Without access to the ERV, they are only assumptions. 24 hour periods of time--that is probably the period of time of importance. To try and adjust for daylight savings time in a long-term test/experiment would be silly. If it were me, I would pick a period of time (i.e., 24 hours) as a period of measurement, call that one day, and stick to that metric throughout the test. My guess is that is how it was conducted and recorded.
Certainly when doing plant seedling fertilisation experiments (long ago) at Sutton Bonington (The Agricultural part of Nottingham Uni) we just stuck to using GMT all the way through the program - and each sequence of experiments lasted 84 days so they did tend to cut across DST quite often. This is particularly important with plants, of course, since they react to daylight - but such an approach has also been used in other long running experiments I have assisted with. Lab-time was always GMT.