1. 6e. Update on NOAA-20 Ozone Mapping and Profiler Suite (OMPS)
L. Flynn (NOAA) and NOAA-20 NOAA and NASA OMPS EDR and SDR Teams
GSICS Annual Research Meeting, UV Subgroup, 21st March 2018

2. Disclaimer
"The contents of this presentation are mine personally and do not necessarily reflect any position of the US Government or the National Oceanic and Atmospheric Administration."

3. Outline Instrument Performance Ozone Product Performance
Solar
Wavelength
Geolocation
Summary for other characteristics
Ozone Product Performance
Ozone profile products
Total column ozone products
Web resources (monitoring, document and data)
Summary and Conclusions

4. Initial solar data are good and stable
N20 NP
N20 NM
Slide from Glen Jaross NASA
Comparisons to prelaunch synthetic proxy data.
Residual variations w.r.t. Reference Solar are comparable for SNPP and N20
NOAA-20 lacks radiometric adjustments, esp. in dichroic region, and corrections for solar activity

5. Shifts detected by using first solar calibration
Shifts detected by using first solar calibration. Wavelength shifts affect performance if not corrected in ground processing: NM shift = nm (+ intra-orbital variations); NP shift = nm
NM sun-normalized radiance comparison:
N20 measured and modeled
Measured – modeled difference before shift correction
Measured – modeled difference after 0.03 nm radiance shift of solar to match Earth radiance
Key retrieval wavelengths
Results from Colin Seftor SSAI
Spectral structure translates to ozone retrieval errors

6. Small along-track geolocation error within requirements
High contrast scenes in Bolivia and Persian Gulf both suggest ~ 0.2 deg. along-track pointing error.
Such errors are not important for key OMPS ozone products (0.8 deg. requirement).
An adjustment could be introduced via the instrument mounting matrix.
Salar de Uyuni, 15 Jan., 2018
(SNPP VIIRS image overlayed with pink OMPS image)

7. Other OMPS performance PLTs
Science Trending
Dark current degradation in family with S-NPP OMPS.
Linearity fluctuations comparable to S-NPP OMPS.
Noise Characterization
Measured noise matches modeled noise.
Sample Table Optimization
No buffer overflows occur when the sampling assumes 2.2x compression efficiency.
OMPS has a candidate sample table for future High Resolution measurements – 140 cross-track FOVs every 2.5S.
Baseline solar measurement and goniometry verification
Utilizes eight solar sequences, each at a non-nominal yaw angle
Solar data from orb 1009 indicates no error at Az = 26.9°. Not as good as verifying at 29° as originally planned.
Slide from Glen Jaross NASA

8. Summary of Findings for Ozone Profile EDR
NOAA-20 OMPS Nadir Profiler / V8Pro EDR
EOF analysis of measurements shows good SNRs. Outliers are larger by linear factors in FOV size not square root factors.
EDRs show good results for this stage of maturity.
Output error for Profile Error Code 8 cases (Excessively large initial residual – Flag is correctly set). Code fix will be delivered to NDE with Provisional Table updates.
Code error in cross-track macropixel computation; fix expected at IDPS in July 2018.
New OMPS Nadir Profile sample table needed to match OMPS Nadir Mapper FOV.
Radiation in the South Atlantic Anomaly has a large effect on radiances for 50x50km^2 FOVs, some just outside of geographic flagged region
Outliers are linear not square root in reduction with increasing pixel aggregation.
New field-angle-map, wavelength scales and solar tables provided by SDR Team.
Some cases of negative radiances found in the auroral oval .
Dichroic effects on wavelength/bandpass and on calibration from nm shift not yet accounted for in SDR calibration tables or EDR adjustments.
Possible overcorrection for stray light (not shown).

13. South Atlantic Anomaly

15. Summary of Findings for Total Ozone EDR
NOAA-20 OMPS Nadir Mapper / V8TOz EDR
EOF analysis of measurements shows good SNRs, similar intra-orbit wavelength shift and expected patterns.
EDRs show good results for this stage of maturity.
EDRs tested at STAR for low and medium FOV granules: 35x5  35x15  103x15.
Stray light – need to check measurements in the overlap region from 300 nm to 310 nm with NP.
16-image granule SDR Processing error due to variable processing times; Code fix tested at STAR.
Mismatch in current Medium to Low SDR FOV Cross-track aggregation. New tables have been prepared.
Team needs 8-day averages with corrected SDRs to compute cross-track bias statistics and validate versus S-NPP.

16. 16-Image Granule Errors
Cross-Track Bias

18. Raytheon-devised Cross-granule retrieval fix for 16-Image RDRs tested at STAR by SDR Team

23. Provisional Delivery to NDE planned for May
Planned code, script and table changes at NDE
Dynamic SDR sizing adjustment for NM SDRs in codes, not control scripts
Improved handling of end of orbit, end of day, and duplicate granules
Fix for Profile Error Code 8 output EDR content for V8Pro
Check FOV alignment accuracy for NM and NP within code.
New adjustment tables for V8TOz and V8Pro
Requires access to SDRs with better sample tables
Requires access to SDRs with better stray light corrections
Test data processing to confirm LFSO2 is ready for smaller FOV V8TOz EDRs from NOAA-20

24. Web Resources
Additional information is available in the OMPS V8TOz and V8Pro algorithm theoretical basis documents (ATBDs) and the SDR and EDR maturity review briefings, which can be accessed at:
and
Provisional NOAA-20 OMPS SDR near-real-time status and performance monitoring web page will become available by using the following websites:
Pre-operational NOAA-20 OMPS EDR near-real-time status and performance monitoring web pages will become available at the following websites:
Archive
Operations
Long-term
Daily maps
Activity

25. Summary & Conclusions
The NOAA-20 OMPS Ozone EDR and BUFR products are ready for users to examine with proper formatting and reasonable content.
Deficiencies in the OMPS SDRs are known, and the SDR Team has paths forward to improve the products.
Minor code and script changes will be provided to NDE for the V8Pro, V8TOz and LFSO2 codes.
Final adjustments for the EDR products will be developed using off- line STAR SDR processing.
Product quality will improve as SDR Team adjustments and corrections enter the IDPS processing system.

26. Backup Provisional Review Presentations…

27. NOAA-20 OMPS Instrument Performance
G. Jaross & OMPS SOC
20 February 2018

28. Agenda Performance issues requiring attention
Ground-to-orbit wavelength shifts
Pixel selection in flight tables
Performance anomalies NOT requiring attention
Sensor degradation shortward of 265 nm
FOD on detector
Along-track geolocation
Pixel saturation
Summary of 8 performance PLTs
All but one completed successfully
No indication of goniometry errors – no repeat of yaw maneuvers
Open Issues (further analysis needed)
readout structure during NP solar cal.

29. NM solar wavelength registration remains a challenge – as on SNPP
N20 NM Wavelengths during solar measurements
POS 1
POS 4
POS 7
POS 2
POS 6
POS 5
POS 3
Problem: NM solar data collected when wavelengths changing at greatest rate
3-orbit solar sequence used on NPP OMPS smears out wavelength registration for each of the 7 diffuser positions
1-orbit solar used on N20 yields more precise registration at each position
Recommendation: Base the N20 NM baseline wavelength registration on 1-orbit solar data
3-orbit sequence 1 2 3 4 5 6 7
1-orbit sequence

30. Pixel selection leads to poor radiometry
NM Total Ozone Measurements
Edge-of-swath error caused by radiance error in marginal pixel
Data gaps caused by ground processing error
Radiance error exists in edge pixels
Edge pixels are not needed for full global coverage
Solution: remove unnecessary edge pixels from telemetered data (long-term), remove edge pixels in ground processing (interim).

31. Revised Sample Table needs testing and approval
N20 OMPS NM Radiance Sensitivity
Current sample tables include pixels with little or no photon sensitivity
NM has been tested with a revised sample table that chops 10 pixels off either edge. More than enough for full equatorial coverage.

32. Interim NM Sample Table at 17x17 km2
Sinabung Eruption 19 Feb., 2018
Demonstrates current instrument operations
Small spatial mismatch with NP
SNPP OMPS 50 x 50 km2
N20 OMPS 17 x 17 km2
Courtesy of Colin Seftor

33. No evidence of continued NP degradation after launch
Albedo Calibration accuracy Waiver
472-CCR
NP Solar Flux compared to Reference
Measured solar irradiance changes include sensor, diffuser, and solar flux changes
Some evidence of the degradation seen before launch
NP Sun-normalized radiance compared to Model
Sun-normalized radiances include only diffuser reflectivity changes
No evidence the diffuser has degraded
Observed pre-launch change

34. No effect of detector FOD observed in science data
No need to take action until a radiometric error is found
Initial post-launch LED aliveness test
Xtrack Position
Time
Unbinned Antarctic signals spectral index 80
Xtrack Position
Position 5
Position 6
Orbit 696 solar irradiance
Wavelength
Xtrack Position

35. Small NP solar signal errors
A Single NP Solar Image
NP CCD Signal Levels
10 – 20 count error in smear- corrected NP solar signals
Caused by timing pattern issue – when is charge integrated in the CCD readout
Currently investigating other NP image types. DN

36. Small along-track geolocation error within requirements
Salar de Uyuni, 15 Jan., 2018
(SNPP VIIRS image overlayed with pink OMPS image)
High contrast scenes in Bolivia and Persian Gulf both suggest ~ 0.2 deg. error.
Such errors are not important for key OMPS ozone products (0.8 deg. requirement)
An adjustment could be introduced via the instrument mounting matrix

37. Small number of pixel saturations is acceptable
OMPS geolocation data was used for the analysis
Comprised of unbinned, no co-added data at 44 wavelengths
Wavelength with highest sensitivity (329.6 nm) was included
Search for instances where signal overflowed the 14 bit register (ie, > 214-1)
Results indicate that 4 NM wavelengths saturate, 0.1% of pixels at 329 nm
As expected, percentage mirrors mean signal levels
Small number of NP saturations is consistent with transient events

38. Data compression leads to granularization issue
Difference between CCSDS time and Image time
RDR granules formed from CCSDS time stamp in packet header
Relationship between CCSDS time and image time affected by on-board data compression
Depending on granule boundaries, a granule can receive 14, 15, or 16 images.
SDR is only sized for 15 images max.
Recommendation: Re-size SDR to handle more images per granule
Time Difference (ms)
Image Number in granule

39. Other OMPS performance PLTs
Science Trending
Dark current degradation in family with SNPP
Linearity fluctuations comparable to SNPP
Noise Characterization
Measured noise matches modeled noise
Sample Table Optimization
No buffer overflows occur when the sampling assumes 2.2x compression
OMPS has a candidate sample table for any future High Resolution measurements
Baseline solar measurement and goniometry verification
Utilizes 8 solar sequences, each at a non-nominal yaw angle
Sign error led to incorrect yaw maneuver angles
Instead of repeating, we will collect annual Reference at new deg.
Solar data from orb 1009 (see 28 Jan. data on Slide 8) indicates no error at Az = 26.9 deg. Not as good as verifying at 29 deg. as originally planned.

40. Summary Other than new flight tables, no changes needed to operations
Sample table update to NM and possibly NP needed to repair edge-of-swath problem
Use Medium rather than Low resolution timing pattern to reduce error in interim pending table load approval
New Ground LUTs needed to correct wavelength shifts
This is expected work; i.e. business as usual
Initial wavelength LUTs are available
Intra-orbital corrections use same code introduced for SNPP OMPS
Additional errors do not require correction but are recommended
0.2 deg. along-track pointing adjustment
Introduce more stringent saturation flags