You will hear it said that February 2010 was the second-warmest February on record, which indeed it may be. (Whether it is the second-warmest based on raw data or the second-warmest based on adjusted data, I do not know.) We are, however, in an El Nino, and "the Kid" typically drives dramatic spikes in the temp record, and are subsequently [and properly] discounted in assessment of trends. "Weather," as they say, "is not climate."
But what is interesting about the warmth of this past February is that it applied only to Canada and Greenland.
This is interesting for two reasons.
a) It is the local anomalies, not the global mean, that drive the location, development, and movement of weather systems [and hence climate events like droughts, floods, etc. So the concentration of the Second Hottest February in Canada and Greenland, if true, is interesting.
b) The Northands have famously had a great many temperature stations dropped in recent years. Which stations have been dropped and why, I do not know; but their lack of inclusions means the coverage of the North is much sketchier than, say the continental US or Europe. In some cases, to run the homogenization* of the data, stations have had to have their values projected more than a thousand kilometers from the actual location of the thermometer. If this is so, a single faulty instrument used to "represent" a wide region can have a disproportionate impact on the appearance of the map and on the global mean, if the global mean is weighted by area rather than actual number of stations.
In any case, if I were to see a concentration chart like this on, say, the locations of delamination on paperboard (at a paper packaging plant in Chennai) or the locations of lost-time accidents (at a wire and cable plant in the bootheel of Missouri), I would immediately suspect a local cause - insufficient glue application at the left edge of the laminator; a missing guard rail on a ramp into the warehouse.
Of course, the first thing to check before looking for geographic factors is the quality of the instrumentation itself, its calibration, maintenance, transcription, etc. Are the high values due to the influence of particular stations used to homogenize the values from other stations? Were there any nearby stations against which values could be compared. (If I had may way, all stations would be triplicated and in strict adherence to NOAA standards, with periodic calibration checks and quality audits. That way if a station goes out of whack, you can use the two-out-of-three rule. (Many TAPPI test standards, for example, call for triplicate testing of specimens.) Many's the time I saved myself some work by checking the measurement system first.
The Plot is Afoot
Here is a plot for the US of US under severe to extreme dry conditions.
The 60-month rolling average is to get rid of the weather so that the climate stands out. That is a five-year roll. You can see the trend, of course. Or not, since there does not appear to be any. But the point I wanted to make is technical. The rolling average should be plotted at the center of the period being rolled. The plot here is at the end of the five-year period. Therefore, you should mentally shift the red line to the left by two and a half years, which would give a much more satisfactory picture.
The same is true of extremely wet weather:
On the Other Hand
The annual amount of Cyclonic Energy must be taken with a grain of salt, since our ability to detect and measure cyclones has improved remarkably. The chart should be labeled "...Hurricane or Cyclonic Activity That We Knew About." Notice that since the sun has become less active, the total cyclonic energy has been dropping. As I understand it, cyclones depend on the delta between air and water temperature, but someone may know better than I.