1. Tim Hewison1 (1) EUMETSAT
Day 3 GRWG Session
Tim Hewison1
(1) EUMETSAT

2. Outline of Day 3 GRWG Session
8:30 Bob Instrument Event Logs Requirements
9:10 Tim intro References and Traceability
9:50 Tim Common Reference Channels
10:30 All Handling multiple references
10:30 Coffee break
10:50 All Strategy for migrating to MetopB/IASI & NPOES/CrIS
11:35 All Strategy for CLARREO/TRUTHS
12:20 All Strategy for LEO-LEO inter-calibration
13:05 Lunch
14:35 All Strategy for Re-Analysis products
15:05 Tim NWP Bias Monitoring Double-Differencing as inter-calibration technique
15:35 ? Microwave Sub-group? O-Bs?
16:05 Travel to NMSC
17:35 Introduction about NMSC
18:35 Dinner near NMSC
20:35 Travel back to Daejeon

3. References and Traceability
Traceability Statement for AIRS/IASI
Outstanding Issues with Traceability Concept
How we can strive to resolve these
Strategy for CLARREO/TRUTHS

4. Traceability Statement for AIRS/IASI
A key aspect to GSICS is to provide evidence to support applicability of the inter-calibration products we develop. A document has been prepared collaboratively by GSICS partners at EUMETSAT, NOAA, NIST and CNES to provide a statement demonstrating the suitability of the hyperspectral spectrometers, AIRS and IASI as inter-calibration references.
This discusses the various pre-flight and in-orbit tests that have been performed on these instruments. Particular attention is paid to their relative stability, as this is critical if they are to be used inter-changeably. Various method are described which show that AIRS and IASI offer consistent radiometric calibration – each with uncertainties ~0.1K (k=1).
The statement also discusses the traceability chain of AIRS and IASI to the SI international reference standards, which is needed to achieve the long-term goal of GSICS. From Article for GSICS Quarterly
Statement:

5. Outstanding Issues with Traceability
Traceability Statement for Aqua/MODIS
Application of Traceability Concept to
Model Results
Aircraft instruments
How to resolve these issues? -Discuss!
Identify experts in traceability issues
Dedicated web meeting?
Raise through CEOS/WGCV?
May need to revise current Traceability Statements…

6. Strategy for CLARREO/TRUTHS
“The (US) President's 2012 budget has essentially eliminated funding for the CLARREO mission.” “But all is not lost. NASA HQ argued back that the CLARREO … science were critical to climate science and could not be lost, so a compromise was to keep CLARREO … in a pre-phase A study mode instead of proceeding as expected to phase A and mission formulation. This means that CLARREO no longer has a launch date or an expectation to launch the configuration that passed Mission Concept Review. There will be funding to continue science (e.g. the CLARREO Science Definition Team will continue being funded), possibly breadboard instrument work, and the ability to consider other options to get portions of the CLARREO science accomplished. Examples include the NASA Venture class small missions (e.g. one small instrument on one launch) or the Venture instrument build opportunities, and/or international collaboration activities.” from David Young

7. So what can we do about CLARREO?
Robustly support proposals for similar missions
Including TRUTHS
Continue to support the science of using inter-calibration to provide traceability to other instruments ?
Discuss!

8. Common Reference Channels
The need for Common Reference Channels
Potential users for GSICS CRC products
Defining Common Reference Channels’ SRFs
Development Strategy for GSICS CRC products

9. Work Plan to Develop CRC Products
GRWG:
Define CRC SRFs – This meeting
Define uncertainties
Develop prototype products
GDWG:
Define file format, name conventions
Create directories on GSICS Data And Products Servers
GCC:
Seek feedback from beta-testers
Educate Users
GPPA

10. Practical Example: Defining Common Channels for GEO IR Imagers
Tim Hewison
EUMETSAT
GRWG Web Meeting

11. Original Calibration Table
Composition of GEO VIS Images Collaborate research with CEReS, Chiba University
New Calibration Table
Defining Common Reference Channels essential for traceability for multiple instruments
Also useful for generation of composite datasets
GMS-5
GOES-10
GOES-8
Original Calibration Table
千葉大高村研との共同研究についてです。
校正技術のapplication として位置付けることができます。
図は５つの衛星の可視画像を連結したものです。
衛星ごとに応答関数や観測時刻が異なるので「継ぎ目」が見えるのは当然なのですが、
校正後のテーブルでは継ぎ目があまり目立たなくなっています。センサ感度の劣化も
補正されています。
GMS-5
GOES-10
GOES-8
From 9th GSICS Exec Panel – Nov 2010
Courtesy of Tomoo Ohno (JMA)
12UTC on 5 Sep 2002
discontinuity 11

12. Ways to Define Common Reference Channels
Choose one GEO instrument as a reference standard
Politically difficult?
Some GEOs will not overlap – no collocations – but that doesn’t matter
Still need transformations for other satellites in same series
Choose one LEO instrument as a reference standard
e.g. MetopA/HIRS
Potentially large uncertainties introduced by spectral correction
Define a Synthetic Standard
Expanded here for further study

13. Requirements to Define a Synthetic Standard
Hybrid of different GEO instruments
Needs to be static?
What mix of GEO satellites? - Current?
How to define mix of channels
Single or Multiple Channels?
Defined by correlation threshold of radiances
– or uncertainty of spectral conversion

14. Options to Define a Synthetic Standard
Manually crafted SRFs
+ Simple
- Arbitrary!
Define a Hybrid of GEO instruments
Rectangular SRF with bandwidth defined by mean wavelengths
where SRFs cross 50%
containing 95% (?) of cumulative radiance of full SRF
+ More accurate reflect SRF of absorbing channels
- Need to convolve with realistic radiance spectra – from where?

15. Simple Choice: 3.9 μm Channels
Define Rectangular SRF by mean wavelength at which constituent SRFs pass 50%

16. Harder Choice: WV Channels
Different instruments have channels of different widths
Some instruments have multiple channels within the same band

17. 6.2μm Channels

18. 10.8μm Channels

19. 12.0μm Channels

20. 13.4μm Channels

21. Error introduced by Spectral Conversion
Compare Radiances obtained from IASI radiance spectra convolved with Actual Meteosat SRFs and Standard Rectangular SRFs (Meteosat-8 and Meteosat-9):

22. Uncertainty introduced by Spectral Conversion
Compare Radiances obtained from IASI radiance spectra convolved with:
Actual Meteosat SRFs
Standard Rectangular SRFs
Calculate Residuals of Quadratic fit over full range of radiances:
Channel
IR3.9
IR6.2
IR7.3
IR8.7
IR9.7
IR10.8
IR12.0
IR13.4
STD SRFs
0.72
0.73
1.10
MSG Only
0.03
0.10
0.08
MSG SRFs
0.77
0.09
0.06
0.05
0.02
0.04
0.14
Uncertainty Analysis shows these map into systematic uncertainties of biases estimated from GSICS Correction evaluated for Standard Radiance scenes with similar order of magnitudes
Dominate the error budget for 3.9μm and WV channels
Need more sophisticated definition of SRFs to reduce errors

23. Defining MSG Common Reference
MSG Common Reference SRF
Defined by mean of inner and outer templates cross SRF=0.5

24. Results of Uncertainty Analysis
Using Specified Template MSG SRFs

25. Results of Uncertainty Analysis
Using Common Reference STD SRFs

26. Discussion Points
Still need products based on “actual” SRF of each instrument
But do we need to define this as a different product class?
What level of uncertainties are ‘acceptable’?
For current GEO imagers? – Near Real-time and Re-Analysis
For historic GEO imagers
How do we select the ‘best’ definition of common reference channels?
Who wants to investigate each option?
How do we compare them?
When?

28. Work Plan to Develop CRC Products
GRWG:
Define CRC SRFs – Discuss!
Define uncertainties – Discuss!
Develop prototype products – Discuss!
Global Composite Images? – Discuss!
GDWG:
Define file format, name conventions
Create directories on GSICS Data And Products Servers
GCC:
Seek feedback from beta-testers
Educate Users
GPPA

29. Handling Multiple References
The advantages of using multiple references
The problems of using multiple references
The solutions to using multiple references!

30. Advantages of multiple references
Robustness
In case of failure of one reference
Allows transition between references – e.g. MetopA->B
Greater coverage of diurnal cycle
Both scene and instrument calibration variability
Important for 3-axis stabilized spacecraft
Political

31. Disadvantages of multiple references
Segal's law:
"A man with a watch knows what time it is. A man with two watches is never sure." [Bloch, Arthur (2003). Murphy's Law. Perigee. p. 36.]
It refers to the potential pitfalls of having too much conflicting information when making a decision.
Metrological Traceability Difficulties
as the references can never be perfectly consistent

32. Solution to using multiple references?
Define only one as the calibration reference
All others are regarded as calibration transfer standards
For example:
Time series of homogenised HIRS data? – Covered later
Use AIRS is used as a transfer standard to complement the inter-calibration against IASI as the calibration reference.
Allows collocations with GEO imagers over the diurnal cycle
because of different orbits of Aqua and Metop satellites carrying these instruments
MODIS + Desert BRDFs characterised by other instruments?
Discuss!

33. Instruments + Models as References
Often models are included in inter-calibration chain
Gap-filling method for AIRS
BRDF in DCC
Can models be used as a calibration reference?
“Yes - if you can calibrate it” [Rüdiger Kessel]
To combine instruments and models as reference:
‘Calibrate’ models against observations from the calibration reference instrument ?
Discuss!

34. Error Propagation with Multiple References
We need to accounted for this when constructing the uncertainty evaluation required for products derived from multiple references
How?
For example:
Evaluate errors when transferring calibration between references
Evaluate model errors as part of model calibration
Discuss!

35. Strategy for Migrating References
MetopA/IASI to MetopB/IASI
Aqua/AIRS to NPOES/CrIS
NOAAn/HIRS to NOAAn-1/HIRS

36. Why Migrate References?
Nothing lasts forever (except GSICS!)
Satellites (and their instruments) have finite life spans
Satellite programs have longer life spans, but still finite
“There can be only one!”
Better, faster, cheaper* instruments come along
(*Choose 2 of the above)

37. Specific Examples of Migrating Reference
MetopA/IASI to MetopB/IASI … to MetopC/IASI
Same instrument, different platforms, ~same orbit
Aqua/AIRS to NPP/CrIS … to JPSS1/CrIS … to JPSS2/CrIS …
Different instrument, different platforms, ~same orbit
NOAAn/HIRS to NOAAn-1/HIRS
~Same instruments, different platforms, different orbits
Homogenise time series by SNO
Reverse chronological order
Tie to IASI as one end
Discuss!

38. Strategy for Re-Analysis Products
HIRS-HIRS
Handling different dataset versions
Tying one end of series …

39. EUMETSAT Re-Analysis of Meteosat Archive
EUMETSAT plans to process a re-analysis of whole Meteosat archive
Requires development of inter-calibration product
using homogenised HIRS data as a reference, [Shi & Bates, 2011]
based on GSICS methods
EUMETSAT agreed to draft a detailed work plan for these developments, which Tim Hewison will present this at the next GRWG meeting in March 2011, where it will be reviewed and attempts made to coordinate these activities with related activities of other groups.

40. Sustained Climate Information Flow
Satellite & In Situ
Observations
Satellite data
Environmental
Data Records
Long-term Information Preservation
Fundamental
Climate Data
Records
Interim
Thematical
Climate Data Records
Climate
Information
Short scale
physical phenomena
monitoring
Operational Climate
Monitoring supporting
Climate Services
Longer term climate variability
& climate change analysis
Adaptation
+ mitigation planning
(decision making)
Sustained Applications
Major
model-based
Reanalysis
Short and Medium Latency
Re-calibration
Inter-calibration
Reprocessing
Observing system
performance
Monitoring
and automated
corrections
Data conversion
User Services
Archived Satellite Data and
Sustained Coordinated Processing
SCOPE-CM
Climate Information Records (CIRs) provide specific information about environmental phenomena of particular importance to science and society (e.g., hurricane trends, drought patterns)
The term Fundamental Climate Data Record (FCDR) is used to denote a long-term satellite data record, generally involving a series of instruments (all platforms), with potentially changing measurement approaches, but with overlaps, calibration and quality control sufficient to allow the generation of homogeneous products providing a measure of the intended variable that is accurate and stable enough for climate monitoring. FCDRs include the ancillary data used to calibrate them. For “one-off” research type measurements, the principles do not apply, but as many of the other principles as possible (e.g., those for rigorous instrument characterisation prior and during operations, complementarity of surface and satellite-based observations, etc.) should be followed.
The term Product denotes, values of fields of Essential Climate Variables derived from FCDRs. Products may be generated by blending satellite observations and in situ data, or by blending multiple in situ or multiple satellite data sources. Some products are generated within model assimilation schemes. Some satellite-based products are created by using the laws of radiative transfer to retrieve ECVs from the FCDRs. Other documents use the term Thematic Climate Data Record (TCDR) for such products; in the GCOS Implementation Plan , the term Integrated Climate Product was used. The development of products may require strong collaboration between organisations responsible for the generation of datasets (e.g., meteorological services, oceanographic centres, environmental agencies, space agencies) and the separate research or operational groups that generate products, to ensure continuous refinement and extension. Adequate details of the product generation approach need to be documented and made available, along with the products, to ensure repeatability and incremental improvement of the products.
For further discussion of the terms Fundamental Climate Records and Thematic Climate Data Records see e.g. National Research Council (2004): Climate Data records from Environmental Satellites, the National Academic press, Washington D.C., USA, 150pp.

41. METEOSAT 1984-2005 Archive evaluation using radiosondes
Upgrade of calibration technique
(van de Berg, et al., 95)
Upgrade of calibration technique
(Schmetz, 1989)
ISCCP DX Normalized Instead of nominal
Comparisons between the METEOSAT BTs and the simulated BTs from radoisoundings: (+) represent the raw data, (◊) represent the homogeneised data. The histogram shows the nb of soundings used for comparison.
Can we do better than that and extend to SEVIRI?
Courtesy of Helene Brogniez and Rémy Roca, LMD

42. Plan for a MVIRI and SEVIRI Infrared Channel Inter-calibrated Data Set within GSICS
EUMETSAT proposed within GSICS the inter-calibration of the Meteosat IR channels for the complete series of satellites to arrive at a homogeneous data series that is consistent with HIRS;
It will use inter-calibrated HIRS from NOAA/NCDC (adjusted to HIRS on NOAA-12) with GSICS collocation and inter-calibration procedures taking HIRS as truth;
Control inter-calibration in overlapping areas after IODC started;
Derive CSR, AMV and other products from inter-calibrated radiance
Document changes with respect to prior reprocessed data;
Aim to ultimately tie homogenized HIRS time series to MetopA/IASI
Development and implementation targeted for 2011/2012.

43. Inter-Calibration Plan for Meteosat Re-Analysis
From archive of data from:
MVIRI & SEVIRI on Meteosat-2 to -9
HIRS on TIROS-N, NOAA-6 to -17 and Metop-A – Clear Sky radiance only from Shi & Bates [2011]
Compile test dataset from 2004:
MVIRI on Meteosat-5 and -7
HIRS on NOAA-16 and -17
Apply GSICS collocation criteria
These may require fine tuning later
Analyze results of regression of collocated radiances
Using GSICS baseline methodology
Investigate impact of using only clear sky radiances
by comparison with MVIRI-IASI and MVIRI-HIRS results
Conduct Uncertainty Analysis

44. Inter-Calibration Plan for Meteosat Re-Analysis
Make recommendations from this analysis:
Are clear-sky only inter-calibrated HIRS time series adequate for:
UTH applications?
Other applications – e.g. Cloud Top Height?
What levels of uncertainty can be expected?
How stable the Meteosat instruments are?
Do we need to change the collocation smoothing period?
Or do we need to homogenize HIRS data for all-sky conditions?
Summer 2011 Milestone decision: How to proceed…
Discuss GSICS support for other re-analysis plans!

45. Re-Analysis with Radiosonde References
EUMETSAT also plan to evaluate the use of the ARSA database
(Analyzed Radio Soundings Archive) from LMD
With RTTOV (fast RTM)
– compare with collocated Meteosat radiances
As validation for inter-calibration
GRUAN-GSICS Coordination:
It could it be good to address the GRUAN and the connect to GSICS? I will forward you relevant discussions from the last GCOS/AOPC meeting. Idea would be to use WV radiance spectra from IASI and AIRS and compare with calculated spectra from GRUAN which have very good humidity measurements for UTLS. [ Jo Schmetz, GSICS Exec Panel]

46. GCOS/AOPC-XVI 7-11 February 2011 Draft Summary and Recommendations
8.2 NOAA Activities related to Climate Monitoring, including the MW-sounding temperature consensus data set initiative
The Panel appreciated the paper by Goldberg (Doc 8.2) showing how comparisons of AIRS and IASI radiances with those simulated from reanalyses can be used for independent validation of the climate reanalyses, in particular to improve understanding and reconcile differences between the reanalyses. It noted that to reduce differences between the reanalyses there needs to be an effort to use radiative transfer code which agree at the 0.2 K level with observed values. Using GRUAN-like radiosondes collocated with AIRS and IASI are critical for reducing uncertainties in radiative transfer and to demonstrate the accuracy of the radiative transfer algorithm. In doing so, the reanalysis and NWP communities will be encouraged to assimilate the original AIRS and IASI data without bias corrections, and remaining differences between reanalyses will be primarily due to model physics.
(Text from June 2010 Notes:)
The Panel encouraged the continuation of these spectrally resolved infrared radiances to inter-compare and assess climate reanalyses from NCEP, ECMWF, JMA, NASA. It further stressed the importance of providing and sustaining high quality in-situ observations through programmes such as GRUAN to improve radiative transfer models.
Discuss!

47. Strategy for LEO-LEO Inter-Calibration
PATMOS-x
HIRS-HIRS
Aqua/MODIS-Terra/MODIS …

48. General Strategy for LEO-LEO Products
Monitored Instruments?
AVHRR, HIRS, AIRS, Landsat TM…
Reference Instruments?
IASI, MODIS, GOME2, …?
Methodology?
Direct Comparison of Collocated Radiances (SNOs)
Other
Products?
GSICS Corrections } for nominal SRFs
GSICS Bias Monitoring } and CRCs
Near Real-Time
and Re-Analysis Migrating references

49. Specific Plans for LEO-LEO Products
PATMOS-x
Andy Heidinger
HIRS-HIRS
Shi & Bates
Aqua/MODIS-Terra/MODIS
Dave Doelling?
Landsat7/ETM-Terra/MODIS
Gyanesh Chander …