Transcript GOES-SST RDAC - Bureau of Meteorology

Geostationary-Sea Surface Temperature (SST)
NOAA/NESDIS
Regional Data Assembly Center
Eileen Maturi Andy Harris
Jonathan Mittaz
The Eighth Science Team Meeting of the Global Ocean Data Assimilation Experiment High Resolution Sea Surface Temperature Pilot Project (GHRSST-PP)
Bureau of Meteorology Head Office, 700 Collins St, Docklands, Melbourne, Australia -Monday 14 May to Friday 18 May, 2007
GOES-SST L2P FORMAT
• Level 2P (L2P – “P” stands for preprocessed) for
the operational GOES-SST.
• The L2P format requires the appending of a
number of ancillary data records to the standard
SST value at each pixel
– latitude, longitude, time, cloud proximity, aerosol optical
depth, wind speed, surface solar irradiance, sea ice
fraction, single-sensor error statistics (SSES)
Single Center Error Statistics (SSES) Estimation
Using SSES to provide additional information that cannot be calculated from
L2P ancillary data
Assumption is that retrieval bias depends on

Clear-sky transmittance (calculated from NCEP → CRTM)
 Air-sea temperature difference (currently NCEP only)
Assume that sensitivity to Air-Sea Temperature Difference (ASTD)
increases with decreasing transmittance
 Derive bias = offset + gradient ASTD for different 11
Post-corrected σ is estimated as a function of transmittance only
Different for sensor/day/night
NOAA GOES-SST L2P
•
•
•
•
•
•
•
•
•
•
The full L2P data set has been implemented.
Products are derived from ½-hourly GOES-East & West North & South sectors in
native satellite projection.
Dataset includes the full L2P ancillary field.
Aerosol field is derived from the NESDIS operational daily aerosol optical depth
analysis, and values are sampled to each GOES retrieval on a nearest-neighbor
basis.
The “age of observation” value that is available with the analysis is also included –
in this case the range may extend to several days rather than the 24 hours.
Meteorological fields are derived from the NCEP Global Forecast System forecast
fields and bi-linearly interpolated to the SST retrieval location.
Surface solar irradiance is derived from the 3-hr average for the period in which the
SST observation was made.
Wind speeds are interpolated in time
“proximity confidence” is actually the Bayesian probability of clear-sky, with a
range from 0.95 to 1 in steps of 0.0002.
Product generates 24 Gb /day uncompressed~8Gb/day compressed
GOES L2P Product
 A GOES L2P netCDF product file has 22 parameters for each pixel:
– SST
– Time, Latitude, Longitude
– Satellite Zenith Angle
– Aerosol Optical Depth*
– Surface Solar Irradiance*
– Wind Speed*
– Uncertainty estimates (bias and S.D.)
– Proximity Confidence Value
– QC flags (including cloud and land)
– Ice concentration*
– Deviation from analysis SST
– Temporal coincidences of ancillary data* c.f. SST observation
– Source codes for ancillary data*
– Probability of clear-sky (optional field)
 Example fields shown on next slide
GOES-12 Level 2P N. Hem Mar 28 15:15
Proximity
Surface
Aerosol
SSES
Wind
Confidence
Bias
Solar
S.D.
Optical
SST
speed
Estimate
Irradiance
Depth
(Pclear)
Future Work
• Correct for Radiance Bias
– Improve the GOES calibration
– Dynamic cloudy priors
• Include cloud probability in the SSES
estimate?
• Moving Toward a Full Physical Retrieval
Methodology
FUTURE L2P Products
• MTSAT –SST L2P
• GOES-SST-13 and beyond