Standard Deviation of Mean Monthly Temperatures in Anglo-America

INVESTIGATION of the variability of temperature in a particular month from one year to another may produce results surprising to the nonclimatologist, and perhaps even to the specialist. At Havre, Mont., for example, the mean January temperature over a 5o-year period is I30 F. But there has been a January (I9I6) with a mean of -I3? F., 260 colder than the 5o-year average, and there has also been a January (I9I9) with a mean temperature of 340 F., 21 warmer than the average. In the first instance, Havre had about the average January mean of Fairbanks, Alaska (i0 F.), in the latter instance, about that of Amarillo, Tex. (350 F.)-two sharply contrasting localities climatically! Let us imagine ourselves inspecting a map ofJanuary mean temperature for North America. We see the I40 isotherm passing a few miles south of Havre and following a particular contour along the mountains. Perhaps because of the exactness with which this isotherm is drawn, we tend to regard it as a fixed line whose position has been established almost with the accuracy of an actual contour; yet, as we have seen, during so years it has in effect oscillated between a point in the Yukon Valley north of Fairbanks and one in the Texas Panhandle. The significance of variability may be even more striking in areas of potential strategic importance. At Dawson, Yukon Territory, for instance, the 45-year mean temperature for December is -I40 F. However, there has occurred in Dawson a December (I9I7) in which the mean temperature was -5i0 F., 370 colder than the average, and a]so a December (I928) in which the mean was 70 F., 2I0 warmer than the average. In other words, in less than half a century Dawson has experienced two Decembers that differed from each other in mean temperature by 580, which is numerically equal to the difference between the normal January and the normal July at Ottawa. By way of comparison, and in order to give greater significance to the values of deviation by relating them to popular experience, it may be noted that, in central New England, January, I948, commonly referred to as an unusually cold month, actually showed a departure of only 60 from the mean, and October, I947, regarded as an exceptionally warm autumn month, also departed by only 60 from the mean. Broadly speaking, it seems