Zur Hauptnavigation Zum Inhaltsbereich Zur Suche Zum Seitenfuß

ICDC Sitemap  Deutsch   English

Sie sind hier:UHH   MIN-Fakultät   CEN   ICDC Data Center    Data    Ocean    Sea surface temperature (Reynolds-SST) 

Integrated Climate Data Center - ICDC


We offer here Version 2 of the well known Reynolds SST data set. It comprises two new high resolution sea surface temperature (SST) analysis products that have been developed using optimum interpolation (OI) --> OISST data set. One product uses Advanced Very High Resolution Radiometer (AVHRR) infrared satellite SST data (Pathfinder data: September 1981 through December 2005; operational AVHRR: January 2006 onwards). The other uses AVHRR and Advanced Microwave Scanning Radiometer (AMSR) on the NASA Earth Observing System satellite SST data (June 2002 onwards). Both products also use in situ data from ships and buoys and include a large-scale adjustment of satellite biases with respect to the in situ data.

ICDC offers the data set that is based on AVHRR data because this is the longest time series available (since 1981).

In addition ICDC has computed the monthly mean SST for 1982-2016 and offers these data.

More information is given at


in the Technical Note by Reynolds, and further down under data quality.

See also the publications below (see references).

Last data update at ICDC: August 25, 2017.

On top of this ICDC has computed anomalies of the monthly mean SST relative to two reference periods: 1982-2001 and 1992-2011 and also offers these anomalies for 1982-2016.

NEW are maps of the SST trend for each month for the period 1982-2016. One example of these maps are displayed in the top right; the others can be clicked below:


New as well are graphs of the SST trend computed from 13-month moving average SST of 10° x 10° boxes. For each of these boxes shown in the sample trend figure at the top right the linear trend, its significance (p-value) and the mean SST are computed. Trend values without a p-value are not significant. Down below we show three sample graphs for 10°W to 20°W for, from left to right, 50°N-60°N, 40°N-50°N, 30°N-40°N. All the Graphs are available on request from stefan.kern(at)uni-hamburg.de or under /data/icdc/ocean/reynolds_sst/DATA/TRENDS/.

To top

Time-series, moving 13-month mean and corresponding trend in the global monthly mean SST. Sea-ice covered areas were discarded using a maximum sea-ice extent mask.

To top


Name Unit Comment
Daily data NCDC Product
Retrieval uncertainty K
Land mask -- 1: ocean, 2: land
Ice mask -- 0 ... 1
Monthly data CliSAP/KlimaCampus Product
SST K includes flag minimum ice extent
Mean retrieval uncertainty K
SST standard deviation K
Flag minimum ice extent -- 0: ice, 1: ocean
Flag maximum ice extent -- 0: ice, 1: ocean
Number of days used per month --
Anomaly CliSAP/KlimaCampus product
SST anomaly K Anomaly of the monthly SST with respect to
mean monthly SST of periods 1982-2001 & 1992-2011

To top

Coverage, temporal and spatial resolution

Period and temporal resolution:

  • Daily: 09/1981 to 07/2017
  • Monthly: 01/1982 to 03/2017 (means and anomalies)

Missing days:

Feb. 28, Mar. 19 and 22, May 10, June 10, Aug. 19, Nov. 2, and Dec. 12 2016

Feb. 4 & 25, Mar. 15, May 18-20, May 22 - June 5 2017

Coverage and spatial resolution:

  • Global
  • Spatial resolution: 0.25° x 0.25°, cartesian grid
  • Geographic latitude: -89.875°N to 89.875°N
  • Geographic longitude: 0.125°E to 359.875°E
  • Dimension: 1440 columns x 720 rows
  • Altitude: 0.0 m


  • NetCDF

To top

Data quality

The offered daily data include an error estimation for each grid cell; see Reynolds et al. (2002; 2007) as well as the recent Technical Note for details.

In addition, the daily data set includes the NCEP/NCAR Reanalysis sea ice concentration.

During the computation of the monthly mean SST ICDC found a number of inconsistencies:

  • The daily SST data contain a jump in the land mask used from 2004-12-31 to 2005-01-01. Apparently a different land mask was used starting from 2005.
  • In Antarctica the land mask used seems to be incomplete and does not extend all the way to the water - for instance in the southwestern Weddell Sea. This creates artificial polynyas (= areas of open water in regions where according to the climatological conditions there should be sea ice) along the coast at places where polynyas are either completely absent or where actually existing polynyas are much smaller in size. ICDC mitigated these artifacts by using a different land mask (actually the one from the HadISST product). The artifacts could not be removed completely, though. Hence it is not recommended to interpret too much into SST along the Antarctic coast line.
  • SST along the sea ice cover is treated as follows: Generally the monthly SST for a pixel is computed once the sea ice cover was < 50% on at least 6 days of the respective month. Pixels with 5 or less such days are flagged. The corresponding flag mask is included in the data set as "minimal ice extent".  If the user would like to be more certain in excluding sea ice influence the second flag mask "maximum ice extent" can be used. This mask flags only those pixels as open water (=1) where on all days of the respective month the sea ice cover was < 50%. Note: Usage of a smaller threshold than 50% would have been desirable but yielded artificial ice cover along the coasts of, e.g. U.K. or Norway, and would hence have caused to flag SST in regions where we can be 100% sure that there is no sea ice. Note further that we relied only on the sea ice information provided along with the daily SST product.
  • Data gaps: Sea ice cover data are missing in December 1987 and January 1988 in both hemispheres and are partly inconsistent in May 2009 in the northern hemisphere. For 1987 and 1988 we used the ice flag masks of the previous and the next months; for 2009 we build our ice flag masks on just 16 days instead of 31 days
  • Our monthly SST product contains the mean monthly SST retrieval uncertainty as simple arithmetic mean over the daily SST retrieval uncertainty values. Any spatialtemporal correlations that might exist were not taken into account. Our monthly SST product contains in addition the temporal variability of the SST per pixel as one standard deviation of the mean SST for the respective month.

To top


Stefan Kern
ICDC / CliSAP / University of Hamburg
email: stefan.kern (at) uni-hamburg.de

Richard W. Reynolds
NOAA/NESDIS/National Climatic Data Center
151 Patton Ave
Asheville, NC 28801
email: Richard.W.Reynolds (at) noaa.gov

Chunying Liu
NOAA/NESDIS/National Climatic Data Center
151 Patton Ave
Asheville, NC 28801
email: Chunying.Liu (at) noaa.gov

To top


Data citation

Please cite Reynolds et al., J. Climate, 20, 2007, when using the data; see also the References above.

Plus: NOAA Optimum Interpolation (Reynolds) Sea Surface Temperatures (https://www.ncdc.noaa.gov/oisst) have been obtained in netCDF file format by the Integrated Climate Data Center (ICDC, icdc.cen.uni-hamburg.de), University of Hamburg, Hamburg, Germany.

To top

Print Print page | Last update  Thursday, 24.08.2017 
| Legal notice | Privacy statement | Top Top of page