Hi Irene and others,
This question comes up a lot. There are multiple forms of inundation (coastal/surge, riverine overbank, and non-channel overland flow/local ponding). The WRF-Hydro model explicitly models this latter type; non-channel overland flow/local ponding. The output variable for that is called 'surface head' and it is represented on the terrain routing grid (or "fine grid"). Currently, the WRF-Hydro model does not explicitly model riverine overbank flow. The 1d channel model (both vector Muskingum Cunge and diffusive wave options) now have a parameterized "compound channel" structure to emulate the impacts of overbank flow on floodwave propagation within the explicit channel model. This capability is new and will be released in the next version release and will be active in the NWM v2.0 when it becomes operational. That compound channel overbank flow parameterization is optional but it is not 2way interacting with the overland flow mentioned above so a combined overland flow-riverine overbank flow inundation surface needs to be calculated offline. Currently, the National Water Center and others (e.g. UT-Austin) calculate a riverine overbank flow inundation surface offline using the HAND method with WRF-Hydro/NWM river flows provided as an input.
When performing hyper-resolution model runs where grid spacings are less than, say, 30m, it is possible to get a full inundation surface with WRF-Hydro when running the model with the explicit channel turned "OFF" and simulating all surface flow as and explicit overland flow process. Typically, additional DEM conditioning to burn in channel geometry structures into the DEM is required for such applications. However, the overland flow depth modeled in such configurations can be used as an inundation surface. Research is currently ongoing to evaluate the skill of such approaches and we have run the model in this configuration at 30m and 10m for some recent high impact events like Hurricanes Harvey, Irma and Florence.
Regarding coastal inundation, WRF-Hydro currently does not have a method to represent coastal tidal/surge inundation. A few proposals have been developed to build this capability but nothing is formally supported at this time.
Hope this helps,
Dave Gochis