10. Future Work

1
Monitoring of IR Clear-sky Radiances over Oceans
for SST (MICROS) from NPP/VIIRS, Terra-Aqua/MODIS
and NOAA-Metop/AVHRR http//www.star.nesdis.noaa.g
ov/sod/sst/micros/ XingMing Liang1,2 ,Alexander
Ignatov1 and Korak Saha1,2 1NOAA/NESDIS/STAR,
2Colorado State University/CIRA
1. SST and MICROS
6. MICROS Overview
Sea surface temperatures (SST) are derived from
brightness temperatures (BT) in the IR window
bands of AVHRR, MODIS, and VIIRS sensors.
Multi-Channel and Non-Linear SST (MCSST/NLSST)
algorithms are unresolved combinations of several
BTs. To attribute anomalies in the retrieved SSTs
to a particular band, a near-real time web-based
Monitoring of IR Clear-sky Radiances over Oceans
for SST (MICROS) was established. MICROS is used
to evaluate the stability of the VIIRS BTs, and
their consistency with AVHRR and MODIS BTs.
Both day and night data are processed and
monitored in MICROS. However, only night data
are used here, to minimize effects of diurnal
cycle solar reflectance.
2. Objectives of MICROS
Monitor clear-sky sensor BTs in SST bands over
global ocean in Near-Real Time (Obs), against
CRTM with first-guess input fields (Model)

• Evaluate sensor radiances for stability and
  cross-platform consistency
• Contribute to Global Space-based
  Inter-Calibration System (GSICS)
• Understand minimize M-O biases in BTs SSTs
• Minimize the need for empirical bias correction
  facilitate physical SST
• Validate CRTM and first-guess input SST and upper
  air fields
• Assist in resolving SST anomalies

3. Platforms/Sensors/Bands Monitored in MICROS
Sensor Band (µm) Platform Time in MICROS Pixel Status
AVHRR Ch3b(3.7) Ch4(11.3) Ch5(12.0) NOAA-16 Jul 2008 pr 4km (GAC) Out of family, likely due to Cal issues
AVHRR Ch3b(3.7) Ch4(11.3) Ch5(12.0) NOAA-17 Jul 2008 Feb 2010 4km (GAC) Scan Motor Stalled in Feb 2010
AVHRR Ch3b(3.7) Ch4(11.3) Ch5(12.0) NOAA-18 Jul 2008 pr 4km (GAC) Somewhat unstable
AVHRR Ch3b(3.7) Ch4(11.3) Ch5(12.0) NOAA-19 Mar 2009 pr 4km (GAC) Stable
AVHRR Ch3b(3.7) Ch4(11.3) Ch5(12.0) Metop-A Jul 2008 pr 4km (GAC) Stable
AVHRR Ch3b(3.7) Ch4(11.3) Ch5(12.0) Metop-A Dec 2009 pr 1km (FRAC) Stable
MODIS Ch20(3.8) Ch31(11.0 Ch32(12.0) Terra Jan 2012 pr 1 km Stable
MODIS Ch20(3.8) Ch31(11.0 Ch32(12.0) Aqua Jan 2012 pr 1 km Stable
VIIRS M12(3.7) M15(10.8) M16(12.0) NPP Jan 2012 pr 0.76 km Stable

• Global M-O distributions are uniform
• Histograms are close to Gaussian
• View zenith angle and other dependencies are
  near-flat
• Warm M-O biases due to inaccuracies in M
  (Missing aerosols Using bulk Reynolds instead of
  skin SST Using daily mean to represent night SST
  in CRTM) and O (Residual cloud)
• Temporal M-O variability is mainly due to
  unstable Reynolds SST input to CRTM

7. AVHRR MODIS VIIRS Consistency

• Use Double Differences (DD) to check BTs and
  SSTs for Cross-Platform Consistency
• DDs cancel out/minimize effect of various
  systematic errors instabilities in M-O arising
  from e.g. Errors and Instabilities in Reynolds
  SST GFS Profiles Missing aerosol Possible
  systemic biases in CRTM Updates to ACSPO
  algorithm Residual Cloud

• Adding ERS/ENVISAT (A)ATSR, Metop-B AVHRR, MSG
  SEVIRI, and GOES-R ABI in MICROS is in
  preparatory stages
• Adding SST optional bands is underway
• MODIS Ch22(3.96 µm), Ch23(4.06 µm), Ch29(8.52
  µm)
• VIIRS M13(4.05 µm), M14(8.55 µm)

DDs _at_3.7 µm (RefMetop-A GAC)
4. Sensors Characteristics

• AVHRRs, Terra/MODIS, NPP/VIIRS are consistent to
  within 0.15K.
• Aqua MODIS out of family by 0.3K
• Cal Change 7 Mar 2012 reset BT_at_M12 by 0.13K
  well in family
• NOAA16 unstable and out of family calibration
  problems

Characteristic AVHRR FRAC MODIS VIIRS
Swath width, km 2,900 2,330 3,000
Pixel size _at_Nadir, km 1 1 0.76
of FOVs (Pixels) per scan line 2,048 1,354 3,200
of Detectors 1 10 16
of rotation per 5 min interval 1,800 203 167.4
of scan lines per 5 min interval 1,80011,800 203102,030 167.4162,679
File compression No Yes No
L1b file aggregation All bands Geo 1 file All bands 1file Geo 1file Each band 1file Geo 1file
of L1b files/Day 28 x half-orbits 576 x 5min 6000 x 86sec (5 bands1 geo)
DDs _at_11 µm (RefMetop-A GAC)

• All AVHRRs and NPP/VIIRS are consistent to within
  0.15K.
• Terra and Aqua/MODIS out of family by 0.6K, due
  to suboptimal CRTM coefficients in vV2.02.
• Cal Change 7 Mar 2012 BT_at_M15 by 0.14K Now
  better in family
• NOAA16 unstable and out of family calibration
  problems

5. System Set-Up, Data Volumes, Processing Time
DDs _at_12 µm (RefMetop-A GAC)
of clear-sky ocean pixels and Processing Time
for 24hrs of L1b data

• All AVHRRs and NPP/VIIRS are consistent to within
  0.15K.
• Terra and Aqua/MODIS out of family by 0.3K, due
  to suboptimal CRTM coefficients in v2.02.
• Cal Change 7 Mar 2012 BT_at_M15 0.14K Now better
  in family
• NOAA16 unstable and out of family calibration
  problems

Clear Ocean Pixels / Night Total Hours ACSPOMICROS
AVHRR GAC (N16, 18-19, Metop-A) 3M 4 3 1
AVHRR FRAC (Metop-A) 45M 6 4 2
MODIS (Terra Aqua ) 35M 6 4 2
VIIRS (NPP) 90M 18 10 8
DDs SST (RefMetop-A GAC)

• All AVHRRs, MODISs and NPP/VIIRS are consistent
  to within 0.15K.
• VIIRS Cal Change 7 Mar 2012 SST 0.15K Out of
  family.
• VIIRS SST went out of family 8 Mar
• New regression coefficients imple-mented 3 May
  2012 - back in family
• NOAA16 unstable and out of family, due to
  unstable BTs

• MICROS processes L1B data using the Advanced
  Clear-Sky Processor for Oceans, performs
  statistical analyses of BTs and SSTs publishes
  on the Web.
• MICROS is fully automated, with back-up
  processing enabled to improve stability

9. Conclusion
10. Future Work

• MICROS monitors M-O biases in clear-sky BTs and
  SSTs over oceans in near-real time. It is an
  end-to-end system, currently functional with 4
  AVHRR, 2 MODIS, and NPP/VIIRS
• All M-O biases are warm, due to a combined effect
  of missing aerosols using bulk SST (instead of
  skin) daily mean Reynolds SST (to represent
  night SST) residual cloud
• All MODIS BTs are biased warm by 0.3-0.6K (due to
  suboptimal CRTM coefficients)
• MODIS 3.7µm Terra/Aqua inconsistent by 0.3K,
  likely due to MODIS calibration problem
• VIIRS BTs are consistent with AVHRR, especially
  after VIIRS IR Cal update in Mar 2012
• VIIRS SSTs are again consistent with AVHRR, after
  new regression coefficients implemented on 3 May
  2012, in response to VIIRS Cal update on 7 Mar
  2012

• Improve accuracy of MICROS DDs Use more accurate
  first guess fields Improve ACSPO cloud mask
  Improve CRTM performance (especially daytime)
  Model diurnal variation (DV) in first-guess SST
  Improve Sensor Radiances (Calibration, Spectral
  Response Functions)
• Extend MICROS to Include Reflectance Bands
• Aerosol Quality Monitor (AQUAM,
  http//www.star.nesdis.noaa.gov/sod/sst/aquam/)
  was set up to prepare for adding aerosol in CRTM
• GOCART and NAAPS 3D aerosol fields have been
  selected as input to CRTM AQUAM
• Will use MICROS DDs to evaluate CRTM in VIS
  GOCART/NAAPS fields Sensors calibration
• Add MSG/SEVIRI, GOES-R ABI, ERS/ENVISAT (A)ATSR ,
  Metop-B AVHRR in MICROS.

Acknowledgement This work is conducted under the
Algorithm Working Group funded by GOES-R Program
Office, NPOESS Ocean Cal/Val funded by IPO, and
Polar PSDI, NDE and ORS Programs funded by NOAA.
ACSPO is provided by NESDIS SST Team CRTM is
provided by NESDIS JCSDA. Thanks to John Stroup,
Yury Kihai, Boris Petrenko, John Sapper, Mark
Liu, Yong Chen, Paul Van Delst, David Groff, Fred
Wu, Fangfang Yu, Fuzhong Weng, Changyong Cao, Tim
Hewison, and Nick Nalli for advice and help. The
views, opinions, and findings contained in this
report are those of the authors and should not be
construed as an official NOAA or U.S. Government
position, policy, or decision.
IEEE International Geoscience and Remote Sensing
Symposium Munich, Germany, 22-27 July 2012
Correspondence Alex.Ignatov_at_noaa.gov, Tel
301-763-8102 x190, Fax [phone]