Validation of HNO3, ClONO2, and N2O5 from the Atmospheric Chemistry Experiment Fourier Transform Spectrometer (ACE-FTS)

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. Validation of HNO3, ClONO2, and N2O5 from the Atmospheric Chemistry Experiment Fourier Transform Spectrometer (ACE-FTS) M. A. Wolff, T. Kerzenmacher, K. Strong, K. A. Walker, M. Toohey, E. Dupuy, P. F. Bernath, C. D. Boone, S. Brohede, Valéry Catoire, et al.

acid aerosols, thereby affecting both HNO 3 concentrations and the ozone budget at mid-latitudes (Hofmann and Solomon, 1989).
Of the three species that are the focus of this work, HNO 3 has been the most widely measured. The first measurements of HNO 3 in the stratosphere were made by Murcray et al. (1968), and were followed by the first space-based measurements made by 10 the Limb Infrared Monitor of the Stratosphere (LIMS) on Nimbus 7 (Gille and Russell, 1984;Gille et al., 1984). Regular ground-based Fourier transform infrared spectrometer (FTIR) measurements of HNO 3 were started in 1980 at the National Solar Observatory McMath solar telescope facility on Kitt Peak, Arizona, USA and in 1986 at the International Scientific Station of the Jungfraujoch (ISSJ) in the Swiss Alps (Rinsland 15 et al., 1991). Since then, other stations have performed continuous FTIR measurements of HNO 3 , most of them as part of the Network for the Detection of Atmospheric Composition Change (NDACC, http://www.ndacc.org). HNO 3 was measured during a series of Space Shuttle missions by the Atmospheric Trace MOlecule Spectroscopy (ATMOS) instrument, flown four times between 1985(Abrams et al., 199620 Gunson et al., 1996;Irion et al., 2002), by the CRyogenic InfraRed Radiance Instrumentation for Shuttle (CIRRIS 1A) (Bingham et al., 1997) in 1991, and by the CRyogenic Infrared Spectrometers and Telescopes for the Atmosphere (CRISTA) in 1994 (Offermann et al., 1999;Riese et al., 1999). With the launch of the Upper Atmosphere Research Satellite (UARS) in 1991, longer-term global distributions of HNO 3 were re-Printer-friendly Version Interactive Discussion EGU HNO 3 dataset to date. More recently, the Improved Limb Atmospheric Spectrometer (ILAS) on the Advanced Earth Observing Satellite (ADEOS) (Koike et al., 2000;Irie et al., 2002;Nakajima et al., 2002) and ILAS-II on ADEOS-II (Irie et al., 2006) both measured HNO 3 using infrared solar occultation.
In addition to the ACE-FTS, there are currently four satellite instruments measuring 5 HNO 3 . The Sub-Millimetre Radiometer (SMR) on Odin has been in orbit since 2001 (Murtagh et al., 2002;Urban et al., 2005), and the Michelson Interferometer for Passive Atmospheric Sounding (MIPAS) on Envisat, since 2002(Mengistu Tsidu et al., 2005Stiller et al., 2005;Wang et al., 2007a,b;Fischer et al., 2007). The Aura satellite, launched in 2004, carries another MLS (Waters et al., 2006;Santee et al., 2007) and the HIgh Resolution Dynamics Limb Sounder (HIRDLS) (Gille et al., 2007;Kinnison et al., 2007). These instruments are described in more detail below, in the context of comparisons with ACE-FTS. Stratospheric ClONO 2 was first measured by Murcray et al. (1979) and Rinsland et al. (1985) using solar infrared absorption spectroscopy from a balloon platform. Zan- 15 der and Demoulin (1988) reported on the retrieval of ClONO 2 column densities from FTIR measurements at the mountain station of the Jungfraujoch. Today, many of the FTIRs affiliated with NDACC perform ClONO 2 measurements. ClONO 2 was measured from space by ATMOS during all four Space Shuttle missions using infrared solar occultation spectroscopy (Zander et al., 1986;Rinsland et al., 1994Rinsland et al., , 1985Rinsland et al., , 1996Zander 20 et al., 1996) and by CRISTA using observations of infrared thermal emission (Offermann et al., 1999;Riese et al., 1999). CLAES was the only instrument on UARS able to detect ClONO 2 , and it provided global profiles between October 1991 and May 1993 (Mergenthaler et al., 1996). It was followed by ILAS, which measured ClONO 2 from October 1996 to June 1997 , providing the first high-latitude cov-Introduction EGU (IMK-IAA) scientific data processor. Spectroscopic measurements of N 2 O 5 are difficult due to the presence of interfering species and aerosol in the 1240 cm −1 band that is typically used for retrievals. The first detection was by King et al. (1976); for a review of early efforts to measure N 2 O 5 from the ground and balloons, see Roscoe (1991). Like ClONO 2 , stratospheric N 2 O 5 5 has been detected from space by ATMOS (Abrams et al., 1996;, CRISTA (Riese et al., 1997, 1999, CLAES (Kumer et al., 1996b, 1997, ILAS (Yokota et al., 2002;Oshchepkov et al., 2006), and ILAS-II (Wetzel et al., 2006). In addition, ISAMS, which operated on UARS from October 1991 to July 1992, detected N 2 O 5 using pressure modulated radiometry (Taylor et al., 1993;Smith et al., 1996;Kumer et al., 1997). MIPAS is again the only instrument, other than ACE-FTS, which is currently measuring N 2 O 5 from space (Mengistu Tsidu et al., 2004).
To date, ACE-FTS v2.2 HNO 3 volume mixing ratio profiles have been compared with data from the following satellite instruments: MIPAS ESA (Wang et al., 2007a), MIPAS IMK-IAA (Wang et al., 2007b), Aura-MLS (Froidevaux et al., 2006;Toohey and Strong, 15 2007; Santee et al., 2007), and HIRDLS (Kinnison et al., 2007). Additionally, they have been compared to balloon-borne measurements carried out during the Middle Atmosphere Nitrogen TRend Assessment (MANTRA) mission (Toohey et al., 2007). Mahieu et al. (2005) compared ACE-FTS v.1.0 ClONO 2 with ground-based measurements at northern latitudes and ACE-FTS v2.2 ClONO 2 profiles have been included in the vali-Stratosphere and Troposphere Retrieved by Occultation (ACE-MAESTRO) (McElroy et al., 2007). Both instruments record solar occultation spectra, ACE-FTS in the infrared (IR), and MAESTRO in the ultraviolet-visible(vis)-near-IR, from which vertical profiles of atmospheric trace gases, temperature, and atmospheric extinction are retrieved. In addition, a two channel near-IR-vis imager (ACE-IMAGER) provides profiles 15 of atmospheric extinction at 0.525 and 1.02 µm (Gilbert et al., 2007). The SCISAT spacecraft is in a circular orbit at 650-km altitude, with a 74 • inclination angle (Bernath et al., 2005), providing up to 15 sunrise and 15 sunset solar occultations per day. The choice of orbital parameters results in coverage from 85 • S to 85 • N with an annually repeating pattern, and a sampling frequency that is greatest over the Arctic and Antarc- 20 tic. The primary scientific objective of the ACE mission is to understand the chemical and dynamical processes that control the distribution of ozone in the stratosphere and upper troposphere, particularly in the Arctic (Bernath et al., 2005;Bernath, 2006, and references therein).
ACE-FTS measures atmospheric spectra between 750 and 4400 cm −1 (2.2-13 µm) 25 at 0.02 cm −1 resolution (Bernath et al., 2005). Profiles as a function of altitude for pressure, temperature, and over 30 trace gases are retrieved from ACE-FTS measure-Introduction EGU ments. The details of ACE-FTS data processing are described by Boone et al. (2005). Briefly, a non-linear least squares global fitting technique is employed to analyze selected microwindows (0.3-30 cm −1 -wide portions of the spectrum containing spectral features for the target molecule). The analysis approach does not employ constraints from a priori information (i.e., it is not an optimal estimation approach). Prior to per-5 forming volume mixing ratio (VMR) retrievals, pressure and temperature, as a function of altitude, are determined through the analysis of CO 2 lines in the spectra.
Issues have been identified in some ACE-FTS profiles and these have been flagged as Do Not Use (DNU). A continuously updated list of the DNU profiles and other data issues can be found at https://databace.uwaterloo.ca/validation/data issues.php. 10 The ACE-FTS instrument collects measurements every 2 s, which yields a typical altitude sampling of 3-4 km within an occultation, neglecting the effects of refraction that compress the spacing at low altitudes. Note that this altitude spacing can range from 1.5-6 km, depending on the geometry of the satellite's orbit for a given occultation. The actual altitude resolution achievable with the ACE-FTS is limited to about 3-4 km, 15 as a consequence of the instrument's field-of-view (1.25-mrad-diameter aperture and 650-km altitude). Atmospheric quantities are retrieved at the measurement heights. It should be noted that no diurnal corrections have been performed for any molecule retrieved from the ACE-FTS observations. For the purpose of generating calculated spectra (i.e., performing forward model calculations), quantities are interpolated from 20 the measurement grid onto a standard 1-km grid using piecewise quadratic interpolation. The comparisons in this study were performed using the 1-km grid data. Forward model calculations employ the spectroscopic constants and cross section measurements from the HITRAN 2004 line list (Rothman et al., 2005).
The precision of the ACE-FTS VMRs is defined as the 1σ statistical fitting errors from Introduction multaneously with HNO 3 . OCS is fixed to the results of an earlier retrieval step. The contribution of CFC-12 in the microwindows contains no structure, and so is accounted for with the baseline (scale and slope) parameters in the fitting routine.
There is a discrepancy between the spectroscopic constants from HITRAN 2004 in the two HNO 3 regions (one near 900 cm −1 and the other band near 1700 cm −1 ) used 10 in the ACE-FTS retrievals. Figure 1 shows the difference between using a set of microwindows near 900 cm −1 versus a set of microwindows near 1700 cm −1 . The profiles shown are an average of 100 occultations. The discrepancy between intensities in the two bands appears to be in the range of 5 to 10%. Note that both regions are required in the retrieval because the region near 900 cm −1 is the only source of information at 15 the lowest altitudes (below 10 km), while the 1700 cm −1 -band provides the only information at the highest altitudes (above 35 km). Both regions contribute information for the retrieval between 10 and 35 km. One consequence of this discrepancy is that retrieved HNO 3 VMR profiles could be noisier than they should be below 12 km. Future versions of ACE-FTS processing will scale the intensities in the band near 1700 cm Introduction EGU ±24 h and 1000 km to provide a reasonable number of coincidences. The correlative datasets, temporal and spatial coincidence criteria, and number of coincidences are summarized in Table 1 for the satellite and airborne instruments. Table 2 gives information on the FTIR locations and instruments used. Differences in vertical resolution can influence comparisons, so these have been 5 taken into account in this study. All the satellite instruments and the FIRS-2 balloon instrument have vertical resolutions that are similar to those of ACE-FTS. In these cases, no smoothing was applied to the data and the correlative profiles were linearly interpolated onto the 1-km ACE-FTS altitude grid. For instruments with lower vertical resolution than ACE-FTS (the aircraft-based 10 ASUR instrument and all ground-based FTIRs) the ACE-FTS profiles were degraded using the averaging kernel matrix and the a priori profile of the comparison instrument (Rodgers and Connor, 2003). Partial columns were calculated from all FTIR and coincident smoothed ACE-FTS profiles and used in the comparisons. The balloon-borne SPIRALE VMR profile was obtained at significantly higher vertical resolution than ACE- 15 FTS, and so was convolved with triangular functions having full width at the base equal to 3 km and centered at the tangent height of each occultation. This approach simulates the smoothing effect of the 3-4 km ACE-FTS resolution, as discussed by Dupuy et al. (2007). The resulting smoothed profiles were interpolated onto the 1-km ACE-FTS grid. Co-located pairs of VMR profiles from ACE-FTS and each validation experiment 20 (referred to as VAL in text and figures below) were identified using the appropriate temporal and spatial coincidence criteria. Then the following procedure was applied to the vertical profile measurements used in this assessment, with some modifications for the individual balloon-borne profile comparisons and the FTIR partial column comparisons (see Sects. 5 and 6 for details).

25
(a) Calculate the mean profile of the ensemble for ACE-FTS and the mean profile for VAL, along with the standard deviations on each of these two profiles. These mean profiles are plotted as solid lines, with ±1σ as dashed lines, in panel (a) of the comparison figures discussed below. The standard error on the mean, also known as the Introduction EGU uncertainty in the mean, is calculated as σ(z)/ N(z), where N(z) is the number of points used to calculate the mean at a particular altitude, and is included as error bars on the lines in panel (a). Note: in some cases, these error bars, as well as those in panels (b) and (c) (see below) may be small and difficult to distinguish. (b) Calculate the profile of the mean absolute difference, ACE-FTS−VAL, and the standard deviation 5 in the distribution of this mean difference (Note that the term absolute, as used in this work, refers to differences between the compared values and not to absolute values in the mathematical sense). To do this, the differences are first calculated for each pair of profiles at each altitude, and then averaged to obtain the mean absolute difference at altitude z: where N(z) is the number of coincidences at z, ACE i (z) is the ACE-FTS VMR at z for the i th coincident pair, and VAL i (z) is the corresponding VMR for the validation instrument. This mean absolute difference is plotted as a solid line in panel ( ence, as a percentage, defined using: where MEAN i (z) is the mean of the two coincident profiles at z for the i th coincident pair. Panel (c) of the comparison figures presents the mean relative difference as a solid blue line, along with the relative standard deviation as dashed lines, and the relative uncertainty in the mean as errors. Equation (3) gives the same weight to ratios with 5 extremely small denominators, which contain, in relative terms more noise, thus overestimating the relative differences for these cases (von Clarmann, 2006). Therefore, we have calculated additionally the relative deviation from the mean using: The relative deviation is added as a solid cyan line with its standard deviation as a dashed cyan line, in panel (c) for the ClONO 2 and N 2 O 5 comparisons, where small VMRs at the lowest and highest altitude levels lead to overestimated relative differences. The Odin satellite was launched in February 2001 into a near-polar, sun-synchronous, 600-km altitude orbit with an 18:00 ascending node (Murtagh et al., 2002). The Submillimetre Radiometer (SMR) observes limb thermal emission from HNO 3 on roughly 5 two measurement days per week using an auto-correlator spectrometer centered at 544.6 GHz. Operational Level 2 HNO 3 retrievals are produced by the Chalmers University of Technology (Göteborg, Sweden).
Here we use Chalmers v.2.0 HNO 3 profiles, which have a horizontal resolution of ∼300-600 km, vertical resolution of 1.5-2 km, and single-scan precision better than 10 1.0 ppbv over the range 18 to 45 km (Urban et al., 2006(Urban et al., , 2007. The estimated total systematic error is less than 0.7 ppbv throughout the vertical range (Urban et al., 2005(Urban et al., , 2006. The ACE-FTS-SMR coincidence criteria employed were ±12 h and 500 km. Whenever multiple SMR measurements were found to be coincident with the same ACE-FTS occultation, the SMR observation closest in distance was used. From these 15 coincident measurements between February 2004 and November 2006 any SMR scan with a data quality flag value not equal to 0 was discarded. Furthermore, pairs of coincident data points were removed when either the ACE-FTS relative error exceeded 100% or the SMR response was below 0.75 (indicating that a priori information contributed significantly to the retrieved value) (Urban et al., 2005;Barret et al., 2006). 20 The number of remaining coincident pairs used in the comparisons are shown along the right hand axis in Fig. 2d. The decrease in the number of comparison pairs below 20 km is due to declining SMR response, while above 32 km it is due to an increasing relative error in the ACE-FTS HNO 3 retrievals. Figure 2 shows the statistical comparisons of all coincident profiles. Seasonal and/or 25 latitude-limited comparisons were found to be of similar character, as were comparisons separated into SMR daytime or nighttime groups (not shown). The SMR and ACE-FTS mean profiles (Fig. 2a)  EGU maximum at different altitudes. The ACE-FTS HNO 3 maximum (∼23 km) is at a higher altitude than the SMR maximum (∼21 km). The magnitude of the standard deviation of the means in Fig. 2 suggests that the SMR data is considerably noisier particularly above 30 km. The ACE-FTS VMR is typically 1.7 parts per billion by volume (ppbv), and at most 2.7 ppbv, smaller than SMR in the lower stratosphere (18-27 km). Above 5 27 km, the ACE-FTS VMR is typically 0.5 ppbv (at most 0.7 ppbv) larger than SMR (Fig. 2b). The mean relative difference (Fig. 2c) exceeds −100% at 17.5 km. This negative difference decreases towards higher altitudes and changes to positive relative differences at 27 km. Typically, it is ∼15% (31%, at most) between 27 and 35 km. This behaviour suggests an altitude shift between the two instruments, as was ob-10 served in MIPAS IMK-IAA -SMR HNO 3 comparisons by Wang et al. (2007b). Wang et al. (2007b) suggested an altitude shift of 1.5 km which is consistent with that found in MLS-SMR comparisons (Santee et al., 2007). To test this, an altitude shift of +1.5 km was applied to all SMR profiles. The shifted SMR profile and the comparison with the ACE-FTS are also shown in Fig. 2. For the shifted SMR mean profile, the HNO 3 15 maximum is at the same altitude as seen by ACE-FTS, around 23 km. That seems to confirm the existence and the size of the altitude shift as seen by the aforementioned satellite comparisons. Santee et al. (2007) suggested that it might be caused by systematic errors in the SMR 544.6 GHz pressure/temperature and pointing retrievals. The ACE-FTS HNO 3 VMRs are still up as much as 20% smaller than the shifted SMR 20 values between 18 and 35 km, corresponding to a mean negative bias of -1 ppbv and a maximum negative bias of −1.9 ppbv at 25 km. These values are similar to the differences between MIPAS IMK-IAA and the altitude-shifted SMR as seen by Wang et al. (2007b), who concluded that other error sources (spectroscopy, calibration) may also contribute to the disagreement. Introduction EGU altitudes between 18 and 22 km for both ACE-FTS and SMR. The 1.5-km altitude shift has been applied to the SMR data used in the plot.

Aura-MLS: HNO 3
The Microwave Limb Sounder (MLS) was launched on the Aura satellite in July 2004. It is in a sun-synchronous orbit at an altitude of 705 km and an inclination of 98 • , with the 5 ascending node crossing the equator at 13:45 (local time) (Waters et al., 2006). Global measurements are obtained daily from 82 • S to 82 • N, with 240 scans per orbit. Like SMR, MLS measures atmospheric thermal emission in the limb. Seven radiometers are used to provide coverage of five spectral regions between 118 GHz and 2.5 THz. The standard MLS HNO 3 product is derived from the 240 GHz retrievals at and below (i.e., at pressures equal to or larger than) 10 hPa and from the 190 GHz retrievals above that level (Livesey et al., 2007). The retrieval is performed on a pressure grid with six levels per decade for pressures greater than 0.1 hPa and three levels per decade for pressures less than 0.1 hPa using the optimal estimation approach described by Livesey et al. (2006). The vertical resolution for the HNO 3 VMR profiles is 3.5-5 km, the along-track horizontal resolution is 300-500 km. Validation of the MLS v2.2 HNO 3 data product is described by Santee et al. (2007). The precision of the individual MLS v2.2 HNO 3 profiles is estimated to be ∼0.6-0.7 ppbv, and the recommended pressure range for the use for scientific studies is 215-3.2 hPa (Livesey et al., 2007). Santee et al. (2007) compared ACE-FTS v2.2 HNO 3 with MLS HNO 3 measurements. 20 They found that ACE-FTS values are slightly larger than those from MLS but agree to within 0.5-1 ppbv on average, corresponding to ∼10% between 19 and 30 km and to ∼30% above. Below 19 km, the differences increased and exceeded 50% where average VMRs are very low. For their study, Santee et al. (2007) Figure 4 shows the statistical comparisons of all coincident ACE-FTS and MLS profiles. In agreement with the results of Santee et al. (2007), the ACE-FTS mean HNO 3 profile is consistently ∼0.6 ppbv (maximum difference = 0.8 ppbv) larger than that of MLS. The mean relative differences of 5 the global comparisons are less than 23% between 18 and 32 km, and reach a minimum of 7% at approximately 25 km. The relative differences reach maxima of ∼30% at the top and bottom of the altitude range where the mean HNO 3 profile reaches its lowest values. The statistical comparison is divided into five latitude bands in Fig. 5. The relative 10 differences in the northern (Fig. 5, part 1, middle row) and southern (Fig. 5, part 2, top row) midlatitude bands are ∼10% between 18 and 27 km, within 20% between 28 and 32 km, and increase to 35% above 32 km. At the lowest altitudes, 15-18 km, the mean relative difference reaches 50% for the northern mid-latitudes and exceeds 100% for the southern mid-latitudes. The HNO 3 profiles in the polar latitude bands (Fig. 5, part 1, 15 top row and part 2, bottom row) agree to within 20% between 18 and 30 km and within 40% above and below this range. The standard deviation of the mean relative difference increases dramatically below 22 km for the 60 • -90 • S latitude band, indicating a large spread in the differences between the individual comparisons. No distinction was made between measurements in and outside the polar vortex. Therefore, the large 20 variance is a result of large spatial gradients in HNO 3 across the polar vortex edge at these altitudes in winter when denitrification drastically reduces the HNO 3 VMRs inside the vortex (Santee et al., 2004). Due to the typically lower HNO 3 values in the tropical regions, the mean relative differences are largest in the 30 • S-30 • N latitude band (Fig. 5 (Fischer et al., 2007). MIPAS provides nearly pole-to-pole coverage (87 • S-89 • N) every day, measuring continuously around an orbit in both day and night.

5
It acquires emission spectra over the range 685-2410 cm −1 (14.5-4.1 µm), which includes the vibration-rotation bands of many molecules of interest. From July 2002 until March 2004, MIPAS was operated at full spectral resolution (0.025 cm −1 ) with a nominal limb-scanning sequence of 17 steps from 68-6 km with 3 km tangent height spacing in the troposphere and stratosphere, generating complete profiles spaced ap-10 proximately every 500 km along the orbit. In March 2004, operations were suspended following problems with the interferometer slide mechanism. Operations were resumed in January 2005 with a 35% duty cycle and reduced spectral resolution (0.0625 cm −1 ). The European Space Agency (ESA) produces profiles of pressure, temperature, and six key species, among them HNO 3 . The algorithm used for the Level 2 analysis is described in detail by Ridolfi et al. (2000), Carli et al. (2004), and Raspollini et al. (2006). Complementary to the ESA operational data products, several different off-line data processors are in use for science-oriented analysis of the MIPAS data (von Clarmann et al., 2003). The MIPAS IMK-IAA data processor was developed at the IMK, Germany, including a component to allow non-local thermodynamic equilibrium treatment 20 from the IAA, Spain (von Clarmann et al., 2003). HNO 3 , ClONO 2 , and N 2 O 5 are three of the trace gases retrieved with the MIPAS IMK-IAA processor and are available at http://www-imk.fzk.de/asf/ame/envisat-data/. EGU HNO 3 comparisons were defined in both papers as ±9 h, 800 km, and a maximum potential vorticity (PV) difference of ±3×10 −6 K m 2 kg −1 s −1 at 475 K potential temperature. Wang et al. (2007a) and Wang et al. (2007b) compared about 600 daytime and nighttime MIPAS profiles to about 350 ACE-FTS coincident profiles, separated into two different latitude bands: 30-60 • and 60-90 • , resulting in a mean distance of 5 280±151 km and a mean time difference of 7.1±8.4 h. The consistency between both MIPAS HNO 3 products (ESA and IMK-IAA) and ACE-FTS HNO 3 was found to be very good. The mean differences were between ±0.1 and −0.5 ppbv for the ACE-FTS versus MIPAS ESA data product comparisons (Wang et al., 2007a) and between ±0.1 and −0.7 ppbv for the ACE-FTS versus MIPAS IMK-IAA data product comparisons (Wang 10 et al., 2007b). That corresponds to relative differences between ±5 and ±10% for altitudes between 10 and 30 km and between ±10 and ±15% for altitudes above (up to 35 km) (Wang et al., 2007a,b).
In both papers, data were analysed for the period 9 February to 25 March 2004, including data from the ACE satellite commissioning period which continued until 21 15 February 2004. We recalculated the comparisons between ACE-FTS sunset observations and MIPAS for the period 21 February to 25 March 2004 using only data from the ACE Science Operations period. Figures 6 and 7 show the results of these revised comparisons.
For the comparison with the MIPAS ESA data used in this work (v4.62), we narrowed 20 the coincidence criteria to ±6 h and 300 km, resulting in 138 coincident profiles, shown in Fig. 6. The mean difference between ACE-FTS and MIPAS ESA HNO 3 is typically −0.1 ppbv and varies between −0.71 ppbv at 27.5 km and +0.33 ppbv at 30.5 km. That corresponds to typically ±2% between 10 and 27 km and to ±9% between 27 and 36 km. A maximum relative difference of −25% is obtained for the highest comparison 25 altitude of 36.5 km.
The comparison between the ACE-FTS and MIPAS (IMK-IAA v8) HNO 3 products was calculated using the same coincidence criteria as defined by Wang et al. (2007b) and is shown in Fig. 7. Between 10 and 31 km, ACE-FTS is typically is 0.2 ppbv smaller ACPD 8,2008 10 and a maximum PV difference of ±3×10 −6 K m 2 kg −1 s −1 at 475 K potential temperature. When combining all coincidences, the mean differences between ACE-FTS and MIPAS ClONO 2 were found to be less than 0.04 ppbv (<5%) up to altitudes of 27 km. At nearly all altitudes, ACE-FTS reported smaller VMR values than MIPAS. Above 27 km, the differences increased to around −0.15 ppbv (−30% at 34.5 km). In the altitude range between 15 and 19 km, slightly enhanced differences of up to −0.03 ppbv could be observed (Höpfner et al., 2007). The high-altitude bias was assumed to be photochemically induced. Therefore, Höpfner et al. (2007)

EGU
Relative differences were used to account for any problems with the absolute values of modeled NO y . The expression used is: In the resulting comparison between ACE-FTS and the CTM-corrected MIPAS ClONO 2 VMRs, the maximum absolute differences were reduced and no systematic 5 bias up to 27 km altitude was seen. At higher altitudes, however, the model overcompensated for the photochemically-induced bias and the corrected MIPAS ClONO 2 values were up to 0.1 ppbv smaller than those measured by ACE-FTS (Höpfner et al., 2007).
For this paper, we recalculated the comparison between ACE-FTS and MIPAS ClONO 2 using IMK-IAA v11 for the period 21 February to 25 March 2004, considering only the ACE-FTS data after the start of the ACE Science Operations period. The results of the comparisons, which do not change significantly the findings of Höpfner et al. (2007), are shown in Fig. 8. The ACE-FTS ClONO 2 values are smaller than the uncorrected MIPAS product for all altitudes. The mean relative differences are better 15 than −7% between 16 and 27 km, and reach −30% at 34 km (Fig. 8, top row). The comparison between ACE-FTS and the CTM-corrected MIPAS ClONO 2 profiles shows no systematic difference between 16 and 27 km. Typically mean relative differences are within ±1%, reaching a maximum of −6% around 16-17 km. Above 27 km, ACE-FTS ClONO 2 is larger than the corrected MIPAS values with a maximum relative difference 20 of 22% around 33 km (Fig. 8, bottom row), suggesting that the model is overcompensating as observed in the previous study.
As explained in Sect. 3, Eq. (3) overestimates the relative differences in the lowest altitude region, 13-16 km, when some denominators are extremely small. Therefore, profiles of the relative deviation of the mean, calculated with Eq. (4), are also included in Fig. 8. The relative deviation of the mean clearly shows that ACE-FTS is very consistent with MIPAS ClONO 2 also at lower altitudes, differing not more than −6% between 2452 Introduction The retrieval method and characteristics of N 2 O 5 profiles inverted from MIPAS observations have been described by Mengistu Tsidu et al. (2004). N 2 O 5 is retrieved from its infrared emission in the ν 12 band in the spectral range from 1239-1243 cm −1 . Spec-5 troscopic data for N 2 O 5 by Wagner and Birk (2003) were taken from the HITRAN 2004 database (Rothman et al., 2005). At the altitude of the N 2 O 5 VMR maximum (around 30 km), ACE-FTS VMRs are ∼0.5 ppbv (75%) smaller than MIPAS IMK-IAA daytime observations and ∼0.4 ppbv (70%) smaller than the MIPAS IMK-IAA nighttime observations. At altitudes below the VMR maximum, these differences decrease in absolute terms. In relative terms, however, largest differences appear at around 18 km and at the highest altitudes, just below 40 km.

EGU
To account for the diurnal cycle of N 2 O 5 and the different local observation times of MIPAS and ACE-FTS, we have performed a correction using the KASIMA CTM (Kouker et al., 1999), as was done for ClONO 2 . Rows 2 and 4 of Fig. 9 show results of the CTM-corrected comparisons for MIPAS IMK-IAA daytime and MIPAS IMK-IAA nighttime measurements, respectively. In both cases, the large differences at the VMR The measured spectra are integrated over 90 s to achieve a sufficient signal-noise-ratio. 15 The horizontal resolution of the HNO 3 profile is ∼20 km, which depends on the ground speed of the aircraft and the integration time. The vertical profiles are retrieved on a 2-km grid, using the optimal estimation method (Rodgers, 2000). The vertical resolution of the HNO 3 measurement is 6-10 km in the lower stratosphere and a retrieval is possible between 15 and 35 km. The precision of a typical measurement is 0.3 ppbv 20 and the estimated accuracy is ∼0.6 ppbv or 15%, whichever is higher (von König et al., 2000;Kleinböhl et al., 2003 Figure 10 shows the results from the comparison. The ACE-FTS VMRs are slightly larger in the lower stratosphere and smaller in the middle stratosphere than those VMRs from ASUR. The ACE-FTS-ASUR differences are up to 2.5 ppbv or 32% in the lower stratosphere (between 18 and 26 km) and are up to −0.3 ppbv or −6% in the middle stratosphere (28-36 km). The sign of these differences changes at 27.5 km. validation campaigns for Odin and Envisat. SPIRALE performs simultaneous in situ measurements for about ten chemical species using six tunable lasers (Moreau et al., 2005). Measurements are done during the balloon ascent from about 10 to 35 km height, with a high frequency sampling (∼1 Hz), thus providing a vertical resolution of only a few meters. The diode lasers emit in the mid-infrared domain (from 3 to 20 8 µm) with beams injected into a 3.5-m-long multipass Heriott cell located under the gondola and largely exposed to ambient air. A total optical path length of 430.78 m is obtained by multiple reflections between the two cell mirrors. Species concentrations are retrieved from direct infrared absorption, by fitting experimental spectra with spectra calculated using the HITRAN 2004 database (Rothman et al., 2005). The 25 species concentration can be converted into VMR using the on-board pressure and temperature measurements. Specifically, the ro-vibrational lines in the microwindow 1701.5-1701.8 cm −1 were used for HNO 3 (Moreau et al., 2005).

EGU
The global uncertainties for the VMRs have been assessed by taking into account the random errors and the systematic errors, and combining them as the square root of their quadratic sum. The two important sources of random errors are the fluctuations of the laser background emission signal and the signal-to-noise ratio. The laser line width and the non-linearity of the detector contribute to the systematic errors. The resulting 5 global uncertainty is estimated to be 25% below 17 km, and 20% above. With respect to the above errors, systematic errors in the spectroscopic data (essentially molecular line strength and pressure broadening coefficients) are considered to be negligible.
The SPIRALE measurements occurred on 20 January 2006 between 17:46 UT and 19:47 UT, with vertical profiles obtained between 13.7 and 27.2 km altitude. The mea-  (Hauchecorne et al., 2002). PV maps in the region of both measurements have been calculated each hour between 17:00 UT on 20 January and 08:00 UT on 21 January on isentropic surfaces, every 50 K from 350 K to 800 K (corresponding to 13-30 km height). From these PV fields, it can be deduced that SPIRALE and ACE-FTS vertical profiles were located in similar air masses in the well-established polar vor-20 tex for the whole range of altitudes sounded by SPIRALE. The dynamical situation was very stable with PV agreement better than 10%. So the geophysical situation is suitable for direct comparisons. As mentioned in Sect. 3, SPIRALE data were smoothed with a triangular weighting function of 3 km at the base (corresponding to ACE-FTS resolution). Consequently, the bottom and the top of the SPIRALE profile have been 25 truncated by 1.5 km. The resulting profile was subsequently interpolated onto the 1km ACE-FTS grid. Possible diurnal variations due to the different times of the day of the measurements (SPIRALE flew at night and ACE-FTS measurements were at sunrise) have been examined with a photochemical box model (McLinden et al., 2000). It Interactive Discussion EGU appears that the diurnal variations in HNO 3 were negligible. Figure 11 shows that the ACE-FTS HNO 3 profile is systematically larger than the SPIRALE profile. Between 15 and 23 km, ACE-FTS and SPIRALE agree to within 45% and within 13% between 23 and 26 km. The low HNO 3 values observed by SPIRALE in the 20.7-22 km layer are probably due to the polar stratospheric cloud (PSC) that 5 SPIRALE encountered from 19.3 to 20.7 km, which was detected by an aerosol counter aboard the gondola. Using HYSPLIT (Draxler and Hess, 1998a,b) backward trajectories above 20.7 km, it appears that the temperature encountered along the trajectories was close to the nitric acid trihydrate equilibrium temperature during the two days before the measurements. The low temperatures encountered by the air parcel probably 10 allowed formation of PSC particles with large size (greater than 1 µm), leading to a denitrified layer. By the time the SPIRALE measurements were made, the PSC had sedimented.

FIRS-2 balloon: HNO 3 , ClONO 2 , N 2 O 5
The balloon-borne Fourier transform infrared spectrometer FIRS-2 (Far-InfraRed 15 Spectrometer-2) was designed and built at the Smithsonian Astrophysical Observatory. It has contributed to previous satellite validation efforts (e.g., Jucks et al., 2002;Nakajima et al., 2002;Canty et al., 2006). FIRS-2 detects atmospheric thermal emission in limb-viewing mode from approximately 7 to 120 µm (∼80-1350 cm −1 ) at a spectral resolution of 0.004 cm −1 (Johnson et al., 1995). Vertical profiles of about 30 trace gases 20 are retrieved from the float alitude (typically 38 km) down to the tropopause using a nonlinear Levenberg-Marquardt least-squares algorithm, with pressure and temperature profiles dervied from the 15 µm band of CO 2 . HNO 3 is retrieved from the ν 9 band between 440 and 470 cm −1 . The retrievals from the ν 5 and 2ν 9 bands, made with the HITRAN 2004 dataset differ systematically by 2%. ClONO 2 is retrieved jointly with the 25 ν 5 band at 560 cm −1 and the q-branch of the 720 cm −1 band. N 2 O 5 is retrieved from the band between 710 and 770 cm −1 . However, the signal-to-noise ratio is not very good EGU for this particular retrieval. Uncertainty estimates for FIRS-2 contain random retrieval error from spectral noise and systematic compontents from errors in atmospheric temperature and pointing angle (Johnson et al., 1995;Jucks et al., 2002). We compare ACE-FTS HNO 3 , ClONO 2 , and N 2 O 5 profiles with the data obtained during a FIRS-2 balloon flight from Esrange, Sweden on 24 January 2007 at 10:11 UT. The average 5 location of the flight was 67.27 • N and 27.29 • E, with some smearing of the longitude footprint as FIRS-2 was observing to the east. The data were recorded before local solar noon with a solar zenith angle of 86.6 • . The float altitude was just under 28 km, limiting the maximum measurement altitude to 31 km. The closest ACE-FTS occultation was sr18561, obtained on 23 January 2007, at 08:25 UT (64.70 • N, 15.02 • E), 10 placing it 26 h earlier and 481 km away from the location of the balloon flight. The FIRS-2 trace gas profiles are reported on a 1-km grid and were interpolated onto the ACE-FTS 1-km grid. Figure 12 shows the comparisons of the VMR profiles of HNO 3 (top row), ClONO 2 (middle row), and N 2 O 5 (bottom row) measured by FIRS-2 and ACE-FTS. Scaled 15 (Dunkerton and Delisi, 1986;Manney et al., 1994) PV values for the times and locations of both measurements indicate that both instruments measured airmasses inside the polar vortex. At the time of the FIRS-2 flight, PSCs could be observed from the ground and the scattering of the upwelling radiation in the spectra indicated that the balloon gondola passed through a PSC during the flight. Also, there is a slight enhancement 20 in the ACE-IMAGER extinction data at 20 km for this occultation, which may have been caused by the presence of PSCs.

HNO 3
The ACE-FTS HNO 3 VMR profile shows values up to 3 ppbv smaller than the FIRS-2 VMR from 15 to 18 km (−25%) and from 26 to 31 km (−55%). In the altitude range 25 between 19 and 25 km, ACE-FTS measured values that were up to 3 ppbv larger than FIRS-2. Relative differences in this altitude range reach at most 50% at 20 km. The low FIRS-2 HNO 3 values between 19 and 25 km are very likely due to a denitrified 2459 ACE-FTS HNO 3 and ClONO 2 measurements were also compared with partial columns retrieved from solar absorption spectra recorded by ground-based Fourier transform infrared spectrometers. All of the FTIR instruments are located at NDACC stations, except the Poker Flat FTIR which is a NDACC candidate instrument, currently waiting 15 for its certification. Table 2 lists the stations, their locations, and further details regarding the instrument type, the spectral resolution, the retrieval code and the microwindows used to retrieve HNO 3 and ClONO 2 . The references in Table 2 provide more information about the instruments, the retrieval techniques and the measurements made at each station. The 20 participating sites span latitudes from 77.8 • S to 76.5 • N. The geographical locations of these sites are shown in Fig. 13.
The FTIR data were analyzed using either the SFIT2 retrieval code Rinsland et al., 1998) or PROFFIT92 (Hase, 2000). Hase et al. (2004) showed that VMR profiles and total columns retrieved Introduction EGU within 1%). Considering the slightly different handling of spectroscopic data for the ClONO 2 retrieval, one can expect that PROFFIT92 and SFIT2 retrieved ClONO 2 agree within ±2%. Both algorithms employ the optimal estimation method (Rodgers, 2000) to retrieve vertical profiles from a statistical weighting between a priori information and the high-resolution spectral measurements. Averaging kernels calculated as part of this analysis quantify the information content of the retrievals, and can be convolved with the ACE-FTS profiles, which have higher vertical resolution. The information required for the retrievals, such as a priori profiles and covariances, treatment of instrument lineshape, and atmospheric temperature and pressure are optimized for each site as appropriate for the local conditions.
The coincidence criteria used for the FTIR comparisons are ±24 h and 1000 km, with three exceptions. For the high-latitude stations Kiruna and Thule, tighter criteria of ±12 h and 500 km were used, in order to minimize the influence of the polar vortex. Note that for Poker Flat and Arrival Heights, these tighter criteria would have reduced the number of coincidences too much. Therefore, the original criteria (±24 h, 1000 km) 15 were kept for these two high latitudes stations. For Reunion Island, the criteria were ±24 h, ±10 degrees latitude, and ±15 degrees longitude, resulting in a maximum spatial difference of 1211 km. These relatively relaxed criteria were necessary to obtain a reasonable number of ACE overpasses for each station (between 5 and 29). In cases where several ACE-FTS occultations met the coincidence criteria for one FTIR mea-20 surement at a site, only the closest ACE occultation (optimized for temporal and spatial differences) was used.
The comparisons include coincident measurements from March 2004 through October 2006. For each station, the ACE-FTS profiles were interpolated onto the FTIR retrieval grid and extended below the lowest retrieved altitude using the FTIR a priori 25 VMR values. This combined profile was smoothed using the FTIR averaging kernels and a priori profile, as described in Sect. 3, to minimize the smoothing error (Rodgers and Connor, 2003). Atmospheric density profiles were calculated based on the pressure and temperature profiles used in the FTIR retrievals. These FTIR density profiles ACPD 8,2008  EGU were used to calculate partial columns from the smoothed ACE-FTS profiles and the retrieved FTIR profiles, except for Jungfraujoch where the ACE-FTS partial columns were calculated using the ACE-FTS density profiles. The altitude ranges of the partial columns were determined separately for each station and species. Each altitude range is limited to the altitude levels that fulfill two 5 criteria: (1) ACE-FTS retrievals have to exist, and (2) the sensitivity of the FTIR measurements has to be 0.5 or greater, indicating that the measurement contributes at least 50% to the retrieved profile. The latter can be determined from the FTIR a priori information and the averaging kernel matrix (Vigouroux et al., 2007). The chosen altitude ranges are listed in Tables 3 and 4. For the retrievals of HNO 3 , all FTIR sites used spectroscopic data from the HITRAN 2004 database (Rothman et al., 2005). All participating sites used microwindows in the region 860-875 cm −1 as listed in Table 2. The Degrees of Freedom of Signal (DOFS) of the FTIR partial columns (equal to the trace of the averaging kernel matrix) are 15 between 1 and 2, thus indicating there is enough independent information for a partial column.
The time series of the HNO 3 partial column comparisons are shown for all stations in Fig. 14, along with the relative differences as a percentage of the FTIR partial columns. The agreement between ACE-FTS and the FTIRs is typically ±20%, and does not 20 exceed ±50% except for two cases measured at Arrival Heights, where the two ACE-FTS partial columns differ by ±150% from the four coincident FTIR partial columns. Table 3 summarizes these results, listing the mean relative differences (mean of the N differences (ACE-FTS-FTIR)/FTIR), the standard deviations, and standard errors on the mean. 25 The largest relative differences (given with standard error), −11.4% ± 8.7% and −12.6% ± 14.4%, are reported for the comparisons with the two high-latitude stations Poker Flat and Arrival Heights, respectively. Both stations include a large number 2463 Introduction EGU of measurements in the polar winter and spring, the period with the greatest vortex activity. The large standard deviation of 59.2% for Arrival Heights is caused by two ACE-FTS occultations showing one exceptionally high and the other exceptionally low partial column values, both occuring in the late austral winter. Neglecting all winter and spring measurements for Arrival Heights improves the agreement between ACE-5 FTS and FTIR to 3.8% and decreases significantly the standard deviation on the mean difference to 20.4%. We suggest that the larger differences reported at these two stations are caused by the higher variability in the polar areas. That is supported by the comparison results of the two other polar stations, Thule and Kiruna, for which tighter coincidence criteria were applied. They show a positive bias of ∼2.5% which is well within the mean relative differences of ±6% reported for the low-and midlatitude stations. At five of the nine stations, the mean relative difference is negative, thus suggesting a small negative bias in the ACE-FTS HNO 3 partial columns relative to the FTIR measurements, which is consistent with a mean relative difference of −1.3% (25.9% standard deviation) calculated from all coincident FTIR comparisons. Figure 15 shows 15 good correlation between ACE-FTS and the FTIR HNO 3 partial columns, with a correlation coefficient R=0.823. The line fitted to the data has slope of 0.88 and intercept of 0.11×10 16 molecules cm −2 . No significant latitudinal dependence of the bias could be identified.
In similar work, FTIR HNO 3 partial columns were compared with MIPAS ESA partial 20 columns by Vigouroux et al. (2007) and were updated by Wang et al. (2007a). Wang et al. (2007a) found a mean relative difference of ±2% with 1σ standard deviations between ±5.4% and 13.2% using coincidence criteria defined as 300 km and ±3 h.

ClONO 2
Coincident data for this comparison with ACE-FTS was available from four FTIR sta- 25 tions. All FTIR retrievals of ClONO 2 used spectroscopic data from the HITRAN 2004 database (Rothman et al., 2005) with supplements from Wagner and Birk (2003), using a two-microwindow approach similar to that described by Reisinger et al. (1995). 2464 8,2008 Validation of HNO 3 , ClONO 2 and N 2 O 5 from ACE-FTS

EGU
The time series of the ClONO 2 partial column comparisons are shown for all stations in Fig. 16, along with the relative differences as a percentage of the FTIR partial columns. The mean difference results are given in Table 4. All stations show a large 1σ standard deviation on the mean relative differences. For the polar stations, the inhomogeneous ClONO 2 distribution in Arctic stratospheric airmasses during periods of 5 high vortex variability will contribute to this variation. However, there appears to be no systematic dependence of the mean relative differences or their standard deviations on latitude. The mean relative difference between ACE-FTS and the midlatitude Jungfraujoch station is 4.7% ± 4.2% with a 1σ standard deviation of 16.3%, which is the lowest standard deviation obtained for the ClONO 2 comparisons. It should be noted that the 10 Wollongong dataset has a DOFS of only 0.3, thus indicating that the partial column contains less than 1 independent piece of information, and that there is a contribution from the noise.
The scatter plot of the complete dataset (Fig. 17) shows a fair correlation between ACE-FTS and the FTIR ClONO 2 partial columns, with a correlation coefficient 15 R=0.815. The line fitted to the data has slope 0.71, thus being significantly lower than 1, and intercept 0.36×10 15 molecules cm −2 . A possible reason for this are the different line parameters used for the ClONO 2 retrievals. Most of the information in the ACE-FTS partial columns is coming from the microwindow centered at 1292.6 cm −1 , whereas the FTIRs used microwindows around ∼780 cm −1 for their retrievals. This is 20 a topic for further investigation. Höpfner et al. (2007) compared MIPAS IMK-IAA ClONO 2 partial columns with ground-based FTIR data using tighter coincidence criteria of 800 km, ±8 h, and additionally, a maximum PV difference of 3·10 −6 Km 2 kg −1 s −1 at 475 K. Relative differences were found to be between −9.2% and 10%. Introduction

Conclusions
In this study, we have undertaken an assessment of the quality of HNO 3 , ClONO 2 , and N 2 O 5 data (ACE-FTS v2.2, including the N 2 O 5 update) prior to its public release. All three molecules belong to the group of 14 baseline species for the ACE mission. HNO 3 is retrieved using 12 microwindows between 867-1728.6 cm −1 , covering an altitude 5 range from 5 to 37 km. The ClONO 2 retrieval employs two microwindows, centered at 780.15 cm −1 and 1292.6 cm −1 , and covers altitudes between 12 and 35 km. N 2 O 5 is retrieved from two microwindows between 1210 and 1270 cm −1 at altitudes from 15 to 40 km. All VMR profiles have a vertical resolution of about 3-4 km.

HNO 3
10 ACE-FTS HNO 3 profiles from the first three years of the mission have been compared with coincident measurements made by the SMR, MLS, and MIPAS (ESA and IMK-IAA data products) satellite instruments, multiple aircraft flights of ASUR, and individual balloon flights of SPIRALE and FIRS-2. ACE-FTS HNO 3 partial columns have been compared with measurements by nine globally distributed ground-based FTIRs. In 15 Fig. 18, the mean absolute differences and the mean relative differences for all of the statistical and indvidual vertical profile comparisons are shown together, while Table 5 provides a summary of the results of these comparisons. The comparison of ACE-FTS HNO 3 with the four satellite data products shows an agreement between −1.9 ppbv and +0.8 ppbv (±25%). On average, ACE-FTS has a 20 negative bias with a maximum value of −0.7 ppbv relative to MIPAS (both the ESA and the IMK-IAA data products) and a slightly larger positive bias with a maximum value of +0.8 ppbv relative to MLS. Relative mean differences with respect to MIPAS and MLS are within ±10% between 19 and 26 km, as seen in Fig. 18. An altitude shift of 1.5 km was applied to the SMR profiles based on the results of previous assessments. 25 The magnitude of the altitude shift and the remaining relatively large negative bias of −1.9 ppbv around 25 km is consistent with results from other satellite comparisons 2466 Introduction EGU (Wang et al., 2007a,b;Santee et al., 2007). Statistical comparisons also involved a set of 16 coincident pairs of ACE-FTS and ASUR aircraft observations. Between 18 and 26 km ACE-FTS HNO 3 VMRs are up to 2.5 ppbv (32%) larger than ASUR. Between 26 and 36 km, the two instruments typically agree within ±0.2 ppbv (±3%) which is consistent with the satellite comparisons.

5
Comparisons were also made with individual profiles obtained from two balloon flights. ACE-FTS HNO 3 VMRs are 1.0-4.2 ppbv (typically 28%) larger than the SPI-RALE VMRs. Larger differences are observed in the comparison with FIRS-2, varying from −3 ppbv at 16 and 28 km to +3 ppbv at 20 km. The mean relative differences oscillate between −55% and +50% and are typically 20%. All three airborne (SPIRALE, FIRS-2, and ASUR) measurements were performed in the Arctic winter during vortex conditions and show a minimum in HNO 3 at about 20 km, which results in a high bias for ACE-FTS of ∼2 ppbv which is not seen in the satellite comparisons. The SPIRALE and FIRS-2 data were affected by the presence of a PSC and may have seen local denitrification. However, the same degree of denitrification was not observed by ACE- 15

FTS.
The last set of comparisons is with HNO 3 partial columns measured by the groundbased FTIRs. The mean relative differences are between −12.6% and +6.0%. The mean relative difference of all 122 FTIR coincidences is −1.3% with a standard deviation of ±25.9%, suggesting a slight negative bias in the ACE-FTS partial columns over 20 the altitude regions being compared (∼15-30 km). No significant latitudinal bias could be detected.
Overall the quality of the ACE-FTS v2.2 HNO 3 VMR profiles is good over the altitude range from 18 to 35 km. At lower altitudes, between 10 and 18 km, good agreement is seen between both MIPAS data products and ACE-FTS. As seen in Fig. 18 ACE-FTS ClONO 2 profiles have been compared with the ClONO 2 measurements from the MIPAS satellite instrument (IMK-IAA data product) and from the FIRS-2 balloon flight. Partial column comparisons were performed with measurements by four 5 ground-based FTIR instruments. Table 5 provides a summary of the profile comparisons. Good agreement between ACE-FTS and MIPAS IMK-IAA ClONO 2 is seen in the mean absolute differences, which are typically within ±0.01 ppbv and reach not more than −0.04 ppbv (±1%) for 16-27 km. ACE-FTS has a positive bias relative to CTMcorrected MIPAS IMK-IAA of about 0.09 ppbv (14%) between 27 and 34 km. We do 10 not have an explanation for the large disagreement between the ClONO 2 profiles from ACE-FTS and the FIRS-2 balloon flight, which reaches −2.8 ppbv (−170%) at 24 km. These differences are under investigation. The ground-based FTIR comparisons show varying degrees of agreement. Good agreement was found for the comparisons with the Jungfraujoch and Thule partial 15 columns. The mean relative differences (given with standard error) are 4.7% ± 4.2% with standard deviation ±16.3% and −0.1% ± 8.7% with standard deviation ±28.9%, respectively. For the two Arctic stations, several coincident measurements in periods with high vortex variability (winter and spring) are included in the comparisons and hence may contribute to a larger scatter in the relative differences.  tained between 27 and 31 km. Over this region, ACE-FTS has a low bias of maximum −0.6 ppbv (∼51%).
To conclude, we have used all available data to assess the quality of three NO y reservoirs measured by ACE-FTS. Only limited coincident measurements existed for ClONO 2 and N 2 O 5 , but a good set of statistical comparisons was obtained for HNO 3 . 10 If new correlative data become available in future, particularly for ClONO 2 and N 2 O 5 , further comparisons are recommended.