Wide, you are quite correct normally, a QNH given at an airport will give a value that should be correct at that altitude for the conditions. which is why this one airport was perplexing, the errors at the approach end of the runway led to a number of odd approaches. Temp correction factors should correct for colder than ISA temperatures for altitudes above the reference datum, which should be the airfield itself. Everywhere else that seems to work quite well, except for one location. That location was very terrain affected on the approach, and was non-radar, so the concerns were related to the initial approach and then the rest up to the final fix. Then came the surprise reports on late finals... Why, not resolved in my time.
Thanks for your confirm.
Actually, that table is a very mean one. The temperature in the vertical is not a "deviation from normalcy" temperature (as one would expect for a "correction" table), though, simply the absolute temperature (in Degrees Celsius, not Kelvin). Add a row with +15C at the top, and the first 4-5 columns will read "0".
IF only a handful of airports are "slightly" off with their PA, my first thought would be "just a calibration error", or, if not persistent, "just rounding". Or maybe something like "humidity" or just up/down drafts due to local geography is relevant ?