1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Performance of the ozone mapping and profiler suite (OMPS) products
L. Flynn, C. Long, X. Wu, R. Evans, C.T. Beck NOAA
I. Petropavlovskikh, G. McConville CIRES
W. Yu, Z. Zhang, J. Niu, E. Beach, Y. Hao IMSG
C. Pan, UMD CICS
B. Sen, M. Novicki Northrop Grumman Aerospace Systems
S. Zhou, Wyle Information Systems
C. Seftor, SSAI
Corresponding author: Lawrence E. Flynn, NOAA, NCWCP E/RA2, 5830 University
Research Court, #2850, College Park, MD 20740-3818 ([email protected])
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 Terms
3360 Remote Sensing
0340 Middle atmosphere: composition and chemistry
3311 Clouds and aerosols
“Manuscript Submitted for the Special Issue of AGU JGR-Atmospheres on Suomi NPP Cal/Val
Science Results”.
2
21
Abstract. NOAA, through the Joint Polar Satellite System (JPSS) program, in partnership with National Aeronautical
22
and Space Administration (NASA), launched the Suomi National Polar-orbiting Partnership (S-NPP) satellite, a risk
23
reduction and data continuity mission, on October 28, 2011. The JPSS program is executing the S-NPP Calibration and
24
Validation (Cal/Val) program to ensure the data products comply with the requirements of the sponsoring agencies.
25
The Ozone Mapping and Profiler Suite (OMPS) consists of two telescopes feeding three detectors measuring solar
26
radiance scattered by the Earth's atmosphere directly and solar irradiance by using diffusers. The measurements are
27
used to generate estimates of total column ozone and vertical ozone profiles for use in near-real time applications and
28
extension of ozone climate data records. The calibration and validation efforts are progressing well, and both Level 1
29
(Sensor Data Records/SDRs) and Level 2 (Ozone Environmental Data Records/EDRs) have advanced to release at
30
Provisional Maturity. This paper provides information on the product performance over the first 22 months of the
31
mission. The products are evaluated through the use of internal consistency analysis techniques and comparisons to
32
other satellite instrument and ground-based products. The initial performance finds total ozone showing negative bias
33
of 2 to 4% with respect to correlative products and ozone profiles often within ±5% in the middle and upper
34
stratosphere of current operational products. Potential improvements in the measurements and algorithms are identified.
35
These will be implemented in coming months to reduce the differences further.
3
36
1. Introduction
37
The OMPS is composed of three instruments, called the Nadir Mapper (OMPS-NM), Nadir Profiler
38
(OMPS-NP), and Limb Profiler (OMPS-LP). The OMPS-NM is a total ozone column sensor and uses a
39
single grating and a Charge-Coupled Device (CCD) array detector to make measurements every 0.42 nm
40
from 300 nm to 380 nm with 1.0-nm Full-Width Half-Maximum (FWHM) resolution. It has a 110°
41
cross-track FOV (~2800 km on the Earth’s surface) and 0.27° along-track slit width FOV. In standard
42
Earth science mode, the measurements are combined into 35 cross-track bins [20 spatial pixels giving
43
3.35° (50 km) at nadir, and 2.84° at ±55° cross-track dimensions for the FOVs]. The resolution is 50 km
44
along-track at nadir, created by using a 7.6-second reporting/integration period. This resolution choice is
45
changeable; and the use of smaller FOVs and shorter integration times is under investigation.
46
The OMPS-NP is an ozone profile sensor and uses a double monochromator and a CCD array detector
47
to make measurements every 0.42 nm from 250 nm to 310 nm with 1.1-nm resolution. It has a 16.6°
48
cross-track FOV, 0.26° along-track slit width. The measurements are combined (100 spatial pixels) into
49
a single spectrum and the reporting period is 38 seconds giving it a 250 km × 250 km cell size collocated
50
with the five central OMPS-NM FOVs.
51
The OMPS-LP is a high vertical resolution ozone profile sensor and uses a prism spectrometer with
52
spectral coverage from 290 nm to 1000 nm. It has three slits separated by 4.25° (viewing back along the
53
nadir track and 250 km cross-track on either side at the tangent point) with a 19-second reporting period
54
that equates to 125 km along-track motion. The slits have 112 km (1.95°) vertical FOVs equating to 0 to
55
60 km coverage at the limb, plus offsets for pointing uncertainty, orbital variation, and Earth oblateness.
56
The CCD array detector provides measurements every 1.1 km with 2.1-km vertical resolution. There are
57
two apertures into the instrument creating a total of six images on the CCD. Further the CCD array is
58
read out with long and short integration times to produce four effective gains.
4
59
The OMPS instruments and measurements are described further in Rodriguez et al. [2003] and
60
Remund et al. [2004] and in three companion papers within this special issue: Seftor et al. [2013], Jaross
61
et al. [2013] and Wu et al. [2013]. Of note is their use of working and reference solar diffusers to
62
monitor instrument throughput degradation and maintain good calibration of the radiance/irradiance
63
ratios used in the retrieval algorithms. Key updates to the SDRs are discussed in the companion papers.
64
Of special interest for the OMPS-NM products are the implementation of a stray light correction,
65
identification of intra-orbit wavelength scale variations, and estimation of calibration offset for OMPS
66
through vicarious and internal consistency checks. For the OMPS-NP the key items are the same as for
67
the OMPS-NM with an additional refinement to the SDR needed to account for solar activity at the time
68
of the measurements relative to that for the Day One solar spectrum.
69
The OMPS-NM and -NP measurements are processed into total column ozone and ozone profile
70
products in near-real-time (NRT) at the NOAA Interface Data Processing Segment (IDPS) and
71
distributed for use in numerical weather models and to assist in forecasting daily ultraviolet (UV) Index
72
values. They are also processed off-line with other retrieval algorithms as a first step in the process of
73
incorporating them into long-term ozone climate data records (CDRs). This paper will discuss results for
74
both types of products but will emphasize the operational ones. The OMPS-LP measurements are
75
processed into ozone profile products by the NASA S-NPP Science Team. The OMPS-LP algorithms
76
are moving toward implementation in the NOAA operations but that topic is not covered in this paper.
77
2. Ozone Retrieval Algorithms
78
The OMPS measurements are used to generate estimates of atmospheric ozone in both operational and
79
off-line systems. This paper will concentrate on the on the products from the NOAA IDPS derived from
80
the OMPS-NM and NP measurements but will also present some information on some preliminary
81
results using the Version 8 total ozone and profile ozone algorithms to generate climate data records
5
82
from the OMPS-NM and NP taking place at NASA and NOAA. A brief introduction of the algorithms
83
appears below. More detailed descriptions of the IDPS algorithms can be found in JPSS Algorithm
84
Theoretical Basis Documents (ATBDs), Operational Algorithm Documents (OADs) and Data Format
85
Control Books (DFCB) available at the S-NPP Library: npp.gsfc.nasa.gov/science/documents.html.
86
A description of research algorithms for the OMPS Limb Profiler is given in Rault and Loughman
87
[2013].
88
2.1. Nadir Mapper Total Column Ozone
89
The spectral measurements from the OMPS-NM of the radiances scattered by the Earth’s atmosphere
90
are used to generate estimates of the total column ozone. The IDPS algorithm uses ratios of Earth
91
radiance to Day 1 Solar irradiance measurements at triplets of wavelengths to obtain estimates of the
92
total column ozone, effective reflectivity, and the wavelength dependence of the reflectivity as affected
93
by aerosols. [The Day 1 Solar spectra are a set of spectra used in the algorithm to compute the
94
radiance/irradiance ratios. The algorithm/processing initially used a set of pre-launch estimates for these
95
and then after real solar measurements were taken new estimates were made (system updates in 6/2012
96
for the OMPS NM and 7/2012 for the OMPS NP). The working diffuser is used every two weeks and
97
the reference diffuser is used every six months to make new measurements. So far the reference diffuser
98
measurements have shown little or no changes above 290 nm and changes from -0.1%/year to -0.5/year
99
growing larger in magnitude from 290 nm to 250 nm. The processing system has daily values for Earth
100
Calibration Factors that can be used to adjust for instrument throughput changes relative to the Day 1
101
values but they have not yet been activated. More information on the instruments performance is in the
102
companion papers.] Table values computed for a set of standard profiles, cloud heights, latitudes and
103
solar zenith angles are interpolated and compared to the measured top-of-atmosphere albedos. The
104
triplets combine an ozone insensitive wavelength channel (at 364, 367, 372 or 377 nm) to obtain cloud
6
105
fraction and reflectivity information, with a pair of measurements at shorter wavelengths. The pairs are
106
selected to have one “weak” and one “strong” ozone absorption channel. The hyperspectral capabilities
107
of the sensor are used to select multiple sets of triplets to balance ozone sensitivity across the range of
108
expected ozone column amounts and solar zenith angles. The "strong" ozone channels are placed at
109
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.
110
They are paired with a longer “weak” channel at 321.0, 329.0, 332.0, or 336.0 nm. The ozone absorption
111
cross-sections decrease from 3 (milli-atm. cm)-1 to 0.3 (milli-atm. cm)-1 over the range of “strong”
112
wavelengths. Typical ozone columns range from 100 Dobson Units (1 DU = 1 milli-atm-cm) to 600 DU.
113
The Multiple Triplet algorithm is applied twice for each FOV. This was done to resolve the “Who
114
goes first?” quandary created by the desires to use information from other sensors in the retrieval
115
algorithms, e.g., OMPS wanted to use the Cross-track Infrared Sounder (CrIS) temperature profile
116
estimates, and CrIS wanted to use the OMPS ozone estimates. The “1st Guess” OMPS-NM products use
117
climatological or forecast fields for surface reflectivity and pressure, snow/ice coverage, cloud optical
118
centroid depth, and atmospheric temperature. They use internally calculated estimates of cloud fractions
119
and effective reflectivity from measurements at non-ozone absorbing UV wavelengths. This product is
120
sometimes called the Total Ozone Intermediate Product (TOZ IP).
121
The 2nd Pass or EDR OMPS products use snow/ice coverage from the Visible Infrared Imaging
122
Radiometer Suite (VIIRS) near-real-time products and temperature profiles from CrIS products. Both
123
products use internally calculated estimates of cloud fractions and effective reflectivity from
124
measurements at non- or weakly absorbing ozone wavelengths. This second product is sometimes called
125
the Total Ozone Environmental Data Record (TOZ EDR). As we will show, both applications of the
126
algorithm are performing well but have room for improvement.
7
127
It was originally planned to use IR measurements to provide the cloud top pressure data for the EDR
128
product but with the development of UV-measurement-based estimates by the EOS Aura Ozone
129
Monitoring Instrument (OMI) Science Team as described in Vasilikov et al.[ 2008] these were found to
130
often (e.g., for thin cirrus) lead to errors as the IR optical cloud tops would be found at a lower pressure
131
than the UV optical cloud tops. Both OMPS total ozone products are currently using a climatology of
132
UV cloud optical centroids compiled from five years of OMI data. Future plans are to include direct
133
calculation of UV cloud optical centroid pressures from the OMPS measurements.
134
The total ozone product files contain both Multiple Triplet and Heritage Version 7 Triplet Total
135
Column Ozone estimates. See McPeters et al., [1996] for information on the Version 7 algorithm. The
136
heritage Version 7 products use the 318, 331 and 364 nm channels in a single triplet. Given the current
137
state of the calibration, each triplet will have its own small biases. We are working on soft calibration
138
adjustments to remove the inter-channel biases and homogenize the multiple triplet results. The nice
139
thing about the Version 7 product is that its behavior is only affected by the relative calibration of the
140
three channels in its single triplet. Its weakness is that this triplet is not ideal under all viewing
141
conditions, particularly for high ozone amounts at large solar zenith or satellite viewing angles.
142
The Version 8 total ozone algorithm used for reprocessing is described in Bhartia and Wellemeyer
143
(2002).
144
2.2. Nadir Profiler
145
The spectral measurements from the OMPS Nadir Profiler and Nadir Mapper of the radiances scattered
146
by the Earth’s atmosphere are used to generate estimates of the ozone vertical profile along the orbital
147
track. The algorithms currently use ratios of Earth radiance to Solar irradiance at a set of 12 wavelengths
148
(at approximately 252, 274, 283, 288, 292, 298, 302, 306, 313, 318, 331 and 340 nm) with the shortest
149
eight taken from the Nadir Profiler and the longest four from the Nadir Mapper to obtain estimates of the
8
150
total column ozone, effective reflectivity, and the ozone vertical profile in 12 Umkehr Layers. (The 12
151
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
152
radiances for the four longer wavelengths are obtained from the 25 Nadir Mapper FOVs co-located with
153
a single Nadir Profiler FOV. The IDPS ozone profile EDR product is made using an implementation of
154
the Version 6 Solar Backscatter Ultraviolet instruments (SBUV/2) algorithm [Bhartia et al. 1996]. The
155
longer channel radiance/irradiance ratios are used to generate estimates of the total column ozone and
156
scene effective reflectivity. This total column ozone estimate is combined with a power law ozone
157
profile in the upper layer from the shorter channels to generate a first guess ozone profile that becomes
158
the A Priori for a maximum likelihood ozone profile retrieval using the ratios for the seven shortest
159
wavelengths (currently omitting the 252 nm channel); adding the 313 nm channel at high solar zenith
160
angles (SZA). Additional information is in the OMPS Nadir Profile Algorithm Theoretical Basis and
161
Operational Algorithm Description Documents, and a volume of the Common Data Format Control
162
Book. The Version 8 Ozone Profile algorithm used for reprocessing and operationally at NOAA for the
163
SBUV/2 is described in Bhartia et al. [2013]. The Version 6 ozone profile is reported as Dobson Units
164
(milli-atm-cm) in Umkehr Layers for 12 layers – with the bottom two layers, layers 0 and 1, combined,
165
and the top layer extending upward including layers 12 and above. It is also reported as ozone mixing
166
ratios at 19 pressure levels. This latter product is obtained by taking the derivative of a spline fit of the
167
cumulative layer ozone amounts.
9
168
3. Products and internal consistency
169
The OMPS total ozone and ozone profile products from the operational system are available as
170
provisional-release data sets through links at
171
3.1. Total Column Ozone
172
The OMPS-NM total column ozone product provides estimates over the full sunlit Earth of total column
173
ozone, effective reflectivity and UV absorbing aerosols (an aerosol index) each day. The measurements
174
are binned into 35 cross-orbital-track fields of view (FOVs) on-board the spacecraft with each FOV
175
using a specific fixed set of CCD pixels. The performance requirements for the Total Ozone EDR are
176
given in Table 1. As can be seen, the requirements are stratified by ozone amount. There is an additional
177
long-term requirement that reprocessed products for climate data records should be stable at the
178
1%/decade level. The initial evaluation of the products in this section looks at the values for the retrieval
179
quantities relative to those expected from past experience as well as the cross-track consistency of the
180
results as internal checks.
http://www.nsof.class.noaa.gov/saa/products/welcome .
181
One of the initial ways to examine the performance of the products is to see if they match the expected
182
range and distribution of values. Figures 1.a, 1.b and 1.c provide a typical example of this check for the
183
primary OMPS-NM total column EDR algorithm products. They show false color global maps of (a)
184
total column ozone, (b) effective reflectivity and (c) an aerosol index for April 3, 2013 from the IDPS
185
EDR total ozone products. Maps for other days for the last 18 months – and for other monitoring
186
products – are available for viewing at the NOAA Satellite Integrated Calibration/ Validation System
187
(ICVS) web site at http://www.star.nesdis.noaa.gov/icvs/index.php.
188
Figure 1.a is for total column ozone. The colors show different levels of ozone in Dobson Units (milli-
189
atm-cm). The large amounts in the Northern Hemisphere are expected at this time of year. The black
10
190
area around the South Pole is a region of polar night where there is no sunlight to make the OMPS
191
measurements.
192
Figure 1.b is for the effective reflectivity. The colors show varying reflectivity in percent. The range
193
from close to 0% over cloud-free land (e.g.,, the Sahara desert) to close to 100% for bright clouds and
194
snow/ice covered surfaces (e.g., Antarctica) is as expected. The EDR product derives this estimate from
195
the longest wavelength of each triplet where the ozone absorption is negligible.
196
Figure 1.c is for an absorbing aerosol index. The colors show different levels of the index computed as
197
a measurement residual for the 364 nm channel using the reflectivity estimate from the 331 nm channel.
198
While some features are expected (e.g., dust over the Sahara) others, such as the North/South bands in
199
the mid-Pacific are not. Some small features over the open ocean near the Equator are produced by sun
200
glint but the large North/South features result from cross-track biases between the two channels and are
201
discussed later in this section.
202
The OMPS-NM is an extremely versatile instrument in that it can be programmed to execute a variety
203
of spatial and temporal aggregations and to select portions of the spectrum to report. Figure 2.a on the
204
left shows the derived UV effective reflectivity from a single wavelength channel (380 nm) measured
205
with high spatial resolution, 5 km by 10 km at nadir. Comparison to the EOS Aqua Moderate Resolution
206
Imaging Spectrometer (MODIS) image in Figure 2.b on the right for the same day documents the
207
capability of the OMPS TC sensor to identify 10 km or smaller surface features and characteristics. It
208
also confirms the Geolocation accuracy of the OMPS products as the overlain white land boundaries
209
follow the land/sea brightness changes.
210
Since the OMPS-NM uses a CCD array, there are essentially thousands of independent detectors. This
211
means that the products from each cross-track FOV are derived from their own set of detectors. To
212
investigate the internal consistency of the products, we compute weekly averages over a
11
213
latitude/longitude box in the Equatorial Pacific defined by 20ºS to 20ºN, 100ºW to 180ºW. This region is
214
selected because of its benign conditions – dark ocean surface, low non-cloud aerosol loading, and
215
relatively homogeneous total column ozone amounts. Averages are created for a variety of retrieval
216
products and their behavior is tracked from week to week. Figure 3 shows four weekly cross-track
217
patterns for April 2013 for the following: (a) average total column ozone, (b) 1-percentile reflectivity,
218
(c) average aerosol index, and (d) average SO2 index. The OMPS total ozone algorithm also computes
219
an SO2 Index. Each figure has the cross-track view position on the horizontal axis. Cross-track 17 is the
220
nadir position and 0 and 34 are the extreme west and east viewing angles, respectively. Plots for
221
additional months and other variables (and retrievals from other instruments) are available for viewing at
222
the ICVS web site cited above.
223
Figure 3.a shows the weekly average total column ozone. The 2% cross-track variations in the relative
224
cross-track estimates are very stable and are related to the small inaccuracies in the current
225
measurements as discussed at length in the companion papers. The shifts up and down from week-to-
226
week are consistent with the slowly varying ozone for this region.
227
Figure 3.b shows the weekly, one-percentile effective reflectivity values. These are computed by
228
collecting the effective reflectivity for each cross-track viewing position for a week and then finding the
229
one-percentile quantity for those values. The one-percentile effective reflectivity values are expected to
230
be approximately 4% for this region of the globe from climatological measurements made by other
231
instruments for cloud free scenes. The larger values observed here are due to the lack of stray light
232
correction in the SDR algorithm, inaccuracies in the radiance/irradiance calibration, and sun glint for the
233
8th to 15th cross-track viewing conditions. Changes in the SZA and the sun glint combine to produce the
234
week-to-week changes for those cross-track positions.
12
235
Figure 3.c shows the weekly average Aerosol Index (the 364 nm channel measurement residuals when
236
the effective reflectivity for 331 nm is used to predict them.) in N-value units (1 N value ~2.3%) for the
237
same region and period. The cross-track deviations represent 1% to 3% inter-channel calibration offsets
238
between the two measurements. This region of the Pacific has very low aerosol loading, so the general
239
positive offset represents an overall inter-channel bias. The bump between the 8th and 15th cross-track
240
positions is related to a failure of the reflectivity model to properly account for the wavelength
241
dependence of sun glint. The large upswing for the higher numbered cross-track positions corresponds
242
to the large North/South features seen in Figure 1.c.
243
Figure 3.d shows the weekly average SO2 Index. This SO2 Index is created by using the measurement
244
residuals from the 317, 331 and 336 nm triplet with total ozone and effective reflectivity estimated from
245
the algorithm. An SO2 amount (in DU) is computed to make the residuals 0 by using the relative SO2
246
versus O2 absorption ratios. As this assumes that the SO2 distribution is the same as the ozone profile in
247
the standard table (as interpolated for latitude and total ozone amount) it is considered an index related
248
to the presence of SO2 as opposed to a true product. The results are very repeatable from week to week
249
but show large variations with cross-track position. The estimates from this triplet are very sensitive to
250
inter-channel calibration biases and the pattern is primarily produced by differences in the relative bias
251
of the measurements at these three wavelengths across the CCD array. We expect that this product’s
252
cross-track averages will flatten out when small inter-channel biases are removed by soft calibration
253
adjustments. We are also investigating the use of triplets with less sensitivity to measurement errors. The
254
SO2 Index is shown here because the multiple triplet algorithms flag all values greater than 6 DU. As
255
this region of the Pacific has very low SO2 loading, the large positive values here are not physically
256
reasonable. The false-alarm flagging will continue to occur frequently until better calibration
257
adjustments are in use. The comparison studies in this paper have included data whether this flag is set
13
258
or not. The NASA OMPS Science Team has developed improved algorithms for estimating SO2 using
259
information in the 308 to 315 nm spectral range [Yang et al., 2009, Yang et al., 2013]. Their use in
260
operational OMPS processing is under investigation.
261
Figure 4 shows time series of daily averages over the Equatorial Pacific box for total column ozone
262
estimates from NOAA-19 SBUV/2, NASA EOS Aura OMI, and S-NPP OMPS First Guess products.
263
The SBUV/2 and OMI estimates are from Version 8 algorithms processed by NOAA and NASA,
264
respectively. The figure shows a bias of approximately –4% (10 DU) between the OMPS and SBUV/2
265
or OMI products for most of the record. This is just within the accuracy performance limit in Table 1 but
266
the primary purpose of this comparison is to check the longer term stability by comparison to two other
267
satellite products. There have been changes over time to the OMPS products in the form of new Day 1
268
Solar Flux data sets, new wavelength scales, and changes in the cloud top pressure climatology to a UV-
269
based one. The OMPS instrument is making a series of calibration measurements on Sundays leading to
270
reduced statistics once per week. These days have been removed from the means used to create this
271
figure.
272
3.2. Nadir Ozone Profile
273
The OMPS-NP ozone profile product provides estimates over the full sunlit Earth of the ozone vertical
274
profile along the nadir satellite track each day. The performance requirements for the Nadir Ozone
275
Profile EDR are given in Table 2. As can be seen, the requirements are stratified by vertical levels.
276
There is an additional long-term requirement that reprocessed products for climate data records should
277
be stable at the 2%/decade level. The initial evaluation of the products in this section looks at the values
278
for initial and final measurement residuals relative to those from Solar Backscatter Ultraviolet (SBUV/2)
279
instruments on NOAA Polar Orbiting Environmental Satellites (POES), as well as internal consistency
280
of the total ozone from summed profiles compared to other direct total ozone estimates.
14
281
Before launch it was known that the OMPS-NP measurements would be greatly affected by charged
282
particles as the satellite passed within the South Atlantic Anomaly (SAA) region. A performance
283
exclusion was prescribed and quality flags are set for retrievals in that region. Figure 5 show the size of
284
the effect on the shortest profiling wavelength at 252 nm. It compares the radiances for May 5, 2013 for
285
orbits outside the SAA with those for four passing through it. The portions of the orbital tracks cover
286
60ºS to the 20ºN. Fortunately, the OMPS-LP looks at the Earth’s atmosphere approximately 3000 km
287
(~25 degrees of latitude South) behind the satellite, so the satellite is outside of the SAA when it looks
288
back and views this region and region it is viewing when it passes through the SAA has already been
289
seen by the OMPS-NP. That is, the OMPS-NP and OMPS-LP cover each other’s deficiencies with
290
regard to SAA measurement contamination viewing regions.
291
Figure 6 compares the daily equatorial zonal mean initial measurement residuals for the six shortest
292
channels for the IDPS V6 ozone retrieval algorithm to those from the same algorithm applied to NOAA-
293
19 SBUV/2 instrument measurements. The region from 0° to 90° West Longitude is excluded to avoid
294
SAA effects. The residuals are in N-values (1 N ~ 2.3%) and are computed as measurement minus
295
radiative transfer model result differences using the first guess profile in the model. The time period is
296
from the end of February 2012 through September 2013. The residuals are computed with respect to
297
differing first guess profiles especially with respect to the use of the shortest channel; the 252 nm
298
channel is not used in generating the first guess profile or in the maximum likelihood retrieval for the
299
OMPS-NP while it is used for the SBUV/2s. The jump in the initial residuals in early June 2012 – most
300
noticeable in the 298 nm residuals – are coincident with the introduction of new Day 1 Solar Spectra for
301
the OMPS-NM SDR. The jumps in July 2012 – most noticeable in the 252, 283 and 288 nm channels –
302
are coincident with the introduction of a new Day 1 Solar spectrum for the OMPS-NP SDR. The jumps
303
in February 2013 – most noticeable in the 252 nm residuals -- are coincident with the start of weekly
15
304
updates to the dark corrections to the OMPS-NP SDRs. While there are offsets in the time series they
305
track well over time with the notable exception of the shorter channels prior to the start of weekly dark
306
updates and other SDR changes. The initial residual offsets are related to the first guess profile
307
differences, calibration biases, and the lack of stray light and intra-orbit wavelength scale corrections for
308
the OMPS-NP SDR. (The V6 uses a “first guess profile” computed from the discrete channel total ozone
309
estimate from the 318 and 331 nm channels and a power-law estimate of upper layer ozone from the 273
310
and 283 nm channels instead of a true A Priori.) Methods to use residuals and other information to inter-
311
calibrate the SBUV/2 instruments are described in DeLand et al. [2012]. These same approaches will be
312
used with OMPS and concurrent SBUV/2 records for CDR creation.
313
The OMPS-NP SDR measurements have significant contributions from out-of-range spectral stray
314
light. These were well characterized prelaunch and a measurement-based correction is under
315
development. Similar signal contamination has been seen for some of the SBUV/2 series of instruments
316
and corrections are applied to those measurements [DeLand et al. 2012]. The current stray light
317
contributions to the radiances produce false correlations in the retrieved ozone at the top of the profile
318
with the underlying surface or cloud reflectivity changes, because radiances at shorter wavelengths
319
contributing to vertical information of ozone at those altitudes are not sensitive to such changes. Figure
320
7 presents the results of an investigation into this phenomenon. The ozone profile retrievals for the
321
latitudes from 20S to 20N for the first five orbits of data for May 15, 2013 were used in the study. Cubic
322
polynomials in latitude are fit to each layer ozone data set and to the effective reflectivity as estimated
323
from the 331 nm channel. The differences between the retrieved values and these fit polynomials are
324
plotted in Figure 7 for the top four retrieval layers on the vertical axes versus the effective reflectivity
325
variations, with respect to its cubic fit, on the horizontal axes. The negative correlation is symptomatic
326
of stray light in the shorter wavelength radiances driving underestimation of ozone amounts in the upper
16
327
portions of the profile; the stray light proportion increases as the signals drop off with increasing ozone
328
absorption for the shorter wavelengths. The measurements at the shorter wavelengths provide the
329
retrieval information for the upper part of the profile. Note that the stray light contribution will be
330
present even for the low reflectivity scenes as the source radiances do not go to zero but are at the pure
331
atmospheric Rayleigh scattering levels, that is, a correction will produce increased ozone in these layers
332
even for low reflectivity scenes.
333
While the instrument throughput degradation for the OMPS-NP SDR (See Seftor et al. [2013].) is
334
estimated to be less than 0.5% over the first year of operations, and can be well approximated by linear
335
trends in time for each of the profile wavelengths, the Day One solar spectrum for the OMPS-NP does
336
not reflect the true changes in the solar flux due to changes in solar activity over a solar cycle.
337
Fortunately, one can estimate the spectral changes by the use of Mg II Indices and scale factors as
338
described in Cebula, et al. [1992]. A set of scale factors showing the relative variations in the OMPS-NP
339
solar spectra for relative changes in an Mg II core-to-wing ratio Index as estimated from the first year of
340
OMPS-NP working diffuser measurements are shown in Figure 8. The Mg II Index used to derive these
341
factors shows up to 3% variations over 27-day solar rotation at peak solar activity. (The period from
342
mid-2012 to mid-2013 has been a very active one for Solar sunspots. There are two contributing
343
influences to the initial residuals in Figure 6 as the Day 1 Solar used in the retrieval algorithm does not
344
follow the true solar and the atmospheric ozone creation and destruction are also tied to the Solar UV
345
levels. The associated variations appear in the 252-nm residuals, Figure 6.a for the OMPS-NP, and in the
346
273-nm residuals, Figure 6.b for the SBUV/2. Notice the highly expanded scale for Figure 6.b.) Options
347
are under consideration to make use of daily Mg II Index values to adjust the solar flux for these
348
variations.
17
349
4. Validation and further comparisons
350
4.1. Total Column Ozone
351
The IDPS OMPS-NM total column ozone products have been compared to similar products from
352
ground-based and satellite-based instruments. Figure 4 gave comparison of the average performance
353
over the Equatorial Pacific Box for the OMPS IDPS products and contemporaneous OMI and SBUV/2
354
ones. Additional examples are discussed below and displayed in Figures 9 and 10.
355
Figure 9 shows time series of differences of OMPS total column ozone (IDPS 1st Guess) and the OMI
356
total column ozone from the Version 8 NASA processing with estimates from measurements made by
357
ground-based Dobson instruments in Boulder Colorado (See Komhyr et al. [1989] for information on
358
this instrumentation), Mauloa Hawaii, and Lauder New Zealand. The 1st Guess product is used because
359
it has been more stable over the period (fewer algorithm changes) and the changes to the EDR product
360
have been to make it more like the 1st Guess product (e.g., switching to the use of measurement-based
361
partial cloud estimates and to the use of a UV climatology for the cloud top pressure). In addition, the
362
EDR product makes an adjustment for the temperature sensitivity of ozone absorption by using CrIS
363
near-real-time temperature profile retrievals and to the surface reflectivity by using VIIRS near-real-time
364
snow/ice retrievals. Both of these ancillary products have been evolving over the period of study. The
365
Dobson instruments values are for direct sun observations. The OMPS data are weighted average
366
overpass (distance weighted for all data with 0.5º Latitude and SEC(Latitude)º Longitude. The OMI data
367
is from the standard gridded NASA product at http://ozoneaq.gsfc.nasa.gov/OMIOzone.md.
368
Data were collected from March 2012 through June 2013. Because there was an adjustment to the Day
369
1 solar from pre-launch to in-orbit based spectra in mid 2012, the statistical study described here only
370
used the data from July 2012 to June 2013. Figures 9.a, 9.c, and 9.e shows time series of the two satellite
371
data sets differences versus the ground-based estimates. Notice the declines in the nine-point moving
18
372
averages in early 2012 for the OMPS differences. Figures 9.b, 9.d, and 9.f show scatter plots comparing
373
the differences of the satellite and ground data sets to their averages. Because all three data sets have
374
measurement and retrieval noise components, we use methods which allow for both in the regression
375
fits. That is, we follow the development in Flynn [2007] and construct a model of the form:
376
(S – AvgS – ε) = m (G – AvgG – δ)
(1)
377
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
378
set of matchup satellite-based estimates and AvgS is its average and ε is the noise term, and m is the
379
slope of the linear relationship between the two. We allow for an implicit difference in the additive term
380
as opposed to a purely multiplicative model because the ground and satellite systems will have different
381
sensitivity to ozone in the lower atmosphere due to backscatter contribution heights for the satellites
382
versus direct solar viewing by the ground system. The sensitivity to the lower layers is further
383
complicated by clouds and especially in the case of Mauna Loa and Boulder by differences in the
384
average surface pressure of the large satellite FOVs compared to the specific location of the ground
385
stations. Adjustments from 6 to 12 DU have been used to account for this difference based on Hilo HI
386
ozonesondes and standard profile lower layer amounts.
387
Subsets of the matchups for each station are created by selecting days from 7/2012 to 6/2013 where all
388
three instruments made estimates. Table 3 gives the results of statistical computations for satellite and
389
ground data set pairs with three assumptions about the relative sizes of each set’s noise term. Column 1
390
of Table 3 give a shorthand name for each station and the number of days with data. Column 2 specifies
391
the satellite data set for each row. Columns 3 and 4 give the means for the Dobson and satellite
392
instrument, respectively. Column 6, 7, and 8 gives estimates from linear regression fits with three noise
393
assumption: mG – no noise in the Dobson data; mS – no noise in the satellite data; and mE – equal-
394
relative-size, uncorrelated noise in both. Column 8 gives, σ, the standard errors on the slope estimates.
19
395
Intervals of the form [mG-σ,mG+σ] give the ranges of slopes that could be found depending on the
396
assumed distribution of the two systems’ measurements noise. Columns 9 and 10 give the estimates of
397
the standard deviations of the ε and δ noise terms in Eq. 1 for the equal-noise cases. Column 11 is the
398
Pearson correlation coefficient. Columns 12 and 13 give the expected differences between the satellite
399
and Dobson data at the maximum and minimum observed data values as determined from the equal-
400
noise model. The OMPS results show good consistency with the OMI ones but with 4 to 16 DU lower
401
values. The OMPS are also lower than the Dobson (after adjusting MLO for below-station column
402
ozone) by 3 to 24 DU. Some of these values lie outside of the allowed accuracy requirements of Table 1.
403
Since some of the differences between the satellite and the Dobson estimates are due to matchup
404
location discrepancies, the equal noise results are reasonable measures of the product performance. The
405
estimated uncertainty of the OMPS results, as reported for ε, range from 4.7 to 6.3 DU and are within
406
the allowed precision requirements of Table 1.
407
Figure 10 extends the OMPS/Dobson comparisons to a collection of 22 Dobson stations. The
408
corresponding results for the operational NOAA-19 SBUV/2 instrument total ozone estimates from the
409
Version 8 algorithm are also shown. One can see that the OMPS calibration adjustments in the early part
410
of 2012 produce a large decrease in the bias relative to both the Dobson network and the SBUV/2
411
record. The small positive bias for the SBUV/2 results is a common result for the NOAA SBUV/2 series
412
of V8 products.
413
4.2. Nadir Ozone Profile
414
The NOAA series of SBUV/2 instruments make a set of measurements similar to those used from
415
OMPS in the ozone profile retrieval. While the IDPS Version 6 algorithm has been replaced in NOAA
416
SBUV/2 operational processing with the Version 8 algorithm, the Version 6 products are still available
417
from the SBUV/2 for comparisons. In particular, the NOAA-19 platform has a similar equator crossing
20
418
time but slightly different spacing in its orbital tracks than the S-NPP platform. This leads to periodic
419
formation flying for the two platforms approximately every 12 days.
420
Figure 11 shows the results of a match up comparison for a pair of orbits on July 10, 2013. The two
421
satellites are flying in formation – orbital tracks within 50 km and sensing times with 15 minutes at the
422
Equator. The ozone profile retrievals are reported in Dobson Units for 12 pressure layers. They are
423
plotted here versus latitude, degrees north. The figures show general agreement between the retrievals
424
for the two instruments but with the OMPS-NP retrieving smaller values at the top of the profiles. As
425
discussed earlier, this is probably due to the lack of a correction for stray light. The mean layer
426
differences for OMPS minus SBUV/2 for 20º S to 20º N from the top layer to the bottom layer are -9.3, -
427
7.9, -7.0, -5.0, -0.7, -0.5, -6.1, -2.6, 6.0, 6.3, 6.1 and 4.0 %.
428
5. Summary and conclusions
429
The OMPS instruments are performing well, and initial evaluation of the products from the OMPS-NM
430
and OMPS-NP finds many positive aspects. The total column ozone has biases with other records on the
431
order of -3% and these biases are stable over periods away from known system changes. The ozone
432
profile retrievals compare well in the middle stratosphere but have significant biases at the top of the
433
profiles where stray light becomes a problem. As the SDR processing continues to mature, e.g.,
434
implementation of adjustments for stray light and wavelength scale shifts, the products will become
435
even better. Given the stability of the measurements, the prognosis for extrapolated characterization of
436
the small OMPS-NP throughput degradation and development of soft calibration adjustments to create
437
homogenized ozone profile products for OMPS and SBUV/2 is favorable.
438
The IDPS OMPS-NM total column ozone and OMPS-NP ozone profile products have been released
439
at the NOAA archive with provisional maturity. Links to the OMPS data products and readme files with
440
product caveats are provided at
http://www.nsof.class.noaa.gov/saa/products/welcome
.
21
441
Both products are undergoing evaluation by near-real-time users and will be provided in near-real-time
442
in Binary Universal Form for Representation of meteorological data (BUFR) files.
443
Acknowledgments. The work was supported by the NOAA JPSS and the NASA S-NPP programs. The
444
manuscript contents are solely the opinions of the authors and do not constitute a statement of policy,
445
decision, or position on behalf of NOAA or the U. S. Government.
446
References
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
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.pdf
Komhyr, 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.
22
476
477
478
479
480
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.
23
481
Table 1. OMPS Total Column Ozone Requirementsa
Attribute
Requirement
a. Horizontal Cell Size
b. Vertical Cell Size
50×50 km2 @ nadir
0 -60 km
c. Mapping Uncertainty 1-σ
5 km at Nadir
d. Measurement Range
50 -650 milli-atm-cm
e. 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-cm
g. Refresh
90% of sunlit globe / 24 hrs
h. Stability
1%/decade
a
Requirements apply under daytime conditions with Solar
Zenith Angles (SZA) up to 80º.
482
Table 2. OMPS Nadir Profile Ozone Requirementsb
Attribute
a. Horizontal Cell Size
b. Vertical Cell Size
Requirement
250 × 250 km2
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 km
c. Mapping Uncertainty 1-σ
25 km
d. Measurement Range
0.1 – 15 ppmv
e. 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 ppmv
f. 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 ppmv
g. Refresh
60% of sunlit globe / 7 days
h. Stability
2%/decade
b
Requirements apply under daytime conditions with Solar
Zenith Angles (SZA) up to 80º. Requirements do not apply within
the South Atlantic Radiation Anomaly region.
483
484
24
485
Table 3. Statistics for Dobson Match-Up Data Sets In Figure 9.
486
487
488
Site
# Days
Sat.
Name
AvgG
DU
AvgS
DU
mG
mS
mE
σ
δ
DU
ε
DU

MinE
DU
MaxE
DU
BOU
OMPS
308.7
293.9
0.90
0.98
0.94
0.02
6.7
6.3
0.97
-10.6
-24.1
N=335
OMI
308.7
306.3
0.93
1.00
0.96
0.02
6.4
6.1
0.97
0.3
-8.1
MLO
OMPS
256.6c
259.4
0.99
1.13
1.06
0.03
4.7
4.9
0.93
0.4c
5.9c
c
15.6c
N=217
OMI
256.6
LAU
OMPS
304.5
c
266.9
1.03
1.17
1.10
0.03
4.4
4.8
0.94
6.0
300.2
0.97
1.00
0.99
0.02
4.8
4.7
0.99
-3.3
-5.8
N=270
OMI
304.5
304.4
0.97
1.01
0.99
0.02
5.3
5.2
0.98
0.6
-1.1
c
The Dobson station is near the top of Mauna Loa. Satellite FOVs include ocean scenes. Adjustments from 6 to 12 DU
have been used to account for these scene differences based on Hilo HI ozonesondes and standard ozone profiles.
25
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
.
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.
26
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
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.
27
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
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.
28
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
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.
29
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
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.
30
666
667
668
669
670
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
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.
31
704
705
706
707
708
709
710
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
.
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.
32
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
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.
33
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
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.
34
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
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.
35
830
831
832
833
834
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.