WRF-Hydro warm up period for flash floods

[hesam goudarzi]

Dear All

  I am planning to simulate some flash flood events using WRF and WRF-Hydro, but I am uncertain about the warm-up settings. From what I understand, a warm-up period of at least one year is required for the model. I have a few questions regarding this:

1. The need for a long warm-up period suggests that the WRF-Hydro model is primarily designed for continuous simulations rather than event-based simulations. This is because introducing warm-up files (such as restart and hydro.rst) requires simulating a period of at least one year. Is there any way to configure the model for event-based simulations? Has any research been conducted in this area, or is the WRF-Hydro model fundamentally intended for continuous simulations?
2. When calibrating the WRF-Hydro model for flood simulations, is it necessary to perform the calibration for the warm-up period as well, or is it sufficient to calibrate only for the event period?

I would greatly appreciate your guidance on these issues.

Best regards, hesam

[aubrey]

Hi Hesam:

The model spinup strategy really depends on your location (e.g., wet vs. dry), your application, and how you are initializing your model. If you are initializing from reanalysis or observed conditions, you may not need to spinup as long. If you are working in a wet environment, you also may not need as much spinup time. If you are interested in baseflow or groundwater is an important mechanism, that may also require longer spinup.

For calibration, we generally recommend calibrating over a wide range of conditions versus a single event. This help prevent over-fitting.

I'm not sure what you mean by "continuous simulations" versus "event-based simulations"? WRF-Hydro can certainly be used for event simulations, but would require reasonable initial conditions. But this is true of any hydrologic model since many water stores have longer memories. For our National Water Model operational configurations, for example, which run short forecast cycles, we run a background analysis simulation and then the forecast simulations grab restarts from that background analysis cycle.

Hopefully others in the community have additional guidance.

Thanks!

Aubrey