I’ve spent a lot of time on this blog showing how badly maintained and situated the stations in the USHCN network are. And rightly so, the majority of them have issues. But, finding the good ones is actually more important, because they are the ones that hold the true unpolluted temperature signal. Unfortunately, the “good ones” are few and far between.
But when one comes along that is a real gem, it deserves to be highlighted. I present the USHCN climate station of record for Tucumcari New Mexico, COOP ID # 299156, located at the Agricultural Experiment station about 3 miles outside of the edge of town.
I “had” (he just moved to St. Louis) a nephew who lived in Tucumcari, and he just happened to be friends with the director of the experiment farm. Before my nephew left they both helped me get this survey done.
Surfacestations.org image gallery link
This station has several advantages:
- Length of continuous record – going back to at least 1946 at this location, possibly to 1905 but NCDC MMS metadata stops at 1946.
- Length of continuous instrumentation – using mercury max/min thermometers
- Length of continuous data record – there doesn’t appear to be any missing years
- Lack of encroachment – 3 miles from the northeast edge of town, little development, little UHI. Tucumcari is well off the beaten path of development. Population actually declined 12% in recent years.
- Good siting – the station rates a CRN2 due to distant trees and sun angle, and one small asphalt road 70 meters away.
Eyeballing, you can see that the temperature data trend for Tucumcari is slightly positive over the last century, about 0.5°C, but there is a “bump” in 2000, which brings it to about 0.9°C. This same bump appears in neighboring stations such as in San Jon (33km away) and in Boys Ranch (135km away). There is nothing in the metadata location or equipment record to suggest a reason for the bump. So, either the bump is naturally occurring, or there is something we don’t know about that changed in the local environment, or we have another data set splicing error like the GISS Y2K debacle from last year.
I plotted the data provided by GISS (which you can find here) to show the effect of the “bump” at year 2000 on the overall trend:
Here is the data plot after the GISS homogeneity adjustment, I’ve hue shifted my saved version to red to help keep the graphs visually separate:
And here is the overlay of the USHCN data from GISTEMP and the data from the GISTEMP homogenization process:
In this case, the GISTEMP homogenization code appears to do what would be reasonably expected; reduce temperatures in the present to account for population growth and UHI. I’ve pointed out more than a few times that the GISTEMP homogenization adjustment often becomes flawed for truly rural sites like this when there are large cities within the 250km up to 1200km (depending on process) adjustment zone that Hansen uses, that have accelerating UHI trends. Due to these cities, often the past of a rural station gets adjusted cooler, resulting in an increased temperature trend, such as what happens at Cedarville, CA. Hopefully we’ll have a detailed analysis of that adjustment from John Goetz soon.
If you look at this list, you’ll see that there are a lot of rural stations within 250km. Tucumcari has the advantage of being truly in the middle of nowhere when it comes to other big cities. The closest big cities are Amarillo and Lubbock, but as I understand the algorithm used, when they are near the edge of the 250 km zone, their weighted value decreases.
In this case though, the GISTEMP homogeneity adjustment doesn’t take Tucumcari’s declining population into account, it only uses nightlights, and while the population may dwindle, town infrastructure usually doesn’t; streetlights counted around the station likely remain.