How can we reindex off line?

1. Index is out of date on www.curriki.org

Task
Can you develop a process where the index is done off line, then copied back to www.curriki.org?

Then I would run a refresh to catch some, but not all the changes that happen between the offline reindex and copying it back.