Richie,
We're looking at a very early European-wide ballad here, based on fairy lore. Why would we be looking for real persons? Okay so Thomas of Ercildoune might have been a real person many centuries ago but in this case it's extremely unlikely. You might as well try to find the real Goldilocks!

Okay, the Scandinavian stream is a red herring. In European Folk Ballads, 1967, Eds. Seemann, Stromback and Jonsson (apologies for lack of twiddly bits above letters) the first ballad studied is our 'The Knight and the Elves'. According to a study carried out by Alfhild Forslin, there are 2 definite sub-divisions of the ballad, and it's difficult to say which influenced the other. One area is west Scandinavia in which a king or knight goes to Elfland and a fairy tries to entice Olaf to stay with them with a long list of gifts which he refuses so they put the death curse on him and he goes home, dies and his mother makes all sorts of excuses to his bride about him being out hunting. Finally she lifts a red cloth and sees his body then his bride and his mother die.

The west Europe tradition includes Breton, French, Italian, Spanish, English and Scottish in which the fairy is definitely water-based. They give a Breton text which has the plant motif at the end albeit oak trees and turtle doves (Chinese influence?)

They also give an Italian version which like Giles Collins only has the second half of the story. I'd like to see a French version.

By the way I have a version in a book of German ballads but it's actually just a shortened Sir Olaf from Denmark.