SST ANOMALY MAPS - Computational Details

Climatology : The anomalies are calculated using the climatology provided in the Ocean Atlas Cds, based on the UWM/COADS analysis.

Data Coverage : Due to a change in how ERD receives data from FNMOC, and changes in the processing and formatting of this data, we believe we now receive more complete data on a more timely basis. However, FNMOC used to do some preliminary quality control before ERD received the data, which is no longer being done.

Method for Discarding Bad Values : The Ocean Atlas data include the standard deviation, by month, for each 1-degree monthly mean SST. For now, the new method flags any value that is 3 or more standard deviations from the observed mean of the data in that 1-degree box and month, where the standard deviation is the one given in the climatology. The old method was based purely on the climatologies. The previous method would have thrown out as unreasonable the very high values that have been recorded in the present El Niño, and thus provide misleading charts. The new method is based on the idea that if the data are consistent and there is enough data, then they are probably correct. The climatological standard deviation is taken as a measure of how much variance we can expect in the data while the observed mean is taken as an indicator of the present level, so we are discarding inconsistent data, but not data that may be well above the climatological mean. The problem with this method is when there are fewer than 5 observations, particularly locations with one or two observations, all the values may be unreasonable. This has definitely caused problems in some areas, and values for areas with very few observations should be used with caution. We are working on a solution for this problem. Also, if a source of many observations, such as a buoy, should become uncalibrated, the resulting mean value will be misleading.