A Call for an Aloft Air Quality Monitoring Network: Need, Feasibility

Subscriber access provided by University of South Dakota

Feature

A Call for an Aloft Air Quality Monitoring Network: Need, Feasibility, and Potential Value Rohit Mathur, Christian Hogrefe, Amir Hakami, Shunliu Zhao, James Szykman, and Gayle Hagler Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.8b02496 • Publication Date (Web): 04 Sep 2018 Downloaded from http://pubs.acs.org on September 6, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 21

Environmental Science & Technology

2

A Call for an Aloft Air Quality Monitoring Network: Need, Feasibility, and Potential Value

3 4

Rohit Mathur1*, Christian Hogrefe1, Amir Hakami2, Shunliu Zhao2, James Szykman1, Gayle Hagler1

1

1

5 6 7

National Exposure Research Laboratory, Office of Research and Development, U.S. Environmental Protection Agency, RTP, NC, USA 2 Department of Civil and Environmental Engineering, Carleton University, Ottawa, Canada

8

*

9

Abstract: Changing precursor emission patterns in conjunction with stringent health protective

10

air quality standards, necessitate accurate quantification of non-local contributions to ozone

11

pollution at a location due to atmospheric transport, that by nature predominantly occurs aloft

12

nocturnally. Concerted efforts to characterize ozone aloft on a continuous basis to quantify its

13

contribution to ground-level concentrations however are lacking. Applying our classical

14

understanding of air pollution dynamics to analyze variations in widespread surface-level ozone

15

measurements, in conjunction with process-based interpretation from a comprehensive air

16

pollution modeling system and detailed backward-sensitivity calculations that quantitatively link

17

surface-level and aloft pollution, we show that accurate quantification of the amount of ozone in

18

the air entrained from aloft every morning as the atmospheric boundary layer grows is the key

19

missing component for characterizing background pollution at a location, and propose a cost-

20

effective continuous aloft ozone measurement strategy to address critical scientific gaps in

21

current air quality management. Continuous aloft air pollution measurements can cost-effectively

22

be achieved through leveraging advances in sensor technology and proliferation of tall

23

telecommunications masts. Resultant improvements in ozone distribution characterization at

24

400-500m altitude are estimated to be 3-4 times more effective in characterizing the surface-

25

level daily maximum 8-hour average ozone (DM8O3) than improvements from surface

26

measurements since they directly quantify the amount of pollution imported to a location, and

27

furnish key-missing information on processes and sources regulating background ozone and its

28

modulation of ground-level concentrations. Since >80% of the DM8O3 sensitivity to

29

tropospheric ozone is potentially captured through measurements between 200-1200m altitude (a

30

possible design goal for future remote sensing instrumentation), their assimilation will

31

dramatically improve air quality forecast and health advisories.

Correspondence to: Rohit Mathur ([email protected])

1

ACS Paragon Plus Environment


32 33

Introduction

34

Information on the state of the atmosphere inferred from ground-based air quality monitoring

35

networks has played a central role in characterizing and improving air quality in many parts of

36

the world. For instance, the Air Quality System (AQS) (1) which contains ambient concentration

37

measurements of criteria air pollutants, collected by EPA, state, local, and tribal air quality

38

agencies, has helped to identify areas where current air quality is unacceptable as well as to

39

prevent deterioration in areas where air is relatively free of contamination. Trends in ground-

40

level ozone (O3) design values (i.e., the three-year average of the annual 4th highest daily

41

maximum 8-hour average O3 mixing ratio) show significant reductions resulting from

42

implementation of control measures and technological advances that have reduced emissions of

43

NOx and volatile organic compounds (VOCs) precursor species across the U.S. (2). However, as

44

air quality standards have tightened from a 1-hour averaged concentration to an 8-hour averaged

45

concentration with a lower threshold value, more rural and suburban regions have fallen into

46

non-attainment (3), emphasizing the need to quantify the contributions of long-range air

47

pollutant transport to local air quality.

48 49

At most land-based locations, ground-level O3 exhibits a characteristic, well-recognized diurnal

50

cycle (4) with a typical mid-afternoon maximum and a night-time minimum that arises from

51

diurnal variations in activity patterns of precursor emission sources, the atmospheric boundary

52

layer, and photo-chemistry in the atmosphere. Surface deposition, titration by NO emissions, and

53

the lack of chemical production in a shallow inversion layer which decouples O3 at the ground

54

from upper levels, lead to the night-time suppression of ground-level O3. The break-up of the

55

nocturnal inversion, entrainment of O3-rich air from above and onset of O3 production via

56

photochemical reactions involving NOx and VOC precursor species, result in increase in O3

57

levels in the morning hours leading to afternoon peak values that then decline as the sun sets and

58

chemical production declines.

59 60

Influence of Aloft Air Pollution on Surface Concentrations

61

The classical view of diurnal O3 variations can alternatively be examined as a rate of change (or

62

tendency) in the ground-level O3, and comprehensive atmospheric chemistry-transport models 2


Page 2 of 21

Page 3 of 21


63

can be instrumented to provide a quantitative view of the relative importance of the various

64

atmospheric dynamical and chemical process in shaping this diurnal evolution of O3 mixing

65

ratios. As an example, Figure 1 presents representative summer-time diurnal variations in median

66

process tendencies regulating ground-level O3 for the contiguous U.S., simulated by the

67

Community Multiscale Air Quality (CMAQ) modeling system (5). While O3 mixing ratios

68

typically peak around mid-day to early afternoon, the rate of increase in ground-level O3 is

69

highest in the morning hours and peaks around 9:00 am local time; similar variations are also

70

seen in ground-level O3 measurements (Figure 1, inset A). Also shown in Figure 1 are the model

71

estimates for rate of change of O3 due to the dominant atmospheric processes. Turbulent mixing

72

and chemical production contribute to the accumulation of O3 at the surface, while dry

73

deposition is the dominant sink. The net vertical O3 flux at the surface is dictated by the turbulent

74

mixing responding to the evolution of the planetary boundary layer and dry deposition at the

75

earth’s surface. This vertical flux is positive in the morning, and is the dominant contributor to

76

the net increasing O3 tendency during the morning period, indicative of the prominent

77

contribution of O3 trapped in the nocturnal residual layer aloft to daytime ground-level O3. The

78

magnitude of the net O3 tendency during the morning period varies spatially and is smaller at

79

locations influenced by fresh NO traffic emissions that titrate O3 as it is entrained from aloft

80

(Figure 1, inset B).

81 82

In its simplest form, ground-level O3 mixing ratios at any given location can thus be considered

83

to be comprised of an in-situ component determined from local source/sink balance, and a local

84

background component representing non-local O3 that was transported to that location. Intrinsic

85

to the concept of background concentrations is a space and time-scale so that it can be expected

86

to vary both spatially and temporally (6). In actuality the background O3 at any location is

87

comprised of O3 photo-chemically produced upwind as well as a natural component that

88

originated in the stratosphere. If background concentration for a geographic region is defined as

89

the concentration attributable to sources (natural or anthropogenic) outside of the region, then the

90

hemispheric O3 background influences the regional background which in turn influences a local

91

background. Thus, understanding the origin of ground-level O3 at a location, requires not only an

92

accurate description of its local sources and sinks but also an accurate attribution of the

93

background over the different space and time scales. Though air masses are advected to a 3



94

location through the course of a day, the significantly larger contribution of O3 in the air

95

entrained from above during the morning hours as the nocturnal inversion breaks up, can be

96

considered to be the starting point (or background) on which the in-situ component builds to

97

shape that day’s peak O3.

98 99

It can thus be argued that accurate characterization of this nocturnal aloft O3 is key for

100

quantifying the impacts of regional to continental scale transport of O3 on ground-level O3

101

mixing ratios. The recognition of such a relationship between ozone storage in the residual layer

102

overnight and evolution of near-surface O3 on the following day is not new. Neu et al. (7) using

103

a model based on transilient turbulence theory and initialized using vertical profiles from

104

tethered balloon soundings, estimated that 50-70% of the maximum near-surface O3 levels were

105

mixed down from the previous night’s residual layer. Using a simple one-dimensional model,

106

Zhang and Rao (8) showed that entrainment of O3 rich air from aloft contributed to ground-level

107

O3 buildup. Through complementary analysis of continuous vertical measurements at a tall tower

108

and results from a comprehensive atmospheric model, Aneja et al. (9) illustrated strong

109

correlations between nocturnal O3 in the residual layer aloft and ground-level maximum O3 the

110

following afternoon. In spite of this well-acknowledged vertical coupling of ground-level

111

daytime O3 with its nocturnal component in the aloft residual layer, and the need to quantify it on

112

a routine basis for areas not attaining the National Ambient Air Quality Standard (NAAQS),

113

little coordinated effort to date has been devoted towards measuring and characterizing regional

114

nocturnal O3 distributions aloft (10).

115 116

Changing emission patterns across the globe (11) are now necessitating an improved

117

characterization of long-range air pollution transport, which by nature occurs aloft. Based on

118

Figure 1, if we assume that the rate of change of ground-level O3 during the morning hours

119

(8:00-9:00 am) is indicative of the influence of the aloft O3 reservoir on ground-level O3, then

120

long-term trends in this morning-time O3 tendency could be suggestive of changes in this

121

reservoir influenced in turn by changes in local, regional, and hemispheric scale O3 distributions.

122

Figure 2 presents this inferred change in seasonal aloft-O3 entrainment rates over the 1990-2010

123

period using hourly measurements at monitors from the Clean Air Status Trends Network

124

(CASTNET; https://www.epa.gov/castnet). The directionality of these observed trends (Figures 4


Page 4 of 21

Page 5 of 21


125

2a and b) is generally consistent with previous analysis of trends of surface-level O3 distributions

126

(11-13), with decreases at most locations in the eastern U.S. indicative of reductions in

127

tropospheric O3 from emission reductions, but a contrasting increasing trend at many monitoring

128

sites in the western U.S. attributed to increasing O3 in air masses entering the western boundaries

129

of the U.S. Figure 2 also suggests that the spatial distribution of the O3 reservoir in the residual

130

layer has changed and will continue to change in response to changes in local and global

131

emissions and any climate-driven changes in stratosphere-troposphere exchange. Also shown in

132

Figures 2c and 2d are the corresponding changes in seasonal aloft O3 entrainment rates inferred

133

from multi-decadal simulations with the CMAQ model (14). Comparison of model and observed

134

trends in these entrainment rates show discrepancies both in magnitude and direction at several

135

locations especially in the western U.S. Availability of continuous aloft O3 measurements would

136

help better assess the ability of current models in representing the aloft O3 reservoir and

137

constrain modeled entrainment rates, thereby improving confidence in model estimates of spatial

138

and temporal variations in background O3 pollution.

139 140

As local precursor emissions (and consequently the in-situ chemical O3 production) change, the

141

relative importance of the local background to NAAQS attainment increases. Consequently,

142

measurements that can help assess changes in this component and enable robust evaluation of the

143

abilities of regional and global scale model in representing it are needed to build confidence in

144

the resulting inferences on source attribution. Satellite measurements from instruments such as

145

the Tropospheric Emission Spectrometer (TES) are now providing new perspectives on large

146

scale tropospheric O3 distributions (15) but the relatively coarse vertical resolution (~6-7 km) is

147

inadequate to quantify O3 in the nocturnal residual layer. Aircraft (16) and balloon based

148

measurements (17) have provided useful information on the vertical O3 structure from the

149

surface to the stratosphere but are resource intensive and impractical for establishing a long-term

150

record of O3 in the residual layer over regional scales. Ground based ozone lidars can provide

151

high temporal resolution of the vertical structure of O3 in the troposphere (18), but are expensive

152

to deploy continuously to cover a region such as the contiguous U.S.

153 154

A Cost-effective Aloft Air Pollution Monitoring Strategy for Now

5



155

Past efforts at instrumenting tall towers with O3 monitors have yielded valuable insights on the

156

diurnal variations in vertical O3 distributions. Resources associated with installation and

157

maintenance of long sampling lines have most likely been a deterrent to the long-term operation

158

of these limited efforts. Nevertheless, analysis of limited measurements from such platforms (9)

159

have provided compelling indication of the dependence of maximum ground-level O3 on

160

nocturnal residual layer amounts, and consequently the potential value of in-situ continuous aloft

161

measurement of the residual layer O3. Additional observations from emerging remote sensing

162

platforms, higher frequency ozonesonde launches, expansion of a ground-based lidar network,

163

and commercial aircrafts will provide valuable information on the large-scale processes dictating

164

baseline O3 (19) and in describing specific episodic cases, but may still not provide adequate

165

constraints to accurately represent and forecast the day to day variability in ground-level O3.

166

Continuous O3 measurements at an altitude of 400-500m above ground level will help directly

167

link the variability in ground level O3 at a location with the transported (or local background)

168

component since a direct measurement of the amount of O3 entrained to the surface will be

169

available. An appropriately spaced network of such measurements could then provide the vital

170

missing information on regional-scale nocturnal transport aloft that links surface measurements

171

with the larger scale free-troposphere column information, as well as a key measurement to

172

evaluate and improve the source attribution models.

173 174

Advances in air pollution monitoring technologies have recently yielded compact and low-power

175

instrumentation that accurately measure ambient ozone. Technological advances are also now

176

making available powerful microprocessors with low power draw and high on-board memory

177

storage, low power cellular modems, increased cellular communications coverage, and lower

178

cost solar panels to power such an instrument package. Together, these technology shifts are

179

already generating a significant change in the spatial and temporal availability of surface air

180

monitoring data (20). Compact air monitoring systems, with self-contained power and wireless

181

data communications, can allow for nearly autonomous monitoring in remote areas, such as tall

182

towers, and circumvent the need to run long power cables or sampling lines that hindered

183

previous efforts. Jiao et al. (21) demonstrated that a solar-powered UV absorbance-based

184

compact ozone monitor was able to provide accurate readings, despite frequent power cycling

185

and operation in an environment without controlled temperature or humidity. Similar 6


Page 6 of 21

Page 7 of 21


186

instrumentation deployed on the top of tall building in Hong Kong was also shown to provide

187

accurate data even after being subjected to frequent power cycling, subtropical conditions, and

188

several major weather events (22), suggesting the feasibility of an instrument package that could

189

provide reliable measurements over a long-term in an autonomous manner.

190 191

The proliferation of tall telecommunication and broadcasting masts and towers across the world

192

provide an existing set of platforms to possibly host such compact, autonomous, high accuracy

193

O3 (and in future other species) sensors. Figure 3 presents the summer-average e-folding distance

194

(23, 24) for night time O3 distributions (representative of spatial scales over which O3 mixing

195

ratios are correlated) at the ground and an altitude of 400-500m based on O3 distributions

196

simulated by CMAQ. Expectedly, the e-folding distances increase with altitude, with values of

197

200 km or greater at altitude of 400-500m AGL. This implies that a potential monitoring network

198

grid of 21 x 13 (east-west x north-south) with a separation of about 200 km could provide a

199

complete view of the spatial distribution of nocturnal O3 in the residual layer aloft over the

200

contiguous U.S. This possible continuous record of O3 measurements aloft would for the first

201

time provide a comprehensive view of nocturnal O3 transport aloft and furnish key missing

202

information to evolve our understanding of the processes and sources regulating background O3,

203

it’s spatial and temporal variability, and most importantly it’s modulation of ground-level O3 and

204

consequent impacts on human and ecological health.

205

Potential Benefits of Improved Aloft Characterization for Managing Air Quality

206

The expected impacts of accurately quantifying aloft O3 for improved surface-level daily

207

maximum 8-hour average O3 (DM8O3) characterization is demonstrated through detailed

208

calculations with the CMAQ-Adjoint model (25) (see supplementary material: CMAQ-Adjoint).

209

Note that since the adjoint calculations are computationally intensive, for demonstration

210

purposes we limit the domain size to the southeastern U.S. Figure 4 illustrates the backward

211

(adjoint) sensitivity of surface DM8O3 > 70ppb (current O3 NAAQS) to a possible 1 ppb change

212

in characterization of O3 at different altitudes. These simulations for the 1-14 July, 2011 period

213

are representative of typical current summertime conditions over the southeastern U.S. As is

214

evident from comparison of Figures 4a and 4b, DM8O3 is 3-4 times more sensitive (and over

215

larger spatial regions) to O3 distributions at 400-500m than that at the surface, implying that

216

measurements that help better characterize O3 aloft would yield far greater benefit than 7



217

enhancements to the surface networks. Removal of O3 by dry deposition and titration by fresh

218

NOx emissions (Figure 1) result in shorter O3 lifetime at the surface and thus the smaller

219

sensitivity. As local NOx emissions decline, the O3 transported nocturnally in the residual layer

220

to a region and entrained to the surface, becomes a more significant contributor to the DM8O3.

221

Figure 4c presents the cumulative DM8O3 sensitivity over the study domain and period to O3 at

222

different altitudes of the model troposphere and suggests that about 80% of the DM8O3

223

sensitivity could be captured by improved characterization of O3 distributions between 200-

224

1200m. Thus, future remote sensing instruments that have greatest sensitivity within this altitude

225

range would yield maximum benefits in improving surface-level O3 (and possibly other criteria

226

pollutants) air quality characterization. As compact, low power draw, and reliable NOx sensors

227

become available additional DM8O3 improvements can also be realized through improvements

228

in aloft NOx characterization (Figure S2).

229 230

The potential benefits of deploying an aloft air quality monitoring network based on the

231

relatively simple idea outlined above can be tested immediately. A relatively good density of tall

232

masts (height > 400m) used by broadcasting companies already exists in the U.S. (especially the

233

eastern U.S., see Figure 3) and other parts of the world. With an estimated cost of ~$5,000-6,000

234

in materials for a rugged monitoring system that includes a UV absorbance-based O3 instrument

235

(~$3500), and supporting equipment (~$1500) including a microprocessor, solar panel and

236

battery power, celluar modem for data communications, and weather-proof enclosure, the

237

deployment of a few hundred monitoring systems on existing masts could be achieved with a

238

relatively modest (and low compared to other conventional vertical profiling techniques) capital

239

investment. Public-private partnerships between Federal, State and private broadcasting

240

organizations could be pursued to help with the deployment and maintenance of this aloft

241

network. These continuous O3 measurements could be assimilated into comprehensive

242

atmospheric models in real-time to develop reliable next-day air quality forecast to guide public

243

health warning and promote restrictive actions (e.g., agricultural burn decisions). Data from the

244

network could also guide episodic deployment of lidars in networks such as TOLNet

245

(http://www-air.larc.nasa.gov/missions/TOLNet) to capture the detailed three-dimensional O3

246

structure and transport over large regions. Aloft measurements of air pollution in conjunction

247

with measurements of diurnal evolution of mixing heights (26) planned at several Photochemical 8


Page 8 of 21

Page 9 of 21


248

Assessment Monitoring Stations (PAMS) would help constrain and improve model

249

representation of vertical pollutant transport and consequently surface-level background

250

pollution. These data can also be used to guide constrained observing system simulation

251

experiments (OSSE) to further refine the aloft network configuration.

252 253

Compared to other factors influencing environmental quality (land, water, built environment,

254

sociodemographic), human exposure to air pollution exhibits the strongest association with all-

255

cause and cause-specific (heart disease, cancer) mortality (27). While significant improvements

256

in air quality have occurred in the U.S., studies suggest that still more than half of the population

257

is exposed to unhealthy levels of particle and ozone pollution (28), point to adverse health effects

258

from to exposure to O3 and PM2.5 concentrations below current NAAQS (29), and recommend

259

air quality standards lower than the current (30). Implementation of the current NAAQS and

260

future revisions would greatly benefit from improvements in our understanding of background

261

pollution levels. Measurements of air pollution in the residual layer will provide key missing

262

information on regional air pollution transport that is prominent at night, regulates the local

263

background pollution, influences subsequent day peak values, and helps connect the variability

264

in ground-level pollution with larger scale pollution features that descend from the free

265

troposphere. Technological advances in air sensor development coupled with the availability of a

266

platform in the form of tall telecommunications and broadcasting masts, provide a timely and

267

unique opportunity for deployment of a cost effective aloft air quality monitoring network whose

268

value can be immediately tested. Continuous measurements of air pollution aloft from such a

269

network hold potential to provide vital information for quantifying and accounting for

270

background air pollution when designing and implementing air quality standards.

271 272

Disclaimer: The views expressed in this paper are those of the authors and do not necessarily represent the view or

273

policies of the U.S. Environmental Protection Agency.

274

Acknowledgements: We thank Thomas Pierce and Nealson Watkins for constructive comments and suggestions on

275

initial versions of this manuscript. We gratefully acknowledge pioneering efforts by the North Carolina Department

276

of Environmental Quality in instrumenting and making available data from a tall tower, analysis and use in

277

modeling of which served as motivation for this contribution.

278

Notes: The authors declare no competing financial interest.

9



279

References

280

1.

281 282

[accessed 3 February 2017]. 2.

3.

285 286

United States Environmental Protection Agency, Our Nation’s Air: Status and trends through 2015, https://gispub.epa.gov/air/trendsreport/2016/.

283 284

United States Environmental Protection Agency, Air Quality System, https://www.epa.gov/aqs, 2017

Chameides, W.L.; Saylor, R.D.; Cowling, E.B. Ozone pollution in the rural United States and the new NAAQS. Science 1997, 276, 916, DOI: 10.1126/science.276.5314.916.

4.

Trainer, M.; Williams, E.J.; Parrish, D.D.; Buhr, M.P.; Allwine, E.J.; Westberg, H.H.; Fehsenfeld, F.C.; Liu,

287

S.C. Models and observations of the impact of natural hydrocarbons on rural ozone. Nature 1987, 329, 705-

288

707.

289

5.

Appel, K.W.; Napelenok, S.; Foley, K.M.; Pye, H.O.T.; Hogrefe, C.; Luecken, D.; Bash, J.O.; Roselle, S.J.;

290

Pleim, J.E.; Foroutan, H.; Hutzell, W.; Pouliot, G.; Sarwar, G.; Fahey, K.; Gantt, B.; Gilliam, R.C.; Kang, D.;

291

Mathur, R.; Schwede, D.; Spero, T.; Wong, D.C.; Young, J. Overview and evaluation of the Community

292

Multiscale Air Quality (CMAQ) model version 5.1. Geosci. Model Dev. 2017, 10, 1703-1732,

293

doi:10.5194/gmd-10-1703-2017.

294

6.

Mathur, R.; Xing, J.; Gilliam, R.; Sarwar, G.; Hogrefe, C.; Pleim, J.; Pouliot, G.; Roselle, S.; Spero, T.L.;

295

Wong, D.C.; Young, J. Extending the Community Multiscale Air Quality (CMAQ) modeling system to

296

hemispheric scales: overview of process considerations and initial applications. Atmos. Chem. Phys. 2017, 17,

297

12449-12474, https://doi.org/10.5194/acp-17-12449-2017.

298

7.

Neu, U.; Künzle, T.; Wanner, H. On the relation between ozone storage in the residual layer and daily

299

variation in the near-surface ozone concentration – a case study. Boundary-Layer Meteorol., 1994, 69, 221-

300

247.

301

8.

302 303

concentrations. J Appl Meteorol., 1999, 38, 1674–1691. 9.

304 305

Aneja, V.P.; Mathur, R.; Arya, S.P.; Li, Y.; Murray, G.C.; Manuszak, T.L. Coupling the vertical distribution of ozone in the atmospheric boundary layer. Environ. Sci. & Technol., 2000, 34, 2324-2329.

10.

National Research Council, Observing weather and climate from the ground up: A nationwide network of networks. Washington, DC, The National Academies Press, https://doi.org/10.17226/12540, 2009.

306 307

Zhang J.; Rao, S.T.; The role of vertical mixing in the temporal evolution of ground-level ozone

11.

Xing, J.; Mathur, R.; Pleim, J.; Hogrefe, C.; Gan, C.; Wong, D.C.; Wei, C.; Gilliam, R.; Pouliot, G.

308

Observations and modeling of air quality trends over 1990-2010 across the northern hemisphere: China, the

309

United States and Europe. Atmos. Chem & Phys. 2015, 15, 2227-2914, doi:10.5194/acp-15-2723-2015.

310

12.

Cooper, O.R.; Gao, R-S.; Tarasick, D.; Leblanc, T.; Sweeney, C. Long-term ozone trends at rural ozone

311

monitoring sites across the United States, 1990-2010. J. Geophys. Res., 2012,117, D22307,

312

doi:10.1029/2012JD018261.

313

13.

Lin, M.; Horowitz, L.W.; Payton, R.; Fiore, A.M.; Tonnesen, G. US surface ozone trends and extremes from

314

1980 to 2014: quantifying the roles of rising Asian emissions, domestic controls, wildfires, and climate.

315

Atmos. Chem. Phys. 2017, 17, 2943-2970, doi:10.5194/acp-17-2943-2017. 10


Page 10 of 21

Page 11 of 21

316


14.

Gan, C.-M.; Pleim, J.; Mathur, R; Hogrefe, C.; Long, C.N.; Xing, J.; Wong, D.; Gilliam, R.; Wei, C.

317

Assessment of long-term WRF–CMAQ simulations for understanding direct aerosol effects on radiation

318

"brightening" in the United States. Atmos. Chem. Phys. 2015, 15, 12193-12209, doi:10.5194/acp-15-12193-

319

2015.

320

15.

Verstraeten, W. W.; Boersma, K. F.; Zörner, J.; Allaart, M. A. F.; Bowman, K. W.; Worden, J. R. Validation

321

of six years of TES tropospheric ozone retrievals with ozonesonde measurements: implications for spatial

322

patterns and temporal stability in the bias. Atmos. Meas. Tech. 2013, 6, 1413–1423, doi:10.5194/amt-6- 1413-

323

2013.

324

16.

Petetin, H.; Thourer, V.; Athier, G.; Blot, R.; Boulanger, D.; Cousin, J.M.; Gaudel, A.; Nedelec, P.; Cooper, O.

325

Diurnal cycle of ozone throughout the troposphere over Frankfurt as measured by MOZAIC-IAGOS

326

commercial aircraft. Elem Sci Anth. 2016, 4,129, doi: http://doi.org/10.12952/journal.elementa.000129.

327

17.

Thompson, A. M., et al. Intercontinental Chemical Transport Experiment Ozonesonde Network Study (IONS)

328

2004: 2. Tropospheric ozone budgets and variability over northeastern North America. J. Geophys. Res. 2007,

329

112, D12S13, doi:10.1029/2006JD007670.

330

18.

Kuang, S.; Newchurch, M.J.; Burris, J.; Wang, L.; Buckley, P. I.; Johnson, S.; Knupp, K.; Huang, G.; Phillips,

331

D.; Cantrell, W. Nocturnal ozone enhancement in the lower troposphere observed by lidar. Atmos. Environ.

332

2011, 45, 33, 6078.

333

19.

334 335

Cooper, O.R.; Langford, A.O.; Parrish, D.D.; Fahey, D.W. Challenges of a lowered U.S. ozone standard. Science 2015, 348, 1096-1097, doi:10.1126/science.aaa5748.

20.

Snyder, E.G.; Watkins, T.H.; Solomon, P.A.; Thoma, E. D.; Williams, R.W., Hagler, G.S.; Shelow, D.; Hinds,

336

D.A.; Kilaru, V.J.; Preuss, P.W. The changing paradigm of air pollution monitoring. Environ. Sci. Technol.

337

2013, 47(20), 11369-11377.

338

21.

Jiao, W.; Hagler, G. S.; Williams, R. W.; Sharpe, R. N.; Weinstock, L.; Rice, J. Field assessment of the

339

Village Green project: an autonomous community air quality monitoring system. Environ. Sci. Technol., 2015,

340

49 (10), 6085–6092.

341

22.

Hagler, G.; Ning, Z.; Louie, P.; Sharpe, R.; Lam, N.; Williams, R. Performance assessment of a solar

342

powered air quality and weather station placed on a school rooftop in Hong Kong. Proc. Of the Air Quality

343

Measurement Methods and Technology Conference, Air & Waste Management Association, Chapel Hill, NC,

344

March 15-17, 2016.

345

23.

346 347

Rao, S.T.; Zalewsky, E.; Zurbenko, I.G. Determining Temporal and Spatial Variations in Ozone Air Quality. J. Air & Waste Manage. Assoc. 1995, 45, 57-61, DOI:10.1080/10473289.1995.10467342.

24.

Godowitch, J. M.; Hogrefe, C.; Rao, S. T. Diagnostic analyses of a regional air quality model: Changes in

348

modeled processes affecting ozone and chemical-transport indicators from NOx point source emission

349

reductions. J. Geophys. Res. 2008, 113, D19303, doi:10.1029/2007JD009537.

350 351

25.

Hakami, A.; Henze, D.K.; Seinfeld, J.H.; Singh, K.; Sandu, A.; Kim, S.; Byun, D.; Li, Q. The adjoint of CMAQ. 2007, Environ. Sci. & Technol., 41, 7807-7817.

11



352

26.

Knepp, T. N.; Szykman, J. J.; Long, R.; Duvall, R.M.; Krug, J.; Beaver, M.; Cavender, K.; Kronmiller, K.;

353

Wheeler, M.; Delgado, R.; Hoff, R.; Berkoff, T.; Olson, E.; Clark, R.; Wolfe, D.; Van Gilst, D.; Neil, D.

354

Assessment of mixed-layer height estimation from single-wavelength ceilometer profiles. Atmos. Meas. Tech.,

355

2017, 10, 3963-3983, https://doi.org/10.5194/amt-10-3963-2017.

356

27.

Jian, Y.; Messer, L.C.; Jagai, J.S.; Rappazo, K.M.; Gray, C.L.; Grabich, S.C.; Lobdell, D.T. Associations

357

between environmental quality and mortality in the contiguous United States, 2000-2005. Environ. Health

358

Perspect. 2017, 125, 355-362, http://dx.doi.org/10.1289/EHP119.

359

28.

air/state-of-the-air/sota-2016-full.pdf [accessed 28 February 2017].

360 361

American Lung Association (ALA), State of the air 2016, http://www.lung.org/assets/documents/healthy-

29.

Di, Q.; Wang, Y.; Zanobetti, A.; Wang, Y.; Koutrakis, P.; Choirat, C.; Dominici, F.; Schwartz, J. D. Air

362

pollution and mortality in the Medicare population. N. Engl. J. Med. 2017, 3376, 2513-2522,

363

DOI:10.1056/NEJMoa1702747.

364

30.

Cromar, K.R.; Gladson, L.A.; Perlmutt, L.D.; Ghazipura, M.; Ewart, G.W. Estimated excess morbidity and

365

mortality caused by air pollution above American Thoracic Society–recommended standards, 2011–2013. Ann

366

Am Thorac Soc. 2016, 13, 1195–1201, doi: 10.1513/AnnalsATS.201602-103AR.

367

12


Page 12 of 21

Page 13 of 21

368 369


Table of Contents Art

370 371

13



372

373 374

Figure 1: Diurnal variation of CMAQ estimated O3 process tendencies regulating variations in ground-level O3.

375

Values shown are median values (of July 2010 monthly average diurnal variations) across all Continental U.S. land

376

cells. CHEM represents the net rate of change due to atmospheric chemical production and loss, DDEP is the rate of

377

change due to dry deposition to the earth’s surface, Mixing represents the rate of change due to turbulent mixing,

378

and Net is the sum of tendencies from all modeled processes regulating O3. Note that the values for the Mixing and

379

DDEP terms are shown on the y-axis on the right. Inset A: Comparison of model and observed average (for July

380

2010 and across all AQS sites in the Continental U.S) diurnal surface O3 tendency. Inset B: Distribution of model

381

O3 tendency across all land cell across the Continental U.S. The lower percentile curve represents locations

382

influenced by fresh NOx emissions as is evident by the large negative tendency associated with the evening traffic

383

rush hours. Traffic related NOx emissions in the morning rush hours result in titration of O3 entrained from aloft and

384

the consequent smaller magnitude of O3 tendency at these locations.

385

14


Page 14 of 21

Page 15 of 21


386

387 388

Figure 2: Inferred changes in seasonal entrainment rates of aloft O3 to the surface over the past two decades. Trends

389

in observed average 8-9 am ground-level O3 tendency in units of ppb hr-1 yr-1 during 1990-2010 at CASTNET

390

monitoring locations are shown for (a) Spring (March-May) and (b) Summer (June-August). (c and d) same as (a

391

and b) but based on model simulations. The hourly tendency at each monitor is estimated as the difference in

392

measured O3 at an hour and at the previous hour. Seasonal-averages of the estimated tendencies at 8:00am and 9:00

393

am (times when the net O3 tendency peak) are computed for each year and the trends for the 21-year period are

394

shown. Squares indicate high-elevation sites (altitude > 1000m ASL), while circles represent sites with altitude
400 m) based on the list of tallest structures between 400-500m height available at:

406

https://en.wikipedia.org/wiki/ List_of_tallest_structures_%E2%80%93_400_to_500_metres.

407

folding distance, for a given model altitude, we first calculate the correlation between the synoptic component of the

408

0700 UTC O3 time series at a grid point with that at each of the other grid points. For each location we then examine

409

the relationship between this correlation and the distance between the points (Note that the relationship between

410

correlation and distance generally follows an exponential decline). The e-folding distance at a location is the

411

distance at which correlations in O3 at this location with that at any remote location drops below 0.37. The e-folding

412

distance is thus an indicator of spatial correlation in the O3 field. 16


To estimate the e-

Page 17 of 21


413

414 415

Figure 4: Adjoint sensitivity of surface-level daily maximum 8-hour average O3 >70ppb (DM8O3) across the

416

southeastern U.S. during 1-14 July, 2011. Estimated potential change in simulated DM8O3 (in ppb) due to (a) a 1

417

ppb change in representation of O3 at the surface (0-20m) and (b) a 1 ppb change in representation of O3 at 400-

418

500m altitude. (c) Sensitivity of DM8O3 across the southeastern U.S. to O3 at different altitudes expressed as percent

419

of total sensitivity integrated through the model depth (in blue). Also shown in red is the cumulative sensitivity of

420

DM8O3 as function of height. Measurements characterizing O3 distribution at 200-1200 m altitude (shaded region)

421

would yield greatest benefit in improving DM8O3 predictions via data assimilation.

17



Supplementary Materials for

1

3

A Call for an Aloft Air Quality Monitoring Network: Need and Feasibility

4 5

Rohit Mathur1, Christian Hogrefe1, Amir Hakami2, Shunliu Zhao2, James Szykman1, Gayle Hagler1

2

6 7 8

1

9

Correspondence to: Rohit Mathur ([email protected])

National Exposure Research Laboratory, Office of Research and Development, U.S. Environmental Protection Agency, RTP, NC, USA 2 Department of Civil and Environmental Engineering, Carleton University, Ottawa, Canada

10

Community Multiscale Air Quality (CMAQ) Model

11

CMAQ is a comprehensive air pollution modeling system that simulates air quality on scales

12

ranging from urban to hemispheric. The model includes detailed representation of processes

13

(emissions, atmospheric chemistry, transport, removal) regulating the fate of airborne pollutants.

14

Information on the dynamical state of the atmosphere is typically derived from simulations of the

15

Weather Research and Forecast (WRF) meteorological model. The model code, documentation,

16

and extensive peer-reviewed literature on model processes, algorithms, applications, and

17

evaluation are publicly available at: https://www.epa.gov/cmaq. The modeling system is widely

18

used for air pollution research and regulatory applications across the world.

19 20

CMAQ-Adjoint

21

A comparative analysis of the influence of ozone concentrations at the surface and aloft on

22

DM8O3 is conducted using an adjoint version of the CMAQ model (23). An adjoint model solves

23

an auxiliary set of sensitivity equations that mimic and are linked to the model’s governing

24

equations. Unlike the base CMAQ model, the adjoint equations are integrated backward in time,

25

and in doing so they trace influences on a scalar concentration function (often referred to as the

26

adjoint cost or objective function) back to concentrations and sources at preceding model time

27

steps. Backward integration of adjoint equations follows a forward simulation of the CMAQ

28

model during which the required concentrations and model variables at all time steps are saved

29

for use in the adjoint simulation. The advantage of the adjoint sensitivities calculated in this

1


Page 18 of 21

Page 19 of 21


30

manner is in their ability to provide time- and location-specific measures of influence on the

31

objective function.

32 33

To quantify how surface or aloft concentrations of ozone affect surface DM8O3, the adjoint

34

objective function is defined as the cumulative DM8O3 over a region – depicted by the inner box

35

within the computational domain for these calculations (depicted in Figure S1), for days on

36

which the DM8O3 value exceeds a 70 ppb threshold. A 20-cell margin from the boundary (the

37

inner box in Figure S1) is considered as the target area for this analysis to avoid influences from

38

the lateral boundaries and to better characterize in-domain influences. The calculation of adjoint

39

sensitivities are initially forced at the times and locations where the objective function

40

materializes, in this case hours and cells that comprise MD8O3 exceedances over the target area.

41

The adjoint “forcing” at each location and time then corresponds to the instantaneous sensitivity

42

of the objective function to ozone concentration at that location and time. This “forcing” can be

43

considered analogous to emissions for the forward model, since it acts as the “source” of

44

influences that drive the adjoint simulations. These influences are then integrated backward in

45

time through the adjoint system of equations to preceding times and for various modeled

46

processes (transport, chemistry, deposition, etc) at each time step.

47 48

With this definition of the adjoint objective function, the values at each location in Figure 4

49

should be interpreted as the total DM8O3 contributions in ppb (for MD8O3 > 70 ppb) in the

50

target area for any ppb ozone at that location. As such, these values can be taken as a measure of

51

information that ozone observations at various altitudes and locations may carry with regards to

52

characterization of surface-level ozone. Note that calculated adjoint contributions to surface

53

MD8O3 concentrations extend beyond the inner box in Figure S1, as concentrations in these

54

upwind locations influence surface concentrations over the target area. Similar adjoint analysis

55

performed to estimate the added value of potential aloft NOx measurements (Figure S2) also

56

suggest significantly larger contributions from aloft NOx levels to regulating DM8O3, than

57

surface values. These results suggest that aloft NOx, either transported from elevated sources, or

58

recycled from reservoir species is also influential in regulating surface DM8O3 especially in

59

regions away from major NOx emissions. The lower DM8O3 sensitivity to surface NOx

2



Page 20 of 21

60

concentrations arises form O3 destruction from reactions with NO in vicinity of the NOx

61

emission sources.

62 63

CMAQ-Adjoint was used to simulate air quality and the backward sensitivities described above,

64

over a domain covering the southeastern U.S. (depicted in Figures 4 and S1) discretized with a

65

horizontal grid of 12km resolution, for the July 1-14, 2011 period. The adjoint version of CMAQ

66

used here includes all gas-phase chemistry, transport, removal processes relevant for

67

representing

68

(https://www.epa.gov/cmaq/cmaq-models-0).

69

Carbon Bond 5 (CB05) chemical mechanism which includes 187 chemical reactions amongst 72

70

gaseous species (5). The vertical extent from the surface to 50hPa was discretized with 35 layers

71

of variable thickness. Meteorological fields were derived from simulations with the Weather

72

Research and Forecast (WRF) model, while anthropogenic emissions of various species were

73

based on the 2011 National Emissions Inventory (NEI). 3-D chemical initial fields and boundary

74

conditions were derived from previously conducted CMAQ simulations over a larger continental

75

domain which have been extensively evaluated against a variety of surface and aloft

76

measurements (5). This CMAQ-adjoint model configuration, domain size and discretization, and

77

simulation duration enable an appropriate attribution of the backward sensitivities of surface-

78

level DM8O3 to O3 and precursor levels at various altitudes

tropospheric

O3.

The

adjoint

is

based

on

CMAQv4.5

Gas-phase chemistry is represented using the

79

3


Page 21 of 21


80

Figure S1: The computational domain for the CMAQ-adjoint simulations. The inner box is the target area over

81

which the adjoint objective function is defined.

82 83

Figure S2: Adjoint sensitivity of surface-level daily maximum 8-hour average O3 >70ppb (DM8O3) to NO mixing

84

ratios at different altitudes across the southeastern U.S. during 1-14 July, 2011. Estimated change in simulated

85

DM8O3 (in ppb) due to (a) a 1 ppb change in representation of NO at the surface (0-20m) and (b) a 1 ppb change in

86

representation of NO at 400-500m altitude. Improvements in characterization of NOx mixing ratios aloft would also

87

yield improved predictive capabilities for DM8O3. NOx aloft, either transported from elevated sources or recycled

88

from reservoir species (such as organic nitrates) is also influential in regulating surface DM8O3 in regions away

89

from major surface-level NOx sources. Thus as measurement accuracy of compact, low-cost NOx sensor is

90

improved, the proposed aloft network could be enhanced to also provide aloft NOx measurements to improve surface

91

O3 predictions.

92

4


A Call for an Aloft Air Quality Monitoring Network: Need, Feasibility

Recommend Documents