How not to measure temperature, part 32

A common theme with official climate stations of record is their placement with city and county fire stations. The reason? An observer is needed to transfer the data from the thermometer to the B91 form sent to NCDC every month. Unfortunately, fire stations are often not good places to measure temperature due to the amount of concrete and equipment around them. Their placement is to better serve the city population, putting them in the middle of UHI. For example here we see Bartow, FL city fire station provided by volunteer surveyor Don Kostuch.


While this placement “is” over grass, the narrow grass strip is also within feet of parking, a major thoroughfare, located downtown near high rise buildings. This is not an ideal place at all to measure temperature, yet it is the official USHCN climate station of record. As shown by the graph provided by NASA GISS, there appears to be a step bias introduced in the early 1980’s.


A possible cause is that the MMTS electronic thermometer was placed in 1984, according to NCDC’s database. It is not uncommon to see such step biases introduced when the MMTS replaces the traditional Stevenson Screen shelter and mercury thermometers, since cable issues often force the MMTS to be closer to buildings. For example, consider Marysville, CA, one of the earliest and worst examples of a place unsuitable for climate measurement.

So far, over 70% of our USHCN climate network uses the MMTS electronic thermometer system.

While fire stations do give the appearance of a regular warm body to record the temperature and send it in to NCDC, sometimes it doesn’t always go as planned. For example, this B91 form from Bartow looks a bit like a workbook assignment from school. There is a lot of data missing, and uncertainty over a couple of days of data:

See the original B91 form in PDF format.

Now while this form lacks neatness, the observer at least tries to provide something each day, in contrast to the very neat but nearly vacuous B91 form provided by the Marysville observer (PDF format).

Please don’t get the idea that I’m putting down the hard work of the amateur climate observers, they perform a valuable and much needed service. The point here is quality control issues and missing data.

These B91 forms are just a couple of random samples, more on that missing data issue soon.


2 thoughts on “How not to measure temperature, part 32

  1. What Don didn’t mention, maybe because it isn’t obvious from ground level is that just across the fire station parking lot, there is a large surface level water storage tank. See Microsoft Live Local for a good view from above looking west. For that photo, put “http:double slash” before:
    Even more curious is that a elevated water tower just behind that tank was apparently recently removed, as it appears in one of the directional views, while only the concrete footings remain in the other three. Is that sort of major nearby change recorded in the station records?
    Finally, the driveway coming out of that facility appears from the markings to be a drive through of some sort, guaranteeing a maximum traffic exposure in addition to the fire trucks.

  2. So now the issue isn’t confined to microsite violation. Just how much of this-here data is being recorded and how much is USHCN interpolating?

Comments are closed.