Across the United States, there are literally thousands of weather stations recording the temperature, precipitation, and various other meteorological variables. Some of these belong to the United States Historical Climatology Network (USHCN). The temperature readings from these stations are collected by NOAA, and are distributed freely by the US government. On the previously linked page, there is a map of the United States with the location of the surface stations used in the USHCN. Many of the stations are close together; many of them are far apart. How many stations does the United States need to accurately diagnose the average US temperature? This post does not answer that question, except that it is far less than the number we currently have.

I’m going to look at three stations today. And since I’m originally from Minnesota, I’m going to use stations from that state. I am also going to use the CRN rating as prescribed by Anthony Watts to show that micro-climate siting issues are small. For this analysis, I chose two stations with a CRN rating of 2 (this highest in Minnesota) and one station with a CRN rating of 5. The ratings are quickly explained on the page above: Basically, the lower the number the better - like golf. The stations chosen were Minneapolis (2), Morris (2), and Detroit Lakes (5).

Surface Station Photographic Evidence

I’ve chosen one image for each of the stations that I thought best represented what was available on the Surfacestations.org database.

Detroit Lakes

This station has the highest CRN rating available (5). Therefore, this represents a poor location for a surface station according to the CRN rating system. As can be seen, this station is very close to an air-conditioning unit. It’s not visible in the photo, but it is also close to a large structure that the air-conditioning unit presumably cools.

Morris

The Morris station is located in a corn/soybean field. In general farmers in Minnesota will rotate their crops from corn one year to soybeans the next. The fact that in this photo the field is corn should not indicate that every year the field is planted with corn. How the different crops affects the temperature is unknown, but likely to be small.

Minneapolis

I really like this photo. It shows the station, barely visible in the background with the huge airplane in the foreground. The obvious implication is that the constant stream of aircraft landing, departing, and taxiing next to the station will have a large impact upon its temperature readings. Note, however, that this station is given a CRN rating of 2, the second highest available.

All Minnesota surface station photos courtesy of Anthony Watts, www.surfacestations.org and Don Kostuch.

First, let’s compare the two stations with the lowest CRN ratings: Minneapolis and Morris. Since these should have the least microclimate influences, the two temperature time series should be similar. (All data plotted below is the raw data from the USHCN archive.)

The results are exactly as predicted. Only I would venture that they are much more highly correlated that most people would think. These two stations are 222 kilometers apart. (Google Maps) Given that there is almost a 0.9 correlation between these two stations, this suggests that the spatial patterns of yearly average temperature are much greater than 222 km, and that surface stations do not need to be this close together to accurately measure a spatially- and temporally-averaged temperature.

One other point is regarding the Urban Heat Island (UHI). The Minneapolis station is located in the metropolitan area of Minneapolis and St. Paul, and thus should be profoundly influenced by the UHI. On the other hand, the Morris station is not located near a large urban area and should not be influenced by a UHI. An interesting result, but one I don’t have time to develop here.

But that is for two stations with CRN ratings of 2. How does the CRN 5 rated station compare to these?

As can be seen, the temperature correlation between the Minneapolis and Detroit Lakes stations is still very high (>0.8). The temperature readings for Detroit Lakes from 1999 onward are higher than at Minneapolis, which the photographer attributes to the addition of the air-conditioning unit near the station. Another possible reason for the smaller correlation is that the Detroit Lakes station is further in distance from the Minneapolis station, almost 300 kilometers.

The above facts suggest that:

• This area of the USHCN is over-sampled
• CRN ratings as applied by surfacestations.org do not contribute a great deal to yearly average temperatures
• The urban heat island (UHI) may not have a large effect in this region
• Local heat sources, such as airplanes and air-conditioners, have only a small influence upon the temperature record

(No Ratings Yet)

 Loading ...

Related Posts:

Trackback URI |