One of the key criteria for placement of weather stations in the COOP, and by extension, the subset USHCN network is the requirement for a “warm body” (an observer) to read the thermometer daily, and write down the max and min (plus rainfall) in the B-91 observers log for monthly submission to the National Climatic Data Center.
By that criteria for a live observer then, manned facilities such as fire stations are often prime locations for NOAA climate monitoring stations.
The problem with many fire stations is that they are placed right in the middle of the things they are designed to protect – towns and cities. That’s the case with Monroe, North Carolina’s USHCN Climate station of record. I visited this station shortly before my visit to NCDC as part of the road trip you all were kind enough to send me on with funding help (again my thanks).
One of the most interesting things is the view the MMTS temperature sensor has at the major highway intersection that the fire station resides on. In addition to the brick wall about 28 feet away acting as a nighttime heatsink, possibly biasing Tmin readings, and the a/c units, the MMTS has a clear view of traffic and the new local Rite Aid Store:
Monroe, NC USHCN Station looking west
The aerial view gives a better idea of the environment that the MMTS temperature sensor sees:
A complete set of station pictures are available here at this link.
The question then becomes: what is being measured? Temperature and along with it climate trends or that plus increasing traffic and city growth? The plot from GISTEMP shows an abrupt increase in temperature starting around 1970, after a period of cooling:
The MMTS has been at this location since June 9th of 2003. Prior to that it was located further south of the city. As the Google map from NCDC’s MMS interactive database shows, the location of the station has been getting steadily closer to the center of the town center:
The blue marker is the current location at the fire station, the southernmost marker is the oldest, dating from 1948-08-01 to 1954-02-01, while the middle marker dates from 1954-02-01 to 2003-06-09. Unfortunately while temperature data for this location extends back to 1888, location data in NCDC’s database only goes back to 1948.
These moves don’t explain the abrupt jump around 1970, which appears to be a discontinuity. However a check of adjacent stations in the are also show a similar jump at the time. So it appears to be natural. The station at Charlotte, NC which you can see here, has a similar jump around 1970. It is 48 kilometers (29.8 miles) away.
So the jump appears to be natural. But notice that Charlotte’s Douglas Airport record shows a cooling trend post 1990 while Monroe, just 30 miles away shows a warming trend. Which one represents the true temperature trend?
The difficulty in figuring this out is part of the problem with the USHCN and COOP network, because sorting out stations moves, population growth, and changes around the sensor all figure into the record. In my opinion, disentangling these elements to get the true climate change signal requires more than just a blanket application of an algorithm and adjustments, it requires a detailed examination of the station history and measurement environment as well.