Performance of the ozone mapping and profiler suite (OMPS) products

L. Flynn, C. Long, X. Wu, R. Evans, C.T. Beck NOAAI. Petropavlovskikh, G. McConville CIRESW. Yu, Z. Zhang, J. Niu, E. Beach, Y. Hao IMSGC. Pan, UMD CICSB. Sen, M. Novicki Northrop Grumman Aerospace SystemsS. Zhou, Wyle Information SystemsC. Seftor, SSAICorresponding author: Lawrence E. Flynn, NOAA, NCWCP E/RA2, 5830 University Research Court, #2850, College Park, MD 20740-3818 (Lawrence.E.Flynn@noaa.gov)

Key Points• OMPS nadir total column ozone and ozone profile products are performing well • Deficiencies have been identified but the remedies are known• Further improvements to the SDR will be implemented in the system soon

Index Terms3360 Remote Sensing0340 Middle atmosphere: composition and chemistry3311 Clouds and aerosols “Manuscript Submitted for the Special Issue of AGU JGR-Atmospheres on Suomi NPP Cal/Val Science Results”. 

11

1

23456789

10

11

12

13

14

15

16

17

18

1920

Abstract. NOAA, through the Joint Polar Satellite System (JPSS) program, in partnership with National Aeronautical

and Space Administration (NASA), launched the Suomi National Polar-orbiting Partnership (S-NPP) satellite, a risk

reduction and data continuity mission, on October 28, 2011. The JPSS program is executing the S-NPP Calibration and

Validation (Cal/Val) program to ensure the data products comply with the requirements of the sponsoring agencies.

The Ozone Mapping and Profiler Suite (OMPS) consists of two telescopes feeding three detectors measuring solar

radiance scattered by the Earth's atmosphere directly and solar irradiance by using diffusers. The measurements are

used to generate estimates of total column ozone and vertical ozone profiles for use in near-real time applications and

extension of ozone climate data records. The calibration and validation efforts are progressing well, and both Level 1

(Sensor Data Records/SDRs) and Level 2 (Ozone Environmental Data Records/EDRs) have advanced to release at

Provisional Maturity. This paper provides information on the product performance over the first 22 months of the

mission. The products are evaluated through the use of internal consistency analysis techniques and comparisons to

other satellite instrument and ground-based products. The initial performance finds total ozone showing negative bias

of 2 to 4% with respect to correlative products and ozone profiles often within ±5% in the middle and upper

stratosphere of current operational products. Potential improvements in the measurements and algorithms are identified.

These will be implemented in coming months to reduce the differences further.

22

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

1. Introduction

The OMPS is composed of three instruments, called the Nadir Mapper (OMPS-NM), Nadir Profiler

(OMPS-NP), and Limb Profiler (OMPS-LP). The OMPS-NM is a total ozone column sensor and uses a

single grating and a Charge-Coupled Device (CCD) array detector to make measurements every 0.42 nm

from 300 nm to 380 nm with 1.0-nm Full-Width Half-Maximum (FWHM) resolution. It has a 110°

cross-track FOV (~2800 km on the Earth’s surface) and 0.27° along-track slit width FOV. In standard

Earth science mode, the measurements are combined into 35 cross-track bins [20 spatial pixels giving

3.35° (50 km) at nadir, and 2.84° at ±55° cross-track dimensions for the FOVs]. The resolution is 50 km

along-track at nadir, created by using a 7.6-second reporting/integration period. This resolution choice is

changeable; and the use of smaller FOVs and shorter integration times is under investigation.

The OMPS-NP is an ozone profile sensor and uses a double monochromator and a CCD array detector

to make measurements every 0.42 nm from 250 nm to 310 nm with 1.1-nm resolution. It has a 16.6°

cross-track FOV, 0.26° along-track slit width. The measurements are combined (100 spatial pixels) into

a single spectrum and the reporting period is 38 seconds giving it a 250 km × 250 km cell size collocated

with the five central OMPS-NM FOVs.

The OMPS-LP is a high vertical resolution ozone profile sensor and uses a prism spectrometer with

spectral coverage from 290 nm to 1000 nm. It has three slits separated by 4.25° (viewing back along the

nadir track and 250 km cross-track on either side at the tangent point) with a 19-second reporting period

that equates to 125 km along-track motion. The slits have 112 km (1.95°) vertical FOVs equating to 0 to

60 km coverage at the limb, plus offsets for pointing uncertainty, orbital variation, and Earth oblateness.

The CCD array detector provides measurements every 1.1 km with 2.1-km vertical resolution. There are

two apertures into the instrument creating a total of six images on the CCD. Further the CCD array is

read out with long and short integration times to produce four effective gains.

33

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

The OMPS instruments and measurements are described further in Rodriguez et al. [2003] and

Remund et al. [2004] and in three companion papers within this special issue: Seftor et al. [2013], Jaross

et al. [2013] and Wu et al. [2013]. Of note is their use of working and reference solar diffusers to

monitor instrument throughput degradation and maintain good calibration of the radiance/irradiance

ratios used in the retrieval algorithms. Key updates to the SDRs are discussed in the companion papers.

Of special interest for the OMPS-NM products are the implementation of a stray light correction,

identification of intra-orbit wavelength scale variations, and estimation of calibration offset for OMPS

through vicarious and internal consistency checks. For the OMPS-NP the key items are the same as for

the OMPS-NM with an additional refinement to the SDR needed to account for solar activity at the time

of the measurements relative to that for the Day One solar spectrum.

The OMPS-NM and -NP measurements are processed into total column ozone and ozone profile

products in near-real-time (NRT) at the NOAA Interface Data Processing Segment (IDPS) and

distributed for use in numerical weather models and to assist in forecasting daily ultraviolet (UV) Index

values. They are also processed off-line with other retrieval algorithms as a first step in the process of

incorporating them into long-term ozone climate data records (CDRs). This paper will discuss results for

both types of products but will emphasize the operational ones. The OMPS-LP measurements are

processed into ozone profile products by the NASA S-NPP Science Team. The OMPS-LP algorithms

are moving toward implementation in the NOAA operations but that topic is not covered in this paper.

2. Ozone Retrieval Algorithms

The OMPS measurements are used to generate estimates of atmospheric ozone in both operational and

off-line systems. This paper will concentrate on the on the products from the NOAA IDPS derived from

the OMPS-NM and NP measurements but will also present some information on some preliminary

results using the Version 8 total ozone and profile ozone algorithms to generate climate data records

44

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

from the OMPS-NM and NP taking place at NASA and NOAA. A brief introduction of the algorithms

appears below. More detailed descriptions of the IDPS algorithms can be found in JPSS Algorithm

Theoretical Basis Documents (ATBDs), Operational Algorithm Documents (OADs) and Data Format

Control Books (DFCB) available at the S-NPP Library: npp.gsfc.nasa.gov/science/documents.html.

A description of research algorithms for the OMPS Limb Profiler is given in Rault and Loughman

[2013].

2.1. Nadir Mapper Total Column Ozone

The spectral measurements from the OMPS-NM of the radiances scattered by the Earth’s atmosphere

are used to generate estimates of the total column ozone. The IDPS algorithm uses ratios of Earth

radiance to Day 1 Solar irradiance measurements at triplets of wavelengths to obtain estimates of the

total column ozone, effective reflectivity, and the wavelength dependence of the reflectivity as affected

by aerosols. [The Day 1 Solar spectra are a set of spectra used in the algorithm to compute the

radiance/irradiance ratios. The algorithm/processing initially used a set of pre-launch estimates for these

and then after real solar measurements were taken new estimates were made (system updates in 6/2012

for the OMPS NM and 7/2012 for the OMPS NP). The working diffuser is used every two weeks and

the reference diffuser is used every six months to make new measurements. So far the reference diffuser

measurements have shown little or no changes above 290 nm and changes from -0.1%/year to -0.5/year

growing larger in magnitude from 290 nm to 250 nm. The processing system has daily values for Earth

Calibration Factors that can be used to adjust for instrument throughput changes relative to the Day 1

values but they have not yet been activated. More information on the instruments performance is in the

companion papers.] Table values computed for a set of standard profiles, cloud heights, latitudes and

solar zenith angles are interpolated and compared to the measured top-of-atmosphere albedos. The

triplets combine an ozone insensitive wavelength channel (at 364, 367, 372 or 377 nm) to obtain cloud

55

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

fraction and reflectivity information, with a pair of measurements at shorter wavelengths. The pairs are

selected to have one “weak” and one “strong” ozone absorption channel. The hyperspectral capabilities

of the sensor are used to select multiple sets of triplets to balance ozone sensitivity across the range of

expected ozone column amounts and solar zenith angles. The "strong" ozone channels are placed at

308.5, 310.5, 312.0, 312.5, 314.0, 315.0, 316.0, 317.0, 318.0, 320.0, 322.5, 325.0, 328.0, or 331.0 nm.

They are paired with a longer “weak” channel at 321.0, 329.0, 332.0, or 336.0 nm. The ozone absorption

cross-sections decrease from 3 (milli-atm. cm)-1 to 0.3 (milli-atm. cm)-1 over the range of “strong”

wavelengths. Typical ozone columns range from 100 Dobson Units (1 DU = 1 milli-atm-cm) to 600 DU.

The Multiple Triplet algorithm is applied twice for each FOV. This was done to resolve the “Who

goes first?” quandary created by the desires to use information from other sensors in the retrieval

algorithms, e.g., OMPS wanted to use the Cross-track Infrared Sounder (CrIS) temperature profile

estimates, and CrIS wanted to use the OMPS ozone estimates. The “1 st Guess” OMPS-NM products use

climatological or forecast fields for surface reflectivity and pressure, snow/ice coverage, cloud optical

centroid depth, and atmospheric temperature. They use internally calculated estimates of cloud fractions

and effective reflectivity from measurements at non-ozone absorbing UV wavelengths. This product is

sometimes called the Total Ozone Intermediate Product (TOZ IP).

The 2nd Pass or EDR OMPS products use snow/ice coverage from the Visible Infrared Imaging

Radiometer Suite (VIIRS) near-real-time products and temperature profiles from CrIS products. Both

products use internally calculated estimates of cloud fractions and effective reflectivity from

measurements at non- or weakly absorbing ozone wavelengths. This second product is sometimes called

the Total Ozone Environmental Data Record (TOZ EDR). As we will show, both applications of the

algorithm are performing well but have room for improvement.

66

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

It was originally planned to use IR measurements to provide the cloud top pressure data for the EDR

product but with the development of UV-measurement-based estimates by the EOS Aura Ozone

Monitoring Instrument (OMI) Science Team as described in Vasilikov et al.[ 2008] these were found to

often (e.g., for thin cirrus) lead to errors as the IR optical cloud tops would be found at a lower pressure

than the UV optical cloud tops. Both OMPS total ozone products are currently using a climatology of

UV cloud optical centroids compiled from five years of OMI data. Future plans are to include direct

calculation of UV cloud optical centroid pressures from the OMPS measurements.

The total ozone product files contain both Multiple Triplet and Heritage Version 7 Triplet Total

Column Ozone estimates. See McPeters et al., [1996] for information on the Version 7 algorithm. The

heritage Version 7 products use the 318, 331 and 364 nm channels in a single triplet. Given the current

state of the calibration, each triplet will have its own small biases. We are working on soft calibration

adjustments to remove the inter-channel biases and homogenize the multiple triplet results. The nice

thing about the Version 7 product is that its behavior is only affected by the relative calibration of the

three channels in its single triplet. Its weakness is that this triplet is not ideal under all viewing

conditions, particularly for high ozone amounts at large solar zenith or satellite viewing angles.

The Version 8 total ozone algorithm used for reprocessing is described in Bhartia and Wellemeyer

(2002).

2.2. Nadir Profiler

The spectral measurements from the OMPS Nadir Profiler and Nadir Mapper of the radiances scattered

by the Earth’s atmosphere are used to generate estimates of the ozone vertical profile along the orbital

track. The algorithms currently use ratios of Earth radiance to Solar irradiance at a set of 12 wavelengths

(at approximately 252, 274, 283, 288, 292, 298, 302, 306, 313, 318, 331 and 340 nm) with the shortest

eight taken from the Nadir Profiler and the longest four from the Nadir Mapper to obtain estimates of the

77

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

total column ozone, effective reflectivity, and the ozone vertical profile in 12 Umkehr Layers. (The 12

Umkehr layers boundaries are at: [0,.25,.50,.99,1.98,3.96,7.92,15.8,31.7,63.3,127,253,1013] hPa.) The

radiances for the four longer wavelengths are obtained from the 25 Nadir Mapper FOVs co-located with

a single Nadir Profiler FOV. The IDPS ozone profile EDR product is made using an implementation of

the Version 6 Solar Backscatter Ultraviolet instruments (SBUV/2) algorithm [Bhartia et al. 1996]. The

longer channel radiance/irradiance ratios are used to generate estimates of the total column ozone and

scene effective reflectivity. This total column ozone estimate is combined with a power law ozone

profile in the upper layer from the shorter channels to generate a first guess ozone profile that becomes

the A Priori for a maximum likelihood ozone profile retrieval using the ratios for the seven shortest

wavelengths (currently omitting the 252 nm channel); adding the 313 nm channel at high solar zenith

angles (SZA). Additional information is in the OMPS Nadir Profile Algorithm Theoretical Basis and

Operational Algorithm Description Documents, and a volume of the Common Data Format Control

Book. The Version 8 Ozone Profile algorithm used for reprocessing and operationally at NOAA for the

SBUV/2 is described in Bhartia et al. [2013]. The Version 6 ozone profile is reported as Dobson Units

(milli-atm-cm) in Umkehr Layers for 12 layers – with the bottom two layers, layers 0 and 1, combined,

and the top layer extending upward including layers 12 and above. It is also reported as ozone mixing

ratios at 19 pressure levels. This latter product is obtained by taking the derivative of a spline fit of the

cumulative layer ozone amounts.

88

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

3. Products and internal consistency

The OMPS total ozone and ozone profile products from the operational system are available as

provisional-release data sets through links at http://www.nsof.class.noaa.gov/saa/products/welcome .

3.1. Total Column Ozone

The OMPS-NM total column ozone product provides estimates over the full sunlit Earth of total column

ozone, effective reflectivity and UV absorbing aerosols (an aerosol index) each day. The measurements

are binned into 35 cross-orbital-track fields of view (FOVs) on-board the spacecraft with each FOV

using a specific fixed set of CCD pixels. The performance requirements for the Total Ozone EDR are

given in Table 1. As can be seen, the requirements are stratified by ozone amount. There is an additional

long-term requirement that reprocessed products for climate data records should be stable at the

1%/decade level. The initial evaluation of the products in this section looks at the values for the retrieval

quantities relative to those expected from past experience as well as the cross-track consistency of the

results as internal checks.

One of the initial ways to examine the performance of the products is to see if they match the expected

range and distribution of values. Figures 1.a, 1.b and 1.c provide a typical example of this check for the

primary OMPS-NM total column EDR algorithm products. They show false color global maps of (a)

total column ozone, (b) effective reflectivity and (c) an aerosol index for April 3, 2013 from the IDPS

EDR total ozone products. Maps for other days for the last 18 months – and for other monitoring

products – are available for viewing at the NOAA Satellite Integrated Calibration/ Validation System

(ICVS) web site at http://www.star.nesdis.noaa.gov/icvs/index.php.

Figure 1.a is for total column ozone. The colors show different levels of ozone in Dobson Units (milli-

atm-cm). The large amounts in the Northern Hemisphere are expected at this time of year. The black

99

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

area around the South Pole is a region of polar night where there is no sunlight to make the OMPS

measurements.

Figure 1.b is for the effective reflectivity. The colors show varying reflectivity in percent. The range

from close to 0% over cloud-free land (e.g.,, the Sahara desert) to close to 100% for bright clouds and

snow/ice covered surfaces (e.g., Antarctica) is as expected. The EDR product derives this estimate from

the longest wavelength of each triplet where the ozone absorption is negligible.

Figure 1.c is for an absorbing aerosol index. The colors show different levels of the index computed as

a measurement residual for the 364 nm channel using the reflectivity estimate from the 331 nm channel.

While some features are expected (e.g., dust over the Sahara) others, such as the North/South bands in

the mid-Pacific are not. Some small features over the open ocean near the Equator are produced by sun

glint but the large North/South features result from cross-track biases between the two channels and are

discussed later in this section.

The OMPS-NM is an extremely versatile instrument in that it can be programmed to execute a variety

of spatial and temporal aggregations and to select portions of the spectrum to report. Figure 2.a on the

left shows the derived UV effective reflectivity from a single wavelength channel (380 nm) measured

with high spatial resolution, 5 km by 10 km at nadir. Comparison to the EOS Aqua Moderate Resolution

Imaging Spectrometer (MODIS) image in Figure 2.b on the right for the same day documents the

capability of the OMPS TC sensor to identify 10 km or smaller surface features and characteristics. It

also confirms the Geolocation accuracy of the OMPS products as the overlain white land boundaries

follow the land/sea brightness changes.

Since the OMPS-NM uses a CCD array, there are essentially thousands of independent detectors. This

means that the products from each cross-track FOV are derived from their own set of detectors. To

investigate the internal consistency of the products, we compute weekly averages over a

1010

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

latitude/longitude box in the Equatorial Pacific defined by 20ºS to 20ºN, 100ºW to 180ºW. This region is

selected because of its benign conditions – dark ocean surface, low non-cloud aerosol loading, and

relatively homogeneous total column ozone amounts. Averages are created for a variety of retrieval

products and their behavior is tracked from week to week. Figure 3 shows four weekly cross-track

patterns for April 2013 for the following: (a) average total column ozone, (b) 1-percentile reflectivity,

(c) average aerosol index, and (d) average SO2 index. The OMPS total ozone algorithm also computes

an SO2 Index. Each figure has the cross-track view position on the horizontal axis. Cross-track 17 is the

nadir position and 0 and 34 are the extreme west and east viewing angles, respectively. Plots for

additional months and other variables (and retrievals from other instruments) are available for viewing at

the ICVS web site cited above.

Figure 3.a shows the weekly average total column ozone. The 2% cross-track variations in the relative

cross-track estimates are very stable and are related to the small inaccuracies in the current

measurements as discussed at length in the companion papers. The shifts up and down from week-to-

week are consistent with the slowly varying ozone for this region.

Figure 3.b shows the weekly, one-percentile effective reflectivity values. These are computed by

collecting the effective reflectivity for each cross-track viewing position for a week and then finding the

one-percentile quantity for those values. The one-percentile effective reflectivity values are expected to

be approximately 4% for this region of the globe from climatological measurements made by other

instruments for cloud free scenes. The larger values observed here are due to the lack of stray light

correction in the SDR algorithm, inaccuracies in the radiance/irradiance calibration, and sun glint for the

8th to 15th cross-track viewing conditions. Changes in the SZA and the sun glint combine to produce the

week-to-week changes for those cross-track positions.

1111

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

Figure 3.c shows the weekly average Aerosol Index (the 364 nm channel measurement residuals when

the effective reflectivity for 331 nm is used to predict them.) in N-value units (1 N value ~2.3%) for the

same region and period. The cross-track deviations represent 1% to 3% inter-channel calibration offsets

between the two measurements. This region of the Pacific has very low aerosol loading, so the general

positive offset represents an overall inter-channel bias. The bump between the 8 th and 15th cross-track

positions is related to a failure of the reflectivity model to properly account for the wavelength

dependence of sun glint. The large upswing for the higher numbered cross-track positions corresponds

to the large North/South features seen in Figure 1.c.

Figure 3.d shows the weekly average SO2 Index. This SO2 Index is created by using the measurement

residuals from the 317, 331 and 336 nm triplet with total ozone and effective reflectivity estimated from

the algorithm. An SO2 amount (in DU) is computed to make the residuals 0 by using the relative SO2

versus O2 absorption ratios. As this assumes that the SO2 distribution is the same as the ozone profile in

the standard table (as interpolated for latitude and total ozone amount) it is considered an index related

to the presence of SO2 as opposed to a true product. The results are very repeatable from week to week

but show large variations with cross-track position. The estimates from this triplet are very sensitive to

inter-channel calibration biases and the pattern is primarily produced by differences in the relative bias

of the measurements at these three wavelengths across the CCD array. We expect that this product’s

cross-track averages will flatten out when small inter-channel biases are removed by soft calibration

adjustments. We are also investigating the use of triplets with less sensitivity to measurement errors. The

SO2 Index is shown here because the multiple triplet algorithms flag all values greater than 6 DU. As

this region of the Pacific has very low SO2 loading, the large positive values here are not physically

reasonable. The false-alarm flagging will continue to occur frequently until better calibration

adjustments are in use. The comparison studies in this paper have included data whether this flag is set

1212

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

or not. The NASA OMPS Science Team has developed improved algorithms for estimating SO2 using

information in the 308 to 315 nm spectral range [Yang et al., 2009, Yang et al., 2013]. Their use in

operational OMPS processing is under investigation.

Figure 4 shows time series of daily averages over the Equatorial Pacific box for total column ozone

estimates from NOAA-19 SBUV/2, NASA EOS Aura OMI, and S-NPP OMPS First Guess products.

The SBUV/2 and OMI estimates are from Version 8 algorithms processed by NOAA and NASA,

respectively. The figure shows a bias of approximately –4% (10 DU) between the OMPS and SBUV/2

or OMI products for most of the record. This is just within the accuracy performance limit in Table 1 but

the primary purpose of this comparison is to check the longer term stability by comparison to two other

satellite products. There have been changes over time to the OMPS products in the form of new Day 1

Solar Flux data sets, new wavelength scales, and changes in the cloud top pressure climatology to a UV-

based one. The OMPS instrument is making a series of calibration measurements on Sundays leading to

reduced statistics once per week. These days have been removed from the means used to create this

figure.

3.2. Nadir Ozone Profile

The OMPS-NP ozone profile product provides estimates over the full sunlit Earth of the ozone vertical

profile along the nadir satellite track each day. The performance requirements for the Nadir Ozone

Profile EDR are given in Table 2. As can be seen, the requirements are stratified by vertical levels.

There is an additional long-term requirement that reprocessed products for climate data records should

be stable at the 2%/decade level. The initial evaluation of the products in this section looks at the values

for initial and final measurement residuals relative to those from Solar Backscatter Ultraviolet (SBUV/2)

instruments on NOAA Polar Orbiting Environmental Satellites (POES), as well as internal consistency

of the total ozone from summed profiles compared to other direct total ozone estimates.

1313

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

Before launch it was known that the OMPS-NP measurements would be greatly affected by charged

particles as the satellite passed within the South Atlantic Anomaly (SAA) region. A performance

exclusion was prescribed and quality flags are set for retrievals in that region. Figure 5 show the size of

the effect on the shortest profiling wavelength at 252 nm. It compares the radiances for May 5, 2013 for

orbits outside the SAA with those for four passing through it. The portions of the orbital tracks cover

60ºS to the 20ºN. Fortunately, the OMPS-LP looks at the Earth’s atmosphere approximately 3000 km

(~25 degrees of latitude South) behind the satellite, so the satellite is outside of the SAA when it looks

back and views this region and region it is viewing when it passes through the SAA has already been

seen by the OMPS-NP. That is, the OMPS-NP and OMPS-LP cover each other’s deficiencies with

regard to SAA measurement contamination viewing regions.

Figure 6 compares the daily equatorial zonal mean initial measurement residuals for the six shortest

channels for the IDPS V6 ozone retrieval algorithm to those from the same algorithm applied to NOAA-

19 SBUV/2 instrument measurements. The region from 0° to 90° West Longitude is excluded to avoid

SAA effects. The residuals are in N-values (1 N ~ 2.3%) and are computed as measurement minus

radiative transfer model result differences using the first guess profile in the model. The time period is

from the end of February 2012 through September 2013. The residuals are computed with respect to

differing first guess profiles especially with respect to the use of the shortest channel; the 252 nm

channel is not used in generating the first guess profile or in the maximum likelihood retrieval for the

OMPS-NP while it is used for the SBUV/2s. The jump in the initial residuals in early June 2012 – most

noticeable in the 298 nm residuals – are coincident with the introduction of new Day 1 Solar Spectra for

the OMPS-NM SDR. The jumps in July 2012 – most noticeable in the 252, 283 and 288 nm channels –

are coincident with the introduction of a new Day 1 Solar spectrum for the OMPS-NP SDR. The jumps

in February 2013 – most noticeable in the 252 nm residuals -- are coincident with the start of weekly

1414

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

updates to the dark corrections to the OMPS-NP SDRs. While there are offsets in the time series they

track well over time with the notable exception of the shorter channels prior to the start of weekly dark

updates and other SDR changes. The initial residual offsets are related to the first guess profile

differences, calibration biases, and the lack of stray light and intra-orbit wavelength scale corrections for

the OMPS-NP SDR. (The V6 uses a “first guess profile” computed from the discrete channel total ozone

estimate from the 318 and 331 nm channels and a power-law estimate of upper layer ozone from the 273

and 283 nm channels instead of a true A Priori.) Methods to use residuals and other information to inter-

calibrate the SBUV/2 instruments are described in DeLand et al. [2012]. These same approaches will be

used with OMPS and concurrent SBUV/2 records for CDR creation.

The OMPS-NP SDR measurements have significant contributions from out-of-range spectral stray

light. These were well characterized prelaunch and a measurement-based correction is under

development. Similar signal contamination has been seen for some of the SBUV/2 series of instruments

and corrections are applied to those measurements [DeLand et al. 2012]. The current stray light

contributions to the radiances produce false correlations in the retrieved ozone at the top of the profile

with the underlying surface or cloud reflectivity changes, because radiances at shorter wavelengths

contributing to vertical information of ozone at those altitudes are not sensitive to such changes. Figure

7 presents the results of an investigation into this phenomenon. The ozone profile retrievals for the

latitudes from 20S to 20N for the first five orbits of data for May 15, 2013 were used in the study. Cubic

polynomials in latitude are fit to each layer ozone data set and to the effective reflectivity as estimated

from the 331 nm channel. The differences between the retrieved values and these fit polynomials are

plotted in Figure 7 for the top four retrieval layers on the vertical axes versus the effective reflectivity

variations, with respect to its cubic fit, on the horizontal axes. The negative correlation is symptomatic

of stray light in the shorter wavelength radiances driving underestimation of ozone amounts in the upper

1515

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

portions of the profile; the stray light proportion increases as the signals drop off with increasing ozone

absorption for the shorter wavelengths. The measurements at the shorter wavelengths provide the

retrieval information for the upper part of the profile. Note that the stray light contribution will be

present even for the low reflectivity scenes as the source radiances do not go to zero but are at the pure

atmospheric Rayleigh scattering levels, that is, a correction will produce increased ozone in these layers

even for low reflectivity scenes.

While the instrument throughput degradation for the OMPS-NP SDR (See Seftor et al. [2013].) is

estimated to be less than 0.5% over the first year of operations, and can be well approximated by linear

trends in time for each of the profile wavelengths, the Day One solar spectrum for the OMPS-NP does

not reflect the true changes in the solar flux due to changes in solar activity over a solar cycle.

Fortunately, one can estimate the spectral changes by the use of Mg II Indices and scale factors as

described in Cebula, et al. [1992]. A set of scale factors showing the relative variations in the OMPS-NP

solar spectra for relative changes in an Mg II core-to-wing ratio Index as estimated from the first year of

OMPS-NP working diffuser measurements are shown in Figure 8. The Mg II Index used to derive these

factors shows up to 3% variations over 27-day solar rotation at peak solar activity. (The period from

mid-2012 to mid-2013 has been a very active one for Solar sunspots. There are two contributing

influences to the initial residuals in Figure 6 as the Day 1 Solar used in the retrieval algorithm does not

follow the true solar and the atmospheric ozone creation and destruction are also tied to the Solar UV

levels. The associated variations appear in the 252-nm residuals, Figure 6.a for the OMPS-NP, and in the

273-nm residuals, Figure 6.b for the SBUV/2. Notice the highly expanded scale for Figure 6.b.) Options

are under consideration to make use of daily Mg II Index values to adjust the solar flux for these

variations.

1616

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

4. Validation and further comparisons

4.1. Total Column Ozone

The IDPS OMPS-NM total column ozone products have been compared to similar products from

ground-based and satellite-based instruments. Figure 4 gave comparison of the average performance

over the Equatorial Pacific Box for the OMPS IDPS products and contemporaneous OMI and SBUV/2

ones. Additional examples are discussed below and displayed in Figures 9 and 10.

Figure 9 shows time series of differences of OMPS total column ozone (IDPS 1st Guess) and the OMI

total column ozone from the Version 8 NASA processing with estimates from measurements made by

ground-based Dobson instruments in Boulder Colorado (See Komhyr et al. [1989] for information on

this instrumentation), Mauloa Hawaii, and Lauder New Zealand. The 1st Guess product is used because

it has been more stable over the period (fewer algorithm changes) and the changes to the EDR product

have been to make it more like the 1st Guess product (e.g., switching to the use of measurement-based

partial cloud estimates and to the use of a UV climatology for the cloud top pressure). In addition, the

EDR product makes an adjustment for the temperature sensitivity of ozone absorption by using CrIS

near-real-time temperature profile retrievals and to the surface reflectivity by using VIIRS near-real-time

snow/ice retrievals. Both of these ancillary products have been evolving over the period of study. The

Dobson instruments values are for direct sun observations. The OMPS data are weighted average

overpass (distance weighted for all data with 0.5º Latitude and SEC(Latitude)º Longitude. The OMI data

is from the standard gridded NASA product at http://ozoneaq.gsfc.nasa.gov/OMIOzone.md.

Data were collected from March 2012 through June 2013. Because there was an adjustment to the Day

1 solar from pre-launch to in-orbit based spectra in mid 2012, the statistical study described here only

used the data from July 2012 to June 2013. Figures 9.a, 9.c, and 9.e shows time series of the two satellite

data sets differences versus the ground-based estimates. Notice the declines in the nine-point moving

1717

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

averages in early 2012 for the OMPS differences. Figures 9.b, 9.d, and 9.f show scatter plots comparing

the differences of the satellite and ground data sets to their averages. Because all three data sets have

measurement and retrieval noise components, we use methods which allow for both in the regression

fits. That is, we follow the development in Flynn [2007] and construct a model of the form:

(S – AvgS – ε) = m (G – AvgG – δ) (1)

Where G is a set of ground-based estimates and AvgG is the set’s average and δ is the noise term, S is a

set of matchup satellite-based estimates and AvgS is its average and ε is the noise term, and m is the

slope of the linear relationship between the two. We allow for an implicit difference in the additive term

as opposed to a purely multiplicative model because the ground and satellite systems will have different

sensitivity to ozone in the lower atmosphere due to backscatter contribution heights for the satellites

versus direct solar viewing by the ground system. The sensitivity to the lower layers is further

complicated by clouds and especially in the case of Mauna Loa and Boulder by differences in the

average surface pressure of the large satellite FOVs compared to the specific location of the ground

stations. Adjustments from 6 to 12 DU have been used to account for this difference based on Hilo HI

ozonesondes and standard profile lower layer amounts.

Subsets of the matchups for each station are created by selecting days from 7/2012 to 6/2013 where all

three instruments made estimates. Table 3 gives the results of statistical computations for satellite and

ground data set pairs with three assumptions about the relative sizes of each set’s noise term. Column 1

of Table 3 give a shorthand name for each station and the number of days with data. Column 2 specifies

the satellite data set for each row. Columns 3 and 4 give the means for the Dobson and satellite

instrument, respectively. Column 6, 7, and 8 gives estimates from linear regression fits with three noise

assumption: mG – no noise in the Dobson data; mS – no noise in the satellite data; and mE – equal-

relative-size, uncorrelated noise in both. Column 8 gives, σ, the standard errors on the slope estimates.

1818

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

Intervals of the form [mG-σ,mG+σ] give the ranges of slopes that could be found depending on the

assumed distribution of the two systems’ measurements noise. Columns 9 and 10 give the estimates of

the standard deviations of the ε and δ noise terms in Eq. 1 for the equal-noise cases. Column 11 is the

Pearson correlation coefficient. Columns 12 and 13 give the expected differences between the satellite

and Dobson data at the maximum and minimum observed data values as determined from the equal-

noise model. The OMPS results show good consistency with the OMI ones but with 4 to 16 DU lower

values. The OMPS are also lower than the Dobson (after adjusting MLO for below-station column

ozone) by 3 to 24 DU. Some of these values lie outside of the allowed accuracy requirements of Table 1.

Since some of the differences between the satellite and the Dobson estimates are due to matchup

location discrepancies, the equal noise results are reasonable measures of the product performance. The

estimated uncertainty of the OMPS results, as reported for ε, range from 4.7 to 6.3 DU and are within

the allowed precision requirements of Table 1.

Figure 10 extends the OMPS/Dobson comparisons to a collection of 22 Dobson stations. The

corresponding results for the operational NOAA-19 SBUV/2 instrument total ozone estimates from the

Version 8 algorithm are also shown. One can see that the OMPS calibration adjustments in the early part

of 2012 produce a large decrease in the bias relative to both the Dobson network and the SBUV/2

record. The small positive bias for the SBUV/2 results is a common result for the NOAA SBUV/2 series

of V8 products.

4.2. Nadir Ozone Profile

The NOAA series of SBUV/2 instruments make a set of measurements similar to those used from

OMPS in the ozone profile retrieval. While the IDPS Version 6 algorithm has been replaced in NOAA

SBUV/2 operational processing with the Version 8 algorithm, the Version 6 products are still available

from the SBUV/2 for comparisons. In particular, the NOAA-19 platform has a similar equator crossing

1919

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

time but slightly different spacing in its orbital tracks than the S-NPP platform. This leads to periodic

formation flying for the two platforms approximately every 12 days.

Figure 11 shows the results of a match up comparison for a pair of orbits on July 10, 2013. The two

satellites are flying in formation – orbital tracks within 50 km and sensing times with 15 minutes at the

Equator. The ozone profile retrievals are reported in Dobson Units for 12 pressure layers. They are

plotted here versus latitude, degrees north. The figures show general agreement between the retrievals

for the two instruments but with the OMPS-NP retrieving smaller values at the top of the profiles. As

discussed earlier, this is probably due to the lack of a correction for stray light. The mean layer

differences for OMPS minus SBUV/2 for 20º S to 20º N from the top layer to the bottom layer are -9.3, -

7.9, -7.0, -5.0, -0.7, -0.5, -6.1, -2.6, 6.0, 6.3, 6.1 and 4.0 %.

5. Summary and conclusions

The OMPS instruments are performing well, and initial evaluation of the products from the OMPS-NM

and OMPS-NP finds many positive aspects. The total column ozone has biases with other records on the

order of -3% and these biases are stable over periods away from known system changes. The ozone

profile retrievals compare well in the middle stratosphere but have significant biases at the top of the

profiles where stray light becomes a problem. As the SDR processing continues to mature, e.g.,

implementation of adjustments for stray light and wavelength scale shifts, the products will become

even better. Given the stability of the measurements, the prognosis for extrapolated characterization of

the small OMPS-NP throughput degradation and development of soft calibration adjustments to create

homogenized ozone profile products for OMPS and SBUV/2 is favorable.

The IDPS OMPS-NM total column ozone and OMPS-NP ozone profile products have been released

at the NOAA archive with provisional maturity. Links to the OMPS data products and readme files with

product caveats are provided at http://www.nsof.class.noaa.gov/saa/products/welcome .

2020

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

Both products are undergoing evaluation by near-real-time users and will be provided in near-real-time

in Binary Universal Form for Representation of meteorological data (BUFR) files.

Acknowledgments. The work was supported by the NOAA JPSS and the NASA S-NPP programs. The

manuscript contents are solely the opinions of the authors and do not constitute a statement of policy,

decision, or position on behalf of NOAA or the U. S. Government.

References

Bhartia, P.K., et al., (1996), Algorithm for the estimation of vertical ozone profiles from the backscattered ultraviolet technique, J. Geophys. Res., 101, 18793-18806.

Bhartia, P.K., and C.G. Wellemeyer (2002), TOMS-V8 total ozone algorithm in OMI Algorithm Theoretical Basis Document Volume II: OMI Ozone Products, edited by P.K. Bhartia, under algorithm documents at http://www.knmi.nl/omi/research/documents/#other .

Bhartia, P.K., et al., (2013) Solar Backscatter UV (SBUV) total ozone and profile algorithm, Atmos. Meas. Tech., 6, 2533-2548, doi:10.5194/amt-6-2533-2013, 2013.

Cebula, R.P., DeLand, M.T., and Schlesinger, B.M., (1992), Estimates of solar variability using the solar backscatter ultraviolet (SBUV) 2 MgII Index from the NOAA 9 Satellite, J. Geophys. Res., 97 (D11), 11613-11620.

DeLand, M. T., et al., (2012), Calibration of the SBUV version 8.6 ozone data product, Atmos. Meas. Tech. Discuss., 5, 5151-5203, doi:10.5194/amtd-5-5151-2012.

Flynn, L.E., (2007), Comparing two sets of noisy measurements, U.S. Dept. Commerce, Washington, DC, NOAA Technical Report NESDIS #123.

http://docs.lib.noaa.gov/noaa_documents/NESDIS/TR_NESDIS/TR_NESDIS_123.pdfKomhyr, W. D., R. D. Grass, and R. K. Leonard (1989), Dobson spectrophotometer 83: A standard for

total ozone measurements, 1962 – 1987, J. Geophys. Res., 94(D7), 9847 – 9861. McPeters, R.D. et al., (1996), Nimbus-7 TOMS Ozone Mapping Spectrometer (TOMS) Data Products

User’s Guide, NASA Reference Publication #1384. Rault, D. and Loughman, R.P., (2013), The OMPS Limb Profiler Environmental Data Record Algorithm

Theoretical Basis Document and Expected Performance. IEEE T. Geoscience and Remote Sensing 51(5-1): 2505-2527.

Remund, Q.P., et al., (2004), The ozone mapping and profiler suite (OMPS): on-orbit calibration design, Proc. SPIE, 5652, 165-173.

Rodriguez, J. V., et al., (2003), “An overview of the nadir sensor and algorithms for the NPOESS ozone mapping and profiler suite (OMPS),” Proc. SPIE, 4891.

Vasilkov, A., et al. (2008), Evaluation of the OMI cloud pressures derived from rotational Raman scattering by comparisons with other satellite data and radiative transfer simulations, J. Geophys. Res., 113, D15S19, doi:10.1029/2007JD008689.

2121

441

442

443

444

445

446

447448449450451452453454455456457458459460461462463464465466467468469470471472473474475

Yang, K., et al., (2009), Improving retrieval of volcanic sulfur dioxide from backscattered UV satellite observations, Geophys, Res. Lett., 36, L03102, doi:10.1029/2008GL036036.

Yang, K., et al., (2013), First Observations of SO2 from the Satellite Suomi NPP OMPS: Widespread Air Pollution Events over China, GRL, Accepted 9/2013.

2222

476477478479480

Table 1. OMPS Total Column Ozone Requirementsa

Attribute Requirementa. Horizontal Cell Size 50×50 km2 @ nadirb. Vertical Cell Size 0 -60 kmc. Mapping Uncertainty 1-σ 5 km at Nadird. Measurement Range 50 -650 milli-atm-cme. Measurement Precision 1. X<250 milli-atm-cm 6.0 milli-atm-cm 2. 250<X<450 milli-atm-cm 7.7 milli-atm-cm 3. X>450 milli-atm-cm 2.8 milli-atm-cm + 1.1%f. Measurement Accuracy 1. X<250 milli-atm-cm 9.5 milli-atm-cm 2. 250<X<450 milli-atm-cm 13.0 milli-atm-cm 3. X>450 milli-atm-cm 16.0 milli-atm-cmg. Refresh 90% of sunlit globe / 24 hrsh. Stability 1%/decade aRequirements apply under daytime conditions with Solar Zenith Angles (SZA) up to 80º.

Table 2. OMPS Nadir Profile Ozone Requirementsb

Attribute Requirementa. Horizontal Cell Size 250 × 250 km2

b. Vertical Cell Size 5-km reporting 1. Below 30 hPa ( ~ < 25 km) 10-20 km 2. 30 -1 hPa ( ~ 25 -50 km) 7-10 km 3. Above 1 hPa ( ~ > 50 km) 10-20 kmc. Mapping Uncertainty 1-σ 25 kmd. Measurement Range 0.1 – 15 ppmve. Measurement Precision 1. Below 30 hPa ( ~ < 25 km) Greater of 20 % or 0.1 ppmv 2. At 30 hPa ( ~ 25 km) Greater of 10 % or 0.1 ppmv 3. 30 -1 hPa ( ~ 25 -50 km) 5% – 10% 4. Above 1 hPa ( ~ > 50 km) Greater of 10 % or 0.1 ppmvf. Measurement Accuracy 1. Below 30 hPa ( ~ < 25 km) Greater of 10 % or 0.1 ppmv 2. 30 -1 hPa ( ~ 25 -50 km) 5% – 10% 3. at 1 hPa (~20 km) Greater of 10 % or 0.1 ppmv 4. Above 1 hPa ( ~ > 50 km) Greater of 10 % or 0.1 ppmvg. Refresh 60% of sunlit globe / 7 daysh. Stability 2%/decade bRequirements apply under daytime conditions with Solar Zenith Angles (SZA) up to 80º. Requirements do not apply within the South Atlantic Radiation Anomaly region.

2323

481

482

483484

Table 3. Statistics for Dobson Match-Up Data Sets In Figure 9. Site Sat. AvgG AvgS mG mS mE σ δ ε MinE MaxE

# Days Name DU DU DU DU DU DUBOU OMPS 308.7 293.9 0.90 0.98 0.94 0.02 6.7 6.3 0.97 -10.6 -24.1N=335 OMI 308.7 306.3 0.93 1.00 0.96 0.02 6.4 6.1 0.97 0.3 -8.1MLO OMPS 256.6c 259.4 0.99 1.13 1.06 0.03 4.7 4.9 0.93 0.4c 5.9c

N=217 OMI 256.6c 266.9 1.03 1.17 1.10 0.03 4.4 4.8 0.94 6.0c 15.6c

LAU OMPS 304.5 300.2 0.97 1.00 0.99 0.02 4.8 4.7 0.99 -3.3 -5.8N=270 OMI 304.5 304.4 0.97 1.01 0.99 0.02 5.3 5.2 0.98 0.6 -1.1 cThe Dobson station is near the top of Mauna Loa. Satellite FOVs include ocean scenes. Adjustments from 6 to 12 DUhave been used to account for these scene differences based on Hilo HI ozonesondes and standard ozone profiles.

2424

485

486487

488

.

Figure 1. Daily global maps with false color images of three IDPS products for April 3, 2013: Top – Total column ozone for April 3, 2013 from the IDPS EDR product. The color bar gives the amounts in Dobson Units (1 DU = 1 milli-atm-cm); Middle – Effective reflectivity (average of the 364, 367, 372 and 377 nm channel estimates) for April 3, 2013 from the IDPS EDR product. The colors show varying reflectivity in percent; and Bottom – Absorbing aerosol index for April 3, 2013 from the IDPS EDR product. The colors show different levels of the index computed as a measurement residual for the 364 nm channel using the reflectivity estimate from the 331 nm channel. The units are in N-values which are approximately equivalent to 2.3% per unit. Daily images for the full record to date are available through links at http://www.star.nesdis.noaa.gov/icvs/index.php.

2525

489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533

Figure 2. High-Spatial-Resolution Geolocation Comparison. The image on the left shows a false color map of the OMPS effective reflectivity (from a single Ultraviolet channel at 380 nm) over the Arabian Peninsula region for January 30, 2012 when the instrument was making a special set of high-spatial-resolution measurements with 5×10 km2 FOVs at nadir. The color scale intervals range from 0 to 2 % in dark blue to 18 to 20 % in yellow. The image on the right is an Aqua Moderate Resolution Imaging Spectroradiometer (MODIS) Red-Green-Blue image for the same day.

2626

534535536537538539540541542543544545546547548549550551552553554

Figure 3. Weekly-Averages of four retrieval parameters (Ozone, 1-percentile Effective Reflectivity, and Aerosol and SO2 Indices ) from the 1st Guess product versus Cross-Track position (Cross-track 17 corresponds to the nadir position and 1 and 35 are the extreme viewing angles.) for April 2013 for a latitude/longitude box (20ºS to 20ºN, 100ºW to 180ºW) in the Equatorial Pacific Box. Figure 3.a shows the weekly average total column ozone in Dobson Units Figure 3.b shows the weekly average, one-percentile effective reflectivity (average of the 364, 367, 372 and 377 nm channel estimates) values. Figure 3.c shows the weekly average Aerosol Index (the 364nm channel measurement residuals when the effective reflectivity for 331 nm is used to predict them.) in N-value units (1 N value ~ 2.3 %). Figure 3.d. shows the weekly average SO2 Index.

2727

555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Figure 4. Time series of daily average total column ozone satellite products in the Equatorial Pacific box. Each symbol is a nine-point time-centered moving average of daily mean ozone in the Equatorial Pacific Latitude/Longitude box for total column ozone estimates from NOAA-19 SBUV/2, NASA EOS Aura OMI, or S-NPP OMPS First Guess products.

2828

592593594595596597598599600601602603604605606607608609610611

Figure 5. OMPS-NP Radiances at 252 nm versus latitude for portions of ten orbits for May 5, 2013. The radiances for the four orbits passing inside the SAA (large deviations) are called out in the figure with different symbols and line styles; the longitudes in degrees West of these four orbits at their Equator crossing points are given as well. The solid lines give the radiances for the other six orbits (three on each side) showing the expected increase in radiances as the solar zenith angles get smaller moving northward along the orbital tracks. The variations among these six are produced by longitudinal fluctuations in the stratospheric ozone field.

2929

612613614615616617618619620621622623624625626627628629630631632633634635636

Figure 6. Initial Measurement Residuals for OMPS-NP Version 6 Algorithm. The six figures show the daily average initial residuals for profile wavelengths (a) 252 nm (b) 274 nm, (c) 283 nm (d) 288 nm, (e) 292 nm, and (f) 298 nm for the V6PRO product from OMPS compared to the V6PRO product for the operational POES NOAA-19 SBUV/2 for the equatorial daily zonal means (20N to 20S) with 0-90W removed to avoid the SAA effects. Every third day is plotted to reduce clutter.

3030

637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670

.

Figure 7. Scatter Plots of Ozone Layer Variations versus 331-nm Effective Reflectivity Variations for the first five orbits of May 15, 2013. The data sets are (a) Umkehr Layer 12 and above (< 0.25 hPa), (b) Umkehr Layer 11 0.25 – 0.50 hPa), (c) Umkehr Layer 10 (0.50 – 0.99 hPa), and (d) Umkehr Layer 9 (0.99 – 1.98 hPa). The variations are computed relative to smooth polynomial fits of each parameter versus latitude. The correlations are not true geophysical variations but are symptomatic of stray light in the shorter profiling channels with longer wavelength sources; the algorithm interprets increased radiance as evidence of decreased ozone.

3131

671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710

Figure 8. OMPS-NP Mg II Index Scale Factors for Solar Flux variations. These factors provide estimates of the expected changes in flux at a wavelength (at the OMPS-NP 1.1-nm FWHM bandpass resolution) due to solar activity as tracked by a Mg II line core-to-wing index. The Mg II Index created from the OMPS NP measurements for this study shows up to 3% variations over a 27-day solar rotation.

3232

711712713714715716717718719720721722723724725726727728729730731732

Figure 9. Comparison of OMPS and OMI total column ozone with Dobson estimates for Boulder CO, Manua Loa HI, and Lauder NZ. The figures on the left show the time series of differences for satellite overpass data minus the ground-based Dobson. The diamonds are for OMI and the plus signs are for OMPS. The solid line is the nine-point moving average for the OMPS data. The figures on the right are the satellite minus Dobson differences versus their averages. The solid line is the fit for OMPS and the dotted line is the fit for OMI as described in the text. Figure pairs (a) and (b), (c) and (d), and (e) and (f) are for Boulder, Mauna Loa and Lauder, respectively.

3333

733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795

Figure 10. Monthly differences between matchup NOAA-19 SBUV/2 Version 8 total column ozone and OMPS 1st Guess total column ozone with a collection of 22 Dobson estimates from the World Ozone Data Center. For OMPS, the data are distance-weighted averages for estimates within 0.5º Latitude and SEC(Latitude)º Longitude of each stations location. For SBUV/2, the data are distance-weighted averages for estimates within 2.0º Latitude and 20º Longitude of each stations location. Each data point is a monthly average difference for the satellite instrument versus the Dobson ones. At least six matchup values are required for a station to be used in the monthly average. As few as five stations may have reported enough data for the later values.

3434

796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829

Figure 11. Chasing orbit comparisons of SBUV/2 and OMPS-NP Version 6 Ozone Profiles for July 10, 2013. Figures (a)-(l) show the 12 Umker layer amounts versus latitude for the two products. The layer boundaries are given in hPa within the figures. The two orbit are within 50 km and 15 minutes of each other at the Equator.

3535

831832833834