1 / 29

Monitoring of SST Radiances in NOAA MICROS System www.star.nesdis.noaa.gov/sod/sst/micros/

Xingming Liang 1,2 and Sasha Ignatov 1 1 NOAA/NESDIS/STAR, 2 CSU/CIRA. Recent Progress with MICROS. Monitoring of SST Radiances in NOAA MICROS System www.star.nesdis.noaa.gov/sod/sst/micros/. GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany .

lloyd
Download Presentation

Monitoring of SST Radiances in NOAA MICROS System www.star.nesdis.noaa.gov/sod/sst/micros/

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Xingming Liang1,2 and Sasha Ignatov1 1NOAA/NESDIS/STAR, 2CSU/CIRA Recent Progress with MICROS Monitoring of SST Radiances in NOAA MICROS System www.star.nesdis.noaa.gov/sod/sst/micros/ GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 1 of 20

  2. ACSPO SST Team J. Stroup, X. Zhou, K. Saha, B. Petrenko, Y. Kihai, J. Sapper Calibration Community T. Hewison, T. Chang, F. Wu, F. Yu, L. Wang, F. Weng, C. Cao, M. Goldberg – GSICS Team A. Wu, J. Xiong– MODIS Calibration Support Team (MCST) NOAA Community Radiative Transfer (CRTM) Team Y. Chen, Q. Liu, Y. Han, F. WengJ. Stroup, X. Zhou, K. Saha, B. Petrenko, Y. Kihai, J. Sapper NOAA Soundings Team S. Yi, W. Wolf Acknowledgements GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 2 of 20

  3. Outline Progress with MICROS since GSICS’2013 Sustained near-real time monitoring of AVHRR GAC (NOAA-16, -18, -19; Metop-A, -B), FRAC (Metop-A, -B), MODIS (Terra, Aqua), and VIIRS (S-NPP). Near-Real Time MICROS now covers period from July 2008 ̶ present Added 12 years of reprocessed AVHRR GAC data (2002 ̶ 2014) from NOAA-15, -16, -17, -18, -19, Metop-A and –B Tested ECMWF vs. GFS profiles as input into fast CRTM, to minimize M-O biases and improve DDs Conclusion and Future Work GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 3 of 20

  4. Sustained Monitoring of 5 AVHRRs, 2 MODISs, 1 VIIRS Nighttime DD’s @3.7 µm (Ref = Metop-A GAC) …… ACSPO ver 1.0 CRTM r577 Reynolds SST ACSPO ver 2.3 CRTM ver 2.1 CMC SST • 5+ years of data available from Jul’2008 – present • AVHRRs on Metop-A, -B; N-17, -19; MODIS &VIIRS are more stable; N-16, -18 less stable • Cross-platform inconsistencies are in part of due to • Sensor characterization (Working with MCST, AVHRR CAL Teams to resolve) • CRTM changes (Code version, coefficients, first-guess SST and atmospheric profiles inputs) • Part of temporal variability is due to ACSPO & CRTM updates -- Reanalysis underway GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 4 of 20

  5. AVHRR GAC Reanalysis: 2002-Present • Included in Reanalysis are data from 7 AVHRRs (N-15, -16, -17, -18, -19 and Metop-A, -B) – GAC only • ACSPO V2.2 & CRTM V2.1 with Reynolds V2.0 Daily SST and GFS 1⁰x 26 levels consistently used • Only nighttime data are used in this presentation (Daytime data are also available in MICROS but they are not used here) GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 5 of 20

  6. NIGHTDouble Differences BT@3.7µm (Ref = N-17/Metop-A) • Stability (Best > ..> Worst): Metop-A > N-17 > Metop-B > N-19 > N-18 > N-16 > N-15 • Most stable Metop-A & N-17 used as references • Stability for longwave IR11&12 (not shown) is similar to IR37 • Least stable are N-15, -16 (after 2006), and -18 (after 2011) GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 6 of 20

  7. NIGHTDouble Differences SST (Ref = N-17/Metop-A) • Artifacts in SSTs are strongly linked to instabilities in BTs • Most stable are Metop-A, -B, N-17, -19 • Least stable are N-15, -16 (after 2006), -18 (after 2011) • Need help of Calibration community to fix AVHRR Fundamental CDR! GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 7 of 20

  8. Testing ECMWF vs. GFS profiles as CRTM input • Based on prior sensitivity analyses (Saha et al, JGR 2012), first-guess daily SST in CRTM was updated from Reynolds to CMC in June 2013 • As a result, time series of mean M-O biases and DDs became more stable, and STDs of M-O biases reduced • Next step is to optimize the atmospheric profiles input to CRTM • Here, we test ECMWF vs. current implementation - NCEP GFS (both real-time products) • Use 1 month of global data from NOAA-16 -18, -19 AVHRRs; Terra and Aqua MODISs; and S-NPP VIIRS from 18 Jan – 17 Feb 2014 • For long term reanalysis in MICROS, we also plan to explore ERA-Interim and MERRA climate reprocecessings GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 8 of 20

  9. ECMWF vs. GFS GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 9 of 20

  10. Histograms Effect of using ECMWF vsGFS GFS • Number of clear-sky pixels increases by ~3% (positive effect on ACSPO cloud mask) • Histograms are slimmer (Global standard deviations of M-O biases smaller) • Unusually warm M-O biases in longwave IR11 &IR12 become closer to 0 • IR11 is affected less than IR12 • IR37 is affected only minimally ECMWF GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 10 of 20

  11. Example Global Map of M-O biases in IR12 • Elevated M-O biases in the tropics may suggest that TPW in GFS is underestimated • Biases are significantly reduced and more spatially flat when ECMWF is used • Smaller cold spots appear in ECMWF implementation, suggesting that ECMWF TPW may be (slightly) overestimated GFS ECMWF GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 11 of 20

  12. Wind Speed & TPW Dependencies of M-O Biases ECMWF GFS High TPW and Wind Speed Dependencies significantly improved when ECMWF is used GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 12 of 20

  13. Time Series of M-O Biases Using GFS Typical M-O biases in IR37 are ~0.2K IR37 ~0.2K IR11 Typical M-O biases in IR11 & IR12 are ~0.5-0.6K ~0.5K IR12 GFS ~0.6K GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 13 of 20

  14. Time Series of M-O Biases Using ECMWF M-O biases in IR37 change very little, as expected IR37 ~0.2K IR11 M-O biases in IR11 & IR12 reduce by ~0.2K ~0.3K IR12 ECMWF ~0.4K GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 14 of 20

  15. Double difference (IR11) GFS and ECMWF DDs are within several 0.01K GFS However, for ECMWF they seem to cluster a little closer to the desired “0” line Also, the stability (measured by STDs) are somewhat better for ECMWF ECMWF GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 15 of 20

  16. Conclusion and Future work • Conclusion and Future Work GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 16 of 20

  17. Conclusion • NRT monitoring of 5 AVHRRs, 2 MODISs and VIIRS in MICROS sustained • 5+ years of data are now available (2008-2014) • AVHRRs on Metop-A, -B, N-17, -19, and MODIS & VIIRS are stable • AVHRRs on N-16 and -18 are less stable • Documented in Liang and Ignatov, JGR, 2013 • 12+ years (2002-pr) of 7 AVHRRs GAC reanalyzed & displayed in MICROS • Use consistent versions of ACSPO & CRTM, and Reynolds/GFS input data • Some AVHRRs are more stable and some less so • Ranked stability: Metop-A > N-17 > Metop-B > N-19 > N-18 > N-16 > N-15 • ECMWF tested as CRTM input as alternative to current GFS • More clear-sky ocean pixels (positive effect of cloud mask) • STDs of M-O biases are smaller for IR11 & 12; Unchanged for IR37 • Warm M-O bias in IR11&12 become closer to 0 • Stability slightly improved for all bands and SST GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 17 of 20

  18. Future Work • Extend AVHRR GAC ACSPO-RAN back to ~1980 & Display in MICROS. Need help from sensor CAL community to Improve AVHRR Sensor Radiances (Calibration, Spectral Responses) • Extend ECMWF vs. GFS comparisons. Additionally analyze ERA-Interim & MERRA profiles • Analyze & Improve daytime M-O biases and DDs • Complete and document consistent CRTM coefficients analyses • Add GEO (MSG/SEVIRI, Himawari/AHI, GOES-R/ABI) in MICROS • Extend MICROS to Include Reflectance Bands (use CRTM in conjunction with GOCART/NAAPS aerosol data) • Work with CAL/GSICS community, to explore adding hyper-spectral (AIRS/IASI) transfer standard, in addition to CRTM GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 18 of 20

  19. Back up

  20. Const. CRTM Coeff. Effect on M-O bias and DDs • Effect of consistent CRTM coefficients on M-O bias and DDS in micros GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 20 of 20

  21. Motivation MODIS CRTM Coefficients Recalculated Metop-B CRTM Coefficients: Calculated Correctly? from www.star.nesdis.noaa.gov/sod/sst/micros • MICROS DDs are well suited to evaluate the sensors stability, • but cross-platform inconsistencies are at least in part due to inconsistent CRTM coefficients for different sensors/platforms. • Sensitivity analysis of CRTM Coefficients is critically important for understanding and minimizing model effect on MICROS DD GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 21 of 20

  22. Const. CRTM Coeff. Effect on M-O bias and DDs • 4 consistent CRTM Coefficients datasets were analyzed in MICROS using 1 day of global data (15 Jan 2013) • ODAS-ORD, ODAS-PW, ODPS-ORD, and ODPS- PW [Y. Chen et al., 2010] • Based on the same baseline LBLRTM v11.7. • Different CFC absorption effect on M-O and DDs were analyzed • CFC-free, 50% CFC, and 100% CFC • Compare 4 different results using the following metrics • Global M-O biases. M-O biases are expected to be (1) closer to zero and (2) coherent across bands. • Global STDs of M-O biases. STDs are expected to be smaller. • Double differences for pairs of platforms in close orbits. • For Hi-Res, (1) Metop-A and -B, and (2) S-NPP and Aqua. • For GAC, (1) Metop-A and -B and (2) NOAA-18 and -19. GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 22 of 20

  23. Comparison of M-O bias and STDs for 4 Coeff. sets • PW-ODPS is recommended • PW-ODPS has smaller or comparable STDs with others • PW-ODPS M-O biases are smaller or comparable with others • Similar performances are observed in IR11&12 (No shown) • Generally,DDs are insensitive to ORD-PW, ODAS-ODPS within ±0.05K except for Aqua-NPP (due to Aqua CAL) GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 23 of 20

  24. Const. CRTM Coeff. Effect on M-O bias and DDs GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 24 of 20

  25. CFC absorption Effect on M-O biases • Change of CFC amount are only minimal effect on M-O biases in IR37, but reduce M-O biases by ~0.2 K for IR11 and ~0.3 K for IR12. • The CFC effect of M-O biases for different platforms are different, and expect as one major factor to influence DDs. GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 25 of 20

  26. CFC absorption Effect on M-O bias and DDs GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 26 of 20

  27. Double differences (IR37) • Stabilities are slightly improved when using ECMWF instead of GFS • Difference from M-O bias, in which Improvement form ECMWF are only for IR11 &12, stabilities get consistently improved in all three bands and SST. GFS One month of data were generated to compare DDs between GFS and ECMWF (Jan. 18 – Feb. 17, 2014) ECMWF GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 27 of 20

  28. Double difference (IR12) • Stabilities are slightly improved when using ECMWF instead of GFS • Difference from M-O bias, in which Improvement form ECMWF are only for IR11 &12, stabilities get consistently improved in all three bands and SST. GFS One month of data were generated to compare DDs between GFS and ECMWF (Jan. 18 – Feb. 17, 2014) ECMWF GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 28 of 20

  29. Double difference (SST) • Stabilities are slightly improved when using ECMWF instead of GFS • Difference from M-O bias, in which Improvement form ECMWF are only for IR11 &12, stabilities get consistently improved in all three bands and SST. GFS One month of data were generated to compare DDs between GFS and ECMWF (Jan. 18 – Feb. 17, 2014) ECMWF GSICS 2014 annual meeting, March 24-28, 2014, EUMETSAT, Darmstadt, Germany Slide 29 of 20

More Related