I'll try moving it into the fridge and watch when the RH starts changing.
Regarding sensors used by OS and Davis:
OS uses a humidity sensor which has several interleaved metallic fingers on a ceramic substrate. They measure the AC resistance between the metal fingers and use a look-up table or some curve fits to convert the resistance reading to humidity. Unfortunately, temperature also has a big effect on the resistance, so they must factor that in too.
The lower the humidity, the higher the resistance. The colder it is at a given humidity level, the higher the resistance. Once the resistance gets too high, an inexpensive unit like OS sells cannot accurately measure resistance any more. The bottom line is that these sensors work great at medium and high humidities but fall apart when it gets really cold and/or really dry.
Perhaps this is why OS chose to stop reading humidity at low temps. For example, I looked on the data sheet for one sensor, the Ghitron HCZ-D5 (just an example, I have no reason to think this is the one OS uses). It does not give data to convert the resistance reading to humidity below a temperature of 5C.
Don't quote me on this, but I believe the Davis units use capacitive humidity sensors, and these have the opposite problem -- they work quite well at low humidity but not as well when the humidity soars.
Seems to me the ideal setup would use two sensors -- a resistive one for high humidity and a capacitive unit for when things get dry and/or cold.