Intercomparison of Aerosol Optical Thickness Derived from MODIS

Jul 9, 2015 - Larger windows can introduce undesirable results because of surface- or aerosol-type heterogeneity, and also, there is a possibility for...
3 downloads 6 Views 2MB Size
Subscriber access provided by UNIV OF MISSISSIPPI

Article

INTERCOMPARISON OF AEROSOL OPTICAL THICKNESS DERIVED FROM MODIS AND IN-SITU GROUND DATASETS OVER JAIPUR -A SEMI ARID ZONE IN INDIA Swagata Payra, Manish Soni, Anikender Kumar, Divya Prakash, and Sunita Verma Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.5b02225 • Publication Date (Web): 09 Jul 2015 Downloaded from http://pubs.acs.org on July 12, 2015

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 31

Environmental Science & Technology

82x45mm (96 x 96 DPI)

ACS Paragon Plus Environment

Environmental Science & Technology

1

INTERCOMPARISON OF AEROSOL OPTICAL

2

THICKNESS DERIVED FROM MODIS AND IN-

3

SITU GROUND DATASETS OVER JAIPUR - A

4

SEMI ARID ZONE IN INDIA

5

Page 2 of 31

Swagata Payra1, Manish Soni1, Anikender Kumar2, Divya Prakash1 and Sunita Verma1,*

6

7 8

1

Centre for Excellence in Climatology, Birla Institute of Technology Mesra, Jaipur Campus, Jaipur 302017, Rajasthan, India

9 10 11

2

Department of Chemical and Environment Engineering, National University of Colombia, Bogota, Colombia.

12

13

14

15

16

ACS Paragon Plus Environment

1

Page 3 of 31

Environmental Science & Technology

17

Abstract

18

First detailed seasonal validation has been carried out for the MODerate Resolution Imaging Spectro-

19

radiometer (MODIS) terra and aqua satellites level 2.0 collection version 5.1 AOT (τMODIS) with Aerosol

20

Robotic Network (AERONET) level 2.0 AOT (τAERONET) for the year 2009-2012 over semi arid region

21

Jaipur, north western India. The correlation between τMODIS vs τAERONET at 550 nm is determined with

22

different spatial and temporal size windows. The τMODIS overestimates τAERONET within a range of

23

+0.06±0.24 during the pre-monsoon (Apr-Jun) season while it underestimates the τAERONET with -

24

0.04±0.12 and -0.05±0.18 during dry (Dec-Mar) and post-monsoon (Oct-Nov) seasons, respectively.

25

Correlation without (with) Error Envelope has been found for pre-monsoon 0.71 (0.89), post-monsoon

26

0.76 (0.94) and dry season 0.78 (0.95). The τMODIS is compared with τAERONET at three more ground

27

AERONET stations in India i.e. Kanpur, Gual Pahari and Pune. Furthermore, the performance of MODIS

28

Deep blue and Aqua AOT550 nm (τDB550nm and τAqua550nm) with τAERONET is also evaluated for all considered

29

sites over India along with a US desert site at White Sand, Tularosa Basin, New Mexico. The statistical

30

results reveal that τAqua550nm performs better over Kanpur, Pune whereas τDB550nm performs better over

31

Jaipur, Gual Pahari and White Sand HELSTF (US site).

32

33

Keywords: MODIS, AERONET, Aerosols, Dust, Climate, Deep Blue

34

*Correspondence:

35

Dr. Sunita Verma, Phone:+ 91-141-2385094 (extn - 306), Email:[email protected], Centre of

36

Excellence in Climatology, Birla Institute of Technology Mesra, Jaipur Campus, 27 Malviya Industrial

37

Area, Jaipur 302017, Rajasthan, India

38

ACS Paragon Plus Environment

2

Environmental Science & Technology

Page 4 of 31

39 40

1. Introduction

41

Aerosols play an important role in the climate of the Earth-atmosphere system by means of their

42

direct and indirect impact on climate.1 Atmospheric aerosols are able to alter Earth’s Radiation budget by

43

scattering and absorption of solar radiation. Due to aerosol’s highly variable spatial and temporal

44

distribution, large uncertainty exists in the aerosol radiative forcing and for that reason the level of

45

scientific understanding is very low.2

46 47

The in-situ as well as satellite measurements can help in the scientific understanding of aerosols.

48

Aerosols observations through remote sensing can provide detailed knowledge on a long time scale

49

covering large spatial area.3 Ground based observation is a point measurement and cannot provide spatial

50

variation unless a large network is working. In satellite observation, same instrument makes observation

51

globally. The aerosol concentration can be compared at different locations which will not be affected by

52

the calibration errors of the instrument.4 Satellite data provide us to quantitatively determine the aerosol

53

optical properties and thus global aerosol budget.

54

considerable errors in the retrieved aerosol products where the surface reflectance is high while for

55

ground based data have no such constraints and therefore it provide the measurement of microphysical

56

and optical properties of ambient columnar aerosols with high accuracy.7 The bias or error in Aerosol

57

Optical Thickness (AOT) retrievals from satellite could arise from various sources e.g. incorrect

58

assumptions about surface reflectance, aerosol type, status of sensor calibration and observation

59

geometry, etc.8

5, 6

However, the satellite measurement may introduce

60

The validation of satellite based measurement with ground based observation is important to

61

establish its quality and suitability to be used in climatology and weather modeling. Ground based data is

62

used for validation of satellite measured aerosol properties. 9 The MODIS-derived aerosol properties over

63

land

4, 7, 9-13

and over the ocean

14, 15

have been validated by a large number of ground based stations

ACS Paragon Plus Environment

3

Page 5 of 31

Environmental Science & Technology

64

worldwide. There are several studies on the validation of aerosol optical thickness data of ground based

65

instrument with satellite based measurement and is in progress since sensors started to provide the data. 4,

66

7, 11-12, 15-23

67

The retrieval of AOT over land continues to be a challenge, due to large surface heterogeneity

68

especially over India. This can introduce an error in estimated surface reflectance, which goes into

69

radiative transfer calculations to retrieve AOT from satellite data. The averaging of pixels and cloud

70

contamination is another source of error in remote sensing. To improve the accuracy of the MODIS data,

71

it is essential to compare and validate the MODIS data with independent ground-based measurements.

72

The spatial, temporal resolution and computation method of both is not same. The ground based

73

AERONET sunphotometer measures AOT directly and no particular aerosol model is considered. The

74

calibration and estimation of τAERONET can be done more reliably as the biases that can alter AOT values

75

are also recorded at AERONET stations.24,25 Several authors have compared between τMODIS and τAERONET

76

to show the seasonal dependence and variations in the correlations. The comparison between ground-

77

based observations and MODIS satellite data retrievals has been carried out by various authors 4, 7, 21, 26-30

78

over India. Mostly an overestimation by MODIS during summer or pre-monsoon, underestimation during

79

winter and post monsoon is reported. 21, 31 Despite the increased proficiency and use of closer-to-realistic

80

models in the retrieval algorithms, several studies have shown that discrepancies still exist between

81

(products) retrieved AOT from satellite over arid, semi-arid regions and urban regions.30,32-33 However, to

82

date a systematic seasonal study of MODIS and ground based observations as done in present study from

83

Jaipur, a semi arid region India is not reported to our knowledge and is an important location for the study

84

of aerosols as its proximity to the Thar desert.34

85

Remote sensing technique is always good for Dark Target (DT) (τMODIS). Since the launch of

86

MODIS, aerosol retrievals over lands uses a DT approach when 2.1 µm surface reflectance is greater than

87

0.15, no aerosol retrievals are performed.30 However, the DT algorithm fails when reflectance is greater

88

than 0.15 (Deserts, snow covered areas). Hsu. et al., 2004(32,33) developed a new algorithm named

89

"Deep Blue (DB) Algorithm" to deal with situations like this. DB algorithm requires polarization

ACS Paragon Plus Environment

4

Environmental Science & Technology

Page 6 of 31

90

correction for Terra Sensor of MODIS. Due to polarization problem of Terra satellite, the only deep blue

91

data available after 2007 is of deep blue aqua AOT 550nm ( http://modis-atmos.gsfc.nasa.gov

92

/products_calendar.html). This method is particularly useful for arid, semi arid and urban areas where

93

AOT retrieval is problematic due to scattering of sun light in the presence of sand particles and greater

94

surface reflectance. 32,33

95

The objective of present study is to validate the MODIS satellite aerosol product with ground

96

based AERONET observations over Jaipur in North Western India. An extensive evaluation and

97

validation has been carried out between the measurements of level 2 Collection Version 5.1 MODIS

98

onboard the Terra and Aqua satellites retrieved product (τMODIS) and the available Deep Blue AOT550nm

99

(τDB550nm) against the seasonal ground based AERONET (2009-12) observations over Jaipur in North

100

Western India and compared with three more sites over India. The study is divided into 5 sections. The

101

section 1, 2 and 3 gives the brief introduction, site description and meteorology, respectively. Section 4

102

describes the details of data used and its procedure of extraction along with introduction to Error

103

Envelope (EE) criterion used for the study. In section 5.1.1 and 5.1.2 of results and analysis, the inter

104

comparison between the τMODIS and τAERONET without and with EE Criterion of MODIS has been done.

105

The τMODIS is also compared with τAERONET at three more ground AERONET stations i.e. Kanpur, Gual

106

Pahari and Pune, India for both criterions. The analysis is further extended in section 5.1.3 to compare the

107

performance of τDB550nm and τAqua550nm with τAERONET for all sites over India seasonally, and then annually

108

with a US desert site at White Sand, Tularosa Basin, New Mexico. This is the first time that seasonal

109

MODIS data has been reported from a site in North western India and inter-compared with more

110

AERONET ground measurement sites over India.

111 112

2. Site Description

113

The measurement station is located at B M Birla Science and Technology Center ,Jaipur (26.9°

114

N, 75.8° E), the capital of Rajasthan state. Jaipur, a hot semi arid region, at eastern boundary of the Thar

115

Desert. This Desert is the major source of dust storm located in western India and Eastern Pakistan for

ACS Paragon Plus Environment

5

Page 7 of 31

Environmental Science & Technology

116

Indian sub-continental during pre-monsoon season. The frequency of occurrence of dust storms is

117

maximum during pre-monsoon season when dust is transported from the Thar desert.35. This site is

118

strategically important due to its proximity to Thar desert.

119 120

The Jaipur AERONET sites (Figure 1)36 chosen for validation purpose of τMODIS, τAqua550nm and τDB550nm against τAERONET observation along with Gual Pahari34 ,Kanpur21 and Pune31.

121

Gual Pahari is an urban background site surrounded by farms and fields, and is located 25 km

122

south of the nearest city, New Delhi. The Delhi pollution and rapid rate of urbanization contributes to

123

aerosols load over Gual Pahari location. This location has sub tropical climate.

124

At Kanpur, the AERONET site is located at Indian Institute of Technology Kanpur campus which

125

is 17 km away from the center of Kanpur .Kanpur is a representative site of the Ganga basin in terms of

126

the weather conditions and atmospheric seasonal variability. This location also has sub tropical climate.21

127 128

At Pune, the site is located at Indian Institute of Tropical Meteorology (IITM), which is 100 km away from west coast of India. It has tropical wet and dry climate.31

129 130

3. Synoptic Meteorology

131

The meteorological parameters for year 2009-12 (Figure. 2) over Jaipur site is retrieved from

132

National Center for Environmental Prediction (NCEP) – National Center for Atmospheric

133

Research (NCAR) reanalysis data. The rainfall values taken from Department of Water

134

Resource, Government of Rajasthan;(http://waterresources.rajasthan.gov.in/Daily_Rainfall_Data

135

/Rainfall_Index.htm).

136 137

According to local meteorology, the seasons has been differentiated in three major classes, viz, dry (December to March), pre-monsoon (April-June), and post-monsoon (October and November).

138

Dry season is generally characterized by decrease in humidity and wind speed. We see in the

139

figure 2 that there is decrease in humidity up-to march and then the humidity starts increasing. The pre-

140

monsoon season in the region is characterized by warm winds originating from western part of Rajasthan.

141

This in turn results in increase in temperature, wind speed and humidity. Monsoon season (July-

ACS Paragon Plus Environment

6

Environmental Science & Technology

Page 8 of 31

142

September) is characterized by decrease in temperature and more increase in humidity as the south west

143

monsoon reaches Rajasthan. During monsoon, region receives scanty to normal rainfall in western and

144

eastern parts respectively. The post monsoon months is characterized by decrease in temperature and low

145

humidity(~40-50%). This continues until the January of dry season after which the temperature gradually

146

starts increasing.

147

4. Data Analysis and Methodology

148

4.1 Aerosol Robotic Network (AERONET)

149

The Aerosol Robotic Network (AERONET) consists of automatic tracking CIMEL Sun photometer/sky

150

radiometers located at more than 500 ground sites around the world currently. The CIMEL sunphotometer

151

is a multi-channel, automatic sun-and-sky scanning radiometer that measures the direct solar irradiance

152

and sky radiance at the Earth's surface only during daylight hours (sun above horizon)

153

application of the cloud screening and quality control procedures described by Smirnov et al. 2000(41).

154

During study period, an AERONET CIMEL sun/sky photometer was operated at the Jaipur ground site.

155

Measurements are taken at pre-determined discrete seven wavelengths in the visible and near-IR parts of

156

the spectrum to determine atmospheric transmission and scattering properties at every fifteen minutes.

157

In this study, in-situ hourly averaged level 2.0 τAERONET data are computed at 0.550 µm to match the

158

spectral resolution of τMODIS.

159

37-40

after

4.2 MODerate resolution Imaging Spectro-radiometer (MODIS)

160

MODIS sensors on board the Earth Observing System (EOS) Terra and Aqua polar-orbiting

161

satellites measures the Earth reflecting solar and emitted terrestrial radiances in 36 high resolution

162

channels from 0.4 – 14.0 µm with a spatial resolution of 250 m, 500 m, 1 km depending on the

163

wavelength. Its large swath of 2330 km and large spectral coverage (36 channels) make it highly suitable

164

for characterization of aerosol properties. 8,42

ACS Paragon Plus Environment

7

Page 9 of 31

165

Environmental Science & Technology

The spatial resolution of MODIS pixel is 10 x 10 km. The overall accuracy of the τMODIS over 15

However, depending on the assumptions on

166

land is estimated to be 15% with a minimum of 0.05.

167

surface types and aerosol optical properties, the accuracy may differ for specific regions of the world.

168

This paper also examines AOT received from observations in blue channel where surface 30

169

reflectance is relatively high

170

products, bias and assumptions, algorithms used in deep blue and standard MODIS products (τMODIS) can

171

be found elsewhere. 9, 15, 30, 32-33, 40, 43-44

172 173

called as Deep Blue AOT 550nm. Details about the MODIS, its retrieval

4.3 Methodology The present paper discusses the validation of level 2 Collection Version 5.1 MODIS retrieved

174

product (τMODIS) and Deep Blue AOT550nm (τDB550nm) against AERONET (2009-12) observations.

175 176 177

4.3.1 Procedure for Extracting AOT

178

The extraction of pixel from MODIS header file is done with HDFlook.45 The AOT derived from

179

MODIS is the spatial average at pixel surface whereas it is of point location for AERONET. The

180

probability of coincidence when the AERONET point location AOT and, AOT derived from center pixel

181

value of MODIS matches is very low.

182

below to get the desired value from MODIS.

183

For spatial window,

184

Length=AERONET_SITE (Latitude) ± SW (Spatial window size in degrees)

185

Width= AERONET_SITE (Longitude) ± SW (Spatial window size in degrees)

186

Where SW (0.05, 0.10, 0.125, 0.25, 0.375, 0.50)

187

Temporal window estimation is done by taking the MODIS overpass scan time ± temporal window size

188

(0.5, 1.0, 2.0, 3.0 hour).

22

So we used different spatial and temporal windows as shown

189

A rectangular box is considered with centre at AERONET site by taking Length and Width. The

190

average pixel values are calculated after considering all the quality checked pixels. Those pixel values

ACS Paragon Plus Environment

8

Environmental Science & Technology

Page 10 of 31

191

which were found within the mean and standard deviations are studied else filtered out. Every Pixel was

192

also sub-sampled to find out the direction. It was found that the most of the pixels lie within North East

193

Direction (Towards Indo-gangetic plain). Larger windows can introduce undesirable results due to surface

194

or aerosol type heterogeneity and also there is a possibility for light cirrus cloud contamination. Smaller

195

windows can results into very less number of pixels as MODIS has a spatial resolution of 10 Km × 10

196

Km. Spatial window of size ±0.05 (10 Km × 10 Km) will only correspond to 1 pixel by 1 pixel, size

197

(±0.10) to 2× 2 pixel (20 Km × 20 Km) and so on. 12 But 2 × 2 is a very small sample for considering

198

the average value of τMODIS.. Moreover, average travel speed of aerosol front is 50 Km/hr. This has been

199

visually estimated by TOMS daily sequence aerosol index images (http://jwocky.gsfc.nasa.gov/aerosols/

200

aermovie.html). So, the 50 Km × 50 Km ( 5 × 5 pixel) or ±0.25 spatial window will match 1-hour

201

sunphotometer data segment well.12 This is also found in our case as most of MODIS observations lies

202

between 1 hour of standard overpass time (10:30 am and 1:30pm IST).

203

The percent change of number of observation when going from spatial window of size ±0.20 to

204

±0.25 is also very large, almost half (~48 %). So spatial windows size of ±0.25 degree is considered

205

overall as it gives better results in terms of correlation, count (Minimum 5 Number of pixels for MODIS

206

and 2 for AERONET), standard deviation, Details are shown in Table S1.

207

All the spatial statistics in the rest of paper are considered based on spatial window size of ±0.25(≈ 50

208

KM × 50 KM) or 5 pixel × 5 pixel averaged over 1 hour time.

209

The same criterion has been followed for extraction of pixels in Deep Blue AOT550nm (τDB550nm).

210 211

For validation purpose, τAERONET (500 nm) has been interpolated to a common wavelength 550 nm using the power law. 28, 31

212

4.3.2 Error Envelope Criterion for inter-comparison of MODIS vs AERONET

213

Since the launch of MODIS it is seen that MODIS measures AOT with an Expected Error or Error

214

Envelope (EE). The previous MODIS validation studies also signify consideration of an EE for τMODIS

215

which is represented by equation (1).

216

AOT – |EE| ≤ τMODIS ≤ AOT + |EE| 15, 16,31

(Equation 1)

ACS Paragon Plus Environment

9

Page 11 of 31

217 218 219 220

Environmental Science & Technology

where and

EE = ± 0.05 ± 0.20× AOT over land EE = ± 0.03 ± 0.05× AOT over ocean

So we have also performed the validation by the direct retrieved AOT from MODIS with EE over Jaipur in section 5.1.1 and 5.1.2, respectively.

221 222

5. Results and Discussions

223

5.1 Statistical Analysis

224

5.1.1 Inter-comparison without Error Envelope Criterion

225

The graphs in Figure 3a show the comparison between τMODIS and τAERONET during the period

226

April 2009 to March 2012 for Jaipur. A significant correlation is found for dry (0.78) and post-monsoon

227

(0.76) when there is almost no dust loading. While a correlation of 0.71 is found for pre-monsoon during

228

heavy dust-loading period. The mean and standard deviation of τ MODIS is 0.53 ±0.33, 0.32 ±0.19 and 0.36

229

±0.28 while τ

230

post-monsoon respectively. The variation in dry season is less compare to pre or post monsoon. Pre and

231

post monsoon retrieval process is often problematic by aerosol in homogeneity especially bright surfaces

232

and cloud contamination.

AERONET

is 0.47 ± 0.17, 0.36 ± 0.15 and 0.40 ±0.24 during pre-monsoon, dry season and

233

Figure 3a also shows that τMODIS overestimates τAERONET during pre-monsoon. More points lie

234

above trend line (red solid line) which shows overestimation (≈59%) during the season. Dust Storms are

235

frequent (source Thar Desert) over Jaipur. The presence of dust particles in the atmosphere reduces the

236

transparency which in turn reduces the reflectance and the τMODIS is overestimated.21 Coarse mode volume

237

concentration also increases during this period as it is a season of dust loading and surface is dry during

238

this time. This may also be the reason for overestimation. Also it is shown in the figure that more points

239

lies below the trend line (blue and green solid line). τMODIS underestimates the τAERONET by 67 % and by 62

240

% approximately during dry and post-monsoon seasons, respectively over Jaipur.

241

overestimates τAERONET during pre-monsoon season where the mean error lies within a range of

ACS Paragon Plus Environment

The τMODIS

10

Environmental Science & Technology

Page 12 of 31

242

+0.06±0.24. MODIS underestimates the τAERONET with a -0.04±0.12 and -0.05±0.18 during the dry season

243

(Dec-Mar) and post-monsoon (Oct-Nov), respectively. All the statistics are significant to 2 decimal places

244

and confidence interval is 95%. Fine mode particles are dominant than coarse mode particles during post

245

monsoon and dry seasons. Surface reflectance is the largest bias during non-dust period. During low dust

246

loading period the atmosphere is more transparent and thus surface reflectance is overestimated which in

247

turn results in underestimation of the τMODIS. The Fractional Bias (FB) for the three seasons are 0.15,-0.11,

248

0.12 for the dry, pre-monsoon and post monsoon, respectively. Negative FB value indicates that MODIS

249

overestimates AERONET whereas positive FB indicates MODIS underestimates AERONET data. The

250

average AOT lies in the range of 0.2-0.6 for all the seasons for both AERONET and MODIS.

251

The τMODIS overestimate τAERONET (Figure 3b) during all the seasons over Gual Pahari. The inter-

252

comparison has been carried out for the duration April 2009 to January 2010 as AERONET site is

253

operational during this time only. More number of points is above trend line. It overestimates with

254

0.10±0.14, 0.10±0.15, 0.06±0.22 for pre-monsoon, post-monsoon and dry season respectively. MODIS

255

overestimates τAERONET for high AE (α=1.1) and underestimate τAERONET for low AE (α=0.8) in most of the

256

cases. The FB for the Dry, Pre and Post seasons are -0.07, -0.16 and -0.15.

257

MODIS overestimate the τAERONET during all the seasons over Kanpur (Figure 3c). It overestimates with

258

0.30±0.23, 0.13±0.19, 0.06±0.14 for pre-monsoon, post-monsoon and dry season respectively. Kanpur

259

site has major industrial activities. Due to the rapid industrialization, the polluted/fine particles and hence

260

the absorption coefficient is large which leads to overestimation of τMODIS. for all the seasons. FB is

261

highest for the Kanpur in all the other sites. The FB for the dry, pre- and post-monsoon seasons are -0.11,

262

-0.41 and -0.16. The work by Tripathi et al 2005 (21) suggests that there is underestimation in post-

263

monsoon (Sept-Nov) and winter (Dec-Feb) and overestimation in pre-monsoon (Mar-May) for the year

264

2004.

265

Pune is less polluted in comparison to Kanpur and Gual Pahari. During the pre-monsoon period

266

MODIS overestimates the τAERONET (0.09±0.14) as coarse mode particles are dominant due to abundant

ACS Paragon Plus Environment

11

Page 13 of 31

Environmental Science & Technology

267

dust loading over Pune (Figure 3d). However, during the post-monsoon (-0.01±0.11) and dry seasons (-

268

0.01 ±0.11), fine mode particles are dominant than coarse mode particles which leads to under-estimation

269

of τMODIS. The FB for the dry, pre- and post- monsoon seasons are 0.09,-0.21 and 0.03 respectively. A

270

good coincidence is shown between MODIS and AERONET observations by More et al., 2013 (31)

271

during pre-monsoon but for winter (Dec-Mar) MODIS underestimation is reported for the same site.

272 273

5.1.2 Inter-comparison with Error Envelope Criterion

274

The correlation is again determined by considering τ MODIS with EE (Equation 1) as suggested by

275

validation studies.19-20 The observations which follows Equation 1 are only considered for statistical

276

analysis.

277

The correlation (0.95, 0.89, 0.94), FB (0.05, 0.00, 0.00), Normalized mean square error i.e.

278

NMSE (0.04,0.03,0.04) between τMODIS and τAERONET with EE improves considerably for dry, pre and

279

post monsoon seasons respectively. Similar procedure has been applied for Kanpur, Pune and Gual

280

Pahari.

281 282

For Jaipur without EE there is overestimation in pre-monsoon and underestimation in other

283

seasons. However, the FB comes very close to 0 and there is underestimation in all seasons namely dry,

284

pre, post (0.05, 0.00, 0.00) when EE criterion is followed. For Pune, there is underestimation in dry

285

(FB=0.01) and overestimation in pre (FB= -0.12) and post season (FB= -0.05). For Gual Pahari and

286

Kanpur, without EE and with EE there is overestimation in all the seasons. The FB for Gual Pahari and

287

Kanpur are -0.08,-0.11,-0.11 and -0.07,-0.16,-0.10 for dry, pre and post monsoon, respectively. The

288

correlations for all the site locations also improved considerably.

289

The average of τMODIS and τAERONET at 550 nm has been shown in Figure 4 over Jaipur, Gual

290

Pahari, Kanpur and Pune, respectively and results have been tabulated. For Jaipur, the average of τMODIS

291

and τAERONET is higher during the pre-monsoon and less during all other seasons. The average AOT i.e

292

τMODIS and τAERONET for all the seasons lies within the range of 0.2-0.6. The average AOT ( τMODIS and

ACS Paragon Plus Environment

12

Environmental Science & Technology

Page 14 of 31

293

τAERONET) lies within the range of 0.4-0.8 for Kanpur and Gual Pahari and 0.2-0.6 for Pune during all the

294

seasons.

295

Figure 5a shows the Taylor diagram without EE and Figure 5b with EE criterion for all the sites.

296

For the purpose of display SD (normalized) values are multiplied by 10 (range is 0.00-0.06) in Figure 5b.

297

After using the EE criterion the SD (normalized) lies in the range of 0.03-0.05 which was earlier 0.1-0.3.

298

For Jaipur, the value of NMSE, MAE and RMSE during pre-monsoon season are 0.24, 0.16 and 0.24,

299

respectively without EE criterion while after applying EE criterion the values comes to 0.031, 0.068 and

300

0.082 respectively. Our results show that the correlation has improved significantly after EE criterion is

301

followed. RMSE error also reduces to 0.05-0.06. Also, the average AOT of Jaipur lies in between 0.2-0.6

302

before and after applying EE criterion i.e Equation 1, which ensures the integrity of data. Inter-

303

comparison of MODIS retrievals with respect to AERONET suggests that the MODIS observations over

304

Jaipur are validated and acceptable. The comparative results between data series give a representative

305

pattern of the uncertainties that needs attention for additional parameterization for present MODIS

306

algorithms.

307 308

Detailed statistical analysis without and with EE is shown in Table 1. Details about the formulas are given in Table S2.

309 310 311 312 313

5.1.3 Inter-comparison with Deep Blue Aqua AOT550nm (τDB550nm) and Aqua AOT

550nm

(τAqua550nm) The correlation between the τMODIS and τAERONET at 550 nm during the pre monsoon season for Jaipur, Gual Pahari, Kanpur and Pune are 0.71, 0.77, 0.72, 0.61 respectively (Figure 3).

314

To find out the performance of deep blue algorithm at these sites, we again compared both

315

τAqua550nm and τDB550nm against τAERONET for all the sites. The correlation between MODIS Deep Blue Aqua

316

AOT550nm (τDB550nm) with τAERONET is found to be 0.75, 0.67, 0.54 and 0.78 within the chosen spatial

317

(±0.25) and temporal window (1 hour average) during pre-monsoon for respective sites. When τAqua550nm

ACS Paragon Plus Environment

13

Page 15 of 31

Environmental Science & Technology

318

is compared with τAERONET the correlation is found to be 0.74, 0.84, 0.66 and 0.75 (Table 2). As per the

319

correlation, for pre-monsoon, τDB550nm performs better for Jaipur and Pune, and for Gual pahari and

320

Kanpur, τAqua550nm performed better. Figure 6(a), 6(b), 7(a) and 7(b) show the scatter plot for τDB550nm and

321

τAqua550nm with τAERONET for Jaipur and Pune respectively.

322

The detailed statistical analysis in terms of correlation, FB and NMSE shows that, for Jaipur and

323

Gual Pahari during the pre-monsoon and post monsoon respectively, τDB550nm performs better then

324

τAqua550nm and for other, situation is reverse. Results for the respective sites are shown in Table 2.

325 326

The result from previous seasonal validation studies on MODIS over India suggests an

327

overestimation during pre-monsoon and underestimations during other seasons over Kanpur, 21

328

India. For Pune

329

and under-estimation in post-monsoon and winter seasons. The present study found

330

overestimation during all seasons in Kanpur and Gual Pahari but Pune and Jaipur has

331

overestimation only in pre-monsoon and underestimation during post-monsoon and dry seasons.

332

Furthermore, as the Jaipur site is a semi arid region, we chose to compare the

333

performance of τDB550nm and τAqua550nm with τAERONET for considered sites over India along with a US desert

334

site for the entire duration (2009-12) to check the consistency of our results. This US location is situated

335

in New Mexico at the White Sands Missile Range (WSMR) High Energy Laser Systems Test Facility

336

(HELSTF). This region of South Central Mexico is high desert with annual precipitation around 20 cm

337

and 350 days sunshine. Strong seasonal winds are very frequent during February to May. So dust aerosols

338

along with gypsum from nearby place can cause frequent dust storms. The annual inter-comparison

339

results over Jaipur, Gual Pahari and White Sand HELSTF (US site) suggests that τDB550nm performs better

340

than τAqua550nm in terms of correlation, root mean square error (RMSE) and FB. Our annual validation

341

results contradicts with recent study by Bibi et al. (46) over Jaipur however agrees in case of Kanpur and

31

previous inter-comparison study suggests a good match during pre-monsoon

ACS Paragon Plus Environment

annual

14

Environmental Science & Technology

342

Pune i.e. τAqua550nm performs better than τ

343

presented in Table 3.

DB550nm.

Page 16 of 31

The details of annual inter-comparison results are

344 345

As per the overall annual performance is considered, for Jaipur, Gual Pahari and White Sand (US

346

site) τDB550nm with correlation (0.79,0.80,0.58), fractional bias (0.08, -0.03 and 0.26), RMSE

347

(0.22,0.22,0.03) performed better than τAqua550nm with correlation (0.78,0.80,0.62), fractional bias (0.13, -

348

0.12 and -1.34), RMSE (0.22,0.22,0.03) respectively. For Kanpur and Pune, τAqua550nm performed better

349

statistical results. Thus, our recommendation is to use τDB550nm for Jaipur, Gaul Pahari and White Sand

350

sites and τAqua550nm for Kanpur and Pune.

351 352

As per the seasons, for pre-monsoon season for Jaipur, τDB550nm has to be used where τAqua550nm

353

faces difficulty in correct retrieval of aerosol. For other seasons and remaining sites standard MODIS

354

product (τAqua550nm) produce good statistical results. Though this also shows there is further improvement

355

that needs to be done in deep blue algorithm for semi-arid, arid and urban regions so that better

356

characteristics can be obtained. Results obtained in this work can also aid in improvement of MODIS

357

aerosols products that need to be done in updated version of algorithms.

358

359

Acknowledgments

360

We would like to thank the Editor and anonymous reviewers for suggestions that helped in

361

improving substantially the presentation of the revised manuscript. MODIS data were obtained

362

from the Level 2 and Atmosphere Archive and Distribution System (LAADS) at Goddard Space Flight

363

Center (GSFC), (http://ladsweb.nascom.nasa.gov/data/). We also acknowledge the MODIS mission

364

scientists and AERONET group, PI’s of respective site, for the production of the data used in this

365

research effort. We also acknowledge NCEP for providing the reanalysis data.

ACS Paragon Plus Environment

15

Page 17 of 31

Environmental Science & Technology

366

Supporting Information Available

367

The Supplementary information contains 2 tables on a total of 4 pages. Spatial and Temporal window

368

selection criterion for MODIS pixels extraction against AERONET observations for Jaipur is illustrated

369

in Table S1. Spatial window size

370

size(Averaged 0.5, 1.0, 2.0, 3.0 ) hr are considered for this evaluation.Table S2 contains the formulas

371

used as performances measures in this study. This information is available free of charge via the Internet

372

at http://pubs.acs.org/ .

±(0.05, 0.10, 0.125, 0.25, 0.375, 0.50) and Temporal window

373 374

References

375

1. Schwartz, S. E., Arnold, F., Blanchet, J. P., Durkee, P. A., Hofmann, D. J., Hoppel, W. A., ... &

376

Wendisch, M. (1995). Group report: Connections between aerosol properties and forcing of

377

climate. Aerosol forcing of climate, 251-280.

378

2. Lemke, P., Ren, R. B., & Alley, I. (2007). The physical science basis. Contribution of Working

379

Group I to the fourth assessment report of the Intergovernmental Panel on Climate

380

Change. Climate Change 2007, 337-383.

381 382

3. Kaufman, Y. J., Tanré, D., & Boucher, O. (2002). A satellite view of aerosols in the climate system. Nature, 419(6903), 215-223.

383

4. Misra, A., Jayaraman, A., & Ganguly, D. (2008). Validation of MODIS derived aerosol optical

384

depth over Western India. Journal of Geophysical Research: Atmospheres (1984–

385

2012), 113(D4).

386

5. King, M. D., Kaufman, Y. J., Tanré, D., & Nakajima, T. (1999). Remote sensing of tropospheric

387

aerosols from space: Past, present, and future, 80, 2229–2259; DOI 10.1175/1520-

388

0477(1999)0802.0.CO;2, 1999.

ACS Paragon Plus Environment

16

Environmental Science & Technology

389

Page 18 of 31

6. Bellouin, N., Boucher, O., Haywood, J., & Reddy, M. S. (2005). Global estimate of aerosol direct radiative forcing from satellite measurements.Nature, 438(7071), 1138-1141.

390

391

7. Choudhry, P., Misra, A., & Tripathi, S. N. (2012, October). Study of MODIS derived AOD at

392

three different locations in the Indo Gangetic Plain: Kanpur, Gandhi College and Nainital.

393

In Annales Geophysicae (Vol. 30, No. 10, pp. 1479-1493). Copernicus GmbH.

394

8.

Levy, R. C., Remer, L. A., Kleidman, R. G., Mattoo, S., Ichoku, C., Kahn, R., & Eck, T. F.

395

(2010). Global evaluation of the Collection 5 MODIS dark-target aerosol products over

396

land. Atmospheric Chemistry and Physics, 10(21), 10399-10420.

397

9. Kaufman, Y. J., Wald, A. E., Remer, L. A., Gao, B. C., Li, R. R., & Flynn, L. (1997). The

398

MODIS 2.1-µm channel-correlation with visible reflectance for use in remote sensing of

399

aerosol. Geoscience and Remote Sensing, IEEE Transactions on, 35(5), 1286-1298.

400

10. Chu, D. A., Kaufman, Y. J., Remer, L. A., & Holben, B. N. (1998). Remote sensing of smoke

401

from MODIS airborne simulator during the SCAR‐B experiment. Journal of Geophysical

402

Research: Atmospheres (1984–2012),103(D24), 31979-31987.

403

11. Ichoku, C., Chu, D. A., Mattoo, S., Kaufman, Y. J., Remer, L. A., Tanré, D., ... & Holben, B.

404

N. (2002). A spatio‐temporal approach for global validation and analysis of MODIS aerosol

405

products. Geophysical Research Letters, 29(12), MOD1-1.

406

12. Chu, D. A., Kaufman, Y. J., Ichoku, C., Remer, L. A., Tanré, D., & Holben, B. N. (2002).

407

Validation of MODIS aerosol optical depth retrieval over land. Geophysical research

408

letters, 29(12), MOD2-1.

409

13. Santese, M., De Tomasi, F., & Perrone, M. R. (2007). AERONET versus MODIS aerosol

410

parameters at different spatial resolutions over southeast Italy. Journal of Geophysical

411

Research: Atmospheres (1984–2012), 112(D10).

ACS Paragon Plus Environment

17

Page 19 of 31

Environmental Science & Technology

412

14. Tanré, D., Kaufman, Y. J., Herman, M., & Mattoo, S. (1997). Remote sensing of aerosol

413

properties over oceans using the MODIS/EOS spectral radiances.Journal of Geophysical

414

Research: Atmospheres (1984–2012), 102(D14), 16971-16988.

415

15. Remer, L. A., Tanre, D., Kaufman, Y. J., Ichoku, C., Mattoo, S., Levy, R., ... & Ahmad, Z.

416

(2002). Validation of MODIS aerosol retrieval over ocean.Geophysical research letters, 29(12),

417

MOD3-1.

418

16. Ichoku, C., Remer, L. A., Kaufman, Y. J., Levy, R., Chu, D. A., Tanré, D., & Holben, B. N.

419

(2003). MODIS observation of aerosols and estimation of aerosol radiative forcing over

420

southern Africa during SAFARI 2000. Journal of Geophysical Research: Atmospheres (1984–

421

2012), 108(D13).

422

17. Chu, D. A., Kaufman, Y. J., Zibordi, G., Chern, J. D., Mao, J., Li, C., & Holben, B. N. (2003).

423

Global monitoring of air pollution over land from the Earth Observing System-Terra Moderate

424

Resolution

425

Atmospheres (1984–2012),108(D21).

426 427

Imaging

Spectroradiometer

(MODIS). Journal

of

Geophysical

Research:

18. Ichoku, C., Kaufman, Y. J., Remer, L. A., & Levy, R. (2004). Global aerosol remote sensing from MODIS. Advances in Space Research, 34(4), 820-827.

428

19. Remer, L. A., Kaufman, Y. J., Tanré, D., Mattoo, S., Chu, D. A., Martins, J. V., ... & Holben,

429

B. N. (2005). The MODIS aerosol algorithm, products, and validation. Journal of the

430

atmospheric sciences, 62(4), 947-973.

431

20. Levy, R. C., Remer, L. A., Martins, J. V., Kaufman, Y. J., Plana-Fattori, A., Redemann, J., &

432

Wenny, B. (2005). Evaluation of the MODIS aerosol retrievals over ocean and land during

433

CLAMS. Journal of the Atmospheric Sciences,62(4), 974-992.

ACS Paragon Plus Environment

18

Environmental Science & Technology

Page 20 of 31

434

21. Tripathi, S. N., Dey, S., Chandel, A., Srivastava, S., Singh, R. P., & Holben, B. N. (2005, June).

435

Comparison of MODIS and AERONET derived aerosol optical depth over the Ganga Basin,

436

India. In Annales Geophysicae (Vol. 23, No. 4, pp. 1093-1101).

437

22. Retalis, A., Hadjimitsis, D. G., Michaelides, S., Tymvios, F., Chrysoulakis, N., Clayton, C. R.,

438

& Themistocleous, K. (2010). Comparison of aerosol optical thickness with in situ visibility

439

data over Cyprus. Natural Hazards and Earth System Science, 10(3), 421-428.

440

23. He, Q., Li, C., Tang, X., Li, H., Geng, F., & Wu, Y. (2010). Validation of MODIS derived

441

aerosol optical depth over the Yangtze River Delta in China. Remote Sensing of

442

Environment, 114(8), 1649-1661.

443

24. Dubovik, O., Smirnov, A., Holben, B. N., King, M. D., Kaufman, Y. J., Eck, T. F., & Slutsker,

444

I. (2000). Accuracy assessments of aerosol optical properties retrieved from Aerosol Robotic

445

Network (AERONET) Sun and sky radiance measurements. Journal of Geophysical Research:

446

Atmospheres (1984–2012),105(D8), 9791-9806.

447

25. Kumar, N., Chu, A. D., Foster, A. D., Peters, T., & Willis, R. (2011). Satellite remote sensing

448

for developing time and space resolved estimates of ambient particulate in Cleveland,

449

OH. Aerosol Science and Technology, 45(9), 1090-1108.

450 451

26. Jethva, H., Satheesh, S. K., & Srinivasan, J. (2005). Seasonal variability of aerosols over the Indo-Gangetic basin. Journal of Geophysical Research: Atmospheres (1984–2012), 110(D21).

452

27. Jethva, H., Satheesh, S. K., & Srinivasan, J. (2007). Evaluation of Moderate‐Resolution

453

Imaging Spectroradiometer (MODIS) Collection 004 (C004) aerosol retrievals at Kanpur,

454

Indo‐Gangetic Basin. Journal of Geophysical Research: Atmospheres (1984–2012), 112(D14).

ACS Paragon Plus Environment

19

Page 21 of 31

Environmental Science & Technology

455

28. Prasad, A. K., & Singh, R. P. (2007). Comparison of MISR-MODIS aerosol optical depth over

456

the Indo-Gangetic basin during the winter and summer seasons (2000–2005). Remote Sensing

457

of Environment, 107(1), 109-119.

458

29. Jethva, H., Satheesh, S. K., Srinivasan, J., & Levy, R. C. (2010). Improved retrieval of aerosol

459

size‐resolved properties from moderate resolution imaging spectroradiometer over India: Role of

460

aerosol model and surface reflectance. Journal of Geophysical Research: Atmospheres (1984–

461

2012), 115(D18).

462

30. Misra, A., Jayaraman, A., & Ganguly, D. (2014). Validation of Version 5.1 MODIS Aerosol

463

Optical Depth (Deep Blue Algorithm and Dark Target Approach) over a Semi-Arid Location in

464

Western India. Aerosol and Air Quality Research.

465

31. More, S., Kumar, P. P., Gupta, P., Devara, P. C. S., & Aher, G. R. (2013). Comparison of

466

aerosol products retrieved from AERONET, MICROTOPS and MODIS over a tropical urban

467

city, Pune, India. Aerosol and Air Quality Research, 13(1), 107-121.

468

32. Hsu, N. C., Tsay, S. C., King, M. D., & Herman, J. R. (2004). Aerosol properties over

469

bright-reflecting source regions. Geoscience and Remote Sensing, IEEE Transactions

470

on, 42(3), 557-569.

471

33. Hsu, N. C., Jeong, M. J., Bettenhausen, C., Sayer, A. M., Hansell, R., Seftor, C. S., ... &

472

Tsay, S. C. (2013). Enhanced Deep Blue aerosol retrieval algorithm: The second

473

generation. Journal of Geophysical Research: Atmospheres, 118(16), 9296-9315.

474

34. Verma, S., Prakash, D., Ricaud, P., Payra, S., Attié, J. L., & Soni, M. (2015). A New

475

Classification of Aerosol Sources and Types as Measured over Jaipur, India. Aerosol and Air

476

Quality Research, 15(3), 985-993, doi: 10.4209/aaqr.2014.07.0143.

ACS Paragon Plus Environment

20

Environmental Science & Technology

Page 22 of 31

477

35. Verma, S., Payra, S., Gautam, R., Prakash, D., Soni, M., Holben, B., & Bell, S. (2013). Dust

478

events and their influence on aerosol optical properties over Jaipur in Northwestern

479

India. Environmental monitoring and assessment, 185(9), 7327-7342.

480

36. http://censusindia.gov.in/2011census/maps/administrative_maps/admmaps2011.html

481

State Map ,Census of India 2011,Office of The Registrar General & Census Commissioner,

482

India New Delhi, Ministry of Home Affairs, Government of India

483

37. Holben, B. N., Eck, T. F., Slutsker, I., Tanre, D., Buis, J. P., Setzer, A., ... & Smirnov, A.

484

(1998). AERONET—A federated instrument network and data archive for aerosol

485

characterization. Remote sensing of environment, 66(1), 1-16.

486

38. Holben, B. N., Tanre, D., Smirnov, A., Eck, T. F., Slutsker, I., Abuhassan, N., ... & Zibordi, G.

487

(2001). An emerging ground-based aerosol climatology: Aerosol optical depth from

488

AERONET. Journal of Geophysical Research: Atmospheres (1984–2012), 106(D11), 12067-

489

12097.

490

39. Eck, T. F., Holben, B. N., Dubovik, O., Smirnov, A., Slutsker, I., Lobert, J. M., & Ramanathan,

491

V. (2001). Column‐integrated aerosol optical properties over the Maldives during the

492

northeast monsoon for 1998–2000. Journal of Geophysical Research: Atmospheres (1984–

493

2012), 106(D22), 28555-28566.

494

40. Dubovik, O., Holben, B., Eck, T. F., Smirnov, A., Kaufman, Y. J., King, M. D., ... & Slutsker,

495

I. (2002). Variability of absorption and optical properties of key aerosol types observed in

496

worldwide locations. Journal of the atmospheric sciences, 59(3), 590-608.

497

41. Smirnov, A., Holben, B. N., Eck, T. F., Dubovik, O., & Slutsker, I. (2000). Cloud-screening

498

and

quality

control

algorithms

499

Environment, 73(3), 337-349.

for

the

AERONET

database.Remote

Sensing

of

500

42. King, M. D., Menzel, W. P., Kaufman, Y. J., Tanré, D., Gao, B. C., Platnick, S., ... & Hubanks,

501

P. A. (2003). Cloud and aerosol properties, precipitable water, and profiles of temperature and

ACS Paragon Plus Environment

21

Page 23 of 31

Environmental Science & Technology

502

water vapor from MODIS. Geoscience and Remote Sensing, IEEE Transactions on, 41(2), 442-

503

458.

504

43. Levy, R. C., Remer, L. A., Mattoo, S., Vermote, E. F., & Kaufman, Y. J. (2007).

505

Second‐generation operational algorithm: Retrieval of aerosol properties over land from

506

inversion of Moderate Resolution Imaging Spectroradiometer spectral reflectance. Journal of

507

Geophysical Research: Atmospheres (1984–2012), 112(D13).

508

44. Levy, R. C., Remer, L. A., & Dubovik, O. (2007). Global aerosol optical properties and

509

application to Moderate Resolution Imaging Spectroradiometer aerosol retrieval over

510

land. Journal of Geophysical Research: Atmospheres (1984–2012), 112(D13).

511

45. Gonzalez, L., & Deroo, C. (2003). HDFLook/HDFLook MODIS Handbook. Laboratoire

512

d’Optique Atmosph! erique, Universit! e des Science et Technologies de Lille, France. Revised

513

28 April, 2003. http://www-loa.univ-lille1.fr/Hdflook/E_HDF.html.

514

46. Bibi, H., Alam, K., Chishtie, F., Bibi, S., Shahid, I., & Blaschke, T. (2015). Intercomparison of

515

MODIS, MISR, OMI, and CALIPSO aerosol optical depth retrievals for four locations on the

516

Indo-Gangetic plains and validation against AERONET data. Atmospheric Environment, 111,

517

113-126.

518

519 520 521 522 523

ACS Paragon Plus Environment

22

Environmental Science & Technology

Gual Pahari Jaipur

Kanpur

Pune

Figure 1. AERONET Site(Jaipur,Gual Pahari,Kanpur,Pune) used in the study.

ACS Paragon Plus Environment

Page 24 of 31

Page 25 of 31

Environmental Science & Technology

Figure 2. Synoptic Meteorology for year 2009-2012 over Jaipur, Northwestern India. 3(a)

ACS Paragon Plus Environment

Environmental Science & Technology

3(b)

3(c)

ACS Paragon Plus Environment

Page 26 of 31

Page 27 of 31

Environmental Science & Technology

3(d)

Figure 3. τ

MODIS

vs τ

AERONET

for (a) Jaipur (b) Gual Pahari, (c) Kanpur (d) Pune (AERONET Sites).

MODIS values have been averaged around ±0.25 latitude x ±0.25 longitude box centered over the corresponding location. Black Dash lines indicates AOD+|EE| and AOD- |EE| where EE = ± 0.05 ± 0.20*AOT and Blue dash line shows 1:1 line. Blue, Red and Green solid line shows trend line for Dry, pre-monsoon, post-monsoon respectively The points below the trend line shows underestimation and above shows overestimation by τ MODIS when compared with τ AERONET .

Figure 4. Average AOT during different seasons over Gual Pahari, Jaipur, Kanpur and Pune, respectively.

ACS Paragon Plus Environment

Environmental Science & Technology

5(a)

Page 28 of 31

5(b)

Figure 5. Taylor diagrams for all the sites a) without EE criterion b) with EE criterion

6(a)

6(b)

Figure 6(a) Deep blue AOT 550nm Vs AERONET AOT550nm (b) Aqua AOT 550nm Vs AERONET AOT550nm For Jaipur

ACS Paragon Plus Environment

Page 29 of 31

Environmental Science & Technology

7(b) Figure 7(a) Deep blue AOT 550nm Vs AERONET AOT550nm (b) Aqua AOT 550nm Vs AERONET AOT550nm For Jaipur

ACS Paragon Plus Environment

Environmental Science & Technology

τ R

   

 

JAIPUR

PRE-MONSOON POST MONSOON

DRY SEASON

     

 

GUAL_PAHARI

PRE-MONSOON

     

POST MONSOON

DRY SEASON

MODIS

Vs

τ

Page 30 of 31

AERONET

FB

FA2

NMSE

W/O EE

W/ EE

W/O EE

W/ EE

W/O EE

W/ EE

W/O EE

W/ EE

0.78 0.71

0.95 0.89

0.15 -0.11

0.05 0.00

87.06 84.35

100.00 100.00

0.17 0.24

0.04 0.03

0.76

0.94

0.12

0.00

81.40

100.00

0.24

0.04

0.72 0.76

0.92 0.87

-0.07 -0.16

-0.08 -0.11

91.67 97.10

100.00 100.00

0.10 0.07

0.03 0.03

0.84

0.95

-0.15

-0.11

97.73

100.00

0.07

0.03

0.84 0.70

0.94 0.93

-0.11 -0.41

-0.07 -0.16

97.92 83.78

100.00 100.00

0.08 0.27

0.03 0.05

0.81

0.94

-0.16

-0.10

98.44

100.00

0.09

0.03

0.72 0.61

0.92 0.86

0.09 -0.21

0.01 -0.12

95.22 96.71

100.00 100.00

0.11 0.16

0.03 0.04

0.78

0.91

0.03

-0.05

98.55

100.00

0.10

0.04

  DRY SEASON

KANPUR

PRE-MONSOON POST MONSOON

     

  DRY SEASON

PUNE

PRE-MONSOON POST MONSOON

 

Table 1. Detailed Statistics Without(W/O) and With(W/ ) Error Envelope criterion.

Where FB

Fractional Bias

NMSE Normalized mean square error

FA2

Factor of 2 observations

R

Correlation

ACS Paragon Plus Environment

Page 31 of 31

Environmental Science & Technology

τAERONET

Vs CORRELATION

τAQUA550nm τDB550nm τAQUA550nm τDB550nm τAQUA550nm τDB550nm τAQUA550nm τDB550nm

DRY PREMONSOON POST MONSOON

0.79 0.74 0.83

0.74 0.75 0.85

   

   

0.18 0.13 0.04

0.43 -0.17 0.09

   

   

0.17 0.22 0.42

0.45 0.17 0.60

JAIPUR

FRACTIONAL BIAS DRY PREMONSOON POST MONSOON NMSE DRY PREMONSOON POST MONSOON

GUAL_PAHARI 0.78 0.84 0.87

KANPUR

0.73 0.67 0.89

0.87 0.66 0.84

PUNE

0.80 0.54 0.81

0.80 0.75 0.87

0.76 0.78 0.78

    -0.04 -0.12 -0.10

-0.12 -0.03 -0.02

0.01 -0.20 -0.10

0.14 -0.42 -0.03

0.06 -0.18 -0.02

0.19 -0.31 0.24

    0.08 0.07 0.06

0.15 0.10 0.08

0.05 0.19 0.07

0.34 0.37 0.18

0.09 0.08 0.04

0.19 0.16 0.25

Table 2. Detailed Statistics of MODIS Aqua AOT (τAqua550nm) and MODIS Deep Blue Aqua AOT550nm (τDB550nm) against AERONET AOT 550nm(τAERONET)

JAIPUR

τAERONET Vs

GUAL_PAHARI

KANPUR

PUNE

WHITE SAND

τAQUA550nm

τDB550nm

τAQUA550nm

τDB550nm

τAQUA550nm

τDB550nm

τAQUA550nm

τDB550nm

τAQUA550nm

τDB550nm

CORRELATION

0.78

0.79

0.80

0.80

0.78

0.69

0.77

0.69

0.62

0.58

FRACTIONAL BIAS

0.13

0.08

-0.12

-0.03

-0.12

-0.14

-0.03

0.02

-1.34

0.26

RMSE

0.19

0.22

0.19

0.22

0.23

0.36

0.11

0.16

0.22

0.03

Table 3. Detailed Annual Statistics (2009-12) of MODIS Aqua AOT (τAqua550nm) and MODIS Deep Blue Aqua AOT550nm (τDB550nm) against AERONET AOT 550nm (τAERONET)

 

ACS Paragon Plus Environment