Long-term comparison of temperatures observed from multiple sensors at the New Brunswick, NJ NWS Cooperative Weather Station

During the mid-1980s, the National Weather Service (NWS) began a mass replacement of traditional liquid-in-glass thermometers mounted in Cotton Region Shelters (LiG/CRS) with electronic thermometers called Maximum Minimum Temperature Systems (MMTS). At the 1993 AMS 8th Applied Climatology Conference, Croft and Robinson reported on the impact of the MMTS changeover at the New Brunswick, NJ, NWS Cooperative Observing Station. They evaluated the previous eight-year period, when parallel LiG/CRS and MMTS observations were gathered. Their findings showed that MMTS maxima averaged 0.6°C lower, minima 0.3°C higher, and means 0.2°C lower than the corresponding LiG/CRS values, with statistically-significant differences (α = 0.01) between the two sets of observations demonstrated by a paired t-test. Fortunately, seventeen additional years of coincident observations (1995-present) have been collected and archived from both thermometer units. This permits us to renew an evaluation of the coincident observations. In addition to standard comparisons of daily maximum and minimum temperatures using techniques from the earlier study, we will evaluate diurnal temperatures and monthly means of all values. Change detection methods will be employed to identify potential changes in records from the individual sensors and the comparative results. These will be assessed relative to any available metadata or perhaps will identify instrument changes that were not documented. Also, for the past five years, a NJ Mesonet station operated by the Rutgers Meteorology Program has been co-located with the MMTS and LiG instruments. These observations will be included in recent comparisons. This direct comparison of thermometry units permits an enhanced understanding of the potential impacts of various units on the temporal homogeneity of temperature records at a given location.