1. VERIFICATION OF POINTING AND ANTENNA PATTERN KNOWLEDGE OF SUPERCONDUCTING SUBMILLIMETER- WAVE Limb-EMISSION SOUNDER (SMILES) Makoto Suzuki 1, Satoshi Ochiai 2, Chihiro Mitsuda 3, Koji Imai 4, Takeshi Manabe 5, Kenichi Kikuchi 1, Toshiyuki Nishibori 1, Naohiro Manago 1, Yoshitaka Iwata 1, Takuki Sano 1 and Masato Shiotani 6 1 Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency, Japan 2 National Institute of Information and Communications Technology, Japan 3 Fujitsu FIP Corporation, Japan, 4 Tome R&D Inc., Japan 5 Department of Aerospace Engineering, Graduate School, Osaka Prefecture University, Japan 6 Research Institute for Sustainable Humanosphere, Kyoto University, Japan 11/07/28IGARSS2011 Suzuki1

2. JEM/SMILES Mission (JEM/SMILES: Superconducting Submillimeter-Wave Limb-Emission Sounder designed to be aboard the Japanese Experiment Module on ISS; Collaboration project of JAXA - Japan Aerospace Exploration Agency - and NICT - National Institute of Information and Communications Technology -) 1. Demonstration of superconductive mixer and 4-K mechanical cooler for the submillimeter limb-emission sounding in space 2. Observation on atmospheric minor constituents in the middle atmosphere [SIS Mixer] RF: 640 GHz, IF: 11-13 GHz; Junction: Nb/AlOx/Nb, ~7 kA/cm 2 ; Fabricated at Nobeyama RO [Mechanical Cooler] Two- stage Stirling and J-T; 20mW @4K, 200mW @20K, 1000mW @100K; Power Consumption: <300 W; Mass: 90 kg [Standard Products] – 1 scan ： O 3, HCl, ClO, CH 3 CN, O 3 isotopes, HOCl, HNO 3 – Multi-scan ： HO 2, BrO [Research Products] UTH, Cirrus Clouds, volcanic SO 2, H 2 O 2 Aug. 23, 20102Presentation at Hokkaido U.

3. JEM/SMILES Payload The SMILES was carried by the H-IIB with the H-II Transfer Vehicle (HTV) (Sep. 11); the HTV was attached to the ISS (Sep. 18); the SMILES was attached to the JEM (Sep. 25) (All dates in JST) SMILES Dimension: 1.85 m x 1 m x 0.8 m Weight: < 500 kg Mission Life: 1 year Aug. 23, 20103Presentation at Hokkaido U.

4. Two Bands among Band A, B, C can be observed. May 11, 2011ACE meeting, M. Suzuki et al4 Band ABand B Band C Frequency region has been selected by engineering interest, as high as possible, but 625-626 GHz region is the only frequency to measure HCl below 1 THz. At 600 GHz troposphere is opaque in limb. Tsys ~ 350 K, and Noise floor is ~0.4 K, given by

5. Scientific targets of SMILES 1. Inorganic Chlorine chemistry ClO to HCl ratio (O 3 trend in the US) HOCl production (O 3 trend in the LS) Global ClO (background ClO) 2. Bromine budget(very short- lived source gas issue) 3. HO x budget etc. May 11, 2011ACE meeting, M. Suzuki et al5 Error estimation for the mid-latitude case based on the single scan measurement Simulated SMILES observation performance

6. version 1.3 SMILES L2 data It is already comparable to other satellite data, Aura/MLS, SciSAt-1/ACE-FTS, ENVISAT/MIAPS, TIMED/SABER etc. 11/07/28IGARSS2011 Suzuki6

7. 7 ver. 1.3 O 3 O 55 deg < latitude < 65 deg Ref – SMILES(Ref – SMILES) / |SMILES| The SMILES L2 O 3 has a negative bias with respect to ACE-FTS and MLS at around 50 km. However O 3 measurements are well determined within 5% differences at 20-40 km.

8. 8 ver. 1.3 HCl The SMILES L2 HCl has a negative bias with respect to ACE-FTS and MLS at around 50 km. However it is almost similar to those of the comparison instruments below 45 km. 55 deg < latitude < 65 deg Ref – SMILES(Ref – SMILES) / |SMILES|

9. 9 MLS vs. SMILES: ClO (daytime) -35 deg < latitude < -25 deg Coincidence events: SMILES (59) for MLS v2.2 (92) SMILES (59) for MLS v3.3 (92) | t | < 24 hour, | r | < 200 km, | sza | < 2 deg Ref – SMILES(Ref – SMILES) / |SMILES| The SMILES L2 ClO is in good agreement with MLS v2.2 and v3.3. O Cl

10. SMILES is expected to provide break-through data, not the yet another satellite data. such as "Global ClO distribution" SMILES provides global ClO distribution with high precision. Furthermore, measurements of ClO, HCl, HOCl, and HO 2 can provide important insights into the Cly chemistry. ? Santee et al., 2008 ? MLS 25 ppt @22km this work 25 ppt at EQ 22km May 11, 201110ACE meeting, M. Suzuki et al

11. SMILES retrieval requires detailed knowledge of instrument function. y = F(x) +  – Y: measurements, x: physical parameters – F: Forward mode, both atmospheric and instrumental – e: noise SMILES has significantly lower noise compared to previous instruments, Aura/MLS etc, due to 4 K cooled detector system. It requires detailed and improved forward model for the retrieval, especially on the instrument function. In the IGARSS 2011, 3 papers on SMILES instrument function have been presented: – S. Ochiai et al, Gain Nonlinearity Calibration of the SMILES reciever. – H. Ozeki et al, RESPONSE CHARACTERISTICS OF RADIO SPECTROMETERS OF THE SUPERCONDUCTING SUBMILLIMETER- WAVE LIMB- EMISSION SOUNDER (JEM/SMILES) – This paper. 11/07/28IGARSS2011 Suzuki11

12. Left: AOS characteristics Right: Nonlinearity correction 11/07/28IGARSS2011 Suzuki12 Fig2: A spectral image of comb signal Fig4 : Temporal variation of resolution (AU1)

13. Pointing Knowledge, Accuracy Pointing knowledge is critical for limb observation. SMILES pointing knowledge: – ISS attitude data – SMILES attitude (star sensor) – SMILES antenna angle resolver 11/07/28IGARSS2011 Suzuki13

14. It was found that there might be jitter among attitude data. ISS attitude data should be corrected, using MAXI-ISS jitter data. 11/07/28IGARSS2011 Suzuki14 Fig. 1 Relative difference between MAXI GPS 1pps signal and the ISS clock telemetry data.

15. ISS attitude gives much smooth retrieved profile. After the jitter correction, systematic bias also disappeared. 11/07/28IGARSS2011 Suzuki15 Fig. 2 Retrieved O 3 profiles using the SMILES Star Sensor data (left) and ISS attitude data (right).

16. Tentative conclusion on pointing knowledge ISS attitude (GPS triangulation) looks to be stable and reliable for SMILES retrieval. SMILES star sensor (raw data) looks noisy, as expected from its specification. Temporal mechanical alignment is calculated from the average difference of ISS attitudes and SMILES star sensor data. There is timing jitter between SMILES (MAXI or JEM) and ISS attitude telemetry. After the timing jitter correction, SMILES O3 profile became very smooth. Actual random error of SMILES O3 data looks to be affected by uncertainty of pointing knowledge (random error of pointing mirror angle resolver, which is ~ 60 m in tangent height, or ~ 1% in L2 O3 value). – Detector noise (Tsys ~340 K) should give random error of L2 O3 << 0.5 %. 11/07/28IGARSS2011 Suzuki16

17. 11/07/28IGARSS2011 Suzuki17 Offset Cassegrain, 400 mm (vertical) x 200 mm (horizontal) elliptical shaped antenna Scan step 0.009375 deg (33.75’’), resolver +/- 0.0015 deg (5.4’’)

18. Fig. 3 Two dimensional (Elevation, Azimuth) SMILES antenna pattern. [7] T. Manabe, et al., “Measurements of the Offset-Cassegrain Antenna of JEM/SMILES Using a Near-Field Phase-Retrieval Method in the 640 GHz Band”, 21 st International Symposium Space Terahertz Technology, Oxford UK, March 23-25, 2010. 11/07/28IGARSS2011 Suzuki18

19. Fig. 4 One dimensional antenna pattern in elevation axis without antenna motion (dark), and with antenna motion (light). 11/07/28IGARSS2011 Suzuki19

20. Fig. 5 Averaged retrieved profiles (red: moving antenna, blue: fixed antenna), relative difference normalized to a priori of L2 ver. 1.3 O 3 (upper left), temperature (upper right), HCl (lower left) and BrO from SMILES Band A (lower right). 261 observations are averaged in Oct. 12, 2009 at equatorial region N10-S10. 11/07/28IGARSS2011 Suzuki20 Band A O 3

21. Fig. 5 Averaged retrieved profiles (red: moving antenna, blue: fixed antenna), relative difference normalized to a priori of L2 ver. 1.3 O 3 (upper left), temperature (upper right), HCl (lower left) and BrO from SMILES Band A (lower right). 261 observations are averaged in Oct. 12, 2009 at equatorial region N10-S10. 11/07/28IGARSS2011 Suzuki21 Band A Temp.

22. Fig. 5 Averaged retrieved profiles (red: moving antenna, blue: fixed antenna), relative difference normalized to a priori of L2 ver. 1.3 O 3 (upper left), temperature (upper right), HCl (lower left) and BrO from SMILES Band A (lower right). 261 observations are averaged in Oct. 12, 2009 at equatorial region N10-S10. 11/07/28IGARSS2011 Suzuki22 Band A HCl Figure in the manuscript was mistake.

23. Fig. 5 Averaged retrieved profiles (red: moving antenna, blue: fixed antenna), relative difference normalized to a priori of L2 ver. 1.3 O 3 (upper left), temperature (upper right), HCl (lower left) and BrO from SMILES Band A (lower right). 261 observations are averaged in Oct. 12, 2009 at equatorial region N10-S10. 11/07/28IGARSS2011 Suzuki23 Band A BrO

24. Conclusion on Antenna Pattern The difference of antenna pattern affects the weighting function, K, in Eq. (1), and it thus changes retrieved profiles. Fig. 5 shows an example of systematic difference between moving and fixed antenna pattern. The systematic differences with considering moving antenna are ~2% for O 3 and < 1% for temperature. Systematic differences for other species are from ~2% for HCl (strong spectral signal) to ~30% for CH 3 CN (very weak spectral signal overlapped by other species). It is concluded that the moving antenna pattern should be used for the L2 retrieval. Systematic errors due to the moving antenna should be negligibly small (multiply ~2% for O 3 by -55 dB for antenna pattern uncertainty). 11/07/28IGARSS2011 Suzuki24

25. Summary Pointing knowledge of the SMILES for the L2 retrieval system can be calculated properly from the ISS attitude with corrections (ISS telemetry timing jitter, and structural difference). The one-dimensional model of antenna pattern based on the physical- optics calculations and the phase-retrieval measurements od the antenna flight model was implemented in the SMILES operational L2 retrieval system. Systematic errors of L2 products are estimated to be ~1% for the pointing and negligibly small for the antenna pattern treatment. The results from 3 papers in IGARSS 2011 are implemented in the next L2 version (ver 2.0, under test run). 11/07/28IGARSS2011 Suzuki25