Re: [wrf-hydro_users] Digest for wrf-hydro_users@ucar.edu - 1 update in 1 topic

[Yongxin Zhang]

HI Tomásio,

     I can develop NCL scripts for you to create the WRF-Hydro forcings using the WRF-SAAG outputs but I am currently quite busy and should have time to do so around the middle of August. I am just wondering if that time frame will work for you.

Thanks,

Yongxin

On Mon, Jul 13, 2026 at 5:13 AM <wrf-hyd...@ucar.edu> wrote:

wrf-hyd...@ucar.edu

Google Groups

Topic digest
View all topics

• WRF-SAAG - 1 Update

WRF-SAAG

Tomásio Januário <tomasio...@gmail.com>: Jul 12 09:12AM -0700 Hello Everyone,   I pretend to force the WRF-Hydro model with WRF-SAAG outputs. Can you help me with how to regrid the data? Has anyone created any ncl-regrid script?

Back to top

You received this digest because you're subscribed to updates for this group. You can change your settings on the group membership page. To unsubscribe from this group and stop receiving emails from it send an email to wrf-hydro_use...@ucar.edu.

[Tomásio Januário]

Dear Yongxin,

Thank you very much for your reply.

Yes, that timeline works well for me. I will wait until mid-August. 

Thank you again for your support. I look forward to working with you.

Best regards,

--
You received this message because you are subscribed to the Google Groups "wrf-hydro_users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to wrf-hydro_use...@ucar.edu.
To view this discussion visit https://groups.google.com/a/ucar.edu/d/msgid/wrf-hydro_users/CAEA-d78tg5w3Ri77joYeQw9P4-zzy2gL0k%3DOubiDnpjde%2Bb4Jg%40mail.gmail.com.

--

Tomásio Eduardo Januário

D.Sc. Candidate in Hydrometeorology / Numerical Modeling

Department of Physics – Eduardo Mondlane University (UEM – Mozambique)

📞 +258 (82) 4102-697 | +258 (85) 6637-595 | +55 (11) 9914-12019

📧 tomasio....@uem.mz | tomasio....@usp.br | tomasio....@alumni.usp.br

🔗 Currículo Lattes | LinkedIn |  

[Tomásio Januário]

Hello Yongxin,

🔗 Currículo Lattes | LinkedIn | UCAR/NCAR–Africa Initiative 

[Yongxin Zhang]

HI Tomásio,

      I am wondering if you could send me a few days' worth of the WRF-SAAG outputs (or point me to the location of the output) and your domain files so I can develop the NCL scripts for you to create the WRF-Hydro forcings.

Thanks,

Yongxin

[Tomásio Januário]

Hi Yongxin,

I am currently preparing a study entitled “Assessing the Performance of the WRF-Hydro Forecasting System to Predict Drought Series in the Piracicaba River Basin.” The main objective is to assess whether WRF-Hydro can adequately reproduce the historical drought event of 2014/2015 that affected the reservoirs of the Cantareira System.

At the moment, I am conducting the calibration and validation of the model using GLDAS data for the historical simulation period. Once the model is calibrated and validated, I would like to use the WRF-SAAG outputs to generate the atmospheric forcings required by WRF-Hydro and evaluate the model performance using the SAAG simulations.

--

You received this message because you are subscribed to the Google Groups "wrf-hydro_users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to wrf-hydro_use...@ucar.edu.

To view this discussion visit https://groups.google.com/a/ucar.edu/d/msgid/wrf-hydro_users/CAEA-d7_%3DTQyghwV782UtE-me40g_Ht%3DdyWZ9N2UTUvYWhUyUww%40mail.gmail.com.

[Tomásio Januário]

The correct domain file for Piracicaba region

[Tomásio Januário]

Hi Yongxin,

I hope you are doing well. Regarding my experiment: “Assessing the Performance of the WRF-Hydro Forecasting System to Predict Drought Series in the Piracicaba River Basin.”

I initially used PyWrfHydroCalib to calibrate the WRF-Hydro model for one station, AMPARO, over a two-month period (October–December 2013), using 30 iterations.

I then used the calibrated parameters to run the model for a six-month period (October 2013–April 2014) and for three subbasins/stations: AMPARO, ATIBAIA, and JAGUARIÚNA.

However, I found a significant problem when using the calibrated parameters. The simulated streamflow values are several hundred times higher than the observed values, even though the parameters were calibrated using PyWrfHydroCalib. The results are shown in the attached files.

I am therefore considering extending the calibration period and increasing the number of iterations. However, running PyWrfHydroCalib for a longer period is very computationally demanding and requires considerably more time and computational resources.

I would appreciate your advice on how to proceed. In particular, do you recommend another calibration approach or method that could be more suitable for this experiment? Alternatively, do you have any experience with running PyWrfHydroCalib for longer calibration periods or on HPC systems, and could you suggest an efficient way to perform the calibration?

Thank you very much for your time and assistance.

[aubrey]

Hi Tomásio:

We generally calibrate the hydrologic model over multiple years to expose the model to a range of different wet/dry/normal conditions. Also 30 iterations is likely not sufficient to explore the full range of parameters and parameter interactions. We generally run calibration over at least 300-400 iterations (since WRF-Hydro is an expensive model with all routing options active), though more lightweight modeling systems recommend using iteration counts in the thousands (depending on the number of parameters). Also, parameters may not transfer well from one basin to another, unless you have confidence the basins are hydrologically similar.

A couple of questions on your region of interest:

1) Are there managed reservoirs, irrigation, or other withdrawals in your watershed? If these withdrawals/storages are not represented in the model, your streamflow bias will be high.

2) Have you checked the WRF precipitation estimates against station obs to make sure they look reasonable in your region? This is always a good first check before tuning hydro model parameters. We know WRF can have significant biases, and these will lead to streamflow biases.

3) Is there significant groundwater interaction in this region? This could mean, for example, surface water losses to recharge a groundwater aquifer in mountainous areas, or discharge to the channel from groundwater in valleys.

4) If you have a persistent high bias but streamflow correlations look reasonable, you might not have enough storage in your configuration. You might try increasing porosity (smcmax) or running with a deeper soil column (i.e., > 2m) to allow more storage. Again, this should be determined based on the geology in your region. There are global depth-to-bedrock products that can help.

Let us know how it goes. It would be exciting to see the SAAG outputs used for hydrologic model applications, so I would be happy to help you get a setup that works.

Thanks!

Aubrey