Humidity sensors are sensitive to the environment, they operate in, in more than one way. Of course there is the obvious way in which a humidity sensor is affected by its environment – it measures the humidity in the air surrounding it, but this humidity can also affect the accuracy of the sensor and, in some cases render the sensor unusable.
If a humidity sensor, such as the low-cost DHT22, is exposed to extreme humidity, for example 100% relative humidity or fog for an extended period, it becomes saturated. This causes it to be return faulty readings and it might even get stuck at 100% relative humidity. The datasheets for the low-cost sensors DHT22 and RHT03 specifically mentions that they are not designed to be used in environments where fog can develop.
I have unwillingly tested this myself, before I became aware of the environmental requirements for the DHT22, by using it to measure outside temperature and relative humidity. The set up worked perfectly for a couple of months in the fall but as we came closer to winter the humidity readings started acting weird and after a very foggy day it got stuck at 100% with only a slight decrease around noon when the temperature were at its highest.
By incident I then stumbled upon the article The Internet of Bees at Make in which Nathan Seidle explains how he wrapped the humidity sensor (HTU21D) in PTFE tape to compensate for extreme conditions, while still getting accurate readings (accuracy is dependent on the chosen sensor). By chance I had some PTFE tape lying around and wrapped a new RHT03 tightly with the tape and used it to replace the now malfunctioning DHT22 (DHT22 and RHT03 are interchangeable).
So far this works, but I am not sure for how long. Luckily, the rest of the sensor hardware is situated indoors and should not be affected significantly by the outdoor environment.
Robert Smith has several thorough tests of various humidity sensors in which he confirms my findings (in a much more scientific way!) as well as elaborate on the choice of humidity sensor to use for extremish environments.
On a side node: during the same extremely humid period my green house sensor gave up. It is basically equivalent to the outdoor sensor, but all the sensor hardware is situated in the green house, which is only poorly shielded from the outdoor environment. I have not restored it completely, but can confirm that RFM12B does not like humid weather as well – the ATTiny84 and AM2302 still works.