I have the Aqualogic chlorinator which is very similar to the Aquarite unit and actually uses the same cell. Mine is about 3 years old so it may work slightly different than yours. After much research and testing on my unit here is what I have found.
First, I confirmed with a Goldline tech that the Aqualogic unit does compensate the salt reading for temperature and I would assume that the Aquarite units also compenstate for temperature although I am not sure about one that is 8 years old. To confirm this myself, I ran my spa from 55 degrees F to over 100 degrees F with the same salt level and found the salt reading not to change by more than one step (i.e. 100 ppm). This would not be possible if the unit did not compenstate the salt reading for temperature. Given that the salt reading is fairly constant over temperature, I would assume that the temperature compensation is sufficient.
I have noticed that when subjected to a very fast change in temperature, the unit will not adjust fast enough to compensate. I only see this problem when I first turn the solar on and I see that the current spikes but the temperature sensor is a bit slower in reading the change in temp so I suspect that is the reason. Once stabilized, the reading returns to the correct value.
Also, I believe that most SWGs do compensate the salt reading but not all compensate the chlorine output for temperature. The pool pilot has long had chlorine output temperature compensation and the current generation Aqualogics have it as well but historically, most units did not do this.
Without temperature compensation in the salt reading, the error would be signifcation at temperatures far from the calibration point. For example, assuming the SWG did not compensate for temperature and the manufactures calibrated at 82 degrees F, then the error at 55 degrees would be close to -900 ppm. At 70 degrees the error would be -500 ppm. At 95 degrees the error would be about 400 ppm.
I also confirmed with the tech that they actually use the current sensor reading to trigger high and low salt readings and not the actual salt display. They primary reason is that their concern was to protect the unit circuits against too much current draw at the high end and too little chlorine production at the low end.
Now this is not to say that there is no error in the SWG salt reading. I have found that my SWG also reads consistently lower than the strips. I believe this is a calibration issue with the SWG although I have yet to confirm this with the Goldline techs.
The SWG measures the salt level by very similar method used in TDS measurements. Using the amps and voltage reading to determine the conductivity of the water, the unit then uses a
formula which converts conductivity and water temperature into salinity. Because the conductivity of the water is dependent on not only the NACL in the water but other solids, the SWG would normally read higher than the actual NACL level. My unconfirmed belief is that Goldline calibrates the SWG to read lower than normal because they cannot compensate for other solids in the water. Probably to ensure the owner puts enough salt in the water. I am not sure they would confirm this even if it were true.
Here is another good source for conductivity.