Very Small Gridded Streamflow Simulation

[Tong Wan]

Hi,

I've been running the wrf-hydro simulation over a small domain 13*9  nearby a relatively dry site in Utah.  My simulation time period is from 2015-2020 and I spin up from 2010-2015. However, my gridded simulation output of streamflow is extremely small and is about 1/100 of the USGS observation. I attached my domain input files, the forecast points csv, and my streamflow output here, could anyone help me with this issue? Any suggestions would be appreciated.

Thank you,

Tong

[Kevin]

Tong,

I took a quick look at your domain, and it seems okay. You mentioned you are looking at gridded simulation output, but your WRF-Hydro input files appear to be for a reach-based configuration. If you are running with gridded-only routing, you will need to set the LINKID grid to -9999 and turn off the reach-based option. Did you check the precipitation forcings you are using, and do they seem to match gage observations? 

You may want to look at what the groundwater is doing during your simulation, and potentially run with the groundwater turned off to see the effect on surface flows.

I also notice that this is a very limited-area domain and you are running at a relatively coarse resolution (1000m/250m). As such, you do not have much variation in your parameters on the LSM grid. You may want to bump up the resolution a bit for both LSM and routing, and use a higher-resolution landcover and soils dataset if possible. Other users may have additional suggestions.

Kevin

[Tong Wan]

Hi Kevin,

Thanks for your response. I really appreciate your time! 

I checked my namelist files and I found that yes, I commented off the reach-base option and the model is running for gridded configuration. 

I also took a look at my forcing input data. I compared the precipitation with a met station's measurement which is about 250km from this site, there seems to be a large difference between the precipitation, which might be a reason why I got very small streamflows. I used the NLDAS-2 data and regridded by the bilinear interpolation using the ESMF regridding tools.  Is there any problem with the method or the data that I used for regridding? Do you have any advice on what other data I can also use to create my forcing data which may provide a better representation of the climatic information? 

In terms of the resolution, yes, my current LSM resolution is 1000m and the routing is 250m. If it is possible, do you think a 500m LSM resolution and 100m or 125m routing are good enough or we may try finer? Currently, my DEM source is 30m, if I set my routing resolution as 100m or 125m, do I also need a finer resolution of DEM?

Thank you,

Tong