Guest essay by Samuel I Outcalt
The USCRN monthly average air temperature records were searched for continuous records covering the 2004 to 2014 decade over the contiguous United States. The average values were selected as these averages represented the average of several sensors. The records represent mid-month averages so a simple interpolation algorithm was used to convert the x-axis of plots from continuous months with an initial index value of zero to fractional years [ x = 2004+(1/24) + (x/12) ]. As mid month data was used, the record stream ran from 1 January 2004 to 1 January 2015. The graph for Elgin, Arizona is displayed as Figure 1.
It was later discovered that Elgin was the only record that could NOT be easily included in within an area of stations having a negative trend during the decade.
The map of the station location slopes in temperature change during the decade is presented as Figure 2.
Figure 2. A map of the average temperature (C) change during the decade. In some cases the station labels were shifted slightly to avoid state border lines and label overlap.
It is apparent that the region between the Appalachian and Rocky Mountains is cooling compared to warning regions along the northern sections of both the Atlantic and Pacific Coasts. It should be noted that in some states with paired station records there is a significant difference in the magnitude of the decade trend. These differences indicate significant site effects especially in the flat agricultural regions in along the Gulf and Atlantic Coastal Regions. Some of these variations may have been introduced by downwind crop rotation and increased aerodynamic roughness produced by the rather massive structures use to shield precipitation gauges.
The absolute magnitude of larger decade trends extrapolated to degrees [ C ] / Century far exceeds the magnitude of the approximate 3 to 5 C global warming since 1870. The warming of 3 C / Century was first documented from an analysis of the geothermal profile at Barrow, Alaska. That analysis, which was completed and reported by the Alaskan Branch of the USGS in 1970. The well documented post 1870 warming is considered by some to mark the end of the Little Ice Age and others to be an artifact of the onset of the industrial revolution. If the present decade trends were to extend for a century, the 2104 air temperature contrast between Arco, Idaho and Wolf Point, Montana would be near 20 C.
A precision spatial interpolation of the data due to the lack of records in many areas. Even with a dense network the geographic interpolation algorithm would have to be conditioned by topography. In order to drive home the pattern of the decade trend contrasts a hand drawn border was used to draft Figure 3.
Figure 3. The area with cooling temperatures is shaded light blue.
In Figure 3 the location of Elgin. AZ is indicated (*) but not labeled. The the trend sign boundary would have become too irregular to include it in the cooling area. However, the contrast between it’s value of -0.25 and the nearby value of +0.16 indicate that the trend values in complex terrain are not only conditioned by local site effects by the local topography.
This brief analysis indicates that the areas of warming and cooling display a strong degree of spatial correlation.
However, there are strong indications that the expansive addition of more USCRN stations will produce more complex contrasts by the end of the next decade.
The pronounced mid continent cooling appears to be the product of the increased frequency and/or lowered temperatures of winter Arctic Air invasions. The pronounced warming New England and the Pacific Coast, which extends over most of the Southwest may be traceable of ocean circulation changes.
This short exploration indicates that simple linear regression applied to all the variables supplied by the USCRN data will provide a robust image of the Geography of Modern Climate Change, when applied to the present data over shorter time spans. The calculation requirements are massive when the large number of stations added during this decade is considered. However, this task as well as the design of topography conditioned extrapolation algorithms is well within the capacity of the computers at the National Center for Atmospheric Research, now with massive computers at both Boulder and Cheyenne.