Using Cellular Communication Networks To Detect ... - ACS Publications

Aug 4, 2016 - Using Cellular Communication Networks To Detect Air Pollution. Noam David .... a few other works have pointed out the technology's poten...
0 downloads 4 Views 2MB Size
Subscriber access provided by Northern Illinois University

Article

Using Cellular Communication Networks to Detect Air Pollution Noam David, and H. Oliver Gao Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.6b00681 • Publication Date (Web): 04 Aug 2016 Downloaded from http://pubs.acs.org on August 5, 2016

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 free 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 accessible to all readers and 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.

Environmental Science & Technology 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 34

Environmental Science & Technology

1

Using Cellular Communication Networks to

2

Detect Air Pollution

3

Noam David† and H. Oliver Gao†‡*

4 5

6 7



The School of Civil and Environmental Engineering, Cornell University, Ithaca, NY,

14853, United States. ‡

One-Hundred Talents Scholar, School of Hydraulics, Changsha University of

Science and Technology, Hunan Province, China

8 9 10 11 12 13 14 15 16 17 18 19 20

21 1 ACS Paragon Plus Environment

Environmental Science & Technology

22

ABSTRACT

23

Accurate real time monitoring of atmospheric conditions at ground level is vital for

24

hazard warning, meteorological forecasting and various environmental applications

25

required for public health and safety. However, conventional monitoring facilities are

26

costly and often insufficient, for example, since they are not representative of the

27

larger space, and are not deployed densely enough in the field.

28

There have been numerous scientific works showing the ability of commercial

29

microwave links that comprise the data transmission infrastructure in cellular

30

communication networks to monitor hydrometeors as a potential complementary

31

solution. However, despite the large volume of research carried out in this emerging

32

field during the past decade, no study has shown the ability of the system to provide

33

critical information regarding air quality.

34

Here we reveal the potential for identifying atmospheric conditions prone to air

35

pollution by detecting temperature inversions that trap pollutants at ground level. The

36

technique is based on utilizing standard signal measurements from an existing cellular

37

network during routine operation.

38 39 40 41 42 43 44 45 46

2 ACS Paragon Plus Environment

Page 2 of 34

Page 3 of 34

Environmental Science & Technology

47

INTRODUCTION

48

A wide spectrum of severe and chronic health impacts are associated with air

49

pollution. The US Environmental Protection Agency (EPA), for instance, is

50

responsible for setting the national ambient air quality standards for six pollutants

51

commonly found in the air: particulate matter (PM), ground-level ozone, carbon

52

monoxide, nitrogen oxides - NOx (a generic term for NO and NO2), sulfur oxides, and

53

lead. Of these contaminators, also known as “criteria pollutants”, PM that is 10

54

micrometers in diameter or smaller (PM10), and particularly that of 2.5 micrometers

55

and below (PM2.5) is of main concern, as these can generally pass through the human

56

respiratory pathways and enter the lungs and bloodstream. Once inhaled, they can

57

cause acute health damage, including lung cancer and other cardiopulmonary

58

mortality.1 According to the World Health Organization (WHO), in the year of 2012,

59

for example, 3.7 million premature deaths worldwide were attributable to ambient air

60

pollution.2

61

Notably, temperature inversion (TINV) within the boundary layer has been found to

62

be a major contributor to events of poor air quality worldwide.3-11 In typical

63

conditions in the lower atmosphere, temperature decreases as altitude increases.

64

TINV, on the other hand, is a case where the temperature in a certain layer increases

65

with an increase in altitude. 11-14 The resulting stagnation inhibits vertical motion in

66

the atmosphere, creating ideal conditions for increased contaminant concentrations

67

when formed close to the surface, since air pollutants released into the atmosphere's

68

lowest layer are trapped near the ground beneath a warmer air stratum. As reviewed

69

by Bailey et al.8, in European cities, for example, both elevated and ground-based

70

TINVs have been found to coincide with events of excessive PM concentration.6

71

Anthropogenic emission of particulates, or the suspension of dust are certainly the 3 ACS Paragon Plus Environment

Environmental Science & Technology

72

root causes of these periods of poor air quality. Still, regardless of whether the high

73

pollution event was in a maritime or continental climate, over complex terrain or

74

elsewhere, a temperature inversion was also present, intensifying the conditions of

75

poor air quality, and preventing dissipation of polluted air. An additional example has

76

often been cited from Logan, Utah, USA, where during the winter of 2004, a series of

77

elevated PM events occurred, coinciding with strong winter TINVs.15 Moreover,

78

TINVs have been linked with respiratory16 and cardiovascular 10, 17 diseases. Thus, the

79

trapping of pollutants, and their precursors at ground level by inversions, has crucial

80

public health implications.

81

Beyond the adverse effects caused by TINVs intensifying air pollution and human

82

health hazards, a combination of an inversion and high Relative Humidity (RH>95%)

83

trapped at the bottom of the stable layer creates favorable conditions for the creation

84

of fog18-20, which is in itself a natural hazard with potentially destructive

85

repercussions.21-23

86

Existing tools for inversion monitoring are based on limited in-situ measurements or

87

remote sensing observations. Instrumented towers, which have temperature sensors

88

installed on them at different altitudes and radiosondes, are the common instruments

89

for direct in-situ monitoring. While these tools provide precise measurements, their

90

spatial coverage and representativeness is poor due to the specific point

91

measurements. Remote sensing systems24-25 are limited in many conditions. Lidar

92

ceilometers, for example, have difficulty in detecting the inversion layer during fog.26

93

Aside from the specific technical limitations of each method reviewed above, the

94

instrument costs are high, and they require continuous maintenance and operation. 27

95

Consequently, deployment of these tools is insufficient in terms of both coverage and

96

resolution. Satellite systems produce measurements at a better spatial resolution,

4 ACS Paragon Plus Environment

Page 4 of 34

Page 5 of 34

Environmental Science & Technology

97

including from various surfaces, such as oceans and the Arctic, where it is difficult to

98

measure.28 However, while the vertical resolution of these systems is good at higher

99

altitudes, remote sensing from satellite systems is lacking for measurement of TINVs

100

that are low to the ground level (e.g. McGrath‐Spangler and Denning 29), which

101

happen to be the inversions that are critical in the creation of the conditions for

102

extreme air pollution.

103

In this study we show the potential for detecting low lying inversions using received

104

signal level (RSL) measurements taken by cellular communication networks. Weather

105

phenomena and variations in atmospheric conditions interfere with the

106

electromagnetic waves transmitted by commercial microwave links (MLs) that

107

comprise the infrastructure for data transport between cellular base stations. Thus,

108

cellular network infrastructure provides, in essence, an already existing atmospheric

109

monitoring facility, as was first demonstrated for rainfall measurements30,31 , although

110

these networks were not originally intended for this purpose. MLs are deployed over a

111

wide geographic expanse and close to ground level, typically at elevations of tens of

112

meters above the surface. These links operate at frequencies of tens of GHz and are

113

widely deployed around the world, including in many developing countries (e.g.

114

Doumounia et al.41). Thus, this technology can provide a low cost monitoring solution

115

in regions where conventional measurement tools are limited or non-existent. So far,

116

most of the research projects on this topic have focused dominantly on the ability of

117

this system to monitor and map rainfall since, comparing to other hydrometeors,

118

raindrops induce predominant transmission losses to the microwave channel.30-42 Only

119

few other works have pointed out the technology's potential for monitoring additional

120

hydrometeors, including water vapor,43, 44 fog45, 46 and dew47. The common factor of

121

all research carried out to this point is that they rely on the signal loss measured by the

5 ACS Paragon Plus Environment

Environmental Science & Technology

Page 6 of 34

122

wireless network to detect and estimate the intensity of the weather phenomenon

123

observed.

124

Notably, low lying inversions induce a unique effect on the microwave beam. In cases

125

of inversion a layering occurs that manifests in a change of the profile of the

126

atmospheric refraction index as elevation increases. As a result, the atmosphere may

127

operate as a waveguide, due to the channels or ducts that form in it. A scenario called

128

multipath may also occur, in which, several waves arrive at the receiver due to

129

reflections from the layers with the different refraction indices.48 These phenomena

130

may lead to an amplification of the received signal when compared to the standard

131

case. Consequently, detection of this effect using commercial microwave networks

132

can be utilized to identify low lying inversions that are prone to hazardous conditions

133

for public health and safety. We demonstrate the novel concept by analyzing extreme

134

air pollution events using standard measurements of existing commercial microwave

135

links. Additional proof of feasibility for detection of low lying inversions using the

136

proposed technique is demonstrated during a heavy fog event.

137 138

METHOD

139

Atmospheric Refractivity

140

Snell's law, given by Equation (1), describes the angle of refraction of a wave passing

141

through a boundary between two different isotropic media:

n1 sin 1  n2 sin 2

142 143

Where:

144

n1 – The refractive index of the first medium.

145

n2 – The refractive index of the second medium.

146

ϕ1 – The angle of incidence.

6 ACS Paragon Plus Environment

(1)

Page 7 of 34

Environmental Science & Technology

147

ϕ2 – The angle of refraction.

148

Figure 1(A) shows the path of a wave according to this law. Note that the larger the

149

refractive index of one medium is than the other, the greater the angle of refraction is

150

in relation to the angle of incidence. Part of the beam is returned at an angle equal to

151

the angle of incidence (the red trajectory).

152

If the wave enters the boundary at an angle greater than the critical angle, total

153

internal reflection will occur (the blue trajectory). The magnitude of the critical angle

154

can be used as a measure of the intensity of guided propagation, i.e. of how much

155

radiant energy is trapped within the duct.49

156

The atmospheric refractive index can be expressed by the following relation:

n  1  N 106

157

(N-Units)

(2)

158

Where N is the refractivity which can be computed according to the following

159

formula:50, 51

N  Ndry  Nwet  77 .6

160 161

e P + 3.732 105 2 T T

(N-units)

(3)

where:

162

P-

total atmospheric pressure (hPa).

163

T-

absolute temperature (K).

164

e-

partial water vapour pressure (hPa).

165

For frequencies up to 100 GHz, the expression above has an error of less than 0.5%.51

166

Let the gradient of a given parameter X with altitude H, be denoted as- X'= dN/dH.

167

Derivating Eq. (3) and substituting typical values of 1013 millibar, 15 C, and 15

168

millibar for the pressure, temperature and water vapor pressure, respectively, we get

169

the following expression describing the dependency of refractivity N on the

170

atmospheric parameters:52

7 ACS Paragon Plus Environment

Environmental Science & Technology

N '  0.27 P'  4.5e'  1.18T '

171

Page 8 of 34

(N-units/km)

(4)

172

Note that temperature and moisture are the dominant factors functioning jointly to

173

generate the meaningful variations in refraction. Typically, the inversion is

174

accompanied by sufficient humidity gradient. The variations of refractivity N due to

175

pressure fluctuations are negligible since wind gusts rebuild equilibrium.

176

It can be shown53 that the gradient of N with altitude (dN/dH) is relatively equal to the

177

radius of curvature of the microwave ray -

dN N  dH H

:

178

 ML  

179

Atmospheric Microwave Propagation

180

The propagation of a wave in lower troposphere can be categorized into four cases

181

according to the changes in refractive index with altitude:52, 54

(N- units/km)

(5)

182

I

N  -76 to 0 (N-Units/km): Normal H

(6)

183

II

N  0 (N-Units/km): Sub refraction H

(7)

184

III -157 

185

IV

N  - 76 (N-Units/km): Super refraction H

N  - 157 (N-Units/km): Ducting H

(8) (9)

186

Figure 1(B) schematically describes the beam propagation in the medium under each

187

of these four cases.

188

Normal Propagation – In standard conditions in the lower troposphere the typical

189

gradient of N is around −40 N-units/km,53 and may vary somewhat51 from area to

190

area, and from season to season. As a result of the moderate gradient of N with

191

altitude in these conditions, a wave propagating through the medium is refracted

192

downwards slightly, according to Snell's law, as shown by the curve signified as 8 ACS Paragon Plus Environment

Page 9 of 34

Environmental Science & Technology

193

"Normal" in Fig. 1(B) that simulates the path of microwave radiation between the

194

transmitter and receiver in standard atmospheric conditions. Commercial MLs are

195

designed to take this effect into consideration.

196

Cases where the gradient of N is different from the standard case (cases II through IV)

197

cause anomalous propagation of waves in the atmosphere.

198

Sub Refraction- A condition where the wave is refracted less (sub-normally) with

199

respect to the standard condition, and in a manner that can often cause the upward

200

bending of the wave causing it to move further away from the surface with distance

201

from the transmitter.

202

The generating factor of the following phenomena - super refraction and ducting, is

203

typically a temperature inversion.49, 54

204

Super Refraction –In this situation, the gradient of the refractive index with altitude

205

in the layer is higher than the standard case, and will cause increased bending of the

206

microwave radiation towards the Earth. As a result, a phenomenon called multipath

207

may occur – the arrival of two or more waves at the receiver.48 This causes

208

constructive and destructive interference which results in relative amplification, or

209

attenuation of the received signal, respectively. A boundary case of super refraction is

210

the case known as ducting.

211

Ducting – Earth's radius is approximately 6,400 km and therefore Earth's curvature

212

is:  Earth  157 (km-1). Thus, by Eqs. (2) and (4) in a case where dN/dH = −157 (N-

213

Units/km), the radio waves will travel parallel to Earth's curvature. When dN/dH
 c ). The trajectory in red describes the refraction angle of the beam (Φ2), and the

245 246 247

angle of reflection (which is equal to the angle of incidence). (B) Beam bending during standard, sub-refractive and super refractive conditions. Under ducting conditions, beam trapping occurs.

248

RESULTS

249

Using central Israel as an illustrative case, we investigated the potential of

250

commercial microwave systems for detecting near-ground inversions that formed a

251

basis for the development of extreme air pollution events and fog. The ML

252

measurements were compared to lidar ceilometer measurements that detected the

253

mixing layer height coupled to the altitude of the inversion.9 This height represented

254

the upper bound of a ground inversion, or the elevation of the inversion base in cases

255

where it occurred at higher altitudes (we also used the radiosonde measurements to

256

assist in this determination).

257

Additionally, comparisons were made to measurements of air pollutants (PM-10 and

258

NOx) that were found in prior research to have increased concentrations during near-

259

ground inversions.6,9,10,58 In their standard configuration, the microwave networks

Typical atmospheric stratification in cases of inversion. The refraction coefficient of the middle layer where the wave propagates (n2) is higher than those of the two layers confining it (n1, n3). According the Snell's law (Eq. 1), when the angle of incidence (Φ1) is greater than the critical angle (indicated as  c ) total internal reflection occurs (the path indicated in blue

11 ACS Paragon Plus Environment

Environmental Science & Technology

260

used for this research are set up to log the minimum and maximum RSL measurement

261

in intervals of 15 minutes. We used the maximum RSL measurements which, by

262

definition, would capture cases where the RSL is increased when compared to the

263

standard case. Accordingly, the proprietary sensor measurements are based on quarter

264

hourly averages.

265

Figure 2(A) shows the relative location of the research area on a map of Israel. Beside

266

it, Fig. 2(B) maps the location of MLs, the air pollution measuring stations,59 and the

267

meteorological ground stations60 from which data was taken. The monitoring stations

268

and links shown for event 1 are indicated in red, those used in event 2 are indicated in

269

green, and those used in event 3 are colored blue. The instruments used to detect the

270

inversion layer were a Lidar ceilometer located at the Beit Dagan ground station (35

271

m ASL), and a radiosonde released from that station twice every 24 hours at 00:00

272

and 12:00 (All times in this paper are stated in Universal Time Coordinated).

273

In cases where the inversion layer was not detected by the ceilometer due to

274

obstruction by low level clouds, the level of the cloud base was taken as a reference

275

altitude (these cases are indicated in the ceilometer figures with a blue line, see Figs.

276

3F and 5D). The calculations of the refraction gradient with elevation (N/H) in the

277

cases of inversion studied were based on the Beit Dagan radiosonde samples (the

278

symbol  indicates a discrete calculation). In the few cases where no specific

279

measurements were taken by the proprietary instruments, an average of adjacent

280

measurements was taken (these time periods were all shorter than or equal to half an

281

hour).

12 ACS Paragon Plus Environment

Page 12 of 34

Page 13 of 34

Environmental Science & Technology

282 283 284 285 286 287 288 289 290 291 292 293 294

Figure 2. A) Israel map. The research area focused on the center of Israel. An enlarged insert map of this area is given in Fig, 2(B). B) Microwave network and atmospheric measurement stations. Each of the three colors (Red, Green, Blue) appearing in the figure indicates the MLs and measuring stations used in each of the three events described below. Red – Event 1, Green – Event 2, Blue – Event 3. Each line represents two links operating over the same propagation path at two different frequencies between 18 and 19 GHz. The line near Tel Itzhak represents a single ML operating at a frequency of 23 GHz. Link lengths are between 6.1 and 14.3 kilometers. The elevations of the MLs indicated in red and blue vary between 41 and 87 meters ASL. The elevations of the MLs indicated in green vary between 116 and 316 meters ASL. Additionally, the meteorological ground stations at Beit Dagan (35 m ASL) and the air quality monitoring station in Holon (24 m ASL), Road 4 (50 m ASL), and Yad Rambam (92 m ASL) are shown. The location of the city of Tel Aviv is indicated with a star.

295

NOx Episode and ML Measurement: the Air Pollution Event on November 14,2010

296

According to the Israeli environmental protection agency, two episodes of extreme air

297

pollution took place in central Israel on November 14th, 2010. On that day several air

298

monitoring stations measured NOx concentrations above the standard of 100

299

(ppb/hour) 1 between 05:00 and 06:00 in the morning, as well as in the evening hours,

300

between 16:00 and 18:00. The radiosonde released at 00:00 between the 13th and 14th

301

of November measured a continuous increase in temperature (inversion) and a

13 ACS Paragon Plus Environment

Environmental Science & Technology

Page 14 of 34

302

continuous decrease in RH as altitude increased, with values between 17.2 and 23.2

303

°C and 88% to 31% for altitudes between 35 and 111 meters ASL, respectively. Using

304

Eq. (4), and given the radiosonde measurements, the change in refraction as a function

305

of altitude indicated a case of atmospheric ducting:

N  616  157 (N-units/km) H

306

(10)

307

For comparison, we reviewed the measurements of the different instruments on

308

November 11th and 11th, when these conditions did not occur and the radiosonde

309

measurements (from 00:00 between 35 and 135 meters ASL) showed:

N  39 H

310

(N-units/km)

(11)

311

Figure 3 shows the measurements of the MLs (indicated by a red line in Fig.2(B)),

312

NOx concentration values, and the ceilometer measurements taken over a period of 36

313

hours between November 13th and 14th when the air pollution episode was observed

314

(left column), and during the same hours between November 11th and 12th (right

315

column) taken as a reference day for comparison. The blue and red lines (Figs. 3(A)

316

and 3(B)) show the maximum RSL measurements of the two different links deployed

317

over the same propagation path. The green and black lines (Figs. 3(C) and 3(D)) show

318

the concentration measurements at the Holon and Road 4 stations, respectively. The

319

lines in Figs. 3(E) and 3(F) show the ceilometer measurements of the mixing layer

320

height.

321

Looking at the left column in Fig. 3, the anomaly resulting in relative amplification of

322

the RSL during episodes of air pollution can be seen (Indicated in Fig. 3(A) in grey).

323

The relative amplification was measured by both links simultaneously. During the

324

three air pollution episodes where RSL measurement anomalies were observed, both

325

ground stations near the link locations (~3.5- 4 km from the middle of the MLs) 14 ACS Paragon Plus Environment

Page 15 of 34

Environmental Science & Technology

326

simultaneously detected NOx levels above the standard value of 100 ppb,1 reaching

327

up to 7 times the standard value at peak concentration (Fig. 3(C)). Additionally, note

328

that the anomaly on the links occurred only at times when the upper boundary of the

329

inversion (Fig. 3(E)) was at its minimal height, which is of the order of the link's

330

altitude, at around 100 meters Above Ground Level (AGL). On the other hand, by

331

looking at the right column in Fig. 3, which represents a reference day, one sees no

332

RSL anomaly for the entire 36 hour period (Fig. 3(B)). The inversion base, in this

333

case, was consistently above 200 meters (Fig. 3(F)). An increase in NOx

334

measurements above standard levels was observed, though, particularly by the Road 4

335

station, between the hours of 17:00 and 22:00 on the evening of November 11th (Fig.

336

3(D)). The time interval when this increase was detected was during period with

337

highly congested traffic (the end of the work day).

338 339 340 341 342 343 344

Figure 3. The microwave network monitoring compared to specialized measurement systems. Maximum RSL measurements (A, B), NOx (C, D) and altitude of the boundary of the inversion (E, F) as measured during the high pollution day (left column), and the reference day (right column). The dashed horizontal line in figures (C, D) represents an hourly baseline level of 100 ppb above which the measured concentration is considered to exceed the standard1. The dashed line in figs. E and F represents an elevation of 100 meters AGL. The 15 ACS Paragon Plus Environment

Environmental Science & Technology

345

areas in grey broadly represent episodes where the links registered a relative increase in RSL

346

measurements and the measurements taken by the other sensors during the same time period.

347

However, these hours are also a high vehicular traffic period on the day discussed

348

(November 14), for the same reason. The possible explanation for the measurable

349

increase in NOx concentration on November 14 is thus the result of a strong near-

350

ground inversion, detected by the MLs and the proprietary instruments, compared to

351

November 11-12 where TINV and rapid lapse in humidity with altitude did not occur.

352

Accordingly, no anomalous propagation was observed by the MLs during November

353

11-12.

354

Figure 4 present scatter plots of the maximum RSL measured by the links (ML1 and

355

ML2) compared to the NOx concentration measurements taken by the Holon and

356

Road 4 monitoring stations during the air pollution event (between 09:00 on

357

November 13th, and 21:00 on November 14th, i.e. a scatter plot based on the

358

measurements presented in Figures 3(A) and 3(C)). The equation at the right bottom

359

of each panel shows the linear fit that was calculated. The Pearson correlation, R,

360

between the link and the NOx measurements was found to be between 0.47 and 0.6

361

during the event (based on 145 data points in each panel, P-value< 0.01). An increase

362

in the pollutant concentration can be seen in each of the figures when RSL

363

amplification is measured. A

B

C

D

364

16 ACS Paragon Plus Environment

Page 16 of 34

Page 17 of 34

Environmental Science & Technology

365 366 367 368

Figure 4: Maximum RSL measurements vs. NOx concentration measurements. Figures 4(A) and 4(B) show the measurements from ML1 compared to the Holon and Road 4 station measurements, respectively. Figures 4(C) and 4(D) show the measurements of ML2 compared with the Holon and Road 4 station measurements, respectively.

369

Note that there is a difference between the locations of the MLs and the locations of

370

the stations. Further, the method of measurement is different: a point measurement of

371

the air pollution sensors, compared to an areal measurement over several kilometers

372

taken by the MLs. Specifically, the phenomenon observed by each of the tools is

373

different. Air pollution sensors provide a direct measurement of NOx concentration,

374

whereas the MLs provide a measurement of received signal levels which are affected

375

by the TINV that exists in the area (and which induces an increase in pollutant

376

concentration). As a result, the figures indicate a correlation in direction, but

377

differences are expected.

378

PM10 Episode and ML Measurement: The air pollution event of 11-12 May, 2009

379

On the night between the 11th and 12th of May 2009, a holiday took place in Israel

380

("Lag Ba'Omer"61), during which, as is custom, thousands of bonfires are lit across the

381

country, starting from the evening hours of the holiday and continuing until the

382

following morning.

383

Figure 5 presents PM10 vs. commercial ML measurements which were taken over a

384

time frame of 36 hours during this air pollution event. The Figure illustrates the

385

ability of commercial ML to detect situations where air quality decreases dramatically

386

as a result of PM10 being captured within a TINV that, combined with increased

387

particulate emissions, caused extreme air pollution (i.e. PM10 concentrations of 150

388

μg / m3 or above1).

389

Radiosonde measurements from 00:00 between the above dates indicated atmospheric

390

layering in the area, and anomalous propagation conditions at the elevations of the

17 ACS Paragon Plus Environment

Environmental Science & Technology

Page 18 of 34

391

links that were examined. A temperature inversion existed in the first 140 meters ASL

392

and between 470 and 530 meters ASL (with an average increase of 1.5ºC per 100

393

meters in both cases).

394

We calculated the refractivity index using a sample from a representative elevation for

395

each layer, and the refraction changes with elevation. These values were found to be:

396

Layer I - between 35 and 140 meters:

397

N  103 (N-Units/km) – subrefraction H

398

N1(35 m ASL)= 340 (N-units)

399

Layer II - between 140 and 470 meters:

400

N  52 (N-Units/km)- standard conditions H

401

N2(240 m ASL)= 345 (N-units)

402

(12)

(13)

Layer III - between 470 and 530 meters:

403

N  502 (N-Units/km) – ducting H

404

N3(470 m ASL)= 333 (N-units)

(14)

405

We note that similar to the description in theoretical Fig. 1(A), the profile of refraction

406

indices with elevation fulfilled the condition: N1N3, for these three layers such

407

that an air waveguide was formed. The resulting atmospheric layering can be

408

observed in the region indicated in grey in Fig. 5(C), according to the Beit Dagan

409

ceilometer measurements.

410

This instrument detected the upper boundary of the ground inversion that was

411

fluctuating between elevations of 140 and 200 meters AGL, between 21:20 and 03:40

412

of 11 and 12 May, respectively (brown graph). In conjunction, during part of this

18 ACS Paragon Plus Environment

Page 19 of 34

Environmental Science & Technology

413

period, the ceilometer detected a cloud base (the graph in blue) at the elevation of the

414

layer III inversion base (around 500 meters ASL).

415

Now let us focus on two MLs that span 14.3 kms in length and operate along a slant

416

path from 116 m to 316 m ASL (marked in green in Fig. 2). The propagation path

417

elevation of these links passes through layer I and is the same, for the most part, as the

418

air waveguide created in layer II.

419

These links measured a relative increase in RSL (Fig. 5 (A)) between the hours of

420

22:45 and 04:15 ( of 11 and 12 May, respectively), overlapping with the period when

421

the peak PM10 concentration was observed at Yad Rambam air monitoring station

422

between 21:45 and 03:30 (Fig. 5(B)).

423

Notably, Yad Rambam ground station is located 92m ASL, i.e. within the ground level

424

TINV. As a result, the increase in PM10 concentration is correlated with the

425

measurements of the microwave system (and ceilometer). When the layering

426

dissipated, the microwave propagation returned to normal (and the pollutant

427

concentration dropped measurably).

428

The Pearson correlation between the measurements of links 1 and 2, and the PM10

429

measurements from Yad Rambam station was found to be 0.56 and 0.59 respectively

430

(based on 145 samples, P-value< 0.01). Scatter plots created for this event in the

431

manner described in Fig. 4 showed similar results.

19 ACS Paragon Plus Environment

Environmental Science & Technology

Page 20 of 34

A

B

C

432 433 434 435 436 437 438 439 440

Figure 5. Proprietary atmospheric instrument measurements compared to microwave network measurements. The observations were taken between 11-12 May 2009 (36 hours). ML measurements (A), PM10 measurements taken by the station at Yad Rambam (B), and Beit Dagan ceilometer measurements (C). Notably, an atmospheric layering state can be observed in the region indicated in grey in Fig. 5(C). A sharp increase in PM10 concentration can been seen at the Yad Rambam station as the top boundary of the inversion layer descends to 200 meters and below (as measured by the ceilometer). Accordingly, a correlation can be seen with the measurements of the MLs operating between 116 and 316 meters ASL.

441

Fog and ML Measurement: The fog event of 19- 20 November, 2010

442

According to the Israeli Meteorological Service, starting in the late evening of

443

November 19th, heavy fog began to develop over Southern and Central Israel,

444

reaching its peak in the early morning hours of November 20th, 2010. A Red Sea

445

Trough20 with a central axis moved Eastward over the ground, accompanied by a

446

barometric high to the south-west at altitude. The change to a westerly flow after a hot

447

dry day, with cold, humid air intruding under a strong, low marine inversion on the

448

shore, caused the event to develop.

449

Figure 6 presents RSL measurements taken by 3 MLs (Fig, 6 (A),(B)), a ceilometer

450

(Fig. 6(C)) and Meteorological Optical Range (MOR27) measurements – (Fig. 6(D)).

20 ACS Paragon Plus Environment

Page 21 of 34

Environmental Science & Technology

451

The ceilometer detected the upper boundary of the inversion that was fluctuating

452

between altitudes of 90 and 170 meters AGL, between 17:30 and 07:30 of November

453

19- 20, respectively. The MLs (Fig. 2, blue) are deployed at elevations between 41-87

454

meters ASL. The measurements of the radiosonde released on the night of the event

455

(00:00) indicated conditions of anomalous propagation. At elevations between 35 and

456

200 meters sub-refraction conditions were measured (N/H = 137 (N-Units/km)),

457

while in the layer above, between 200 and 240 meters ducting was measured (N/H

458

= −792 (N-Units/km)). A temperature inversion at an average rate of 3.6ºC per 100

459

meters existed continuously in these elevations (between 35 and 240 meters ASL).

460

The Beit Dagan station (35 ASL) measured high RH values (96%-100%) during 20:00

461

and 08:00 of November 19-20, respectively.

462

According to the above, we calculated the refraction index according to the

463

radiosonde samples from three representative altitudes in the 240 meters above sea

464

level. The values were found to be:

465

N1 at ground level (35 meters ASL)= 352 (N-units)

466

N2 (170 meters ASL)= 375 (N-units )

467

N3 (240 meters ASL)= 340 (N-units)

468

Note that the atmospheric condition – N1< N2> N3 is fulfilled in this case as well.

469

An anomaly (relative amplification) can be seen simultaneously in the measurements

470

of the 3 links prior to the peak of the fog event. The beginning of the peak fog level is

471

indicated by the vertical dashed line in Fig. 6, and from that moment on the

472

ceilometer detected severe visibility limitations, and indicated conditions called “Sky

473

obscured” (according to the Synop code). This period is indicated by a dashed blue

474

line in Fig. 6(C). The effective meaning is that during this interval thick fog existed at

21 ACS Paragon Plus Environment

Environmental Science & Technology

Page 22 of 34

475

ground level. The MOR, taken by a visibility sensor at Beit Dagan during the same

476

period, dropped to a few tens to hundreds of meters (Fig. 6(D)).

A

B

C

D

477 478 479 480 481 482 483

Figure 6. The fog event between 19- 20 November 2010 (36 hours): proprietary instrument measurements vs. microwave network measurements. The measurements from the two MLs which were operating over the same propagation path near Herzliya (Fig. 2(B), blue) are presented in Fig. 6(A). The measurements from the Tel Itzhak ML (Fig. 2(B), blue) are shown in Fig. 6(B). Ceilometer measurements (Fig. 6(C)) and MOR observations (Fig. 6(D)) were taken from Beit Dagan station.

484

We note that the anomalies in the link measurements ended simultaneously with the

485

beginning of the fog's peak. A plausible explanation for this observation is the

486

inversion layer sinking down to an altitude below the link's line of sight (as is

487

supported by the ceilometer measurements), and thus above the layer adjacent to the

488

ground, standard atmospheric conditions that allowed for normal propagation now

489

returned. Prior research54 has shown that conditions of anomalous propagation are

490

typical in precursor conditions to fog, and can provide a measure for early warning

491

against its occurrence.

492 493 494

The proof-of-concept results presented above point to the feasibility of using commercial microwave infrastructure as an environmental sensor network for 22 ACS Paragon Plus Environment

Page 23 of 34

Environmental Science & Technology

495

detecting anomalies in the atmospheric refraction index near ground level. The

496

interest in detecting these cases derives from the fact that the phenomenon of

497

atmospheric layering is associated with TINVs, a leading cause to a plunge in air

498

quality, and to fog.

499

The validation of the concept presented here draws robust scientific rationale from

500

several different aspects. The verification was supported by observations during

501

different events, different episodes, in varying locations, and with different

502

microwave systems while, generally, in events where no ducting occurred, no

503

anomalies in link measurements were observed. More specifically, in cases where

504

these conditions were detected by the specialized instruments at elevations in the

505

order of magnitude of the propagation path, a relative increase in RSL was observed.

506

We note that, in particular, a relative increase in RSL is a unique effect which is linked

507

directly with ducting and multi-pathing, phenomena that occur during inversions.48, 54

508

From a different vantage point, previous research has shown marked increase in

509

specific air pollutant concentrations, similar to those observed here, during TINVs.6,

510

10, 58

511

the variance of the observed NO, NO2, PM10 (and benzene) concentrations was

512

caused by changes in the mixing layer height (which is bounded by the inversion

513

base). The results revealed here from the proposed novel method of using cellular

514

communication infrastructure as a sensor network are therefore corroborated by those

515

arrived at in prior research.

516

The extreme visibility limitations caused by fog can lead to tragic accidents and

517

widespread economic damages, such as disruption of schedules and flight

518

cancellations.21-23 However, the tools that exist today, do not always provide sufficient

A recent study conducted by Schäfer et al.9 concluded that an important part of

23 ACS Paragon Plus Environment

Environmental Science & Technology

519

response.45, 46 Thus the motivation to develop techniques to improve the ability to

520

monitor and forecast this phenomenon.

521

The third event discussed above provides additional proof of feasibility for the ability

522

of commercial MLs to identify low lying TINVs and more specifically the potential of

523

this technology to detect and forecast fog, a precondition of which is the existence of

524

this phenomenon.20

525

We note that the potential for monitoring fog using commercial microwave links was

526

revealed in recent works,45, 46 but the physical principle on which the measurements

527

were based was different. In these recent works fog was measured directly, based on

528

the attenuation of the microwave signal by the fog droplets suspended in the air, and

529

specifically, only after the phenomenon had already occurred. In the principle we

530

demonstrate here, what is detected is anomalous propagation caused by a low lying

531

TINV, which is one of the prerequisite conditions for the creation or strengthening of

532

fog.54 Thus, use of the method proposed here, can provide information crucial to

533

early warning about, hence proactive mitigation or preparation for the phenomenon.

534

During the second and third events studied here, proximity to the proprietary

535

instruments (radiosonde, ceilometer, visibility sensor) was slightly lower when

536

compared to the first case. However, the comparison between the different

537

instruments is reasonable, because ducting and inversions take place over a larger

538

spatial scale than the difference in locations between the instruments and links. The

539

typical spatial scale of the phenomenon can reach tens and even hundreds of

540

kilometers,62 and particularly in the test area where the climate and geographic

541

conditions are similar.19 Thus, the phenomenon was observed simultaneously by the

542

proprietary instruments and the MLs in events 2 and 3 as well. That being said, clearly

543

the altitude and thickness of the various layers differs from one place to the next (e.g.

24 ACS Paragon Plus Environment

Page 24 of 34

Page 25 of 34

Environmental Science & Technology

544

Bean63), and may contribute to the discrepancy in measurements between the different

545

methods.

546

While in case 1 ducting was observed directly at link elevation, in cases 2 and 3 sub-

547

refraction conditions were observed with ducting conditions occurring above. Sub-

548

refraction conditions cause an upwards bending of the wave, so in these cases it is

549

possible that the amplification was a result of an upward deflection of the beam and

550

its capture in the duct layer, or constructive interference due to multipath as a

551

consequence of the beam being subnormally bent upwards from the lower layer and

552

then reflecting downwards off the higher layer. Alternately, it is possible that the

553

atmospheric waveguide was at the elevation of the links in the area where they were

554

deployed thus, causing the amplification. However, due to the difference in locations

555

between the links and the proprietary measuring instruments the elevation detected at

556

the measuring station (Beit Dagan) did not correctly represent the elevation of the air

557

waveguide at the link location.

558

Overall, our results indicate the potential of commercial microwave systems to

559

provide indirect detection of air pollution and fog. The widely existing MLs, in a

560

unique manner with inherent advantages over the commonly used conventional

561

techniques, provide near-ground atmospheric measurements at a high tempo-spatial

562

resolution with simple implementation of minimal extra cost. The work reported in

563

this paper shows feasibility of this novel idea, and possible future research can be

564

conducted to fully investigate relevant practical details and societal implications along

565

the new path presented here.

566 567

25 ACS Paragon Plus Environment

Environmental Science & Technology

568

DISCUSSION

569

The novelty of this work stems from revealing the great potential of cellular

570

communication networks for detecting near-ground temperature inversions, and the

571

adverse atmospheric conditions that can exist during such occurrence. Due to their

572

deployment at ground level, and their wide geographic distribution and the type of

573

measurements they offer, commercial MLs are particularly suited for detection and for

574

provision of real time warning of air pollution episodes and fog. The microwave

575

network engineering literature refers to the phenomenon of atmospheric stratification

576

as interference that more intensively affects longer microwave links (e.g., 5km or

577

longer in length),48 and as a result, such longer links are the MLs with the highest

578

potential that can be exploited for detecting near-ground hazardous atmospheric

579

conditions. Future research can be done to analyze additional events in different

580

regions using ML networks in conjunction with ground truth instruments adjacent to

581

the link. Robust signal processing algorithms and comprehensive analysis methods for

582

the large volume of ML data can then be developed to support practical deployment of

583

ML-based high-resolution atmospheric sensor networks at near-ground level. The

584

proposed technique can provide benefit to many developing countries (e.g. India,

585

countries in Africa, etc.) where the percentage of ML facilities is high,41, 64 while on

586

the other hand, conventional monitoring resources are limited, if they exist at all. The

587

current paper concentrated on the signal amplification effect as a result of the

588

atmosphere behaving as a wave guide. However the multipath effect, as well as

589

ducting, that are identified with cases of TINV, may also cause signal attenuation.54-56

590

Furthermore, in some specific cases, signal amplification may be happening due to

591

other reasons, such as multipath due to terrestrial objects in the close vicinity of the

26 ACS Paragon Plus Environment

Page 26 of 34

Page 27 of 34

Environmental Science & Technology

592

propagation path, which can lead to misdetection. Further investigation, outside the

593

scope of the current paper, is required on these topics, in future research.

594

It appears that the field of wireless environmental sensor networks had progressed less

595

quickly than was expected about a decade ago, as a result of a multitude of obstacles,

596

including – deployment costs, communication and data processing requirements,

597

maintenance and reliability.67 The main characteristics of an efficient wireless sensor

598

network require: low cost, large spatial dispersion and autonomous sensing.68

599

Commercial MLs operate continuously by default, the system is routinely maintained

600

and monitored, the measuring instruments are already deployed in the field, and the

601

relevant data sent and collected efficiently, by many of the cellular providers, as part

602

of normal operation. Thus, this infrastructure can be considered as an "opportunistic

603

wireless sensor network"68 for environmental monitoring. The challenge, then, is the

604

alternative use of such resources. The possibilities arise from the diversity of the

605

opportunistic sensors, reflected in the different operating frequencies of the system,

606

the variation in signal polarization, different link lengths and varying link elevations

607

above ground. By definition, the system is a communication network, and thus,

608

naturally, the measurement quality of a single sensor (link) may at times not be

609

optimal, but the number of measurements is large and diverse – and thus compensates

610

for the measurement quality of each single link, or, alternatively, allows for the

611

selection of specific, suitable links from the large array available for monitoring the

612

phenomenon of interest.

613

We note that in addition to MLs, data transmission in the network can be carried out

614

by other means, such as fiber optic (or copper) cables. There are some locations where

615

MLs are replaced with fiber optic cables. However, the use of microwave

616

infrastructure is expected to continue to be widespread in the future.64

27 ACS Paragon Plus Environment

Environmental Science & Technology

617

Considering commercial microwave systems as an innovative method for monitoring

618

atmospheric conditions creates many opportunities to provide valuable information

619

regarding air quality. To this point, tools for the creation of 2D rainfall maps (e.g.

620

Overeem et al.42) and fog monitoring (David et al.45, 46) were developed. Thus, the

621

system can provide information about relatively clean air regions due to washout

622

effects as a result of rainfall65 in a known area mapped by the system or pollutant

623

scavenging processes due to fog66 – new capabilities that have yet to be demonstrated

624

using this technology. However, the goal of the current research was to take the whole

625

approach one step further, and to show proof of concept of the potential commercial

626

MLs may have to monitor air quality and precursors for fog.

627

AUTHOR INFORMATION

628

Corresponding Author

629

*Email: [email protected], Phone: +1607-2548334.

630

Present Addresses

631

§School of Civil and Environmental Engineering, Cornell University, Ithaca, NY

632

14853.

633

ACKNOWLEDGMENTS

634

This work is dedicated to Ella David Helfman, the daughter of Adi and Noam (the

635

lead author), in celebration of her first birthday.

636

The microwave data were provided by Cellcom, Pelephone and PHI to the research

637

team of Tel Aviv University (TAU) to whom Dr. David was affiliated. The authors

638

are grateful to Y. Koriat, I. Alexandrovitz, E Levi and B. Bar (Cellcom), N. Dvela, A.

28 ACS Paragon Plus Environment

Page 28 of 34

Page 29 of 34

Environmental Science & Technology

639

Hival and Y. Shachar (Pelephone), Y. Bar Asher, O. Tzur, Y. Sebton, A. Polikar and

640

O. Borukhov (PHI). We extend our sincere acknowledgments to Prof. Hagit Messer,

641

Prof. Pinhas Alpert and their research team from TAU for fruitful and valuable

642

discussions. Finally, we deeply thank the Israeli Meteorological Service (IMS) and

643

the Israel Ministry of Environmental Protection for supplying meteorological and air

644

pollution data. This work was supported by Cornell University's David R. Atkinson

645

Center for a Sustainable Future (ACSF). The second author acknowledges partial

646

funding support from the Natural Science Foundation of China (NSFC) project #

647

71428001.

648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677

REFERENCES 1. US EPA; http://www.epa.gov/ttn/naaqs/criteria.html, accessed in February 2016. 2. WHO; http://www.who.int/mediacentre/factsheets/fs313/en/, 2014. 3. Schrenk, H. H.; Heimann, H.; Clayton, G. D.; Gafafer, W. M.; Wexler, H. Air pollution in Donora, Pa. Epidemiology of the unusual smog episode of October 1948. Preliminary Report. Pub. Health Bull. 1949, 306. 4. Kallos, G.; Kassomenos, P.; Pielke, R. A.; Synoptic and mesoscale weather conditions during pollution episodes in Athens, Greece. Bound.-Layer Meteorol. 1993, 62, 163–184. 5. Incecik, S. Investigation of atmospheric conditions in Istanbul leading to air pollution episodes. Atmos. Environ. 1996, 30, 2739–2749. 6. Kukkonen, J.; Pohjola, M.; Sokhi, R. S.; Luhana, L.; Kitwiroon, N.; Fragkou, L.; Rantamäki, M.; Berge, E.; Ødegaard, V.; Slørdal, L.H.; Denby, B; Denby, B.; Finardi, S. Analysis and evaluation of selected local-scale PM10 air pollution episodes in four European cities: Helsinki, London, Milan and Oslo. Atmos. Environ. 2005, 39, 2759–2773. 7. Malek, E.; Davis, T.; Martin, R. S.; Silva, P. J. Meteorological and environmental aspects of one of the worst national air pollution episodes (January, 2004) in Logan, Cache Valley, Utah, USA. Atmos. Res. 2006, 79, 108–122. 8. Bailey, A.; Chase, T. N.; Cassano, J. J.; Noone, D. Changing temperature inversion characteristics in the US southwest and relationships to large-scale atmospheric circulation. J. App. Met. Climatol. 2011, 50(6), 1307-1323.‫‏‬ 9. Schäfer, K.; Wagner, P.; Emeis, S.; Jahn, C.; Muenkel, C.; Suppan, P. Mixing layer height and air pollution levels in urban area. In International Society for Optics and Photonics (SPIE) Remote Sensing. 2012, 853409-853409.‫‏‏‬ 10. Shmool, J. L.; Michanowicz, D. R.; Cambal, L.; Tunno, B.; Howell, J.; Gillooly, S.; Roper, C.; Tripathy, S.; Chubb, L.G.; Eisl, H.M; Gorczynski, J.E.; Holguin, F.E.; Shields, K.N.; Gorczynski, J. E. Saturation sampling for spatial variation in multiple air pollutants across an inversion-prone metropolitan area of complex terrain. Environ. Health. 2014, 13(1), 28. 29 ACS Paragon Plus Environment

Environmental Science & Technology

678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727

11. Haikin, N.; Galanti, E.; Reisin, T. G.; Mahrer, Y.; Alpert, P. Inner Structure of Atmospheric Inversion Layers over Haifa Bay in the Eastern Mediterranean. Bound.-Layer Meteorol. 2015, 156(3), 1-17.‫‏‬ 12. André, J. C.; Mahrt, L. The Nocturnal Surface Inversion and Influence of ClearAir Radiative Cooling. J. Atmos. Sci. 1982, 39, 864–878. 13. Schnelle, K. B.; Brown, C. A. The Air Pollution Control Technology Handbook. CRC Press: Boca Raton, London, New York, Washington D.C., 2002. 14. Vihma, T.; Hartmann, J.; Lüpkes, C. A case study of an on-ice air flow over the Arctic marginal sea-ice zone. Bound.-layer meteorol. 2003, 107(1), 189-217.‫‏‬ 15. Silva, P. J.; Vawdrey, E. L.; Corbett, M.; Erupe, M. Fine particle concentrations and composition during wintertime inversions in Logan, Utah, USA. Atmos. Environ. 2007, 41(26), 5410-5422.‫‏‬ 16. Beard, J. D.; Beck, C.; Graham, R.; Packham, S. C.; Traphagan, M.; Giles, R. T.; Morgan, J. G. Winter temperature inversions and emergency department visits for asthma in Salt Lake County, Utah, 2003-2008. Environ. health perspect. 2012, 120(10). 1385-1390.‫‏‬ 17. Abdul-Wahab, S. A.; Bakheit, C. S.; Siddiqui, R. A. Study the relationship between the health effects and characterization of thermal inversions in the Sultanate of Oman. Atmos. Environ. 2005, 39(30), 5466-5471.‫‏‬ 18. Johnstone, J. A.; Dawson, T. E. Climatic context and ecological implications of summer fog decline in the coast redwood region. Proc. Natl. Acad. Sci. 2010, 107(10), 4533-4538. 19. Gultepe, I.; Tardif, R.; Michaelides, S. C.; Cermak, J.; Bott, A.; Bendix, J.; Muller, M.D.; Pagowsky, M.; Hansen, B.; Ellord, G.; Jacobs, W.; Toth, G.; Cober, S. G. Fog research: A review of past achievements and future perspectives. Pure Appl. Geoph. 2007, 164(6-7), 1121-1159. ‫‏‬ 20. Yair, Y; Ziv, B. An Introduction to Meteorology. The open University of Israel Press: Israel, 2012. 21. Pagowski, M.; Gultepe, I.; King, P. Analysis and modeling of an extremely dense fog event in southern Ontario. J. appl. meteor. 2004, 43(1), 3-16. 22. Gultepe, I.; Pearson, G.; Milbrandt, J. A.; Hansen, B.; Platnick, S.; Taylor, P.; Gordon, M.; Oakley, J.P.; Cober, S. G.. The fog remote sensing and modeling field project. Bull. Amer. Meteorol. Soc. 2009, 90(3), 341-359. 23. Gadher, D.; Baird, T. Airport dash as the fog lifts. The Sunday Times, posted online 24 December 2006. [Available online at: http://www.thesundaytimes.co.uk/sto/news/uk_news/article56426.ece] 24. Emeis, S.; Schäfer, K.; Münkel, C. Surface-based remote sensing of the mixinglayer height–a review. Meteorologische Zeitschrift. 2008, 17(5), 621-630. 25. Emeis, S.; Schäfer, K.; Münkel, C. Observation of the structure of the urban boundary layer with different ceilometers and validation by RASS data. Meteorologische Zeitschrift. 2009, 18.2, 149-154. 26. De Haij, M.; Wauben, W.; Baltink, H. K. Determination of mixing layer height from ceilometer backscatter profiles. In Remote Sensing International Society for Optics and Photonics, 2006, 63620R-63620R. 27. World Meteorological Organization (WMO). Guide to Meteorological Instruments and Methods of Observation, 7th, ed.; Chairperson, Publications Board: Geneva, Switzerland, 2008. 28. Liu, Y.; Key, J. R.; Schweiger, A.; Francis, J. Characteristics of satellite-derived clear-sky atmospheric temperature inversion strength in the Arctic, 1980-96. J. climate. 2006, 19(19), 4902-4913. 30 ACS Paragon Plus Environment

Page 30 of 34

Page 31 of 34

728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777

Environmental Science & Technology

29. McGrath‐Spangler, E. L.; Denning, A. S. Global seasonal variations of midday planetary boundary layer depth from CALIPSO space‐borne LIDAR. J. Geophys. Res.-Atmos. 2013, 118(3), 1226-1233. 30. Messer, H.; Zinevich, A.; Alpert, P. Environmental monitoring by wireless communication networks. Science. 2006, 312(5774), 713-713. 31. Leijnse, H.; Uijlenhoet, R.; Stricker, J. Rainfall measurement using radio links from cellular communication networks. Water Resour. Res. 2007, 43, W03201, DOI 10.1029/2006WR005631. 32. Leijnse, H.; Uijlenhoet, R.; Stricker, J. N. M. Microwave link rainfall estimation: Effects of link length and frequency, temporal sampling, power resolution, and wet antenna attenuation. Adv. Water Resour. 2008, 31(11), 1481-149. 33. Zinevich, A.; Alpert, P.; Messer, H. Estimation of rainfall fields using commercial microwave communication networks of variable density. Adv. Water Resour. 2008, 31, 1470-1480. 34. Zinevich, A.; Messer, H.; Alpert, P. Frontal rainfall observation by a commercial microwave communication network. J. Appl. Meteor. Climatol. 2009, 48, 13171334. 35. Zinevich, A.; Messer, H.; Alpert, P. Prediction of rainfall intensity measurement errors using commercial microwave communication links. Atmos. Meas. Tech. 2010, 3, 1385-1402. 36. Chwala, C.; Gmeiner, A.; Qiu, W.; Hipp, S.; Nienaber, D.; Siart, U.; Eibert, T.; Pohl, M.; Seltmann, J.; Fritz, J.; Kunstmann, H. Precipitation observation using microwave backhaul links in the alpine and pre-alpine region of Southern Germany. Hydrol. Earth Syst. Sci. 2012, 16, 2647-2661, DOI 10.5194/hess-162647-2012. 37. Rayitsfeld, A.; Samuels, R.; Zinevich, A.; Hadar, U.; Alpert, P. Comparison of two methodologies for long term rainfall monitoring using a commercial microwave communication system. Atmos. Res. 2012, 104-105, 119-127. 38. Overeem, A.; Leijnse, H.; Uijlenhoet, R. Country-wide rainfall maps from cellular communication networks. Proc. Natl. Acad. Sci. 2013, 110, 2741 – 2745, DOI 10.1073/pnas.1217961110. 39. David, N.; Alpert, P.; Messer, H. The potential of cellular network infrastructures for sudden rainfall monitoring in dry climate regions. Atmos. Res. 2013, 131, 1321, DOI 10.1016/j.atmosres.2013.01.004. 40. Schleiss, M.; Rieckermann, J.; Berne, A. Quantification and modeling of wetantenna attenuation for commercial microwave links. IEEE Geosci. Remote Sens. Lett. 2013, 10(5), 1195-1199. ‫‏‬ 41. Doumounia, A.; Gosset, M.; Cazenave, F.; Kacou, M.; Zougmore, F. Rainfall monitoring based on microwave links from cellular telecommunication networks: First results from a West African test bed. Geophys. Res. Lett. 2014, 41(16), 60166022. 42. Liberman, Y.; Samuels, R.; Alpert, P.; Messer, H. New algorithm for integration between wireless microwave sensor network and radar for improved rainfall measurement and mapping. Atmos. Meas. Tech. 2014, 7(10), 3549-3563. 43. David, N.; Alpert, P.; Messer, H. Technical Note: Novel method for water vapour monitoring using wireless communication networks measurements. Atmos. Chem. Phys. 2009, 9, 2413-2418. 44. David, N.; et al. Humidity measurements using commercial microwave links. In Advanced Trends in Wireless Communications; Khatib, M. Ed.; InTech publications: Rijeka, Croatia 2011; pp 65-78; DOI 10.5772/15292. 31 ACS Paragon Plus Environment

Environmental Science & Technology

778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827

45. David, N.; Alpert, P.; Messer, H. The potential of commercial microwave networks to monitor dense fog- feasibility study. J. Geophys. Res.-Atmos. 2013, 118 (20), 11,750–11,761. 46. David, N.; Sendik, O.; Messer, H.; Alpert, P. Cellular network infrastructure-the future of fog monitoring?. Bull. Amer. Meteorol. Soc. 2015, 96, 1687-1698,‫ ‏‏‬DOI 10.1175/BAMS-D-13-00292.1. 47. Harel., O; David, N.; Alpert, P.; Messer, H. The potential of microwave communication networks to detect dew using the GLRT- experimental study. IEEE J. Sel. Topics Appl. Earth Observ. in Remote Sens. 2015, 8(9), 4396- 4404, DOI 10.1109/JSTARS.2015.2465909. 48. Rec. ITU-R P.452-13. Prediction procedure for the evaluation of microwave interference between stations on the surface of the Earth at frequencies above about 0.7 GHz, 2007. 49. Burrows, C. R. Radio Wave Propagation: Consolidated Summary Technical Report of the Committee on Propagation of the National Defense Research Committee, Radio Wave Propagation Consolidated Summary Technical Report of the Committee on Propagation of the National Defense Research Committee; Academic press inc: New York, USA, 1949. 50. Smith, E. K.; Weintraub, S. The constants in the equation for atmospheric refractive index at radio frequencies. Proceedings of the IRE. 1953, 41(8), 10351037. 51. Rec. ITU-R P.453-10. The radio refractive index: its formula and refractivity data, 2012. 52. Von Engeln, A.; Teixeira, J. A ducting climatology derived from the European Centre for Medium‐Range Weather Forecasts global analysis fields. J. Geophys. Res.-Atmos. 2004, 109(D18), 1984–2012. 53. Bakshi, U. A.; Bakshi, A. V.; Bakshi, K. A. Antenna and Wave Propagation; Technical Publications: Pune, India, 2009. 54. Raghavan, S. Radar meteorology; Springer Science and Business Media: Dordrecht, Netherlands, 2003. 55. Mishra, A. R. Fundamentals of cellular network planning and optimization: 2G/2.5 G/3G... evolution to 4G; John Wiley and Sons: Chichester, England, 2004. 56. Rec. ITU-R P.453-11. The radio refractive index: its formula and refractivity data, 2015. 57. Valtr, P.; Pechac, P.; Kvicera, V.; Grabner, M. Estimation of the refractivity structure of the lower troposphere from measurements on a terrestrial multiplereceiver radio link. IEEE Trans. Antennas Propag. 2011, 59(5), 1707-1715. 58. Wallace, J.; Kanaroglou, P.; The effect of temperature inversions on ground-level nitrogen dioxide (NO2) and fine particulate matter (PM2.5) using temperature profiles from the Atmospheric Infrared Sounder (AIRS). Sci. Tot. Environ. 2009, 407, 5058–5095. 59. Israel Ministry of Environmental Protection; Http://www.sviva.gov.il/English (the interface language for access to the data itself is in Hebrew) 60. Israeli Meteorological Service; https://ims.data.gov.il/ims/7 (the site is in Hebrew). 61. Adler, G.; Flores, J. M.; Abo Riziq, A.; Borrmann, S.; Rudich, Y. Chemical, physical, and optical evolution of biomass burning aerosols: a case study. Atmos. Chem. Phys. 2011, 11(4), 1491-1503. 62. Brooks, I. M.; Goroch, A. K.; Rogers, D. P. Observations of strong surface radar ducts over the Persian Gulf. J. Appl. Meteor. 1999, 38(9), 1293-1310. 32 ACS Paragon Plus Environment

Page 32 of 34

Page 33 of 34

828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843

Environmental Science & Technology

63. Bean, B. R. On the Climatology of Ground-based Radio Ducts. J. Res. Natl. Bur. Stand. 1959, 63D(1), 29-34. 64. Ericsson. Delivering high capacity and cost efficient backhaul for broadband networks today and in the future; Microwave towards 2020; Stockholm, Sweden, 2015. 65. Yoo, J. M.; Lee, Y. R.; Kim, D.; Jeong, M. J.; Stockwell, W. R.; Kundu, P. K.; Oh, S.M.; Shin, D.B.; Lee, S. J. New indices for wet scavenging of air pollutants (O 3, CO, NO 2, SO 2, and PM 10) by summertime rain. Atmos. Environ. 2014, 82, 226-237.‫‏‬ 66. Herckes, P.; Chang, H.; Lee, T.; Collett Jr, J. L. Air pollution processing by radiation fogs. Water, air, and soil Pollution. 2007, 181(1-4), 65-75. ‫‏‬ 67. Corke, P.; Wark, T.; Jurdak, R.; Hu, W.; Valencia, P.; Moore, D. Environmental wireless sensor networks. Proc. IEEE. 2010, 98(11), 1903-1917. 68. Sendik, O.; Messer, H. A New Approach to Precipitation Monitoring: A critical survey of existing technologies and challenges. IEEE Signal Process. Mag. 2015, 32(3), 110-122.

844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 33 ACS Paragon Plus Environment

Environmental Science & Technology

864

Table of Contents Graphic

865 866

34 ACS Paragon Plus Environment

Page 34 of 34