By this are you saying that they would need it to be at -70C for long term storage but that it's possible it is delivered out to inoculation centres from a regional hub and kept at -20C for a few hours while the vaccine doses are being administered?
The typical way this is done is to define a long term storage condition and period and test the product at start and end.
Also an "in use" period is defined - where the patient/pharmacist has the treatment. That might be, for instance with a bottle of pills designed to last a week, to demonstrate the contents are stable after the seal has been broken for a week. In this case it could be 5 days at fridge temperature.
Finally, transport studies are done, which demonstrate stability for limited periods of time outside of normal conditions eg high temperature excursions if sat on an airport in the sunshine in Texas for that bottle of pills, or perhaps a few hours at room temperature in this case. These can include freeze/thaw cycles as that can affect product performance more than temperature change per se.
Stability may need to be demonstrated with all of these concurrently ie period of shelf life at long term storage PLUS a transport excursion PLUS a period in use - as that's what will happen.
Exactly what studies are done depends on what is already known about the fundamental stability, the intended supply chain and end use. Typically, regulatory agencies will insist on new stability data for each site of manufacture, as subtle changes in manufacture can affect stability.
Now some speculation: This vaccine has not had very long in development at all, so stability data and understanding will be much more limited than usual. There may be almost none for some manufacturing sites. Also, it might be very difficult to characterise degradation for a long RNA strand - I doubt it is possible to precisely characterise, and I've no idea what criteria might be used. It might be that -80 is used because of the difficulty of defining how much degradation is acceptable, or to standardise between sites of manufacture.