Yeahp, I've been geeting similar results. If you "quality control" the area (of this hot-spot), you'll find a bouy in that region giving very high readings. I delete the temperature on that bouy and re do the analysis.
Yeah, well, I've noticed that for some problematic buoys off the US east coast, you can't delete the station off the station list. The buoys are plotted as lat/long, not off the list. One would have to find and edit out the report. You have to use a good text editor to avoid losing LINE FEEDS.