The method modeling the lumped elements in FD-TD has bee studied by many researchers. I remember Piket-May's paper in IEEE MTT, 1994 would be the first work.
Actually, her method is that one-cell of FD-TD is replaced with the differential equation of the lumped model. The detail is in her paper.
In RF, lumped devices are easily modeled by simple networks of ideal RLC componencts.
Take a look on section 1.2 of this book " Signal Integrity: From High Speed to Radiofrequency Applications - Fabien Ndagijimana ISBN: 978-1-84821-550-4 Wiley-ISTE http://eu.wiley.com/WileyCDA/WileyTitle/productCd-1848215509.html
This section describes the behaviour and model of components in high frequencies.