. C2 2010 PEL height time series 53
Figure 38. C22012 PEL height time series 54
Figure 39. B2 2010 wind speed time series 58
Figure 40. B2 2011 wind speed time series 59
Figure 41. B2 2010 wind direction time series 60
Figure 42. B2 2011 wind direction time series 61
Figure 43. B2 2010 air temperature time series 62
Figure 44. B2 2011 air temperature time series 63
Figure 45. B2 2010 sea surface temperature time series 64
Figure 46. B2 2011 sea surface temperature time series 65
Figure 47. B2 2010 air-sea temperature difference time series 66
Figure 48. B2 2011 air-sea temperature difference time series 67
Figure 49. B2 2010 relative humidity time series 68
Figure 50. B2 2011 relative humidity time series 69
Figure 51. B2 2010 PEL height time series 70
Figure 52. B2 2011 PBL height time series 71
Figure 53. B3 2010 wind speed time series 75
Figure 54. B32011 wind speed time series 76
Figure 55. B3 2012 wind speed time series 77
Figure 56. B3 2010 wind direction time series 78
Figure 57. B32011 wind direction time series 79
Figure 58. B3 2012 wind direction time series 80
-------
Figure 59. B32010 air temperature time series 81
Figure 60. B32011 air temperature time series 82
Figure 61. B32012 air temperature time series 83
Figure 62. B3 2010 sea surface temperature time series 84
Figure 63. B3 2011 sea surface temperature time series 85
Figure 64. B3 2012 sea surface temperature time series 86
Figure 65. B3 2010 air-sea temperature difference time series 87
Figure 66. B32011 air-sea temperature difference time series 88
Figure 67. B3 2012 air-sea temperature difference time series 89
Figure 68. B32010 relative humidity time series 90
Figure 69. B32011 relative humidity time series 91
Figure 70. B32012 relative humidity time series 92
Figure 71. B32010 PEL height time series 93
Figure 72. B32011 PEL height time series 94
Figure 73. B32012 PEL height time series 95
Figure 74. Overwater meteorological measurement sites and corresponding WRF inner-domain
extraction points 102
Figure 75. Source locations and structures 107
Figure 76. Visual of inner-most receptor rings 108
Figure 77. Robust high concentration results for MMIF.RCALF AERMOD 1-hour averaging
times 117
Figure 78. Robust high concentration results for MMIF.RCALT AERMOD 1-hour averaging
times 118
Figure 79. Robust high concentration results for AERC.RCALF AERMOD 1-hour averaging
times 119
Figure 80. Robust high concentration results for AERC.RCALT AERMOD 1-hour averaging
times 120
Figure 81. Fraction-factor-of-two MMIF.RCALF AERMOD results for 1-hour averaging time. .123
Figure 82. Fraction-factor-of-two MMIF.RCALT AERMOD results for 1-hour averaging time. .124
Figure 83. Fraction-factor-of-two AERC.RCALF AERMOD results for 1-hour averaging time.. 125
Figure 84. Fraction-factor-of-two AERC.RCALT AERMOD results for 1-hour averaging time.. 126
Figure 85. Concentration maxima vs. distance, Site B2, 2011, Source Group #4, 1-hravg 127
-------
Figure 86. PEL height corresponding to concentration maxima, Site B2, 2011, Source Group #4,
1-hravg 127
Figure 87. Wind speeds corresponding to concentration maxima, Site B2, 2011, Source Group
#4, 1-hravg 128
Figure 88. Wind speed corresponding to concentration maxima, Site B2, 2011, Source Group
#2, 1-hravg 128
Figure 89. PEL height corresponding to concentration maxima, Site B2, 2011, Source Group #2,
1-hravg 129
Figure 90. Robust high concentration results for MMIF.RCALF AERMOD 3-hour averaging
times 130
Figure 91. Robust high concentration results for MMIF.RCALT AERMOD 3-hour averaging
times 131
Figure 92. Robust high concentration results for AERC.RCALF AERMOD 3-hour averaging
times 132
Figure 93. Robust high concentration results for AERC.RCALT AERMOD 3-hour averaging
times 133
Figure 94. Fraction-factor-of-two MMIF.RCALF AERMOD results for 3-hour averaging times. 135
Figure 95 Fraction-factor-of-two MMIF.RCALT AERMOD results for 3-hour averaging times..136
Figure 96. Fraction-factor-of-two AERC.RCALF AERMOD results for 3-hour averaging times.
137
Figure 97. Fraction-factor-of-two AERC.RCALT AERMOD results for 3-hour averaging times.
138
Figure 98. Concentration vs. distance, Site B2, 2010, Source Group #4, 3-hr, avg 139
Figure 99. PEL height of concentration maxima vs. distance for Site B2, Source Group #4,
3-hr avg 140
Figure 100. WRF and Profiler soundings (temperature vs. height) at Endeavor Island Sept. 18th,
2010 141
Figure 101. Robust high concentration results for MMIF.RCALF AERMOD 8-hour averaging
times 143
Figure 102. Robust high concentration results for MMIF.RCALT AERMOD 8-hour averaging
times 144
Figure 103. Robust high concentration results for AERC.RCALF AERMOD 8-hour averaging
times 145
Figure 104. Robust high concentration results for AERC.RCALT AERMOD 8-hour averaging
times 146
-------
Figure 105. Fraction-factor-of-two MMIF.RCALF AERMOD results for 8-hour averaging times.
147
Figure 106. Fraction-factor-of-two MMIF.RCALT AERMOD results for 8-hour averaging times.
148
Figure 107. Fraction-factor-of-two AERC.RCALF AERMOD results for 8-hour averaging times.
149
Figure 108. Fraction-factor-of-two AERC.RCALT AERMOD results for 8-hour averaging times.
150
Figure 109. Concentration vs. distance, Site B2, 2010, Source Group #3, 8-hr, avg 151
Figure 110. PEL height of concentration maxima vs. distance for Site B2, 2010, Source Group
#3, 8-hr avg 151
Figure 111. 1/L of the concentration maxima vs. distance for Site B2, 2010, Source Group #3,
8-hr avg 152
Figure 112. Robust high concentration results for MMIF.RCALF AERMOD 24-hour averaging
times 154
Figure 113. Robust high concentration results for MMIF.RCALT AERMOD 24-hour averaging
times 155
Figure 114. Robust high concentration results for AERC.RCALF AERMOD 24-hour averaging
times 156
Figure 115. Robust high concentration results for AERC.RCALT AERMOD 24-hour averaging
times 157
Figure 116. Q-Q plot for Site C1, 2011, Source Group #5, 24-hr avg 158
Figure 117. Air-sea temperature difference corresponding to concentration maxima, Site C1,
2011, Source Group #5, 24-hr avg 158
Figure 118. PEL height corresponding to concentration maxima for Site C2, 2012, Source
Group #5, 24-hr avg 159
Figure 119. Fraction-factor-of-two MMIF.RCALF AERMOD results for 24-hour averaging times.
160
Figure 120. Fraction-factor-of-two MMIF.RCALT AERMOD results for 24-hour averaging times.
161
Figure 121. Fraction-factor-of-two AERC.RCALF AERMOD results for 24-hour averaging times.
162
Figure 122. Fraction-factor-of-two AERC.RCALT AERMOD results for 24-hour averaging times.
163
Figure 123. Concentration maxima vs. distance, Site C2, 2012, Source #3, 24-hr avg 164
-------
Figure 124. PEL height corresponding to concentration maxima, Site C2, 2012, Source Group
#3, 24-hr avg 164
Figure 125. Robust high concentration results for MMIF.RCALF AERMOD Period averaging
times 166
Figure 126. Robust high concentration results for MMIF.RCALT AERMOD period averaging
times 167
Figure 127. Robust high concentration results for AERC.RCALF AERMOD period averaging
times 168
Figure 128. Robust high concentration results for AERC.RCALT AERMOD period averaging
times 169
Figure 129. Concentration maxima vs. distance, Site C1, 2012, Source Group #1, Period
average 170
Figure 130. Concentration maxima vs. distance, Site B3, 2012, Source Group #1, Period
average 170
Figure 131. Fraction-factor-of-two MMIF.RCALF AERMOD results for period averaging times.
172
Figure 132. Fraction-factor-of-two MMIF.RCALT AERMOD results for period averaging times.
173
Figure 133. Fraction-factor-of-two AERC.RCALF AERMOD results for Period averaging times.
174
Figure 134. Fraction-factor-of-two AERC.RCALT AERMOD results for Period averaging times.
175
Figure 135. Concentration maxima vs. distance for the Period average, Site B2, 2010, Source
Group #3 176
Figure 136. Concentration maxima vs. distance, Site B2, 2010, Source Group #4 178
Figure 137. PEL height corresponding to concentration maxima, Site B2, 2010, Source Group
#4 179
Figure 138. Inverse L corresponding to concentration maxima, Site B2, 2010, Source Group #4
179
Figure 139. Wind speed corresponding to concentration maxima, Site B2, 2010, Source Group
#4 180
Figure 140. Concentration maxima vs. distance, Site B3, 2011, Source Group #3 183
Figure 141. Q-Q plot of WRF-based AERMOD concentration results versus observation-based
results, Site B3, 2011, Source Group #3 184
-------
Figure 142. PEL height vs. distance for concentration maxima, Site B3, 2011, Source Group #3
184
Figure 143. Wind speed vs. distance for concentration maxima, Site B3, 2011, Source Group
#3 185
Figure 144. Concentration maxima vs. distance for Site C1, 2010, Source Group #3 187
Figure 145. PEL height corresponding to concentration maxima, Site C1, 2011, Source Group
#3 188
Figure 146. Wind speed corresponding to concentration maxima, Site C1, 2011, Source Group
#3 188
Figure 147. Far-source at 10,000 m 1-hour average maximum concentrations MMIF.RCALF
simulations 192
Figure 148. Far-source at 10,000 m 1-hour average maximum concentrations MMIF.RCALT
simulations 193
Figure 149. Far-source at 10,000 m 1-hour average maximum concentrations AERC.RCALF
simulations 194
Figure 150. Far-source at 10,000 m 1-hour average maximum concentrations AERC.RCALT
simulations 195
Figure 151. Far-source at 10,000 m Period average maximum concentrations for MMIF.RCALF
simulations 196
Figure 152. Far-source at 10,000 m Period average maximum concentrations for MMIF.RCALT
simulations 197
Figure 153. Far-source at 10,000 m Period average maximum concentrations for AERC.RCALF
simulations 198
Figure 154. Far-source at 10,000 m Period average maximum concentrations for AERC.RCALT
simulations 199
-------
LIST OF TABLES
Table 1. WRF vertical grid structure 7
Table 2. WRF physics parameterization schemes used for the four-year Arctic simulation 8
Table 3. Performance benchmarks for simple and complex conditions 13
Table 4. Overwater datasets used in METSTAT evaluation 18
Table 5. Summary statistics for WRF meteorology extractions 96
Table 6. AERMOD meteorology fields 99
Table 7. Overwater measurement site details 102
Table 8. Hypothetical source and stack parameters 106
Table 9. WRF AERMOD meteorology extraction methods 110
Table 10. AERMOD simulation periods 112
Table 11. Site B2 1 -hour average statistics scores 176
Table 12. Site B3 1 -hour average statistics scores 181
Table 13. Site C1 1-hour average statistics scores 186
Table 14. Site C2 1 -hour average statistics scores 189
-------
[Blank]
-------
LIST OF ABBREVIATIONS AND ACRONYMS
AERC WRF meteorology extraction cases processed by AERCOARE
AERMIC American Meteorological Society/Environmental Protection Agency
Regulatory Model Improvement Committee
AERMOD American Meteorological Society/Environmental Protection Agency
Regulatory Model
AERCOARE AERMOD-COARE
ASTD Air-Sea Temperature Difference
AIDJEX Arctic Ice Dynamics Joint Experiment
6 Bowen ratio
BOEM Bureau of Ocean Energy Management
BSSE Bureau of Safety and Environmental Enforcement
c Model constant
c0 Observed concentration value
cp Predicted concentration value
c Average concentration value
cn nth highest concentration
°C Degrees Centigrade
COARE Coupled Ocean Atmospheric Response Experiment
DOI Department of the Interior
ECMWF European Center for Medium-Range Weather Forecasting
EPA Environmental Protection Agency
ERA ECMWF Reanalysis Project
ERA-40 ERA 45-year global atmospheric reanalysis
ERA-I ECMWF Re-Analysis project global atmospheric reanalysis
eta Vertical pressure coordinate in WRF
f Coriolis parameter
FF2 Fraction-factor-of-two
FNMOC Fleet Numerical Meteorology and Oceanography Center
g Grams
H Sensible heat flux
ISC3 Industrial Source Complex 3
K Degrees kelvin
kg Kilograms
km Kilometers
-------
L Monin-Obukhov length
LCC Lambert Conformal Conic
m Meters
METSTAT Meteorological Statistics
MG Geometric mean bias
MIXH PEL height or "mixing height"
MMIF Mesoscale Model Interface Program
MYJ Mellor-Yamada-Janjic PEL parameterization scheme used in WRF
NAAQS National Ambient Air Quality Standards
NARR North-American Regional Reanalysis
NBDC National Buoy Data Center
NCAR National Center for Atmospheric Research
NCEP National Center for Environmental Prediction
NOAA National Oceanic and Atmospheric Administration
NSR New Source Review
O Observed value
OBS, obs Label for observation-based AERMOD simulations
OCD Offshore and Coastal Dispersion
OCS Outer Continental Shelf
OLM Ozone Limited Method
P Sea level atmospheric pressure (also used to indicate "predicted" value in
statistical calculations)
p Predicted value
PEL Planetary boundary layer
PFL Profile file input to AERMOD
PRIME Plume Rise Model Enhancements
PVMRM Plume Volume Molar Ratio Method
Q-Q Quantile-Quantile (statistical plot)
r Albedo
RCALF Label for WRF-MMIF AERMOD simulations where PEL height was not
recalculated by MMIF
RCALT Label for WRF-MMIF AERMOD simulations where PEL height was
recalculated MMIF
rg Geometric correlation coefficient
RH Relative humidity
RHC Robust High Concentration
-------
RMSE Root Mean Square Error
RPO Regional Planning Organization
RTG Real Time Global sea-surface temperature analysis (from NCEP)
s Seconds
SFC AERMOD surface meteorology input file
SST Sea Surface Temperature
T Temperature
TKE Turbulent Kinetic Energy, Thermal Kinetic Energy
TMS Total Model Score statistical measure
U Zonal wind component
UW-PBL University of Washington Shallow Convection PEL
i/* Friction velocity
V Meridional wind component
VG Geometric variance
VPTLR Virtual Potential Temperature Lapse Rate
w*. Convective scaling velocity
W Watts
WD Wind direction
WRF Weather Research and Forecasting
WS Wind Speed
YSU Yonsei University
z Height above the surface
z0 Surface roughness length
z,c Convective PEL height
Zim Mechanical PEL height
0 Degrees angular
jjg Geometric mean
OQ Standard deviation of wind direction
ow Standard deviation of vertical wind speed
(// Stability correction parameter
-------
[Blank]
-------
1 INTRODUCTION
This study aims to evaluate alternative methods for supplying meteorological variables to the
American Meteorological Society/Environmental Protection Agency Regulatory Model
(AERMOD) for regulatory air quality modeling of sources located over the ocean. It is
hypothesized, given an appropriate overwater meteorological dataset, AERMOD can be used
for New Source Review (NSR) following the same procedures as used for sources over land. A
summary of all the elements of the study is contained within Volume I. Volume II summarizes
Task 2 of the study which focused on evaluation of AERMOD against historical tracer study
measurements using meteorology inputs derived from Weather Research and Forecasting
(WRF) model datasets. Similarly, Volume 3 summarizes Task 3 of this study and focuses on
evaluation of AERMOD using WRF and observational meteorology datasets over the Chukchi
and Beaufort Seas along the Arctic coasts of Alaska. Like the Task 2 methodology, Task 3
evaluates a combined modeling approach where the meteorological variables are provided by
WRF, processed by a combination of the Mesoscale Model Interface Program (MMIF)1 and
optionally, the AERMOD-Coupled Ocean-Atmosphere Response Experiment (COARE) or
AERMOD-COARE (AERCOARE)2. WRF meteorology is used to drive AERMOD for several test
cases. The results are compared to results of AERMOD driven by meteorological observations.
The purpose of this study is to provide evidence to help answer some of the following questions:
• How well does WRF predict overwater surface meteorology in the Arctic?
• Are pollutant concentrations predicted by AERMOD driven by WRF meteorology as
conservative as those predicted by AERMOD driven by meteorological measurements?
• What WRF modeling configurations and meteorology extraction methods provide the best
AERMOD inputs?
• How sensitive is AERMOD to differences between the WRF meteorology and observations
for simulations of typical OCS sources?
A WRF meteorological dataset suitable for dispersion modeling in the Arctic was developed for
this study. WRF meteorology is extracted from the WRF output files and used to drive AERMOD
simulations for ice-free periods of 2009-2012, where overwater-based observational datasets
were available. Results from the observation-based and WRF-based AERMOD simulations are
compared and contrasted in an attempt to answer these questions.
Meteorological observation datasets from four overwater locations were obtained for this study.
Two of the locations were in the Beaufort Sea and two were in the Chukchi Sea. Data were
1 MMIF-Beta provided as a "related" alternative software for regulatory dispersion modeling by the U.S.
EPA at the website: http://www.epa.gov/scram001/dispersion related.htm
2 AERCOARE is made publically available by the U.S. EPA at the website:
http://www.epa.gov/ttn/scram/dispersion related.htm
-------
available at these locations for various time-spans during the ice-free summer and autumn
periods of 2010, 2011, and 2012. The 2010-2012 periods were selected to overlap previous
modeling efforts and to take advantage of the vertical temperature profiler data collected at
Endeavor Island during this period. The profiler was a passive microwave radiometer operating
from 2010 to 2012 at the offshore Endeavor Island facility near Prudhoe Bay, Alaska. The
profiler data were used to assist in the estimates of atmospheric planetary boundary layer (PBL)
height at each of the sites. The term "PBL height" is used to indicate the height or depth of the
mixing layer and is synonymous with "mixing height."
This report summarizes the methodology and results for each element of the Task 3
investigation, including:
i) The methodology used for the WRF simulations
ii) Evaluation of the WRF performance
iii) The methodologies used to prepare AERMOD meteorology from both the observational
datasets and the WRF simulations.
iv) The AERMOD modeling approach and methodology.
v) Evaluation of the AERMOD results and comparisons of observation-based and WRF-
based AERMOD results.
vi) Examination of the influence of the meteorological data on AERMOD performance.
-------
2 ARCTIC WRF SIMULATIONS
Mesoscale atmospheric modeling of the Arctic is challenging in many regards. The region is
characterized by a climate prone to extremes. Very cold winter temperatures, powerful
cyclones, and extreme overwater surface inversions are typical features of the weather and
climate in the region. Modeling of the surface energy balance is challenging considering unique
features such as seasonal ice cover, frozen tundra ground surfaces, and solar insolation
extremes. These extremes may result in conditions beyond the capabilities or bounds of the
parameterization schemes used in WRF.
2.1 Polar WRF
The Ohio State University Polar Meteorology Group has developed a polar-optimized version of
WRF, "Polar WRF." This version of the model was developed to improve WRF Arctic modeling
capabilities (Mines & Bromwich, 2008)(Bromwich, et al., 2009). Their modifications have
focused on optimizing the surface energy budget and parameterization of sea-ice and
permanent ice surfaces within the Noah land surface model (Chen, et al., 1997). The
modifications have included implementation of a variable sea-ice and snow thickness and
seasonally-variable sea-ice albedo. Polar WRF was used for this study.
The WRF simulations were run using the Mellor-Yamada-Janjic (MYJ) (Mellor & Yamada,
1982)(Janjic, 1994) PEL parameterization scheme. MYJ is a robust one-dimensional local-
closure turbulent kinetic energy (TKE) scheme widely used in the modeling community. The
MYJ PEL scheme was selected over its alternatives because it was the preferred PEL scheme
for Polar WRF after the original Polar WRF benchmark study by Mines and Bromwich (2008).
Bromwich et al. (2013) also confirmed MYJ performance was satisfactory for Arctic simulation.
An evaluation of Polar WRF performance under different sets of options and settings was not
conducted and beyond the scope of this study. Significant effort has been made to optimize
Polar WRF performance by the developers of the model, as documented in Mines and
Bromwich (2008), Wilson et al. (2009), Bromwich et al. (2009), Bromwich et al.(2013), and
Mines etal. (2015).
2.2 WRF Simulation Methodology
WRF modeling was conducted generally according to a modeling protocol reviewed and
accepted by EPA and BOEM representatives prior to the study and is included in Appendix A.
The WRF simulations were conducted using the National Center for Atmospheric Research
(NCAR) community-developed WRF model dynamical core version 3.4.1 in conjunction with the
Ohio State University Polar WRF version 3.4.1 modules (Mines & Bromwich, 2008). WRF is a
limited-area non-hydrostatic, terrain-following eta-coordinate mesoscale model (NCAR, 2014).
WRF is the state-of-the-art mesoscale model used today to forecast regional weather, diagnose
historical weather events, and provide meteorological datasets for regional air quality dispersion
modeling.
-------
The simulations were conducted on a polar-stereographic map projection. The outer-most 36
kilometer (km) domain encompassed all of Alaska and parts of Northern Canada and Russia, as
shown in Figure 1. A 12 km nested domain included most of interior Alaska and the Bering,
Chukchi, and Beaufort Seas. A 4 km nested domain focused on the regions of the Chukchi Sea,
Beaufort Sea, and the North Slope of Alaska. The 4 km domain was sized to contain all of the
OCS Lease Blocks and was built with a 70 km buffer, as shown in Figure 2 to account for a
50 km buffer around sources and receptors. The additional 20 km buffer was used to account
for WRF "edge-effect" contamination, an artifact of numerical downscaling. The additional 20 km
buffer encompasses five grid points on the edge of the nested domain - the five grid point buffer
is typically used to account for edge-effects (NCAR, 2014). The CALPUFF 50 km buffer is
required by EPA long-range transport modeling guidance to account for possible recirculation of
pollutants (EPA, 1998).
The 36 km domain was comprised of 110 by 120 grid points, south to north and west to east,
respectively. The 12 km domain was 130 by 157 grid points and the 4 km domain was 151 by
271 grid points. The WRF vertical grid structure was built using 37 levels, disproportionately
stacked towards the surface. The boundary layer resolution used finer vertical spacing than
typically used for most simulations over land to help the meteorological fields respond more
explicitly to dynamical influences. The vertical grid structure is described in Table 1, including
layer average height and thickness estimates based on the hypsometric equation, Eq. (1):
dz
The ERA-lnterim (ERA-I) global atmospheric reanalysis (Simmons, et al., 2006) was used as
the driving reanalysis dataset for the 4-year WRF Arctic dataset. The ERA-I is a global
atmospheric reanalysis produced by the European Centre for Medium Range Weather
Forecasts (ECMWF). ERA-I includes 6-hourly output, 37 pressure levels, and 0.75° x 0.75°
spatial resolution. It is a widely-used dataset with the appropriate coverage for simulations of the
Arctic. The North American Regional Reanalysis (NARR) dataset could not be used for the
study because the outer WRF domain exceeds the boundaries of the NARR domain.
WRF can use temporal and spatial data assimilation methods to "nudge" gridded wind,
temperature, and water vapor towards observations or analysis data. When nudging is applied,
meteorological variables at adjacent grid points are relaxed towards the observed or analysis
value, weighted by distance. Observation nudging was not used for the WRF simulations in this
study. Analysis nudging was used for the WRF simulations on the 36 and 12 km domains. PEL
nudging of wind, moisture, and temperature was not used to comply with advice given in
(Stauffer, Seaman, & Binkowski, 1991).
-------
Observational nudging can also be used in a preliminary step using the WRF preprocessor
"obsgrid" in an attempt to improve the analysis dataset. Observation nudging was used on the
analysis dataset in this study.
130°E
140°E
150°E
160°E -
170°E
ALASKA BOEM
150°E 150°W 90°W
80°W
90°W
- 100°W
110°W
120°W
180C
170°W 160°W 150°W 140°W 130°W
1 25 75 200 500 1000 1500 2000 3000
Figure 1. WRF four-year Arctic dataset nested modeling domains: 36km, 12km, and 4km.
-------
Active Lease Sites
180°
165°W
150°W
135°W
72°N
70°N
68°N
66°N
64°N
62°N
*•
72°N
70°N
68°N
66°N
64°N
62°N
170°W 165°W 160°W 155°W 150°W 145°W 140°W
1 25 75 200 500 1000 1500 2000 3000
• Lease site (magenta)
• Arctic National Wildlife Refuge (orange)
Denali Wilderness, National Preserve, and National Park (red)
• 70 km buffer from active lease sites (yellow arrows)
Figure 2. 12 km and 4 km WRF domains.
-------
Table 1. WRF vertical grid structure.
Level
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
26
27
28
29
30
31
32
33
34
35
36
37
"eta" level
1
0.9985
0.997
0.995
0.993
0.991
0.988
0.985
0.98
0.97
0.96
0.95
0.94
0.93
0.91
0.89
0.87
0.84
0.8
0.76
0.72
0.68
0.64
0.6
0.55
0.5
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.06
0.027
0
Pressure (mb)
1000
999
997
995
993
991
989
986
981
972
962
953
943
934
915
896
877
848
810
111
734
696
658
620
573
525
478
430
383
335
288
240
193
145
107
76
50
Height (m)*
0.0
12.2
24.5
40.8
57.2
73.6
98.3
123.0
164.3
247.4
331.2
415.7
500.8
586.6
760.5
937.2
1117.1
1392.8
1772.4
2166.7
2577.0
3005.0
3452.2
3921.0
4540.7
5203.7
5917.1
6690.5
7536.4
8472.3
9522.5
10724.1
12136.7
13866.9
15621.6
17503.4
19594.2
Mid Height (m)*
—
6.1
18.4
32.7
49.0
65.4
85.9
110.6
143.6
205.9
289.3
373.4
458.2
543.7
673.5
848.8
1027.1
1254.9
1582.6
1969.6
2371.9
2791.0
3228.6
3686.6
4230.8
4872.2
5560.4
6303.8
7113.5
8004.4
8997.4
10123.3
11430.4
13001.8
14744.2
16562.5
18548.8
Layer Thickness (m)*
—
12.2
12.2
16.4
16.4
16.4
24.7
24.7
41.3
83.1
83.8
84.5
85.1
85.8
173.8
176.8
179.8
275.8
379.6
394.3
410.3
427.9
447.3
468.7
619.8
662.9
713.4
773.4
846.0
935.8
1050.2
1201.6
1412.6
1730.1
1754.7
1881.8
2090.8
"Standard height and thickness estimated using P0=1000mb, Ptop=50mb, T0=20.15 °C, and dT/dz=-6.5 °C/km.
-------
The 4-year WRF simulation (2009-2012) was conducted in 5.5-day simulation blocks with 12
hours of overlapping time to account for model "spin-up." The spin-up time allows for the model
to develop sub-grid scale processes, including mature vorticity, convection, and moisture fields.
The list of physics parameterization schemes used in the WRF modeling is included in Table 2.
The same schemes and settings used for the Task 2 study WRF modeling were used except
only the MYJ PEL scheme was used for this study.
Table 2. WRF physics parameterization schemes used for the four-year Arctic simulation.
Parameterization
Micro-physics
(mp_physics)*
PEL physics
(bl_pbl_physics)*
Option
selected
Thompson
MYJ
WRF
option #
8
2
Description
Moisture physics parameterization.
Thompson scheme: 6-class
hydrometeors
MYJ: local TKE scheme
Source
Thompson et al. (2008)
(Mellor and Yamada,
1982 and Janjic, 1994)
Cumulus / convection
(cu_physics)*
Kain-Fritsch 1
Sub-grid convection scheme using
mass-flux approach. Used on 36
and 12 km domains only: resolved
explicitly on high resolution
domains. Also used Kain-Fritsch
"Eta" moisture advection trigger
Kain (2004)
Radiation
(ra_sw_physics)*
(ra Iw physics)*
Land surface
(sf_surface_physics)*
Surface layer
(sf sfclay physics)*
RRTMG
Unified
Noah LSM
ETA M-O
similarity
4
2
2
Rapid radiative transfer model
using cloud overlap schemes
4-layer soil model with fractional
snow cover, frozen soil physics,
and ice sheet cover physics
Monin-Obukhov similarity theory
based scheme
lacono etal. (2008)
Tewari et al. (2004)
Janjic (1994)
*WRF model keywords and options names.
2.3 Sea Surface Temperature Datasets
The simulation originally used the National Center for Environmental Prediction (NCEP) Real
Time Global (RTG) SST analysis dataset. The RTG dataset contains satellite-derived SST at
0.5°C resolution. Arctic WRF simulations may be quite sensitive to the accuracy of the sea-ice
or sea-surface temperature (SST) dataset used. Preliminary investigation revealed deficiencies
within the NCEP RTG SST analysis dataset over a span of the open water periods of interest.
Warm water surface plumes from Mackenzie River outflow resulted in overpredictions of SST
across the Beaufort Sea in the RTG dataset. The overprediction of SST was attributed to
smoothing techniques used in RTG data analysis.
The U.S. Naval Fleet Numerical Meteorology and Oceanography Center (FNMOC) Global SST
analysis dataset (USN, 2014) was identified as a sufficient alternative to correct for the biases
observed in the RTG dataset. Preliminary investigation of the FNMOC dataset revealed superior
accuracy and depiction of the Mackenzie River plume when compared to the RTG dataset.
-------
The FNMOC dataset is created using satellite-derived SST data at 0.25°C resolution and in situ
SST data from ships and buoys and updated every 6 hours. Remotely-sensed SST data using
passive infra-red sensors may be prone to error due to difficulties estimating temperature when
low cloud cover or partial sea-ice cover is present (Xu & Ignatov, 2010). As a result, the time
series of SST data may be discontinuous, marked by sudden shifts in SST magnitude during
weather regime changes or in the case where in-situ data becomes available as ships or buoys
enter the grid cell. The dataset is also temporally coarse with an update frequency of every six
hours that also contributes to the discontinuous nature of the dataset. Despite these
deficiencies, the FNMOC dataset represents one of the highest quality SST datasets available.
-------
[Blank]
10
-------
3 EVALUATION OF WRF PERFORMANCE OVER OPEN-WATER PERIODS
FNMOC-SST-based WRF performance was assessed in two ways: quantitatively with statistics
relating WRF meteorology to measurements and qualitatively by graphical comparison of WRF
meteorology to measurements. A portion of the quantitative analysis was conducted using the
publically-available METSTAT software (ENVIRON Int. Corp., 2014). METSTAT calculates a
suite of model performance statistics using wind speed and direction, temperature, and moisture
observations. WRF predictions are extracted from the nearest grid cell for comparison to the
observed values. METSTAT computes metrics for bias, error, and correlation and compares
them to a set of performance benchmarks set for ideal model performance (Emery, et al., 2001).
3.1 METSTAT Statistics
Statistical measures calculated by METSTAT include observation and prediction means,
prediction bias, and prediction error. The METSTAT analyses are valuable for evaluating the
performance of the WRF simulations on a domain-wide level.
Mean observation (M0) is calculated using values from all sites for a given time period by
Eq. (2):
2
7=1 i=l
where O'j\s the individual observed quantity at site /and time/, and the summations are over all
sites (t) and overtime periods (J).
Mean Prediction (MP) is calculated from simulation results that are interpolated to each
observation used to calculate the mean observation for a given time period by Eq. (3):
3
7 = 1 i=l
where P] is the individual predicted quantity at site /and time/. Note the predicted mean wind
speed and mean resultant direction are derived from the vector-average (for east-west
component z/and north-south component v).
Bias (B) is calculated as the mean difference in prediction-observation pairings with valid data
within a given analysis region and for a given time period by Eq. (4):
<«>
w-°;)
7=1 i=l
11
-------
Gross Error (E) is calculated as the mean absolute difference in prediction-observation pairings
with valid data within a given analysis region and for a given time period by Eq. (5):
1
—yy
//ZjZj
7 = 1 i=l
(5)
PJ-°J
Note the bias and gross error for winds are calculated from the predicted-observed residuals in
speed and direction (not from vector components u and v). The direction error for a given
prediction-observation pairing is limited to range from 0 to ±180°.
Root Mean Square Error (RMSE) is calculated as the square root of the mean squared
difference in prediction-observation pairings with valid data within a given analysis region and
for a given time period by Eq. (6):
1
; / 12
RMSE =
h
7=1i=l
(6)
The RMSE, as with the gross error, is a good overall measure of model performance. However,
since large errors are weighted heavily (due to squaring), large errors in a small sub-region may
produce a large RMSE even though the errors may be small and quite acceptable elsewhere.
Additional WRF performance analyses were conducted using the time series of meteorological
variables extracted at the buoy measurement sites. The approach and analysis is discussed in
Section 4.0. The time series comparisons are a valuable tool for understanding the influence of
meteorology on AERMOD prediction performance because the meteorology at the extracted
point is used to drive AERMOD.
3.2 METSTAT Benchmarks
The METSTAT benchmarks were developed using the results of about 30 meteorological model
performance simulations performed to support air quality studies of urban areas (Emery, et al.,
2001). Another set of model performance benchmarks were developed for complex conditions
as part of the Western Regional Air Partnership (WRAP) meteorological modeling of the
western United States, including the Rocky Mountain Region as well as the complex conditions
in Alaska (Kemball-Cook, et al., 2005).
Table 3 lists the meteorological model performance benchmarks for simple and complex
(Kemball-Cook, et al., 2005) terrain. The benchmarks provide a measure of WRF model
performance with regards to other modeling cases in the U.S. However, given the wide variety
of landforms, weather, and climatic regions in the U.S. it is likely these benchmarks are
applicable to most regions of the world. Less stringent criteria have been developed for complex
12
-------
terrain conditions based on the higher degree of variance found in regions of heterogeneous
terrain and microclimate. Point measurements along the coast can be influenced by marine or
land-based boundary layers, depending on the conditions at any given moment. Strong
gradients of temperature, relative humidity (RH), and cloud can exist at the interface between
the marine and land PBLs. Large differences in meteorology between WRF and measurement
data may occur if WRF grid resolution is not dense enough to resolve these tight gradients, or if
the WRF grid cell location is not located within the same PEL (marine or land-based) as the
measurement location. Given the complexity of meteorological conditions in the Arctic, it may be
assumed the complex criteria provide a more suitable set of performance goals.
Table 3. Performance benchmarks for simple and complex conditions.
Parameter Simple Complex
Temperature Bias <±0.5 K <±2.0 K
Temperature Error < 2.0 K < 3.5 K
Humidity Bias < ±0.8 g/kg < ±1.0 g/kg
Humidity Error <2.0 g/kg <2.0 g/kg
Wind Speed Bias < ±0.5 m/s < ±1.5 m/s
Wind Speed RMSE < 2.0 m/s < 2.5 m/s
Wind Direction Bias < ±10 degrees < ±10 degrees
Wind Direction Error < 30 degrees < 55 degrees
Although METSTAT analysis can be applied to individual meteorological station datasets, it is
typically used to evaluate performance against a group of stations within the WRF domain. This
approach is advantageous because it evaluates performance across the entire domain and
dampens bias that can occur at any individual site (advantageous if the climate at the site is
heavily influenced by small-scale local terrain or roughness features not resolved in the WRF
domain). However, if too many stations are used in the analysis the statistics may be unduly
smoothed and not truly representative of WRF performance.
METSTAT statistical results are typically displayed in a "soccerplot." This type of plot contains
the statistical results for selected periods plotted with respect to the simple and complex "goals"
listed in Table 3. If a point is located within the goals, it indicates the METSTAT results from the
given period satisfy the criteria benchmarks for the meteorological variable in question. If the
point is outside the goals it does not satisfy the criteria, indicating WRF performance was poor
for the particular period and region evaluated, if the observational data used is sufficiently
representative of the regional meteorology. If the point is within the complex terrain criteria goals
but outside the simple criteria goals, WRF performance is satisfactory for complex terrain and
meteorological conditions, but not necessarily for simple terrain and meteorological conditions. If
the point is inside both the simple and complex criteria goals, WRF performance is considered
satisfactory if the observational datasets used are accurate and representative. This result
indicates good agreement between observations and simulated surface meteorology.
13
-------
METSTAT was used to evaluate the performance of each WRF simulation using surface
meteorological data from the U.S. National Climatic Data Center DS-3505 database (NOAA-
NCDC, 2014). The database contains records of most official surface meteorological stations
from airports, military bases, reservoirs/dams, agricultural sites, and other sources dating from
1901 to the present.
14
-------
4 EVALUATION OF WRF SIMULATION RESULTS
This section evaluates WRF performance with respect to observed meteorology using the
methods described in Section 3. The evaluation includes both time series comparisons and
METSTAT analyses. Time series of observed and extracted meteorology were plotted for each
site for the open-water periods used in this study, together for qualitative comparison. The
qualitative analysis provided insight into the capabilities of WRF for each scenario at each
measurement site. A full description of the four measurement sites used in this study is provided
in Section 5.2.
The data plotted are from the direct MMIF extractions (without AERCOARE processing).
AERCOARE and MMIF PEL height recalculation does not change wind speed and direction,
RH, or sea-surface temperature. PEL height time series plots contain both direct WRF PEL
height heights (referred to hereafter as "RCALF" values) and MMIF-calculated PEL heights
(referred to hereafter as "RCALT" values). Both WRF and MMIF can produce unphysically low
PEL heights during highly stable conditions. A minimum PEL height limit of 25 m was enforced
for all the datasets used in this study. The minimum PEL height value of 25 m was set after
deliberations with EPA Region 10. It was noted that the 25 m value is a reasonable minimum
mixing height for the most extreme stable periods, based on measurements reported in (Hsu,
1988) (Garratt, 1992). A minimum absolute value of 5 m was established for the Monin-
Obukhov length (L), synonymous with the L limits used in the OCD model (DiCristofaro &
Hanna, 1989). AERCOARE reprocessing of WRF meteorology did not substantially change
PEL height.
4.1 Overland METSTAT Results
METSTAT was used to evaluate WRF performance using observations from land-based
measurement stored in the NOAA DS-3505 database (NOAA-NCDC, 2014). The inner-domain
(Domain 3) METSTAT results were selected for evaluation in this study. A total of 19 surface
station datasets were obtained and used for the METSTAT analysis for all periods analyzed.
The periods included the open-water seasons (Aug. -Oct.) of 2010, 2011, and 2012. These
results were useful for evaluating WRF performance on a domain-wide level.
The land-based METSTAT soccer plot results for temperature performance are plotted in Figure
3. Results were favorable, exhibiting low bias and error falling within the criteria for simple
conditions. Temperature simulation performance was favorable for each month and year
simulated.
Results for humidity performance are plotted in Figure 4. WRF humidity was slightly positively
biased, but the statistical scores were also within the simple conditions criteria. Results were
favorable for all periods simulated.
Results for wind speed and wind direction performance are plotted in Figure 5 and Figure 6,
respectively. Wind speed was biased slightly low, but generally within the simple conditions
criteria. Wind speed RMSE exceeded the criteria for simple terrain conditions for 2010 and 2011
15
-------
October simulation periods. The RMSE for these periods were within the complex conditions
criteria, however. Wind direction bias and error were favorable, falling within the simple
conditions criteria for all periods simulated.
BOEM/EPA AERCOARE Alaska d03 Temperature Performance
ERA+FNMOC SST-all
O CO-
Lll
£
p
Tempera)
) 1 2
I
Complex Conditions
* Aug
^ Sep
as Oct
1
-3
* 2010
* 2011
* 2012
1
-2
!
Simple Conditions
i i i
-1012
i
Temperature Bias (K)
Figure 3. Land-based temperature METSTAT results.
16
-------
BOEM/EPA AERCOARE Alaska d03 Humidity Performance
ERA+FNMOC SST-all
o
CO
in
s
C5 ° _
o
£
"0 0
E - "
Z3
•^- in
d
o
d
* Aug
Sep
® Oct
1
-2.0
* 2010
» 201 1
* 20 I £
I
-1.5
Simple Conditions
*
"/%
I 1 1 1 I
-1.0 -0.5 0.0 0.5 1.0 1.5
Humidity Bias (g/kg)
Figure 4. Land-based humidity METSTAT results.
BOEM/EPA AERCOARE Alaska d03 Wind Speed Performance
ERA+FNMOC SST-all
LU
CO
CC CM
GO o
13 T^
q
6
Compfex Conditions
* Aug
* Sep
a Oct
* 2010
* 2011
* 2012
\^
-2
Conditions
\
-1
Wind Speed Bias (m/s)
Figure 5. Land-based wind speed METSTAT results.
17
-------
BOEM/EPA AERCOARE Alaska d03 Wind Direction Performance
ERA+FNMOC SST-all
-------
Sep. 48536 1
Oct. 48536 1
a 101: ConocoPhillips Klondike buoy 1 located at 70.9N, 165.2W
CK: Shell Chukchi-Klondike buoy located at 71.5N, 164.1W (corresponds to Site C2 dataset)
CB: Shell Camden Bay buoy located at 70.4N, 146.OW (corresponds to Site B3 dataset)
102: ConocoPhillips Klondike buoy 2 located at 70.99N, 164.99W
MOB1: Joint (ConocoPhillips, Shell) Hanna Shoal buoy located at72.2N, 161.5W
MOB2: Joint Chukchi-Klondike buoy (2011 only) located at 70.9N, 165.2W (corresponds to Site C1 dataset)
MOBS: Joint Chukchi-Burger buoy (2011 only) located at 71.5N, 164.1W
48536: ICEX-AIR free arctic ice buoy; position varied
HB: Shell Harrison Bay buoy located at 70.9N, 150.3W
The METSTAT results for temperature are plotted in Figure 7. The results suggested WRF was
able to accurately predict temperature given the error and bias were within the more stringent
simple conditions criteria for 2010 and 2011. The 2012 results, from a single buoy, fell within the
complex conditions criteria for 2012.
Figure 8 shows the METSTAT results for humidity. Humidity bias and error were very low for
2010 and 2011, resulting in statistics that fell within the criteria. Note these criteria were
developed for land-based datasets that typically exhibit a larger variance in humidity. Sea-based
humidity in the Arctic tended to be near water vapor saturation most of the time.
The wind speed METSTAT results are shown on Figure 9. Although August and September
statistics were within the simple conditions criteria, WRF tended to underpredict wind speeds by
exhibiting a degree of negative bias. The October simulations were more scattered, with scores
exceeding the RMSE criteria, but falling within the complex conditions criteria for bias.
Figure 10 shows the METSTAT results for wind direction. The wind direction error and bias is
misrepresented because the Arctic buoy datasets contained in the National Buoy Data Center
(NBDC) database are aligned to magnetic north. The correction of measurements to true north
in this region is about a +14° adjustment. This adjustment would result in WRF bias falling within
the "simple conditions" criteria and would likely reduce error into the "simple conditions" criteria
also.
19
-------
BOEM/EPA AERCOARE Alaska d03 Temperature Performance
ERA+FNMOC SST-all
I
o co -
LU
CD
Q_
E
0) r- —
|—
Comctex Conditions
* Aug
* Sep
w Oct
1
-3
* 2010
* 2011
* 2012
1
-2
Simple Conditions
%«»
® $
e
*
i r i
-1012
Temperature Bias (K)
Figure 7. Overwater temperature METSTAT results.
BOEM/EPA AERCOARE Alaska d03 Humidity Performance
ERA+FNMOC SST-all
o
CO
in
o
5 5-
!5 o
E - '
Z3
I in
0
0
o
* Aug
* Sep
o Oct
I
-2.0
* 2010
# 2011
* 2012
1
-1.5
Comnlex Conditions
• .inn •[- Conditions
"V
III I
-1.0 -0.5 0.0 0.5 1.0 1.5
Humidity Bias (g/kg)
Figure 8. Overwater humidity METSTAT results.
20
-------
BOEM/EPA AERCOARE Alaska d03 Wind Speed Performance
ERA+FNMOC SST-all
"ta
•5. n -
LU
i
a:
-D W -
s
CO
"0 ,- -
1
0 -
* Aug
•$ Sep
» Oct
* 2010
* 2011
* 2012
I
-2
*
e
Complex Conditions
*
Simple Conditions
i i i i
-1012
Wind Speed Bias (m/s)
Figure 9. Overwaterwind speed METSTAT results.
BOEM/EPA AERCOARE Alaska d03 Wind Direction Performance
ERA+FNMOC SST-all
"D O
IT ^
o
LU
c
g
o _
CM
* Aug
* Sep
9 Oct
*
jg
*
1
•20
Co
Sir
* 2010
* 2011
» 2012
-10
nple Conditbn^
* •
I !
0 10 20
Wind Direction Bias (deg)
Figure 10. Overwaterwind direction METSTAT results.
21
-------
4.3 Meteorological Time Series at Site C1: Chukchi-Burger
4.3.1 Wind speed
Time series plots of site C1 2011 and 2012 wind speed are shown in Figure 11 and Figure 12,
respectively. Note WRF wind speed is at 10 m while the observed wind speed was measured at
3.5 m. In general, the WRF simulations results followed the 2011 and 2012 trends in observed
wind speed. However, it might be expected that the observed wind speeds would be slightly
less on average given the lower measurement height. On the contrary, there was a tendency for
WRF to underpredict wind speed on average, by 0.5 m/s and 0.4 m/s over the 2011 and 2012
periods, respectively. The standard deviation of wind speed was 1.2 and 1.0 m/s for these
periods, respectively. The WRF negative wind speed bias corresponded to the same bias
indicated by the METSTAT analysis. For most near surface sources, lower wind speeds would
tend to result in overprediction of concentration by AERMOD.
4.3.2 Wind direction
Time series plots of site C1 2011 and 2012 wind direction are shown in Figure 13 and Figure
14, respectively. WRF wind directions compared very well to the measurements, as can be
seen in the figures. The average wind direction deviation for both periods was 5.1° and -3.6°,
respectively.
4.3.3 Air temperature
Time series plots of site C1 2011 and 2012 air temperature are shown in Figure 15 and Figure
16, respectively. Air temperature WRF extractions were at a height of 2.0 m above the surface
and the buoy probe measured temperature at 3.0 m, so some slight deviation in temperature
could be expected. WRF underpredicted temperatures for most of the 2011 season. The
average hourly temperature error was -0.85°C for 2011. The 2012 WRF average temperature
error was -0.36°C. In 2012, a period of significant deviation was evident from Sept. 19th to Sept.
22nd with WRF overpredicting temperature by 2-3°C.
4.3.4 Sea surface temperature
Time series plots of sited 2011 and 2012 SST are shown in Figure 17 and Figure 18,
respectively. The FNMOC SST analyses underpredicted SSTs for all of 2011, substantially so
for most of August. The 2012 SST analysis matched closely with measurements from mid-
September through October except for a short period in late September where the SST was
underpredicted. The SST analysis underpredicted SST from August to early September 2012,
by almost 6 °C at some points. The deviations may be due to errors within the FNMOC dataset
possibly due to cloud contamination or partial sea-ice cover. The period of August 13th -August
28th 2011 was the longest period of underpredicted SST and corresponded to a period of
22
-------
northeasterly winds (correctly simulated by WRF). The period of high SST error abruptly begins
with the shift from east to northeasterly winds and abruptly ends after a shift to more southerly
winds. Given the relatively high winds and constant RH before, during, and after this period, it is
less likely low cloud cover was the cause of the error. The northerly wind component and the
fact that the period was relatively early in the ice-free period suggests partial ice-cover may
have been to blame for the error.
Abrupt shifts to erroneously low FNMOC SST also occur for shorter periods throughout 2011
and 2012, not necessarily corresponding to any wind direction or shift in wind direction. Average
WRF deviations were -0.88°C and -0.97°C for 2011 and 2012, respectively.
4.3.5 Air-sea temperature difference
Time series plots of site C1 2011 and 2012 air-sea temperature difference (ASTD) are shown in
Figure 19 and Figure 20. In general, the "sign" of the WRF ASTD matched the observations for
most of 2011 and 2012. Only for a few short periods did WRF support a negative ASTD while
the observations supported a significant positive ASTD. The magnitude of the ASTD did vary
quite substantially over some periods, with an average root mean square error of about 1.0°C
for both periods. Hours with significant difference between WRF and measurement ASTD
(significant considered as an absolute value of 0.5°C or more difference) and opposite ASTD
sign occurred 13% and 6% of the time for 2011 and 2012, respectively.
4.3.6 Relative humidity
Time series plots of site C1 2011 and 2012 RH are shown in Figure 21 and Figure 22,
respectively. The 2011 WRF predictions tended to overpredict RH on average (+4% on
average), significantly so over periods of September where observed RH dropped into the 70-
80% range. It is unknown why WRF produces a mode of 94% RH values.
For 2012, the buoy consistently measured RH of 100% for extended periods. It is unknown why
similar periods were not evident in the 2011 dataset. Although WRF RH did peak at 100% for a
few hours in 2012, WRF did not favor the mode of 100% values evident in the measurement
dataset. Again, as in 2011, a mode of 94% humidity occurred in the WRF results.
In 2011, two periods were identified where measured RH was relatively low and WRF RH was
high. From Sep. 2-9 and Sep. 16-24, measured RH ranged generally from 70 - 90 % while
WRF maintained RH around 94%. WRF ASTD predictions were favorable during this period and
both WRF and measurements supported unstable conditions.
4.3.7 PBL height
Time series plots of site C1 2011 and 2012 PBL heights are shown in Figure 23 and Figure 24,
respectively. Note no measurements of PBL were available at this site. The MMIF recalculated
PBL heights (RCALT) were of similar magnitude to the WRF RCALF PBL heights most of the
23
-------
time except for several periods when WRF PEL heights (RCALF) were at the minimum 25 m
height. The MMIF recalculation produced much higher PEL heights during many of these
periods. Notable cases included Oct. 1-7, 2011 where WRF PEL heights were sustained at 25
m through the period and MMIF recalculated heights ranged generally from 400 to 600 m.
Several notable cases in 2012 included Sept. 7-8 and Sept. 13-14, where WRF sustained the
minimum 25 m height for long periods and MMIF resulted in PEL heights ranging from 200 -
1000 m during these periods.
4.3.8 Discussion
The C1 WRF meteorology featured lower average wind speed than measured 2011 and 2012.
The underpredicted wind speed was a factor in conservative AERMOD predictions because
higher wind speeds generally result in lower concentrations downwind of the source. Both WRF
air temperature and SST values were biased cold on average. The sign and magnitude of ASTD
predicted by WRF consistently agreed with measured values for much of the periods modeled.
However, the magnitude of the predicted ASTD did vary substantially by up to 4 °C for several
short periods (about a day each instance) in late August and early September, 2012.
Differences in ASTD may affect the PEL scaling and mixing parameters. The ASTD differences
would likely result in PEL height differences also if measured PEL heights were available for this
site.
24
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug052011 Aug 12 2011 Aug 19 2011 Aug 26 2011 Sep 02 2011 Sep 09 2011 Sep 16 2011 Sep 23 2011 Sep 30 2011 Oct 07 2011
Figure 11. C1 2011 wind speed time series.
25
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug 22 2012 Aug 29 2012 Sep 05 2012 Sep 12 2012 Sep 19 2012 Sep 26 2012 Oct 03 2012
Figure 12. C1 2012 wind speed time series.
Oct 10 2012
26
-------
360
330
O Site Measurement
A WRF-MMIF (RCALF)
Aug052011 Aug 12 2011 Aug 19 2011 Aug 26 2011 Sep 02 2011 Sep 09 2011 Sep 16 2011 Sep 23 2011 Sep 30 2011 Oct 07 2011
Figure 13. C1 2011 wind direction time series.
27
-------
360
330
O Site Measurement
A WRF-MMIF (RCALF)
Aug 22 2012 Aug 29 2012 Sep 05 2012 Sep 12 2012 Sep 19 2012 Sep 26 2012 Oct 03 2012
Figure 14. C1 2012 wind direction time series.
Oct 10 2012
28
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug052011 Aug 12 2011 Aug 19 2011 Aug 26 2011 Sep 02 2011 Sep 09 2011 Sep 16 2011 Sep 23 2011 Sep 30 2011 Oct 07 2011
Figure 15. C1 2011 air temperature time series.
29
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug 22 2012 Aug 29 2012 Sep 05 2012 Sep 12 2012 Sep 19 2012 Sep 26 2012 Oct 03 2012
Figure 16. C1 2012 air temperature time series.
Oct 10 2012
30
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug052011 Aug 12 2011 Aug 19 2011 Aug 26 2011 Sep 02 2011 Sep 09 2011 Sep 16 2011 Sep 23 2011 Sep 30 2011 Oct 07 2011
Figure 17. C1 2011 sea-surface temperature time series.
31
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug 27 2012
Sep032012 Sepl02012 Sep 17 2012 Sep 24 2012 Oct 01 2012
Figure 18. C1 2012 sea-surface temperature time series.
32
-------
O Site Measurement
A WRF-MMIF (RCALF)
' Aug 05 2011 Aug 12 2011 Aug 19 2011 Aug 26 2011 Sep 02 2011 Sep 09 2011 Sep 16 2011 Sep 23 2011 Sep 30 2011 Oct 07 2011
Figure 19. C1 2011 air-sea temperature difference time series.
33
-------
12-
10
O Site Measurement
A WRF-MMIF (RCALF)
u
Aug 22 2012 Aug 29 2012 Sep 05 2012 Sep 12 2012 Sep 19 2012 Sep 26 2012 Oct 03 2012
Figure 20. C1 2012 air-sea temperature difference time series.
Oct 102012
34
-------
100
90
80
O Site Measurement
A WRF-MMIF (RCALF)
Aug052011 Aug 12 2011 Aug 19 2011 Aug 26 2011 Sep 02 2011 Sep 09 2011 Sep 16 2011 Sep 23 2011 Sep 30 2011 Oct072011
Figure 21. C1 2011 relative humidity time series.
35
-------
100
90
Ss 80
Aug 22 2012 Aug 29 2012
5ep 05 2012 Sep 12 2012 Sep 19 2012 Sep 26 2012 Oct 03 2012
Figure 22. C1 2012 relative humidity time series.
O Site Measurement
A WRF-MMIF (RCALF)
Oct 102012
36
-------
A WRF-MMIF (RCALF)
D WRF-MMIF (RCALT)
Aug 05 2011 Aug 12 2011 Aug 19 2011 Aug 26 2011 Sep 02 2011 Sep 09 2011 Sep 16 2011 Sep 23 2011 Sep 30 2011 Oct 07 2011
Figure 23. C1 2011 PBL height time series.
37
-------
A WRF-MMIF (RCALF)
D WRF-MMIF (RCALT)
Aug 22 2012 Aug 29 2012 Sep 05 2012 Sep 12 2012 Sep 19 2012 Sep 26 2012 Oct 03 2012
Figure 24. C1 2012 PBL height time series.
Oct 10 2012
38
-------
4.4 Site C2: Chukchi-Klondike
4.4.1 Wind speed
Time series plots of site C2 during 2010 and 2012 are shown in Figure 25 and Figure 26,
respectively. WRF wind speed predictions resulted in average error of 0.1 m/s and -0.2 m/s and
RMSEof 1.1 and 1.0 m/s for the two periods, respectively. Qualitatively, the figures show WRF
predicted wind speed quite well. The negative bias of 2012 was caused mainly by short periods
of underpredicted wind speed, generally -1 to -2 m/s, occurring Sep. 1st - 3rd, Sep. 9th - 11th, and
Sep. 20th to 22nd. However, given WRF wind speed was calculated at 10 m height and the
observed wind speed was at 3.5 m, the WRF wind speed should have been biased high in
comparison to the measurements.
4.4.2 Wind direction
Time series plots of site C2 during 2010 and 2012 are shown in Figure 27 and Figure 28,
respectively. Average wind direction error for 2010 was +0.7° and RMSE was 19.5°. Average
2012 wind direction error was +10° and RMSE was 16.2°.
4.4.3 Air temperature
Time series plots of site C2 during 2010 and 2012 are shown in Figure 29 and Figure 30,
respectively. The WRF simulation results followed the temperature magnitude and trends
closely, but periods of significant deviation were evident. WRF overpredicted temperature by 2-
4°C early August 2010 and by 1 - 2 °C short periods in late September 2010 and 2012. WRF
also underpredicted temperature in 2010 by 1 - 2 °C Aug 8-14 and Sep. 12-19. Overall, the
WRF simulation average deviation was +0.1 °C and +0.2 °C for 2010 and 2012, respectively.
Given the high frequency of stable conditions, the temperature bias may have been an artifact
of the differences in height: WRF temperature was extracted at 2.0 m and the measurement
height was 3.0 m.
4.4.4 Sea surface temperature
Time series plots of site C2 for 2010 and 2012 are shown in Figure 31 and Figure 32,
respectively. The early 2010 FNMOC SST analysis estimates were different than the
observations. The SST estimates improved in September 2010, but were biased about +1°C for
the rest of the season. The FNMOC analysis data matched the measurements closely with very
little error all of 2012. Overall, the average 2010 bias was -0.06°C and 0.02°C for 2012.
4.4.5 Air-sea temperature difference
Time series plots of site C2 during 2010 and 2012 are shown in Figure 33 and Figure 34,
respectively. WRF simulated the correct sign of the ASTD for most of the simulated time for
both the 2010 and 2012 periods. Hours with significant difference between WRF and
measurement ASTD (significant considered as an absolute value of 0.5°C or more difference)
and opposite ASTD sign occurred 15% and 11% of the time for 2010 and 2012, respectively.
39
-------
A significant deviation occurred August 21-22, 2010 when WRF predicted a positive +1°C ASTD
while the measurements indicated a negative difference of about -2°C. In 2010, WRF tended to
predict a lower magnitude of ASTD more often than the measurements until late September. In
2012, WRF ASTD error was low except for short periods Sep. 1st - 3rd and Sep. 21st, where
WRF predicted ASTD near 0 °C and measurements indicated an ASTD of -2 to -4 °C. . WRF
average ASTD bias was -0.13° in 2010 and +0.16° in 2012.
4.4.6 Relative humidity
Time series plots of site C2 during 2010 and 2012 are shown in Figure 35 and Figure 36,
respectively. The 2010 WRF RH results were quite accurate overall, with an average error of
+1.3%. WRF RH was biased high on average during the low-RH period occurs late September
and early October. In 2012, WRF consistently produces an RH of 94% and rarely produces an
RH above 94%, while measurements exceed 94% in these periods. WRF overpredicts RH
during periods where measured RH dips into the 80 - 90 % range. Overall, the 2012 WRF bias
is +0.43%.
4.4.7 PBL height
Time series plots of site C2 PBL height for 2010 and 2012 are shown in Figure 37 and Figure
38, respectively. There is no set of measurements to compare to WRF predictions, so the plots
contain only WRF (RCALF) and MMIF PBL height (RCALT) predictions. In 2010, minimum PBL
heights of 25 m occur most of the time, as predicted by WRF. The MMIF rediagnosis results in
the same minimum 25 m PBL heights for most of these same hours. The maximum MMIF PBL
heights agree well in time and relative magnitude. In 2012, WRF predicts minimum PBL heights
most of the time also. The MMIF rediagnosis results in higher PBL heights of 30-100 m fora
majority of these periods. The timing and magnitude of the maximum PBL heights agree well
between WRF and MMIF.
4.4.8 Discussion
Overall, WRF period bias was relatively small for all variables. Wind speed was consistently
within 1 m/s of the measurements except for a few short periods. The sign and magnitude of
ASTD predicted by WRF was the same as the measurements for the majority of the hours
modeled. Given the relative accuracy of the WRF predictions at C2, it was assumed AERMOD
predictions using WRF and measured meteorology would be similar. WRF-MMIF PBL heights
are used for AERMOD modeling for both the WRF and measurement-based simulations.
40
-------
O Site Measurement
A WRF-MMIF (RCALF)
L_/3_i i i , ~ Q , /T« , 1
Aug012010 Aug082010 Aug 15 2010 Aug 22 2010 Aug 29 2010 Sep 05 2010 Sep 12 2010 Sep 19 2010 Sep 26 2010 Oct 03 2010
Figure 25. C2 2010 wind speed time series.
41
-------
Aug272012 Sep 03 2012 Sep 10 2012 Sep 17 2012 Sep 24 2012 Oct 01 2012
Figure 26. C2 2012 wind speed time series.
O Site Measurement
A WRF-MMIF (RCALF)
42
-------
360
330
O Site Measurement
A WRF-MMIF (RCALF)
Aug012010 Aug 08 2010 Aug 15 2010 Aug 22 2010 Aug 29 2010 Sep 05 2010 Sep 12 2010 Sep 19 2010 Sep 26 2010 Oct 03 2010
Figure 27. C2 2010 wind direction time series.
43
-------
360
330
O Site Measurement
A WRF-MMIF (RCALF)
Aug 27 2012 Sep 03 2012 Sep 10 2012 Sep 17 2012 Sep 24 2012 Oct 01 2012
Figure 28. C2 2012 wind direction time series.
44
-------
u
O Site Measurement
A WRF-MMIF (RCALF)
Aug012010 Aug082010 Aug 15 2010 Aug 22 2010 Aug 29 2010 Sep 05 2010 Sep 12 2010 Sep 19 2010 Sep 26 2010 Oct032010
Figure 29. C2 2010 air temperature time series.
45
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug 27 2012 Sep032012 Sep 10 2012 Sep 17 2012 Sep 24 2012 Oct 01 2012
Figure 30. C2 2012 air temperature time series.
-------
16
14
12
10
u
O Site Measurement
A WRF-MMIF (RCALF)
Aug012010 Aug082010 Aug 15 2010 Aug 22 2010 Aug 29 2010 Sep 05 2010 Sep 12 2010 Sep 19 2010 Sep 26 2010 Oct032010
Figure 31. C2 2010 sea surface temperature time series.
47
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug 27 2012
Sep032012 Sepl02012 Sep 17 2012 Sep 24 2012 Oct 01 2012
Figure 32. C2 2012 sea surface temperature time series.
48
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug 01 2010 Aug 08 2010 Aug 15 2010 Aug 22 2010 Aug 29 2010 Sep 05 2010 Sep 12 2010 Sep 19 2010 Sep 26 2010 Oct 03 2010
Figure 33. C2 2010 air-sea temperature difference time series.
49
-------
12
10
O Site Measurement
A WRF-MMIF (RCALF)
U
Aug 27 2012 Sep 03 2012 Sep 10 2012 Sep 17 2012 Sep 24 2012 Oct 01 2012
Figure 34. C2 2012 air-sea temperature difference time series.
50
-------
100
90
80
Site Measurement
WRF-MMIF (RCALF)
Aug 01 2010 Aug 08 2010 Aug 15 2010 Aug 22 2010 Aug 29 2010 Sep 05 2010 Sep 12 2010 Sep 19 2010 Sep 26 2010 Oct 03 2010
Figure 35. C2 2010 relative humidity time series.
51
-------
100
5s 80
Aug 27 2012
5ep03 2012
Sep 10 2012 5epl72012 Sep 24 2012 Oct 01 2012
Figure 36. C2 2012 relative humidity time series.
O Site Measurement
A WRF-MMIF (RCALF)
52
-------
A WRF-MMIF (RCALF)
D WRF-MMIF (RCALT)
Aug 01 2010 Aug 08 2010 Aug 15 2010 Aug 22 2010 Aug 29 2010 Sep 05 2010 Sep 12 2010 Sep 19 2010 Sep 26 2010 Oct 03 2010
Figure 37. C2 2010 PBL height time series.
53
-------
Aug272012 Sep 03 2012 Sep 10 2012 Sep 17 2012 Sep 24 2012 Oct 01 2012
Figure 38. C2 2012 PBL height time series.
A WRF-MMIF (RCALF)
D WRF-MMIF (RCALT)
54
-------
4.5 Site B2: Reindeer Island
4.5.1 Wind speed
Time series plots of site B2 during 2010 and 2011 are shown in Figure 39 and Figure 40,
respectively. WRF underpredicted wind speed over portions of both seasons and rarely
overpredicted wind speed. As a result, the average wind speed was biased low. The average
wind speed deviation was -0.7 and -0.9 m/s for 2010 and 2011, respectively. The measurement
height was 10.7 m and the WRF wind speed was calculated at 10.0 m. As a result, a small
amount of negative bias would be expected. The negative bias at Site B2 correlated with the
regional negative bias found by the METSTAT analysis (note Reindeer Island data were not
included in the METSTAT database).
4.5.2 Wind direction
Time series plots of site B2 in 2010 and 2011 are shown in Figure 41 and Figure 42,
respectively. During both seasons, WRF wind directions correlated well with the observations,
based on qualitative comparison of the WRF data and site measurements. No extended period
of significant wind direction deviation was evident.
4.5.3 Air temperature
Time series plots of site B2 for 2010 and 2011 are shown in Figure 43 and Figure 44 ,
respectively. WRF consistently overpredicted air temperature for 2010 with a peak
overprediction of 5.2° occurring at the beginning of the record on August 18th. The average
temperature was 0.9° warmer than the observed temperature. Temperature predictions for 2011
were more accurate overall, resulting in an average temperature deviation of only 0.25°C.
However, significant deviations did occur for 2011. WRF overpredicted temperature by 2-5° on
September 1st and underpredicted temperature by 2-3° on August 11th.
4.5.4 Sea surface temperature
Time series plots of site B2 during 2010 and 2011 were shown in Figure 45 and Figure 46,
respectively. The FNMOC SST analysis data used by WRF were warmer than observed for
most of 2010 with the exception of a period of significant underprediction in mid-August. The
average 2010 deviation in SST was +0.28°C. For 2011, the SSTs provided to WRF were cooler
than observed, resulting in an average SST deviation of-0.63°C.
4.5.5 Air-sea temperature difference
Time series plots of site B2 for 2010 and 2011 are shown in Figure 47 and Figure 48,
respectively. For much of 2010 and 2011, the observed SST was near 0°C.This was
disadvantageous for atmospheric stability prediction because small errors in ASTD could lead to
an opposite sign of stability. However, ASTD magnitude is small and would likely tend to
support near-neutral stability conditions, regardless of sign. Hours with significant difference
55
-------
between WRF and measurement ASTD ('significant' considered as an absolute value of 0.5° or
more difference) and opposite ASTD sign occurred 28% and 48% of the time for 2010 and
2011, respectively.
Overall, the WRF ASTD was warmer than observed due to overpredicted air temperatures in
2010 and underpredicted water temperatures in 2011. The average ASTD deviation from the
observations was +0.62°C in 2010 and +0.88°C in 2011.
A significant period of opposite ASTD sign occurred from Aug. 22-25, 2010 as WRF predicted
positive ASTD while the measurements indicated negative ASTD. In this case, WRF supported
stable atmospheric conditions while the measurements supported unstable conditions. A
significant period of opposite ASTD sign also occurred from Aug. 26 - Sep. 7, 2012, where
WRF predicted an average ASTD of +0.5°C while the measurements indicated an average
ASTDof-0.9°C.
Periods of reversed heat flux appeared frequently in August 2011, but mainly because the
ASTD was near 0°C throughout the period. WRF consistently overpredicts ASTD by predicting
ASTD values near 0°C while the measurements average near -1°C.
4.5.6 Relative humidity
Time series plots of site B2 in 2010 and 2011 are shown in Figure 49 and Figure 50,
respectively. The observations revealed a nearly saturated surface layer for most of 2010 and
2011. The WRF predictions followed this trend closely and correctly simulated the drier periods
that occur late in 2010 and early in 2011.
4.5.7 PBL height
Time series plots of site B2 during 2010 and 2011 are shown in Figure 51 and Figure 52,
respectively. WRF tended to overpredict the frequency of highly stable conditions based on the
high frequency of minimum PBL heights evident for both 2010 and 2011, resulting in
underprediction of average PBL height. The average deviation of the WRF (RCALF) predictions
was -94 m and -46 m, respectively. WRF tended to overpredict PBL height during a portion of
the unstable periods, evident in late September 2010 and 2011. The MMIF recalculated PBL
heights were more accurate on average, but still excessively low, resulting in an average bias of
-62 m and -23 m for 2010 and 2011, respectively.
High PBL error did occur during the significant period of opposite sign ASTD identified in
Section 5.5.5 from Aug. 22 - 25 and Aug. 30 - Sep. 1, 2010. Both WRF (RCALF) and MMIF
(RCALT) underpredict PBL height during these periods. The primary cause of these differences
is the incorrect sign of ASTD which resulted in the prediction of stable atmospheric conditions
during an unstable period.
PBL height error was as prevalent in 2011 as 2010 as both WRF (RCALF) and MMIF (RCALT)
predicted minimum PBL heights of 25 m during extended periods with measured PBL heights of
50 -200 m. Again, incorrect ASTD sign is to blame. WRF consistently predicts slightly positive
56
-------
ASTD, supporting stable conditions and the measurements support unstable conditions given
the negative ASTD.
4.5.8 Discussion
WRF temperatures support stable conditions through much of 2010 and 2011, while
measurements support unstable conditions. This results in significant underprediction of PEL
height for large portions of 2010 and 2011. Far-source (> 1,000m from source to receptors)
AERMOD results using WRF-derived PEL heights would likely be conservative overall due to
the persistence of minimum 25 m heights during stable periods. The persistence of stable
conditions predicted by WRF in 2011 due to ASTD error likely would result in more conservative
long-term average concentrations in the mid- and far-source. It is unlikely the PEL height bias
would affect short term near-source maximum concentrations since these will tend to occur
during unstable conditions. WRF overpredicted PEL height during unstable conditions on
average, suggesting maximum near-source concentrations could be conservative. The
consistent underprediction of wind speed could also result in overprediction of concentrations by
AERMOD at both the near-source and far-source receptors.
57
-------
Aug 22 2010 Aug 29 2010 Sep 05 2010 Sep 12 2010 Sep 19 2010
Figure 39. B2 2010 wind speed time series.
O Site Measurement
A WRF-MMIF (RCALF)
58
-------
Aug 05 2011 Aug 12 2011 Aug 19 2011 Aug 26 2011 Sep 02 2011 Sep 09 2011
Figure 40. B2 2011 wind speed time series.
O Site Measurement
A WRF-MMIF (RCALF)
Sep 16 2011
-------
360
330
Aug 22 2010
Aug 29 2010 Sep 05 2010 Sep 12 2010 Sep 19 2010
Figure 41. B2 2010 wind direction time series.
O Site Measurement
A WRF-MMIF (RCALF)
60
-------
360
330
O Site Measurement
A WRF-MMIF (RCALF)
Aug 05 2011 Aug 12 2011 Aug 19 2011 Aug 26 2011 Sep 02 2011 Sep 09 2011
Figure 42. B2 2011 wind direction time series.
Sep 16 2011
61
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug 22 2010
Aug 29 2010 Sep 05 2010 Sep 12 2010 Sep 19 2010
Figure 43. B2 2010 air temperature time series.
62
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug 05 2011 Augl22011 Aug 19 2011 Aug 26 2011 Sep022011 Sep 09 2011
Figure 44. B2 2011 air temperature time series.
Sep 16 2011
63
-------
-2
-4
-6
O Site Measurement
A WRF-MMIF (RCALF)
Aug 22 2010
Aug 29 2010 Sep 05 2010 Sep 12 2010 Sep 19 2010
Figure 45. B2 2010 sea surface temperature time series.
64
-------
-2
-4
-6
O Site Measurement
A WRF-MMIF (RCALF)
Aug 05 2011 Augl22011 Aug 19 2011 Aug 26 2011 Sep022011 Sep 09 2011 Sep 16 2011
Figure 46. B2 2011 sea surface temperature time series.
65
-------
12-
10
U
-6
-8
-10
-12L
O Site Measurement
A WRF-MMIF (RCALF)
Aug 22 2010
Aug 29 2010 Sep 05 2010 Sep 12 2010 Sep 19 2010
Figure 47. B2 2010 air-sea temperature difference time series.
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug 05 2011 Aug 12 2011 Aug 19 2011 Aug 26 2011 Sep 02 2011 Sep 09 2011 Sep 16 2011
Figure 48. B2 2011 air-sea temperature difference time series.
67
-------
100 F
90
Ss 80
70
60 L
Aug 22 2010
Aug 29 2010 Sep 05 2010 Sep 12 2010 Sep 19 2010
Figure 49. B2 2010 relative humidity time series.
O Site Measurement
A WRF-MMIF (RCALF)
-------
100
O Site Measurement
A WRF-MMIF (RCALF)
Aug 05 2011 Augl22011 Aug 19 2011 Aug 26 2011 Sep 02 2011 5ep 09 2011
Figure 50. B2 2011 relative humidity time series.
Sep 16 2011
-------
O- OS) <3B> O
OOQSK) OS)
OGOED <303D
O Site Measurement
A WRF-MMIF (RCALF)
D WRF-MMIF (RCALT)
Aug 22 2010
Aug 29 2010 Sep 05 2010 Sep 12 2010 Sep 19 2010
Figure 51. B2 2010 PBL height time series.
70
-------
O Site Measurement
A WRF-MMIF (RCALF)
D WRF-MMIF (RCALT)
Aug 05 2011 Aug 12 2011 Aug 19 2011 Aug 26 2011 Sep 02 2011 Sep 09 2011
Figure 52. B2 2011 PBL height time series.
Sep 16 2011
71
-------
4.6 Site B3: Beaufort-Sivulliq
4.6.1 Wind speed
Time series plots of site B3 for 2010, 2011, and 2012 are shown in Figure 53, Figure 54, and
Figure 55, respectively. WRF predicted the wind speed trend accurately through 2010, with the
exception of periods of maximum wind speed where WRF slightly underpredicted wind speed.
WRF consistently underpredicted wind speeds through most of 2011 and 2012. The average
wind speed deviation was -0.42, -1.07, and -0.54 m/s for the three periods, respectively. The
negative bias was likely magnified by the fact WRF wind speeds were calculated at the 10 m
height and the buoy measurements were at 3.3 and 3.5 m height for 2010 and 2011/2012,
respectively. The negative bias at the measurement location corresponded with the negative
bias for wind speed found by the regional METSTAT analysis.
4.6.2 Wind direction
Time series plots of site B3 in 2010, 2011, and 2012 are shown in Figure 56, Figure 57, and
Figure 58, respectively. WRF appears to have accurately predicted wind direction for most
periods. The standard deviation was 24°, 19°, and 31° for the three periods, respectively. Again,
the wind direction bias would likely not affect the AERMOD results because the methodology
used in this study limits the influence of wind direction.
4.6.3 Air temperature
Time series plots of site B3 during 2010, 2011, and 2012 are shown in Figure 59, Figure 60, and
Figure 61, respectively. WRF overpredicted air temperature through August and September
2010 on average. The 2011 WRF predictions appear to match the magnitude and trend of
temperature well with the exception of a period of underprediction in early October. Periods of
underprediction were also evident in mid-September and early October 2012. Average
temperature deviations were +0.49°, -0.39°, and -0.57° for the three years, respectively.
4.6.4 Sea surface temperature
Time series plots of site B3 in 2010, 2011, and 2012 are shown in Figure 62, Figure 63, and
Figure 64 , respectively. Each year SST predictions are more erroneous at the beginning of the
open-water season and more accurate near the end of the season. The transition appears to
occur around mid-September. The average WRF SST biases were +0.55°, +0.33°, and -0.67°,
for 2010, 2011, and 2012, respectively.
4.6.5 Air-sea temperature difference
Time series plots of site B3 for 2010, 2011, and 2012 are shown in Figure 65, Figure 66, and
Figure 67, respectively. The WRF air temperature and SST warm biases in 2010 coincide,
resulting in relatively low error in ASTD. As seen in the figure, the magnitude, trend, and sign of
72
-------
heat flux in 2010 tracked well with the observations. ASTD error was greater in 2011 and
substantial periods of opposite ASTD sign were evident. In 2012, the ASTD trends and heat flux
sign were predicted well, but the magnitude of ASTD was underpredicted during several notable
periods.
Hours with significant difference between WRF and measurement ASTD (significant considered
as an absolute value of 0.5°C or more difference) and opposite ASTD sign occurred 11%, 35%,
and 11% of the time for 2010, 2011, and 2012, respectively. Significant periods of opposite
ASTD sign occurred in 2011, notably Sep. 21 - 27 and Oct. 15-20. During these periods the
site measurements predicted a positive ASTD while WRF predicted a negative ASTD. The
measurements supported stable atmospheric conditions while WRF results supported unstable
atmospheric conditions.
For 2012, WRF underpredicted ASTD from Sep. 17 - Sep. 21 and overpredicted ASTD from
Sep. 25 - Sep. 28. However, the sign of ASTD was predicted favorably during these periods.
Overall, WRF predictions favored a negative bias (underprediction), resulting in average ASTD
deviations of-0.06°C, -0.72°C, and +0.10°C, respectively. This bias was caused mainly by WRF
predicting ASTD near 0°C during periods in all three years where ASTD measurements were
+1-2°C.
4.6.6 Relative humidity
Time series plots of site B3 2010, 2011, and 2012 are shown in Figure 68, Figure 69, and
Figure 70, respectively.
WRF appears to predict the magnitude and trend of RH for 2010 and 2011 well. Average RH
error was +0.9% and +2.3% for 2010 and 2011, respectively. The 2011 time series plot reveals
WRF RH trends and magnitude followed the measurements quite well. However, WRF did
overpredict RH by 10 - 20% periodically during low-RH episodes in late October 2011,
corresponding to short periods of northeasterly winds (therefore, not associated with offshore
flows).
For 2012, the measurement dataset maintained an almost constant RH near 99% the entire
open water season. It is possible these data were erroneous either due to faulty equipment or a
problem with data transfer. Though WRF RH trends did correspond with the measurements, RH
remained high, greater than 90% for most of the season except for a few short periods where
RH dipped down to near 80%. These events corresponded to periods of offshore flow, with
south to southwesterly winds (180-210°). Southwesterly wind events were very rare in 2010 and
2011 and quite common in 2012. The average 2012 RH bias is -6.4%.
4.6.7 PBL height
Time series plots of site B3 in 2010, 2011, and 2012 are shown in Figure 71, Figure 72, and
Figure 73, respectively. Again, WRF PBL height estimates are low, favoring the minimum 25 m
height through most of the three years. The MMIF RCALT rediagnosis results in a lower
73
-------
frequency of minimum 25 m PEL heights. However, the average PEL height is still lower than
measured. Average PEL height deviations were -46, -113, and -31 m for the three years,
respectively. Average MMIF-recalculated PEL height deviations were -23, -27, and -14 m,
respectively.
In Section 4.6.5 several periods of significant differences in ASTD sign and magnitude were
identified for 2011, notably the Sep. 21 - 27, and Oct. 15-20 episodes where WRF supported
unstable atmospheric conditions while the measurements supported stable conditions. Despite
a positive ASTD, measured PEL heights are relatively high from 200 - 600 m during this period.
The unstable conditions in WRF result in PEL heights of similar magnitude, resulting in good
agreement with measurements.
For the Sep. 21 - 27, 2011 period the average measured PEL height was 350 m. WRF RCALF
and MMIF RCALT average PEL heights were 140 m and 250 m, respectively. Measured PEL
height was relatively high despite the positive ASTD of +1 to +2 °C that would tend to support
stable atmospheric conditions. It is possible PEL structure in the region was influenced more by
synoptic conditions than local forcing during this period.
For the Oct. 15 - 20, 2011 period the average measured PEL height was 330 m while average
WRF RCALF and RCALT PEL heights were 320 and 350 m, respectively. Again, relatively high
PEL heights were measured despite ASTD values that supported stable conditions. The
relatively high WRF PEL heights corresponded to the unstable conditions supported by negative
ASTD.
4.6.8 Discussion
The tendency for WRF to underpredict wind speed at this site should contribute to higher
AERMOD concentrations. However, the wind speeds were biased low most often during peak
winds, so it is unknown whether the bias would have an influence on the maximum short term
average concentrations. Both the WRF and MMIF (RCALT and RCALF) PEL height predictions
were consistently underpredicted during stable periods. Measured PEL heights were less than
50 m on brief occasions, while WRF and MMIF supported extensive periods of minimum 25 m
PEL height. The tendency of WRF to underpredict PEL height may lead to concentration
overpredictions by AERMOD. Near-source short term maximum concentrations predicted using
WRF meteorology may be more accurate (in terms of comparison to predictions using
measured meteorology) because PEL heights were more accurate during unstable periods. The
influence of these biases on AERMOD concentration predictions are reviewed in Section 6.0.
74
-------
O Site Measurement
A WRF-MMIF (RCALF)
Augl92010 Aug262010 Sep 02 2010 Sep 09 2010 Sep 16 2010 Sep 23 2010 Sep 30 2010 Oct 07 2010
Figure 53. B3 2010 wind speed time series.
75
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug 14 2011
Aug282011 Sepll2011 Sep 25 2011 Oct 09 2011
Figure 54. B3 2011 wind speed time series.
Oct 23 2011
76
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug2l 2012
Aug 28 2012
Sep042012 Sep 11 2012 Sep 18 2012 Sep 25 2012
Figure 55. B3 2012 wind speed time series.
Oct02 2012
77
-------
360
330
O Site Measurement
A WRF-MMIF (RCALF)
Aug 19 2010 Aug 26 2010 Sep 02 2010 Sep 09 2010 Sep 16 2010 Sep 23 2010 Sep 30 2010 Oct 07 2010
Figure 56. B3 2010 wind direction time series.
78
-------
360
330
Augl42011 Aug282011 Sep 11 2011 Sep 25 2011 Oct092011
Figure 57. B3 2011 wind direction time series.
O Site Measurement
A WRF-MMIF (RCALF)
Oct 23 2011
79
-------
360
330
Aug 21 2012 Aug 28 2012 Sep 04 2012 Sep 11 2012 Sep 18 2012 Sep 25 2012
Figure 58. B3 2012 wind direction time series.
Oct 02 2012
O Site Measurement
A WRF-MMIF (RCALF)
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug 19 2010 Aug 26 2010 Sep 02 2010 Sep 09 2010 Sep 16 2010 Sep 23 2010 Sep 30 2010
Figure 59. B3 2010 air temperature time series.
Oct 07 2010
81
-------
O Site Measurement
A WRF-MMIF (RCALF)
Augl42011 Aug282011 Sep 11 2011 Sep 25 2011 Oct092011
Figure 60. B3 2011 air temperature time series.
Oct232011
-------
O Site Measurement
A WRF-MMIF (RCALF)
i\ug212012 Aug282012 Sep 04 2012 Sep 11 2012 Sep 18 2012 Sep 25 2012
Figure 61. B3 2012 air temperature time series.
Oct 02 2012
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug 19 2010 Aug 26 2010 Sep 02 2010 Sep 09 2010 Sep 16 2010 Sep 23 2010 Sep 30 2010 Oct 07 2010
Figure 62. B3 2010 sea surface temperature time series.
84
-------
Augl42011 Aug 28 2011 Sep 11 2011 Sep 25 2011 Oct 09 2011
Figure 63. B3 2011 sea surface temperature time series.
O Site Measurement
A WRF-MMIF (RCALF)
Oct 23 2011
-------
16-
14
12
u
O Site Measurement
A WRF-MMIF (RCALF)
Hug 21 2012
Aug282012
Sep042012 Sep 11 2012 Sep 18 2012 Sep 25 2012 Oct 02 2012
Figure 64. B3 2012 sea surface temperature time series.
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug 19 2010 Aug 26 2010 Sep 02 2010 Sep 09 2010 Sep 16 2010 Sep 23 2010 Sep 30 2010 Oct072010
Figure 65. B3 2010 air-sea temperature difference time series.
87
-------
O Site Measurement
A WRF-MMIF (RCALF)
Aug 14 2011 Aug 28 2011 Sep 11 2011 Sep 25 2011 Oct 09 2011 Oct 23 2011
Figure 66. B3 2011 air-sea temperature difference time series.
-------
O Site Measurement
A WRF-MMIF (RCALF)
21 2012 Aug 28 2012 Sep 04 2012 Sep 11 2012 Sep 18 2012 Sep 25 2012 Oct 02 2012
Figure 67. B3 2012 air-sea temperature difference time series.
-------
100
Ss 80
Aug 19 2010 Aug 26 2010 Sep 02 2010 Sep 09 2010 Sep 16 2010 Sep 23 2010 Sep 30 2010 Oct 07 2010
Figure 68. B3 2010 relative humidity time series.
O Site Measurement
A WRF-MMIF (RCALF)
90
-------
100
Aug 14 2011 Aug282011 Sep 11 2011 Sep 25 2011 Oct092011
Figure 69. B3 2011 relative humidity time series.
O Site Measurement
A WRF-MMIF (RCALF)
Oct232011
91
-------
100
-------
O Site Measurement
A WRF-MMIF (RCALF)
D WRF-MMIF (RCALT)
Aug 19 2010 Aug 26 2010 Sep 02 2010 Sep 09 2010 Sep 16 2010 Sep 23 2010 Sep 30 2010
Figure 71. B3 2010 PBL height time series.
Oct 07 2010
93
-------
Aug 13 2011 Aug 27 2011 Sep 10 2011 Sep 24 2011 Oct082011
Figure 72. B3 2011 PBL height time series.
O Site Measurement
A WRF-MMIF (RCALF)
D WRF-MMIF (RCALT)
Oct 22 2011
94
-------
CXECXOO CO O
OSD OOO
O Site Measurement
A WRF-MMIF (RCALF)
D WRF-MMIF (RCALT)
1212012 Aug282012 Sep 04 2012 Sep 11 2012 Sep 18 2012 Sep 25 2012
Figure 73. B3 2012 PBL height time series.
Oct 02 2012
95
-------
4.7 Summary
The statistics for all meteorological values analyzed in this section are listed in Table 5. Overall,
it is evident WRF tends to underpredict wind speed, especially at site B3 given B3 winds were
measured at 3.5 m and WRF winds were extracted at 10 m. SST was consistently overpredicted
at site B2. WRF simulations resulted in less ASTD sign error for the Chukchi Sea locations than
the Beaufort Sea locations, primarily because ASTD at the Beaufort Sea locations was near 0°C
for a majority of the open-water seasons.
PEL Heights were overpredicted overall at site B3. Predictions of PEL height at site B2 were the
most favorable, resulting in relatively low error and bias.
Table 5. Summary statistics for WRF meteorology extractions.
Parameter
Wind speed
Wind
Direction
Air Temp.
Site
C1
C1
C2
C2
B2
B2
B3
B3
B3
C1
C1
C2
C2
B2
B2
B3
B3
B3
C1
C1
C2
C2
B2
B2
Year
2011
2012
2010
2012
2010
2011
2010
2011
2012
2011
2012
2010
2012
2010
2011
2010
2011
2012
2011
2012
2010
2012
2010
2011
Hours
1591
1266
1612
1067
912
1224
1391
811
1145
1591
1266
1612
1067
912
1224
1391
811
1145
1591
1266
1612
1067
912
1224
Obs.
Mean
6.97
6.82
6.04
6.08
4.69
5.65
5.51
6.29
5.33
108
131
148
157
146
118
150
110
172
4.76
2.28
3.37
0.52
2.21
4.12
WRF
Mean
6.46
6.37
6.11
5.89
3.98
4.78
5.10
5.19
4.79
106
149
151
159
116
110
141
105
164
3.91
1.92
3.30
0.70
3.11
4.37
Bias
-0.51
-0.44
0.07
-0.19
-0.70
-0.87
-0.41
-1.10
-0.54
5.2
-16.5
0.7
10.0
3.6
0.8
4.5
1.7
-4.2
-0.85
-0.36
-0.08
0.18
0.90
0.25
Std. Error
1.17
1.00
1.06
0.97
1.24
1.38
1.14
1.75
1.32
17.7
23.6
19.5
16.2
25.4
21.3
24.5
19.2
31.1
1.04
1.01
0.92
0.63
1.22
0.95
96
-------
Table 5, continued. Summary statistics for WRF meteorology extractions.
Parameter
Air Temp.
SST
ASTD
RH
PBL Height -
(WRF
RCALF
comparison) -
Site
B3
B3
B3
C1
C1
C2
C2
B2
B2
B3
B3
B3
C1
C1
C2
C2
B2
B2
B3
B3
B3
C1
C1
C2
C2
B2
B2
B3
B3
B3
C1
C1
C2
C2
B2
B2
B3
B3
Year
2010
2011
2012
2011
2012
2010
2012
2010
2011
2010
2011
2012
2011
2012
2010
2012
2010
2011
2010
2011
2012
2011
2012
2010
2012
2010
2011
2010
2011
2012
2011
2012
2010
2012
2010
2011
2010
2011
Hours
1391
811
1145
1591
1266
1612
1067
912
1224
1391
811
1145
1591
1266
1612
1067
912
1224
1391
811
1145
1591
1266
1612
1067
912
1224
1391
811
1145
1591
1266
1612
1067
912
1224
1391
811
Obs.
Mean
1.40
-1.02
4.61
6.60
4.50
4.04
1.36
2.39
4.7
1.47
-0.53
4.77
-1.83
-2.22
-0.67
-0.84
-0.18
-0.57
-0.06
-0.49
-0.16
88
95
93
93
98
97
93
90
99
NA
NA
NA
NA
143
176
194
305
WRF
Mean
1.89
-1.40
4.03
5.70
3.53
3.98
1.38
2.67
4.07
2.02
-0.20
4.10
-1.80
-1.60
-0.68
-0.68
0.44
0.31
-0.13
-1.20
-0.06
92
91
95
93
97
95
94
93
93
293
270
213
195
49
94
147
193
Bias
0.49
-0.39
-0.57
-0.88
-0.97
-0.06
0.02
0.28
-0.63
0.55
0.33
-0.67
0.03
0.62
-0.01
0.16
0.62
0.88
-0.06
-0.71
0.1
3.8
-3.7
1.3
0.4
-1.1
-1.9
0.9
2.34
-6.4
NA
NA
NA
NA
-94
-82
-46
-113
Std. Error
1.04
0.70
0.93
0.90
1.12
3.98
1.38
0.96
1.21
0.93
0.58
0.96
0.86
1.14
0.86
0.82
1.16
1.25
0.77
1.01
0.82
6.1
6.4
3.3
4.1
2.5
3.3
3.5
5.0
6.5
NA
NA
NA
NA
98
115
124
216
97
-------
Table 5, continued. Summary statistics for WRF meteorology extractions.
Parameter Site Year Hours Obs. WRF Bias Std. Error
Mean Mean
B3 2012 1145 194 163 -31 143
C1 2011 1591 NA 397 NA NA
C1 2012 1266 NA 401 NA NA
PBL Height
(WRF-MMIF
RCALT
comparison)
C2
C2
B2
B2
2010
2012
2010
2011
1612
1067
912
1224
NA
NA
143
176
239
239
81
94
NA
NA
-62
-82
NA
NA
84
115
B3 2010 1391 194 171 -22 103
B3 2011 811 305 278 -27 177
B3 2012 1145 194 181 -14 151
98
-------
5 AERMOD SIMULATION METHODOLOGY
The AERMOD modeling system was developed as the next generation regulatory air quality
dispersion model, designed to incorporate state-of-the-art PEL parameterizations based on
Monin-Obukhov Similarity theory. Monin-Obukhov theory uses scaling factors based on the rate
of heat and momentum flux to describe the structure and evolution of the PEL. AERMOD
replaced the ISC modeling system that used Pasquill-Gifford stability classes and corresponding
lookup tables to estimate dispersion scaling parameters. AERMOD requires a complex set of
meteorological input to characterize the PEL structure and the turbulence parameters used to
estimate rates of dispersion. A full description of the formulas and parameterization schemes
used in AERMOD and its meteorological pre-processor AERMET can be found in USEPA
(2004a).
5.1 AERMOD Meteorological Input Files
AERMOD requires two meteorological files as input: a surface meteorological file (SFC file) and
a near-surface profile of temperature and wind profile file (PFL file). Each file contains a time
series of hourly-averaged meteorological variables. The PFL file need only include wind and
temperature information at a single height, but turbulence measurements and information at
additional heights may improve the accuracy of the simulation by providing a more complete
description of the atmospheric structure for AERMOD. The meteorological variables contained
in the SFC and PFL files relevant to this study are listed and described in Table 6.
Table 6. AERMOD meteorology fields.
Meteorological
variable
Units3 Abbreviation
Description
Sensible heat flux
(Surface)
Friction velocity
The rate of heat transfer to the atmosphere from the ground,
W/m2 H positive H the ground is heating the PBL, negative H the ground is
cooling the PBL.
Characteristic velocity scaling factor used to describe the rate of
transfer of energy from atmospheric momentum to the surface
* through turbulent motions. Mechanical turbulence and the rate of
pollutant dispersion in the PBL is a function of u*.
m/s
Convective scaling
velocity
m/s
Characteristic vertical velocity scaling factor used to describe the
transfer of momentum due to convective processes in the PBL. It is
used to estimate turbulence and corresponding rates of dispersion
in the convective PBL.
Vertical potential
temperature
gradient above the
mixed layer
K/m
d@/dz
Used in convective conditions only: describes the gradient of
potential temperature (temperature a parcel of air would have at
sea-level pressure) at the interfacial layer above the well-mixed
layer. This value specifies the "strength" of the top of the well-
mixed layer for describing the fraction of plume penetration into
and above the interfacial layer.
99
-------
Table 6, continued. AERMOD meteorology fields.
Meteorological
variable
PEL height
potential under
convective
processes
PEL height
potential under
mechanical
processes
Monin-Obukhov
length
Surface roughness
length
Bowen ratio
Albedo
Wind speed
Wind direction
Temperature
Standard deviation
of wind direction
Standard deviation
of vertical wind
speed
Units3 Abbreviation Description
m z Depth of the mixed layer possible under convective forcing.
AERMOD uses the maximum of zic orzimforthe PEL height.
m Zim Depth of the mixed layer possible under mechanical forcing.
m AERMOD uses the maximum of zic orzimforthe PEL height.
The fundamental scaling parameter of Monin-Obukhov similarity
theory that is used to define the influence of buoyancy-induced and
mechanical turbulence on the structure of the surface layer of the
atmosphere. In stable conditions it can be considered as the
, relative height at which buoyant production of turbulent energy is
equal to that produced by mechanical/wind-shear processes.
A negative L indicates unstable, convective conditions while a
positive L indicates stable conditions. Large absolute value of L is
indicative of neutral conditions while small absolute values of L are
indicative of strongly stable or unstable conditions.
A scaling parameter used to describe the influence of ground
surface friction on the structure of the PEL. Values of z0 over the
ocean are very low (10"5- 10"3m) and are a function of wave height
(Arya, 1988).
The ratio of sensible heat flux to latent heat flux from the ground.
Values > 1 occur in drier conditions when most heat flux is in the
form of sensible heat. Values < 1 occur in moist conditions, when
B sufficient surface moisture is available for evaporation. In marine
environments, the Bowen ratio is always small. It is used by
AERMET to estimate sensible and latent heat fluxes during
unstable conditions and passed thru to AERMOD to estimate the
deposition of gases.
The fraction of total incident solar radiation reflected by the earth's
r surface. It is used by AERMET to estimate the surface radiation
balance but only used by AERMOD for the deposition of gases.
The average scalar wind speed at a specified measurement height.
m/s ws Typical measurement height is 10 m (the meteorological file
provides a column to specify measurement height), but may be as
low as a few meters on meteorological buoys.
degrees WD The average wind direction at a specified measurement height.
T or The average atmospheric temperature at a specified measurement
K O (absolute height, typically 2 m (the meteorological file provides a column to
temp.) specify measurement height).
decrees a (Provided in the PFL file only). Standard deviation of the wind
direction during the period.
m/s a (Provided in the PFL file only). Standard deviation of the vertical
wind speed during the period.
'meters (m), seconds (s), watts (W), kelvin (K).
100
-------
The meteorology files used by AERMOD are typically built by a preprocessor program.
AERMET is the accepted preprocessor for regulatory modeling of land-based air pollutant
sources. MMIF and AERCOARE are the alternative pre-processors examined in this study. The
MMIF program provides a method to extract WRF data directly from the WRF output files and
create SFC and PFL meteorology files for AERMOD. It uses the fields available in WRF to
estimate the meteorological variables listed in Table 6 or to create input files for AERCOARE.
Meteorological preprocessors produce the AERMOD meteorological parameters from a set of
raw meteorological measurements or model data: wind speed, wind direction, temperature,
solar radiation, differential temperature, humidity, cloud cover, and atmospheric pressure are
typical inputs to the pre-processor. The quality of the AERMOD meteorological input is therefore
highly dependent on the representativeness of the raw meteorology fed to the preprocessor.
AERCOARE requires overwater ASTD, RH, and WS to characterize the surface layer energy
fluxes. The resulting static stability of the overwater atmospheric PEL is highly dependent on the
"sign" of the ASTD, especially during light to moderate winds speeds. RH is important because
it is used to determine the rate of sensible heat flux from the sea to the PEL. Lower RH
promotes greater latent heat flux from the sea to the PEL and less sensible heat flux. Therefore,
the rate of PEL heating is reduced with lower RH as more energy is invested into sea-surface
evaporation.
5.2 Overwater Measurement Datasets and AERCOARE Processing
Meteorological datasets collected at four overwater sites from 2010-2012 were selected for this
study (Air Sciences Inc., 2010) (AECOM Environment, 2009). Meteorological data collected at
these sites were sufficient to provide inputs for AERCOARE. Two of the sites, "Burger" and
"Klondike," are the locations of meteorological buoys that collected data during open-water
periods on the Chukchi Sea. The other two sites, "Sivulliq" and "Reindeer Island," are the
locations of a meteorological buoy and an island-based meteorological station that collected
measurements during open-water periods on the Beaufort Sea. The temperature profiler dataset
collected on Endeavor Island was used in conjunction with the buoy data to estimate PEL
heights at the Beaufort sites. The temperature profiler was operated as part of a PSD-quality
meteorological monitoring campaign conducted at Endeavor Island for Shell Offshore, Inc.
(SLR, Inc., 2011). The sites are shown in Figure 74 and described in detail in Table 7.
101
-------
300-
200-
100-
0-
-100-
-200-
: Chukchi-Burger, 2010, 2012
Chukchi-Klondike,
—'•
B2:
Point Lay
# Buoy Site
•fr WRF grid point
extracted
• WRF grid points
I—-^
B3: Beaufort-Sivuliiq, 2010-2012
ENDV: Endeavor-Island, 2010-2012
••'
I
2100
2000
1900
1800
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
1
'-100
— elevation (m)
-500 -400
-300 -200
-100
100
200
300
400
Figure 74. Overwater meteorological measurement sites and corresponding WRF
inner-domain extraction points.
Table 7. Overwater measurement site details.
Site
Identifier a
C1
C2
B2
B3
Location
Site
Chukchi-
Burger
Chukchi-
Klondike
Beaufort-
Reindeer
Island
Beaufort-
Sivulliq
Lat/Lon
70.9 N,
165.3W
71. 5 N,
164.1 W
70.5 N,
148.3W
70.4 N,
146.0 W
Year
2011,
2012
2010,
2012
2010,
2011
2010,
2011,
2012
PBL height
method
from WRF
from WRF
critical bulk-
Richardson
layer
method
critical bulk-
Richardson
layer
method
Measurement height (m)
Air
Temp.
3.0
3.0
1.4
1.4
(2010)
3.0
Sea
Temp.
-1.2b
-1.2
-1.2
-1.2
Wind
3.5
3.5
10.7
3.3
(2010)
3.5
3 The site identifiers used in Task 1 and the Task 3 protocol of this study were retained. Note that B1 was a buoy site
in the Beaufort Sea that collected data in 2009 but not the 2010-2012 period of Task 3.
b Negative value indicates temperature measurement is below the sea surface.
AERCOARE requires hourly WS, WS, T, SST, RH, air pressure, oe (optionally), and PBL height
to create the overwater input meteorology for AERMOD.
102
-------
For the Reindeer Island 2011 dataset, the Endeavor Island RH measurements were used for
substitution because the RH sensor had missing or invalid data for most of the period. Although
the Endeavor Island measurement site was technically land-based, the RH was near saturation
(>90%) for most of the 2011 period and therefore would not have likely varied much from the
true RH at the buoy locations. The Endeavor Island monitoring program Quality Assurance
Project Plan (QAPP) (SLR, Inc., 2011) states the RH sensor height was selected to monitor RH
within the marine layer.
PEL height is a parameter used by AERMOD. It was estimated in this study using a method
described in Section 5.3 for the Beaufort Sea sites. No measurements were available for the
Chukchi Sea sites to estimate PEL height directly. Although simplistic parameterizations such
as the Venkatram (1980) method could be used to estimate PEL height at the Chukchi buoy
locations, it was deemed inappropriate for this study. The WRF-MMIF (RCALT) PEL heights
were used at the Chukchi locations in the AERMOD meteorology files. Therefore, PEL height
was only a factor in the comparison of AERMOD performance at the Beaufort Sea sites.
The temperature profiler dataset was collected using a Kipp & Zonen MTP-5 passive microwave
radiometer. The device was operated from 2010 to 2012 at the Endeavor Island facility near
Prudhoe Bay, Alaska. The profiler was oriented to collect estimates of temperature at 31 levels
from 0 to 1000 m above instrument height over the Beaufort Sea. Although the Endeavor Island
meteorological station was technically a land-based station, it was located on a small man-made
island located at the end of a thin peninsula that extends out about 6 kilometers (km) into the
Beaufort Sea.. The profiler site was located, with approval by EPA Region 10, in the most
practical site for measuring the marine boundary layer on the Beaufort Sea (SLR, Inc., 2011).
The profiler data and accompanying Endeavor Island meteorological tower dataset were
collected under an EPA-approved QAPP (SLR, Inc., 2011) satisfying PSD-quality data collection
requirements.
The Endeavor Island meteorological tower was deployed on a small man-made island that was
surrounded by the sea. Despite this, the wind, temperature, and humidity data measured at the
site may be influenced by nearby industrial structures and operations or the landscape of the
island depending on the meteorological conditions.
The Reindeer Island station was operated under an EPA-approved QAPP (AECOM
Environment, 2009) to provide PSD-quality meteorological data for air emissions permitting.
Reindeer Island is essentially a sandbar (about 250 m wide at its widest point and 3 km long)
located about 16 km north of Prudhoe Bay in the Beaufort Sea. The station was deployed at this
location to collect marine meteorological measurements during both ice and ice-free conditions.
The monitoring project specifically included a marine buoy located about 2 km south of the
Island to collect the data necessary for overwater dispersion modeling during the ice-free
periods.
AERCOARE processing was conducted using a PEL height of 600 m for the COARE
"gustiness" calculation (USEPA, 2012). Minimum PEL height and absolute value of L limits were
set to 25 m and 5 m, respectively, as described in Section 4.
103
-------
5.3 PBL Height Diagnosis Methods
Ideally, direct measurements of PBL height should be used for validation of WRF data and for
use in the observation-driven AERMOD simulations. However, direct PBL height measurements
were not available at the sites. The Endeavor Island temperature profiler data provided the
means for estimating PBL height at the Beaufort Sea sites. There was no equivalent dataset for
estimating PBL heights for the Chukchi Sea sites.
The 2010-2012 temperature profiler dataset provided the best means for diagnosing PBL height
at the Beaufort Sea sites. The temperature profile data can be used to estimate PBL height
using a method based on Richardson number theory. The Richardson number (Ri) is a measure
of the ratio of potential to kinetic energy in an atmospheric layer, represented by the vertical
stability and vertical wind shear, respectively. Small values of Ri indicate weak static stability or
strong wind shear, indicative of conditions where vertical mixing is prevalent. Large values of Ri
are indicative of a layer where the strength of the static stability is greater than the energy
provided by vertical wind shear. Vertical mixing is restrained in such conditions.
The bulk-Richardson number, Rib, is a form of Richardson number that provides an estimate of
the energy ratio over a layer between heights zi and z2. The Rib is calculated using Equation (7)
(Vogelezang & Holtslag, 1996):
R. =
(u(z2) - u(
where b\s a parameterization constant, recommended by Vogelezang and Holtslag (1996) to
be 100. The profiler does not measure wind speed so not all of the information needed to
estimate the Rib at each layer is provided. Using Monin-Obukhov similarity theory, profile of wind
speed can be estimated using values of u», z0, and L provided by AERCOARE:
(8)
where k\s the von Karman constant (0.4) and \\> is the stability correction function. Equation (8)
provides a means for estimating the wind shear between the two layers.
The height of the PBL can be assumed to be the height where Rib exceeds a critical value,
referred to as the critical Richardson number (7?/OTf). A Ricrit of 0.03 was used to diagnose the
PBL height, based on the recommendations of Gryning and Batchvarova (2003). It must be
emphasized this method relies on assumptions and parameterizations that may not result in
accurate PBL height estimates in all cases. WRF PBL heights that do not compare well to the
104
-------
"observed" PEL heights estimated using this method and may not necessarily indicate WRF
performance is poor.
It is unlikely the hourly profiler data were representative of conditions at the Chukchi Sea buoy
sites, given the sites were over 500 km away from the profiler location. Since there were no
alternative datasets in the immediate region, PEL height estimates for sites C1 and C2 could
only be made using an empirical parameterization scheme. However, a decision was made not
to use a parameterization scheme to estimate PEL heights at these locations. Initial
investigation of applicable parameterization schemes found no reliable method to produce
consistently accurate PEL height estimates. Instead, the WRF-MMIF PEL heights at the
Chukchi buoy locations were provided to AERMOD for modeling. Therefore, PEL height was
only a factor in the comparison of AERMOD performance at the Beaufort Sea sites.
5.4 Hypothetical Sources
EPA provided five unique source group configurations for this study (Wong, 2012). Each group
represented a hypothetical OCS source with stack characteristics typical of drill ship sources
that have operated on the OCS in the recent past or have been proposed in recent permit
applications for the Arctic as shown in Table 8. Each source contained multiple vertical stacks
with warm, buoyant plumes. Emissions from Source Group #1 and #2 are lowest to the ground,
with stacks averaging about 15 m in height. Emissions from Source Group #3 are concentrated
at a height of 23 m. Emissions from Source Group #4 are concentrated at a height of 27 m.
Given these values, Sources #1 and #2 can be considered representative of "short" stacks.
Source #4 can be considered representative of "tall stacks." Source #3 average stack height
falls within the average heights of the tall and short stack groups.
AERMOD can account for the influence of the wakes of structures on downwind concentrations
using the Prime Rise Model Enhancements (PRIME) algorithm (Schulman, et al., 2002).
Building/structure downwash effects were not considered for Source Groups #1-4. Source
Group #5 was comprised of the same stacks and parameters as Source Group #2, but with
structure downwash applied. A hypothetical set of structures based on typical OCS source
structure dimensions were provided for Source Group #5. The structure layout is shown in
Figure 75. All sources were located at the same central location with respect to the receptors
and structures (indicated as position 0 m, 0 m in the x- and y- coordinate system shown on
Figure 75).
105
-------
Table 8. Hypothetical source and stack parameters.
Source
Group
1
2
3
4
5
Unit
Diesel engine
Incinerator
Diesel engine
Boiler
Incinerator
Propulsion engine
Generator
Boiler
Diesel engine
Winch
Heater
Diesel engine
Boiler
Incinerator
Stack
height
(m)
16
14
18
17
10
25
20
15
39
25
23
18
17
10
Stack
temp.
(K)
700
550
680
500
525
570
610
420
580
580
510
680
500
525
Stack exit
velocity
(mis)
30
20
28
10
17
30
22
2
21
14
42
28
10
17
Stack
diameter
(m)
0.50
0.40
0.40
0.45
0.40
0.60
0.25
0.30
0.70
0.20
0.15
0.40
0.45
0.40
Downwash
effects
No
No
No
No
No
No
No
No
No
No
No
Yes
Yes
Yes
106
-------
100
BO-
60-
1m)
20-
D-
2-,-
K-
-B"
\
All Sources
-120
-ail
-60
n:;
(m)
Figure 75. Source locations and structures.
5.5 Receptor Grid
AERMOD predicts pollutant concentration at assigned receptors based on their distance from
sources and proximity to terrain features. For this study, a network of 50 receptor rings was
used. Each ring contained 360 receptors at 1° spacing. The rings were centered at the same
origin (origin of 0, 0 in the x- and y- directions, respectively, as shown in Figure 75) with
incremental radial spacing of downwind distance based on a geometric series from 30 m to 10
km. The geometric series equation was formulated using Equation (9):
Distance of ring N (m) = 30(m) *
(9)
where N is the number of the receptor ring. The value of 1.1259 was determined iteratively to fit
a geometric series of 50 rings spaced between 30 m and 10 km.
Receptors were placed at a height of 0.0 m (no flagpole receptors). The inner-most rings are
shown with respect to the structures and sources in Figure 76. The vessel to the south had no
107
-------
200m
-200 -150 -100 -50
(m)
Figure 76. Visual of inner-most receptor rings
source nor downwash influence on the drill rig sources. The vessel was considered part of the
ambient air.
5.6 WRF Meteorology Extraction and Processing
The variables listed in the SFC and PFL files can be calculated directly or indirectly by MMIF
using the fields available in the WRF output files. The three ways MMIF can create
meteorological files for AERMOD are:
a) Create onsite, upper air, and land use data and run AERMET
b) Create AERCOARE input files and run AERCOARE: AERCOARE produces SFC and
PFL files for input into AERMOD.
c) Create AERMET-like SFC and PFL files directly.
Method a would be inappropriate for overwater dispersion studies because AERMET is only
configured for overland meteorology. Methods b and c are both tested in this study. For
Method b, the WRF simulations provide the variables that might be measured by a buoy, ship,
or offshore platform. The meteorological fields from WRF were ported to AERCOARE.
AERCOARE produces the AERMOD SFC and PFL files using its specialized overwater
108
-------
algorithms. For Method c, MMIF passes through or calculates all variables directly from the
WRF output files. If not available from WRF output, MMIF estimates the similarity scaling
variables L and w» from Richardson-number methodology defined in Louis (1979). In the current
study, L is calculated and supplied directly by WRF under all scenarios simulated. The variable
w» is not calculated by WRF but calculated by MMIF.
Limits were placed on values of L and PEL height to prevent the occurrence of unphysical
extreme values in both the observation-based and WRF-based meteorological datasets, as
mentioned in Section 4. A minimum absolute value of 5.0 m was set as the limit for L in both
MMIF and AERCOARE. A minimum PEL height of 25 m was established for all datasets. The
PEL height and L limits used for both WRF- and observation- based simulations ensure the
most stable atmospheric conditions result in similar parameters. The limits on L are the same
limits used in the OCD model (Hanna, et al., 1985) (DiCristofaro & Hanna, 1989).
WRF estimates PEL height through the PEL parameterization scheme and each scheme uses a
different method to formulate PEL height. The WRF PEL heights are also fixed to the nearest
vertical grid cell center and can vary abruptly over small distances. MMIF can also rediagnose
the PEL height using the bulk-Richardson approach of Vogelezang and Holtslag (1996) as
described in Section 5.3. AERMOD simulations can be very sensitive to the PEL height
(Richmond & Morris, 2012). Therefore, MMIF-predicted PEL heights may provide significantly
different predicted concentrations than the PEL height used internally by WRF.
Given the options described above, four different MMIF extraction methods were tested:
1. MMIF was applied to extract and prepare data sets for direct use by AERMOD
(MMIF produces the AERMOD SFC and PFL input files directly). The PEL height
predicted by WRF is used in the SFC file.
2. As in Option 1), but the PEL height was rediagnosed from the wind speed and
potential temperature profiles using the Bulk-Richardson algorithm within MMIF.
3. MMIF was applied to extract the key meteorological variables of overwater wind
speed, wind direction, temperature, humidity, and PEL height from WRF results. The
MMIF extracted data were used to build an AERCOARE input file. AERCOARE used
these variables to predict the surface energy fluxes, surface roughness length and
other variables needed for the AERMOD simulations. For the current study,
AERCOARE was applied using the defaults recommended in the AERCOARE model
evaluations study (Richmond & Morris, 2012).
4. As in Option 3), but the PEL height was rediagnosed using the bulk-Richardson
algorithm within MMIF.
The naming convention and description of the four extraction methods are listed in Table 9.
Note "RCALT" refers to extractions with MMIF rediagnosis of PEL height and "RCALF" refers to
direct use of WRF PEL height (with the minimum 25 m PEL height applied). Also, note "AERC"
109
-------
refers to simulations with additional AERCOARE processing after MMIF extraction, and "MMIF"
refers to simulations using the WRF meteorology directly without further processing by
AERCOARE.
Table 9. WRF AERMOD meteorology extraction methods.
MH Process Path
Method
1 ) MMIF.RCALF WRF •*• MMIF -» AERMOD
2) MMIF.RCALT WRF -» MMIF (with PEL diagnosis) -» AERMOD
3) AERC.RCALF WRF -» MMIF -» AERCOARE -» AERMOD
4) AERC.RCALT WRF -» MMIF (with PEL diagnosis) -» AERCOARE -» AERMOD
MMIF identifies the nearest WRF grid point to the coordinates of the overwater measurement
site and extracts the data time series from this point (no interpolation between points). The
extraction points used in Task 3 were selected to correspond with the meteorological
measurement sites as closely as possible. The extraction points are shown in Figure 74.
5.7 AERMOD Evaluation Methodology
In this study, AERMOD concentrations were calculated using meteorology from overwater
observations and meteorology extracted from WRF simulations. The maximum predicted
concentrations at each receptor ring were extracted and the observation-driven AERMOD
results were compared directly to the WRF-driven AERMOD results. This approach simplifies
the investigation of the bias of the WRF simulations and removes the influence of wind direction
differences.
AERMOD version 14134 was used for this study, using all regulatory defaults except the
"VECTORWS" flag was used for WRF wind speed. The regulatory default for AERMOD version
14134 assumes scalar wind speed. AERMOD retains a non-regulatory algorithm that corrects
for the use of vector wind speeds, initiated using the "VECTORWS" flag. The algorithm,
described in USEPA (2004a) estimates the scalar wind from a vector wind using turbulent
fluctuations as an estimate or measurement of oe. This correction is most useful for low wind
speed conditions where the scalar and vector wind speeds are most likely to diverge. It was also
necessary to specify the "Beta" option in AERMOD to use the MMIF extracted data.
5.7.1 Simulation scenarios
The periods of available measurement data varied depending on the length of the open-water
season and buoy deployments. The open-water season is the period during the summer and
early autumn when the polar ice has melted and retreated northward enough to allow exposure
of the sea surface to the open air. A transitional period occurs between the ice- and ice-free
periods when a substantial portion of the sea surface is covered by broken ice. The buoys were
110
-------
typically deployed at the end of the transitional period to prevent damage to the buoys from
floating ice. Table 10 lists the periods of continuous overwater data extracted for this study. The
seasonal length varied from 811 hours at site B3 in 2011, to 1612 hours at site C2 in 2010. As
shown in Table 10, very few hours of calm wind occurred in the measurement- and extracted-
WRF datasets. The site B3 2011 observational dataset had the most hours of calm wind or
missing data, but such hours were still less than 3% of the dataset.
Five different averaging periods were simulated for each combination of site, year, and source
type: 1-hour, 3-hour, 8-hour, 24-hour, and period-long averaging periods. These averaging
times were selected because they correspond to the averaging periods applicable to the
NAAQS. To ensure tracer emission rate independence, AERMOD simulations were conducted
using a stack unit emission rate of 1 g/s. The resulting AERMOD concentrations were divided by
the tracer release rates to provide normalized concentration with units of us/m3.
This study involved a large number of AERMOD simulations. A total of 1,125 AERMOD
simulations were conducted to satisfy all of the possible scenarios:
• 3 site datasets per year (two years of data at sites C1, C2, and B2 and three years of
data at site B3),
• 3 years (2010-2012),
• 5 sources,
• 5 averaging periods,
• 5 meteorological datasets:
i) Observations
ii) MMIF.RCALF WRF extractions
iii) MMIF.RCALT WRF extractions
iv) AERC.RCALF WRF extractions
v) AERC.RCALT WRF extractions
The definitions the WRF extractions in ii) through v) are explained in Table 9.
A set of summary statistics were calculated for each simulation to evaluate the performance of
WRF-based simulations compared to the observation-based simulations, included in
Appendix B. All AERMOD input and output files and analysis plots are included in electronic
form on a data disk attached as Appendix C.
111
-------
Table 10. AERMOD simulation periods.
Site
Identifier
-i
f^O
DO
B3
Date range
Aug. z, zu I I , Hour z — uci.
7 9f"M -\ Mmir ft
Aug. lo, z(j\z, Hourzo —
Or-t "in OHIO Mniir 1 fi
jui. zo, zu ID, Hour lo — uct.
3on-l n \-\r\i ir 1 7
Aug. Zo, zu \z, Hour y — uci.
6on-IO Urn ir -1 Q
Aug. lo, zu iu, Hour i —
Qan lA Om n Mm ir O4
jui. ou, zu i i, Hour i — oep.
•1ft Om 1 Mrtiir lA
Aug. i^, zu iu, Hour i — uci.
-in on-in MC-UH-O^
oep. zi, zUi 1 , Hour o — uct.
O^ Om 1 Mrtiir-1^1
Ann -1 Q OO-1O Mrtiir -17
Total
Possible
Hours
1591
1266
1612
1067
912
1224
1391
811
1145
Scenario
Observations
MMIF.RCALF
MMIF.RCALT
AERC.RCALF
AERC.RCALT
Observations
MMIF.RCALF
MMIF.RCALT
AERC.RCALF
AERC.RCALT
Observations
MMIF.RCALF
MMIF.RCALT
AERC.RCALF
AERC.RCALT
Observations
MMIF.RCALF
MMIF.RCALT
AERC.RCALF
AERC.RCALT
Observations
MMIF.RCALF
MMIF.RCALT
AERC.RCALF
AERC.RCALT
Observations
MMIF.RCALF
MMIF.RCALT
AERC.RCALF
AERC.RCALT
Observations
MMIF.RCALF
MMIF.RCALT
AERC.RCALF
AERC.RCALT
Observations
MMIF.RCALF
MMIF.RCALT
AERC.RCALF
AERC.RCALT
Observations
MMIF.RCALF
MMIF.RCALT
AERC.RCALF
AERC.RCALT
Total
Hours
Modeled a
1584
1591
1591
1586
1586
1266
1266
1266
1266
1266
1591
1612
1612
1612
1612
1067
1067
1067
1067
1067
911
912
912
907
907
1219
1224
1224
1216
1216
1379
1391
1391
1384
1384
795
811
811
811
811
1145
1145
1145
1145
1145
3 Total hours modeled - hours without calms or missing data.
112
-------
5.7.2 Statistical measures and methods
The statistical measures and methods are similar to the techniques applied in the EPA
evaluation of AERMOD (USEPA, 2003). The statistical scores were calculated using the
maximum concentration results from each receptor ring. The maximum concentration at each
receptor ring for each AERMOD scenario was extracted resulting in a set of 50 values (one
value for each ring) for each scenario. The results from the WRF meteorology based AERMOD
simulations were compared to the results from the meteorological measurement based
AERMOD simulations. The primary purpose of the investigation was to judge whether the WRF-
based method provided results similar to maximum predictions from observation-based
methods, and not biased towards underprediction. The tools used for the evaluation are
described below:
• Quantile-quantile (Q-Q) plots. Q-Q plots were prepared to test the ability of the WRF-
based concentration predictions to represent the frequency distribution of the
observation-based AERMOD concentration predictions. Q-Q plots are simple ranked
pairings of predicted and observed concentration, such that any rank of the predicted
concentration is plotted against the same ranking of the observed concentration. The Q-
Q plots can be inspected to examine whether the predictions are biased towards
underestimates at the important upper-end of the frequency distribution.
• Log-log scatter diagrams. Log-log diagrams were prepared in which each plot contains a
plot of the WRF-extraction AERMOD predictions versus the observation-based
AERMOD predictions paired in time.
• Robust Highest Concentration (RHC). RHC has been used in most EPA model
evaluation studies to measure the model's ability to characterize the upper end of the
frequency distribution. Note this can also be accomplished by visual inspection of the Q-
Q plots.
3n-l\ (10)
v '
where cn is the nth highest concentration and c is the average of the (n-1) highest
concentrations. For the small sample size data sets in the current analysis, n was taken
to be 5.
• Fraction-factor-of-two (FF2). FF2 is the ratio of the number of WRF-based concentration
predictions within a factor-of-two from the observation-based predictions to the total
number of predictions.
• Geometric correlation coefficient (rg). rg is a standard correlation coefficient computed
using the natural log of the predictions and measurements and is calculated as follows:
113
-------
rg =
- lnO))(ln(y) - ln(y)) (11)
• Geometric mean (jtyg). jtyg is the nth root of the product of n numbers. The geometric mean
provides a method to evaluate a general expected value with dampened outlier
influence. Geometric mean is calculated as follows:
(12)
Bias of the geometric mean (MG). MG is a symmetric measure independent of the
magnitude of the concentration. A perfect model would result in MG = 1. MG is
calculated as follows:
MG = ev
where c0 and cp are the observed and predicted concentrations, respectively.
• Geometric variance (VG). VG is a measure of the precision of the dataset. A perfect
model would result in VG = 1. VG is calculated as follows:
Total modeling score (TMS). To summarize the modeling results with one composite
score, a "model score" was calculated for each AERMOD case. The formula for this
score is basically an average of five statistics: the FF2, geometric correlation coefficient,
geometric mean, RHC, and VG with equal weighting. MG is not included in the model
score because (ig is an equivalent measure. The value ranges from 0 to 1, with 1 being a
"perfect" model:
, . „„„..„,„„, . mm(RHCp,RHCobs) i.Q) (15)
+ rn + ,.. .. ^ + max(RHCp,RHCobsyVGp)
114
-------
6 AERMOD SIMULATION RESULTS
Given the large number of AERMOD simulations conducted for this study, it would be tedious to
review each set of simulation results individually in this report. Instead, summary statistics are
used to compare AERMOD performance. Several individual AERMOD simulations with
interesting results are investigated further in this report. For each of the 1,125 AERMOD
simulations, a set of plots were produced to evaluate and compare concentration results and
corresponding meteorological conditions. The scenarios simulated were described in Section
5.7.1, using the source groups described in Table 8 over the periods outlined in Table 10.
Meteorology for each AERMOD simulation was produced using the methods described in Table
9. All results plots were included on a data disk in Appendix C
The summary statistics for each simulation, defined in 5.7.2, provide an effective set of tools for
evaluating the simulations. The summary statistics were calculated for each of the source
groups identified in Table 8. As described in Section 5.4, Source Groups #1 and #2 can be
considered representative of "short" stacks. Source Group #4 can be considered representative
of "tall stacks." Source Group #3 average stack height falls within the average heights of the tall
and short stack groups. Some interesting cases were selected for further investigation using the
concentration and meteorology plots developed for each simulation.
For each simulation, the maximum concentration at each receptor ring was extracted. The
statistics are computed for the data pairs comprised of the maximums values from each ring
from the WRF-based and measurement-based AERMOD simulations. Each statistic is
computed using the 50 data pairs representing the maximum values from each receptor ring.
Therefore, each data pair is not necessarily extracted from the same receptor but is extracted
from receptors with the same downwind distance from the sources. Q-Q plots and scatter plots
of the results from each scenario are included in Appendix C.
The RHC may be the most relevant statistic for this study. The capability of WRF-driven
AERMOD simulations to produce conservative concentration estimates (as conservative as the
measurement-based AERMOD simulations) is inferred by the RHC. RHC provides a tool to
investigate the accuracy of results on the high end of the concentration spectrum for each
simulation. The FF2 statistic provides a means to evaluate the accuracy of the entire distribution
of concentration for each WRF-driven AERMOD simulation. All other statistics are implied
through its value. Plots were prepared for direct comparison of the RHC and FF2 statistic
values.
In the figures of the statistical score, different features were used to identify each separate
simulation. A color scheme was used to identify the results by site (B2: green, B3: black, C1:
red, and C2: blue). An outline symbol was used to identify each simulation by year (circle for
2010 simulations, triangle for 2011 simulations, and square for 2012 simulations). The results
for each source group were identified by the symbol inner number. This symbol/color scheme
was used to compare the results of all simulations together in a single plot for each WRF
meteorology extraction method.
115
-------
The statistical scores were examined for each averaging period followed by summaries for each
measurement site. A few interesting cases were selected for deeper analysis. The full set of
statistical scores at each site is reported in Sections 6.6, 6.7, 6.8, and 6.9 for Sites B2, B3, C1,
and C2, respectively for the 1-hour averaging periods only. Tables of the statistical scores for
the 3-hour, 8-hour, 24-hour, and period averaging periods are included in Appendix B.
6.1 1 -hour Averages
The 1-hour averaging period is the shortest averaging period used in this study. The 1-hour
averaging period is important because it is the shortest averaging period used in the NAAQS
used for peak SC>2, NC>2, Os, and CO concentrations.
6.1.1 Robust high concentration
RHC values were calculated using the AERMOD results from all four WRF-meteorology
extraction methods as shown in Table 7 and for each of the emission sources listed in Table 6.
The values from all scenarios are shown in Figure 77 - Figure 80 for the 1-hour averaging
period simulations for each WRF extraction method (MMIF.RCALF, MMIF.RCALT,
AERC.RCALF, and AERC.RCALT, respectively). Figure 77 - Figure 80 demonstrate RHC
values were greatest for Source Group #5 and lowest for Source Group #1. Since emission rate
was not normalized between the groups, magnitude of the concentrations was irrelevant except
for when comparing between Source Groups #2 and #5 (same stack groups, no-downwash vs.
downwash). The downwash cases resulted in the highest RHC due to early transport of the
plume to the surface that resulted in high concentrations at the near-source receptors.
Most of the WRF-based RHC values were similar in magnitude to the observation-based RHC
predictions: all WRF RHC values were within a factor of two of the measurement RHCs. The
overall WRF-based average RHC value was slightly underpredicted in comparison to the
observation-based runs. WRF-based RHCs at site B2 were consistently lower than observation-
based RHCs, although still within a factor of two. Site C1 RHCs compared the most favorably.
Source Group #4, representative of emissions from "tall" stacks as described in Section 4.4,
appeared the most sensitive to differences in extracted meteorology. The RHC was
underpredicted in the AERC and MMIF.RCALF cases. The 2010 and 2011 site B2 simulations
for Source Group #4 produced the WRF RHC values that differed most from the observation-
based RHC values.
From the results of the meteorological analysis, it was hypothesized underpredicted wind
speeds at sites C1, B2, and B3 could result in overprediction of concentration. The
hypothesized trends were not evident in the 1-hour average RHC results. However, the bias
over the entire period would not likely relate to the short term concentration bias (which is
inferred through RHC) since worst case concentrations would occur during short term periods of
extreme weather: generally low wind speed, highly stable or unstable conditions.
116
-------
10J
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
0
2011
2012
Obs-based
Figure 77. Robust high concentration results for MMIF.RCALF AERMOD 1-hour averaging times.
117
-------
Obs-based (:-s!:-. )
icr
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
0
2°11
2012
Figure 78. Robust high concentration results for MMIF.RCALT AERMOD 1-hour averaging times.
118
-------
10J
«
i
II!
a
10'
Obs-based (.-s/-.' )
103
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
0
2011
2012
Figure 79. Robust high concentration results for AERC.RCALF AERMOD 1-hour averaging times.
119
-------
10-
s
ra
ul
tc
10'
icr
Obs-based (.-s/ .' )
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
Figure 80. Robust high concentration results for AERC.RCALT AERMOD 1-hour averaging times.
120
-------
6.1.2 Fraction-factor-of-two
The FF2 results for 1-hour averaging times are plotted in Figure 81 - Figure 84 for the four WRF
extraction simulation sets, respectively. Most of the AERMOD simulations using WRF
meteorology performed similarly to the observation-based AERMOD simulations. The FF2
scores tended to be between 0.7 and 1.0. Site B2 simulations fare the worst under all cases,
particularly the tall stack source group (#3 and #4) B3 simulations. Overall, there is no
significant difference in FF2 between the WRF extraction methods.
The lowest FF2 scores occurred for simulations involving Source #4. The average FF2 for all
Source Group #4 simulations was 0.77 compared to 0.85, 0.87, 0.80, and 0.99 for Source
Groups #1, 2, 3, and 5, respectively. Again as with the RHC results, concentration results from
units with greater stack height resulted in the lowest scores.
Chukchi Sea sites average FF2 results are higher on average compared to the Beaufort Sea
sites, which is not surprising given the WRF-MMIF PEL heights are used for C1 and C2.
6.1.3 Discussion
Source Group #4 2011 simulations were selected for further investigation, based on the low
RHC and FF2 scores. The concentration maxima are shown in Figure 85, plotted with respect to
distance from the origin. The maximum observation-based concentrations occurred about 500
m downwind of the source. The RCALF maximum concentrations occurred at the same
distance, but were almost a factor of two lower in magnitude. The MMIF.RCALT simulation
resulted in a more accurate maximum concentration, but it occurred farther downwind at about
1000 m from the source.
The PEL height and wind speed for this case are shown in Figure 86 and Figure 87,
respectively. In Figure 86 and Figure 87, the meteorological parameter values shown are the
values from the hour that results in the maximum concentration at each receptor ring. Therefore,
the meteorological variables at each point do not necessarily occur at the same time as those at
the other points. At the 500 m distance (distance of the maximum observation-based
concentrations) RCALF simulations have a high wind speed of about 7 m/s compared to the
observed and RCALT wind speeds of 4 m/s. The RCALT wind speeds, PEL heights, L, and u*
(Appendix C is the data disk for plots of L and ujor this case) are all similar in magnitude to the
observation-based values around the 500 m downwind distance. It is therefore surprising
concentrations do not agree more favorably at this point.
The Source Group #5 simulations resulted in the best FF2 scores and most favorable RHC
scores. Examination of the meteorology reveals large differences in PEL height, wind speed,
and other factors despite good agreement in concentration. It is evident that when downwash is
used, the influence of differences in meteorology is minimal.
The Source Group #2 simulations were the second best performing set of simulations in terms
of FF2 and RHC score. For the B2 2011 case, the WRF-based runs resulted in RHC values
121
-------
ranging from 86 - 93 |js/m3 compared to the observation-based RHC value of 116. FF2 values
ranged from 0.8 to 1.0. The Source Group #2 maximum concentrations occurred at roughly 350
m downwind for the observation-based and WRF cases. The PEL height and wind speed for
this case are shown in Figure 88 and Figure 89, respectively. The observed wind speed is about
5.7 m/s and the WRF-based wind speeds are around 8 m/s for the concentration maxima cases
at 350 m distance. Observed PEL height is about 180 m compared to the RCALT WRF heights
at about 150 m and RCALF cases at 210 m.
122
-------
1.0
0.8
0.6
0.4
0.2
0.0
Source Number: 1-5
Site by Color:
• B2
• Bo
• C1
• C2
Year by Outline:
2010
0
2011
2012
Figure 81. Fraction-factor-of-two MMIF.RCALF AERMOD results for 1-hour averaging time.
123
-------
1.0
0.8
0.6
0.4
0.2
0.0
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
Figure 82. Fraction-factor-of-two MMIF.RCALT AERMOD results for 1-hour averaging time.
124
-------
1.0
0.8
0.6
0.4
0.2
0.0
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
Figure 83. Fraction-factor-of-two AERC.RCALF AERMOD results for 1-hour averaging time.
125
-------
1.0
0.8
0.6
0.4
0.2
0.0
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
Figure 84. Fraction-factor-of-two AERC.RCALT AERMOD results for 1-hour averaging time.
126
-------
10
10
103
Distance {. )
Figure 85. Concentration maxima vs. distance, Site B2, 2011, Source Group #4, 1-hr avg.
500
_ 400
E,
£
_E71
I 300
Dt
C
'x
Z
200
100
<
it
i .
00
)
v(sA
oc
db&i
a
<*«•
X
VI
c
1*
r*)(
<
t
:iocioooc
1 1
XXE
I4!
XJU
DO
^^
TY
IN *
VVy n
00
1
dB
^
3
U
DO
DCOO
3O
IwftvfcA
iboc
Sv^.>
ft
s
ec
™
t
M
*T
vr
V
CCD obs-based
A(V\ AERC.RCALF
[HH AERC.RCALT
MMIF.RCALF
COD MMIF.RCALT
104
Distance ( )
Figure 86. PBL height corresponding to concentration maxima, Site B2, 2011, Source
Group #4,1-hr avg.
127
-------
1
^
*"l
ft
<
o
500000;
:occoa
— "^-r
jjjjg
C
xx>
n
'VI
4?
ifi
^
•t!
•^
IX,
"*>
SE
I
a
j
<
D
J
xn
XXX)
O
soft
A
^
C
o
oo
nt
"ft
oca
A
^^
CCD obs-based
AAA AERC.RCALF
DH AERC.RCALT
000 MMIF.RCALF
OOO MMIF.RCALT
104
Distance {. )
Figure 87. Wind speeds corresponding to concentration maxima, Site B2, 2011, Source
Group #4,1-hr avg.
i
I
•o
1^
«ee
cxxocoooc
cxo
ocobcxx)
.0>
£li,i
CXX)
DOX)
cxxxxo
10"
Distance (••.)
Figure 88. Wind speed corresponding to concentration maxima, Site B2, 2011, Source
Group #2,1-hr avg.
128
-------
600
500
£ 400
300
200
100
c
c
A
\ '
)ou
vMi
i
UL
rtx*
LX
<*»•
|
4"
X
^
!
ft
JL
' I
0
Ot
.1
^jA^^A^
T ,- T , ,,
oooc
'« 1 1 1 1 1
:>
*w
XDOO
III!
ft
OOC
•'"V
I
A.
XX
t
1
n
^
J
*
X
X
0
0
K
n
»
3
-i
>
X
¥
)XXXX)C
)X)
J^^fe*
CXX)
^*;
ep
a
10-=
104
Distance {. )
Figure 89. PBL height corresponding to concentration maxima, Site B2, 2011, Source
Group #2,1-hr avg.
6.2 3-hour Averages
6.2.1 Robust high concentration
The RHC results for all 3-hour average simulations are plotted in Figure 90 - Figure 93 for the
four WRF extraction methods, respectively. All of the RHC results were within a factor of two of
the observation-based RHC results. Again, the downwash simulations (Source Group #5)
resulted in the highest normalized concentrations and best agreement with the observation-
based simulations. The Source Groups #4 and #3 simulations at site B2 consistently
underpredicted RHC values, similar to the 1-hour averages, but were still within a factor of two.
Overall the WRF meteorology rediagnosed PBL heights (MMIF.RCALT) provided the most
accurate and conservative results, when compared to the observation driven AERMOD
predictions. The AERC.RCALF simulations resulted in the least conservative results, based on
the number of cases where WRF-based RHC was lower than observation-based RHC.
129
-------
I
iL
Bfi
10'
Figure 90. Robust high concentration results for MMIF.RCALF AERMOD 3-hour averaging times.
130
-------
•o
01
I/I
ID
Source Number 1-5
Site by Color:
• B2
| B3
• C1
• C2
Year by Outline:
2010
2011
2012
Obs-based (j
-------
s
•5?
1
LL
ct
10'
Obs-based
103
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
Figure 92. Robust high concentration results for AERC.RCALF AERMOD 3-hour averaging times.
132
-------
cc.
Obs-based (/is/™3 )
10J
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
S\
2011
2012
Figure 93. Robust high concentration results for AERC.RCALT AERMOD 3-hour averaging times.
133
-------
6.2.2 Fraction-factor-of-two
The FF2 results for 3-hour averaging times are plotted in Figure 94 - Figure 97. Source
Group #5 simulations had the highest average FF2 (0.99), and Source Group #4 simulations
had the lowest average FF2 (0.76). The AERC simulations result in higher FF2 scores overall at
site C1. The AERC site C1 and C2 FF2 scores were higher overall than the B2 and B3 AERC
cases, but the Chukchi Sea FF2 scores were similar to the Beaufort Sea FF2 scores for the
MMIF cases.
6.2.3 Discussion
The 2011 B2 Source Group #4 FF2 scores were consistently low and 2010 B2 Source Group #4
RHC scores consistently varied the most from the observation-based RHCs. Source Group #4
tall stacks were a poorer performer for the short term average simulations, particularly at site
B2. These cases were selected for a deeper investigation.
A plot of the maximum concentrations with distance for the 2010 B2 Source Group #4
simulations is shown in Figure 98. The near-source concentrations from the WRF simulations
were much lower than the observation-based simulations. After about 300 m, the WRF-based
AERMOD concentrations matched the observation-based concentrations more closely. The
wide differences in the near-source are the cause of the comparatively low FF2 scores for the
2010 B2 Source Group #4 simulations. The plot of PEL heights relating to the concentrations
plotted in Figure 98 are shown in Figure 99. From 0 to 80 m distance, the WRF maximum
concentrations occur during highly stable periods (PEL heights of 25 m for both RCALT and
RCALF cases) while the observation-based near-source maximums occurred during unstable
periods (PEL heights of 250 m).
A single event on August 31 near sunset resulted in the unstable conditions that lead to the
higher observation-based concentrations. The observed ASTD was about -1.7°C while WRF
ASTD was about +0.3°C. The observed ASTD supported unstable atmospheric conditions that
lead to the higher PEL heights observed. The erroneous WRF ASTD supported stable
conditions. The observation-based meteorology during this period resulted in lvalues of-5.0 m
and light northerly winds about 1 m/s. The WRF meteorology included L values about + 70 m
and northerly winds about 3.5 m/s. This type of event, where the measurements result in a
negative ASTD while WRF supports a positive ASTD, occurred frequently in August of 2010 at
site B2, as seen in Figure 47.
134
-------
1.0
0.8
0.6
0.4
0.2
0.0
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
0
2011
2012
Figure 94. Fraction-factor-of-two MMIF.RCALF AERMOD results for 3-hour averaging times.
135
-------
1.0
0.8
0.6
0.4
0.2
0.0
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
Q 2010
0
2011
2012
Figure 95 Fraction-factor-of-two MMIF.RCALT AERMOD results for 3-hour averaging times.
136
-------
1.0
0.8
0.6
0.4
0.2
0.0
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
/\
2011
2012
Figure 96. Fraction-factor-of-two AERC.RCALF AERMOD results for 3-hour averaging times.
137
-------
1.0
0.8
0.6
0.4
0.2
0.0
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
Figure 97. Fraction-factor-of-two AERC.RCALT AERMOD results for 3-hour averaging times.
138
-------
Periods of unstable conditions occur in the WRF dataset also and it was expected that these
periods would result in maximum near-source concentrations. A period of comparable
meteorology was identified in the WRF RCALT meteorology on Sept. 18, 2010 where L values
were consistently - 5.0 m and winds were light, about 1-1.5 m/s, and from the north. Despite
the strongly unstable conditions at the surface, the PEL heights were maintained at the
minimum 25 m throughout the period. During this same period, the observations resulted in L
values in the range of -10 to -20 m and PEL heights around 100 m. The L values of -5.0 indicate
strongly unstable conditions, supported by the negative ASTD during this period. The low PEL
heights seemed unphysical given the conditions.
Plots of WRF and profiler temperature gradients during this period at the Endeavor Island
profiler site are shown in Figure 100. The plot shows that the region was dominated by a warm
layer aloft. WRF surface temperatures were much cooler, resulting in a much stronger inversion
near the surface. Despite the strongly unstable conditions right at the surface, the related
forcing is unable to penetrate significantly into the strong inversion and the low PEL height is
retained.
10J
104
Distance {)
Figure 98. Concentration vs. distance, Site B2, 2010, Source Group #4, 3-hr. avg.
139
-------
400
350
300
•§i 250
1
— '
c 200
X
2
150
100
50
<
YTl
1 1 1
nr
Cf
v
r
E
nr
LJ 1 1 1 1
OOOC£
JOOOOCX
<™f*
fe\AAAA
ft
XXXD
&
OOC
X
r
^^
QjJ,
ooc
A^
1
c:
XX
A
:
r
V
X
i
p^N
US
d
-ij
J
3
-iJ
r
XT
104
Distance {. )
Figure 99. PBL height of concentration maxima vs. distance for Site B2, Source
Group #4, 3-hr avg.
140
-------
Profiler Sounding 2640 - 2010-09-18 23:00 LST
Profiler Sounding 2640 - 2010-09-18 23:00 LST
'•ZiTop-'&CO j^
U - 11 ZO - 0,05 L - 8888 RH - 1 00
Profiler - Solid, WRF -Dashed
ZiMMIF -40
12345
Temperature (C)
275 280 285
Potential Temperature (K)
Figure 100. WRF and Profiler soundings (temperature vs. height) at Endeavor Island
Sept. 18th, 2010. Temperature (left) and potential temperature (right) are plotted for both
the profiler (solid line) and WRF solution (dashed line). PBL heights estimates by WRF
(ZiWRF), WRF-MMIF recalculation (ZiMMIF), bulk Richardson method for the profiler data
(ZiRib), and subjective "hand analyzed" PBL height as determined by a qualified
meteorologist (ZiTop) are listed on the plot.
6.3 8-hour Averages
6.3.1 Robust high concentration
The RHC results for all of the 8-hour averaging time simulations are shown in Figure 101 -
Figure 104 for the four WRF extraction methods, respectively. There was very little difference in
scores between the simulations using different meteorological extraction methods. The most
notable difference was a slight increase of RHC score with the recalculation of PBL height for all
source groups except #5. RHC scores for the RCALT (both AERC and MMIF) simulations were
slightly less accurate, but more conservative on average.
141
-------
6.3.2 Fraction-factor-of-two
The 8-hour average FF2 scores from all simulations are plotted in Figure 105 - Figure 108.
Many of the site B2 simulations perform poorly based on these scores, especially Source Group
#3 2010 simulations. Despite the low FF2 scores at some of these sites, the magnitude of the
maximum concentrations agree well, resulting in the good agreement between WRF and
observation-based RHC scores. Further analysis of the site B2 Source Group #3 2010 case
revealed WRF-based concentrations were overpredicted downwind of 1000 m and
underpredicted upwind of 200 m, as shown in Figure 109. In the 200 - 1000 m range, where all
the maximum concentrations occur, the concentrations agree well. This pattern was the result of
WRF and WRF-MMIF predicting lower PEL heights, as shown in Figure 110. The low PEL
heights lead to underpredicted near-source concentrations and overpredicted concentrations in
the far-source. However, as seen in the near-source region, both observation-based and WRF-
based maximum concentrations occurred during unstable periods based on L values shown in
Figure 111. Again, as seen with the 1-hour and 3-hour results, WRF restricted the growth of the
PEL due to a warm layer aloft. An implication of these results is that if RHC was evaluated for
just the near-source or far-source the RHC results would be less favorable since WRF-based
concentrations did not match the observation-based concentrations at these distances at this
site.
142
-------
I
10'
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
/\
2011
2012
Obs-based
Figure 101. Robust high concentration results for MMIF.RCALF AERMOD 8-hour averaging times.
143
-------
I
\
1
10
J3
u.
tt
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
Obs-based
Figure 102. Robust high concentration results for MMIF.RCALT AERMOD 8-hour averaging times.
144
-------
10=
"6
i
Obs-based
Figure 103. Robust high concentration results for AERC.RCALF AERMOD 8-hour averaging times.
145
-------
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
/\
2011
2012
Obs-based (.••s/-,-. )
Figure 104. Robust high concentration results for AERC.RCALT AERMOD 8-hour averaging times.
146
-------
1.0
0.8
0.6
0.4
0.2
0.0
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
0
2011
2012
Figure 105. Fraction-factor-of-two MMIF.RCALF AERMOD results for 8-hour averaging times.
147
-------
1.0
0.8
0.6
0.4
0.2
0.0
(D
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2012
Figure 106. Fraction-factor-of-two MMIF.RCALT AERMOD results for 8-hour averaging times.
148
-------
1.0
0.8
0.6
0.4
0.2
0.0
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
Figure 107. Fraction-factor-of-two AERC.RCALF AERMOD results for 8-hour averaging times.
149
-------
1.0
0.8
0.6
0.4
0.2
0.0
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
Figure 108. Fraction-factor-of-two AERC.RCALT AERMOD results for 8-hour averaging times.
150
-------
10
103
Distance {. )
Figure 109. Concentration vs. distance, Site B2, 2010, Source Group #3, 8-hr. avg.
250
200
150
100
50
(
XDO
O
c
u<
c
J
c
ex
<
m*
cm
:tm
ooc
xs^/
cue
>
|
JDO
•JQQQ
3
OOC
Y>
-y
o
"X
~x
1
xo
ooc
oca
xx:
oc
a
c
D
X
a
lir
10"
Distance I )
Figure 110. PBL height of concentration maxima vs. distance for Site B2, 2010, Source
Group #3, 8-hr avg.
151
-------
0.15
0.10
1
£ 0.05
K
c
01
0 0.00
I
i -°-°5
c
0 10
-0.15
-0 70
<
P
x>
0
$
*
*
c
>
Pi
o
Pi
",-V-f-iP^
occ
"HAA^
PPC
1 1 L
JOCC
A
J
oa
00-
'**
XX
»
X
:••:
ffiBE
0
v\
3
^
y
5i
**i
300QQC
xoooo
f^^\y^^>
OOCx
rw*
XX
JV
SB0BE
0
M
C
V
*
X
:?
M
~v ,
oc
v
kV
10"
Distance {. )
Figure 111. ML of the concentration maxima vs. distance for Site B2, 2010, Source
Group #3, 8-hr avg.
6.4 24-hour Averages
6.4.1 Robust high concentration
The 24-hour average RHCs from all simulations are plotted in Figure 112 - Figure 115, with
more variance in RHC scores, on average, than the simulations at shorter averaging times. All
RHC scores were within the factor-of-two boundaries, suggesting WRF-based simulations
provided a similar range of maximum concentrations as the observation-driven AERMOD
simulations. There were no substantial differences in the average RHC values between the
simulations using different WRF meteorology extraction methods.
Site C1 2011, Source Group #5 was selected for further investigation since the RHC scores
from this case varied the most from the observation-based RHC scores. The Q-Q plot of
concentration for this case is shown in Figure 116. The WRF-based runs overpredicted
concentration throughout the entire distribution. All concentration maxima occurred in the near-
source within 30 - 60 m of the source. As seen in Figure 117, both the observations and WRF
exhibited strong negative ASTD for near-source maxima but the far-source observation-based
maxima occurred during stable conditions (positive ASTD) while the WRF-based far-source
maxima occurred during unstable conditions (negative ASTD). Despite this difference, observed
and WRF PEL heights (shown in Figure 118) were low, from 25 - 100 m, characteristic of stable
conditions. The low PEL height during strongly unstable conditions (as indicated by the value of
L) is the same phenomenon identified at site B2 for the 1-hour and 3-hour cases examined.
Both the observation-based and WRF-based maximums occurred on the same day, Sept. 5th,
2011. The range of L and PEL height in the AERMOD surface meteorology files were similar for
152
-------
both the observation-based and WRF-based runs. The sensible heat flux varied with a WRF-
based value averaging about +45 W/m2 and observation-based value averaging about 60 W/m2.
Also, the WRF wind speed average was lower at about 8 m/s compared to the observation-
based 9 m/s. The additional heat flux and wind speed contributed to more mixing resulting in
lower concentrations in the observation-based simulations. Despite the differences, the WRF-
based maximum concentrations were within a factor of two of the observation-based
concentrations.
6.4.2 Fraction-factor-of-two
The FF2 scores for the 24-hour averaging simulations are plotted in Figure 119 - Figure 122.
Highest average FF2 values were achieved by the AERC.RCALT simulations. Site C1
simulations had notably better average FF2 scores using recalculated PEL height values. Site
C2 simulations had notably better average FF2 scores using AERCOARE and recalculated PEL
heights. This was most evident for Source Group #3 2012 simulations where FF2 jumped from
0.44 in the MMIF.RCALF simulation to 0.78 in the AERC.RCALT simulation. Concentration
maxima for this case are plotted on Figure 123. All WRF-based simulations resulted in
underpredicted concentrations from 30 - 150 m and overpredicted concentrations beyond 1000
m. The AERC.RCALT simulation resulted in less overprediction and underprediction in the near-
source and far-source, respectively. Figure 124 shows the PEL heights corresponding to the
concentration maxima. The MMIF rediagnosed PEL heights corresponded better with PEL
heights used for the observation-based simulations in the near-source and far-source (note that
MMIF PEL heights were used at sites C1 and C2, however). These results highlight the
sensitivity of AERMOD results to PEL height.
153
-------
1
ro
•
102
Obs-based (/
-------
I
I
«••*
1
1CT
Obs-based
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
Figure 113. Robust high concentration results for MMIF.RCALT AERMOD 24-hour averaging times.
155
-------
10-'
•=
io2
Obs-based (••s/.-/. • )
10J
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
Figure 114. Robust high concentration results for AERC.RCALF AERMOD 24-hour averaging times.
156
-------
io2
t£.
IO2
Obs-based (.••s/.".
10d
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2012
Figure 115. Robust high concentration results for AERC.RCALT AERMOD 24-hour averaging times.
157
-------
10=
5 10'
101
10'
10-
ID'1 10° 101
Observation-Based Max. AERMOD Cone. I
10'
10J
Figure 116. Q-Q plot for Site C1, 2011, Source Group #5, 24-hr avg.
o
3: i
o
Sea Temp.
u
<
_3
f
inn
nr
nc
V
C
rv
a
XDOC
<
00
a
)
X
o
3
oc
0
c
a
10J
Distance {
10*
Figure 117. Air-sea temperature difference corresponding to concentration maxima, Site
C1, 2011, Source Group #5, 24-hr avg.
158
-------
700
600
•g 500
I
« 400
01
c
| 300
200
100
.
^
(
i .
KVl
YTl
iV)
m
ft
nr
y
V
SV
1 f
o
1 1 ii J 1
ayavyt.
0000
«
C)
ca
oc
>
nr
«oc
^AA
JU
T *•
t***ti
OJL
MtMfMd
Ml
T<
CK
-»
-i
5n
xoo
ooc
XX
O
X
nSxS^e^
1.
1_
-r-
V
XA
T-1"
y^
10"
Distance (
Figure 118. PBL height corresponding to concentration maxima for Site C2, 2012, Source
Group #5, 24-hr avg.
159
-------
1.0
0.8
0.6
0.4
0.2
0.0
3
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2012
Figure 119. Fraction-factor-of-two MMIF.RCALF AERMOD results for 24-hour averaging times.
160
-------
1.0
0.8
0.6
0.4
0.2
0.0
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
Figure 120. Fraction-factor-of-two MMIF.RCALT AERMOD results for 24-hour averaging times.
161
-------
1.0
0.8
0.6
0.4
0.2
0.0
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
Figure 121. Fraction-factor-of-two AERC.RCALF AERMOD results for 24-hour averaging times.
162
-------
1.0
0.8
0.6
0.4
0.2
0.0
-------
10
103
Distance {. )
Figure 123. Concentration maxima vs. distance, Site C2, 2012, Source #3, 24-hr avg.
800
f 600
u
£
U)
i
I400
z
200
i
pq
•rtfUl
pq
««
a
ftr
C
v
i_
C
tf
I_L
a
£*
• i '
OCX
:ooo
**>*,
IfW
ooo
XX)0
1 1
4i
(
oo
1 i
tl
3
a
, t
+
X
i
1
>
a
rt
.
D
^1
-
i
M
3
^
i
3
X>
UD
XQQOC
^/VVV^A
AAA/
.•r^jQQg
\A/\
ses
A/"
es
A
X?
^/
iJ^
Nl'
_;
A/
ee
10J
10"
Distance ( )
Figure 124. PBL height corresponding to concentration maxima, Site C2, 2012, Source
Group #3, 24-hr avg.
164
-------
6.5 Period Averages
6.5.1 Robust high concentration
The period-average RHC scores are plotted on Figure 125 - Figure 128. All WRF-based RHC
values were within the factor-of-two of the observation-based RHC values. A majority of the
RCALF simulations predicted lower RHCs than the observation-based simulations. The RCALT
simulations were more conservative and matched the observation-driven RHC better. There
was no discernible difference in RHC value between the MMIF and AERC based runs.
For RCALF simulations, PEL height was biased low during stable and unstable periods. The
high frequency of minimum PEL heights (25 m) that occurred with RCALF simulations
contributed to the negative RHC bias. The highly stable conditions limited vertical mixing,
resulting in lower average concentrations in the near-source, given that most of the stack
heights were well above the surface. Plume height was also above the minimum PEL height of
25 m in cases with tall stacks, such as Source Group #4, or cases with sufficient thermal and
mechanical buoyancy, such as Source Groups #1 and #2.
Source Group #1 2012 WRF-based RHC values were consistently lower than observation-
based values at Sites C1 and B3. Plots of maximum concentration with distance are plotted in
Figure 129 and Figure 130. Both set of results have a similar trend, with near-source and
far-source WRF-based concentrations of similar magnitude as the observation-based
concentrations and the most deviation from 100 to 500 m. At both sites, WRF underpredicts
seasonal wind speed (by -0.5 m/s at both sites). Observed seasonal PEL height at site B3 was
about 200 m. WRF underpredicted PBL height with seasonal averages of 180 m and 160 m for
RCALT and RCALF simulations, respectively. At site C1 the observation-based dataset had the
same seasonal PBL height as the RCALT simulations (about 400 m) and the RCALF
simulations resulted in a seasonal average of 270 m.
Combined, the WRF seasonal underprediction of wind speed and PBL height would presumably
result in overprediction of concentration. At site B3, this was the case in the near-source and
far-source but not in the 100 - 500 m range. At site C1 WRF-based concentrations were
overpredicted only in the far-source. The pattern cannot be entirely an artifact of WRF low PBL
predictions because the C1 RCALT simulations resulted in a similar pattern as the RCALF
simulations (the observation-based simulations used the RCALT PBL heights).
The seasonal concentration profiles for these sites for the tall stack groups (Source Groups #3
and #4) appear similar, with greatest underprediction in the 100 - 1000 m range. Therefore, the
pattern isn't a trait of the short stack group (#1). Further investigation would be warranted to
discover the cause of the differences in maximum concentration in the mid-range of distance
from the source.
165
-------
10
ID1
Obs-based (/ s/./, )
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
/\
2011
2012
Figure 125. Robust high concentration results for MMIF.RCALF AERMOD Period averaging times.
166
-------
101
102
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
Obs-based
Figure 126. Robust high concentration results for MMIF.RCALT AERMOD period averaging times.
167
-------
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
10
Figure 127. Robust high concentration results for AERC.RCALF AERMOD period averaging times.
-------
I/I
ra
I 101
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
101
Obs-based (."S/ •. )
Figure 128. Robust high concentration results for AERC.RCALT AERMOD period averaging times.
169
-------
10
10
103
Distance (. )
Figure 129. Concentration maxima vs. distance, Site C1, 2012, Source Group #1, Period
average.
10
Distance (in)
Figure 130. Concentration maxima vs. distance, Site B3, 2012, Source Group #1, Period
average.
170
-------
6.5.2 Fraction-factor-of-two
The period-average FF2 results are plotted in Figure 131 - Figure 134. Again, the Source
Group #4 results have the lowest FF2 values and Source Group #5 results have the highest
values. This correlated with the previous observation that sources with higher stack heights
were more sensitive to differences in meteorology. The RCALT simulations result in a large
increase in FF2 score for most of the site C1, compared to the RCALF simulations. FF2 scores
are relatively low at site B2 with Source Groups #3 and #4 simulations as low as 0.3 - 0.4.
The Site B2 2010 plot of concentration by distance for the season is shown in Figure 135. All
WRF-based simulations resulted in an underestimate of concentration in the near-source and
overprediction of concentration in the far-source. These results are consistent with the case
where PEL height is consistently underpredicted. The repressed mixing rate in stable conditions
prevents the exhaust plume from reaching the ground-level receptors in the near-source and
results in higher concentrations in the far-source. The WRF seasonal wind speed was lower
than the observed value (4.7 and 4.0 m/s, respectively).
6.6 Site B2 Results
The statistical score results for site B2 1-hour averaging time simulations are listed in Table 11.
The TMS provided an effective summary statistic to compare the performance of the four WRF
meteorology extraction methods. The accuracy of the upper-end of the concentration
distribution can be judged based on RHC score.
Most of the WRF-based B2 simulations slightly underpredicted RHC overall with an average
RHC value 81% that of the observation-based RHC, and the recalculated PEL height (RCALT)
simulations performed slightly better that RCALF runs. The RCALT simulations predicted an
average RHC 84% that of the observation-based RHC, while the average RCALF simulation
was 77% of the observation-based RHC. The simulations using the WRF meteorology directly
(MMIF. RCALF and MM IF. RCALT) performed slightly better overall with an average TMS score
of 0.74 and 0.69 between all applicable scores, respectively. The simulations where WRF
meteorology was processed with AERCOARE (AERC. RCALF and AERC. RCALT) resulted in
overall average scores of 0.72 and 0.69, respectively.
171
-------
1.0
0.8
0.6
0.4
0.2
0.0
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
0
2011
2012
Figure 131. Fraction-factor-of-two MMIF.RCALF AERMOD results for period averaging times.
172
-------
1.0
0.8
0.6
0.4
0.2
0.0
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
0
2011
2012
Figure 132. Fraction-factor-of-two MMIF.RCALT AERMOD results for period averaging times.
173
-------
1.0
0.8
0.6
0.4
0.2
0.0
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
0
2011
2012
Figure 133. Fraction-factor-of-two AERC.RCALF AERMOD results for Period averaging times.
174
-------
1.0
0.8
0.6
0.4
0.2
0.0
[2
o
s
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
S\
2011
2012
Figure 134. Fraction-factor-of-two AERC.RCALT AERMOD results for Period averaging times.
175
-------
10
Distance [ •.)
Figure 135. Concentration maxima vs. distance for the Period average, Site B2, 2010,
Source Group #3.
Table 11. Site B2 1-hour average statistics scores.
Year Src. Run
obs
AERC.RCALT
1 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
2 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
3 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
4 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
Geo
Mean
17.20
7.38
7.68
7.90
7.96
42.74
18.58
19.75
19.85
20.63
65.40
29.37
27.37
17.70
18.39
22.75
2.24
1.72
1.78
2.35
Geo. St.
Dev.
2.57
9.31
8.73
8.38
8.59
2.31
8.53
7.56
7.46
7.58
2.40
10.84
10.90
17.44
17.58
4.47
111.84
113.43
90.33
89.86
MG
1.00
2.33
2.24
2.18
2.16
1.00
2.30
2.16
2.15
2.07
1.00
2.23
2.39
3.69
3.56
1.00
10.14
13.21
12.81
9.70
VG
1.00
19.93
15.04
13.84
14.83
1.00
18.59
11.37
11.28
11.04
1.00
42.05
46.94
1078.48
1092.87
1.00
>5000
>5000
>5000
>5000
RHC
50.02
41.24
39.17
38.67
40.47
101.31
87.11
86.14
85.99
88.59
190.48
170.50
166.69
133.95
138.34
79.69
86.05
47.48
45.01
79.21
Geo. R
1.00
0.85
0.86
0.85
0.84
1.00
0.86
0.87
0.86
0.86
1.00
0.80
0.81
0.73
0.72
1.00
0.80
0.81
0.77
0.79
FF2
1.00
0.70
0.72
0.72
0.72
1.00
0.76
0.76
0.76
0.76
1.00
0.72
0.72
0.68
0.66
1.00
0.60
0.62
0.60
0.60
IMS
0.90
0.57
0.61
0.62
0.61
0.80
0.59
0.63
0.63
0.63
0.81
0.58
0.59
0.50
0.53
0.82
0.49
0.41
0.48
0.41
176
-------
Table 11, continued. Site B2 1-hour average statistics scores.
Year Src. Run
obs
AERC.RCALT
5 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
1 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
2 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
2011 3 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
4 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
5 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
Geo
Mean
237.59
247.27
247.71
253.77
253.77
10.67
14.05
9.00
9.94
14.13
30.51
32.28
20.56
22.43
32.11
64.45
30.24
14.68
17.12
29.14
10.60
11.40
3.02
3.61
10.95
220.04
237.45
237.53
220.92
218.45
Geo. St.
Dev.
3.27
3.09
3.08
3.23
3.23
5.43
3.39
7.00
6.62
3.72
3.98
3.30
6.85
6.37
3.59
2.88
6.54
19.57
16.42
7.68
16.71
8.78
45.55
34.86
10.06
3.26
3.00
2.99
3.01
2.99
MG
1.00
0.96
0.96
0.94
0.94
1.00
0.76
1.19
1.07
0.76
1.00
0.95
1.48
1.36
0.95
1.00
2.13
4.39
3.77
2.21
1.00
0.93
3.51
2.93
0.97
1.00
0.93
0.93
1.00
1.01
VG
1.00
1.01
1.01
1.01
1.01
1.00
1.42
1.53
1.40
1.32
1.00
1.06
2.13
1.78
1.04
1.00
6.81
3469.96
958.47
11.02
1.00
1.92
199.55
62.10
1.54
1.00
1.02
1.02
1.01
1.01
RHC
1224.60
1131.49
1131.49
1221.71
1221.71
48.74
42.31
42.08
45.18
45.63
116.19
89.03
85.86
90.39
90.33
186.80
122.01
122.01
130.24
138.14
84.14
46.12
45.36
46.62
62.37
1135.76
1021.43
1021.43
1017.11
1008.37
Geo. R
1.00
1.00
1.00
1.00
1.00
1.00
0.99
0.95
0.95
0.99
1.00
1.00
0.94
0.95
0.99
1.00
0.83
0.64
0.65
0.81
1.00
0.98
0.87
0.88
0.99
1.00
1.00
1.00
1.00
1.00
FF2
1.00
1.00
1.00
1.00
1.00
1.00
0.82
0.88
0.90
0.86
1.00
1.00
0.82
0.84
1.00
1.00
0.68
0.58
0.62
0.70
1.00
0.76
0.66
0.60
0.76
1.00
1.00
1.00
1.00
1.00
IMS
0.80
0.98
0.99
0.97
0.98
0.87
0.83
0.86
0.89
0.87
0.80
0.93
0.77
0.81
0.98
0.81
0.56
0.49
0.49
0.60
0.82
0.75
0.56
0.56
0.82
0.80
0.96
0.98
1.00
0.99
The 2010 Source Group #4 simulations were selected for further investigation. All of the 2010
Source Group #4 WRF simulations resulted in low FF2 scores of 0.6, and TMS scores of 0.41 to
0.49. However, the RCALT simulations resulted in RHC scores very near to the observation-
based value of 79 us/m3. The two RCALF simulations resulted in RHC scores of 45 and
47 us/m3.
The concentration maxima are shown in Figure 136 for these simulations. Note that the RCALF
simulations underpredicted concentration within the 1000 - 3,000 m range of receptor distances
while the RCALT simulations overpredicted concentration within this range. All of the
simulations underpredict concentration from 30 to 500 m. The difference in RHC occurs
because the RCALT simulations produced a peak concentration at about 1,000 m downwind of
the source that matches the magnitude of the observation-based peak that occurs at roughly
450m.
177
-------
Figure 137 reveals that both the RCALT and RCALF maxima occurred when PEL heights were
the minimum 25 m. Figure 138 shows the inverse L values corresponding to the maximum
concentrations. The values are at or near the minimum 1/L value of -0.2 m, indicating highly
unstable conditions. Observation -based maxima occurred during neutral stability conditions at
these distances. Figure 139 shows the wind speed for the same cases. The RCALT
concentration maxima occurred mostly during wind speeds of about 1 m/s, while RCALF
maxima occurred during wind speeds ranging from 0.5 to 3.5 m/s. The combination of low wind
speed and low PEL height resulted in higher concentrations for RCALT simulations.
10
Distance {•• )
Figure 136. Concentration maxima vs. distance, Site B2, 2010, Source Group #4.
178
-------
Mixing Height (m)
SM NJ PvJ UJ O
Ui O t-n O U
O O O O O O C
\
r
L
A
AJL
TT
•LL
W~
UL
r*r^
Ut
*^~
A.
^
•fl
E
jj^ji^jm^jijjE
TTTTW
^j^^j^
AAJn
• , rl
O
«Vfi^
~i~r
fW
oc
•^
[
V*
XX
•
n
c
Y
X
ft
'
o
D<
3
>
3
Dl
DOOOO
C
°oa
YY"
or
o
h
T
rl-
^&4
¥l^
r ;x
l^M |J
104
Distance (. )
Figure 137. PBL height corresponding to concentration maxima, Site B2, 2010, Source
Group #4.
0 15
0 10
1
£ 0 05
m
c
I
UJ o 05
0 10
0 15
?
<
J_U
*>v
A
-•J
1
oc
o2
x£b
0
•
a
0
x>
DC
1
J
jQGQO
AA
o
)3
AA
/u
)
(V
••*:
i
j-v
M
1 1
0
-
1
?
|
a
^
1C
Distance (••
Figure 138. Inverse L corresponding to concentration maxima, Site B2, 2010, Source
Group #4.
179
-------
z
c
I
in
JUU
*v<
UL
u
X
JL
U
CX
:ooooa
X3OO
<>
0
00
3
1
a
0
XX
>
X
n
V*
s
ji
DOOOO
AA
0
0
. , A C
ataW*^
AA
ooai
RTT
oC
\A/s
x>
0=
A.
>
)&
M
••
^
X
,
> ?*
r
10*
10"
Distance {. )
Figure 139. Wind speed corresponding to concentration maxima, Site B2, 2010, Source
Group #4.
6.7 Site B3 Results
The statistical score results for site B3 1-hour averaging time simulations are listed in Table 12.
For site B3, the MMIF.RCALT simulations resulted in slightly better TMS scores than the other
simulations. MMIF.RCALT simulations resulted in average RHC values of 0.84, compared to the
average scores of 0.81, 0.80 and 0.81 for MMIF.RCALT, AERC.RCALF, and AERC.RCALT,
respectively. Source Group #5 performed the best overall (average TMS of 0.99) and Source
Group #4 performed the worst overall (average TMS of 0.69) in terms of TMS score.
Most of the simulations resulted in RHC scores that were close in magnitude to the observation-
driven RHC score results. The average WRF RHC score was 92% of the observation-based
RHC (slight underprediction on average). Source Group #5 average WRF RHCs were 98% of
the observation-based scores. Source Group #3 average RHC scores were the lowest at 86%
of the observation-based value. Source Group #4 (the tallest set of stacks) simulations resulted
in an average RHC score 92% of the observation-based RHC, but resulted in the greatest
extremes with 2012 MMIF.RCALF at 66% of the observation-based RHC and 2010
MMIF.RCALT at 118% of the observation-based RHC.
The 2011 Source Group #3 simulations were selected for further investigation due to a large
gap in TMS score between the different runs. The MMIF.RCALT simulation resulted in a TMS
score of 0.95 while the MMIF.RCALF simulation resulted in a TMS score of 0.67 Both these
simulations resulted in a similar TMS score, but the MMIF.RCALF had a much lower FF2 score
180
-------
(0.76 versus 0.94) and a geometric mean of 24.6 compared to the MMIF.RCALT value of 41.3
(observation-based geometric mean was 42.1).
The concentration maxima values are plotted with respect to distance in Figure 140. It is evident
that the RCALT simulations agree better with the observation-based simulations in the near-
source region. At 200 m and beyond, all of the WRF simulations resulted in concentrations that
agree well with the predictions of the observation-based simulations. In the near-source, the
RCALF simulations underpredicted concentration. The Q-Q plot of these simulations is shown in
Figure 141, further illustrating the RCALF underprediction leading to the low FF2. The
observation-based simulation maximum occurred at about 200 m downwind of the source, while
the WRF simulation maxima occurred roughly 500 - 700 m downwind.
Figure 142 shows PEL height versus distance for the Source Group #3 B3 2011 simulations.
The RCALT simulations resulted in high PEL heights, greater than 500 m, in the near-source
region corresponding to unstable conditions (L is negative in RCALT simulations, confirming
this). However, these PBL heights were greater than the observation-based PEL heights that
ranged from 100 to 200 m in the near-source region. The RCALF PBL heights agreed better
with the observations with PBL heights of 150 m at a distance of 50 to 200 m and both RCALF
and observation-based inverse L was -0.2.
Figure 143 is a plot of wind speed versus distance for these simulations. The RCALT
simulations near-source maxima occurred during light wind speeds of 0.5 m/s, while the RCALF
simulations maxima occurred during winds greater than 3.0 m/s (observation-based maxima
occurred during wind speeds of 1.0 -1.5 m/s).
Again, it is evident that WRF's tendency to predict minimum PBL heights even during unstable
periods (when L < 0) is the cause of the discrepancies in near-source maxima. The MMIF
rediagnosis resulted in PBL heights more characteristic of unstable conditions, improving the
AERMOD performance.
Table 12. Site B3 1-hour average statistics scores.
Year
2010
Src. Run
obs
AERC.RCALT
1 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
2 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
Geo
Mean
12.54
26.79
25.74
21.14
21.93
34.58
57.85
54.89
46.38
48.48
Geo. St.
Dev.
4.47
1.50
1.50
1.88
1.96
3.41
1.47
1.51
1.75
1.77
MG
1.00
0.47
0.49
0.59
0.57
1.00
0.60
0.63
0.75
0.71
VG
1.00
7.59
6.94
3.08
2.98
1.00
3.47
2.95
1.85
1.90
RHC
47.32
44.94
41.33
40.89
44.56
96.26
90.62
83.60
82.83
87.36
Geo. R
1.00
0.79
0.81
0.95
0.95
1.00
0.71
0.80
0.94
0.93
FF2
1.00
0.76
0.76
0.80
0.80
1.00
0.80
0.80
0.84
0.84
IMS
0.86
0.62
0.62
0.73
0.72
0.80
0.67
0.70
0.81
0.79
181
-------
Table 12, continued. Site B3 1-hour average statistics scores.
Year Src. Run
obs
AERC.RCALT
3 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
2010 4 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
5 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
1 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
2 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
2011 3 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
4 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
5 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
1 AERC.RCALF
Geo
Mean
50.37
84.50
76.22
63.99
72.18
10.63
40.30
33.80
27.16
31.76
249.72
248.69
248.75
248.76
248.95
13.89
18.54
9.95
10.31
15.94
32.76
42.60
23.89
24.57
37.24
42.07
54.66
23.39
24.57
41.32
11.72
26.24
4.99
5.37
18.90
163.62
169.68
189.57
190.37
160.45
14.65
15.56
15.58
Geo. St.
Dev.
3.59
1.61
1.72
2.07
1.90
14.23
1.74
1.69
2.15
2.42
3.16
3.12
3.12
3.20
3.20
3.98
2.07
5.73
5.78
2.80
3.42
2.01
5.22
5.28
2.66
3.90
2.42
11.78
11.11
3.99
13.49
2.74
30.16
25.76
4.61
3.39
3.11
3.14
3.14
3.08
3.26
2.95
3.02
MG
1.00
0.60
0.66
0.79
0.70
1.00
0.26
0.32
0.39
0.34
1.00
1.00
1.00
1.00
1.00
1.00
0.75
1.40
1.35
0.87
1.00
0.77
1.37
1.33
0.88
1.00
0.77
1.80
1.71
1.02
1.00
0.45
2.35
2.18
0.62
1.00
0.96
0.86
0.86
1.02
1.00
0.94
0.94
VG
1.00
3.68
2.86
1.55
1.91
1.00
1117.46
785.50
122.15
116.46
1.00
1.00
1.00
1.00
1.00
1.00
1.83
1.87
1.84
1.28
1.00
1.52
1.69
1.67
1.18
1.00
1.40
8.21
6.88
1.12
1.00
27.69
18.59
15.12
4.44
1.00
1.04
1.07
1.07
1.03
1.00
1.06
1.08
RHC
157.18
139.96
142.12
122.45
125.70
62.18
68.67
69.92
56.53
73.54
1236.04
1244.94
1244.94
1190.27
1193.66
45.51
45.11
41.35
43.44
45.10
90.78
90.68
83.61
90.70
91.87
139.03
120.83
121.60
130.21
135.59
61.43
62.91
53.40
53.07
68.31
1116.54
1072.03
1016.64
1007.06
1066.92
42.63
36.87
39.34
Geo. R
1.00
0.68
0.76
0.96
0.93
1.00
0.72
0.71
0.91
0.91
1.00
1.00
1.00
1.00
1.00
1.00
0.96
0.92
0.92
0.96
1.00
0.96
0.94
0.94
0.97
1.00
0.98
0.92
0.92
0.97
1.00
0.98
0.91
0.90
0.99
1.00
0.99
0.99
0.99
0.99
1.00
0.98
0.98
FF2
1.00
0.82
0.82
0.86
0.84
1.00
0.72
0.72
0.74
0.74
1.00
1.00
1.00
1.00
1.00
1.00
0.84
0.80
0.80
0.86
1.00
0.86
0.76
0.78
0.90
1.00
0.88
0.74
0.76
0.94
1.00
0.76
0.68
0.68
0.82
1.00
1.00
1.00
1.00
1.00
1.00
0.98
0.94
IMS
0.81
0.65
0.71
0.82
0.79
0.82
0.52
0.55
0.57
0.55
0.80
1.00
1.00
0.99
1.00
0.86
0.82
0.78
0.79
0.89
0.81
0.85
0.79
0.80
0.92
0.81
0.84
0.67
0.67
0.95
0.84
0.64
0.58
0.62
0.69
0.80
0.98
0.95
0.95
0.98
0.85
0.94
0.94
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
2 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
13.47
13.42
38.33
36.21
35.84
31.69
32.27
3.62
3.64
2.65
2.60
2.65
3.12
3.22
1.09
1.09
1.00
1.06
1.07
1.21
1.19
1.05
1.06
1.00
1.02
1.02
1.10
1.11
40.40
38.47
90.59
85.46
80.39
80.23
94.42
0.99
0.99
1.00
0.99
0.99
0.99
0.98
1.00
1.00
1.00
1.00
1.00
0.98
0.96
0.97
0.96
0.80
0.97
0.97
0.94
0.91
182
-------
Table 12, continued. Site B3 1-hour average statistics scores.
Year
2012
Src. Run
obs
AERC.RCALT
, AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
4 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
5 AERC.RCALF
MMIF.RCALF
Geo
Mean
67.63
51.15
42.07
35.88
38.31
17.15
15.61
16.03
11.21
11.51
243.59
242.22
242.20
243.06
Geo. St.
Dev.
2.19
2.99
4.18
5.43
5.35
7.59
7.35
7.49
10.92
11.54
3.16
3.09
3.09
3.18
MG
1.00
1.32
1.61
1.89
1.77
1.00
1.10
1.07
1.53
1.49
1.00
1.01
1.01
1.00
VG
1.00
1.37
2.53
5.30
4.59
1.00
1.04
1.09
1.49
1.50
1.00
1.00
1.00
1.00
RHC
162.17
125.46
128.83
137.85
132.11
74.14
54.01
52.29
48.99
56.02
1187.88
1200.45
1199.78
1195.56
Geo. R
1.00
0.92
0.87
0.83
0.85
1.00
1.00
0.99
0.99
0.99
1.00
1.00
1.00
1.00
FF2
1.00
0.82
0.76
0.72
0.76
1.00
1.00
0.98
0.88
0.78
1.00
1.00
1.00
1.00
IMS
0.81
0.80
0.72
0.64
0.67
0.82
0.92
0.96
0.83
0.80
0.80
1.00
1.00
1.00
MMIF.RCALT
243.08
3.18
1.00
1.00
1196.11
1.00
1.00
1.00
10
10
103
Distance {. )
Figure 140. Concentration maxima vs. distance, Site B3, 2011, Source Group #3.
183
-------
5 10'
101
10'
10'
10-
1QU 101
Observation-Based Max. AERMOD Cone. \
10J
Figure 141. Q-Q plot of WRF-based AERMOD concentration results versus
observation-based results, Site B3, 2011, Source Group #3.
500
400
£
» 300
at
E
'x
Z
200
100
r
(
A
I
XX
^k^K
"T
OC
.
V
.«
•?
CCD obs-based
Mi^ AERC.RCALF
DH AERC.RCALT
MMIF.RCALF
CCO MMIF.RCALT
It)2
Distance (- )
104
Figure 142. PBL height vs. distance for concentration maxima, Site B3, 2011, Source
Group #3.
184
-------
3.5
3.0
2.5
1.5
1.0
c
*'
XX)
^&
'
coc
)
a
0
a
A
n
xx
D
X
lYv
^A
OD
C
oO
^ ^^
DOOOOOC
WW
000
uw
ocxx
i't
AA-
.
XT
;**
/
._*_
OC
'
M
'
O
1
A
X
A.
X
oc
,_
XX
103
10"
Distance (• )
Figure 143. Wind speed vs. distance for concentration maxima, Site B3, 2011, Source
Group #3.
6.8 Site C1 Results
The 1-hour average statistic scores for site C1 are listed in Table 13. The average TMS score of
the AERC simulations was 0.94 while the average TMS score of the MMIF simulations was
0.84. Again, the taller stack groups (Source Groups #3 and #4) result in the lowest TMS scores,
with an average of 0.80 and 0.83, respectively while the other source groups result in average
TMS scores of 0.93.
None of the simulations resulted in RHC scores less than 85% of the respective observation-
based RHC score. Geometric correlation coefficients are high, greater than 0.9, except for 2012
Source Group #3 simulations.
The 2010 Source Group #3 simulations were selected for further examination because these
resulted in the lowest FF2 scores and TMS scores, despite RHC scores that were roughly 90%
of the observation-based RHC scores. . Figure 144 shows the concentration maxima with
respect to distance. The AERC simulations resulted in concentrations nearer to the observation-
based simulation results, but underpredicted. The MMIF simulations highly underpredicted
concentration in the near-source region.
PEL heights plotted in Figure 145 and wind speeds plotted in Figure 146 are shown to be
favorable for producing high concentrations near the source for all of the WRF simulations. PEL
heights are high, greater or equal to 600 m, compared to the measured PEL heights of 170 m in
the near-source region. Although all of the WRF-based maximum concentrations occur at about
185
-------
the same wind speed, the AERC maxima occur at u. values of about 0.11 m/s compared to the
MMIF values of about 0.08 m/s (observation-based u. values were about 0.06 m/s).
It appears in this case that the AERC simulation near-source results may be fortuitous. Both
MMIF and AERC maxima occurred in highly unstable conditions at about the same wind speed
and PEL height. Higher u. in the AERC simulations resulted in increased plume spread that
resulted in higher maximum concentrations in the near-source. The observation-based
simulation near-source maxima occurred with lower PEL heights and lower u. than the WRF-
based simulations.
Table 13. Site C1 1-hour average statistics scores.
Year Src. Run
obs
AERC.RCALT
1 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
2 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
2011 3 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
4 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
5 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
1 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
2 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
Geo
Mean
22.90
23.25
22.82
18.57
19.03
53.56
54.47
53.82
44.54
45.53
76.33
75.99
79.95
67.49
62.97
37.58
34.16
34.14
24.58
26.47
203.91
235.84
236.12
238.63
238.57
18.40
17.56
17.13
14.55
14.72
42.13
39.85
38.68
32.34
33.95
Geo.
St.
Dev.
1.64
1.60
1.59
2.16
2.13
1.57
1.53
1.48
1.90
1.93
1.66
1.76
1.72
2.02
2.21
1.58
1.84
1.86
2.90
2.82
3.30
3.09
3.09
3.18
3.18
2.31
2.64
2.73
3.59
3.44
2.21
2.47
2.55
3.49
3.15
MG
1.00
0.99
1.00
1.23
1.20
1.00
0.98
1.00
1.20
1.18
1.00
1.00
0.96
1.13
1.21
1.00
1.10
1.10
1.53
1.42
1.00
0.87
0.86
0.85
0.86
1.00
1.05
1.07
1.27
1.25
1.00
1.06
1.09
1.30
1.24
VG
1.00
1.01
1.01
1.20
1.16
1.00
1.01
1.01
1.12
1.12
1.00
1.03
1.03
1.14
1.30
1.00
1.09
1.11
2.09
1.85
1.00
1.08
1.08
1.08
1.08
1.00
1.03
1.06
1.38
1.29
1.00
1.03
1.05
1.39
1.24
RHC
42.62
44.02
43.41
44.76
45.99
91.27
88.60
83.99
84.81
90.63
137.67
133.93
153.77
135.53
134.48
64.63
63.70
63.74
63.05
62.64
1214.21
1231.18
1234.44
1229.09
1226.75
44.65
43.26
41.33
43.96
45.00
89.88
87.34
82.20
85.39
88.30
Geo.
R
1.00
0.98
0.98
0.92
0.94
1.00
0.98
0.97
0.93
0.93
1.00
0.96
0.96
0.89
0.82
1.00
0.91
0.88
0.81
0.84
1.00
0.98
0.98
0.98
0.98
1.00
0.99
0.99
0.97
0.98
1.00
0.99
0.99
0.98
0.98
FF2
1.00
1.00
1.00
0.86
0.92
1.00
1.00
1.00
0.94
0.94
1.00
1.00
1.00
0.92
0.88
1.00
0.96
0.96
0.86
0.86
1.00
0.92
0.92
0.94
0.94
1.00
0.98
0.96
0.84
0.84
1.00
0.98
0.98
0.84
0.86
TMS
0.90
0.99
0.99
0.88
0.90
0.81
0.99
0.98
0.92
0.91
0.81
0.98
0.95
0.89
0.86
0.83
0.94
0.93
0.76
0.79
0.80
0.93
0.94
0.94
0.94
0.86
0.97
0.96
0.85
0.87
0.81
0.97
0.96
0.85
0.88
186
-------
Table 13, continued. Site C1 1-hour average statistics scores.
Year Src. Run
obs
AERC.RCALT
3 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
2012 4 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
5 AERC.RCALF
MMIF.RCALF
Geo
Mean
71.36
51.14
50.63
35.43
36.52
21.74
18.77
18.37
12.29
12.93
167.73
213.14
215.91
204.43
Geo.
St.
Dev.
1.83
3.23
3.32
6.25
5.98
3.90
5.10
5.50
10.08
9.21
3.26
3.11
3.11
3.14
MG
1.00
1.40
1.41
2.02
1.95
1.00
1.16
1.18
1.77
1.68
1.00
0.79
0.78
0.82
VG
1.00
1.96
2.06
13.94
11.77
1.00
1.12
1.18
3.90
3.16
1.00
1.11
1.12
1.07
RHC
149.29
130.75
127.26
131.36
136.11
58.66
51.73
61.86
57.34
49.27
1193.22
1258.86
1248.61
1235.46
Geo.
R
1.00
0.83
0.83
0.71
0.72
1.00
1.00
1.00
0.98
0.98
1.00
0.98
0.98
0.99
FF2
1.00
0.82
0.82
0.74
0.74
1.00
0.94
0.90
0.74
0.74
1.00
1.00
1.00
1.00
IMS
0.81
0.75
0.76
0.60
0.60
0.82
0.92
0.88
0.69
0.70
0.80
0.92
0.93
0.95
MMIF.RCALT
198.55
3.12
0.85
1.06
1243.41
0.99
1.00
0.95
10'
101
Id1
10"
102
Distance {. )
104
Figure 144. Concentration maxima vs. distance for Site C1, 2010, Source Group #3.
187
-------
600
500
£ 400
Dt
i
= 300
X
Z
200
100
C
xn
oc
a
X
c
0
D
OO^v
ooo
:> dSa
\
nr
Wtf
oo
3D
°°f
°<
T
/~V~1
ttarvi
eeet
•wwi
X)G
-F
V
n-*
u^
Oc
Q
-0
o
e
H-
f
3
*2
st-
Y*
a
a
io4
Distance (m
Figure 146. Wind speed corresponding to concentration maxima, Site C1, 2011, Source
Group #3.
188
-------
6.9 Site C2 Results
The site C2 statistical score results are listed in Table 14. The RCALT simulations produced
more favorable results with an average TMS score of 0.86 compared to the RCALF TMS score
of 0.78. Most simulations produced RHC scores that were accurate and slightly conservative,
resulting in an overall average RHC 101% of corresponding observation-based RHC values.
The maximum RHC value was 145% of the corresponding observation-based RHC for Source
Group #4, 2010 RCALT simulations at site C2. The tall stack group, Source Group #4, resulted
in the lowest TMS scores (0.64) compared to the other stack groups (0.80 to 0.97). The Source
Group #1 simulations slightly underestimated RHC on average, resulting in an average RHC
score 96% of the corresponding observation-based RHC scores while the other stack groups
resulted in average RHC scores greater than observation-based RHC scores.
Table 14. Site C2 1-hour average statistics scores.
Year Src.
Run
Geo
Mean
Geo.
St.
Dev.
MG
VG
RHC
Geo.
R
FF2 TMS
obs
18.04
2.34
1.00
1.00
45.51
1.00
1.00
1.00
AERC.RCALT
17.67
2.90
1.02
1.10
45.06
0.97
1.00
0.97
1 AERC.RCALF
13.36
4.56
1.35
1.92
42.25
0.95
0.82
0.79
2
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
13.09
16.09
41.16
39.87
31.60
31.34
4.66
3.62
2.30
2.51
3.65
3.83
1.38
1.12
1.00
1.03
1.30
1.31
2.02
1.30
1.00
1.03
1.39
1.46
42.46
45.55
88.95
91.15
89.15
90.76
0.95
0.97
1.00
0.99
0.98
0.98
0.76
0.82
1.00
1.00
0.84
0.82
0.79
0.88
0.81
0.98
0.86
0.85
MMIF.RCALT
obs
AERC.RCALT
2010 3 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
4 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
5 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
2012 1 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
36.33
47.91
60.21
43.77
42.05
48.02
20.92
21.01
8.48
8.81
14.59
197.72
216.21
219.05
211.01
206.13
15.12
11.86
9.27
9.20
11.29
3.12
3.04
2.67
5.04
5.52
4.06
3.56
4.71
18.59
15.31
8.95
3.03
3.07
3.08
3.10
3.08
2.77
4.22
6.02
5.85
4.60
1.13
1.00
0.80
1.10
1.14
1.00
1.00
1.00
2.47
2.38
1.43
1.00
0.92
0.90
0.94
0.96
1.00
1.27
1.63
1.64
1.34
1.15
1.00
1.08
1.33
1.49
1.11
1.00
1.14
43.51
25.88
2.96
1.00
1.01
1.02
1.01
1.01
1.00
1.34
2.76
2.67
1.52
91.75
120.94
134.54
130.93
130.10
130.72
52.92
76.10
46.15
45.74
76.99
1164.51
1228.07
1229.25
1225.01
1224.13
41.90
40.58
36.82
38.37
42.99
0.98
1.00
1.00
0.99
0.99
0.99
1.00
0.99
0.97
0.95
0.98
1.00
1.00
1.00
1.00
1.00
1.00
0.98
0.95
0.95
0.98
0.86
1.00
0.96
0.90
0.88
0.92
1.00
0.94
0.76
0.74
0.74
1.00
1.00
1.00
1.00
1.00
1.00
0.78
0.76
0.76
0.78
0.92
0.81
0.91
0.91
0.88
0.96
0.83
0.90
0.55
0.63
0.67
0.80
0.97
0.98
0.99
0.99
0.88
0.85
0.72
0.73
0.81
189
-------
Table 14, continued. Site C2 1-hour average statistics scores.
Year Src. Run
obs
AERC.RCALT
2 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
Geo
Mean
36.19
28.58
24.22
23.86
27.74
Geo.
St.
Dev.
2.62
3.81
5.11
5.02
4.11
MG
1.00
1.27
1.49
1.52
1.31
VG
1.00
1.27
1.96
1.94
1.38
RHC
92.68
90.74
86.95
91.43
92.20
Geo.
R
1.00
0.98
0.98
0.98
0.98
FF2
1.00
0.78
0.80
0.80
0.78
IMS
0.80
0.86
0.78
0.78
0.85
obs
AERC.RCALT
3 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
4 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
5 AERC.RCALF
MMIF.RCALF
MMIF.RCALT
58.62
38.19
29.51
30.66
35.12
19.08
8.98
4.42
5.34
8.16
191.96
234.80
235.54
203.59
201.47
2.13
3.72
9.95
8.86
4.29
4.23
12.10
43.19
27.99
13.16
3.09
3.05
3.05
3.08
3.08
1.00
1.54
1.99
1.91
1.67
1.00
2.12
4.32
3.58
2.34
1.00
0.82
0.82
0.94
0.95
1.00
1.96
24.94
16.22
2.61
1.00
5.94
4042.10
408.55
8.78
1.00
1.06
1.07
1.02
1.01
131.37
125.13
137.90
135.46
135.27
62.56
55.29
51.89
49.03
53.37
1112.40
1106.07
1106.12
1102.10
1102.02
1.00
0.91
0.89
0.90
0.91
1.00
0.99
0.93
0.92
0.98
1.00
0.99
0.99
1.00
1.00
1.00
0.68
0.82
0.82
0.70
1.00
0.66
0.70
0.70
0.64
1.00
1.00
1.00
1.00
1.00
0.81
0.74
0.63
0.66
0.72
0.82
0.63
0.56
0.57
0.62
0.80
0.95
0.95
0.98
0.99
6.10 Far-Source Results
The results of this study have shown the highest concentrations typically occur immediately
downwind of a source, generally within a kilometer of the stacks. Offshore air pollutant sources
may be located a greater distance from shore. To address far-source model performance, the
maximum concentrations at 10,000 m (the most distant ring of receptors from the stacks) were
extracted for evaluation. The 1-hour averaging time simulation results are shown in Figure 147 -
Figure 150.
The WRF 1-hour average simulations resulted in far-source concentrations that were accurate
and slightly conservative on average That is, the simulations predicted concentrations that were
similar and slightly higher than concentrations predicted by the observation-based simulations.
In the far-source, there was little difference in results between AERC and MMIF simulations or
RCALT and RCALF simulations. No simulations underpredicted concentration and site B2 and
B3 simulations were the most conservative, resulting in overpredicted concentrations for all
source groups. The overpredicted concentrations were likely due to the high frequency of
minimum PEL heights predicted by WRF and MMIF at Sites B2 and B3. The high frequency
provided more opportunities for high far-source concentrations.
The period-average simulations at 10,000 m, shown in Figure 151 - Figure 154, resulted in
larger differences in WRF- and measurement-based AERMOD concentrations than found for
the 1-hour average simulations maxima. For the period average simulations, the influence of
stack height was less of a factor for simulation accuracy. There was little difference between the
190
-------
RCALF and RCALT simulations at all Sites except C1. The C1 predictions were improved
(values are nearer to the observation-based concentrations) using the MMIF rediagnosed
mixing heights (RCALT). There was little noticeable difference between AERC and MMIF
simulations. Site B2 and B3 maximum concentrations were overpredicted consistently. These
results were consistent with the negative wind speed and PEL height bias at these sites.
The 2010 WRF-based C2 concentration predictions were less than the observation-based
concentrations. WRF predicted lower wind speed on average at C2 which could account for the
lower concentrations in the far-source due to the lower u*.
191
-------
30
25
IA
a.
5
o
o
S
CC.
20
15
10
te.
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2011
2012
5 10 15 20
Observation-Based Max. AERMOD Cone,
25
30
Figure 147. Far-source at 10,000 m 1-hour average maximum concentrations MMIF.RCALF simulations.
192
-------
5
0
o
te.
I
VI
as
(0
t£.
30
25
20
15
10
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
2012
5 10 15 20
Observation-Based Max. AERMOD Cone,
25
30
Figure 148. Far-source at 10,000 m 1-hour average maximum concentrations MMIF.RCALT simulations.
193
-------
o
u
Q
O
S
-------
30
25
IA
a.
5
o
o
S
20
15
10
-
oc
Source Number: 1-5
Site by Color:
• B2
• Bo
• C1
• C2
Year by Outline:
2010
/\
201 1
2012
10 15 20
Observation-Based Max. AERMOD Cone,
25
30
Figure 150. Far-source at 10,000 m 1-hour average maximum concentrations AERC.RCALT simulations.
195
-------
I
Jt
i
o
o
tt
%
1<
re
m
cp
LJL
GC
0.7
0.6
0.5
0.4
0.3
0.2
0.1
°-8
'
/I
,
^-
' i
jj,;) i-Lgi
/ Mw
' .A v
4 /
' ' f
^
s£»
/
3/
/
a;-
/'
®
3
/X
(D
^. -^
4
7
f ^j\
\ ^ji 'j
,-"'
/
''"
/
, - ''
/
,--'
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
Q 2010
(\ 2011
Q 2012
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Observation-Based Max. AERMOD Cone. (.••:;/••. )
Figure 151. Far-source at 10,000 m Period average maximum concentrations for MMIF.RCALF simulations.
196
-------
0.8
0.1 0.2 0.3 0.4 0.5 0.6
Observation-Based Max. AERMOD Cone. (.••Si1--.- )
0.7
Source Number: 1-5
Site by Color:
• B2
• B3
• C1
• C2
Year by Outline:
2010
/\
2011
2012
0.8
Figure 152. Far-source at 10,000 m Period average maximum concentrations for MMIF.RCALT simulations.
197
-------
0.
0.2 0.3 0.4 0.5 0.6
Observation-Based Max. AERMOD Cone. (.-•s/-.-.' )
0.7
Source Number: 1-5
Site by Color:
• B2
| B3
• C1
• C2
Year by Outline:
O 2010
0
2011
2012
0.8
Figure 153. Far-source at 10,000 m Period average maximum concentrations for AERC.RCALF simulations.
-------
0 8
0 7
«. n fi
^
u
= 05
a D*a
Q
0
a: Q 4
<
! 03
I
IL
a o 2
On
0 0
U.jj
/
/
/
1
'' x4
/• J/
0 0
/T\ /j
Ci< C^
^Vy. V3
w/
/
J
, rt
rW"
191 ^
,, LSP
<.!> /
m.
,-'
1 0
®
f
^®
/
/
/
b
A A
(D
^ •"
.2 0
/
/
.,'
(2
/
.- "
3
/
f
f
f
/
/
i
f
t
D
*
/
/
X
X
^
\ "ft /
.I*
0
1
7
(S»
1*
4 0
/
5 0
/
6 0
/
^ "
,-'
7 0
8
Source Number: 1-5
Site by Color:
• B2
| B3
• C1
• C2
Year by Outline:
Q 2010
2011
rn 2012
Observation-Based Max. AERMOD Cone. (..'S/\-.- )
Figure 154. Far-source at 10,000 m Period average maximum concentrations for AERC.RCALT simulations.
199
-------
[Blank]
200
-------
7 CONCLUSIONS
The METSTAT and site-specific meteorological analyses suggest WRF was able to produce
hourly meteorological datasets that compared favorably to land-based and overwater
measurements. The regional METSTAT analyses found temperature and wind speed were
within simple-terrain criteria for the majority of periods. Wind speed at overwater sites was
biased high in October only. Overall, WRF meteorology at all four sites agreed with measured
data, but with some biases. Average wind speed was underpredicted at all sites, more so at
Beaufort buoy locations. PEL height was underpredicted on average. Minimum PEL heights (25
m) occurred too frequently in WRF results at the Beaufort buoy locations. WRF had a tendency
to produce minimum PEL heights even during strongly unstable conditions when L was -5.0 m.
It was found in these cases that although the local ASTD supported unstable conditions, a warm
layer aloft prevented the growth of the PEL. The MMIF rediagnosed PEL height (RCALT)
improved the PEL height predictions, especially at site B3.
Overall, most AERMOD concentration results using WRF meteorology were favorable, falling
within a factor of two of the observation-based concentrations and producing RHC values that
corresponded well with observation-driven AERMOD results. Maximum concentrations in the
far-source (greater than 1,000 m) tended to be conservative. Maximum concentrations in the
near-source (within 1,000 m) were underpredicted for the tall stack simulations (Source Groups
#3 and #4) due to the persistence of overly stable conditions that prevented near-source mixing
to the surface. Maximum concentrations tended to occur in the near-source anywhere from 100
m to 1,000 m. Near-source maxima tended to occur during unstable conditions characterized by
higher PEL heights due to the increased rate of vertical mixing. The WRF-based concentration
predictions of Source Group #5 (downwash cases) were consistently the most comparable to
the observation-based predictions of all the source groups.
WRF-based AERMOD concentrations agreed better with observation-based concentrations at
Sites C1 and C2 overall. This was mostly attributed to the fact that the WRF-MMIF RCALT PEL
heights were used for the observation-based simulations. Site B2 consistently had the lowest
FF2 scores of the four sites at all averaging periods due to the persistence of minimum PEL
heights both in the RCALF and RCALT cases. The persistence of low PEL heights resulted in
far-source period averages at B2 that were highly conservative resulting in RHC values that
were more than a factor of two than the observation-based RHC values. Period average
far-source RHC values at site C2 were underpredicted with respect to the observation-based
RHC values, but still within a factor of two of the observation-based values.
The negative wind speed bias at sites B2 and B3 supported conservative maximum
concentrations values in the far-source at long-term averaging periods due to the lower rate of
diffusion associated with lower speeds. The MMIF recalculated PEL heights (RCALT) were
more comparable than the WRF (RCALF) PEL heights to the observed PEL heights.
The main conclusions of the study are summarized as responses to the set of questions below:
201
-------
• Is there a consistent bias across source type and/or location (e.g. Chukchi vs. Beaufort)?
In particular are there any instances where the WRF simulations result in a bias towards
underprediction compared to using the buoy observations?
For the sources considered in this study, the absolute maximum concentrations occurred
in the near-source during unstable conditions. There was no consistent bias at the
Chukchi or Beaufort sites for maximum short term average concentrations. Prediction
accuracy (with respect to observation-based predictions) was better at the Chukchi sites
because the WRF-MMIF PEL heights were used for the observation-based simulations
(no PEL height measurements were available for these sites). With respect to averaging
time, the long-term period-averaged concentrations were underpredicted in some
instances. For example, the simulations using the MMIF recalculated PEL heights
underpredicted the long-term maximum concentrations at the site C2.
The FF2 scores were persistently lower at site B2 than the other sites. Concentrations
were typically overpredicted at this site in the far-source and underpredicted in the
near-source due to the high frequency of minimum PEL height.
WRF-based Source Group #4 simulations underpredicted concentration in the
near-source consistently, with respect to observation-based simulations. It was found
that the taller stack groups (Source Groups #3 and #4) were more sensitive to
differences in meteorology than the other groups. Far-source tall-stack maximum
concentrations were underpredicted in cases where PEL height was overpredicted. High
PEL height corresponded to unstable conditions that promote vertical mixing. These
conditions could therefore promote higher concentration maxima for tall stacks in the
near-source but promote lower concentrations in the far-source. Near-source tall-stack
concentrations were underpredicted when PEL height was underpredicted.
The MMIF-rediagnosis (RCALT) of PEL height tended to improve WRF-based AERMOD
performance by producing more accurate PEL heights.
• For locations where WRF performed better, does that ultimately translate to different
dispersion model results?
The short term maximum concentrations were less sensitive to bias in the WRF results.
This was likely because the concentration maxima occurred during the extreme
atmospheric stability conditions (either stable or unstable). The MMIF PEL height and L
limits result in observation- and WRF-based meteorological simulations that are quite
similar during the most extreme conditions.
The long-term far-source maximum concentrations were the most sensitive to the
meteorological bias over the period modeled. Underpredicted wind speeds at Sites B2
and B3 favored conservative period-averaged concentrations in the far-source. Site B2
period average far-source concentrations were highly conservative with RHC values
greater than a factor of two of the observation-based concentrations.
202
-------
It is highly recommended that FNMOC SST analysis, or a similar high-resolution SST
dataset based on both remotely-sensed and in-situ measurements, be used instead of
alternative datasets such as the NCEP RTG for simulations of open-water periods in the
Beaufort Sea. The MacKenzie River warm-water outflow plume is a prevalent feature on
the Beaufort Sea in summer. Low resolution SST analysis or excessive smoothing may
result in erroneous air-sea temperature difference estimates. The FNMOC SST analyses
gave a better spatial and temporal description of the SST distribution and gradient
across the Beaufort Sea over the 2010-2012 periods analyzed.
• Did it make any difference when WRF predictions were processed by AERCOARE as
opposed to directly for predictions of the surface energy fluxes?
Overall, there was little discernable advantage in using AERCOARE. Considering
average TMS, the "MMIF" runs (direct extraction from WRF without AERCOARE
processing) resulted in slightly higher scores. Total average TMS was a fraction higher
for AERC simulations at site C1 only (0.94 versus 0.84) The MMIF recalculation of PEL
height (RCALT) has a much greater influence than AERCOARE processing on the
accuracy of the simulations.
• Does it make any difference when PEL heights are rediagnosed by MMIF?
Overall, maximum concentration results were more accurate and more conservative
when the MMIF rediagnosis was applied.
The concentration results from the shorter stack groups and downwash-affected sources
were less sensitive to differences in the PEL height. If the plume is already near ground
level, maximum concentrations at ground level will occur in the near-source and are less
sensitive to the height of the PEL. Concentration maxima from taller stacks are much
more sensitive to the PEL height. Note that the height of the tall stacks used in this study
is near to the minimum PEL height (25 m). If the minimum PEL height was greater than
the tallest stack, it is likely that concentration estimates would be more comparable.
Given the results of this study, a few recommendations can be made:
The use of WRF meteorology for AERMOD dispersion modeling resulted in similar
concentrations compared to the measurement-based modeling. The WRF-based
concentrations were within a factor-of-two of the predictions from the measured
meteorology simulations. WRF tended to underpredict PEL height during unstable
periods and underpredict wind speed. These biases, in general, contributed to
overpredictions of concentration in the far-source (> 1 km). In the tall stack cases, these
biases contributed to underprediction of concentration in the near-source. However,
there was no scenario modeled in this study where the maximum RHC values predicted
by WRF were underpredicted (in comparison to measurement-based RHC results) by
more than a factor of two. This suggests WRF extracted meteorology can be used as an
alternative to offshore observations for air permitting in such areas. It is likely that
203
-------
near-source underprediction could be impeded if a higher minimum PEL height was
applied.
The WRF simulated datasets should undergo considerable scrutiny prior to their
application. We recommend at minimum, an evaluation against measurements in the
offshore or coastal areas of the study domain using METSTAT. Comparisons of air-sea
temperature difference to measurements should be made where possible.
The MMIF PEL height rediagnosis (RCALT) option should be applied to obtain more
accurate and conservative maximum AERMOD predictions. The rediagnosis option
provides a consistent way to define the PEL height as opposed to the multiple definitions
used by the various WRF PEL schemes. In several instances the rediagnosed PEL
heights also agreed more closely with observations and also resulted in conservative
predicted maxima concentrations, despite a tendency to overpredict PEL height during
unstable periods on average.
The downwash algorithm should be used where applicable. It is assumed that most
offshore stacks will be mounted on ships or platforms that will form a wake.
Concentration estimates using downwash were less sensitive to meteorological biases.
Underprediction of concentration in the near-source will be prevented if downwash is
accounted for.
There is no discernable benefit in using AERCOARE to process meteorology extracted
from WRF. AERMOD results were similar overall with and without AERCOARE
processing. The only exception might be when offshore sigma-theta observations are
available. WRF does not provide either lateral or vertical turbulence parameters that
might better characterize dispersion in some instances. However, in this study other
differences between measured and simulated meteorology tended to mask the benefits
of having such measurements.
A high resolution SST dataset is recommended to capture near-shore temperature
gradients. Avoid using the coarse SST data typically available in the meteorological
reanalysis datasets. In this study, the SST spatial gradients are high in the Beaufort Sea
due to the prevalence of the Mackenzie River outflow plume. Due to the smoothing in
the reanalysis datasets, the influence of the river plume affected a much larger area than
suggested by buoy measurements and more refined SST datasets. This bias resulted in
WRF predicting a different PEL structure than was observed in some cases.
When used by AERMOD, we recommend WRF-extracted meteorology be filtered to
avoid extreme conditions not typically observed over water. In our study, we defined
calm conditions as 0.5 m/s, required mixing heights to be greater than 25 m, and did not
allow the absolute value of the Monin-Obukhov length (L) to be less than 5 m.
In conclusion, this study compared WRF meteorological predictions and WRF-driven AERMOD
simulations to AERMOD applications prepared with measurements. Such datasets in the Arctic
204
-------
are limited to a few locations, a couple seasons, and in some instances patched together with
assumptions that were difficult to assess. WRF should be used to account for locations and
seasons outside of the available datasets and the MMIF extractions likely provide a more robust
and consistent meteorological database to simulate sources in the Arctic.
205
-------
[Blank]
206
-------
8 REFERENCES
AECOM Environment, 2009. Reindeer Island Ambient Air Quality and Meteorological Monitoring
Station Quality Assurance Project Plan, Anchorage, Alaska: Prepared for Shell Exploration
and Production Company.
Air Sciences Inc., 2010. Chukchi and Beaufort Seas Oceanographic and Meteorological
Monitoring Station Quality Assurance Project Plan, Anchorage, Alaska: prepared for Shell
Exploration and Production Company.
Air Sciences Inc., 2011. Supplemental Outer Continental Shelf (OCS) Operating Permit
Application, Shell Beaufort Sea, Alaska Exploratory Program: Conical Drilling Unit Kulluk,
Attachment C: ENVIRON Profiler, Anchorage: Shell Offshore Inc.
Arya, P., 1988. Introduction to Micrometeorology. London: Academic Press.
Brashers, B. & Emery, C., 2014. The Mesoscale Model Interface Program (MMIF) Draft User's
Manual, Novato, CA: ENVIRON Int. Corp. Air Sciences Group, Prepared for U.S. EPA Air
Quality Assessment Division.
Bridgers, G., 2011. Model Clearinghouse Review ofAERMOD-COARE as an Alternative Model
for Application in an Arctic Marine Ice Free Environment. Research Triangle Park(North
Carolina): U.S. EPA.
Bromwich, D., Mines, K. & Bai, L.-S., 2009. Development and Testing of Polar WRF, Part 2:
Arctic Ocean. J. Geophys. Res., p. D08122.
Bromwich, D. et al., 2013. Comprehensive Evaluation of Polar Weather Research and
Forecasting Performance in the Antarctic. J. Geophys. Res., Volume 118, pp. 274-292.
Carsey, F., 1978. Character of Arctic PEL Structure as Determined by Acoustic Radar.. AIDJEX
Bulletin, pp. 132-150.
Chen, F., Janjic, Z. & Mitchell, K., 1997. Impact of atmospheric surface layer parameterization in
the new land-surface scheme of the NCEP Mesoscale Eta numerical model.. Bound.-Layer
Meteor., Volume 185, pp. 391-421.
Cimorelli, A. et al., 2004. AERMOD: Description of Model Formulation, s.l.: Environmental
Protection Agency, EPA-454/R-03-004.
Cole, J. & Summerhays, J., 1979. A Review of Techniques Available for Estimation of Short-
Term NO2 Concentrations. Journal of the Air Pollution Control Association, 29(8), pp. 812-
817.
DiCristofaro, D. & Hanna, S., 1989. OCD: The Offshore and Coastal Dispersion Model, s.l.:
Prepared for U.S. Dept. of Interior MMS, , Report #A085-1.
Emery, C., Tai, E. & Yarwood, G., 2001. Enhanced meteorological modeling and performance
evaulation for two Texas ozone episodes, Novato, CA: Prepared for the Texas Nat. Res.
Cons. Commission by ENVIRON Int. Corp..
207
-------
ENVIRON Int. Corp., 2014. METSTAT. [Online]
Available at: http://www.camx.com/download/support-software.apx
EPA, 1998. Interagency Workgroup on Air Quality Modeling (IWAQM) Phase 2 Summary
Report and Recommendations for Modeling Long Range Transport Impacts, Research
Triangle Park, NC, EPA-454/R-98-019: U.S. EPA Office of Air Quality Planning and
Standards.
Fairall, C. et al., 2003. Bulk Parameterization of Air-Sea Fluxes: Updates and Verification for the
COARE Algorithm. J. Climate, Volume 16, pp. 571-591.
Garratt, J., 1992. The Atmospheric Boundary Layer. New York: Cambridge University Press.
Gryning, S. & Batchvarova, E., 2003. Marine atmospheric boundary-layer height estimated from
NWP model output. Int. Jour, of Environ, and Pollution, pp. 147-153.
Hanna, S. et al., 1985. Development and Evaluation of the Offshore and Coastal Dispersion
Model. J. Air Poll. Contr. Assoc., Volume 35, pp. 1039-1047.
Hanrahan, P., 1999. The Plume Volume Molar Ratio Method for Determining NO2/NOx Ratios
for Modeling, Part 1: Methodology. Journal of the Air & Waste Management Association,
Volume 49, pp. 1324-1331.
Mines, K. & Bromwich, D., 2008. Development and testing of Polar WRF, Part 1: Greenland ice
sheet meteorology.. Mon. Wea. Rev., pp. 1971-1989.
Mines, K. etal., 2015. Sea ice enhancements to Polar WRF. Mon. Wea. Rev., (in press).
Hsu, S., 1988. Coastal Meteorology. San Diego: Academic Press, Inc..
Janjic, Z., 1994. The step-mountain eta coordinate model: further developments of the
convection, viscous sublayer and turbulence closure schemes. Mon. Weather Rev., Volume
122, pp. 927-945.
Kemball-Cook, S., Jia, Y., Emery, C. & Morris, R., 2005. Alaska MM5 Modeling for the 2002
Annual Period to Support Visibility Modeling, Novato, CA: Prepared for the Western
Regional Air Partnership by ENVIRON Int. Corp..
Leavitt, E., Albright, M. & Baumann, R., 1978. Variations in Planetary Boundary Layer
Parameters Observed During AIDJEX. AIDJEXBulletin, Volume 39, pp. 149-163.
Louis, J., 1979. A Parameteric Model of Vertical Eddy Fluxes in the Atmosphere. J. Atmos. Sci.,
Volume 35, pp. 187-202.
Mellor, G. & Yamada, T., 1982. Development of a turbulence closure model for geophysical fluid
problems. Rev. Geophys. Space Phys., Volume 20, pp. 851-875.
NCAR, 2014. Weather Research & Forecasting (WRF) ARW Version 3 Modeling System User's
Guide, s.l.: Nat. Center for Atmos. Research Mesoscale & Meteorology Division.
208
-------
NOAA-NCDC, 2014. NOAA A/CDC Integrated Surface Database. [Online]
Available at: www.ncdc.noaa.gov/isd
Richmond, K. & Morris, R., 2012. Evaluation of the Combined AERCOARE/AERMOD Modeling
Approach for Offshore Sources, s.l.: ENVIRON Int. Corp. Prepared for USEPA R.10, EPA-
910-R-12-007.
Schulman, L, Strimaitis, D. & Scire, J., 2002. Development and Evaluation of the PRIME plume
rise and building downwash model.. Journal of the Air & Waste Management Association,
Volume 50, pp. 278-390.
Simmons, A., Uppala, S., Dee, D. & Kobayashi, S., 2006. ERA-lnterim: New ECMWF reanalysis
products from 1989 onwards.. ECMWF Newsletter, pp. 26-35.
SLR, I nc., 2011. Quality Assurance Project Plan for the Endeavor Island (Endicott)
Meteorological Monitoring Program, Anchorage, Alaska: Prepared for Shell Offshore, Inc..
Stauffer, D., Seaman, N. & Binkowski, F., 1991. Use of four-dimensional data assimilation in a
limited-area mesoscale model. Part II: effects of data assimilation within the planetary
boundary layer. Mon. Weather Rev., Volume 119, pp. 734-754.
US Navy, 2014. Fleet Numerical Meteorology and Oceanography Center (FNMOC) Naval
Oceanography Portal - Met. and Ocean. Products. [Online]
Available at: http://www.usno.navy.mil/FNMOC/
USEPA, 2003. AERMOD: Latest Features and Evaluation Results, Research Triangle Park, NC:
U.S. EPA, OAQPA, EPA-454/R-03-003.
USEPA, 2004a. AERMOD: Description of model formulation, Research Triangle Park, North
Carolina: U.S. Environ. Protection Agency, EPA-454/R-03-004.
USEPA, 2004b. User's Guide for the AERMOD Meteorological Preprocessor AERMET,
Research Triangle Park, North Carolina: U.S. Environ. Protection Agency, EPA-454/B03-
002.
USEPA, 2004c. User's Guide fortheAMS/EPA Regulatory Model AERMOD, Research Triangle
Park, North Carolina: U.S. Environ. Protection Agency, EPA-454/B-03-001.
USEPA, 2012. User's Manual AERCOARE Version 1.0, Seattle, WA: U.S. Environ. Protection
Agency Region 10, EPA-910-R-12-008.
USN, 2014. Fleet Numerical Meteorology and Oceanography Center (FNMOC) Naval
Oceanography Portal - Met. and Ocean. Products. [Online]
Available at: http://www.usno.navy.mil/FNMOC/
Venkatram, A., 1980. Estimating the Monin-Obukhov Length in the Stable Boundary Layer for
Dispersion Calculations. Boundary Layer Meteor., Volume 19, pp. 481-485.
Vogelezang, D. & Holtslag, A., 1996. Evaluation and model impacts of alternative boundary-
layer height formulations. Boundary Layer Meteor., Volume 81, pp. 245-269.
209
-------
Wilson, A., Bromwich, D., Mines, K. & Landis, C., 2009. Enhancement of Polar WRF Arctic
Atmospheric and Surface Processes.. 18-21 May 2009, Madison, Wl, American
Meteorological Society.
Wong, H., 2011. CO ARE Bulk Flux Algorithm to Generate Hourly Meteorological Data for use
withAERMOD. Seattle(WA): U.S. EPA Region 10.
Wong, H., 2012. (personalcommunication). s.Ls.n.
Xu, F. & Ignatov, A., 2010. Evaluation of in situ sea surface temperatures for use in the
calibration and validation of satellite retrievals. J. Geophys. Res., Volume 115, pp. 1-18.
210
-------
APPENDIX A: TASK 3 PROTOCOL
-------
-------
Evaluate the AERMOD Performance Using
Predicted or Measured Meteorology
Task 3 Protocol
AMEC RFP # 12-6480110233-TC-3902
Federal Prime Contract # EP-W-09-028
Prepared for:
AMEC Environment & Infrastructure, Inc.
502 W. Germantown Pike, Suite 850
Plymouth Meeting, PA 19462-1308
Attention: Thomas Carr
Prepared by:
ENVIRON International Corporation
773 San Marin Drive, Suite 2115
Novato, California, 94945
www.environcorp.com
P-415-899-0700
F-415-899-0707
July 2, 2013
ENVIRON
-------
Overwater Dispersion Modeling
Task 3 Protocol
INTRODUCTION
The primary objective of the current study is to test and evaluate AERMOD on the outer
continental shelf (OCS). The current modeling procedures for sources on land use the
AERMOD modeling system. The meteorological AERMET processor included in the system is
inappropriate for OCS sources because the energy fluxes over water are not strongly driven by
diurnal heating and cooling. In addition, the meteorological observations necessary to drive the
dispersion models are commonly not available, especially in the Arctic Ocean. For applications
in the Arctic, the remote location and seasonal sea-ice pose logistical problems for the
deployment of buoys or offshore measurement platforms.
This study evaluates a combined modeling approach where the meteorological variables are
provided by a numerical weather prediction model, and then processed by a combination of a
new Mesoscale Model Interface program (MMIF) and, optionally, AERCOARE (a replacement
for AERMET suitable for overwater conditions). Given an appropriate overwater meteorological
dataset, AERMOD can then be applied for New Source Review following the same procedures
as used for sources over land.
The remainder of this document presents a protocol for Task 3 of the study. Task 3 generates a
WRF meteorological dataset for 2009-2011 suitable for dispersion modeling in the Arctic,
employs various combinations of MMIF and AERCOARE to extract modeled and observational
meteorology overwater, and uses that to drive AERMOD simulations.
The modeling period in the current protocol is 2009-2011 allowing for one year of overlap with
the BOEM/UAF 30-year WRF and Observational dataset. This overlapping period would allow
for a "reanalysis vs. hind-cast" comparison. It also allows for an approximately 1.5 year overlap
with the profiler data collected at Endeavor Island (Jun 2010 to Dec 2011). Extending the
simulation and meteorological analysis through 2012 would provide an extra year of comparison
with the profiler and may be considered if additional funding is available.
Task 3: Evaluate the use of WRF Solutions with AERMOD
AMEC and ENVIRON prepared a Work Plan outlining the various tasks and objectives of the
current study. As directed by EPA and AMEC, the third task protocol includes additional
information on data, options, and issues that were not fully described in the Work Plan. With an
approved protocol, ENVIRON staff will perform the following subtasks:
Task 3a: Generate WRF simulations for calendar years 2009 through 2011.
For these simulations, ENVIRON has selected the National Center for Atmospheric Research's
(NCAR's) community-developed WRF model (dynamical core version 3.4.1). WRF is a limited-
area, non-hydrostatic, terrain-following eta-coordinate mesoscale model.
ENVIRON's 3-year WRF simulation (2009-2011) will include 5.5 day simulation blocks (starting
December 15th, 2008), with 12 hours overlapping to account for model spin-up. The spin-up time
allows for the model to develop sub-grid scale processes, including vorticity and moisture fields.
Given the high latitude, the domains are defined on a polar stereographic map projection. The
outermost 36 km domain encompasses all of Alaska and parts of Northern Canada and Russia;
July 2, 2013 1 ENVIRON
-------
Overwater Dispersion Modeling
Task 3 Protocol
a 12 km nest including most of interior Alaska, the Bering, Chukchi, and Beaufort seas; and the
4 km nest focuses on the regions of the Chukchi and Beaufort seas containing active lease sites
and the entirety of Alaska's North Slope (see Figure 1). EPA guidance recommends a 50 km
buffer around CALPUFF sources and receptors, to allow for re-circulation of the puffs.
Additionally, the first five grid points on the edge of a nested WRF grid are contaminated by the
numerical down-scaling in WRF, and should not be used. Figure 2 illustrates 70 km (5 x 4km +
50km) buffers around the active lease areas with yellow dots (National Park Service, 2010).
ALASKA BOEM
150°E 150°W 90°W
100°W
160°E
- 110°W
170°E
120°W
180° 170°W 160°W 150°W 140°W 130°W
1 25 75 200 500 1000 1500 2000 3000
Figure 1. Proposed WRF 36km, 12km, and 4 km domains.
July 2, 2013
ENVIRON
-------
Overwater Dispersion Modeling
Task 3 Protocol
Active Lease Sites
72°N
70°N
68°N
66°N
64°N
62°N
180C
165°W
150°W
135°W
170°W
165°W 160°W 155°W 150°W 145°W
72°N
- 70°N
68°N
66°N
64°N
62°N
140°W
1 25 75 200 500 1000 1500 2000 3000
Figure 2.12 km and 4 km WRF domains, with lease sites (magenta), the Arctic National
Wildlife Refuge (orange), Class I Areas (red), and 70 km buffers from active lease sites
(yellow dots).
The planned model vertical structure maintains the 37 vertical levels from the Task 2 WRF
modeling. Layers are stacked toward the surface to capture the coastal boundary layer and
sharp arctic wintertime temperature inversions (see Table 1). ENVIRON anticipates that the fine
vertical spacing will help winds and temperatures respond more explicitly to dynamical
influences.
July 2, 2013
ENVIRON
-------
Overwater Dispersion Modeling
Task 3 Protocol
Table 1. WRF model 37 vertical levels with approximate heights AGL.
Level
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
26
27
28
29
30
31
32
33
34
35
36
37
eta
1
0.9985
0.997
0.995
0.993
0.991
0.988
0.985
0.98
0.97
0.96
0.95
0.94
0.93
0.91
0.89
0.87
0.84
0.8
0.76
0.72
0.68
0.64
0.6
0.55
0.5
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.06
0.027
0
Pressure
(mb)
1000
999
997
996
994
992
989
987
982
973
964
955
946
937
919
901
883
856
820
784
748
712
676
640
595
550
505
460
415
370
325
280
235
190
154
124
100
Level Height
(m)
0.0
10.8
21.6
36.0
50.5
65.0
86.7
108.5
145.0
218.3
292.1
366.5
441.5
517.1
670.0
825.3
983.2
1225.0
1557.1
1901.3
2258.5
2630.0
3016.9
3421.0
3952.7
4518.1
5122.3
5771.8
6475.0
7242.8
8090.5
9039.1
10120.5
11385.0
12585.4
13761.3
14907.1
Mid-layer
Height
(m)
5.4
16.2
28.8
43.3
57.7
75.9
97.6
126.8
181.6
255.2
329.3
404.0
479.3
593.5
747.6
904.3
1104.1
1391.0
1729.2
2079.9
2444.3
2823.5
3219.0
3686.9
4235.4
4820.2
5447 .1
6123.4
6858.9
7666.7
8564.8
9579.8
10752.8
11985.2
13173.4
14334.2
Layer
Thickness
(m)
10.8
10.8
14.4
14.5
14.5
21.8
21.8
36.4
73.3
73.8
74.4
75.0
75.6
152.9
155.3
157.9
241.8
332.2
344.2
357.2
371.4
387.0
404.0
531.8
565.4
604.2
649.5
703.2
767.9
847.6
948.7
1081.4
1264.5
1200.4
1175.9
1145.8
Note: Calculated using P0=1 000mb, Ptop=100mb, T0=O.OOC, and dT/dx=-6.5C/km.
July 2, 2013
ENVIRON
-------
Overwater Dispersion Modeling
Task 3 Protocol
WRF must be optimized to simulate coastal arctic weather. To do this, ENVIRON's model
configuration should include the most accurate initial inputs, regionally applicable physics
choices, and effective nudging, combined with the best SST's, sea ice, and land surface models
available. ENVIRON's WRF will build upon the successful application of WRF to reanalyze 30
years (1979-2009) of arctic meteorology prepared by BOEM-UAF specifically to study surface
winds (Krieger, Zhang, Shulski, Fuhong, & Tao, 2012). UAF's method used data assimilation to
generate hourly reanalyses. By contrast, ENVIRON proposes running WRF as a hind-cast,
initializing the model from ECMWF reanalysis grids and running it for a 5.5-day period, using a
combination of 3D analysis nudging and observational nudging. ENVIRON's approach is very
similar to many other WRF model applications used to support photochemical grid modeling in
many parts of the country.
Table 2 shows ENVIRON's proposed WRF hind-cast treatments relative to the BOEM-UAF
reanalysis. The treatment of sea ice is critical to WRF modeling success. The BOEM-UAF
reanalysis employs modifications from a variant of WRF named "Polar WRF (Byrd Polar
Research Center, 2013) to the standard WRF package, which ENVRON also proposes to use.
Although WRFv3.5 has been released, the Polar WRF modifications have not yet been made to
v3.5, and ENVIRON proposes to use Polar WRFv3.4.1 instead.
ENVIRON will also improve upon the 24 km CMC sea ice by ingesting ~4 km gridded snow and
sea ice datasetfrom the National Ice Center (NIC) Ice Mapping System (IMS) available post-
2004 (National Ice Center, 2008). ENVIRON will employ the Morrison microphysics scheme,
which was designed specifically for arctic applications but has documented success at mid-
latitudes as well. ENVIRON concurs with the BOEM-UAF selection of the Rapid Radiative
Transfer Model for GCMs (RRTMG) radiation option, Monin-Obukhov (Janjic) surface layer
scheme, NOAH land surface model (with polar modifications), and the TKE-based Mellor-
Yamada-Janjic (MYJ) planetary boundary layer scheme. Other ENVIRON sensitivity studies for
stable boundary layers in Alaska (for a confidential client) and Wyoming (Hahn, Brashers,
Emery, & McNally, 2012) indicated superior vertical profiles of temperature and moisture using
MYJ compared to YSU and other planetary boundary layer schemes. The BOEM-UAF
reanalysis employs the relatively un-tested Grell-3D cumulus scheme. ENVIRON performed
sensitivity studies in the four corners region, and found the Grell-3D scheme produced
extremely minimal convection during the summer compared to PRISM data. Thus, ENVIRON
proposes to use the Kain-Fritch cumulus scheme on the 36 and 12 km domains, with explicit
convection (no parameterized scheme) on the 4km domain. ENVIRON will update SST's daily,
calculate the skin SST, and update deep soil temperatures following the usual WRF procedures.
Model inputs will use the ERA-interim European Center for Medium Range Weather Forecast's
ERA-lnterim dataset (ERA-I, 6-hourly analysis output, ~0.75°x0.75° degree resolution).
Traditionally, ENVIRON recommends 3-D nudging toward analysis grids for wind, temperature,
and moisture for the 36 and 12 km domains. Analysis nudging within the PEL can prevent the
natural, dynamic development of boundary layer processes. Therefore, ENVIRON's strategy
July 2, 2013 5 ENVIRON
-------
Overwater Dispersion Modeling
Task 3 Protocol
involves 3-D analysis nudging above the PEL for the 36km domain, and observational nudging
against DS-35051 data on the 4km domain. 3-D analysis nudging of the 12 km using the 0.75°
ERA-I data directly would likely degrade model performance. Recent ENVIRON experience
with the WRF system's OBSGRID program in the data-sparse Four Corners region showed that
model performance was degraded when analysis nudging with OBSGRID output was used.
Table 2. BOEM-UAF 30-year vs ENVIRON proposed Model Options
Treatment
WRF Version
Snow/sea ice
Boundary Conditions
Snow
Sea Ice
Microphysics
Radiation
Surface Layer
Land Surface Model
PEL
Cumulus
Time-varying SST
Calculate skin SST
Update Deep Soil
Temp
Fractional Sea Ice
Tice2tsk if2cold
Nudging
FDDA
Obs. nudging
BOEM-UAF REANALYSIS
V3.2.1
BOEM modified version of Polar WRF
codes for snow/sea ice processes
ERA-lnterim (0.75° grid spacing)
Canadian Meteorological Centre
(CMC) daily snow depth (24 km)
AMSR-E daily sea ice
concentration/thickness (12.5 km)
Morrison
RRTMG shortwave and longwave
Monin-Obukhov (Janjic) scheme
Noah with Polar WRF modifications
MYJ TKE
Grell-3D
On
On
Yes
Yes
True
Spectral
Yes
No
ENVIRON 2009-2011
V3.4.1
All Polar WRF modules
ERA-lnterim (0.75° grid spacing)
IMS 4-km NH daily snow
IMS 4-km NH daily (sea ice)
Morrison
RRTMG shortwave and longwave
Monin-Obukhov (Janjic) scheme
Noah with Polar WRF modifications
MYJ TKE
Kain-Fritch (36/1 2km only)
On
On
Yes
IMS 4-km dataset
True
Spectral (u, v, theta, geopotential, and
moisture)
No
Nudge toward DS-3505 data
1 DS-3505 integrated surface hourly (ISH) worldwide station data includes extensive automated QC on all data and
additional manual QC for USAF, NAVY, and NWS stations. It integrates all data from DS9956, DS3280, and DS3240.
10,000 currently active stations report wind speed and direction, wind gust, temperature, dew point, cloud data,
sea level pressure, altimeter setting, station pressure, present weather, visibility, precipitation amounts for various
tie periods, snow depth, and various other elements as observed by each station. (NOAA/NCDC, 2010)
July 2, 2013
ENVIRON
-------
Overwater Dispersion Modeling
Task 3 Protocol
Successful application of spectral nudging in the BOEM-UAF 30-year reanalysis dataset
warrants an attempt in this study. Spectral nudging is relatively new in WRF, and ENVIRON
proposes a limited nudging sensitivity test for February and August 2009 evaluating spectral
versus analysis nudging (Table 3). Spectral nudging configurations will be guided by the
parameters in Otte et al. (2012), that suggest limiting nudging above the tropopause and
adopting a ~6 hourtimescale. One other limited sensitivity study will be performed comparing
analysis nudging both with and without using objectively analyzed input files during February
and August 2009. ENVIRON will also perform observational nudging (winds only) using NCDC
DS-3505 data on the 4 km domain, using a 50 km radius of influence. ENVIRON excludes
nudging of temperature because in coastal areas it may weaken land-sea temperature contrasts
and adversely affect model performance; the majority of North Slope observational assets
reside along the coastline. Table 3 presents proposed relevant nudging parameters.
Table 3. Proposed WRF nudging coefficients
Nudging
Spectral
3-D Analysis
(if required)
2-D Surface
Observational
Domains
Applied
36/1 2/4 km
36 km
None
4km
Nudging Strengths (1/s)
Wind
~6 h timescale
35 10~4
65 10'4
Temperature (no PBL)
~6 h timescale
35 10~4
None
Humidity (no PBL)
~6 h timescale
35 10~4
None
The 2009-2011 WRF simulation will be subjected to a model performance evaluation using the
METSTAT program to evaluate temperatures, winds, and humidity. METSTAT uses surface
meteorological observations and extracted WRF data (paired in time and space) to calculate a
series of statistical measures designed to examine WRF's ability to characterize the
observations. ENVIRON will evaluate the model against as full an observed dataset as feasible,
including DS-3505 data for 2009-2011 and the BOEM-UAF observational dataset for 2009
(extended as feasible) if obtained from UAF. Data contained in the BOEM-UAF dataset, but not
in the DS-3505 data will serve as an independent verification of model performance as this data
was not used for nudging. Additionally, the off-shore buoys analyzed in Task 1 represent
independent verification as they are not contained in the DS-3505 dataset.
Additionally, ENVIRON will employ METSTAT to evaluate the single overlapping year, 2009,
from the 30-year BOEM-UAF simulation. This will enable a direct comparison of the
meteorology for that year.
Task 3b: Extract meteorological datasets using WRF solutions for sites with overwater
observations in the Arctic
ENVIRON will extract WRF solutions from five buoy locations in the Arctic (two sites in the
Chukchi Sea and three in the Beaufort Sea) as well as for the Endeavor Island profiler location
July 2, 2013
ENVIRON
-------
Overwater Dispersion Modeling
Task 3 Protocol
(Figure 3). As proposed in Task 1, ENVIRON will evaluate the four extractions and processing
options as follows:
1. MMIF will be applied to extract and prepare data sets for direct use by AERMOD. All
variables will be as-predicted by the WRF simulations including the surface energy fluxes,
surface roughness and planetary boundary layer (PBL) height.
2. As in Option 1), but the PBL height will be re-diagnosed from the wind speed and potential
temperature profiles using the Bulk-Richardson algorithm within MMIF. Based on the
August to October 2010 monitoring data collected by Shell in the Beaufort Sea, PBL
heights range from 10 m to 700 m, with a median height of 80 m. AERMOD simulations
can be very sensitive to the PBL height (Richmond & Morris, 2012) and MMIF-processed
PBL heights may provide significantly different predicted concentrations than the PBL
height diagnosed by WRF.
3. MMIF will be applied to extract the key meteorological variables of overwater wind speed,
wind direction, temperature, humidity, and PBL height. AERCOARE will use these
variables to predict the surface energy fluxes, surface roughness length and other
variables needed for the AERMOD simulations. AERCOARE has a surface layer scheme
developed specifically to predict surface fluxes from overwater measurements. In this
application, the WRF simulations provide model-derived alternatives for variables
measured by a buoy, ship or offshore platform. AERCOARE can also be applied using a
number of different options. For the current study, we propose to apply AERCOARE using
the defaults recommended in the AECOARE model evaluations study (Richmond & Morris,
2012).
4. As in Option 3), but the PBL height will be re-diagnosed using the Bulk-Richardson
algorithm within MMIF. AERCOARE will be applied as in Option 3.
July 2, 2013 8 ENVIRON
-------
Overwater Dispersion Modeling
Task 3 Protocol
Figure 3. Buoy sites in the Chukchi and Beaufort Seas; Endeavor Island profiler location.
Task 3c: AERCOARE using buoy and profiler data
ENVIRON will use surface input from the buoys in Figure 3 to drive AERCOARE. The buoys
provide wind speed, wind direction, temperature (at various heights), relative humidity, and sea
surface temperature.
The vertical temperature profiler at Endeavor Island will be used to extract hourly mixing heights
from April 2010 (when the profiler was installed) through 2011. These mixing heights will be
provided to AERCOARE to replace WRF-diagnosed and AERCOARE-diagnosed mixing heights
options 3 and 4, respectively, of Task 3b.
Task 3d: AERMOD
ENVIRON will conduct AERMOD simulations for the six OCS hypothetical sources using the
output from the modeling-based approaches (MMIF/AERCOARE as described in Task 3b) to
drive AERMOD. These will be compared directly against AERMOD driven by the buoy and
profiler extraction in Task 3c. Simulations involving AERCOARE will be confined to the open
water time periods of 2009-2011, whereas MMIF/AERMOD runs will be performed for all
months. The analysis will calculate the predicted concentrations over relevant averaging periods
for the criteria pollutants (e.g., 1-hour, 8-hour, 24-hour, and period) for OCS sources. As in Task
1, the sources will be modeled at the buoy and profiler location, with a polar grid of receptors
located along 360 equidistant radii at radial distances (30 m, 50 m, 75 m, 100 m, 125 m, 150 m,
July 2, 2013
ENVIRON
-------
Overwater Dispersion Modeling
Task 3 Protocol
175 m, 200 m, 300 m, 400 m, 500 m, 750 m, 1 km, 1.5 km, 2 km, 3 km, 4 km, 5 km, 6 km, 7 km,
8 km, 9 km, and 10 km) from the source.
Task 3e: AERMOD evaluation
ENVIRON will evaluate the model-driven AERMOD performance in Task 3d against the
observationally-driven AERMOD runs using contour plots, scatter diagrams, Q-Q plots, and
statistics as necessary.
Task 3f: Conclusion from evaluation
ENVIRON will diagnose how various aspects of the modeling procedures influenced the
prediction of concentration (in particular maximum concentration) by assessing (1) the
meteorology and modeling performance influence, (2) MMIF recalculated PEL height vs WRF
PEL height, and (3) MMIF with AERCOARE or MMIF fed directly into AERMOD.
July 2, 2013 10 ENVIRON
-------
Overwater Dispersion Modeling
Task 3 Protocol
REFERENCES
NOAA/NCDC. (2010, 09 15). Retrieved 02 15, 2013, from
http://www.ncdc.noaa.gov/oa/climate/rcsg/datasets.htmlWsurface
Bowden, J. H., Otte, T. L, Nolte, C. G., & Otte, M. J. (2012). Examining interior grid nudging
techniques using two-way nesting in the wrf model for regional climate modeling. J.
Climate, 25, 2805-2823.
Byrd Polar Research Center. (2013, 03 06). The Polar WRF. Retrieved 06 24, 2013, from Ohio
State University: http://polarmet.osu.edu/PolarMet/pwrf.html
Environ. (2010). Evaluation of the COARE-AERMOD Alternative Modeling Approach Support for
Simulation of Shell Exploratory Drilling Sources In the Beaufort and Chukchi Seas.
Hahn, R., Brashers, B., Emery, C., & McNally, D. (2012). Winter 2008 WRF Modeling of the Upper
Green River Basin. Novato, CA: ENVIRON.
Krieger, J., Zhang, J., Shulski, M., Fuhong, L., & Tao, W. a. (2012). Toward Producing a
Beaufort/Chukchi Seas Regional Reanalysis. Retrieved 06 12, 2013, from
Beaufort/Chukchi Seas Mesoscale Meteorology Modeling Study: http://mms-
meso.gi.alaska.edu/pub/amssl2-krieger-presentation.pdf
McNally, D., & Wilkinson, J. G. (2011). Model Application and Evaluation: ConocoPhillips Chukchi
Sea WRF Model Application. Arvada, Colorado: Alpine Geophysics, LLC.
National Ice Center, 2008. u. (n.d.). IMS daily Northern Hemisphere snow and ice analysis at 4
km and 24 km resolution. Retrieved 06 18, 2013, from National Snow and Ice Data
Center.: http://dx.doi.org/10.7265/N52R3PMC
National Park Service. (2010). Federal Land Managers' Air Quality Related Values Work Group
(FLAG) Phase 1 Report—Revised. Retrieved 06 19, 2013, from
http://www.nature.nps.gov/air/Pubs/pdf/flag/FLAG_2010.pdf
Otte, T. L., Otte, M. J., Bowden, J. H., & Nolte, C. G. (2012). Sensitivity of Spectral Nudging
Toward Moisture. US Environmental Protection Agency.
Richmond, K., & Morris, R. (2012, 10). Evaluation of the Combined AERCOARE/AERMOD
Modeling Approach for Offshore Sources. Retrieved 06 19, 2013, from EPA:
http://www.epa.gov/ttn/scram/models/relat/aercoare/AERCOARE-Model-
Evaluation.pdf
Zhang, J. (2011). Beaufort and Chukchi Seas Mesoscale Meteorology Model Study.
July 2, 2013 11 ENVIRON
-------
-------
APPENDIX B: AERMOD RESULTS STATISTICAL SCORES
-------
AERMOD results statistical scores: 1-hour average concentrations.
site
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
year
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
source
1
1
1
1
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
avg
time
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
Simulation
type
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
geo
mean
17.20
7.38
7.68
7.90
7.96
42.74
18.58
19.75
19.85
20.63
65.40
29.37
27.37
17.70
18.39
22.75
2.24
1.72
1.78
2.35
237.59
247.27
247.71
253.77
253.77
geo
stddev
2.57
9.31
8.73
8.38
8.59
2.31
8.53
7.56
7.46
7.58
2.40
10.84
10.90
17.44
17.58
4.47
111.84
113.43
90.33
89.86
3.27
3.09
3.08
3.23
3.23
MG
1.00
2.33
2.24
2.18
2.16
1.00
2.30
2.16
2.15
2.07
1.00
2.23
2.39
3.69
3.56
1.00
10.14
13.21
12.81
9.70
1.00
0.96
0.96
0.94
0.94
VG
1.00
19.93
15.04
13.84
14.83
1.00
18.59
11.37
11.28
11.04
1.00
42.05
46.94
1078.48
1092.87
1.00
>5000
>5000
>5000
>5000
1.00
1.01
1.01
1.01
1.01
RHC
50.02
41.24
39.17
38.67
40.47
101.31
87.11
86.14
85.99
88.59
190.48
170.50
166.69
133.95
138.34
79.69
86.05
47.48
45.01
79.21
1224.60
1131.49
1131.49
1221.71
1221.71
geo
R
1.00
0.85
0.86
0.85
0.84
1.00
0.86
0.87
0.86
0.86
1.00
0.80
0.81
0.73
0.72
1.00
0.80
0.81
0.77
0.79
1.00
1.00
1.00
1.00
1.00
ff2
1.00
0.70
0.72
0.72
0.72
1.00
0.76
0.76
0.76
0.76
1.00
0.72
0.72
0.68
0.66
1.00
0.60
0.62
0.60
0.60
1.00
1.00
1.00
1.00
1.00
TMS
0.90
0.57
0.61
0.62
0.61
0.80
0.59
0.63
0.63
0.63
0.81
0.58
0.59
0.50
0.53
0.82
0.49
0.41
0.48
0.41
0.80
0.98
0.99
0.97
0.98
-------
site
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B3
B3
year
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2010
2010
source
1
1
1
1
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
avg
time
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
Simulation
type
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
geo
mean
10.67
14.05
9.00
9.94
14.13
30.51
32.28
20.56
22.43
32.11
64.45
30.24
14.68
17.12
29.14
10.60
11.40
3.02
3.61
10.95
220.04
237.45
237.53
220.92
218.45
12.54
26.79
geo
stddev
5.43
3.39
7.00
6.62
3.72
3.98
3.30
6.85
6.37
3.59
2.88
6.54
19.57
16.42
7.68
16.71
8.78
45.55
34.86
10.06
3.26
3.00
2.99
3.01
2.99
4.47
1.50
MG
1.00
0.76
1.19
1.07
0.76
1.00
0.95
1.48
1.36
0.95
1.00
2.13
4.39
3.77
2.21
1.00
0.93
3.51
2.93
0.97
1.00
0.93
0.93
1.00
1.01
1.00
0.47
VG
1.00
1.42
1.53
1.40
1.32
1.00
1.06
2.13
1.78
1.04
1.00
6.81
3469.96
958.47
11.02
1.00
1.92
199.55
62.10
1.54
1.00
1.02
1.02
1.01
1.01
1.00
7.59
RHC
48.74
42.31
42.08
45.18
45.63
116.19
89.03
85.86
90.39
90.33
186.80
122.01
122.01
130.24
138.14
84.14
46.12
45.36
46.62
62.37
1135.76
1021.43
1021.43
1017.11
1008.37
47.32
44.94
geo
R
1.00
0.99
0.95
0.95
0.99
1.00
1.00
0.94
0.95
0.99
1.00
0.83
0.64
0.65
0.81
1.00
0.98
0.87
0.88
0.99
1.00
1.00
1.00
1.00
1.00
1.00
0.79
ff2
1.00
0.82
0.88
0.90
0.86
1.00
1.00
0.82
0.84
1.00
1.00
0.68
0.58
0.62
0.70
1.00
0.76
0.66
0.60
0.76
1.00
1.00
1.00
1.00
1.00
1.00
0.76
TMS
0.87
0.83
0.86
0.89
0.87
0.80
0.93
0.77
0.81
0.98
0.81
0.56
0.49
0.49
0.60
0.82
0.75
0.56
0.56
0.82
0.80
0.96
0.98
1.00
0.99
0.86
0.62
-------
site
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
year
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2011
2011
2011
2011
source
1
1
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
avg
time
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
Simulation
type
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
geo
mean
25.74
21.14
21.93
34.58
57.85
54.89
46.38
48.48
50.37
84.50
76.22
63.99
72.18
10.63
40.30
33.80
27.16
31.76
249.72
248.69
248.75
248.76
248.95
13.89
18.54
9.95
10.31
geo
stddev
1.50
1.88
1.96
3.41
1.47
1.51
1.75
1.77
3.59
1.61
1.72
2.07
1.90
14.23
1.74
1.69
2.15
2.42
3.16
3.12
3.12
3.20
3.20
3.98
2.07
5.73
5.78
MG
0.49
0.59
0.57
1.00
0.60
0.63
0.75
0.71
1.00
0.60
0.66
0.79
0.70
1.00
0.26
0.32
0.39
0.34
1.00
1.00
1.00
1.00
1.00
1.00
0.75
1.40
1.35
VG
6.94
3.08
2.98
1.00
3.47
2.95
1.85
1.90
1.00
3.68
2.86
1.55
1.91
1.00
1117.46
785.50
122.15
116.46
1.00
1.00
1.00
1.00
1.00
1.00
1.83
1.87
1.84
RHC
41.33
40.89
44.56
96.26
90.62
83.60
82.83
87.36
157.18
139.96
142.12
122.45
125.70
62.18
68.67
69.92
56.53
73.54
1236.04
1244.94
1244.94
1190.27
1193.66
45.51
45.11
41.35
43.44
geo
R
0.81
0.95
0.95
1.00
0.71
0.80
0.94
0.93
1.00
0.68
0.76
0.96
0.93
1.00
0.72
0.71
0.91
0.91
1.00
1.00
1.00
1.00
1.00
1.00
0.96
0.92
0.92
ff2
0.76
0.80
0.80
1.00
0.80
0.80
0.84
0.84
1.00
0.82
0.82
0.86
0.84
1.00
0.72
0.72
0.74
0.74
1.00
1.00
1.00
1.00
1.00
1.00
0.84
0.80
0.80
TMS
0.62
0.73
0.72
0.80
0.67
0.70
0.81
0.79
0.81
0.65
0.71
0.82
0.79
0.82
0.52
0.55
0.57
0.55
0.80
1.00
1.00
0.99
1.00
0.86
0.82
0.78
0.79
-------
site
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
year
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2012
2012
2012
2012
2012
2012
source
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
avg
time
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
Simulation
type
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
geo
mean
15.94
32.76
42.60
23.89
24.57
37.24
42.07
54.66
23.39
24.57
41.32
11.72
26.24
4.99
5.37
18.90
163.62
169.68
189.57
190.37
160.45
14.65
15.56
15.58
13.47
13.42
38.33
geo
stddev
2.80
3.42
2.01
5.22
5.28
2.66
3.90
2.42
11.78
11.11
3.99
13.49
2.74
30.16
25.76
4.61
3.39
3.11
3.14
3.14
3.08
3.26
2.95
3.02
3.62
3.64
2.65
MG
0.87
1.00
0.77
1.37
1.33
0.88
1.00
0.77
1.80
1.71
1.02
1.00
0.45
2.35
2.18
0.62
1.00
0.96
0.86
0.86
1.02
1.00
0.94
0.94
1.09
1.09
1.00
VG
1.28
1.00
1.52
1.69
1.67
1.18
1.00
1.40
8.21
6.88
1.12
1.00
27.69
18.59
15.12
4.44
1.00
1.04
1.07
1.07
1.03
1.00
1.06
1.08
1.05
1.06
1.00
RHC
45.10
90.78
90.68
83.61
90.70
91.87
139.03
120.83
121.60
130.21
135.59
61.43
62.91
53.40
53.07
68.31
1116.54
1072.03
1016.64
1007.06
1066.92
42.63
36.87
39.34
40.40
38.47
90.59
geo
R
0.96
1.00
0.96
0.94
0.94
0.97
1.00
0.98
0.92
0.92
0.97
1.00
0.98
0.91
0.90
0.99
1.00
0.99
0.99
0.99
0.99
1.00
0.98
0.98
0.99
0.99
1.00
ff2
0.86
1.00
0.86
0.76
0.78
0.90
1.00
0.88
0.74
0.76
0.94
1.00
0.76
0.68
0.68
0.82
1.00
1.00
1.00
1.00
1.00
1.00
0.98
0.94
1.00
1.00
1.00
TMS
0.89
0.81
0.85
0.79
0.80
0.92
0.81
0.84
0.67
0.67
0.95
0.84
0.64
0.58
0.62
0.69
0.80
0.98
0.95
0.95
0.98
0.85
0.94
0.94
0.97
0.96
0.80
-------
site
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
C1
C1
C1
C1
C1
C1
C1
C1
year
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2011
2011
2011
2011
2011
2011
2011
2011
source
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
avg
time
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
Simulation
type
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
geo
mean
36.21
35.84
31.69
32.27
67.63
51.15
42.07
35.88
38.31
17.15
15.61
16.03
11.21
11.51
243.59
242.22
242.20
243.06
243.08
22.90
23.25
22.82
18.57
19.03
53.56
54.47
53.82
geo
stddev
2.60
2.65
3.12
3.22
2.19
2.99
4.18
5.43
5.35
7.59
7.35
7.49
10.92
11.54
3.16
3.09
3.09
3.18
3.18
1.64
1.60
1.59
2.16
2.13
1.57
1.53
1.48
MG
1.06
1.07
1.21
1.19
1.00
1.32
1.61
1.89
1.77
1.00
1.10
1.07
1.53
1.49
1.00
1.01
1.01
1.00
1.00
1.00
0.99
1.00
1.23
1.20
1.00
0.98
1.00
VG
1.02
1.02
1.10
1.11
1.00
1.37
2.53
5.30
4.59
1.00
1.04
1.09
1.49
1.50
1.00
1.00
1.00
1.00
1.00
1.00
1.01
1.01
1.20
1.16
1.00
1.01
1.01
RHC
85.46
80.39
80.23
94.42
162.17
125.46
128.83
137.85
132.11
74.14
54.01
52.29
48.99
56.02
1187.88
1200.45
1199.78
1195.56
1196.11
42.62
44.02
43.41
44.76
45.99
91.27
88.60
83.99
geo
R
0.99
0.99
0.99
0.98
1.00
0.92
0.87
0.83
0.85
1.00
1.00
0.99
0.99
0.99
1.00
1.00
1.00
1.00
1.00
1.00
0.98
0.98
0.92
0.94
1.00
0.98
0.97
ff2
1.00
1.00
0.98
0.96
1.00
0.82
0.76
0.72
0.76
1.00
1.00
0.98
0.88
0.78
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.86
0.92
1.00
1.00
1.00
TMS
0.97
0.97
0.94
0.91
0.81
0.80
0.72
0.64
0.67
0.82
0.92
0.96
0.83
0.80
0.80
1.00
1.00
1.00
1.00
0.90
0.99
0.99
0.88
0.90
0.81
0.99
0.98
-------
site
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
year
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
source
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
2
avg
time
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
Simulation
type
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
geo
mean
44.54
45.53
76.33
75.99
79.95
67.49
62.97
37.58
34.16
34.14
24.58
26.47
203.91
235.84
236.12
238.63
238.57
18.40
17.56
17.13
14.55
14.72
42.13
39.85
38.68
32.34
33.95
geo
stddev
1.90
1.93
1.66
1.76
1.72
2.02
2.21
1.58
1.84
1.86
2.90
2.82
3.30
3.09
3.09
3.18
3.18
2.31
2.64
2.73
3.59
3.44
2.21
2.47
2.55
3.49
3.15
MG
1.20
1.18
1.00
1.00
0.96
1.13
1.21
1.00
1.10
1.10
1.53
1.42
1.00
0.87
0.86
0.85
0.86
1.00
1.05
1.07
1.27
1.25
1.00
1.06
1.09
1.30
1.24
VG
1.12
1.12
1.00
1.03
1.03
1.14
1.30
1.00
1.09
1.11
2.09
1.85
1.00
1.08
1.08
1.08
1.08
1.00
1.03
1.06
1.38
1.29
1.00
1.03
1.05
1.39
1.24
RHC
84.81
90.63
137.67
133.93
153.77
135.53
134.48
64.63
63.70
63.74
63.05
62.64
1214.21
1231.18
1234.44
1229.09
1226.75
44.65
43.26
41.33
43.96
45.00
89.88
87.34
82.20
85.39
88.30
geo
R
0.93
0.93
1.00
0.96
0.96
0.89
0.82
1.00
0.91
0.88
0.81
0.84
1.00
0.98
0.98
0.98
0.98
1.00
0.99
0.99
0.97
0.98
1.00
0.99
0.99
0.98
0.98
ff2
0.94
0.94
1.00
1.00
1.00
0.92
0.88
1.00
0.96
0.96
0.86
0.86
1.00
0.92
0.92
0.94
0.94
1.00
0.98
0.96
0.84
0.84
1.00
0.98
0.98
0.84
0.86
TMS
0.92
0.91
0.81
0.98
0.95
0.89
0.86
0.83
0.94
0.93
0.76
0.79
0.80
0.93
0.94
0.94
0.94
0.86
0.97
0.96
0.85
0.87
0.81
0.97
0.96
0.85
0.88
-------
site
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
year
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
source
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
2
3
avg
time
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
Simulation
type
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
geo
mean
71.36
51.14
50.63
35.43
36.52
21.74
18.77
18.37
12.29
12.93
167.73
213.14
215.91
204.43
198.55
18.04
17.67
13.36
13.09
16.09
41.16
39.87
31.60
31.34
36.33
47.91
geo
stddev
1.83
3.23
3.32
6.25
5.98
3.90
5.10
5.50
10.08
9.21
3.26
3.11
3.11
3.14
3.12
2.34
2.90
4.56
4.66
3.62
2.30
2.51
3.65
3.83
3.12
3.04
MG
1.00
1.40
1.41
2.02
1.95
1.00
1.16
1.18
1.77
1.68
1.00
0.79
0.78
0.82
0.85
1.00
1.02
1.35
1.38
1.12
1.00
1.03
1.30
1.31
1.13
1.00
VG
1.00
1.96
2.06
13.94
11.77
1.00
1.12
1.18
3.90
3.16
1.00
1.11
1.12
1.07
1.06
1.00
1.10
1.92
2.02
1.30
1.00
1.03
1.39
1.46
1.15
1.00
RHC
149.29
130.75
127.26
131.36
136.11
58.66
51.73
61.86
57.34
49.27
1193.22
1258.86
1248.61
1235.46
1243.41
45.51
45.06
42.25
42.46
45.55
88.95
91.15
89.15
90.76
91.75
120.94
geo
R
1.00
0.83
0.83
0.71
0.72
1.00
1.00
1.00
0.98
0.98
1.00
0.98
0.98
0.99
0.99
1.00
0.97
0.95
0.95
0.97
1.00
0.99
0.98
0.98
0.98
1.00
ff2
1.00
0.82
0.82
0.74
0.74
1.00
0.94
0.90
0.74
0.74
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.82
0.76
0.82
1.00
1.00
0.84
0.82
0.86
1.00
TMS
0.81
0.75
0.76
0.60
0.60
0.82
0.92
0.88
0.69
0.70
0.80
0.92
0.93
0.95
0.95
1.00
0.97
0.79
0.79
0.88
0.81
0.98
0.86
0.85
0.92
0.81
C2 2010
1hr AERC.RCALT 60.21
2.67 0.80 1.08 134.54 1.00 0.96 0.91
-------
site
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
year
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
source
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
2
3
3
3
3
avg
time
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
Simulation
type
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
geo
mean
43.77
42.05
48.02
20.92
21.01
8.48
8.81
14.59
197.72
216.21
219.05
211.01
206.13
15.12
11.86
9.27
9.20
11.29
36.19
28.58
24.22
23.86
27.74
58.62
38.19
29.51
30.66
geo
stddev
5.04
5.52
4.06
3.56
4.71
18.59
15.31
8.95
3.03
3.07
3.08
3.10
3.08
2.77
4.22
6.02
5.85
4.60
2.62
3.81
5.11
5.02
4.11
2.13
3.72
9.95
8.86
MG
1.10
1.14
1.00
1.00
1.00
2.47
2.38
1.43
1.00
0.92
0.90
0.94
0.96
1.00
1.27
1.63
1.64
1.34
1.00
1.27
1.49
1.52
1.31
1.00
1.54
1.99
1.91
VG
1.33
1.49
1.11
1.00
1.14
43.51
25.88
2.96
1.00
1.01
1.02
1.01
1.01
1.00
1.34
2.76
2.67
1.52
1.00
1.27
1.96
1.94
1.38
1.00
1.96
24.94
16.22
RHC
130.93
130.10
130.72
52.92
76.10
46.15
45.74
76.99
1164.51
1228.07
1229.25
1225.01
1224.13
41.90
40.58
36.82
38.37
42.99
92.68
90.74
86.95
91.43
92.20
131.37
125.13
137.90
135.46
geo
R
0.99
0.99
0.99
1.00
0.99
0.97
0.95
0.98
1.00
1.00
1.00
1.00
1.00
1.00
0.98
0.95
0.95
0.98
1.00
0.98
0.98
0.98
0.98
1.00
0.91
0.89
0.90
ff2
0.90
0.88
0.92
1.00
0.94
0.76
0.74
0.74
1.00
1.00
1.00
1.00
1.00
1.00
0.78
0.76
0.76
0.78
1.00
0.78
0.80
0.80
0.78
1.00
0.68
0.82
0.82
TMS
0.91
0.88
0.96
0.83
0.90
0.55
0.63
0.67
0.80
0.97
0.98
0.99
0.99
0.88
0.85
0.72
0.73
0.81
0.80
0.86
0.78
0.78
0.85
0.81
0.74
0.63
0.66
-------
site
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
year
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
source
3
4
4
4
4
4
5
5
5
5
5
avg
time
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
1hr
Simulation
type
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
geo
mean
35.12
19.08
8.98
4.42
5.34
8.16
191.96
234.80
235.54
203.59
201.47
geo
stddev
4.29
4.23
12.10
43.19
27.99
13.16
3.09
3.05
3.05
3.08
3.08
MG
1.67
1.00
2.12
4.32
3.58
2.34
1.00
0.82
0.82
0.94
0.95
VG
2.61
1.00
5.94
4042.10
408.55
8.78
1.00
1.06
1.07
1.02
1.01
RHC
135.27
62.56
55.29
51.89
49.03
53.37
1112.40
1106.07
1106.12
1102.10
1102.02
geo
R
0.91
1.00
0.99
0.93
0.92
0.98
1.00
0.99
0.99
1.00
1.00
ff2
0.70
1.00
0.66
0.70
0.70
0.64
1.00
1.00
1.00
1.00
1.00
TMS
0.72
0.82
0.63
0.56
0.57
0.62
0.80
0.95
0.95
0.98
0.99
-------
AERMOD results statistical scores: 3-hour average concentrations.
site
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
year
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2011
2011
source
1
1
1
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
avg
time
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
Simulation
type
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
geo
mean
5.85
5.75
6.29
6.46
26.13
15.98
15.16
15.78
17.30
39.12
18.79
17.33
12.78
14.16
14.24
1.60
1.08
1.19
1.60
162.50
223.13
223.10
231 .47
231 .55
7.04
10.15
geo
stddev
12.46
11.97
11.00
11.14
2.64
10.08
9.80
9.60
9.30
2.86
14.90
14.90
24.66
23.38
5.52
156.61
162.74
133.89
127.76
4.00
3.08
3.08
3.25
3.25
7.31
4.82
MG
1.93
1.96
1.80
1.75
1.00
1.64
1.72
1.66
1.51
1.00
2.08
2.26
3.06
2.76
1.00
8.89
13.16
11.93
8.88
1.00
0.73
0.73
0.70
0.70
1.00
0.69
VG
26.10
23.50
16.80
16.69
1.00
15.12
14.95
13.50
10.64
1.00
69.55
78.62
1380.89
828.81
1.00
>5000
>5000
>5000
>5000
1.00
1.21
1.21
1.21
1.21
1.00
1.44
RHC
38.56
37.20
36.54
38.74
88.59
85.79
82.16
80.51
83.86
157.50
132.33
130.23
125.11
128.93
57.22
78.38
34.29
34.32
62.26
1121.56
1014.09
1014.09
1116.13
1116.13
44.06
41.45
geo
R
0.85
0.85
0.84
0.85
1.00
0.85
0.85
0.85
0.86
1.00
0.83
0.83
0.80
0.81
1.00
0.84
0.83
0.81
0.83
1.00
0.99
0.99
0.99
0.99
1.00
0.99
ff2
0.64
0.70
0.74
0.68
1.00
0.76
0.74
0.76
0.78
1.00
0.64
0.66
0.64
0.64
1.00
0.62
0.60
0.62
0.64
1.00
0.94
0.94
0.98
0.98
1.00
0.82
TMS
0.58
0.61
0.64
0.62
1.00
0.65
0.64
0.65
0.67
0.98
0.56
0.59
0.55
0.56
0.94
0.46
0.39
0.50
0.43
0.98
0.88
0.90
0.88
0.90
0.99
0.83
-------
site
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B3
B3
B3
B3
year
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2010
2010
2010
2010
source
1
1
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
avg
time
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
Simulation
type
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
geo
mean
6.76
7.45
10.36
19.70
23.89
16.06
17.49
23.72
37.52
21.30
10.57
12.42
20.84
6.19
7.36
2.14
2.55
6.92
137.45
163.26
171.12
166.92
155.72
8.17
21.69
20.79
16.55
geo
stddev
8.69
7.90
5.06
5.07
4.68
8.55
7.53
4.95
3.43
9.14
26.59
23.07
10.70
20.38
12.31
63.36
50.11
13.79
3.84
3.13
3.17
3.20
3.15
6.14
1.91
1.94
2.56
MG
1.04
0.95
0.68
1.00
0.83
1.23
1.13
0.83
1.00
1.76
3.55
3.02
1.80
1.00
0.84
2.90
2.43
0.90
1.00
0.84
0.80
0.82
0.88
1.00
0.38
0.39
0.49
VG
1.50
1.40
1.38
1.00
1.12
1.95
1.58
1.11
1.00
9.45
4594.08
1390.74
14.76
1.00
1.71
132.40
44.94
1.38
1.00
1.09
1.11
1.10
1.09
1.00
12.00
10.67
3.79
RHC
39.78
40.28
43.94
93.54
85.14
80.57
82.91
89.91
165.01
98.08
99.89
117.82
117.82
68.23
44.60
39.13
39.88
44.50
1024.79
993.64
1001.01
1003.52
994.70
45.03
44.52
41.80
38.72
geo
R
0.96
0.96
0.99
1.00
0.99
0.95
0.96
0.99
1.00
0.82
0.67
0.69
0.82
1.00
0.98
0.90
0.91
0.99
1.00
0.99
0.99
0.99
0.99
1.00
0.93
0.93
0.98
ff2
0.82
0.86
0.82
1.00
1.00
0.84
0.86
1.00
1.00
0.72
0.64
0.72
0.74
1.00
0.82
0.72
0.76
0.88
1.00
1.00
1.00
1.00
1.00
1.00
0.68
0.68
0.72
TMS
0.87
0.89
0.83
0.99
0.92
0.81
0.86
0.93
0.97
0.56
0.52
0.52
0.64
0.98
0.78
0.57
0.62
0.88
1.00
0.94
0.94
0.94
0.96
1.00
0.61
0.61
0.68
-------
site
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
year
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2011
2011
2011
2011
2011
2011
source
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
avg
time
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
Simulation
type
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
geo
mean
17.40
21.49
46.22
43.74
35.75
37.81
26.63
61.88
57.06
45.67
50.50
5.37
29.24
24.83
19.05
21.86
186.57
203.41
203.53
202.35
202.18
8.88
14.15
8.02
8.77
12.70
21.57
geo
stddev
2.63
4.58
1.80
1.87
2.32
2.32
4.21
1.87
2.01
2.66
2.40
17.02
2.23
2.17
2.93
3.26
3.63
3.24
3.23
3.30
3.31
5.42
2.65
6.54
6.47
3.59
4.45
MG
0.47
1.00
0.47
0.49
0.60
0.57
1.00
0.43
0.47
0.58
0.53
1.00
0.18
0.22
0.28
0.25
1.00
0.92
0.92
0.92
0.92
1.00
0.63
1.11
1.01
0.70
1.00
VG
3.83
1.00
4.93
4.17
2.18
2.27
1.00
4.73
3.69
1.75
2.17
1.00
1673.38
1090.49
132.88
140.42
1.00
1.03
1.03
1.04
1.04
1.00
2.15
1.38
1.45
1.41
1.00
RHC
44.56
87.71
93.01
82.09
80.05
91.27
101.57
119.00
116.99
104.42
121.69
44.37
59.17
58.62
49.42
59.75
1107.59
1065.28
1065.28
1056.55
1056.55
43.94
44.17
40.45
42.84
44.75
90.23
geo
R
0.99
1.00
0.92
0.94
0.98
0.98
1.00
0.90
0.91
0.98
0.98
1.00
0.91
0.91
0.97
0.96
1.00
1.00
1.00
0.99
0.99
1.00
0.99
0.96
0.95
0.99
1.00
ff2
0.72
1.00
0.76
0.76
0.80
0.78
1.00
0.72
0.74
0.86
0.84
1.00
0.54
0.64
0.66
0.56
1.00
1.00
1.00
1.00
1.00
1.00
0.82
0.84
0.80
0.86
1.00
TMS
0.66
1.00
0.66
0.66
0.76
0.73
0.96
0.62
0.67
0.78
0.73
0.92
0.48
0.55
0.55
0.52
0.99
0.97
0.98
0.97
0.98
0.99
0.78
0.87
0.87
0.84
1.00
-------
site
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
year
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2012
2012
2012
2012
2012
2012
2012
2012
source
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
avg
time
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
Simulation
type
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
geo
mean
32.37
19.42
21.00
29.50
25.43
38.83
16.63
17.71
29.73
6.16
16.61
3.65
4.01
12.13
103.68
140.91
156.32
162.36
140.50
9.43
11.57
11.75
10.55
10.34
24.30
27.59
26.86
geo
stddev
2.58
5.87
5.91
3.45
4.87
3.15
15.24
13.87
5.31
16.38
3.47
40.75
32.83
5.99
3.62
3.33
3.39
3.27
3.26
4.45
3.79
3.83
4.54
4.78
3.49
3.32
3.57
MG
0.67
1.11
1.03
0.73
1.00
0.66
1.53
1.44
0.86
1.00
0.37
1.69
1.54
0.51
1.00
0.74
0.66
0.64
0.74
1.00
0.82
0.80
0.89
0.91
1.00
0.88
0.91
VG
1.63
1.40
1.43
1.26
1.00
1.53
8.88
7.23
1.26
1.00
31.18
11.39
8.91
4.66
1.00
1.13
1.22
1.25
1.13
1.00
1.21
1.21
1.09
1.11
1.00
1.13
1.14
RHC
91.82
82.60
81.45
93.64
117.38
111.38
102.49
103.99
111.08
47.21
49.34
47.01
45.60
46.79
754.22
1056.16
974.76
973.79
1050.48
41.38
35.74
36.33
38.64
36.54
88.12
75.43
77.31
geo
R
0.99
0.95
0.95
0.98
1.00
0.98
0.92
0.91
0.96
1.00
0.99
0.94
0.92
0.99
1.00
0.99
0.99
1.00
0.99
1.00
0.97
0.97
0.98
0.98
1.00
0.97
0.96
ff2
0.82
0.78
0.80
0.86
1.00
0.90
0.74
0.72
0.86
1.00
0.72
0.76
0.76
0.78
1.00
0.96
0.96
0.96
0.96
1.00
0.90
0.88
1.00
1.00
1.00
1.00
1.00
TMS
0.81
0.85
0.88
0.85
0.97
0.83
0.67
0.69
0.88
0.94
0.61
0.67
0.68
0.69
0.94
0.86
0.87
0.88
0.90
0.99
0.87
0.89
0.95
0.95
0.99
0.92
0.95
-------
site
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
year
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
source
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
2
avg
time
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
Simulation
type
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
geo
mean
23.98
25.29
39.52
35.77
28.13
25.07
29.78
8.53
10.30
10.45
7.89
7.70
138.38
177.72
182.22
184.40
181.10
15.04
18.15
17.41
14.17
15.30
35.08
41.75
41.02
33.85
35.18
geo
stddev
4.27
4.07
2.49
3.90
5.39
7.42
6.99
9.19
9.50
9.64
14.67
15.65
3.53
3.35
3.37
3.34
3.33
1.92
2.05
2.01
2.87
2.69
1.76
1.94
1.86
2.53
2.56
MG
1.01
0.96
1.00
1.11
1.41
1.58
1.33
1.00
0.83
0.82
1.08
1.11
1.00
0.78
0.76
0.75
0.76
1.00
0.83
0.86
1.06
0.98
1.00
0.84
0.86
1.04
1.00
VG
1.15
1.14
1.00
1.51
2.92
6.65
4.36
1.00
1.11
1.18
1.40
1.48
1.00
1.10
1.11
1.11
1.10
1.00
1.08
1.07
1.32
1.21
1.00
1.06
1.05
1.22
1.22
RHC
73.02
84.44
128.79
119.37
91.94
122.37
121.95
57.93
49.37
46.28
49.12
46.54
1019.43
1136.96
1136.26
1131.73
1132.36
38.51
43.11
40.95
42.48
44.95
77.67
84.73
80.66
85.22
90.97
geo
R
0.97
0.97
1.00
0.92
0.88
0.86
0.91
1.00
0.99
0.99
0.99
0.99
1.00
0.99
0.99
0.99
0.99
1.00
0.97
0.95
0.92
0.94
1.00
0.97
0.96
0.93
0.95
ff2
0.98
1.00
1.00
0.80
0.76
0.68
0.74
1.00
1.00
0.98
0.88
0.86
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.84
0.84
1.00
1.00
1.00
0.86
0.86
TMS
0.95
0.93
0.99
0.84
0.69
0.62
0.73
0.99
0.91
0.91
0.89
0.88
0.97
0.92
0.93
0.93
0.93
0.97
0.92
0.94
0.89
0.91
0.97
0.93
0.94
0.90
0.91
-------
site
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
year
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
source
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
2
3
avg
time
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
Simulation
type
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
geo
mean
48.58
52.39
55.40
47.20
44.43
21.55
22.54
23.31
16.53
18.00
138.06
169.12
181.49
179.21
172.22
12.26
13.45
13.01
11.01
11.18
28.32
30.84
29.34
24.45
26.50
40.94
geo
stddev
1.77
1.97
1.85
2.50
2.84
1.73
2.26
2.28
3.86
3.70
3.34
3.29
3.21
3.27
3.31
3.08
3.32
3.49
4.49
4.55
2.77
3.12
3.24
4.50
4.23
2.24
MG
1.00
0.93
0.88
1.03
1.09
1.00
0.96
0.92
1.30
1.20
1.00
0.82
0.76
0.77
0.80
1.00
0.91
0.94
1.11
1.10
1.00
0.92
0.97
1.16
1.07
1.00
VG
1.00
1.04
1.04
1.23
1.42
1.00
1.10
1.12
2.24
1.98
1.00
1.06
1.13
1.11
1.07
1.00
1.08
1.10
1.39
1.32
1.00
1.11
1.14
1.58
1.41
1.00
RHC
92.74
108.08
114.99
130.50
131.02
40.76
53.81
55.79
52.12
50.65
1034.29
1176.23
1174.92
1174.31
1175.49
42.16
41.81
41.08
41.49
44.81
88.40
87.04
79.97
82.04
91.76
126.30
geo
R
1.00
0.98
0.98
0.92
0.90
1.00
0.98
0.96
0.94
0.95
1.00
0.99
0.99
0.99
0.99
1.00
0.98
0.97
0.95
0.97
1.00
0.97
0.96
0.94
0.95
1.00
ff2
1.00
1.00
1.00
0.88
0.84
1.00
0.96
0.96
0.76
0.76
1.00
1.00
0.96
1.00
1.00
1.00
1.00
0.98
0.86
0.86
1.00
1.00
0.98
0.82
0.78
1.00
TMS
0.94
0.94
0.95
0.89
0.87
0.93
0.91
0.94
0.77
0.80
0.97
0.93
0.92
0.93
0.95
0.99
0.96
0.96
0.88
0.88
1.00
0.95
0.94
0.85
0.85
0.99
C1 2012
3hr AERC.RCALT 36.70 3.86 1.12
2.01
112.73 0.82 0.86 0.79
-------
site
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
year
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
source
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
2
3
3
3
3
avg
time
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
Simulation
type
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
geo
mean
35.69
25.69
26.53
12.35
13.04
12.96
8.42
8.50
114.85
156.59
164.48
164.71
153.98
12.41
13.85
9.97
9.90
13.39
28.03
32.18
23.32
23.28
31.40
27.48
40.09
27.82
28.17
geo
stddev
4.07
8.20
7.84
4.93
6.45
7.25
13.72
12.14
3.65
3.16
3.17
3.13
3.15
3.30
3.58
5.51
5.62
3.97
3.21
3.07
4.76
4.91
3.42
3.82
3.46
6.35
6.73
MG
1.15
1.59
1.54
1.00
0.95
0.95
1.47
1.45
1.00
0.73
0.70
0.70
0.75
1.00
0.90
1.24
1.25
0.93
1.00
0.87
1.20
1.20
0.89
1.00
0.69
0.99
0.98
VG
2.19
15.75
13.28
1.00
1.19
1.33
4.53
3.57
1.00
1.13
1.18
1.19
1.12
1.00
1.04
1.50
1.54
1.06
1.00
1.04
1.30
1.32
1.03
1.00
1.19
1.36
1.47
RHC
106.28
115.84
113.88
49.59
46.56
49.67
48.71
46.77
981.53
997.98
997.52
994.08
995.07
44.06
45.11
40.67
43.96
46.95
85.21
86.43
82.44
81.07
91.99
99.26
122.11
102.13
103.84
geo
R
0.82
0.75
0.75
1.00
0.98
0.98
0.96
0.96
1.00
1.00
0.99
0.99
1.00
1.00
0.99
0.98
0.98
0.99
1.00
0.99
0.98
0.98
1.00
1.00
0.99
0.99
0.99
ff2
0.84
0.70
0.70
1.00
0.94
0.80
0.64
0.62
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.84
0.80
1.00
1.00
1.00
0.80
0.78
1.00
1.00
0.88
0.90
0.88
TMS
0.79
0.61
0.63
1.00
0.93
0.88
0.70
0.70
0.96
0.92
0.91
0.91
0.93
0.99
0.97
0.84
0.83
0.96
0.99
0.96
0.87
0.87
0.95
0.95
0.84
0.89
0.90
-------
site
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
year
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
source
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
2
3
3
3
3
3
4
avg
time
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
Simulation
type
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
geo
mean
34.81
12.50
14.00
5.48
5.58
11.35
138.22
157.00
166.33
164.81
156.12
11.08
9.24
6.83
7.11
9.08
25.43
22.14
16.88
17.47
22.32
36.89
26.29
17.83
19.17
25.93
11.08
geo
stddev
5.30
4.74
6.26
24.13
19.60
10.21
3.34
3.19
3.24
3.24
3.19
3.67
4.94
6.34
6.23
5.16
3.43
4.63
5.84
5.76
4.76
2.60
4.92
11.19
9.54
5.60
5.30
MG
0.79
1.00
0.89
2.28
2.24
1.10
1.00
0.88
0.83
0.84
0.89
1.00
1.20
1.62
1.56
1.22
1.00
1.15
1.51
1.46
1.14
1.00
1.40
2.07
1.92
1.42
1.00
VG
1.25
1.00
1.14
35.98
20.45
1.95
1.00
1.02
1.05
1.04
1.02
1.00
1.19
2.37
2.27
1.28
1.00
1.14
1.76
1.70
1.18
1.00
1.75
17.23
9.93
2.16
1.00
RHC
132.81
42.73
63.15
39.84
40.68
62.48
1017.06
1107.19
1107.50
1101.95
1101.91
40.96
39.80
33.88
34.33
42.08
82.58
80.30
68.47
72.75
82.11
91.74
97.45
91.22
86.55
95.10
42.45
geo
R
0.99
1.00
0.99
0.98
0.96
0.99
1.00
1.00
1.00
1.00
1.00
1.00
0.99
0.93
0.93
0.98
1.00
0.99
0.97
0.97
0.99
1.00
0.99
0.96
0.96
0.98
1.00
ff2
0.84
1.00
0.96
0.78
0.74
0.86
1.00
1.00
1.00
1.00
1.00
1.00
0.84
0.78
0.78
0.82
1.00
0.86
0.78
0.74
0.80
1.00
0.84
0.80
0.80
0.84
1.00
TMS
0.84
0.91
0.88
0.57
0.64
0.78
0.97
0.95
0.96
0.96
0.97
0.99
0.89
0.72
0.76
0.84
0.98
0.91
0.77
0.79
0.88
0.94
0.81
0.65
0.67
0.78
0.96
-------
site
C2
C2
C2
C2
C2
C2
C2
C2
C2
year
2012
2012
2012
2012
2012
2012
2012
2012
2012
source
4
4
4
4
5
5
5
5
5
avg
time
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
3hr
Simulation
type
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
geo
mean
6.19
2.88
3.60
5.50
145.35
172.77
175.77
159.54
153.97
geo
stddev
16.28
48.71
30.66
15.80
3.35
3.33
3.32
3.16
3.17
MG
1.79
3.85
3.08
2.02
1.00
0.84
0.83
0.91
0.94
VG
5.18
1698.01
170.11
6.18
1.00
1.07
1.08
1.06
1.03
RHC
41.21
37.98
38.65
39.77
1029.30
932.64
936.43
936.25
932.35
geo
R
1.00
0.95
0.93
0.99
1.00
0.99
0.99
0.99
0.99
ff2
0.78
0.66
0.66
0.70
1.00
1.00
1.00
1.00
1.00
TMS
0.70
0.56
0.58
0.66
0.99
0.93
0.95
0.97
0.98
-------
AERMOD results statistical scores: 8-hour average concentrations.
site
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
year
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2011
source
1
1
1
1
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
avg
time
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
Simulation
type
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
geo
6.93
4.46
4.31
4.74
4.72
16.08
12.29
11.00
11.69
13.30
19.82
12.01
11.26
8.45
9.40
6.62
0.92
0.68
0.77
0.94
93.13
143.87
143.82
152.79
152.49
4.72
geo
4.05
13.55
13.41
12.52
13.47
3.50
10.80
10.97
10.92
10.44
3.39
16.83
16.99
29.46
27.67
6.13
176.71
206.08
174.23
146.86
4.72
3.38
3.39
3.51
3.52
8.54
MG
1.00
1.55
1.61
1.46
1.47
1.00
1.31
1.46
1.38
1.21
1.00
1.65
1.76
2.35
2.11
1.00
7.23
9.78
8.64
7.04
1.00
0.65
0.65
0.61
0.61
1.00
VG
1.00
13.27
13.80
9.99
12.31
1.00
8.12
9.82
9.08
6.90
1.00
57.41
69.60
1231.88
700.16
1.00
>5000
>5000
>5000
>5000
1.00
1.35
1.35
1.40
1.39
1.00
RHC
36.29
36.00
34.85
33.56
34.65
72.89
81.99
78.44
78.43
80.29
101.58
94.09
93.32
94.52
95.27
38.22
34.49
31.71
30.26
33.28
857.68
845.55
845.55
942.90
942.90
40.38
geo
1.00
0.87
0.87
0.88
0.87
1.00
0.87
0.86
0.87
0.88
1.00
0.82
0.81
0.79
0.80
1.00
0.85
0.84
0.83
0.84
1.00
1.00
1.00
1.00
1.00
1.00
ff2
1.00
0.44
0.44
0.44
0.42
1.00
0.48
0.48
0.48
0.50
1.00
0.24
0.20
0.22
0.28
1.00
0.44
0.42
0.46
0.48
1.00
0.72
0.72
0.72
0.72
1.00
TMS
0.99
0.60
0.59
0.61
0.60
0.97
0.63
0.62
0.64
0.67
0.96
0.52
0.52
0.49
0.51
0.92
0.47
0.46
0.47
0.47
0.95
0.82
0.82
0.79
0.81
0.98
-------
site
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B3
B3
B3
year
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2010
2010
2010
source
1
1
1
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
avg
time
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
Simulation
type
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
geo
6.81
4.56
5.10
6.94
12.91
16.51
11.31
12.58
16.79
21.01
13.86
6.91
8.26
13.76
3.44
4.47
1.34
1.63
4.23
84.38
115.09
122.31
121.42
114.04
5.40
15.47
14.45
geo
6.61
10.77
9.38
6.87
6.04
6.30
10.23
8.93
6.61
3.82
11.77
34.23
29.35
13.75
24.33
15.96
86.66
68.20
18.11
4.13
3.11
3.20
3.11
3.04
7.40
2.28
2.28
MG
0.69
1.04
0.93
0.68
1.00
0.78
1.14
1.03
0.77
1.00
1.52
3.04
2.55
1.53
1.00
0.77
2.57
2.11
0.81
1.00
0.73
0.69
0.70
0.74
1.00
0.35
0.37
VG
1.32
1.58
1.47
1.29
1.00
1.19
2.09
1.73
1.21
1.00
9.97
4318.91
1302.06
15.79
1.00
1.74
149.37
49.91
1.46
1.00
1.22
1.28
1.28
1.23
1.00
13.32
11.30
RHC
40.00
35.48
36.67
40.31
85.39
80.58
70.70
71.41
81.63
94.91
86.94
69.14
83.91
81.59
42.23
39.28
34.69
35.42
35.95
740.68
748.85
753.36
753.85
748.77
39.14
42.26
40.09
geo
0.99
0.96
0.96
0.99
1.00
0.98
0.95
0.95
0.98
1.00
0.87
0.75
0.76
0.87
1.00
0.98
0.91
0.92
0.99
1.00
0.99
0.99
0.99
0.99
1.00
0.97
0.98
ff2
0.88
0.84
0.82
0.88
1.00
0.88
0.70
0.72
0.86
1.00
0.40
0.40
0.30
0.32
1.00
0.80
0.72
0.76
0.80
1.00
0.88
0.82
0.86
0.86
1.00
0.60
0.62
TMS
0.86
0.86
0.87
0.85
0.99
0.89
0.78
0.84
0.86
0.96
0.59
0.46
0.46
0.58
0.99
0.81
0.58
0.63
0.85
0.95
0.88
0.85
0.86
0.88
0.98
0.58
0.60
-------
site
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
year
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2011
2011
2011
2011
2011
source
1
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
avg
time
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
Simulation
type
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
geo
11.66
12.48
13.78
34.15
31.07
25.86
28.20
15.37
40.24
37.58
29.63
32.34
3.05
18.17
16.67
12.65
13.67
101.97
138.18
140.34
141.03
139.30
6.13
9.36
5.91
6.17
8.31
geo
3.08
3.15
5.58
2.16
2.21
2.82
2.83
4.81
2.12
2.22
3.12
2.85
19.95
2.49
2.51
3.57
3.79
4.40
3.25
3.26
3.29
3.29
6.41
3.29
6.97
6.57
4.34
MG
0.46
0.43
1.00
0.40
0.44
0.53
0.49
1.00
0.38
0.41
0.52
0.48
1.00
0.17
0.18
0.24
0.22
1.00
0.74
0.73
0.72
0.73
1.00
0.66
1.04
0.99
0.74
VG
4.13
4.36
1.00
6.07
4.95
2.64
2.84
1.00
5.70
4.46
1.99
2.45
1.00
2503.48
1703.87
168.78
177.99
1.00
1.21
1.22
1.22
1.21
1.00
2.04
1.46
1.61
1.39
RHC
37.31
37.56
81.51
86.99
81.15
74.30
78.01
80.26
100.68
94.62
91.43
104.93
35.67
47.78
47.39
46.06
46.62
950.13
849.58
849.58
850.73
850.73
33.05
37.48
36.19
32.88
34.55
geo
0.99
0.99
1.00
0.97
0.97
0.97
0.98
1.00
0.94
0.96
0.98
0.98
1.00
0.95
0.96
0.98
0.98
1.00
1.00
1.00
1.00
1.00
1.00
0.98
0.95
0.93
0.98
ff2
0.62
0.62
1.00
0.56
0.56
0.60
0.58
1.00
0.46
0.48
0.58
0.56
1.00
0.42
0.52
0.56
0.46
1.00
0.80
0.80
0.80
0.80
1.00
0.84
0.82
0.76
0.88
TMS
0.65
0.65
0.98
0.61
0.62
0.68
0.67
0.93
0.55
0.60
0.71
0.66
0.92
0.46
0.53
0.55
0.53
0.98
0.85
0.87
0.87
0.87
0.95
0.77
0.88
0.84
0.85
-------
site
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
year
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2012
2012
2012
2012
2012
2012
2012
source
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
avg
time
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
Simulation
type
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
geo
14.89
20.12
13.74
14.72
19.08
15.29
23.79
11.04
11.85
19.13
3.59
11.23
2.72
2.72
8.00
72.91
107.63
124.78
131.01
113.73
6.40
8.17
7.98
7.74
7.71
15.69
20.59
geo
5.05
3.15
6.16
6.01
4.20
5.39
3.73
16.59
15.23
6.56
18.95
4.43
48.91
42.63
7.54
3.98
4.08
3.88
3.61
3.79
5.46
4.74
4.78
5.22
5.62
4.22
3.69
MG
1.00
0.74
1.08
1.01
0.78
1.00
0.64
1.39
1.29
0.80
1.00
0.32
1.32
1.32
0.45
1.00
0.68
0.58
0.56
0.64
1.00
0.78
0.80
0.83
0.83
1.00
0.76
VG
1.00
1.45
1.35
1.45
1.21
1.00
1.56
8.00
6.62
1.36
1.00
35.37
9.20
9.64
5.36
1.00
1.20
1.40
1.45
1.23
1.00
1.14
1.11
1.09
1.08
1.00
1.15
RHC
67.31
82.58
73.08
66.62
74.93
68.32
101.66
84.12
78.65
96.89
30.46
44.12
42.08
39.85
42.48
672.27
993.71
940.98
927.39
989.71
34.28
35.13
32.82
33.71
35.23
65.46
64.50
geo
1.00
0.99
0.96
0.94
0.98
1.00
0.98
0.93
0.92
0.97
1.00
0.98
0.95
0.93
0.98
1.00
0.99
0.99
0.99
1.00
1.00
0.99
0.99
0.99
0.99
1.00
0.99
ff2
1.00
0.88
0.84
0.72
0.94
1.00
0.80
0.62
0.64
0.86
1.00
0.72
0.62
0.46
0.76
1.00
0.94
0.78
0.68
1.00
1.00
0.94
0.96
0.98
1.00
1.00
0.96
TMS
0.94
0.82
0.87
0.85
0.88
0.92
0.75
0.64
0.68
0.83
0.93
0.55
0.68
0.64
0.66
0.93
0.82
0.80
0.78
0.88
0.99
0.91
0.92
0.94
0.94
0.96
0.91
-------
site
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
C1
C1
C1
C1
C1
C1
C1
year
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2011
2011
2011
2011
2011
2011
2011
source
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
avg
time
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
Simulation
type
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
geo
17.64
17.04
19.27
21.50
23.92
17.90
15.98
18.95
4.82
6.64
6.90
5.51
5.07
79.31
131.58
135.11
137.46
133.43
9.65
12.11
11.21
9.07
10.57
22.02
27.35
geo
4.60
5.17
4.30
3.05
4.79
6.89
8.86
7.85
10.96
12.11
12.56
18.57
19.64
3.82
3.79
3.81
3.75
3.73
2.58
2.59
2.65
3.83
3.36
2.37
2.53
MG
0.89
0.92
0.81
1.00
0.90
1.20
1.35
1.14
1.00
0.73
0.70
0.88
0.95
1.00
0.60
0.59
0.58
0.59
1.00
0.80
0.86
1.06
0.91
1.00
0.81
VG
1.09
1.14
1.11
1.00
1.53
3.31
6.01
3.67
1.00
1.16
1.28
1.48
1.49
1.00
1.33
1.38
1.41
1.35
1.00
1.08
1.07
1.35
1.17
1.00
1.08
RHC
62.41
65.01
65.95
82.37
109.15
80.49
77.64
107.21
34.24
44.90
38.09
36.62
44.83
696.49
1067.6
5
1045.5
0
1041.2
2
1063.6
8
34.09
40.45
39.52
39.82
43.92
69.20
84.78
geo
0.99
0.98
0.99
1.00
0.94
0.88
0.88
0.92
1.00
1.00
0.99
0.99
1.00
1.00
0.99
0.99
0.99
0.99
1.00
0.99
0.97
0.95
0.96
1.00
0.99
ff2
0.98
0.94
0.98
1.00
0.80
0.68
0.72
0.74
1.00
1.00
0.76
0.84
0.88
1.00
0.86
0.78
0.76
0.84
1.00
1.00
1.00
0.84
0.96
1.00
1.00
TMS
0.95
0.94
0.93
0.94
0.81
0.69
0.70
0.71
0.95
0.87
0.82
0.87
0.86
0.92
0.77
0.81
0.81
0.83
0.95
0.91
0.95
0.89
0.92
0.95
0.91
-------
site
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
year
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2012
2012
2012
2012
2012
2012
2012
source
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
avg
time
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
Simulation
type
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
geo
25.25
21.41
24.65
26.04
31.08
32.52
27.28
25.62
11.54
13.44
14.24
9.90
10.66
89.18
126.21
134.91
134.11
126.52
8.18
8.92
8.70
7.55
7.34
19.46
20.35
geo
2.45
3.38
3.20
2.09
2.36
2.18
3.06
3.48
2.13
2.67
2.86
4.96
4.19
3.96
3.98
3.84
3.84
3.91
3.88
4.17
4.19
5.09
5.75
3.41
3.86
MG
0.87
1.03
0.89
1.00
0.84
0.80
0.96
1.02
1.00
0.86
0.81
1.17
1.08
1.00
0.71
0.66
0.67
0.71
1.00
0.92
0.94
1.08
1.11
1.00
0.96
VG
1.07
1.27
1.19
1.00
1.08
1.09
1.31
1.49
1.00
1.13
1.24
2.49
1.82
1.00
1.14
1.22
1.20
1.14
1.00
1.10
1.11
1.32
1.36
1.00
1.13
RHC
76.36
76.57
87.25
77.67
81.72
81.31
86.62
89.61
34.73
40.66
41.52
42.28
40.73
898.20
1141.8
7
1141.1
1
1137.4
1
1137.8
7
39.90
38.46
34.34
36.02
38.66
78.12
76.17
geo
0.97
0.95
0.97
1.00
0.97
0.97
0.93
0.93
1.00
0.97
0.95
0.93
0.94
1.00
1.00
0.99
1.00
1.00
1.00
0.98
0.98
0.96
0.97
1.00
0.97
ff2
1.00
0.84
0.94
1.00
1.00
1.00
0.88
0.82
1.00
0.96
0.86
0.66
0.84
1.00
1.00
0.94
1.00
1.00
1.00
0.98
0.98
0.86
0.86
1.00
0.98
TMS
0.94
0.91
0.90
0.92
0.94
0.94
0.89
0.87
0.94
0.91
0.88
0.77
0.84
0.95
0.87
0.88
0.90
0.92
0.98
0.95
0.94
0.89
0.88
0.97
0.95
-------
site
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C2
C2
C2
C2
C2
C2
C2
C2
C2
year
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2010
2010
2010
2010
2010
2010
2010
2010
2010
source
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
avg
time
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
Simulation
type
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
geo
19.80
16.96
17.36
27.35
23.86
22.68
15.46
16.82
7.66
7.99
8.32
5.46
5.17
86.00
127.45
132.61
134.19
127.35
8.62
9.69
7.07
7.53
9.89
19.91
23.62
17.11
18.45
geo
4.00
5.16
5.21
2.52
4.81
4.97
9.62
9.64
6.12
8.02
9.15
17.54
15.29
3.82
3.44
3.47
3.51
3.46
4.09
4.19
6.09
5.97
4.57
3.86
3.68
5.34
5.27
MG
0.98
1.15
1.12
1.00
1.15
1.21
1.77
1.63
1.00
0.96
0.92
1.40
1.48
1.00
0.68
0.65
0.64
0.68
1.00
0.89
1.22
1.14
0.87
1.00
0.84
1.16
1.08
VG
1.15
1.48
1.43
1.00
2.00
2.23
14.71
12.60
1.00
1.20
1.42
4.67
3.49
1.00
1.20
1.24
1.26
1.19
1.00
1.05
1.41
1.33
1.07
1.00
1.07
1.29
1.23
RHC
71.84
72.44
76.51
101.54
94.91
78.58
80.55
98.60
45.98
40.93
37.66
40.13
41.27
755.03
925.14
925.65
924.40
924.00
42.44
44.11
40.43
42.44
45.00
83.47
89.80
80.53
81.07
geo
0.97
0.95
0.96
1.00
0.91
0.90
0.86
0.88
1.00
0.99
0.98
0.97
0.97
1.00
1.00
1.00
0.99
1.00
1.00
0.99
0.97
0.98
0.99
1.00
0.99
0.97
0.98
ff2
0.98
0.82
0.82
1.00
0.88
0.88
0.74
0.76
1.00
0.94
0.70
0.60
0.70
1.00
1.00
1.00
0.90
1.00
1.00
1.00
0.82
0.86
1.00
1.00
1.00
0.80
0.86
TMS
0.95
0.86
0.86
0.98
0.82
0.78
0.64
0.63
1.00
0.92
0.84
0.69
0.72
0.95
0.86
0.89
0.87
0.90
0.98
0.96
0.85
0.88
0.95
0.98
0.94
0.86
0.91
-------
site
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
year
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
source
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
2
3
avg
time
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
Simulation
type
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
geo
24.18
17.91
26.35
17.99
18.22
23.51
7.19
7.74
3.61
4.03
6.67
106.45
130.00
135.45
135.79
130.94
6.66
6.70
4.67
4.72
6.43
14.95
15.46
11.31
11.72
15.42
19.19
geo
4.01
4.44
4.25
7.44
7.49
6.36
5.21
7.26
29.40
22.16
10.85
3.69
3.24
3.28
3.25
3.21
4.55
5.25
6.87
7.17
5.94
4.19
5.22
6.61
6.96
5.65
3.16
MG
0.82
1.00
0.68
1.00
0.98
0.76
1.00
0.93
1.99
1.79
1.08
1.00
0.82
0.79
0.78
0.81
1.00
0.99
1.43
1.41
1.04
1.00
0.97
1.32
1.28
0.97
1.00
VG
1.07
1.00
1.24
1.50
1.49
1.42
1.00
1.21
44.22
17.33
1.94
1.00
1.07
1.08
1.09
1.07
1.00
1.08
2.04
1.96
1.16
1.00
1.09
1.64
1.67
1.16
1.00
RHC
89.63
84.07
90.04
71.82
74.86
91.54
33.99
36.62
37.14
36.91
36.18
886.98
971 .59
973.56
969.45
967.85
32.33
36.08
32.22
33.87
38.66
68.10
75.16
63.00
68.94
78.33
70.40
geo
0.99
1.00
0.99
0.98
0.98
0.97
1.00
0.99
0.97
0.96
0.99
1.00
1.00
1.00
1.00
1.00
1.00
0.99
0.93
0.94
0.99
1.00
0.99
0.96
0.97
0.99
1.00
ff2
1.00
1.00
0.90
0.90
0.90
0.80
1.00
0.88
0.78
0.78
0.86
1.00
1.00
1.00
1.00
1.00
1.00
0.96
0.82
0.82
0.92
1.00
0.98
0.80
0.80
0.88
1.00
TMS
0.93
0.93
0.86
0.87
0.90
0.81
0.91
0.91
0.65
0.67
0.85
0.96
0.93
0.94
0.94
0.95
0.95
0.95
0.77
0.79
0.92
0.97
0.95
0.79
0.81
0.92
0.95
-------
site
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
year
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
source
3
3
3
3
4
4
4
4
4
5
5
5
5
5
avg
time
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
8hr
Simulation
type
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
geo
18.60
11.22
12.79
18.82
6.17
4.45
1.84
2.21
3.89
91.87
112.38
118.46
121.05
113.30
geo
6.19
13.11
11.86
7.08
6.54
20.04
57.28
36.94
19.23
4.10
3.64
3.70
3.58
3.60
MG
1.03
1.71
1.50
1.02
1.00
1.38
3.36
2.79
1.59
1.00
0.82
0.78
0.76
0.81
VG
1.75
15.02
10.69
2.38
1.00
4.50
1147.62
144.72
4.77
1.00
1.10
1.16
1.18
1.11
RHC
93.82
67.77
70.86
88.26
34.45
35.63
32.64
32.49
32.96
892.22
862.70
869.39
871 .84
865.40
geo
0.97
0.93
0.92
0.95
1.00
0.99
0.94
0.93
0.98
1.00
0.99
0.98
0.98
0.99
ff2
0.84
0.70
0.54
0.64
1.00
0.70
0.56
0.64
0.74
1.00
1.00
0.88
0.86
1.00
TMS
0.82
0.60
0.64
0.76
0.97
0.72
0.54
0.59
0.71
0.99
0.94
0.90
0.89
0.94
-------
AERMOD results statistical scores: 24-hour average concentrations.
site
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
year
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2011
source
1
1
1
1
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
avg
time
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
Simulation
type
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
geo
mean
3.51
2.24
2.20
2.55
2.46
8.16
5.99
5.62
6.28
6.51
9.57
5.40
4.78
3.79
4.48
3.13
0.44
0.00
0.41
0.47
43.76
71.66
73.15
76.07
73.98
2.48
geo
stddev
5.20
15.99
13.82
11.68
14.97
4.54
13.07
11.90
10.69
12.56
3.91
20.75
21.14
35.49
30.02
7.35
225.54
NA
201.14
182.97
5.16
3.57
3.65
3.72
3.66
10.65
MG
1.00
1.57
1.60
1.38
1.43
1.00
1.36
1.45
1.30
1.26
1.00
1.77
2.00
2.52
2.14
1.00
7.18
NA
7.55
6.62
1.00
0.61
0.60
0.58
0.59
1.00
VG
1.00
11.88
8.94
5.06
8.79
1.00
8.87
8.10
5.60
7.29
1.00
64.26
94.35
1384.36
436.12
1.00
>5000
NA
>5000
>5000
1.00
1.47
1.49
1.54
1.50
1.00
RHC
28.99
23.59
23.10
23.63
23.43
61.21
54.49
52.69
53.88
54.57
54.31
60.01
48.81
49.31
54.83
23.51
24.00
21.74
21.22
22.66
483.47
487.65
487.65
539.44
539.44
31.61
geoR
1.00
0.89
0.88
0.89
0.90
1.00
0.88
0.86
0.87
0.88
1.00
0.87
0.85
0.85
0.86
1.00
0.88
0.93
0.86
0.87
1.00
1.00
1.00
1.00
1.00
1.00
ff2
1.00
0.52
0.38
0.50
0.52
1.00
0.50
0.34
0.40
0.52
1.00
0.32
0.28
0.30
0.34
1.00
0.44
0.40
0.42
0.48
1.00
0.70
0.64
0.62
0.70
1.00
TMS
0.97
0.59
0.60
0.66
0.64
0.95
0.62
0.60
0.64
0.66
0.91
0.54
0.49
0.51
0.51
0.94
0.49
NA
0.48
0.49
0.90
0.80
0.78
0.75
0.79
0.96
-------
site
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B3
B3
B3
year
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2010
2010
2010
source
1
1
1
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
avg
time
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
Simulation
type
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
geo
mean
4.06
2.73
3.14
4.10
6.55
9.89
7.01
7.88
9.96
10.52
7.71
3.80
4.30
7.32
1.71
2.43
0.70
0.84
2.28
50.85
86.87
88.62
88.17
86.93
3.14
8.84
8.35
geo
stddev
8.75
13.59
11.23
8.86
7.57
8.46
12.64
10.64
8.68
4.81
15.45
45.26
37.12
17.50
30.78
21.14
115.37
89.98
23.64
5.01
3.28
3.30
3.28
3.27
10.05
2.89
2.91
MG
0.61
0.91
0.79
0.60
1.00
0.66
0.94
0.83
0.66
1.00
1.37
2.77
2.44
1.44
1.00
0.70
2.44
2.03
0.75
1.00
0.59
0.57
0.58
0.59
1.00
0.36
0.38
VG
1.61
1.99
1.77
1.54
1.00
1.60
2.71
2.15
1.58
1.00
11.16
>5000
1371.90
16.51
1.00
1.82
155.14
51.01
1.46
1.00
1.67
1.70
1.70
1.68
1.00
15.52
13.54
RHC
28.32
25.47
26.08
30.54
62.73
56.10
49.85
53.49
61.32
74.38
60.84
55.60
58.43
64.14
31.66
28.20
27.70
29.89
28.43
489.44
603.29
599.08
605.92
602.98
31.61
34.42
31.85
geoR
0.98
0.95
0.96
0.99
1.00
0.97
0.93
0.94
0.97
1.00
0.89
0.79
0.80
0.89
1.00
0.98
0.92
0.93
0.99
1.00
0.99
0.99
0.99
0.99
1.00
0.98
0.98
ff2
0.54
0.48
0.50
0.56
1.00
0.58
0.44
0.46
0.62
1.00
0.34
0.30
0.36
0.38
1.00
0.78
0.70
0.70
0.76
1.00
0.54
0.54
0.54
0.54
1.00
0.50
0.52
TMS
0.73
0.75
0.76
0.73
0.95
0.75
0.71
0.73
0.75
0.98
0.57
0.47
0.50
0.59
0.98
0.78
0.60
0.61
0.83
0.93
0.70
0.74
0.74
0.74
0.97
0.56
0.57
-------
site
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
year
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2011
2011
2011
2011
2011
source
1
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
avg
time
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
Simulation
type
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
geo
mean
6.51
6.92
8.19
19.30
17.92
14.45
15.62
8.25
21.67
20.82
15.56
16.80
1.64
9.20
9.30
6.80
6.61
54.21
76.70
80.74
80.72
77.36
3.37
5.27
3.90
4.08
4.70
geo
stddev
3.99
4.04
7.37
2.74
2.80
3.67
3.65
6.19
2.73
2.83
3.98
3.67
26.04
3.13
3.27
4.73
4.77
4.66
3.41
3.40
3.42
3.40
8.49
4.56
7.85
7.40
6.00
MG
0.48
0.45
1.00
0.42
0.46
0.57
0.52
1.00
0.38
0.40
0.53
0.49
1.00
0.18
0.18
0.24
0.25
1.00
0.71
0.67
0.67
0.70
1.00
0.64
0.87
0.83
0.72
VG
4.47
4.70
1.00
6.28
5.43
2.79
2.87
1.00
5.57
4.71
2.01
2.34
1.00
2295.20
1799.97
156.36
144.38
1.00
1.30
1.41
1.39
1.31
1.00
2.01
1.50
1.71
1.39
RHC
29.41
30.86
64.14
70.53
65.90
59.01
65.20
59.50
80.46
76.45
70.31
78.02
28.29
38.60
38.42
35.42
36.60
571 .44
432.19
432.15
436.30
436.28
26.99
29.43
26.78
28.61
30.27
geoR
0.98
0.99
1.00
0.97
0.97
0.96
0.97
1.00
0.97
0.98
0.98
0.99
1.00
0.96
0.98
0.99
0.99
1.00
0.99
0.98
0.98
0.99
1.00
0.98
0.96
0.94
0.99
ff2
0.52
0.52
1.00
0.50
0.50
0.48
0.50
1.00
0.46
0.38
0.50
0.52
1.00
0.52
0.38
0.42
0.58
1.00
0.78
0.64
0.64
0.78
1.00
0.80
0.76
0.66
0.82
TMS
0.63
0.63
0.96
0.59
0.61
0.65
0.65
0.91
0.55
0.58
0.69
0.66
0.92
0.48
0.51
0.52
0.56
0.93
0.80
0.80
0.80
0.85
0.96
0.77
0.83
0.79
0.84
-------
site
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
year
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2012
2012
2012
2012
2012
2012
2012
source
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
avg
time
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
Simulation
type
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
geo
mean
8.34
11.26
8.53
9.19
10.66
7.70
12.62
6.35
6.73
10.00
1.80
5.57
1.58
1.68
4.24
48.48
63.47
73.35
75.53
67.58
3.70
3.50
3.59
3.58
3.31
9.39
8.76
geo
stddev
6.78
4.28
7.09
6.94
5.59
6.66
4.91
21.22
18.58
8.53
23.35
5.91
65.53
50.92
10.23
4.72
4.23
3.86
3.54
3.91
6.82
5.89
5.59
6.10
6.98
5.17
4.79
MG
1.00
0.74
0.98
0.91
0.78
1.00
0.61
1.21
1.14
0.77
1.00
0.32
1.14
1.07
0.43
1.00
0.76
0.66
0.64
0.72
1.00
1.06
1.03
1.03
1.12
1.00
1.07
VG
1.00
1.46
1.31
1.48
1.20
1.00
1.52
7.12
5.49
1.28
1.00
29.37
8.90
7.64
5.01
1.00
1.14
1.32
1.36
1.17
1.00
1.13
1.10
1.05
1.06
1.00
1.12
RHC
53.60
62.21
55.55
59.10
59.10
49.54
72.79
65.72
63.93
72.73
22.95
35.83
31.32
30.38
33.53
506.73
479.86
454.63
454.58
477.57
17.59
16.06
16.89
17.34
16.98
40.19
40.08
geoR
1.00
0.99
0.97
0.95
0.99
1.00
0.99
0.95
0.94
0.99
1.00
0.98
0.96
0.94
0.99
1.00
0.99
0.98
0.99
1.00
1.00
0.99
0.99
1.00
0.99
1.00
0.98
ff2
1.00
0.88
0.86
0.64
0.92
1.00
0.76
0.56
0.48
0.92
1.00
0.68
0.56
0.52
0.62
1.00
1.00
0.82
0.68
1.00
1.00
0.92
0.94
0.98
0.98
1.00
0.92
TMS
0.94
0.83
0.89
0.82
0.90
0.90
0.74
0.68
0.69
0.87
0.91
0.53
0.68
0.70
0.63
0.90
0.92
0.83
0.81
0.90
0.90
0.93
0.95
0.97
0.96
0.92
0.94
-------
site
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
C1
C1
C1
C1
C1
C1
C1
C1
C1
year
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2011
2011
2011
2011
2011
2011
2011
2011
2011
source
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
avg
time
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
Simulation
type
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
geo
mean
7.76
7.80
8.25
12.53
11.26
8.02
7.09
9.64
2.72
2.84
2.92
2.39
2.18
60.17
72.83
78.20
78.36
72.85
4.36
5.87
6.32
4.95
4.95
9.93
12.81
13.81
11.17
geo
stddev
5.45
6.05
5.56
3.67
6.03
8.03
10.17
10.04
14.17
14.44
14.23
21.19
23.25
4.09
4.58
4.66
4.60
4.55
3.22
3.25
3.40
4.87
4.27
3.02
3.07
3.03
4.17
MG
1.21
1.20
1.14
1.00
1.11
1.56
1.77
1.30
1.00
0.96
0.93
1.14
1.25
1.00
0.83
0.77
0.77
0.83
1.00
0.74
0.69
0.88
0.88
1.00
0.78
0.72
0.89
VG
1.11
1.13
1.13
1.00
1.47
3.19
5.79
3.65
1.00
1.14
1.17
1.38
1.48
1.00
1.07
1.16
1.15
1.07
1.00
1.17
1.32
1.48
1.22
1.00
1.15
1.28
1.35
RHC
39.01
41.76
41.72
47.48
81.11
43.77
42.77
77.00
22.36
32.31
19.11
19.00
30.36
547.74
766.40
766.57
764.86
764.70
25.20
27.48
29.40
29.61
28.48
54.77
54.62
58.26
58.51
geoR
0.99
0.99
0.98
1.00
0.97
0.94
0.93
0.96
1.00
0.99
0.99
0.99
0.99
1.00
1.00
0.99
0.99
1.00
1.00
0.97
0.95
0.94
0.97
1.00
0.97
0.94
0.94
ff2
0.88
0.94
0.90
1.00
0.82
0.68
0.68
0.82
1.00
0.94
0.98
0.90
0.80
1.00
1.00
0.90
0.96
1.00
1.00
1.00
0.70
0.66
0.92
1.00
1.00
0.68
0.88
TMS
0.91
0.92
0.93
0.89
0.79
0.62
0.67
0.68
0.90
0.89
0.87
0.90
0.78
0.90
0.89
0.90
0.92
0.95
0.91
0.90
0.81
0.83
0.91
0.93
0.92
0.81
0.89
-------
site
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
year
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
source
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
2
avg
time
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
Simulation
type
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
geo
mean
10.95
11.86
15.79
18.00
13.75
12.02
5.37
7.03
7.56
5.41
5.17
45.89
79.88
95.54
89.49
76.12
4.35
4.69
5.94
4.75
4.00
10.36
10.75
12.50
10.02
9.32
geo
stddev
4.01
2.75
2.97
2.67
3.68
4.23
2.80
3.53
3.69
6.45
5.54
4.90
4.78
4.27
4.60
5.10
5.33
5.14
5.14
5.62
6.81
4.82
4.74
4.90
5.60
6.42
MG
0.91
1.00
0.75
0.66
0.86
0.99
1.00
0.76
0.71
0.99
1.04
1.00
0.57
0.48
0.51
0.60
1.00
0.93
0.73
0.92
1.09
1.00
0.96
0.83
1.03
1.11
VG
1.23
1.00
1.23
1.43
1.47
1.55
1.00
1.29
1.52
2.98
2.03
1.00
1.39
1.86
1.67
1.33
1.00
1.20
1.64
1.43
1.37
1.00
1.26
1.58
1.47
1.54
RHC
56.60
57.23
57.95
55.85
58.24
56.36
25.52
28.62
29.47
31.63
29.94
509.64
975.60
975.32
972.25
972.25
34.07
27.03
27.62
29.58
27.79
75.77
54.11
57.20
61.26
55.01
geoR
0.96
1.00
0.95
0.91
0.90
0.91
1.00
0.95
0.92
0.90
0.93
1.00
1.00
0.99
0.99
1.00
1.00
0.97
0.93
0.94
0.96
1.00
0.95
0.92
0.93
0.94
ff2
0.90
1.00
0.98
0.60
0.60
0.84
1.00
0.72
0.62
0.46
0.86
1.00
0.82
0.44
0.68
0.88
1.00
0.84
0.70
0.68
0.70
1.00
0.76
0.64
0.64
0.62
TMS
0.91
0.93
0.90
0.77
0.80
0.87
0.93
0.82
0.78
0.72
0.84
0.89
0.73
0.69
0.75
0.85
0.98
0.87
0.79
0.83
0.85
1.00
0.84
0.79
0.83
0.80
C1 2012
24hr
obs
13.79 3.79
1.00
1.00
101.54 1.00 1.00 0.99
-------
site
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
year
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
source
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
2
3
3
3
avg
time
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
Simulation
type
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
geo
mean
12.97
15.51
9.57
9.04
4.02
4.31
5.46
3.67
2.86
51.68
75.94
92.41
87.08
75.73
4.91
6.09
4.24
4.52
6.18
11.19
14.38
9.90
10.66
14.56
9.32
12.73
8.68
geo
stddev
6.20
6.62
11.97
12.40
8.40
10.28
12.31
22.78
19.84
5.05
3.90
3.58
4.09
3.88
5.67
5.74
7.71
7.32
6.05
5.20
4.98
6.92
6.56
5.24
5.68
5.36
9.28
MG
1.06
0.89
1.44
1.53
1.00
0.93
0.74
1.10
1.40
1.00
0.68
0.56
0.59
0.68
1.00
0.81
1.16
1.09
0.79
1.00
0.78
1.13
1.05
0.77
1.00
0.73
1.07
VG
2.08
3.63
15.01
12.28
1.00
1.28
2.46
6.49
3.52
1.00
1.27
1.66
1.46
1.27
1.00
1.12
1.39
1.28
1.11
1.00
1.12
1.35
1.24
1.11
1.00
1.14
1.47
RHC
89.78
64.10
67.25
88.92
45.98
34.66
32.52
32.23
34.61
497.77
633.47
633.64
632.25
632.17
39.15
40.96
39.10
38.67
44.25
76.98
82.95
75.07
76.33
87.52
61.26
75.39
70.16
geoR
0.90
0.81
0.81
0.86
1.00
0.98
0.94
0.94
0.97
1.00
1.00
0.99
0.99
1.00
1.00
0.99
0.97
0.98
0.99
1.00
0.99
0.97
0.98
0.99
1.00
1.00
0.98
ff2
0.78
0.44
0.42
0.62
1.00
0.78
0.58
0.48
0.62
1.00
0.78
0.56
0.62
0.76
1.00
1.00
0.80
0.84
1.00
1.00
1.00
0.78
0.84
1.00
1.00
0.96
0.88
TMS
0.80
0.63
0.59
0.60
0.98
0.85
0.72
0.69
0.70
0.91
0.81
0.74
0.78
0.85
0.97
0.93
0.86
0.90
0.91
0.97
0.92
0.86
0.91
0.91
0.93
0.88
0.88
-------
site
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
year
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
source
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
2
3
3
3
3
3
avg
time
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
Simulation
type
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
geo
mean
8.97
11.27
3.66
4.18
1.95
2.17
3.62
59.23
69.95
69.22
70.85
69.80
3.06
3.41
2.61
2.56
3.34
6.92
7.97
6.09
6.33
8.08
9.11
8.98
5.72
6.15
8.59
geo
stddev
9.09
7.83
6.82
9.47
37.61
27.98
13.67
4.39
3.60
3.66
3.63
3.59
5.62
6.12
8.31
7.93
6.58
5.23
6.21
7.85
7.81
6.54
3.96
7.76
15.75
13.84
8.72
MG
1.04
0.83
1.00
0.88
1.87
1.69
1.01
1.00
0.85
0.86
0.84
0.85
1.00
0.90
1.17
1.19
0.92
1.00
0.87
1.14
1.09
0.86
1.00
1.02
1.59
1.48
1.06
VG
1.39
1.28
1.00
1.24
41.77
15.54
1.90
1.00
1.08
1.07
1.08
1.08
1.00
1.16
2.20
1.93
1.22
1.00
1.20
1.84
1.75
1.26
1.00
2.02
16.28
10.61
2.67
RHC
68.76
75.44
29.09
36.07
35.76
35.02
36.23
584.72
460.18
459.52
460.29
460.94
21.80
26.63
24.69
26.25
27.20
45.93
55.65
50.76
53.96
56.42
50.95
63.72
54.65
53.64
57.27
geoR
0.99
0.98
1.00
0.99
0.97
0.96
0.98
1.00
1.00
1.00
1.00
1.00
1.00
0.98
0.92
0.93
0.97
1.00
0.98
0.94
0.94
0.98
1.00
0.96
0.91
0.91
0.94
ff2
0.90
0.86
1.00
0.86
0.78
0.78
0.86
1.00
1.00
1.00
1.00
1.00
1.00
0.88
0.68
0.72
0.84
1.00
0.84
0.64
0.68
0.80
1.00
0.78
0.54
0.44
0.50
TMS
0.91
0.87
0.96
0.87
0.66
0.68
0.87
0.92
0.91
0.96
0.95
0.96
0.91
0.89
0.77
0.79
0.90
0.92
0.87
0.78
0.81
0.88
0.92
0.80
0.60
0.62
0.74
-------
site
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
year
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
source
4
4
4
4
4
5
5
5
5
5
avg
time
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
24hr
Simulation
type
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
geo
mean
2.81
2.11
0.94
1.15
1.84
50.34
73.76
75.92
76.48
73.90
geo
stddev
8.10
24.70
70.41
45.98
23.12
5.72
4.36
4.33
4.29
4.31
MG
1.00
1.34
2.99
2.44
1.53
1.00
0.68
0.66
0.66
0.68
VG
1.00
4.93
1041.86
137.94
4.83
1.00
1.28
1.32
1.34
1.29
RHC
24.85
27.89
26.04
27.56
28.05
707.21
755.45
756.41
755.74
754.92
geoR
1.00
0.98
0.94
0.93
0.98
1.00
1.00
0.99
0.99
0.99
ff2
1.00
0.66
0.38
0.42
0.70
1.00
0.78
0.78
0.78
0.80
TMS
0.95
0.70
0.52
0.54
0.70
0.96
0.83
0.84
0.84
0.85
-------
AERMOD results statistical scores: 24-hour average concentrations.
site
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
year
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2011
sourc
e
1
1
1
1
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
avg time
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
Simulation
type
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
geo
mean
0.27
0.20
0.21
0.25
0.23
0.66
0.57
0.57
0.66
0.65
0.79
0.50
0.44
0.37
0.45
0.23
0.00
0.00
0.00
0.00
6.27
11.15
11.05
11.33
11.41
0.23
geo
Stddev
7.33
19.46
16.16
13.28
17.27
6.56
16.70
14.11
12.05
15.00
6.24
33.36
37.16
55.68
44.70
12.37
NA
NA
NA
NA
7.02
4.82
4.79
4.77
4.79
13.23
MG
1.00
1.34
1.28
1.09
1.19
1.00
1.16
1.16
1.00
1.03
1.00
1.58
1.79
2.13
1.76
1.00
NA
NA
NA
NA
1.00
0.56
0.57
0.55
0.55
1.00
VG
1.00
5.55
4.08
2.69
3.99
1.00
4.83
3.95
2.81
3.65
1.00
48.56
90.02
690.36
192.95
1.00
NA
NA
NA
NA
1.00
1.61
1.60
1.66
1.66
1.00
RHC
2.54
2.32
2.23
2.21
2.32
5.81
5.64
5.34
5.36
5.66
6.64
7.49
7.35
7.34
7.51
2.70
2.53
2.27
2.23
2.49
127.75
136.88
131.94
134.53
138.55
3.21
geoR
1.00
0.94
0.94
0.94
0.95
1.00
0.94
0.92
0.93
0.94
1.00
0.93
0.93
0.92
0.92
1.00
0.92
0.91
0.91
0.91
1.00
1.00
1.00
1.00
1.00
1.00
ff2
1.00
0.40
0.36
0.52
0.46
1.00
0.46
0.36
0.48
0.48
1.00
0.36
0.34
0.36
0.38
1.00
0.40
0.40
0.44
0.44
1.00
0.58
0.58
0.56
0.56
1.00
TMS
0.82
0.64
0.66
0.75
0.69
0.82
0.69
0.67
0.75
0.72
0.82
0.57
0.56
0.55
0.57
0.82
NA
NA
NA
NA
0.85
0.74
0.75
0.74
0.74
0.82
-------
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B2
B3
B3
B3
B3
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2010
2010
2010
2010
1
1
1
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
0.25
0.19
0.23
0.27
0.62
0.65
0.52
0.61
0.69
0.84
0.54
0.28
0.33
0.54
0.13
0.00
0.00
0.00
0.14
7.29
12.39
12.32
12.39
12.43
0.19
0.43
0.41
0.34
14.14
15.97
12.24
12.90
10.68
13.20
14.69
11.57
12.34
8.55
27.00
67.10
51.08
28.54
47.63
NA
NA
NA
39.11
7.22
5.44
5.45
5.39
5.41
12.99
4.72
4.75
6.24
0.93
1.20
1.00
0.87
1.00
0.95
1.20
1.02
0.89
1.00
1.55
3.04
2.51
1.55
1.00
NA
NA
NA
0.94
1.00
0.59
0.59
0.59
0.59
1.00
0.44
0.48
0.56
1.26
1.92
1.61
1.23
1.00
1.48
2.52
1.94
1.45
1.00
10.39
1914.18
369.64
13.18
1.00
NA
NA
NA
1.46
1.00
1.44
1.43
1.45
1.46
1.00
5.73
5.36
2.66
2.61
2.05
2.23
2.63
7.12
5.66
4.40
4.57
5.59
8.93
6.84
5.38
5.43
6.69
3.73
2.83
2.04
2.18
2.75
137.57
162.48
163.10
162.70
162.45
2.24
2.12
1.90
1.95
0.98
0.96
0.96
0.99
1.00
0.97
0.94
0.94
0.97
1.00
0.94
0.89
0.89
0.94
1.00
0.92
0.96
0.91
0.99
1.00
1.00
1.00
1.00
1.00
1.00
0.99
0.99
0.99
0.78
0.64
0.62
0.82
1.00
0.64
0.52
0.56
0.70
1.00
0.46
0.36
0.36
0.46
1.00
0.68
0.52
0.64
0.72
1.00
0.66
0.64
0.64
0.66
1.00
0.66
0.68
0.70
0.86
0.75
0.82
0.87
0.82
0.81
0.69
0.79
0.81
0.83
0.58
0.47
0.53
0.59
0.83
NA
NA
NA
0.83
0.85
0.76
0.78
0.78
0.79
0.81
0.64
0.65
0.72
-------
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2011
2011
2011
2011
2011
2011
2011
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
0.35
0.51
1.00
0.92
0.81
0.85
0.51
1.13
1.04
0.83
0.93
0.10
0.41
0.39
0.30
0.31
6.66
9.46
9.46
9.54
9.54
0.37
0.42
0.36
0.35
0.36
0.96
0.98
6.79
10.18
4.45
4.62
5.92
6.12
9.99
4.38
4.68
7.01
6.32
38.10
5.23
5.38
8.13
8.34
6.43
5.32
5.29
5.32
5.34
12.84
7.45
8.70
8.39
9.66
11.15
6.98
0.55
1.00
0.50
0.55
0.63
0.59
1.00
0.45
0.49
0.61
0.55
1.00
0.23
0.25
0.32
0.31
1.00
0.70
0.70
0.70
0.70
1.00
0.90
1.03
1.06
1.04
1.00
0.98
2.31
1.00
3.59
3.31
2.09
1.90
1.00
4.15
3.37
1.68
1.97
1.00
517.88
383.22
48.37
48.10
1.00
1.19
1.19
1.19
1.19
1.00
1.62
1.59
1.76
1.29
1.00
1.37
2.17
4.46
4.28
3.77
3.83
4.34
4.17
4.75
4.23
4.27
4.86
2.05
2.37
2.23
2.28
2.40
118.96
141.50
141.72
143.29
143.07
4.62
4.27
3.33
3.58
4.30
10.22
8.96
1.00
1.00
0.98
0.97
0.97
0.99
1.00
0.99
0.98
0.98
0.99
1.00
0.99
0.99
0.99
0.99
1.00
1.00
1.00
1.00
1.00
1.00
0.98
0.97
0.97
0.99
1.00
0.99
0.68
1.00
0.64
0.66
0.68
0.66
1.00
0.78
0.78
0.80
0.86
1.00
0.60
0.66
0.62
0.62
1.00
0.82
0.82
0.84
0.84
1.00
0.78
0.74
0.70
0.78
1.00
0.84
0.71
0.81
0.67
0.67
0.75
0.73
0.81
0.67
0.69
0.80
0.76
0.81
0.54
0.57
0.59
0.58
0.85
0.84
0.87
0.87
0.88
0.83
0.84
0.82
0.82
0.87
0.83
0.88
-------
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2012
2012
2012
2012
2012
2012
2012
2012
2012
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
0.84
0.85
0.88
0.97
1.11
0.60
0.63
0.90
0.00
0.41
0.12
0.13
0.30
12.83
10.25
11.27
11.16
10.41
0.22
0.22
0.23
0.22
0.21
0.56
0.54
0.51
0.50
8.44
8.13
8.94
11.77
8.68
29.59
25.69
14.13
NA
10.50
82.03
69.57
17.69
6.29
6.16
5.35
5.48
6.14
10.15
7.39
6.39
6.44
8.12
7.99
6.42
6.31
6.53
1.13
1.12
1.09
1.00
0.87
1.62
1.54
1.07
NA
0.38
1.28
1.23
0.51
1.00
1.25
1.14
1.15
1.23
1.00
1.00
0.95
0.98
1.05
1.00
1.05
1.12
1.13
1.44
1.52
1.18
1.00
1.26
4.93
3.88
1.19
NA
1135.41
73.10
85.59
95.98
1.00
1.08
1.07
1.07
1.07
1.00
1.29
1.52
1.44
1.24
1.00
1.29
1.42
1.35
6.96
7.42
8.75
10.46
11.86
8.73
8.58
11.76
4.11
5.17
3.57
3.72
5.04
201 .86
170.88
159.27
164.49
172.76
2.16
1.64
1.20
1.21
1.62
4.54
4.01
2.89
3.01
0.97
0.97
0.99
1.00
0.99
0.97
0.97
0.99
1.00
0.94
0.89
0.89
0.95
1.00
1.00
1.00
1.00
1.00
1.00
0.98
0.98
0.98
0.98
1.00
0.97
0.96
0.97
0.80
0.72
0.92
1.00
0.90
0.80
0.82
0.88
0.98
0.70
0.78
0.78
0.74
1.00
1.00
1.00
1.00
1.00
1.00
0.76
0.62
0.62
0.84
1.00
0.78
0.76
0.74
0.83
0.84
0.90
0.83
0.89
0.67
0.74
0.88
NA
0.56
0.63
0.69
0.59
0.88
0.91
0.95
0.95
0.94
0.83
0.86
0.79
0.85
0.87
0.82
0.87
0.81
0.86
B3 2012
PERIOD MMIF.RCALT 0.51
7.21 1.11
1.27
4.13 0.97 0.80
0.84
-------
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
B3
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
2
3
3
3
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
0.76
0.65
0.54
0.47
0.55
0.15
0.17
0.00
0.14
0.00
6.28
6.43
7.18
7.07
6.54
0.40
0.38
0.35
0.31
0.33
0.90
0.85
0.80
0.72
0.76
1.09
0.98
1.00
5.76
8.68
10.95
13.93
12.86
22.98
19.93
NA
24.87
NA
6.61
6.24
5.65
5.73
6.18
5.93
5.99
6.70
8.26
7.84
5.60
5.66
6.02
7.58
7.23
5.25
5.84
5.66
1.00
1.17
1.40
1.63
1.38
1.00
0.88
NA
1.02
NA
1.00
0.98
0.87
0.89
0.96
1.00
1.06
1.14
1.29
1.22
1.00
1.07
1.13
1.25
1.18
1.00
1.11
1.10
1.00
1.34
2.31
4.00
2.55
1.00
1.16
NA
1.31
NA
1.00
1.01
1.06
1.05
1.02
1.00
1.04
1.14
1.41
1.26
1.00
1.04
1.14
1.39
1.22
1.00
1.08
1.15
5.53
6.15
4.46
4.36
5.98
2.49
2.41
1.47
1.47
2.30
110.61
115.19
111.46
111.63
115.32
4.26
3.77
3.62
3.76
3.89
8.93
8.00
7.35
7.78
8.04
10.55
10.07
8.87
1.00
0.99
0.96
0.96
0.98
1.00
0.99
0.91
0.99
0.94
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.99
0.98
0.99
1.00
1.00
0.98
0.97
0.99
1.00
0.99
0.98
1.00
0.84
0.70
0.64
0.82
1.00
0.92
0.70
0.80
0.92
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.90
0.82
0.86
1.00
1.00
1.00
0.84
0.86
1.00
0.96
0.98
0.81
0.87
0.71
0.69
0.73
0.82
0.92
NA
0.91
NA
0.83
0.98
0.96
0.97
0.98
0.83
0.96
0.92
0.85
0.89
0.83
0.96
0.93
0.86
0.90
0.84
0.95
0.92
-------
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
3
3
4
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
2
3
3
3
3
3
4
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
0.81
0.79
0.45
0.40
0.37
0.28
0.30
7.32
7.70
8.46
8.28
7.65
0.31
0.22
0.26
0.24
0.21
0.73
0.52
0.56
0.52
0.49
0.83
0.53
0.56
0.41
0.43
0.27
7.73
8.59
5.71
6.99
7.71
13.16
11.35
7.93
7.56
6.91
7.05
7.61
7.63
6.90
6.15
6.33
8.05
6.71
6.31
6.13
6.49
7.27
5.69
8.59
9.52
15.58
14.25
11.61
1.34
1.39
1.00
1.13
1.21
1.62
1.49
1.00
0.95
0.86
0.88
0.96
1.00
1.41
1.23
1.32
1.51
1.00
1.41
1.30
1.39
1.47
1.00
1.57
1.49
2.03
1.93
1.00
1.61
1.71
1.00
1.13
1.29
3.52
2.37
1.00
1.01
1.05
1.04
1.01
1.00
1.18
1.39
1.48
1.26
1.00
1.21
1.47
1.54
1.30
1.00
1.90
2.70
10.45
6.46
1.00
8.74
9.90
4.95
4.81
4.15
4.09
4.79
138.69
151.71
148.98
148.19
150.82
4.32
2.49
2.27
2.40
2.55
8.68
5.12
4.69
4.96
5.28
8.93
5.77
5.15
5.19
6.01
4.41
0.97
0.97
1.00
0.99
0.98
0.96
0.97
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.97
0.96
0.99
1.00
0.99
0.96
0.95
0.99
1.00
0.96
0.93
0.91
0.94
1.00
0.78
0.82
1.00
0.94
0.84
0.74
0.78
1.00
1.00
1.00
1.00
1.00
1.00
0.88
0.82
0.72
0.82
1.00
0.86
0.70
0.72
0.78
1.00
0.86
0.68
0.70
0.80
1.00
0.82
0.80
0.83
0.93
0.86
0.72
0.74
0.83
0.97
0.96
0.97
0.99
0.83
0.80
0.85
0.81
0.84
0.83
0.79
0.80
0.80
0.83
0.82
0.73
0.71
0.64
0.65
0.83
-------
C1
C1
C1
C1
C1
C1
C1
C1
C1
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
2012
2012
2012
2012
2012
2012
2012
2012
2012
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
4
4
4
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
2
3
3
3
3
3
4
4
4
4
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
0.18
0.18
0.13
0.14
5.27
4.35
4.95
4.74
4.40
0.29
0.27
0.21
0.23
0.27
0.70
0.67
0.49
0.54
0.67
0.64
0.71
0.46
0.48
0.62
0.21
0.20
0.00
0.10
14.34
17.24
29.22
23.91
8.02
7.11
5.95
6.22
6.96
9.19
8.07
10.77
9.88
8.56
8.56
6.98
9.84
9.02
7.46
11.75
7.87
16.61
15.72
11.34
14.41
15.00
NA
47.00
1.49
1.46
2.02
1.96
1.00
1.21
1.06
1.11
1.20
1.00
1.06
1.40
1.28
1.05
1.00
1.05
1.43
1.30
1.05
1.00
0.91
1.40
1.34
1.05
1.00
1.04
NA
2.21
1.29
1.82
6.36
3.15
1.00
1.07
1.14
1.12
1.07
1.00
1.11
1.36
1.30
1.09
1.00
1.16
1.31
1.21
1.12
1.00
1.37
1.34
1.22
1.12
1.00
1.08
NA
9.88
2.86
2.51
2.48
2.82
101.40
97.07
95.30
95.43
97.09
2.27
2.66
2.22
2.43
2.72
4.83
5.59
4.53
4.86
5.77
5.64
6.81
5.48
5.31
6.79
2.22
2.79
2.13
2.27
1.00
0.98
0.97
0.99
1.00
1.00
1.00
1.00
1.00
1.00
0.99
0.98
0.98
0.99
1.00
0.99
0.99
0.99
0.99
1.00
0.99
1.00
1.00
0.99
1.00
1.00
0.95
0.99
0.90
0.72
0.62
0.78
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.78
0.82
1.00
1.00
0.96
0.84
0.86
1.00
1.00
0.80
0.86
0.88
1.00
1.00
1.00
0.70
0.72
0.80
0.76
0.65
0.70
0.83
0.94
0.96
0.96
0.95
0.81
0.94
0.81
0.85
0.95
0.81
0.93
0.82
0.88
0.94
0.81
0.85
0.82
0.88
0.92
0.81
0.94
NA
0.64
-------
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
C2
2010
2010
2010
2010
2010
2010
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
2012
4
5
5
5
5
5
1
1
1
1
1
2
2
2
2
2
3
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
PERIOD
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
obs
0.17
8.28
6.50
7.02
7.07
6.54
0.26
0.21
0.17
0.18
0.20
0.63
0.48
0.41
0.42
0.47
0.77
22.03
5.40
6.14
5.79
5.77
6.13
7.62
7.48
8.30
7.58
7.75
6.84
7.73
8.01
7.34
7.87
6.21
1.23
1.00
1.27
1.18
1.17
1.27
1.00
1.24
1.51
1.49
1.30
1.00
1.31
1.54
1.52
1.34
1.00
1.42
1.00
1.08
1.04
1.03
1.08
1.00
1.12
1.82
1.72
1.17
1.00
1.16
1.71
1.55
1.20
1.00
2.72
124.53
110.17
108.08
107.79
110.04
2.68
2.15
1.69
1.80
2.18
5.82
4.72
3.55
3.74
4.68
6.84
0.99
1.00
1.00
1.00
1.00
1.00
1.00
0.99
0.95
0.95
0.99
1.00
0.99
0.96
0.97
0.99
1.00
0.88
1.00
1.00
1.00
1.00
1.00
1.00
0.94
0.74
0.78
0.88
1.00
0.88
0.80
0.82
0.88
1.00
0.85
0.85
0.92
0.96
0.96
0.94
0.82
0.89
0.74
0.79
0.86
0.82
0.86
0.75
0.81
0.85
0.82
C2 2012
PERIOD AERC.RCALT 0.47 12.60 1.64
2.51
6.71 0.98 0.76
0.75
C2 2012
PERIOD AERC.RCALF 0.33 22.03 2.36
18.23
4.72 0.95 0.70
0.57
C2 2012
PERIOD MMIF.RCALF 0.33 18.54 2.36
11.19
4.55 0.96 0.72
0.63
C2 2012
PERIOD MMIF.RCALT 0.44 13.13 1.74
2.85
6.50 0.98 0.76
0.67
C2 2012
PERIOD
obs
0.22 13.24 1.00
1.00
2.88 1.00 1.00 0.82
C2
C2
C2
C2
2012
2012
2012
2012
4
4
4
4
PERIOD
PERIOD
PERIOD
PERIOD
AERC.RCALT
AERC.RCALF
MMIF.RCALF
MMIF.RCALT
0.11
0.00
0.06
0.10
37.52
NA
63.53
37.13
2.05
NA
3.87
2.24
5.77
NA
234.87
7.13
2.79
1.80
1.84
2.65
0.99
0.96
0.95
0.99
0.74
0.62
0.64
0.72
0.67
NA
0.56
0.60
C2 2012
PERIOD
obs
5.69 6.66
1.00
1.00
86.70 1.00 1.00
0.82
C2 2012
PERIOD AERC.RCALT 5.31
6.68 1.07
1.01
85.23 1.00 1.00
0.98
-------
C2 2012 5 PERIOD AERC.RCALF 5.76 6.05 0.99 1.02 85.51 1.00 1.00 0.99
C2 2012 5 PERIOD MMIF.RCALF 5.58 6.24 1.02 1.02 85.59 1.00 1.00 0.99
C2 2012 5 PERIOD MMIF.RCALT 5.22 6.81 1.09 1.01 85.07 1.00 1.00 0.98
-------
[Blank]
-------
APPENDIX C: REPORT DISK
-------
-------
Volume 3 results can be requested from
Eric Wolvovsky
BOEM/OEP/DEA
Mail Stop: VAM-OEP
45600 Woodland Road
Sterling VA 20166
703-787-1719
Email: eric.wolvovsky@boem.gov
------- |