1. 1 Partial and total column SFIT2 retrievals from Eureka DA8 and PARIS-IR FTIR spectra in spring 2004 – 2005, including comparisons with the ACE Satellite measurements Keeyoon Sung 1,2, Kaley Walker 1,3, James Drummond 1,4, DeJian Fu 3, Peter F. Bernath 3,5, Richard Mittermeier 6, Hans Fast 6, and Kimberly Strong 1 International Symposium on Molecular Spectroscopy 62 nd Meeting, June 18-22, 2007 1 University of Toronto, Canada ; 2 Jet Propulsion Laboratory, NASA, USA 3 University of Waterloo, Canada; 4 Dalhausie University, Canada 5 University of York, UK; 6 Environment Canada, Canada

2. 2 Eureka (80ºN, 86ºW), Nunavut Eureka Weather Station the Arctic PEARL Eureka Weather Station The Polar Environmental Arctic Research Laboratory (PEARL) One of primary Arctic stations of the Network for Detection of Stratospheric Change (NDSC, now NDACC)

3. 3 Column density measurements over Eureka Infrared active trace gases  H 2 O ( ~ 1 %), CO 2 (370 ppm), all others ( ~ < 1 ppm) Long-term study of composition change  Total and partial columns of the trace gases  FT-IR is productive.  15 years of long term data archive in Eureka are available  Atmospheric chemistry, dynamics, and long-term climate change The results from 2004 -5 measurements  What is specialty - ACE-FTS observation  Contrast in stratospheric temperature  CH 4, N 2 O, O 3, HCl, HF, HNO 3, ClONO 2, NO, NO 2 Chemical O 3 destruction in the Arctic spring  HF as a dynamic tracer  Focus on O 3, HCl, HF, HNO 3, ClONO 2 Surface Temperature: 1861 – 2000 The UN’s Intergovernmental Panel on Climate Change (IPCC) IPCC 2005 G. L. Manney, et al. 2006

4. 4 ACE-FTS Arctic Validation campaigns Three FTS’s  Bomem DA8 (0.004 cm -1 ) MCT, InSb  PARIS-IR (0.02 cm -1 ) a groundbased version of ACE-FTS Sandwiched detector 0.02 cm -1 and 0.05 cm -1  ACE-FTS (0.02 cm -1 ) within 200 km bet. 200 – 500 km switching mirror  The 2004 and 2005 data analyses VMR updates with HALOE, SPARC, Ozonesonde archive Retrieval with SFIT2 (v3.91) HITRAN 2004 Solar lines from F. Hase compilation (Hase, F. et al. 2006)

5. 5 A composite image of DA8 spectra with various optical filters CO 2 H2OH2O H2OH2O PARIS-IR DA8

6. 6 Geographic distribution of atmospheric samplings (relative to the observation site, Eureka on Mar. 3, 2004) PARIS spectra on a single day (DoY=63, 2004) DA8 spectra Resolute Bay All ACE-FTS occultations within 200 km and 200 – 500 km Distance O3-sonde balloon flight track (< 40 km )

7. 7 List of MWs used in this work DA8: SFIT2PARIS 2004PARIS 2005 O3O3 1120(1104,1119,1120,1121), 2775, 0782, 1147, 3040 2775(2775,2779, 2781), 3040 2775(2775,2779,2781), 3040, 0782, 0981 HCl2725(2727, 2775), 292529252843(2925,2843,2963) HF4038, 410940383999(3999,4038) HNO 3 0870(0868, 0872)08690869, 0873 NO 2 2914 (narrow MW)2915- NO1900, 1903-1899(1899,1901) ClONO 2 0780- CH 4 2655(2651, 2667), 1202, 2855(2835, 2860, 2903) 2835(2835,2860, 2898), 2903 2835(2836,2858,2903) N2ON2O1205(1202, 1204), 2523, 2806 2523, 2806, 1183 Note (1) the four digit numbers roughly represent location of MWs in wavenumber. (2) the same MW name does not necessarily represent its counterpart MW for PARIS.

8. 8 Good agreement in O 3 total columns among DA8, PARIS-IR, and ACE-FTS (z>6.5km) 20040303 2004 and 2005 DA8 (open symbols) PARIS( ■, ■) ACE-FTS ( ✶, ✶, ✶,✶ ) Ozonesonde( ▲, ▲) Note that DA8 points fall on the track of PARIS-IR measurement points. MW0782

9. 9 HCl: Comparisons with ACE-FTS/PARIS-IR during the intensive phase In 2005, DA8 falls right on the trend observed by PARIS. Note the contrast in sampled airmasses in 2004 and 2005. 2004 and 2005 DA8 (open symbols) PARIS( ■, ■) ACE-FTS ( ✶, ✶, ✶, ✶ ) O3-sonde( ▲, ▲) MW2725

10. 10 HNO 3 : Comparisons with ACE-FTS/PARIS-IR 2004 and 2005 DA8 (open symbols) PARIS( ■, ■) ACE-FTS ( ✶, ✶, ✶, ✶ ) O3-sonde( ▲, ▲) Day 63 Day 58 Day 64 Day 68 MW0870

11. 11 Partial columns: DA8 and ACE-FTS N2O (6.5 – 58.5 km)CH4 (6.5 – 66.5 km) 2004 and 2005 DA8 (□, ⊲, □, ⊲ ), PARIS (■, ■), ACE-FTS, d < 200km ( ✶, ✶ ), < 500 km ( ✶, ✶ )

12. 12 The linear error analysis technique By the source  Smoothing error o I. Retrieval = an estimate of a state smoothed by A o II. Retrieval = an estimate of the true state + smoothing error  Retrieval noise  easy  model parameter error o Retrieved model parameter (interference, baseline, calibration,…) o Non-retrieved model parameters o T o spectroscopic line par. o ILS (EAP) o SZA  Non-parameterized forward model error  not easy By the property  Random and systematic error Representing covariances measurement error Retrieval error Total error

13. 13 O 3, Errors (in %) for total and partial columns by MW SpeciesMWE˝ cm -1 ranges (in km)σ(sm)σ(rmp)σ(rn) forward model par. σ(T)σ(θ)σ(θ)σ(S)σ(γ)σ(γ) O31120 1104 1119 1120 1121 ~282 ~205 ~203 ~707 ~325 ~21 0.6-1004.70.22.00.91.5 0.6 – 80.90.20.50.31.7 8 – 173.00.11.31.01.7 17 – 246.2< 0.12.65.61.3 24 – 3414.1< 0.15.613.71.2 34 - 5661.2< 0.121.614.50.9 0782~3770.6-1005.20.42.13.61.42 1147~3720.6-1004.80.42.02.31.4- 2775~500.6-1007.60.32.70.21.35 3040~300.6-1005.80.22.30.41.55 Note that (1) σ(fmp) is estimated from synthetic spectra using observation conditions on Mar. 15, 2005. (2) σ(S)’s are assumed based on the values reported in the HITRAN2004. (3) < 5% in broadening produces insignificant errors in columns. (Ref. B. Barret) uncertainties not available 2% in broadening, insignificant in col.

14. 14 Reminder: Cl-budget and HF HF  no known sink in the stratosphere  increase with age of sampled airmass  a dynamic tracer (better than N 2 O) in the Arctic stratosphere Cl budget (HCl, ClONO 2, ClO, ….)  (HCl + ClONO 2 ) explains >95 % of it. (DM Wilmouth, et al., JGR, 2006)  HCl, rapid adsorption and slow recovery  ClONO 2, rapid adsorption and rapid recovery HCl, ClONO 2, (HCl+ClONO 2 )/HF  To minimize the contamination from transport effects  typical values of HCl, ClONO 2, (HCl+ClONO 2 )/HF = 2, 1, 3  indicator of location of the sampled airmasses w.r.t. vortex

15. 15 Characterizing sampled airmass: HCl/HF and ClONO 2 /HF HCl, slow recovery ClONO 2, rapidly recovered?! transition moment of escaping vortex 2004 and 2005 DA8 (open symbols), PARIS (■,■), ACE-FTS : distance < 200km ( ✶, ✶ ), < 500 km ( ✶, ✶ ) HCl ClONO2

16. 16 O 3 /HF: Chemical O 3 destruction We observed substantial chemical O 3 destruction during the intensive phase in 2005 (in a few days !), while no significant changes was observed in 2004. 2004 and 2005 DA8 (open symbols) PARIS (■, ■) ACE-FTS (hexsagons) d < 200km (hex, hex) d < 500 km (hex, hex)

17. 17 Summary Total and partial columns presented with error budgets PARIS-IR captures temporal/regional column variations Good agreements among the three FTS’s  DA8 falls on the track of PARIS-IR  ACE-FTS beyond 200 km, better with ground-based FTS’s. Distinctive different trends were observed in the O 3 columns in 2004 and 2005 Observed chemical O 3 destruction in 2005, but virtually no destruction in 2004

18. 18 Acknowledgements Canadian Foundation for Climate and Atmospheric Sciences CANDAC Canadian Space Agency Natural Sciences and Engineering Research Council (NSERC), Canada Eureka Weather Station CalTech/JPL (Dr. Linda R. Brown) for travel support PEARL Inukshuk

19. 19 Eureka VMR 2006 Constructed all seasons’ VMR set (except H 2 O) Started with the VMR set of Kiruna (Ref.2) Direct updates H 2 O ( 10 – 65 km), HALOE(v.19), H 2 O ( troposphere), daily sonde (Eureka) CH 4 (15 – 60 km), HALOE(v.19) NO (25 – 60 km), HALOE(v.19) NO 2 (15 – 50 km), HALOE(v.19) HF (15 – 60 km), HALOE(v.19) HCl (15 – 60 km), HALOE(v.19) N 2 O (20 – 50 km), SPARC2000(CLAES) HNO 3 (15 – 50 km), SPARC2000(CLAES) ClO (15 – 55 km), SPARC2000(MLS) O 3 (below 35 km), (Mar.-Sep.) Eureka O3sonde archive (1992-2006) O 3 (above 35 km), HALOE(v.19) Derived updates  Isotopologues of H 2 O, O 3, CH 4 Fractionation profiles, derived from the Kiruna set Applied it to the Eureka VMR 2006 Ref. 2: A. Meier et al. Spectroscopic Atlas of Atmospheric MS in mid-IR, 2004

20. 20 Chemistry in the Artic Polar Votex ClONO2 HCl