UTF-8 interpreted as ISO-8859-1/Windows-1252 and then encoded back into UTF-8. Probably came from a copy/paste of material in a different encoding, via a program that was not charset-aware.

http://www.i18nqa.com/debug/utf8-debug.html