Ultimately, novel and original are up to reviewers and funding agencies, but you can think about this from a rule of thumb perspective:
Original has not been published before and is not submitted somewhere else currently.
Novel typically refers to a result being "not obvious." You may have heard of the phrase "incremental advances" before. Most incremental advances (though original) would not be considered novel. Some incremental advances, however, might be considered novel based on the methodologies used to accomplish the work.
One example of this: I am conducting research into the successive composition of topological relations (for context, I am a researcher in geographic information science). Because composition is well-known, you could consider successive composition as an incremental advance, which is original. What sets that work apart is the method by which analysis can be performed upon it (Markov Chains), providing a profile of information loss. That method allows for understanding something more about a reasoning system and provides a methodology for comparison between types of relation vocabularies. That might be considered novel.
Someone asked this question relative to grants, so here is a link to a discussion regarding that: