Hi, Shirlea! Happy GDPR Day! :-) While I can't specifically answer your question, I would imagine that our own trees at Ancestry have little to do with the changing sets of genotyped data used (you can read a little more about that here). But none of the testing companies are forthcoming about precisely which data sets and algorithms are used; they all covet that as proprietary information because the flood of testing over the past few years is what's been fueling the financial engine.
They are all in something of an arms race to try to lay claim to having the most far-reaching and specific ethnic origins calculations, so the resulting reports change from time to time (and why AncestryDNA has that little indicator to tell you if your results are up-to-date or whether you should let them recalculate). You can read more about my curmudgeonly opinion at the link above, but I believe with the current state-of-the-art these reports should be, at best, be considered "for entertainment purposes only."
At worst the reports can be misleading and divisive; divisive not only within families around the dinner table, but in continuing to talk about "ethnicity" they reinforce a completely false concept of "race" within homo sapiens sapiens. I'd have a smaller burr under my saddle if the companies would start referring to these as reports of "ancient geographic origins" rather than--as Ancestry does in big letters--"ethnicity." Ahem.
By the way, while your chunk of Scandinavian went away, mine doubled at Ancestry but stayed the same at 23andMe and FTDNA. Go figure. You can see my results at that link above.