1. MODIS-Terra cross-calibration for ocean color bands Ewa Kwiatkowska Bryan Franz, Gerhard Meister, Gene Eplee OBPG 30 January 2008

2. MODIS calibration changes since launch band 8 412nm SD door permanently opened B side electronics A side electronics pre-launch damage to mirror coating, MS2 Terra SD – Solar Diffuser frame 979, mirror AOI 50.3 0 SV – Space View frame 23, mirror AOI 11.4 0 MS1 – Mirror Side 1 MS2 – Mirror Side 2

3. MODIS trends in ocean color products Terra/Aqua ratios normalized water-leaving radiances L wn L wn (412nm) band 8 L wn (443nm) band 9 L wn (488nm) band 10 L wn (531nm) band 11 L wn (551nm) band 12

4. MODIS-Terra ocean color RVS band 8 412nm Terra response versus scan angle (RVS) in terms of normalized water-leaving radiances L wn L wn (412nm) band 8 Mirror side 1 MS1 Mirror side 2 MS2 -10% +10%

5. TOA sensor cross-calibration Goal Extract MODIS-Terra – RVS, and – polarization sensitivity For all ocean-color bands, per mirror side and detector Through the entire Terra mission Optionally, derive temporal calibration change Approach Vicarious calibration / cross-calibration using SeaWiFS SeaWiFS – stable spatially and temporally SeaWiFS – insensitive to polarization

6. TOA sensor cross-calibration nL w ( ’) TOAgaspolarizglint whitecap air aerosol L’ t ( ) = [ L r ( ) + L a ( ) + tL f ( ) + TL g ( ) + t d ( )L w ( ) ] · t g ( ) · f p ( ) water MODIS – vicarious SeaWiFS AOT( ’) ’  (fit in the ln-ln space) ’  (fit based on bio-optical models) L t ( ) MODIS – measured TOA radiance

7. Current implementation nL w ( ’) TOAgaspolarizglint whitecap air aerosol L’ t ( ) = [ L r ( ) + L a ( ) + tL f ( ) + TL g ( ) + t d ( )L w ( ) ] · t g ( ) · f p ( ) water MODIS – vicarious SeaWiFS MODIS operational assuming accuracy of MODIS NIR band calibration ’  (fit based on bio-optical models)

8. Polarization f p I t – Stokes vector exiting the TOA Mueller matrix M – instrument characterization, including polarization  - rotation angle I m – Stokes vector measured by a sensor I m = M R(  ) I t M 11 … M 14 M 21 … M 24 M 31 … M 34 M 41 … M 44 1 0 0 0 0 cos2  -sin2  0 0 sin2  cos2  0 0 0 0 1 I t Q t U t 0 I m = Polarization correction f p = I m / I t I m = M 11 I t + M 12 (Q t cos2  – U t sin2  ) + M 13 (Q t sin2  + U t cos2  )

9. Polarization of the atmosphere degree of atmospheric polarization d p air molecule (Rayleigh) and glint scattering polarization correction f p pre-launch MODIS characterization  Q 2 + U 2 I d p = MODIS Terra swath 412nm band 8

10. Time series and global optimization Global optimization for each day to derive M 11, M 12, and M 13 M 11, M 12, M 13 = fn(mirror AOI) Global time series throughout the entire Terra mission – MODIS Terra – one day a month global set of granules (90 days) – SeaWiFS – 9-day global L3s centered on each month’s day from MODIS – Strict screening criteria for SeaWiFS L3 bins and MODIS L2 pixels Known: L t – MODIS-measured L’ t – MODIS-vicarious Q t and U t – atmospheric molecular and glint polarization, Rayleigh tables, glint derivation  – theoretical calculation L t = M 11 L’ t + M 12 (Q t cos2  – U t sin2  ) + M 13 (Q t sin2  + U t cos2  ) Unknown: M 11 – RVS and absolute calibration M 12 and M 13 – polarization sensitivity

11. 6 March 2000 1.0 0.99 0.98 1.0 1.01 0.99 1.01 1.0 0.99 1.0 1.01 1.02 1.01 1.02 1.01 1.006 1.014 1.015 1.005 1-1.5% 1-2% 2% 1.5% 1% Detector 4 Mirror side 1 Mirror side 2 412nm 443nm 488nm 531nm 551nm 667nm 678nm polynomial across scan RVS M12 M13 3-rd degree linear linear

12. 1.015 1.025 1.005 1.01 1.02 1.0 1.02 1.0 1.02 1.04 0.98 1.02 0.94 0.98 0.85 0.95 1.05 Detector 4 Mirror side 1 Mirror side 2 16 October 2007 15% 7% 6% 3% 2% 1.3% 1.4% 412nm 443nm 488nm 531nm 551nm 667nm 678nm polynomial across scan RVS M12 M13 3-rd degree linear linear

13. Blue band temporal trends Mirror side 2 Detector 4 Mirror side 1 Space View (lunar) frame Nadir frame Solar Diffuser frame RVS

14. Blue band temporal trends Mirror side 2 Detector 4 Mirror side 1 M12 Space View (lunar) frame Nadir frame Solar Diffuser frame

15. Mirror side 2 Detector 4 412nm 443nm 488nm 531nm 551nm 667nm 678nm 1.00 1.04 0.96 1.08 1.00 1.04 0.96 1.08 1.00 1.04 0.96 1.08 1.00 1.04 0.96 1.08 1.00 1.04 0.96 1.08 1.00 1.04 0.96 1.08 1.00 1.04 0.96 1.08 Space View (lunar) frame Nadir frame Solar Diffuser frame Temporal trends RVS M12 M13

16. Pre-launch M13 RVS M12 M13

17. Temporal smoothing Polynomial fit through time into derived RVS and M12 per individual frame

18. Smoothed results October 2007 March 2000 Mirror side 1 Detector 4 Daily RVS and M12 use temporally smoothed values for all frames RVS M12 412nm 443nm 488nm 531nm 551nm 667nm 678nm

19. Smoothed results October 2007 March 2000 Mirror side 1 Detector 4 Daily RVS and M12 use temporally smoothed values for all frames RVS M12 412nm 443nm 488nm 531nm 551nm 667nm 678nm

20. Verification of the vicarious cross-calibration with lunar measurements band 8 412nm Terra detector ratios detector 10 / detector 1 lunar measurements — vicarious cross-cal 412nm band 8 Mirror side 1 MS1 Mirror side 2 MS2

21. Future plans Why seasonality in M13? Is extracted polarization sensitivity too large? RVS? Global time-series testing for water-leaving radiances Validation of RVS and detector calibration Application of the cross-calibration to Aqua