are indeed temperature sensitive.
They are, after all, just a fuel cell. And chemical reactions (which is how they work) are temperature dependant.
Most sensors contain internal temperature components (typically a calibrated thermistor) which attempts to compensate for ambient temperature differences. The problem is that the ambient temperature around the sensor is not the temperature of the gas impinging on the membrane, and immediately after a fill it could be QUITE HOT.
For example, its pretty common around here to get a fill of Nitrox in an AL80 that reads 3200 psi when the guy filling the tank is complete. A few hours later, with the valve closed, in 80F ambient temperatures, the gauge reads 3000 psi.
Simple application of Boyle's and Charle's law tells us how how the gas was - since we know how hot it IS.
To go from 3200 to 3000 psi there was a 6.6% reduction in temperature in KELVIN.
Since the endpoint was 299.66 K (~300K), the gas was at 318K, or 113F - at the completion of the tank fill.
Yes, that's almost 30F hotter.
This could easily account for an "off" reading if the sensor is not at gas temperature (and it won't be unless its immersed in the gas - the temperature compensating resistor is typically in the back of the shell, behind the membrane)
) 
