SST Monthly Means and Anomalies

The plots in this data group are of monthly means in 1-degree lat/lon squares of sea surface temperature and anomalies. They were computed from GTS surface observations received twice a day by PFEL from FNMOC (Fleet Numerical Meteorology and Oceanography Center). These are global real-time surface observations from platforms such as ships and buoys (both fixed and drifting). They are a subset of the data which will be included in an update to COADS (Comprehensive Ocean-Atmosphere Data Set, also see COADS on CD-ROM) .

Monthly means and statistics are calculated and posted to this site a few days after the first of every month. The variables are as follows:

1. Raw 1 degree SST Monthly Mean - monthly mean computed from all observations received during the month. The actual number of observations averaged in each 1 degree square is plotted in #3.
2. Trimmed 1-degree SST Monthly Mean - monthly mean after discarding out of range values as described below (#8 and #9). The number of observations averaged in each 1 degree square is plotted in #4.
3. Raw Number of SST Observations - map of the total number of observations received during the month
4. Trimmed Number of SST Observations - number of observations after discarding out of range values as described below
5. Climatological mean SST - climatological monthly mean from the Atlas of Surface Marine Data 1994 (file SST_CLM.NC)
6. Climatological mean SST St Dev - climatological standard deviation for each month from the Atlas of Surface Marine Data 1994 (file SST_STD.NC)
7. Monthly SST anomaly - deviation of the trimmed 1-degree SST monthly mean (#2) from the climatological mean SST (#5)
8. No. of Obs. Trimmed based on climatology - number of sst values in each 1 degree box for each month which deviate from the climatological mean by more than three standard deviations
9. No. of Obs. Trimmed based on Observations - The number of sst values which deviate from the data mean in each 1 degree box for each month by more than three standard deviations. This is only checked for values that are already flagged as out of range of the climatology. Only values flagged as out of range of both the data mean AND the climatological mean are discarded ("trimmed") to get the mean plotted in #2. This method in an attempt to avoid discarding as unreasonable the extreme values that have been recorded in, for example the most recent El Niño.

Please note that these are observational data. No analysis has been performed on the data other than the averaging and trimming described above. Users needing more complete coverage may use one of the many interpolated products available (for example see the Climate Modeling Branch SST Analyses page ).

Technical note: The GTS data received from FNMOC is in the extremely cumbersome BUFR format. The software provided to read the data is large, complicated, and slow to run. We convert the data to HDF-EOS structure which allows quick extracts of the data and greatly reduces file sizes. The raw data is then read from the HDF-EOS file, means and anomalies are calculated, and the results are stored in a NetCDF file as required by the visualization program FERRET which is used by the Live Access Server.