template for agu journals: word 2000, version 1 · web viewet al. [2013]. the version 6 ozone...
TRANSCRIPT
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 ([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 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