With Lifepo4 pulling high current at or about 0c can cause long term capacity reduction. Its supposed to be caused by copper being plated into the carbon matrix, into which the ions / charge move through the cells. So the trick is to limit this condition, ie get them warm before you use them.
I have not heard that one before. Do you have a reference on the chemistry behind it? Is it LiPo, or LiFePO4, or pretty much any lithium chemistry?
I ran my pack for a bit at -26C. The pack started out fully charged, and was not happy about moving the Ranger at all. I got the 'jitters' ... where it surges a bit as you press harder on the throttle ... at under 3 mph.
As I used some energy and the pack warmed up a bit, the 'jitters' happened at higher speed.
After I got it into the heated garage and left it over night, the pack was still at 54V. So not FULL but over 80%. It took a bit of a charge, but not too long.
They also take less charge when at or about 0c so when you charge, again do it when the pack is warm, ie just finished using it.
What Jack over at EVTV explains is that, during charging below 0C, the lithium metal plates out of the electrolyte. So charging is doing permanent damage at that point, since the lithium is no longer available to exchange electrons. I went through his footnotes for the published papers and could only see the summaries without paying. But it seemed to check out.
This is sort of consistent with my buddy's Mitsubishi Outlander hybrid. It has NO regen braking until the battery warms up. And youtube ... as much as you can trust youtube information ... has many stories of no regen on Teslas when they are cold.
So ultimately its about keeping the pack warm ie above 0c by a few degrees, when charging, and before you use it.
My charger (DeltaQ) won't charge at all unless the pack is at 10C or higher. I think that's a bit paranoid, but I guess they don't know where the temperature sensor will be installed, or how accurate the sensor is.
I have not checked if regenerative braking works below 0C. The SEVCON knows what temperature it is ... I think ...