252
H.M. Beggs
On a pixel-by-pixel basis, each SST retrieval is augmented with estimates of the
bias error and standard deviation error typically derived from statistical databases
of in-situ and satellite data. Other ancillary fields for each pixel are surface wind
speed, aerosol optical depth, sea ice concentration, time of measurement, difference
from previous reference SST analysis field, and quality control flags, including a
cloud proximity confidence flag. The standard deviation errors associated with the
highest quality SST observations (denoted by a proximity_confidence flag of 5) are
typically in the range 0.3–0.5 ◦ C for the various satellite infrared SST GHRSST-L2P
products currently available (Table 15.1).
High Resolution Picture Transmission (HRPT) AVHRR data has been received
from NOAA’s operational polar orbiting satellites at local L-band receiving stations
by various operational meteorological and oceanographic agencies since 1981. This
data is used as a real-time source of 1.1 km resolution SST into operational systems.
Figure 15.1a–d show plots of the new Integrated Marine Observing System 3
(IMOS) 1 km resolution, HRPT AVHRR L2P SST for various values of cloud
proximity flag, defined in this product as a function of distance to cloud, satellite
Zenith angle and day/night (Beggs et al., 2009c). Figure 15.1e–h present values of
the derived SST standard deviation calculated for each cloud proximity flag value
(Beggs et al., 2009c). These figures illustrate one of the particular strengths of the
GHRSST L2P format for level 2 satellite SST data that users of L2P products
can trade off between spatial coverage and accuracy, depending on the requirements of their application. Less spatial coverage in this IMOS AVHRR L2P product
(Fig. 15.1) corresponds to less chance of cloud contamination and lower estimated
errors.
The design of the L2P format was mainly driven by the requirements of the new
“foundation” SST analyses for the “best” cloud-free SST with sufficient auxiliary
data to enable conversion of SST observations within or immediately below the
cool skin to foundation SST. By definition, the “foundation” SST (SSTfnd) product
provides an SST that is free of any diurnal variations (daytime warming or nocturnal
cooling) and is considered equivalent to the “subskin” ocean temperature (at the
base of the conductive laminar sub-layer of the surface ocean) in the absence of any
diurnal signal (Donlon et al., 2007). The L2P format has proven to be sufficiently
universal that GHRSST L2P files from a range of IR and MW sensors (AATSR,
MODIS, AVHRR, AMSR-E, TMI, SEVIRI and GOES Imagers) are now used in
some operational systems for ocean forecasting and seasonal prediction.
By accessing L2P products in netCDF format via OPeNDAP 4 servers both
in real-time 5 and delayed mode 6 it is possible to download only the variables
or domain that the application requires, therefore saving both on communication
costs and storage requirements. The consistent formats for GHRSST products also
facilitate addition of new data streams into operational systems.
3 http://www.imos.org.au
4 http://www.opendap.org
5 http://dods.jpl.nasa.gov
6 http://data.nodc.noaa.gov/ghrsst
H.M. Beggs
On a pixel-by-pixel basis, each SST retrieval is augmented with estimates of the
bias error and standard deviation error typically derived from statistical databases
of in-situ and satellite data. Other ancillary fields for each pixel are surface wind
speed, aerosol optical depth, sea ice concentration, time of measurement, difference
from previous reference SST analysis field, and quality control flags, including a
cloud proximity confidence flag. The standard deviation errors associated with the
highest quality SST observations (denoted by a proximity_confidence flag of 5) are
typically in the range 0.3–0.5 ◦ C for the various satellite infrared SST GHRSST-L2P
products currently available (Table 15.1).
High Resolution Picture Transmission (HRPT) AVHRR data has been received
from NOAA’s operational polar orbiting satellites at local L-band receiving stations
by various operational meteorological and oceanographic agencies since 1981. This
data is used as a real-time source of 1.1 km resolution SST into operational systems.
Figure 15.1a–d show plots of the new Integrated Marine Observing System 3
(IMOS) 1 km resolution, HRPT AVHRR L2P SST for various values of cloud
proximity flag, defined in this product as a function of distance to cloud, satellite
Zenith angle and day/night (Beggs et al., 2009c). Figure 15.1e–h present values of
the derived SST standard deviation calculated for each cloud proximity flag value
(Beggs et al., 2009c). These figures illustrate one of the particular strengths of the
GHRSST L2P format for level 2 satellite SST data that users of L2P products
can trade off between spatial coverage and accuracy, depending on the requirements of their application. Less spatial coverage in this IMOS AVHRR L2P product
(Fig. 15.1) corresponds to less chance of cloud contamination and lower estimated
errors.
The design of the L2P format was mainly driven by the requirements of the new
“foundation” SST analyses for the “best” cloud-free SST with sufficient auxiliary
data to enable conversion of SST observations within or immediately below the
cool skin to foundation SST. By definition, the “foundation” SST (SSTfnd) product
provides an SST that is free of any diurnal variations (daytime warming or nocturnal
cooling) and is considered equivalent to the “subskin” ocean temperature (at the
base of the conductive laminar sub-layer of the surface ocean) in the absence of any
diurnal signal (Donlon et al., 2007). The L2P format has proven to be sufficiently
universal that GHRSST L2P files from a range of IR and MW sensors (AATSR,
MODIS, AVHRR, AMSR-E, TMI, SEVIRI and GOES Imagers) are now used in
some operational systems for ocean forecasting and seasonal prediction.
By accessing L2P products in netCDF format via OPeNDAP 4 servers both
in real-time 5 and delayed mode 6 it is possible to download only the variables
or domain that the application requires, therefore saving both on communication
costs and storage requirements. The consistent formats for GHRSST products also
facilitate addition of new data streams into operational systems.
3 http://www.imos.org.au
4 http://www.opendap.org
5 http://dods.jpl.nasa.gov
6 http://data.nodc.noaa.gov/ghrsst