Water Budget

[young...@gmail.com]

Hello, 

Does anybody here do water budget analysis of WRF-Hydro simulations? I am currently running a model of a small basin in Iowa. I am using AORC forcing data and the NWM model setup. My domain contains the calibrated NWM parameters. I am doing a 5-year spin up and then a 3-year simulation. 

I am running the model like this for two different simulations, one with tile drainage and one without tile drainage. When I have the tile drainage component turned off, I have significantly more volume as runoff compared to when I run with TD. When doing a mass balance and accumulating the evapotranspiration, rainfall and outlet discharge over the basin, the ET and discharge is much higher than precipitation. This is still the case when accounting for changes in storage (soil column volume, aquifer depth, SWE, etc.). 

I got outlet discharge to mm but converting m3/s to mm3 per timestep and then dividing by the area of the basin. ET and precipitation are averaged over the basin. ET values are from LDASOUT files, precipitation from LDASIN and outlet discharge from CHRTOUT.

ET rates are similar between simulations. I don't know where I'm losing runoff volume between the two different simulations so if anybody has any idea I'd really appreciate it!

Thank you!

Kaleb

[aubrey]

Hi Kaleb:

This is the equation I use for water budget calculations:

Precip - ET - change in Soil Moisture - change in SWE - change in Canopy Store - change in Surface Head - change in GW Storage - Surface Runoff - Baseflow = 0

For NWM applications, you can use the spatialweights.nc file to map the gridded values to the catchments. Alternatively, you can create your own spatial weights between your watershed unit and the grids, just note that the WRF-Hydro water budget is built on the NHDPlus catchments, to it may not sync up if you use units that are not consistent. For watershed water budgets, I generally stop at the channel (just because we do not have an easy way to estimate change in channel storage... though over the long-term this should be minimal). So I determine all of the reaches within my watershed, sum up the incoming (qBucket and qSfcLatRunoff) fluxes over all reaches, then divide that by the total watershed area so that I can compare it to the other land fluxes as a depth. I can usually close the water budget within 2% or so, but that will get worse in watersheds with waterbodies (since these end up being a sink in the model) and in flat areas where there are potentially non-negligible flows going into or out of the "watershed" through lateral terrain flow, since "watersheds" can be poorly defined in flat areas.

I can't speak to the tile drainage formulation since I haven't used it and we are not actively supporting this formulation.

Hope that helps!

Aubrey