UPDATE February 15, 2013: Reinstalled gif animation that had been replaced with another image at TinyPic.
UPDATE January 15, 2010: Warwick Hughes suggested that I check the other datasets. They are included at the end of the post.
While I was preparing the maps included in the Animation Of Weekly Global SST Anomaly Maps – November 1, 1981 To January 6, 2010, I noticed an unusual “wave” of cool SST anomalies that appeared to circle the globe in the mid latitudes of the Southern Hemisphere, Figure 1.
It forms about October 30, 1991, near 40S, wanders 10 degrees North toward the equator until mid December, then turns around and heads poleward. It finally disappears about the end of January 1992. Refer to the .gif animation, Figure 2.
I could not find the effect again in the animation. The El Nino that season wasn’t a major event, but Mount Pinatubo erupted about five months earlier, in early June 1991. Is it an aftereffect of Mount Pinatubo? I have no other explanation, but I also haven’t seen this effect discussed in any papers.
Here’s the video again for those wanting to confirm that the “wave” doesn’t appear at any other time. If you do notice it again, please let me know.
User-defined OI.v2 SST Anomaly maps can be created at the “Full Version” of the NOAA NOMADS webpage here:
UPDATE January 15, 2010: Warwick Hughes noted in a comment:
Bob, it looks like an artifact of the datset construction to me. My inclination would be to check against other SST datsets. I do not want to suggest more work for you.
I’ve included my response in the following so that I can display the maps of the other datasets. Otherwise they’d appear as tinypic addresses that you’d have to cut and paste to your browser.
Warwick: Thanks for the thought. I’ve used KNMI Climate Explorer to create the maps of the other datasets. The other datasets are presented only in monthly format, so as a spotcheck of the second source, I’ve also plotted the OI.v2 data. The monthly OI.v2 shows the line of negative anomalies in November 1991:
OI.v2 data for November 1991
HADISST (satellite based like the OI.v2 data) shows a similar pattern but not as defined, and it only appears in December:
HADISST data for December 1991
The HADSST2 data in that part (your part) of the world is fragmented, as you know, but it seems to contradict the OI.v2 data:
HADSST2 data for December 1991
Since the ERSST.v2 and ERSST.v3b data are also based on the incomplete ICOADS data and are infilled by the NCDC, I wouldn’t expect the pattern to show. Here’s the ERSST.v3b plot for November 1991 anyway:
ERSST.v3b data for Nov 1991