In the hour or so since I last posted on this topic, Emily Lakdawalla from The Planetary Society blog sent me an email; she has solved this whole mystery!
Turns out the 2005 image was actually a set of data taken by Clementine in 1998 and released in 2005. OK, that's one thing. She also found a crater in the Chang'e-1 image that appeared to have a new crater near it that was not in the NASA data! OK, that's two. But then she realized that the Chang'e-1 data was not assembled correctly; it's a mosaic of several smaller images, and two of the mosaic pieces were offset. When she realigned them, poof! The corrected image matched the old one perfectly! The new feature disappeared, and all is right with the world. Or even a different world.
Go check out her really amazing sleuthing. This is how it's done, folks. Case closed!