Tim Hewison (EUMETSAT) (GRWG Chair) - · Tim Hewison (EUMETSAT) (GRWG Chair) ... GSICS / CEOS-IVOS ... Establishment of a Lunar Observation Database

Strategy for calibration references

Tim Hewison (EUMETSAT)

(GRWG Chair)

Special Issue of the IEEE TGRS on “Inter-Calibration of Satellite Instruments”: Special Issue of the IEEE TGRS on “Inter-Calibration of Satellite Instruments”:

02 September 2015 Slide: 2

Overview

• Update on Prime GSICS Corrections – For information

• How to select inter-calibration reference instruments

– For information

• Improving the Moon

– As a calibration reference – A case for international collaboration – Recommendation needed!


Overview



– For information




1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Time /years

Availability of References

Mon

Ref1

Ref2

Ref3

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Time /years

Modified References

Mon

Ref1

Ref2

Ref3

New Prime GSICS Corrections • Define one Primary GSICS Reference

– For each spectral band/application – By consensus agreement within GSICS

• Use others as Transfer References • Blend corrections from all references

– After modifying Corrections to Primary GSICS Reference

• Ensures long-term continuity – Without calibration jumps

• Ensures Traceability – back to single Primary Reference

• Simplifies users’ implementation • Could also blend multiple methods? 0

20406080

100

1 2 3 4 5 6 7 8 9 1011121314151617181920

Wei

ghtin

g /%

Time /years

Simple Weighting of Each Reference

%Ref1

%Ref2

%Ref3


Applied by User

Correcting the Corrections and Blending References

Reference-1 (Primary)

Monitored Instrument

GSICS Correction, g1

Mon1

Reference-2 (Secondary)

GSICS Correction, g2

Mon2

MonRef1

Delta Correction, g1/2

21

Derived by GSICS

Modified Correction, g2,1/2

Mon 21

Prime GSICS Correction, g0

Mon1

-

g ̅

+


Users’ Application of Prime GSICS Correction

Monitored Instrument

Prime GSICS Correction, g0

Mon1 MonRef1


Overview



– For information




Reference Instrument Selection/Migration

• GSICS products need an inter-calibration reference • Inter-calibration Reference is one instrument

– Other instruments may be used as transfer references

• We need a way to select the reference – From a list of candidates

• First question should be: “Is it suitable at all?” – Does it meet minimum threshold requirements (availability, coverage, ...)

• Started asking “How good is it as a Reference?” – Impact on overall uncertainty – Score scheme


Example for SEVIRI-IASI IR

• OneDrive Spreadsheet • All values for illustration only • And subject to change • Some variables more important for

different inter-cal products • Timeliness important for NRTC • Date range critical for Archive Re-

Calibration • Assumed time-stamp is correct! • Scheme needs road testing with other

References

9

Scoring Scheme for GSICS Re-Analysis Correction for SEVIRI IR Channels Threshold Saturation MetopA/IASI Unit Min Max Min Max Weight Min Max OK? Perfect Score Date Range Year 2013 2013 2006 2030 90 2007 2020 Pass 54% 48.8 Geographic Coverage: Lat deg -10 10 -90 90 2 -90 90 Pass 100% 2.0 Geographic Coverage: Lon deg -10 10 -180 180 2 -180 180 Pass 100% 2.0 Dynamic Range K 270 300 180 330 5 180 310 Pass 87% 4.3 Spectral Range cm-1 746 2564 650 2800 10 645 2760 Pass 98% 9.8 Geometric Range: VZA deg 5 15 0 90 2 0.5 55 Pass 61% 1.2 Geometric Range: VAA deg 0 Pass 0.0 Geometric Range: SZA deg 0 Pass 0.0 Geometric Range: SAA deg 0 Pass 0.0 Geometric Range: Pol deg 0 Pass 0.0 Diurnal Coverage hr 9 10 0 12 20 7.8 11.2 Pass 28% 5.6 Field of View km 300 3 1 12 Pass 100% 1.0 Number of obs/day /d 0 Pass 0.0 Number of Collocations /d 1 10000 5 30000 Pass 100% 5.0 Geolocation accuracy km 10 0.1 10 3.3 Pass 3% 0.3 Polarisation knowledge deg Pass 0.0 Radiometric Stability K/yr 1 0.001 10 0.05 Pass 2% 0.2 Orbital Stability hr/yr 12 0.1 0 0.001 Pass 100% 0.0 Radiometric Noise K 10 0.1 1 0.15 Pass 67% 0.7 Uncertainty from SBAF K 1 0.01 20 0.008 Pass 100% 20.0 Spectral Resolution cm-1 100 0.5 0 0.25 Pass 100% 0.0 Spectral Stability cm-1/yr 2 0.01 0 2E-06 Pass 100% 0.0 SBAF Uncertainty K 1 0.001 10 0.15 Pass 1% 0.1 Absolute Calibration Acc K 1 0.001 10 0.05 Pass 2% 0.2 Inter-channel calibration K 2 Pass 0.0 Data Availability 1 1 Pass Documentation 1 1 Pass Community adoption 1 1 Pass Traceability 1 0 Fail Total 200.0 96% 51% 101.1

https://onedrive.live.com/redir?resid=A3160061748CB9AF!1016&authkey=!AMUtwqOsSIod_Ew&ithint=file,xlsx


Expanding WMO OSCAR Capabilities

• WMO OSCAR – Observing Systems Capability Analysis and Review Tool

• Adding Landing Pages with links to: – Instrument Specifications, Calibration Events, Data Outages, Instrument

Monitoring

• WMO expanding OSCAR – automatically assess instruments’ capability to address requirements – Based on pre-defined Expert Rules – Could be used to assess GSICS References – Action: Manik to interact with Jérôme to push forward the development of

the expert system as a tool to select inter-calibration reference instruments

http://www.wmo-sat.info/oscar


GSICS - New Delhi- OSCAR 16-20/03/2015

Instrument view

11


Overview



– For information




Lunar Calibration Workshop

Participants in the Lunar Calibration Workshop Hosted at EUMETSAT, 1-4 Dec 2014

• Organised by EUMETSAT in collaboration with USGS, CNES, NASA

• Almost 30 people from 14 different agencies and departments attending (physically or remotely) GSICS / CEOS-IVOS

• All participants used GSICS Implementation of ROLO model (GIRO) developed at EUMETSAT in collaboration with T. Stone (USGS)

• To develop inter-calibration algorithm, need accurate oversampling factors difficulty encountered by all participants

• Major outcomes: • the calibration community agreed the GIRO to be the

international reference for lunar calibration, traceable to the ROLO model from USGS

• Establishment of a Lunar Observation Database • Recommend investigating possible funding to establish

infrastructure to traceable lunar observations

• A GIRO usage policy and a data policy will be defined in collaboration with the participants

• USGS expected to have a new version of the ROLO by the end of 2015 updated GIRO by mid-2016

• All participants agreed about the need to organise another Lunar Calibration Workshop (~1 year time)

https://gsics.nesdis.noaa.gov/wiki/Development/LunarCalibrationWorkshop


Team Satellite Sensor G/L Dates

Number of Obs

(GSICS dataset)

Phase angle range (°)

CMA FY-3C MERSI LEO 2013-2014 9 [43 57] CMA FY-2D VISSR GEO 2007-2014 CMA FY-2E VISSR GEO 2010-2014 CMA FY-2F VISSR GEO 2012-2014 JMA MTSAT-2 IMAGER GEO 2010-2013 62 [-138,147] JMA GMS5 VISSR GEO 1995-2003 50 [-94,96] JMA Himawari-8 AHI GEO 2014- -

EUMETSAT MSG1 SEVIRI GEO 2003-2014 380/43 [-150,152] EUMETSAT MSG2 SEVIRI GEO 2006-2014 312/54 [-147,150] EUMETSAT MSG3 SEVIRI GEO 2013-2014 45/7 [-144,143] EUMETSAT MET7 MVIRI GEO 1998-2014 128 [-147,144]

CNES Pleiades-1A PHR LEO 2012 10 [+/-40] CNES Pleiades-1B PHR LEO 2013-2014 10 [+/-40]

NASA-MODIS Terra MODIS LEO 2000-2014 136 [54,56] NASA-MODIS Aqua MODIS LEO 2002-2014 117 [-54,-56] NASA-VIIRS NPP VIIRS LEO 2012-2014 20 [50,52]

NASA-OBPG SeaStar SeaWiFS LEO 1997-2010 204 (<10 , [27-66])

NASA/USGS Landsat-8 OLI LEO 2013-2014 3 [-7] NASA OCO-2 OCO LEO 2014

NOAA-STAR NPP VIIRS LEO 2011-2014 19 [-52,-50] NOAA GOES-10 IMAGER GEO 1998-2006 33 [-66, 81] NOAA GOES-11 IMAGER GEO 2006-2007 10 [-62, 57] NOAA GOES-12 IMAGER GEO 2003-2010 49 [-83, 66] NOAA GOES-13 IMAGER GEO 2006 11 NOAA GOES-15 IMAGER GEO 2012-2013 28 [-52, 69] VITO Proba-V VGT-P LEO 2013-2014 25 [-7] KMA COMS MI GEO 2010-2014 60 AIST Terra ASTER LEO 1999-2014 1 -27.7 ISRO OceanSat2 OCM-2 LEO 2009-2014 2 ISRO INSAT-3D IMAGER GEO 2013-2014 2

Instruments with lunar observation capabilities, with the minimum number of Moon observations provided to the GSICS Lunar Observation Dataset (GLOD) - more observations may be available.

Instruments


Examples of Moon observations from participating instruments


16

LEO (CWL) GEO (band)

According to OSCAR (WMO)

Courtesy B. Fougnie, CNES

Landsat-8


17

LEO GEO


Landsat-8


18

LEO GEO


Landsat-8


Improving the Moon (as a reference)

• Where we are: – GSICS Implementation of ROLO (“GIRO”) v1.0.0 now available, validated against ROLO – Starting to compile GSICS Lunar Observations Database (“GLOD”) – Already found significant dependencies in relative biases of Observed-GIRO irradiances – But still good enough to inter-calibrate 0.6µm band to <1% uncertainty (NIR not so good) – Absolute calibration using GIRO limited to 5-10% by original ROLO dataset

• Short-term plan: – Develop inter-calibration algorithm – Integrate results with DCCs in new GSICS products – Continue to expand GSICS Lunar Observations Database (“GLOD”) – Form community consensus on absolute scale of GIRO

• Long-term plan: – Tie GIRO to SI-traceable standards – Ideally through long-term campaign of new hyperspectral lunar observations


Proposal for activities to address inter-calibration using the Moon

In EUMETSAT’s proposal, points to be addressed: A. Estimating the over-sampling factor B. Drift correction C. Estimation of the SBAFs from GIRO and verification of the

results. D. Establishing an absolute scale for lunar calibration

E. Medium and long term activities on lunar calibration


ROLO/GIRO and absolute calibration • KNOWN DEFICIENCIES OF THE CURRENT REFERENCE

• SI traceability • Absolute scale uncertainty (5 – 10 % uncertainty in absolute irradiance scale) • Residual geometry dependencies • Multi-band spectral coverage

• ESTABLISHING THE REQUIREMENTS • Defining potential improvements and requirements on user applications • Independent method to verify that operational visible calibration meets requirements • Support climate applications and past instruments (e.g. no onboard calibration available)

• GOAL

• Getting absolute uncertainties to under ~1% • Surface stability + predictability of the Moon’s brightness = lunar reflectance model valid for all times • Level of accuracy of 0.5% is achievable

• WHAT IS NEEDED? Need for a new lunar measurement program (= many years of observations, to characterize the

variations of the lunar brightness with phase and librations) Minimum 3 years to capture the libration dependence within the bounds of its relative effect

(ROLO operated for more than 8 years) Traceable high-spectral resolution observations over several years with specific measurements to

validate the atmospheric correction.

• FUNDING: Currently under investigation (inter-agency [international] collaboration?)


E. Medium and long term activities on lunar calibration

1. Investigate the possibility to participate to an international consortium to fund dedicated measurement campaigns.

Review planned activities in this area: 1. CMA: Obtained funding for new ground-based observations from CAS 2. CNES? 3. ESA: interest in contributing but no existing mechanism for such an

international funding. Further investigations needed. 4. EUMETSAT: no existing mechanism for international funding but possibility still

under investigation. Science Working Group defines users requirements. 5. ISRO? IMD? 6. JMA: no possibility for funding 7. KMA? 8. NASA: Proposed ARCSTONE mission for Earth Venture funding 9. NOAA? 10. CEOS/IVOS members: AIST, JAXA, KIOST, VITO?

https://gsics.nesdis.noaa.gov/pub/Development/20150316/1p_ARCSTONE_2015_03.pptx


Conclusions

• Advice of CEOS-WGCV for GSICS: – Develop inter-calibration using current GIRO as soon as possible – Support SI-traceable lunar observations to tie ROLO to absolute scale – Encourage independent reproduction of ROLO observations

• Requirements: – System suitable for operational application (source code) – Globally accessible common lunar irradiance model

• Recommendation is sought from GSICS/CGMS: – How to establish international cooperation to establish, operate and process – A campaign of new lunar observations over full range of conditions (>3yr) – Covering full reflected solar band at high spectral resolution (hyperspectral) – Directly traceable to SI standards


Thank You

Documents

Tim Hewison (EUMETSAT) (GRWG Chair) - · Tim Hewison (EUMETSAT) (GRWG Chair) ... GSICS / CEOS-IVOS ... Establishment of a Lunar Observation Database