1. 1 State Implementation Plan (SIP) Modeling for 8-hour Ozone Preliminary 2002 Results For Triangle and Rocky Mount Stakeholders Mike Abraczinskas, NCDAQ Laura Boothe, NCDAQ George Bridgers, NCDAQ May 31, 2005

2. 2 Outline Ozone overview SIP Modeling overview Meteorological modeling Emissions modeling Air Quality modeling Future year emissions summary Menu of possible control options Next steps

3. 3 Ozone and SIP Modeling Overview Laura Boothe, NCDAQ Attainment Planning Branch Chief

4. 4

5. 5 Ozone – Public Health Risks When inhaled, even at low levels, ozone can: –Cause inflammation of lung tissue –Cause acute or chronic respiratory problems –Aggravate, possibly trigger asthma –Decrease lung capacity –Repeated exposure in children may lead to reduced lung function as adults

6. 6 Background 8-hour ozone standard –If a monitored design value is > 0.08 ppm (84 ppb), that monitor is violating the standard –The design value is defined as: 3-year average of the annual 4 th highest daily maximum 8-hour average

7. 7 2001-2003 Ozone Design Values (Highest Value Per County)

8. 8 Violating Ozone Monitors Based on 01-03 data Green dots = attaining monitors Red dots = violating monitors

9. 9 NC 8-hr Ozone Nonattainment Areas

10. 10 Triangle 8-hr Ozone Design Values Monitor Millbrook Butner Duke St Franklinton Bushy Fork Tower W Johnston Fuquay-Varina Pittsboro County Wake Granville Durham Franklin Person Wake Johnston Wake Chatham 01-03 92 94 89 90 91 85 88 82 02-04 88 89 86 88 85 86 83 87 79 2005* 90 83 96 90 99 91 100 88 111 * 4 th highest 8-hr max in 2005 can be no higher than this value in order to attain by the end of the 2005 ozone season. ** Number of times the 4 th highest has been this value or lower in the last 5 years. # ** 4 of 5 1 of 5 4 of 5 5 of 5 3 of 5 5 of 5

11. 11 Rocky Mount 8-hr Ozone Design Values Monitor Leggett County Edgecombe 01-03 89 02-04 85 2005* 94 * 4 th highest 8-hr max in 2005 can be no higher than this value in order to attain by the end of the 2005 ozone season. ** Number of times the 4 th highest has been this value or lower in the last 5 years. # ** 4 of 5

12. 12 Ozone Nonattainment Timeline Immediate (June 15, 2004) –New source review One year –Transportation conformity Three years –State Implementation Plan (SIP) – attainment demonstration Five years (or as expeditiously as practicable) –Basic areas attain standard (Triangle, RMT, GSMNP) Six years (or as expeditiously as practicable) –Moderate areas attain standard (Metrolina)

13. 13 Ozone Nonattainment Timeline Definitions for Triangle and RMT Areas Effective date = Transportation conformity date = SIP submittal date = Attainment date = Data used to determine attainment = (Modeling) Attainment year = Redesignation base years = Maintenance years = June 15, 2004 June 15, 2005 June 15, 2007 June 15, 2009* 2006-2008 2008 2005 or 2006 TBD * Or as early as possible

14. 14 State Implementation Plan (SIP) Need a SIP submittal to EPA within three years –Attainment Demonstration that details the States plan to bring the area into attainment of the Federal standard –Triangle and RMT areas…must include: VOC & NOX Reasonably Available Control Technology (RACT) Reasonably Available Control Measures (RACM) Reasonable Further Progress (RFP)

15. 15 State Implementation Plan (SIP) RACM/RACT Requirements –Applies to all source sectors (point, area, highway mobile & off-road mobile sources) –Only what is necessary to attain NAAQS –NC has already adopted some RACM/RACT type rules Open burning ban during ozone events Expanded I/M program RACT rules for Wake and Durham Counties as contingency measures for 1-hr ozone maintenance RFP Requirements –Must show reductions in future year emissions

16. 16 State Implementation Plan (SIP) Need a SIP submittal to EPA within three years –Attainment Demonstration that details the States plan to bring the area into attainment of the Federal standard Most significant emission controls are already underway –Clean Smokestacks Act –Vehicle emissions testing –Ultra-Low sulfur fuels –Cleaner engines

17. 17 VISTAS Visibility Improvement State and Tribal Association of the Southeast Regional Planning Organization established under the 1999 Regional Haze Rule Collaborative effort of States and Tribes to support management of regional haze and related air quality issues in the Southeastern US No independent regulatory authority and no authority to direct or establish State or Tribal law or policy.

18. 18

19. 19

20. 20 Met, Emissions and AQ Model performance and protocol Emissions Inventories 2002 & 2009

21. 21 Modeling Application Process Select areas or domains of interest Select representative ozone season/episodes Prepare and refine meteorological simulations Prepare and refine emission model inputs Apply air quality modeling system Performance evaluation on episodes Prepare current and future year emissions (Projected and Potential Control Strategies) Re-apply air quality modeling system Analyze the effectiveness of control strategies Apply the attainment test

22. 22 Air Quality Modeling System Meteorological Model Emissions Processor Air Quality Model MM5 SMOKE CMAQ Sparse Matrix Operator Kernel Emissions Community Multiscale Air Quality System Temporally and Spatially Gridded Air Quality Output predictions

23. 23 Modeling Domains 36 km 12 km

24. 24 Grid Structure Horizontal: 36 km & 12 km Vertical: MM5 = 34 layers SMOKE & CMAQ = 19 layers Layer 1 = 36 m deep Ground ~48,000 ft

25. 25 Modeling Season / Episode Full Year of 2002 selected for VISTAS modeling –Regional Haze / Fine Particulate: Full Year –Ozone: Late May – End Of August The higher portion of the 2002 ozone season selected for the Ozone SIP and Attainment Demonstration modeling.

26. 26 Meteorological Modeling Overview George Bridgers, NCDAQ Meteorologist

27. 27 Meteorological Modeling Penn State / NCAQ MM5 meso-scale meteorological model –Version 3.6.1+ –Widely used in the research and regulatory communities –VISTAS Contracted With Barons Advanced Meteorological Systems (BAMS) –Run at both 36km (Nationwide) and 12km (Southeastern US) resolutions

28. 28 Met Model Performance Model Performance For Key Variables: –Temperature –Moisture (Mixing Ratio & Relative Humidity) –Winds –Cloud Cover –Precipitation Comparisons With Other Met Modeling Studies Summary Of Met Model Performance

29. 29 Model Performance Statistics Meteorology In North Carolina May, June, July, August, and September (MJJAS)

30. 30 Overall diurnal pattern captured very well Slight cool bias in the daytime Slight warm bias overnight Temperature

31. 31 MayJune JulyAugust

32. 32 Moisture (Mixing Ratio) Tracks observed trends fairly well Low bias in the morning through the early afternoon High bias in the late afternoon and at night

33. 33 MayJune JulyAugust

34. 34 High bias in the daytime Low bias at night RH is linked to temperature and moisture biases Moisture (Relative Humidity)

35. 35 ~1 mph high bias day, ~2 mph high bias at night –Partly due to relative inability of winds in the model to go calm (There is always some wind) –Also due to starting thresholds of observation network… network cant measure winds < 3 mph, so winds < 3 mph are reported as calm Wind Speed

36. 36 MayJune JulyAugust

37. 37 MayJune JulyAugust

38. 38 General overestimation of clouds in the met model Greatest bias overnight & smallest bias early afternoon Nighttime cloud observations questionable Bias ~4% in May, peaks at ~15% in July, and declines to ~3% in September Cloud Cover

39. 39 General over prediction of clouds (example – July 18 2PM) Cloud Cover

40. 40 Mixed precipitation performance… typical of any summertime weather pattern / forecast Good performing day (Spatially and magnitude): Precipitation

41. 41 Poorer performing day (Magnitude okay is spots, but significant precip I-95 corridor that is false): Precipitation

42. 42 Observed Precip MAY Observed Precip JUNE Modeled Precip MAY Modeled Precip JUNE

43. 43 Observed Precip JULY Observed Precip AUGUST Modeled Precip JULY Modeled Precip AUGUST

44. 44 Comparisons With Other Met Modeling Studies The next series of slides are adapted from Alpine Geophysics documentation for the VISTAS AQ Modeling project. The bar charts are comparisons of VISTAS Phase I (Sensitivities) MM5 modeling to other national and Southeast regional MM5 simulations The performance characteristics of VISTAS Phase I MM5 modeling is very similar to VISTAS Phase II (Annual) MM5 Modeling

45. 45 National MM5 Comparisons

46. 46 The 3 green bars: - VISTAS 1 = January 2002 episode - VISTAS 2 = July 2001 episode - VISTAS 31 = July 1999 episode The yellow bars: - USEPAs 2001 Annual MM5 simulation

47. 47 The 3 green bars: - VISTAS 1 = January 2002 episode - VISTAS 2 = July 2001 episode - VISTAS 31 = July 1999 episode The yellow bars: - USEPAs 2001 Annual MM5 simulation

48. 48 The 3 green bars: - VISTAS 1 = January 2002 episode - VISTAS 2 = July 2001 episode - VISTAS 31 = July 1999 episode The yellow bars: - USEPAs 2001 Annual MM5 simulation

49. 49 The 3 green bars: - VISTAS 1 = January 2002 episode - VISTAS 2 = July 2001 episode - VISTAS 31 = July 1999 episode The yellow bars: - USEPAs 2001 Annual MM5 simulation

50. 50 The 3 green bars: - VISTAS 1 = January 2002 episode - VISTAS 2 = July 2001 episode - VISTAS 31 = July 1999 episode The yellow bars: - USEPAs 2001 Annual MM5 simulation

51. 51 Southeast Regional MM5 Comparisons

52. 52 North Carolina MJJAS 2002 T Error = 1.55 for all pairs

53. 53 North Carolina MJJAS 2002 WS RMSE = 1.84 for all pairs WS RMSE = 1.54 for no calms

54. 54 Closer to 1.0 indicates better performance North Carolina MJJAS 2002 WS IA = 0.73 for all pairs WS IA = 0.74 for no calms

55. 55 Take Away Messages The 2002 meteorological model performance: –Compares favorably to the performance in similar modeling projects / studies, including that of EPA –Can be considered State Of The Science The daytime biases would tend to contribute to lower ozone concentrations in the AQ model: –Cooler afternoon high temperatures –Higher relative humidity –Rapid atmospheric moisture increase late day –Greater cloud and precipitation coverage –Slightly higher wind speeds –Generally, a little too much atmospheric mixing

56. 56 2002 Emissions Overview Mike Abraczinskas, NCDAQ Environmental Engineer II

57. 57 Emissions Inventory Definitions ActualActual = the emissions inventory developed to simulate what happened in 2002 TypicalTypical = the emissions inventory developed to characterize the current emissions… It doesnt include specific events, but rather averages or typical conditions (e.g. Electric Generating Units and fires) FutureFuture = the emissions inventory developed to simulate the future (e.g. 2008 for Triangle and Rocky Mount modeling) ***Note… Actual is used for model performance evaluation only! Typical and Future are used to determine future attainment status.

58. 58 Emission Source Categories –Point sources: utilities, refineries, industrial sources, etc. –Area sources: gas stations, dry cleaners, farming practices, fires, etc. –Motor vehicles: cars, trucks, buses, etc. –Nonroad mobile sources: agricultural equipment, recreational marine, lawn mowers, construction equipment, etc. –Biogenic: trees, vegetation, crops

59. 59 VISTAS 2002 Inventory Actual inventory developed for model evaluation Utilize June 2004 State Consolidated Emissions Reporting Rule (CERR) submittals –Actual 2002 calendar year inventories (Annual 2002) Augment State data where pollutants missing Process onroad mobile through MOBILE6 module of SMOKE emissions system Generate fires as specific daily events Improved temporal and spatial allocation for modeling –Use of actual Continuous Emissions Monitor (CEM) distributions –New CMU monthly ammonia (NH3) profiles by county/SCC

60. 60 VISTAS 2002 Inventory - Point Annual 2002 –Includes Electric Generating Units (EGUs), non-EGU point source data –Reviewed by stakeholders Hourly EGU data generated to temporally allocate emissions during appropriate episodes –Used United State Environmental Protection Agency (USEPA) CEM and stakeholder provided data

61. 61 VISTAS 2002 Inventory - Fire Annual 2002 –Includes agricultural, prescribed, land clearing and wildfire data Modeling files generated using more specific raw data –Includes acres, dates, and locations of fire activity –Generated elevated fire file for sources with appropriate data elements (large wildfires and prescribed burns) –Non-elevated sources retained in county-level area source file

62. 62 VISTAS 2002 Inventory - Area Annual 2002 CMU NH3 model v.3.6 –Provides NH3 estimates from agricultural practices and other animal waste

63. 63 VISTAS 2002 Inventory – Onroad and Nonroad Onroad –Annual 2002 VMT and MOBILE6 inputs collected from States / Locals Nonroad –Annual 2002

64. 64 Emission Processing GriddingSpeciationTemporalEmission Inventory SMOKE Emission Model Air Quality Model

65. 65 Gridding 36 km 12 km

66. 66 36 km 12 km Speciation Converts emissions inventory VOCs to Carbon Bond IV Species

67. 67 Temporal 36 km 12 km Adjusts the annual emissions/data to the month of the year, day of the week and to the hour of the day Weekday diurnal profile for On-road Mobile

68. 68 Emission Processing GriddingSpeciationTemporalEmission Inventory SMOKE Emission Model Air Quality Model

69. 69 2002 Air Quality Modeling Overview George Bridgers, NCDAQ Meteorologist

70. 70 Air Quality Modeling Community Multiscale Air Quality Model (CMAQ) –Version 4.4 (With SOA Modifications) –Widely used in the research & regulatory communities –VISTAS Contracted With UC-Riverside, Alpine Geophysics LLC, and ENVIRON International Corp –Run at both 36km (Nationwide) and 12km (Southeastern US) resolutions

71. 71 AQ Model Performance Triangle Modeled Ozone Performance –1 & 8 Hour Statistical Tables –1 & 8 Hour Time Series And Statistical Plots Rocky Mount Modeled Ozone Performance –1 & 8 Hour Statistical Tables –1 & 8 Hour Time Series And Statistical Plots Ozone Spatial Plots and Animations Summary Of AQ (Ozone) Model Performance

72. 72 Triangle AQ Monitoring Network Overview Model Performance Statistical Tables –1 Hour Ozone Statistics –8 Hour Ozone Statistics Monitor Time Series And Statistical Plots –Rural Site: Pittsboro –Urban Site: Duke Street –Annual Site: Millbrook

73. 73 AQ Monitor Network Overview

74. 74 Model Performance Statistics 1 Hour Ozone

75. 75 Model Performance Statistics 8 Hour Ozone

76. 76 Pittsboro – 1 Hour Time Series

77. 77

78. 78

79. 79

80. 80

81. 81 Pittsboro – 8 Hour Time Series

82. 82

83. 83

84. 84

85. 85

86. 86 Duke Street – 1 Hour Time Series

87. 87

88. 88

89. 89

90. 90

91. 91 Duke Street – 8 Hour Time Series

92. 92

93. 93

94. 94

95. 95

96. 96 Millbrook – 1 Hour Time Series

97. 97

98. 98

99. 99

100. 100

101. 101

102. 102

103. 103

104. 104

105. 105

106. 106

107. 107

108. 108

109. 109 Millbrook – 8 Hour Time Series

110. 110

111. 111

112. 112

113. 113

114. 114

115. 115

116. 116

117. 117

118. 118

119. 119

120. 120

121. 121

122. 122 Rocky Mount AQ Monitoring Network Overview Model Performance Statistical Tables –1 Hour Ozone Statistics –8 Hour Ozone Statistics Monitor Time Series And Statistical Plots –Leggett

123. 123 AQ Monitor Network Overview

124. 124 Model Performance Statistics 1 Hour Ozone

125. 125 Model Performance Statistics 8 Hour Ozone

126. 126 Leggett – 1 Hour Time Series

127. 127

128. 128

129. 129

130. 130

131. 131 Leggett – 8 Hour Time Series

132. 132

133. 133

134. 134

135. 135

136. 136 Spatial Plots And Animations Daily 1 Hour Peak Model Ozone Spatial Plots With Observations Overlaid –June 8 – 18 –July 14 – 20 –August 17 – 29

137. 137 June 8 – 18, 2002 Daily 1 Hour Peak Plots

138. 138

139. 139

140. 140

141. 141

142. 142

143. 143

144. 144 July 14 – 20, 2002 Daily 1 Hour Peak Plots

145. 145

146. 146

147. 147

148. 148

149. 149 August 17 – 29, 2002 Daily 1 Hour Peak Plots

150. 150

151. 151

152. 152

153. 153

154. 154

155. 155

156. 156

157. 157 Take Away Messages Under-predictions of the afternoon peak modeled ozone concentrations account for the majority of the negative bias and error. There are not significant spatial or temporal errors with the modeled ozone that held consistently throughout the 2002 Ozone Season. Episodic air quality (ozone) cycles are well captured by the CMAQ air quality model with reasonable buildup and clean-out of ozone concentrations.

158. 158 Take Away Messages Thinking ahead to Typical and Future year modeling, Relative Reduction Factor (RRF) calculations, and the Modeled Attainment Test: –The relative sense of the modeling will make the afternoon peak under-predictions of ozone less significant and not influence strategy decisions. –There are a sufficient number of modeled days in this Base or Actual year modeling at each monitoring location that exceeds the 70ppb threshold to compute RRFs without the need for additional modeling.

159. 159 2002 Typical and 2009 Emissions Overview Mike Abraczinskas, NCDAQ Environmental Engineer II Note !!! –2008 emissions are being developed for the Triangle and Rocky Mount nonattainment areas Preliminary 2008 emissions and air quality modeling to be performed by NCDAQ –2009 emissions are being presented here today as a surrogate for 2008

160. 160 Emissions Inventory Definitions ActualActual = the emissions inventory developed to simulate what happened in 2002 TypicalTypical = the emissions inventory developed to characterize the current (2002) emissions… It doesnt include specific events, but rather averages or typical conditions (e.g. EGUs and fires) FutureFuture = the emissions inventory developed to simulate the future (e.g. 2008 for Triangle and Rocky Mount modeling) ***Remember… Actual is used for model performance evaluation only! Typical and Future are used to determine future attainment status.

161. 161 2002 & 2009 Emissions Comparison

162. 162 2002 & 2009 Emissions Comparison

163. 163 2002 typical and 2009 Point Source Summary Triangle nonattainment area –NOx and VOC bar charts Rocky Mount nonattainment area –NOx and VOC bar charts Plots of emission differences 2009-2002

164. 164

165. 165

166. 166

167. 167 * ** * Triangle and Rocky Mount nonattainment areas

168. 168 Point Source NOx 2009 minus 2002 (daily max difference, all layers) Increases only Scale 0 to 0.1 moles/s

169. 169 Point Source NOx 2009 minus 2002 (daily max difference, all layers) Decreases only Scale 0 to -0.1 moles/s

170. 170

171. 171

172. 172 Point Source VOC 2009 minus 2002* (daily max difference, all layers) Increases only Scale 0 to 0.1 moles/s

173. 173 Point Source VOC 2009 minus 2002* (daily max difference, all layers) Decreases only Scale 0 to -0.1 moles/s

174. 174 2002 typical and 2009 Area Source Summary Triangle nonattainment area –NOx and VOC bar charts Rocky Mount nonattainment area –NOx and VOC bar charts Statewide breakdown of area source NOx and VOC sources

175. 175 Insert bar charts here

176. 176

177. 177

178. 178

179. 179

180. 180

181. 181 2002 typical and 2009 Nonroad Source Summary Triangle nonattainment area –NOx and VOC bar charts Rocky Mount nonattainment area –NOx and VOC bar charts Plots of emission differences 2009-2002 Statewide breakdown of area source NOx and VOC sources

182. 182

183. 183

184. 184 NONROAD NOx 2009 minus 2002* (max difference) Reductions only Scale 0 to –0.1 moles/s

185. 185

186. 186

187. 187

188. 188

189. 189 2002 typical and 2009 Onroad Mobile Source Summary Triangle and Rocky Mount nonattainment areas –NOx and VOC Plots of emission differences 2009-2002 Animation of 2009 NOx Triangle NOx per county per vehicle type

190. 190

191. 191

192. 192 ONROAD Mobile NOx 2009 minus 2002* (max difference) Reductions only Scale 0 to –0.5 moles/s

193. 193 Chatham County 2009 NOx Emissions 2002 NOx Emissions

194. 194 Durham County 2002 NOx Emissions 2009 NOx Emissions

195. 195 Franklin County 2009 NOx Emissions2002 NOx Emissions L;;llllll

196. 196 Granville County 2009 NOx Emissions2002 NOx Emissions

197. 197 Johnston County 2009 NOx Emissions2002 NOx Emissions

198. 198 Orange County 2009 NOx Emissions 2002 NOx Emissions

199. 199 Person County 2009 NOx Emissions2002 NOx Emissions

200. 200 Wake County 2009 NOx Emissions 2002 NOx Emissions

201. 201

202. 202

203. 203 Identification of Potential NOx and VOC Control Measures Laura Boothe, NCDAQ Attainment Planning Branch Chief

204. 204 What is Needed to Show Attainment? –Will review preliminary air quality results to see how close we are to meeting the 8-hour ozone NAAQS –If not attaining, will look for additional NOx controls Will have to address RACM/RACT requirements Will review emission inventories and potential control measures to get greatest reductions for the cost –Need Stakeholders to assist in coming up with potential cost effective control measures

205. 205 Schedule/Next Steps When do we expect to have preliminary future year air quality modeling results? –2009 in mid-June –2008 in mid-September July 13, 2005 meeting –Review 02-09 emissions –Present preliminary air quality modeling results Well present the 2009 results as an indicator of how close well be in the 2008 run Attainment test –Control Strategy discussion (if needed) Controls needed for 8-hr ozone NAAQS –Outline next steps

206. 206 Contributors South Carolina Department of Health and Environment Conservation Pat Brewer, VISTAS Greg Stella, Alpine Geophysics Cyndi Loomis, Alpine Geophysics Don Olerud, Baron Advanced Meteorological Systems Bill Barnard, MACTEC Ed Sabo, MACTEC Kristen Theising, PECHAN Ralph Morris, ENVIRON Gail Tonneson, University of California-Riverside Dennis McNally, Alpine Geophysics Jim Boylan, Georgia Environmental Protection Department Sheila Holman, NCDAQ Bebhinn Do, NCDAQ Nick Witcraft, NCDAQ Phyllis Jones, NCDAQ Vicki Chandler, NCDAQ Pat Bello, NCDAQ Bob Wooten, NCDAQ Matt Mahler, NCDAQ Janice Godfrey, NCDAQ Ming Xie, NCDAQ Mildred Mitchell, NCDAQ VISTAS Stakeholders

207. 207 Questions/Comments http://ncair.org Laura Boothe, Chief of Attainment Planning 919-733-1488 Laura.Boothe@ncmail.net Mike Abraczinskas, Environmental Engineer II 919-715-3743 Michael.Abraczinskas@ncmail.net George Bridgers, Meteorologist 919-715-6287 George.Bridgers@ncmail.net

208. 208 Thank You!