Thanks! I'm afraid my background, like yours, is more in methods that are not specialized for time series, and so I couldn't credibly give any comprehensive references. My understanding is that a lot of methods designed specifically for time series draw heavily on the theory of stochastic processes. For example http://en.wikipedia.org/wiki/Autoregressive_model, http://en.wikipedia.org/wiki/Autoregressive%E2%80%93moving-a.... I once took a course called Signals, Systems, and Inference that covers some of these ideas (full course notes here http://ocw.mit.edu/courses/electrical-engineering-and-comput...), but that's about as far as I've gotten along that road.