For your NOAA reports, what calculation method do you use, the High-Low or Integration method?
I switched to the integration method after initially using the high-low method because I think it better matches the averages of a nearby county-operated weather station. The county station has been in operation for nearly a decade and I used their raw historical data to calculate my NOAA set up page's monthly normal temperatures and rainfalls. Admittedly less than ten years of records isn't a very long history, but I'm at a much higher elevation than the local airport and the Reno airport is known to have a serious heat island problem so I couldn't use those records. Heck, NWS even cites it in their manuals as a case study of where not to site a weather station in order to avoid biased temperature readings.
