Europe did NOT have a sudden spike the last 3 weeks. And a lot of technical people in the press and on twitter Really Ought To Know Better.
This is important, because if you ask "what changed 3-4 weeks ago" looking for the cause of the increases, you're going to get the wrong answers because it started in July and has been growing uncontrolled since then.
This is the graph you always see. Clearly, a huge spike in the last 3 weeks, before that the infection rate was under control...according to the narrative.
But if you understand exponential growth, you realize this is the typical "hockey stick" chart you get when exponential growth has been going on under the radar for a while, and finally reaches crisis levels.
If we look at the same curve on a log scale, we see the real story...
This is the same data set.
Starting in July, Europe as a whole saw rapidly increasing cases. The rates varied a bit, but were always rapidly growing.
From mid July onward (indicating changes from late June) we have a doubling every 3 weeks.
Through September, mostly a 4 week doubling, from whatever changes happened in mid August. (End of holiday vacations slowing the spread a bit?)
From about October 5, we have a 2 week doubling, reflecting changes from mid to late September. Schools reopening? Weather? Both? Other?
So the rate of growth got worse in late September, but it was already wildly out of control.
This isn't a one-country thing. The major countries in Europe, and the UK, have had growing rates for months now. I don't think the latest increase is largely weather driven; Norway and Sweden have been climbing for months (mid August is hardly cold and flu season).
Looking at this chart it seems the primary cause is whatever changes happened around the start of July, and I'm guessing that was reopening combined with holiday travel and festivities.
Anyway, if you're not asking the right questions, the answers probably don't matter. This isn't an October thing.