I'm with ancient_scars on this one. For this version of the game the Q2 engine will be sufficient, these kinds of games are more about gameplay than graphics, and for me at least the graphics are more than good enough to create an enjoyable atmosphere...
Having said that, for a sequel I would support using a combination of Ogre3D, LUA, OpenAL and one of the many available physics engines to create a game engine from scratch. Using these systems, the dev team would not have to worry so much about keeping the graphics and sound engines up to date with current technologies, allowing the team to focus more on the in-game scripting, GUI's and storyline events.
The development of the sequel could begin once the codebase for this version is almost complete, meaning that once UFO:AI is released, UFO:AI2 is already under development.
Thats my 2c worth anyway.