These SST datasets are prepared at NCEP and NCDC by Dick Reynolds, Diane Stokes, and Tom Smith.

This dataset has been replaced by the Smith and Reynolds Extended Reconstructed SST Dataset.

Monthly 2x2 global analyses for 1950-1999 are available from NCAR's DSS in a collection termed "Reconstructed Historical Monthly Analyses". EOF methods are used to create these grids using marine surface data for 45S-69N with climatology fill in areas north and south of these limits.

NOTE! Because the satellite data makes the OI (see link below) superior to any in situ analysis and because the reconstructed fields are only defined over a limited area, the OI SST is recommended instead of the reconstructed fields from 1982 onward.

(The following is taken from NCAR's DSS description of the data.)
A new interpolation method was developed using empirical orthogonal functions (EOFs) to produce in situ analyses of SSTs from 1950-81. The first step was to produce the spatial EOFs from the optimum interpolation (OI) analyses of Reynolds and Smith (1994). The dominant EOF modes were used as basis functions and were fitted to the in situ data for the period 1950-92 to determine the time dependence of each mode. A global field of SSTs was then reconstructed from these spatial and temporal modes. For details of the EOF reconstruction method see Reynolds et al. (1995) and Smith et al. (1995).

All fields were computed on the COADS 2-degree grid. The grid has 180 zonal points on 1E, 3E, ..., 359E, and 90 meridional points on 89S, 87S, ..., 89N. The climatology (see Reynolds and Smith, 1995) was averaged onto the COADS grid. For convenience, the land values were filled by a Cressman technique. The ocean and land areas are defined by a land sea mask described below. All reconstructed fields were computed as anomalies. However, because of limited in situ data, the reconstructed anomalies were only computed from 45S to 69N with some missing values in the North Atlantic and North Pacific (see Fig. 2 in Smith et al., 1995). To produce the mean presented here, the anomaly was added to the climatology after multiplying the anomaly by weights. The weights were 1 over defined ocean areas, otherwise 0. To provide a smoother transition between defined and undefined open ocean areas, the weights along these open ocean boundaries were set to 0.5 (e.g. at 45S). In addition any SST mean value less than -1.8C was set to -1.8C.

Smith, T. M., R. W. Reynolds, R. E. Livezey, and D. C. Stokes, 1996: Reconstruction of historical sea surface temperatures using empirical orthogonal functions. J. Climate, 9, 1403-1420.

T. M. Smith and Reynolds, R. W., 1998: A high-resolution global sea surface temperature climatology for the 1961-90 base period. J. Climate, 11, 3320-3323.

Click on the links below to view data coverage maps for a particular time period. Percentage of non-missing data per time period is plotted. Coverage is consistent throughout the period of record.

(1950-2000)

Updated: 1/10/03
Maintained by asphilli@ucar.edu