Lets just say you use a simple pt100 sensor for temperature. What happens if someone places a structure nearby, and the windows during certain parts of the day where you happen to sample reflect light onto the sensor/sensor housing. Repeat ad nauseum for all kinds of reflective sources.
Pleebo's article states some discrepancies arise due to movement of test locations.
That's to be accounted for. That isn't an error, but a prevention of an error to get better data. The previous data isn't compromised.
Perhaps the sensor has had too much condensation or rain/freeze cycles, and now has a thick coating of ice.
Perhaps birds have found a nice place to make a nest.
Perhaps birds have found a good place to perch overhead, and give the sensor a regular coating of excrement, leading towards ever slower transitions towards ambient temperature.
Perhaps you live in a place where people like to vandalize or shoot things, and have damaged the station.
Unless there is lack of maintenance and/or equipment review, I don't think that would be a big issue.
If nobody checks on the equipment and/or performs regular maintenance/calibration, then the data will obviously be skewed/tainted. That would be recording error, not equipment error.
Doing things consistently, correctly, and accurately in all circumstances is HARD.
I understand that. But if people are going to swear by the data, then the data should be verified.