West-Central Idaho USHCN Temperature Adjustments By Decade

This is just a quick post. Someone posted a Tweet about record high temperatures in the 1930s and why those hadn't been adjusted out now that we've had the "hottest year in history." I had looked at adjustments to USHCN temperature records previously and was working on a post about several local stations, so I was pretty sure there were substantial adjustments to data in the 1930s. So I gathered the data from a few local stations, computed the differences between the raw data and the official data, and computed the average differences by decade (1890s-2010s). I picked 5 sites, including the one used in the previous post (New Meadows).

The added sites, their elevations, and their distances from New Meadows are:

The data records run from about 1893 to 2014, but not all sites have the same length record and some raw data are missing from most records so decadal means vary in sample size; the 1890s decade has the least data. I regard this as relatively unimportant for this coarse look, but it should be noted. It is also important to know that the pattern of adjustments is not uniform among sites or among decades; I hope to investigate this more thoroughly in an upcoming post.

The results demonstrate a distinct pattern of reduction in size of adjustment from older to newer data:

On average, adjustments resulted in lower temperatures than the raw data. The largest adjustments were to the oldest data. While the 1890s decade had the least data (with one site missing) the magnitude of the adjustment was similar to the 1900s decade. From the 1910s decade through the 1940s decade the adjustments were about 40% less but still fairly large; and yes, there were substantial reductions in official temperature in the 1930s. Adjustments since 1950 were about 50% of the previous 4 decades and less than 1°F on average.

I intend to look at the differences in the way data were adjusted at each of these sites in an upcoming post, because they aren't uniform. Two of the sites (PY and MO) were adjusted relatively uniformly among decades, and one, (PY) had relatively smaller adjustments. Clearly, though, the general effect appears to be a cooling of the past relative to measurements.