Validation of the HIRDLS HNO3 Measurements

1
Validation of the HIRDLS HNO3 Measurements
D. E. Kinnison1, J. Gille1,2, J. Barnett3, C.
Randall2,4, L. Harvey4, A. Lambert5, R.
Khosravi1, M. J. Alexander6, P. F. Bernath7,8, C.
D. Boone7, C. Cavanaugh1, M. Coffey1, C. Craig1,
V. C. Dean2, T. Eden1, D. Ellis2, D. W. Fahey9,
G. Francis1, C. Halvorson1, J. Hannigan1, C.
Hartsough1, C. Hepplewhite3, C. Krinsky2, H.
Lee1, B. Mankin1, T. P. Marcy9,10,11, S.
Massie1, B. Nardi1, D. Packman1, P. J. Popp9,10,
M. L. Santee5, and K. A. Walker7,12.
Summary We show the first evaluation of the HNO3
data product (version 2.04.09) from the High
Resolution Dynamics Limb Sounder (HIRDLS) on the
Earth Observing System (EOS) Aura satellite. The
HIRDLS instrument obtains approximately 5500 HNO3
vertical profiles per day (Figure 1). The
resolution of the retrieved data can be described
using averaging kernels. In Figure 2, the FWHM of
a typical averaging kernel is 1.2 km in the
stratosphere. HIRDLS HNO3 data are generally good
over the latitude range of 64ºS to 80ºN and
pressure range 100 hPa to 10 hPa, with some
profiles, depending on latitude, having useful
information between 100 hPa to 161 hPa and 10 hPa
to 1.9 hPa. The individual profile precisions,
both theoretical (Figures 3-4) and measured
(Figure 5) are between 10-30, under optimal
conditions, but can be much larger, e.g., if the
HNO3 abundance is low or outside the 100 hPa to
10 hPa region. HIRDLS observations are biased
10-30 low compared to ACE-FTS (Figure 6).
HIRDLS HNO3 is in better agreement with Aura MLS
data in the 100 hPa to 10 hPa region (Figure 7).
Comparison are also shown for one Aura Validation
Experiment (AVE) flight in June 2005 (Figure 8).
For this flight, HIRDLS HNO3 is broadly
consistent with the NOAA CIMS instrument in the
100 hPa - 75 hPa region. Below this region, HNO3
filaments as measured by CIMS are not captured.
The reason for this is currently not understood.
1National Center for Atmospheric Research,
Boulder, CO, USA 2University of Colorado at
Boulder, USA 3Atmospheric, Oceanic and Planetary
Physics Department, Oxford University,
UK 4Laboratory for Atmospheric and Space Physics,
Boulder, CO, USA 5Jet Propulsion Laboratory,
Pasadena, CA, USA 6NorthWest Research Associates,
Boulder 7Department of Chemistry, University of
Waterloo, Waterloo, Ontario, Canada. 8Department
of Chemistry, University of York, Heslington,
York, UK. 9National Oceanic and Atmospheric
Administration, Earth System Research Laboratory,
Chemical Science Division, Boulder, CO,
USA. 10Also at Cooperative Institute for Research
in Environmental Sciences, University of
Colorado, Boulder, USA 11Now at 730 N. 23rd St.
300, Milwaukee, Wisconsin, USA. 12Department of
Physics, University of Toronto, Toronto, Ontario,
Canada
Figure 1 Top HIRDLS scan pattern for a 24-hr
period in June 2005 (5500 profiles shown by green
circles). Global coverage is obtained in 12
hours. Bottom Expanded view of HIRDLS
measurement locations.The ascending (magenta) and
descending (blue) HIRDLS orbits are shown. These
orbits are separated by 9.73 hours.
Figure 4 Statistical evaluation of the
theoretical precision from the HIRDLS Level-2
processor (black line in Figure 3) is shown for
three vertical regions. The top, middle, and
bottom rows show the theoretical precision PDFs
for HIRDLS observations taken with operational
scan tables 30, 13, and 23 respectively. HIRDLS
scan table operational details are described in
Gille et al., 2007. Columns left, middle, and
right show the precision PDFs for three vertical
regions.
Figure 2 HNO3 averaging kernels. Every third
kernel is shown for three different latitudes
(1.5?N 45.3?N and 63?S). For most of the
stratosphere the HIRDLS FWHM is 1.1 to 1.2 km.
Figure 3 HIRDLS Level-2 fractional uncertainty
(theoretical precision) for the same three
profiles in Figure 2. Blue dashed line is the
Forward model error green dot-dashed line is the
A Priori error red solid line is the measurement
error. Black line is the total error.
Figure 6 HNO3 profile differences of HIRDLS
minus the Atmospheric Chemistry Experiment (ACE)
- FTS instrument are shown. Absolute and
differences are in the left and right columns
respectively. The top row is for NH coincidences.
The bottom row is for SH coincidences.
Coincidence differences are for all HIRDLS and
ACE-FTS observations taken within a 2-hour period
and 500 km distance. Coincident HIRDLS profiles
were averaged and subtracted from single ACE-FTS
profiles. There are 500 and 92 HIRDLS/ACE-FTS
coincidences in NH and SH respectively. The
individual differences along with the mean and
standard deviation are shown (1-?). These
comparisons include coincident differences from
scan tables 30, 13, and 23. HIRDLS data was
filtered using only data with a theoretical
precision of ?30 (see Figures 3 and 4).
Figure 7 Comparison of MLS v2.2 HNO3 with
HIRDLS. The left panel is a comparison of over
278,000 coincidences, approximately 1000 per day.
The HIRDLS-MLS results are shown for the three
main HIRDLS scan tables (ST30 grey ST13 blue
and ST30 red). The dashed lines are the standard
deviation of the differences (1-?). Each MLS and
HIRDLS coincidence is within 1? latitude and 4?
longitude of the measured tangent points. The
right panel shows the coincident differences for
only scan table 23. For this panel, the
coincidence differences were binned into a 10?
zonal average grid. In both panels, HIRDLS data
were filtered using the theoretical precision
criteria of ?30 (see Figures 3 and 4). HIRDLS
data was not screened for clouds.
Figure 8 Profile comparisons of HIRDLS (red)
and NOAA CIMS (blue) HNO3 (ppbv). The four
profile regions shown in top panel. The UT and
latitude range of the CIMS data is listed in each
panel. Coincident HIRDLS profiles are only taken
between the given CIMS latitude ranges. The
numbers of HIRDLS profiles for this latitude
range are listed in each panel. The mean HIRDLS
profile is shown. The HIRDLS observations at each
level are included in the average if they meet a
given ?30 theoretical error criteria. The error
bars shown are equal to ? 1-std deviation of the
HIRDLS measurements that are included in the
mean.
Figure 5 HIRDLS HNO3 measured standard
deviation for two day in 2006. This result
includes both the contribution from the
theoretical precision and atmospheric
variability. The observations were binned in
potential temperature / equivalent latitude
coordinates. Using this coordinate system ensures
that the similar air masses are averaged
together. Results are shown in percentage of HNO3
mixing ratio. The black lines highlight the 10
contour.