Here are some interesting quotes from
The Weather Observer's Handbook by Christopher Burt (2012 Kindle Version; locations 4569-4631):
"Aspirated screens are ideal for accurate temperature measurements, and when fitted with RH sensors will generally give more representative RH values too, but the greater volume of air movement over humidity sensors tends to exacerbate dust and salt ingress problems and
probably ultimately shortens their working life." [emphasis mine]
"Calibration drift is a problem with humidity sensors, particularly in less expensive systems where it can exceed 5% per year. Whatever type of equipment is used, regular checking over a range in humidities is essential if reasonably accurate long-period humidity measurements are sought. Calibration checking is best carried out annually, or more frequently if spot checks indicate the sensor is regularly more than about 5% different from independent instruments."
"The errors inherent in the measurement of humidity - whether by dry- and wet-bulb or by electronic sensor - mean that the
RH is at best accurate only to about 2-3% in its mid-range.... At high humidities response will be slow, while at low humidities and low temperatures, errors increase and the accuracy falls off further. The same goes for dew point - although often quoted to a precision of 0.1 degC, in reality the measurement is probably no better than +/- 0.5-1 degC when derived from humidity measurements, with still wider error ranges at low temperatures and humidities." [emphasis mine]
"
Combined temperature/RH sensors are popular, but can become expensive and inconvenient if the relatively short working lifetime of the humidity component mandates replacement (and recalibration) of the temperature sensor too." [emphasis mine]
"The calculation varies somewhat depending upon the airflow over the sensors, and for accurate work this needs to be taken into account. For this reason there are different psychometric formulae and tables for sensor exposed in a passively ventilated shelter such as a Stevenson screen and for those in a forced airflow, such as an aspirated or whirling psychrometer. There are also slight differences in the method of calculation for temperatures below 0 degC, owing to differences in the saturation vapour pressure over liquid water and ice surfaces."
*************
Essentially what he's been saying, and it's been known for years even before he wrote it, is that, at least in non-arid climates, our sensor packages are
disposable since they cannot really be calibrated at multiple points (which is required because they aren't linear). I would add, if you have a fan-aspirated Davis unit, expect to order replacements well before the sensors fail (yearly? every other year?) if you want to be within the ballpark on humidity. Also, the manufacturers specs are marketing, not reality, as we've been discovering. This isn't a problem limited to Davis. It's just that Davis is one of very few manufacturers selling aspirated stations. The extra air movement exacerbates the wear and tear on the humidity sensor. Over time, they all drift, though.
Humidity checks should only be made against known reference-quality weather stations. One-off comparisons are not advised because even very expensive AWOS systems at airports can sometimes be out of calibration (which is currently the case at my local airport). But if you look at enough NWS/RAWS stations (
https://www.wrh.noaa.gov/map/) within about 50 - 100 miles (assuming synoptic weather patterns and relatively flat terrain, no mesoscale influences, etc) a consensus dew point range begins to emerge. Plus you can check the software models for projected dew points in your area (weathernerds.org) and, likewise, perform an average (i.e. HRRR + NAM + Euro divided by 3). Do NOT use other neighborhood weather stations (unless you know them to be accurate) because, at least in my area, all of them run well above the reference stations. Very few station owners properly maintain their stations due to the expense and complexity. So the stations humidities and DPs almost always drift up over time and run wet.