[Note, this was an answer to a user working in the Arctic.]
The dark correction routine should be run at the start of your intense field work. It should be done when the instrument is at, or near, the temperature where it will be recording data. The instrument should have been on for longer than 5 min to allow internal temperatures to equilibrate.
For optimum results during recording of normal profiles, you should let the system adjust to the ambient thermal conditions for about 5 min before starting to collect data. If you don't have time to do this, then you will probably still get adequate data, but perhaps not the most optimum for the existing conditions.
All that being said, the dark current drifts the most over the 25–50°C range, which should not be an issue in your case.
After you run this first dark correction, record a short dark "profile." If you look at the dark profile, you should see data randomly varying about zero.
It is good field protocol to collect a short dark profile with the instrument capped every day because in post-processing, you can examine these dark profiles to see that data from all of the channels are randomly distributed about zero. In the case of conditions in the arctic cruise, we anticipate that you will not see significant variation at the cold temperatures at which you are working.