GHRSST-PP L2P Near Atlantic Regional (NAR) Sea Surface Temperature (SST-sub-skin) for 2005-11-24 (NCEI Accession 0002856)

The EUMETSAT OSI-SAF NAR SST products are SST fields derived from NOAA/AVHRR data and available over 6 pre-defined zones, 4 times per day. After acquisition and calibration at CMS, the NOAA/AVHRR data are re-mapped onto a stereopolar grid to produce mosaics of consecutive orbits over Europe and the Near Atlantic at a 2 km resolution. Mosaics are made over a 'Rectification Grand Domaine' area (RGD). For re-mapped grid points observed by more than one overpass, the smallest satellite zenith angle data are selected. In the normal conditions when two satellites are operational, four mosaics per day are produced for all the AVHRR channels. SSTs are then calculated from these data. Six zones are then extracted . They are 1024 * 1024 pixels in size to offer the users (presumably interested by local phenomena) the possibility of handling a restricted volume of data. Note that SSTs are calculated over the entire RGD, but only the predefined zones are delivered as NAR products. The operational CMS AVHRR cloud mask in use is based on a multi-spectral thresholding algorithm (Derrien and Le Gleau 1999). Some refinements specific to the marine conditions have been introduced including Use of fine scale climatology and a fine gradiant climatology to assist in the detection of clouds in areas characterised by strong thermal gradients. SSTs are derived from the 11 and 12 micron brightness temperatures (T11 and T12) . The NOAA /AVHRR has three IR channels : channel 3 (3.6-3.8 micron), channel 4 (10.2-11.2 micron) and channel 5 (11.5-12.5 micron) which provide IR data at 1-km spatial resolution at the satellite subpoint (Kidwell, 1997). NOAA-14 and NOAA-16 have similar characteristics, but with distinct radiometer filter functions that necessitate distinct algorithm coefficients. The algorithms coefficients are derived from multilinear regression on a simulation database . At night the use of the 3.7 channel in a triple window algorithm improves the SST retrieval performances under high water vapor content conditions. The expected accuracy is defined as the expected bias and standard deviation of the primary calculations against drifting buoy measurements determined on a monthly or yearly basis. The following performance evaluation was derived from a detailed validation of NOAA/AVHRR derived SSTs (Brisson et al 2001). Bias: Apart for the first 2 months after the launch of a new satellite, the monthly biases should remain within +/-0.4 K by night. By day local biases of several K are possible because of shallow diurnal heating. Standard deviation: see table below: low-cloud cases (cloud coverage less than 10% of the validation box) have been separated from all cases (cloud coverage less than 60%). The 'expected maximum standard deviation' or the expected maximum biases are the maximum values that can be observed under normal conditions. When these values are overpassed, some anomaly may have occurred and the problem must be identified. Standard deviation Table Yearly, all Monthly, all Yearly, low-cloud Monthly, lowcloud 0.7 0.8 0.6 0.7 Physical definition subskin SST : Comparable to in situ (buoy) measurements at night. Relation to bulk SST: equivalent to bulk SST by night. By day, a bias of several Kelvin may be found under favorable diurnal heating conditions, Relation to skin SST: By day and by night the subskin SST is convertible to skin temperature by subtracting 0.2K. Units and range Units: Centi-Kelvins When using GRIB format, a scaling factor of 100 must be applied. No lower limit is imposed on the SST calculations, providing the sea is ice free. The expected range is thus from about -3 degC till 35 degC, corresponding to actual values of 27000 to 30800. Unprocessed data (for whatever reason) show a negative value (-32768) in HDF format, or is flagged as missing in GRIB section3. Origin SST calculated from the IR channel data of the NOAA/AVHRR re-mapped onto a stereopolar grid at 2 km resolution. SST are extracted over 6 pre-defined zones Geographical definition Projection: Polar stereographic projection defined with an elliptical earth (equatorial radius: 6378.388 km; polar radius: 6356.912 km), y axis is meridian 0. Resolution: 2 km at 45 N Size of each zone: 1024 columns, 1024 lines L2P data products are then derived following the GHRSST-PP Data Processing Specification (GDS) version 1.5 avaialable from http://ghrsst-pp.metoffice.com/documents/GDS-v1.0-rev1.5.pdf. Single Sensor Error Statistics (SSES) are provided based on the GDS-v1.5 specification.