AMTAtmospheric Measurement TechniquesAMTAtmos. Meas. Tech.1867-8548Copernicus PublicationsGöttingen, Germany10.5194/amt-10-3539-2017Ozone comparison between Pandora #34, Dobson #061, OMI, and OMPS in Boulder, Colorado, for the period December 2013–December 2016HermanJayjay.r.herman@nasa.govhttps://orcid.org/0000-0002-9146-1632EvansRoberthttps://orcid.org/0000-0002-8693-9769CedeAlexanderAbuhassanNaderPetropavlovskikhIrinahttps://orcid.org/0000-0001-5352-1369McConvilleGlennMiyagawaKojiNoirotBrandonUniversity of Maryland Baltimore County (JCET) at Goddard Space Flight Center, Greenbelt, MD, USANOAA/ESRL/GMD, Boulder, CO, USALuftBlick, Austria and Goddard Space Flight Center, Greenbelt, MD, USANOAA Earth System Research Laboratory, Boulder, CO, USACooperative Institute for Research in Environmental Sciences (CIRES), University of Colorado, Boulder, CO, USAvisiting scientist at: at NOAA/ESRL/GMD, Boulder, CO, USAretiredJay Herman (jay.r.herman@nasa.gov)27September20171093539354517May201723May201728August201729August2017This work is licensed under the Creative Commons Attribution 3.0 Unported License. To view a copy of this licence, visit https://creativecommons.org/licenses/by/3.0/This article is available from https://amt.copernicus.org/articles/10/3539/2017/amt-10-3539-2017.htmlThe full text article is available as a PDF file from https://amt.copernicus.org/articles/10/3539/2017/amt-10-3539-2017.pdf
A one-time-calibrated (in December 2013) Pandora spectrometer instrument (Pan
#034) has been compared to a periodically calibrated Dobson
spectroradiometer (Dobson #061) co-located in Boulder, Colorado, and
compared with two satellite instruments over a 3-year period (December
2013–December 2016). The results show good agreement between Pan #034 and
Dobson #061 within their statistical uncertainties. Both records are
corrected for ozone retrieval sensitivity to stratospheric temperature
variability obtained from the Global Modeling Initiative (GMI) and Modern-Era
Retrospective analysis for Research and Applications (MERRA-2) model
calculations. Pandora #034 and Dobson #061 differ by an average of
2.1 ± 3.2 % when both instruments use their standard ozone
absorption cross sections in the retrieval algorithms. The results show a
relative drift (0.2 ± 0.08 % yr-1) between Pandora
observations against NOAA Dobson in Boulder, CO, over a 3-year period of
continuous operation. Pandora drifts relative to the satellite Ozone
Monitoring Instrument (OMI) and the Ozone Mapping Profiler Suite (OMPS) are
+0.18 ± 0.2 % yr-1 and -0.18 ± 0.2 % yr-1,
respectively, where the uncertainties are 2 standard deviations. The drift
between Dobson #061 and OMPS for a 5.5-year period (January 2012–June 2017)
is -0.07 ± 0.06 % yr-1.
Introduction
A Pandora spectrometer instrument #034 (PSI, Pan #034) located on top of the NOAA
building in Boulder, Colorado, has been operating since December 2013 with
little maintenance and using the original calibration. The purpose of this
paper is to present a comparison between two co-located ozone measuring
instruments, Pandora #034 and Dobson #061, for the period December 2013
to December 2016. Additional comparisons are made with satellite overpass
data from OMI (Ozone Monitoring Instrument on board the Aura spacecraft) and
OMPS (Ozone Mapping Profiler Suite on board the Suomi NPOESS satellite). This
paper is an extension of a previously published paper (Herman et al., 2015)
that presented just 1 year of data. The results demonstrate the accuracy and
stability of both the Dobson and PSI for retrieval of total column ozone (TCO) and
serve as a validation demonstration at one location for both the fairly new
PSI and for satellite ozone data from OMI and OMPS. Part of the experiment
comparing Pandora #034 to Dobson #061 was to see if Pandora #034
would perform well over a long period without additional calibration or
adjustments. The only change made during the period 2014 to the present
(August 2017) was to replace a broken motor on the sun tracker that caused a
data gap in early 2016.
The characteristics of both the PSI and the Dobson spectroradiometer are
described in Herman et al. (2015). Briefly, the PSI consists of a small
Avantes low stray-light spectrometer (280–525 nm with 0.6 nm spectral
resolution with 5 times oversampling) connected to an optical head by a 400 μ
core diameter single-strand fiber optic cable. The spectrometer is
temperature-stabilized at 20 ∘C inside of a weather-resistant
container. The optical head consists of a collimator and lens giving rise to
a 2.5∘ FOV (field-of-view) FWHM (full width at half maximum) with light
passing through two filter wheels containing diffusers, an open hole, a UV340
filter (which blocks visible light), neutral density filters, and an opaque
position (dark current measurement). The optical head is connected to a small
sun tracker capable of accurately following the sun's center using a small
computer-data logger contained in a weatherproof box along with the
spectrometer. Pandora #034 is capable of obtaining NO2 and TCO amounts sequentially over a period of 80 s. The
integration time in bright sun is about 4 ms, which is repeated and
averaged for 30 s to obtain a very high signal-to-noise ratio and an ozone
precision of less than 1 DU, or 0.2 %
(1 DU = 2.69 × 1016 molecules cm-2).
Calculated TE using model estimates of O3 and temperature
profiles. The trend is calculated from the difference of TE from its
4-year daily mean, which is also used for year 2017, labelled Avg.
The Dobson record in Boulder started in 1966 based on an improved design from
the instrument first deployed in the 1920s (Dobson, 1931). The Dobson instrument
uses a differential absorption method to derive total column ozone from
direct-sun measurements using two UV wavelength pairs in the 300–340 nm
range (see Herman et al., 2015). The extensive Dobson network uses the
Bass–Paur (BP) ozone absorption cross sections (Bass and Paur, 1985) for
operational data processing (Komhyr et al., 1993).
All NOAA Dobson instruments are periodically calibrated against WMO world
standard Dobson #083, which in turn uses Langley method calibrations at
the Mauna Loa Observatory station (Komhyr et al., 1989). Standard lamps are
used to check Dobson spectral registration stability. Recently, in July 2017,
intermediate calibrations from Dobson #083 were applied to the Dobson
#061 ozone data record that improved its comparison with satellite data
(the calibration updates were processed by one of the co-authors, Koji
Miyagawa).
(a) shows the retrieved ozone time series
(December 2013–June 2017) for Pandora (red) and Dobson (black). (b) shows a Lowess (0.1) fit to
the each time series. (c) shows the percent difference, a linear
least-squares fit, and a Lowess (0.1) fit showing seasonal residuals.
The main sources of noise in the PSI measurement comes from the presence of
clouds or haze in the FOV, which increases the exposure time needed to fill
the CCD wells to 80% and reduces the number of measurements in 30 s.
For this comparison study, data were selected for scenes under
clear-sky conditions as determined from the Dobson A–D pair direct-sun data
record.
Comparisons of Pandora (BDM) with Dobson (BP and BDM) retrieved ozone
for Boulder, Colorado, in percent differences of retrieved ozone and
comparisons with OMI and OMPS. Slope is the value of the linear least-squares
fit, ±N is 1 SD, and p is the probability (0 to 1) that the slope is
statistically different from 0 relative to p= 0.05. The solid lines are a
Lowess (0.1) fit and a linear least-squares fit.
Accuracy in the PSI spectral fitting retrieval is obtained using careful
measurements of the spectrometer's slit function, wavelength calibration, and
knowledge of the solar spectrum at the top of the atmosphere. The current
operational PSI ozone retrieval algorithm used in this study is based on
extraterrestrial solar flux from a combination of the Kurucz spectrum
(wavelength resolution λ/Δλ= 500 000)
radiometrically normalized to the lower-resolution shuttle ATLAS-3 SUSIM
spectrum (Van Hoosier, 1996; Bernhard et al., 2004, 2005),
Brion–Daumont–Malicet (BDM) ozone cross sections (Brion et al., 1993, 1998;
Malicet et al. 1995), corrections for stray light, and an effective
ozone-weighted temperature.
The Dobson data used in this study contain the individual measurements (more
than one per day between 09:00 and 15:00 local time (LT) with almost all of the
data between 10:00 and 14:00 LT) for clear-sky direct-sun observations using
the quartz plate and A–D wavelength pairs for ozone retrieval
(Dobson label ADDSGQP). These were made available by one of the co-authors
(I. Petropavlovskikh, private communication, 2017 “Data availability” section). The NOAA Dobson total
ozone data are typically archived at WOUDC (World Ozone and Ultraviolet
Radiation Data Centre) or NDACC (Network for the Detection of Atmospheric
Composition Change) with one representative ozone value per day.
Temperature sensitivity
The PSI ozone retrieval algorithm is more sensitive to the effective
ozone-weighted average temperature than is the four-wavelength Dobson retrieval
(Redondas et al., 2014). Neglecting the temperature sensitivity creates a
seasonal difference between the two instruments. To correct for this, we use
an effective ozone temperature TE based on daily ozone-profile-weighted
altitude temperature averages (Redondas et al., 2014). The temperature and
ozone profile data were obtained from the GMI (Global Modeling Initiative)
model calculation for 2012 to 2016 (https://gmi.gsfc.nasa.gov/merra2hindcast/).
The GMI model provides atmospheric composition hindcasts using MERRA-2
(Modern-Era Retrospective analysis for Research and Applications, Version 2,
meteorology (Strahan et al., 2013; Wargan and Coy, 2016);
https://gmao.gsfc.nasa.gov/reanalysis/MERRA-2/). The simulation with
2 × 2.5∘ resolution uses the CCMI (Chemistry–Climate Modelling
Initiative; Morgenstern et al., 2017) emissions and boundary
conditions. MERRA-2 uses assimilation schemes based on hyperspectral
radiation, microwave observations, and ozone satellite measurements. The
resulting seasonal cycle for TE shows variations over the 4-year
period, while day-to-day variability is enhanced during winter and spring
seasons (Fig. 1). An estimated fifth year (2017) has been added (Fig. 1)
by forming the average of the daily temperatures from the 2013–2016 period.
The TE time series data are used for an ozone retrieval temperature
correction (TCOcorr coefficient) given in the form
TCOcorr= TCO (1 + C(T)) and O3(corr) = O3
TCOcorr (Herman et al., 2015), where C(TE) is given by Eqs. (1) and
(2).
CPandora-BDM(TE)=0.00333(TE-225)(Hermanetal.,2015),CDobson-BP(TE)=-0.0013(TE-226.7)(Redondasetal.,2014),CDobson-BDM(TE)=0.00042(TE-226.7)(Redondasetal.,2014).
As mentioned earlier, the Dobson TCO retrieval normally uses the Bass and
Paur (BP) ozone absorption coefficients, while Pandora uses the
BDM coefficients. A change in TE of
+1∘ leads to TCO changes for the Pandora (BDM), Dobson (BP),
and Dobson (BDM) instruments of +0.33, -0.13, and 0.042 %,
respectively. For a nominal TCO value of 325 DU, the change would be +1.1
and -0.4 DU, a net relative change of 1.5 DU for a 1∘K change between
Pandora (BDM) and Dobson (BP).
While BDM cross sections are not currently recommended for use in standard
Dobson processing, their use yields slightly different values of TCO and a
smaller sensitivity to temperature. The basic Dobson algorithm, based on
pairs of wavelengths, is intrinsically less sensitive to TE than
Pandora's spectral fitting retrieval.
Correlation between Pandora #034 and Dobson #061 for 2014–2016.
Correlation of Pandora #034 and Dobson #061 with OMI and
OMPS for 2014–2016.
Percent difference summary of linear fit slopes and mean differences
in Fig. 3.
Percent diff (A, B)Slope (% yr-1)ProbabilityMean (%)PointsPanelPan, Dob (BP)-0.2 ± 0.04P < 0.001-2.1 ± 1.62020APan, Dob (BDM)-0.2 ± 0.04P < 0.001-2.8 ± 1.62020BOMPS, Dob (BP)-0.09 ± 0.08P= 0.3-1.4 ± 2.1854COMI, Dob (BP)-0.18 ± 0.08P= 0.03-1.4 ± 1.9654DOMPS, Pan-0.18 ± 0.098P= 0.060.96 ± 2.7952EOMI, Pan+0.18 ± 0.096P= 0.061.1 ± 2.1624FTCO comparisons between Pandora, Dobson, OMI, and OMPS
Comparing retrieved TCO from the PSI, Dobson, OMI, and OMPS instruments shows
that there are small but significant differences between the PSI and Dobson
instruments and between the ground-based instruments and satellite-derived
values of TCO. The difference is calculated using 3-year estimates of
secular change based on a linear least-squares fit to the percent
differences (PDs) between the instruments. The cloud-free direct-sun A–D pair
Dobson ozone data are selected for comparison with time-matched
Pandora #034 retrieved ozone data (Herman et al., 2015). The
Pandora #034 retrieved ozone (every 80 s) are matched to the less
frequent Dobson #061 retrieval times that are obtained for midday solar
zenith angles (SZAs) and averaged over ± 8 min (Fig. 2a).
Each clear-sky PSI data point is an average of 2000 (early morning to
evening SZAs) to 4000 (midday SZAs) measurements obtained during 30 s.
All data for this study were clear sky within the instrument's
field of view based on the Dobson criteria for A–D-pair direct-sun clear
sky. In addition, the PSI data are averaged over a period of ±8 min
surrounding the Dobson time of measurement (two to three times per day). Since PSI
measurements are obtained every 80 s, there were an additional 10 PSI
data points averaged together to compare to each Dobson, OMI, or OMPS
measurement. The result is high signal-to-noise values for Pandora and high
precision (0.1 %). The same procedure using cloud-screened PSI data was
used for comparisons with OMI and OMPS, where they measure once or twice per
day over Boulder, Colorado. Some of the variations in the day-to-day ozone
values are driven by changes in the local weather over Boulder, Colorado
(see Fig. 14 in Herman et al., 2015), with weekly averages having much
smaller variation.
Figure 2b shows a Lowess (0.1) fits to the two time series in Fig. 2a that is
approximately equivalent to a 3-month running average. The Lowess (f)
procedure is based on local least-squares fitting using low-order
polynomials applied to a specified fraction f of the data (Cleveland, 1979)
that reduces the effect of outlier points from the mean. The smooth curves
show a small variable difference between the Dobson and Pandora time series.
Figure 2c shows the PD between the time series in Fig. 2a
and the residual seasonal variation in PD. Estimating the slope of the
least-squares fit to the percent difference can be sensitive to the selection of
the end points of the time series. This effect can be minimized by removing
the seasonal time dependence (Fig. 2c) using a low-pass filter function with
zero slope derived from the Lowess (0.1) fit. The result is shown in Fig. 3a.
Figure 3 shows the de-seasonalized PD (A, B) for six pairs
between Pandora #034, Dobson #061, OMI, and OMPS for the 3-year period
2014–2016 (summarized in Table 1). The slightly curvy Lowess (0.1) lines
about each linear fit show the residual seasonal cycles, which are too small
to have an effect on slope determination. Error estimates (Fig. 3 and Table 1)
for the linear least-squares slopes and averages are 1 standard
deviation (SD). Some of the error estimates are large enough to make the
statistical significance of the slopes marginal (see Fig. 3e OMPS vs.
Pandora; 0.18 ± 0.098, p= 0.06), while others are significant (see
Fig. 3d OMI vs. Dobson: -0.18 ± 0.08, p= 0.03) at the 2 SD level.
The significance probability parameter p is given, where p is the
probability (0 to 1) that the slope is statistically different from 0
relative to p= 0.05. Also shown are the numbers of data points in each
time series.
After removal of the residual seasonal variation in the calculated percent
differences, there still is a statistically significant drift of 0.2 % yr-1
(p < 0.001) between the Pandora #034 and Dobson #061
(Fig. 3a, b) using either BP or BDM ozone cross sections for
Dobson #061. The differences in the mean values (-2.1 and -2.8 %) are
not significant at the 2 SD level.
The linear trend (Fig. 3c, -0.09 ± 0.08 % yr-1, p= 0.3)
between the Dobson and OMPS is not significantly different from zero, while
the drift with OMI (Fig. 3d, -0.18 ± 0.08 % yr-1, p= 0.03) is
significant. This suggests that OMI ozone retrievals are drifting with
respect to OMPS and the Dobson. Extending the period from 2012 to June 2017
gives a very small but significant trend, -0.07 ± 0.03 % yr-1,
p= 0.047, for PD (OMPS, Dobson).
Calculations for Pandora #034 (Fig. 3e, f) show marginally
significant (p= 0.06) trends for Pandora #034 compared to OMPS (Fig. 3e,
-0.18 ± 0.098 % yr-1) and OMI (Fig. 3f,
+0.18 ± 0.096 % yr-1). If the Pandora #034 time series is
extended into 2017 to minimize the effect of missing Pandora data in 2016,
then the trends for Pandora compared to OMPS
(-0.2 ± 0.08 % yr-1, p= 0.013) and OMI
(0.15 ± 0.076, p= 0.05) are significant, but not
different from the shorter 2014–2016 period. The secular trends for the
difference between Pandora #034 and Dobson #061 (-0.2 % per year) are
almost the same for both Dobson BP and BDM ozone absorption coefficients even
though the temperature sensitivity using the Dobson BDM ozone absorption
coefficients is small (0.042 % ∘C-1). This suggests that
the stratospheric effective ozone temperature change is not a source for the
small differences between Pandora #034 and Dobson #061.
Figure 4 shows that the TCO between Pandora #034 and Dobson #061 is
highly correlated with 1:1 slope, and the correlation coefficient r2= 0.97
for the 3-year period 2014 to 2016. Similar correlation plots (Fig. 5)
for Pandora #034 and Dobson #061 with OMI and OMPS also show very high
correlations. The correlations in TCO are obtained after only temperature
corrections to Pandora #034 and Dobson #061 using TE (TCO pairs
similar to Fig. 2a).
The Pandora, OMI, and OMPS data used in this study are from the overpass
files located on the public websites (“Data availability” section).
Summary
Temperature-corrected Pandora #034 and Dobson #061 differ by an average
of 2.1 ± 3.2 %, with Pandora using its standard retrieval BDM ozone
absorption cross sections and Dobson using the recommended BP ozone
absorption cross sections. Pandora, as compared to Dobson, shows a small but
significant drift (-0.2 ± 0.08 % yr-1,
p < 0.001) for the 2014–2016 period. Comparisons of Pandora with
OMI and OMPS are marginally significant drifts of 0.18 ± 0.2 and
-0.18 ± 0.2 (p= 0.06) for 2014–2016, but they are significant
(0.15 ± 0.15 % yr-1, p= 0.05, and
-0.2 ± 0.16 % yr-1, p= 0.013, respectively) if the
period is extended to mid-2017 to minimize the effect of missing Pandora data
during 2016. The small Pandora and Dobson trends compared to OMPS suggest
that both instruments are stable. The conclusion is that the periodically
calibrated Dobson #061 is able to detect smaller ozone trends than a
Pandora instrument with no intermediate calibration during a 3-year period.
The longer-term trend for Dobson compared to OMPS for a 5.5-year period
(2012–June 2017) is -0.07 ± 0.06 % yr-1, p= 0.047.
All error estimates are 2 SD.
The data used in this study are available from the following sources:
OMI O3 Overpass Data (GSFC OMI Project): https://avdc.gsfc.nasa.gov/index.php?site=1593048672&id_=_28/aura_omi_l2ovp_omto3_v8.5_boulder.co_067.txt.
NPP O3 Overpass Data (GSFC NPP Project): ftp://toms.gsfc.nasa.gov/pub/omps_tc/overpass/suomi_npp_omps_l2ovp_nmto3_v02_boulder.co_067.txt.
Pandora Data (Herman, 2017): https://avdc.gsfc.nasa.gov/pub/DSCOVR/Pandora/DATA/Boulder/Pandora34/L3c/.
Dobson Data (Petropavlovskikh, 2017): ftp://aftp.cmdl.noaa.gov/data/ozwv/Dobson/WinDobson/Pandora comparisons/Dobson61 Boulder Ad-dsgqp 120213-032717_w_Header.txt.
The authors declare that they have no conflict of
interest.
This article is part of the special issue “Quadrennial Ozone
Symposium 2016 – Status and trends of atmospheric ozone (ACP/AMT
inter-journal SI)”. It is a result of the Quadrennial Ozone Symposium 2016,
Edinburgh, United Kingdom, 4–9 Sep 2016.
Acknowledgements
The authors would like to thank Susan Strahan and the MERRA-2 team for
supplying the atmospheric temperature data for Boulder, Colorado.
Edited by: Stefan Reis
Reviewed by: Robert Chatfield and two anonymous referees
References
Bass, A. M. and Paur, R. J.: The ultraviolet cross-sections of ozone, I, The
measurements, in: Atmospheric Ozone, edited by: Zerefos, C. S., Ghazi, A.,
and Reidel, D., Norwell, Mass., 606–610, 1985.Bernhard, G., Booth, C. R., and Ehramjian, J. C.: Version 2 data of the
National Science Foundation's Ultraviolet Radiation Monitoring Network: South
Pole, J. Geophys. Res., 109, D21207, 10.1029/2004JD004937, 2004.Bernhard, G., Evans, R. D., Labow, G. J., and Oltmans, S. J.: Bias in Dobson
total ozone measurements at high latitudes due to approximations in
calculations of ozone absorption coefficients and air mass, J. Geophys. Res,
110, D10305, 10.1029/2004JD005559, 2005.
Brion, J., Chakir, A., Daumont, D., Malicet, J., and Parisse, C.:
High-resolution laboratory absorption cross section of O3 Temperature effect,
Chem. Phys. Lett., 213, 610–612, 1993.
Brion, J., Chakir, A., Charbonnier, J., Daumont, D., Parisse, C., and
Malicet, J.: Absorption spectra measurements for the ozone molecule in the
350–830 nm region, J. Atmos. Chem., 30, 291–299, 1998.
Cleveland, W. S.: Robust Locally Weighted Regression and Smoothing
Scatterplots, J. Am. Stat. Assoc., 74, 829–836, 1979.
Dobson, G. M. B.: A photoelectric spectrophotometer for measuring the amount
of atmospheric ozone, Proc. Phys. Soc., 43, 324–339, 1931.Herman, J. R., Evans, R. D., Cede, A., Abuhassan, N. K., Petropavlovskikh,
I., and McConville, G.: Comparison of Ozone Retrievals from the Pandora
Spectrometer System and Dobson Spectrophotometer in Boulder Colorado, Atmos.
Meas. Tech., 8, 3407–3418, 10.5194/amt-8-3407-2015, 2015.Komhyr, W. D., Grass, R. D., and Leonard, R. K.: Dobson spectrophotometer 83:
A standard for total ozone measurements, 1962–1987, J. Geophys. Res., 94,
9847–9861, 10.1029/JD094iD07p09847, 1989.Komhyr, W. D., Mateer, C. L., and Hudson R. D.: Effective Bass-Paur 1985
ozone absorption coefficients for use with Dobson ozone spectrophotometers,
J. Geophys. Res., 98, 20451–20465, 10.1029/93JD00602, 1993.
Malicet, J., Daumont, D., Charbonnier, J., Parisse, C., Chakir, A., and
Brion, J.: Ozone UV spectroscopy. II. Absorption cross sections and
temperature dependence, J. Atmos. Chem., 21, 263–273, 1995.Morgenstern, O., Hegglin, M. I., Rozanov, E., O'Connor, F. M., Abraham, N.
L., Akiyoshi, H., Archibald, A. T., Bekki, S., Butchart, N., Chipperfield, M.
P., Deushi, M., Dhomse, S. S., Garcia, R. R., Hardiman, S. C., Horowitz, L.
W., Jöckel, P., Josse, B., Kinnison, D., Lin, M., Mancini, E., Manyin, M.
E., Marchand, M., Marécal, V., Michou, M., Oman, L. D., Pitari, G.,
Plummer, D. A., Revell, L. E., Saint-Martin, D., Schofield, R., Stenke, A.,
Stone, K., Sudo, K., Tanaka, T. Y., Tilmes, S., Yamashita, Y., Yoshida, K.,
and Zeng, G.: Review of the global models used within phase 1 of the
Chemistry–Climate Model Initiative (CCMI), Geosci. Model Dev., 10, 639–671,
10.5194/gmd-10-639-2017, 2017.Redondas, A., Evans, R., Stuebi, R., Köhler, U., and Weber, M.:
Evaluation of the use of five laboratory-determined ozone absorption cross
sections in Brewer and Dobson retrieval algorithms, Atmos. Chem. Phys., 14,
1635–1648, 10.5194/acp-14-1635-2014, 2014.Strahan, S. E., Douglass, A. R., and Newman, P. A.: The contributions of
chemistry and transport to low Arctic ozone in March 2011 derived from Aura
MLS Observations, J. Geophys. Res., 118, 1563–1576,
10.1002/jgrd.50181, 2013.
Van Hoosier, M. E.: Solar ultraviolet spectral irradiance data with increased
wavelength and irradiance accuracy, SPIE Proc., 2831, 57–64, 1996.Wargan, K. and Coy, L.: Strengthening of the Tropopause Inversion Layer
during the 2009 Sudden Stratospheric Warming: A MERRA-2 Study, J. Atmos.
Sci., 73, 1871–1887, 10.1175/JAS-D-15-0333.1, 2016.