(3) RESULTS OF MID-LATITUDE MIPAS VALIDATION MEASUREMENTS OBTAINED BY THE SAFIRE-A AIRBORNE SPECTROMETER U. Cortesi (1), G. Bianchini (1), L. Palchetti (1), E. Castelli (2), B.M. Dinelli (2), G. Redaelli (3) (1) Istituto di Fisica Applicata "Nello Carrara" (IFAC-CNR), Firenze, Italy, u cortesi@ifac.cnr.it (2) Istituto per le Scienze della Atmosfera e del Clima (ISAC-CNR), Bologna, Italy, e.castelli@isac.cnr.it Università degli Studi di L Aquila, Dipartimento di Fisica, L Aquila, Italy, gianluca.redaelli@aquila.infn.it ABSTRACT/RESUME Far infrared emission measurements acquired by the SAFIRE-A limb sounder aboard the M-55 Geophysica high altitude aircraft, during dedicated ENVISAT validation campaigns, primarily aimed at validating MIPAS operational products. O 3 and HNO 3 observations obtained by the airborne instrument during the midlatitude flight on 24 th October, 2002 and intercomparison with MIPAS off-line operational data are discussed in this paper. We also present results of further analysis carried out with the support of modeling tools developed at University of L Aquila and making use of forward and backward trajectories calculations, to identify matching aircraft and satellite data recorded at different times and geographical locations. 1. INTRODUCTION The Fourier transform far-infrared (FT-FIR) spectrometer SAFIRE-A (Spectroscopy of the Atmosphere by using Far InfraRed Emission Airborne) has been involved in field campaigns carried out with the M-55 Geophysica stratospheric aircraft in 2002-2003, aimed at validating the level-2 products of the ENVISAT instruments MIPAS, GOMOS and SCIAMACHY. Three campaigns have been conducted, as part of the ESABC (ENVISAT Stratospheric Aircraft and Balloon Campaigns) activities [1], with the Geophysica platform operating from Forlì, Italy (Lat. 44 N, Lon. 12 E) in July and October 2002 and from Kiruna, Sweden (Lat. 68 N, Lon. 20 E) in February- March 2003 [2]. A total of 11 flights and 45 flight hours was devoted, in these periods, to the validation of the ENVISAT chemistry payload, trying to match, as a first priority, MIPAS observations in the lower stratosphere. SAFIRE-A executed limb sounding observations of upper tropospheric and lower stratospheric emission, that made possible to retrieve volume mixing ratio (VMR) vertical profiles of O 3, HNO 3, N 2O, H 2O and other minor atmospheric constituents in the altitude range between approximately 10 and 20 km. The good quality of the profiles obtained by the SAFIRE-A instrument and their excellent spatial and temporal matching with MIPAS-ENVISAT overpasses already offered a valuable validation data-set of nearcoincident measurements. Moreover, the possibility to further exploit the data acquired during the ENVISAT validation campaigns beyond the simultaneous aircraft and satellite records, by using modeling tools specifically developed by University of L Aquila and based on trajectory calculations, was demonstrated. Proc. of the 2004 Envisat & ERS Symposium, Salzburg, Austria 6-10 September 2004 (ESA SP-572, April 2005) As a result, intercomparison of VMR values of various species retrieved from non coincident MIPAS and SAFIRE-A measurements could also be made, thus leading to a substantial increase in the overall number of useful data points. In this paper, we report the results of MIPAS O 3 and HNO 3 validation based on the data acquired by SAFIRE-A during the mid-latitude flight on the 24 th October, 2002 and on MIPAS L2 operational products generated by the Instrument Processing Facility (IPF version 4.61). In section 2, a brief overview of the airborne spectrometer and of its measurement capabilities is given, whilst in section 3 full details of the M-55 flight on 24 th October, 2002 and the quality of the matching between SAFIRE-A and MIPAS observations are discussed. Results from intercomparison of aircraft and satellite near-coincident observations of O 3 and HNO 3 are presented in section 4. The approach adopted to identify pairs of matching aircraft and satellite observations by means of lagrangian coincidences is detailed in section 5, where promising results are displayed for comparison of both O 3 and HNO 3 distributions. 2. SAFIRE-A INSTRUMENTAL FEATURES AND MEASUREMENT PRODUCTS 2.1 Instrument configuration and observing strategy The SAFIRE-A high resolution FT-FIR spectrometer is a passive remote-sensor operating aboard the M-55 Geophysica aircraft and capable to perform limb sounding observations of the atmospheric emission in the far-infrared region, in narrow spectral bands (bandwidth 2 cm -1 ) between 20 and 200 cm -1, with a spectral resolution of 0,004 cm -1 unapodized. A full description of the spectrometer and details of its upgraded configuration are provided, respectively, in [3] and in [4]. In recent years (1997-2003), the instrument has been deployed aboard the M-55 stratospheric platform in more than 30 engineering and scientific flights (approximately 150 hours of total flight time) at mid-latitudes and in the polar regions and progressively achieved a high degree of reliability. During the 2002-2003 scientific mission with the M-55 Geophysica aircraft dedicated to ENVISAT validation, SAFIRE-A was equipped with its long-wavelength detection channel centred on the [22.0-23.5 cm -1 ] spectral interval, to detect O 3, HNO 3, N 2O and ClO, and with the short-wavelength channel covering either the window [117.0-119.0 cm -1 ] for detection of H 2O and OH
(at mid-latitudes) or [124.0-126.0 cm -1 ] for H 2O and HCl (at high latitudes). The observation strategy is based on a series of limb sounding sequences acquired while flying at maximum altitude and looking along a line of sight perpendicular to the flight direction. Individual sequences, combining limb and upward viewing, are recorded in about 5 minutes, resulting, at an average aircraft speed of 700 km, in a horizontal resolution along the flight direction of approximately 50 km for each VMR profile; the resolution in the direction of the line of sight is typically of the order of 300 km. The vertical resolution of the retrieved mixing ratio profiles can be estimated from the Full Width Half Maximum (FWHM) of the averaging kernels (AK) of each retrieval. HITRAN2k; HNO 3 spectroscopic data have been taken from JPL database [6]. Pressure and temperature profiles were obtained by ECMWF (European Centre for Medium-range Weather Forecast) data processed at University of L Aquila. Temperature and geopotential height values at different pressure levels (from 1 to 1000 mbar) on a latitude-longitude grid (latitude step 1.125, longitude step 1.125 ) are provided every 6 hours (at 00, 06, 12 and 18). These values were linearly interpolated in latitude and in time, in order to make use of the most suitable temperature and pressure values for each sequence. VMR profiles coming from a standard mid-latitude atmospheric model were used either as initial guess of the retrieval and to model interfering gases. The vertical distribution of O 3 and HNO 3 concentration resulting from the retrieval process and reported in this paper for intercomparison with MIPAS v4.61 data are expressed as VMR values versus pressure on a vertical grid given by the tangent pressure levels of SAFIRE-A limb scans. Error bars associated to the SAFIRE-A profiles represent only the contribution of random errors. It should be noted that the upward looking measurements of each SAFIRE-A limb sequence are very poorly resolved in the vertical direction and provide, in fact, an estimate of the column content of the examined molecule above the flight level. In the RAS code, the VMR value corresponding to the highest retrieval height is used to scale the initial guess profile in the altitude range above it. 3 THE M-55 GEOPHYSICA MID-LATITUDE ENVISAT VALIDATION FLIGHT ON 24.10.2002 Fig. 1. SAFIRE-A averaging kernels for O 3 (a) and HNO 3 (b) for sequence 15 on 24 th October 2004. 2.2 Data processing In order to retrieve the vertical VMR distribution for the selected species from the limb sounding sequences recorded during the ENVISAT validation flights, an inversion algorithm developed at ISAC-CNR for the analysis of the airborne measurements (RAS, Retrieval Algorithm for SAFIRE-A) was used. The radiative transfer calculation implemented in RAS is based on a line-by-line and layer-by-layer model including curvature and refraction effects and the retrieval process relies on the global-fit technique described by Carlotti in [5]. The reference spectroscopic database adopted for the line-by-line calculations is On 24 th October, 2002 the M-55 Geophysica carried out a night-time flight from Forlì, Italy (Lat. 42 N, Lon. 12 E), in coincidence with an overpass of the ENVISAT satellite (orbit 3403) along a route that had been studied to optimize the overlapping between the air masses observed by the airborne limb-sounders and in situ sensors and those covered by MIPAS scans 14, 15 and 16. The aircraft flight track and altitude profile are shown, respectively, in Fig. 2 (a) and (b). During the flight the SAFIRE-A spectrometer acquired 20 limb scanning sequences, obtaining several profiles of the target species at approximately the same time and location of MIPAS measurements. As evident from the figure, the best overlapping was obtained with the MIPAS scan at 21:23 UT (scan 15), whose tangent points in the altitude range 10-20 km correspond to latitude and longitude region covered by SAFIRE-A observations during both the North-South and the South-North leg of the flight. Our analysis focused therefore on the intercomparison with MIPAS level-2 products from scan 15 (Lat. 42 N, Lon. 12 E) and particularly on O 3 and HNO 3, for which most of SAFIRE-A scans provided useful results.
Scan 16 (21:22) Scan 15 (21:23) Scan 14 (21:24) (a) (b) Fig. 2. M-55 Geophysica flight track (a) and altitude profile (b) on 24 th October, 2002
An estimate of the spatial and temporal overlapping of aircraft and satellite profiles was made by calculating the distance s between the average location of MIPAS tangent points in the range 10-25 km and the one of each SAFIRE-A scan and the time difference t between corresponding acquisition times. Based on this estimate, SAFIRE-A scans 15 and 16 meet the strictest criteria of s < 200 km and t < 10 min. 3. RESULTS In Fig. 3 (a) and (c), a comparison between the O 3 and HNO 3 profile retrieved by MIPAS and O 3 and HNO 3 VMR data obtained by SAFIRE-A for the best coincidences of scan 15 and 16 of the airborne instrument are shown (error bars for both instruments represent only the random error), highlighting a substantially good agreement, with the largest differences corresponding, in the case of O 3, as well as for HNO 3, to MIPAS lowest tangent pressure. In Fig. 3 (b) and (d), a similar intercomparison is made, considering a larger number of SAFIRE-A profiles, as derived from relaxed time-matching requirements. These plots provide an indication of the variability of the VMR vertical distribution measured by SAFIRE-A over a wider region that can still be considered, however, in close proximity with the location of MIPAS measurements. Fig. 1a and Fig. 1b show, respectively, the results of the calculations of the AK of O 3 and HNO 3 (scan number 15). The vertical sampling step of the measurements (that is the distance between contiguous tangent altitudes) for the examined sequence is less than 1 km for the higher altitudes and 1.5 km for the lower ones. The FWHM of the O 3 retrieved profile is about 1.5 km for tangent altitudes near the flight level and about 2 km for the lower altitudes. This means that the O 3 retrieval is not affected by over-sampling effects. In the case of HNO 3 the averaging kernels are less well shaped; the FWHM is about 1.8 km for altitudes near the flight level and about 4 km for lower altitudes. In this case the retrieval is affected by over-sampling at lower altitudes. 4. MODELING SUPPORT Modeling tools can be used to support MIPAS validation. In particular, the number of MIPAS and SAFIRE-A data points useful to perform intercomparison can be extended beyond those that are simply co-located in space and time, by using a lagrangian approach. Backward and forward isentropic trajectories, starting from all the available SAFIRE-A tangent points, are calculated and used for selecting those air masses sampled by both satellite and the airborne instrument, even if at different times and locations. Trajectory calculations are based on United Kingdom Met Office (UKMO) meteorological fields, and performed using the University of L Aquila Global Trajectory Model (GTM) [7]. The GTM was also routinely operated during the airborne validation campaigns to fine-tune the flight pattern, using forecasts of the direction and intensity of the winds from the NCEP (National Center for Environmental Prediction) Aviation Model and therefore a number of lagrangian correspondences between SAFIRE-A and MIPAS tangent points are expected to be found. For the show comparison, 5 days backward and forward trajectories are launched from the location of SAFIRE-A measurements - i.e. from each of the tangent points of the 20 limb sequences acquired long the flight route - for 24 th October, 2002. Air parcels sampled at least once also from MIPAS within a prescribed match criterion ( time 1h, latitude 1, longitude 1, altitude 1km) are then selected and their O 3 and HNO 3 contents measured from satellite compared to the correspondent SAFIRE- A values at the trajectory starting points. In Fig. 4 (a) and (b), couples of O 3 and HNO 3 VMR values by MIPAS and SAFIRE-A associated to the same air parcel, as defined by the matching criteria and derived by trajectory calculations, are plotted as a function of the retrieval altitude of the SAFIRE-A measurement and superimposed to couples of O 3 and HNO 3 data resulting from direct coincidences that satisfy similar criteria for geographical and vertical overlapping. By combining the two data-sets, we obtained a total number of useful matches more than a factor of 2 larger than the original one. Additional validation can be also performed by comparing SAFIRE-A data with global fields resulting from assimilation of MIPAS files into a 3D Chemical Transport Model (CTM). For this purpose, a sequential assimilation approach is used to assimilate available Ozone MIPAS profiles for October 2002 into the STRATAQ CTM [8] for the stratosphere and resulting fields time-space interpolated onto SAFIRE-A tangent points [9]. 5. CONCLUSIONS We have reported the results of MIPAS-ENVISAT validation by mid-latitude measurements with the airborne FT-FIR spectrometer SAFIRE-A. In particular, we have focused on the validation of lower stratospheric O 3 and HNO 3, that was achieved by directly intercomparing co-located VMR profiles of the target species almost simultaneously acquired by the aircraft and by the satellite limb sounders and by using lagrangian coincidences to further extend the number of matching SAFIRE-A and MIPAS data points. Straight intercomparison of co-located O 3 and HNO 3 profiles has shown a good agreement between VMR values of both species retrieved by MIPAS and SAFIRE-A.
(a) (b) (c) (d) Fig. 3. Comparison between SAFIRE-A and MIPAS O 3 and HNO 3 VMR profiles: (a) O 3 profiles comparison. Mismatch conditions: distance < 200 km, delay < 10 min (b) O 3 profiles comparison. Mismatch conditions: distance < 200 km, delay < 2 h 30 min (c) HNO 3 profiles comparison. Mismatch conditions: distance < 200 km, delay < 10 min (d) HNO 3 profiles comparison. Mismatch conditions: distance < 200 km, delay < 2 h 30 min
(a) (b) Fig. 4. Plot of the two MIPAS validation datasets derived by SAFIRE-A measurements 24 th October, 2002, for (a) O 3 and (b) HNO 3. VMR values from direct coincidences are marked with circles. Triangles represent additional data obtained by trajectory matching.
We have found a similar result, by comparing O 3 and HNO 3 mixing ratios obtained by the two instruments, when looking at the same air masses, as determined from backward and forward isentropic trajectories initialized at each of the SAFIRE-A tangent points. This not only reinforced the validation results obtained for the two species from previous intercomparison of near-coincident measurements, but also served to confirm that the technique based on the lagrangian approach adopted here to support MIPAS O 3 midlatitude validation can be applied to all other species and flights available in the SAFIRE-A ENVISAT validation data-set. ACKNOWLEDGMENTS The SAFIRE-A instrument operation, data analysis and modeling activity described in this paper has been supported by the Italian Space Agency in the frame of the APE-ENVISAT project (Airborne Platform for Earth observation: Observations from stratospheric aircraft: study of stratospheric chemistry and contribution to ENVISAT validation). REFERENCES 1. Wursteisen P., The validation of the ENVISAT chemistry instruments by use of stratospheric balloon and aircraft, Proceedings of ENVISAT Validation Workshop, Frascati, 9 13 December 2002, ESA SP- 531, August 2003. 2. Blom C.E., Cortesi U. and Redaelli G., ENVISAT Validation: introduction to the correlative measurements by the chemistry payload on-board the M-55 Geophysica, Proceedings of 16 th ESA Symposium on European Rocket and Balloon Programmes and Related Research, St. Gallen, Switzerland, June 2003. 3. Carli B., et al., SAFIRE-A: Spectroscopy of the Atmosphere using Far-InfraRed emission /Airborne, Journal of Atmospheric and Oceanic Technology, Vol. 16, p.1313, October 1999. 4. Bianchini G., Cortesi U., Palchetti L. and Pascale E., SAFIRE-A: optimised instrument configuration and new assessment of spectroscopic performances, Applied Optics, Vol. 43, N. 14, pp. 2962-2977, May 2004. 5. Carlotti M., Global fit approach to the analysis of limb-scanning atmospheric measurements, Applied Optics, Vol. 27, pp. 3250 3254, 1988. 6. Flaud J.-M., Perrin A., Orphal J., Kou Q., Flaud P.M., Dutkiewicz Z. and Piccolo C, New analysis of the 5 + 5 9 hot band of HNO 3, Journal of Quantitative Spectroscopy and Radiative Transfer, Vol. 77, pp. 355-364, 2003. 7. Redaelli G., Lagrangian techniques for the analysis of stratospheric measurements, PhD thesis, Univ. of L Aquila, Italy, 1997. 8. Grassi B., Redaelli G. and Visconti G., Assimilation of stratospheric ozone in the chemical transport model STRATAQ, to appear on Annales Geophysicae, 2004. 9. Grassi B., Redaelli G., Cortesi U., Bianchini G. and Castelli E., Assimilation of ozone profiles from MIPAS in the STRATAQ CTM, this issue.