wrfinput in warm simulations

[Sara B]

Dear WRF-Hydro users, 

I would like to ask you some clarifications about the usage of the wrfinput file in case of warm simulations with WRF-Hydro.

I am using WRF-Hydro to produce hydrological forecasts. WRF-Hydro is forced by WRF forecasts (FORC_TYP = 3) and starts every day with the restart files (RESTART.*_DOMAIN4 and HYDRO_RST.*_DOMAIN4) produced the day before (warm start).

From the documentation, I read that wrfinput is always required by WRF-Hydro to run, also in the case of warm start. However, I could not find which variables exactly are used from the wrfinput file (all variables or only a selection). Therefore, in order to generate forecasts, WRF-Hydro uses every day the wrfinput file produced by WRF on that day with real.exe. Now, I am wondering if I am doing something useless...

I would like to know if I could always use every day the same wrfinput or if it is important to have every day the initial conditions of the considered day (as I am doing now). I guess that the key point is to know if WRF-Hydro (for warm starts) uses only the static variables or also the dynamic variables (such as soil moisture and temperature) from the wrfinput file.

Thank you very much!

Best regards,

Sara B

[Alfred Zhong]

Running WRF Hydro from a restart file is technically considered a "warm start" but practically you would want to restart from a simulation that's had a few years of model spin up time for the values to reach equilibrium.   Maybe something is getting lost in translation, but it doesn't sound like you're doing that. (see https://ral.ucar.edu/projects/wrf_hydro/faqs)

Regarding the wrfinput file you mentioned, it is required to provide the initial conditions for the land surface model. I don't fully understand your question, but it might be easier for you to create the wrfinput file from the R script (or their newer Python script) in their WRF Hydro preprocessing tools. You only need to specify the model start month so that it can approximate the initial conditions using that given month.

Pre-processing Tools: https://ral.ucar.edu/projects/wrf_hydro/pre-processing-tools

Pre-processing Tools (Python): https://github.com/NCAR/wrf_hydro_gis_preprocessor/

https://wrf-hydro.readthedocs.io/en/latest/model-inputs-preproc.html#initial-land-surface-conditions:

"The R script will create a simplified version of the wrfinput_d0x.nc file including all necessary fields for the Noah-MP land surface model, but with spatially uniform initial conditions that are prescribed within the script and requires only the geogrid file geo_em.d0{x}.nc as input. Step-wise instructions and detailed requirements are included in the documentation distributed with the script. Users should be aware that the model will likely require additional spin-up time when initialized from this file."

[aubrey]

Hi Sara:

Just to add to Alfred's nice response, a couple of points:

1) When we run WRF-Hydro offline (uncoupled to WRF), the version of the HRLDAS driver that we use for Noah-MP requires the wrfinput file to define some static fields for Noah-MP (e.g., land cover type, soil type) in addition to (on cold start only) LSM starting states for soil moisture, snow, etc. When you run with a restart file, the model will NOT use the initial soil moisture, etc. from wrfinput, but will still use that file to grab the static information (land cover type, etc.).

2) There is a bit of legacy duplication in that the hydro routing modules rely on the geo_em file for some of this static information, but Noah-MP shifted to pulling everything (static and initial condition fields) from wrfinput. As some point we will migrate the hydro codes to use wrfinput so only one file will be needed for both, but until then we still require both the geo_em and wrinput files.

Hope that helps...

Thanks!

Aubrey