Is It Normal for Nudging Results in WRF-Hydro to Be This Perfect?

[zed li]

Hello everyone,

I am currently running the nudging case in WRF-Hydro, specifically the NWM test case. The model runs normally without any errors, but the results surprised me. The simulation results with nudging are almost perfect, nearly identical to the observed data, and I haven’t even adjusted any model parameters.

When I ran the reach_lake case, the results were poor, as I expected.

 The only difference between these two cases is whether nudging is applied. Is it normal for nudging to perform so well? If I run the model on my own watershed in the future, will I achieve similar results, or is this just due to the specificity of the test watershed?

Thanks,

Zed Li

[Arezoo RafieeiNasab]

Hi Zed, 

You could check out the documentation in here: https://wrf-hydro.readthedocs.io/en/latest/nudging.html# 

Short answer is yes, with the nudging you expect an almost perfect simulation as the model simulations at the gauge locations gets updated using the nudge in equation 4.1 and becomes almost perfect. The impact of these replacements will be seen downstream of the location as the new values gets routed downstream. If you are interested, you could also look at this paper, They have used the nudging approach implemented in the NWM.

 Seo, Bong‐Chul, Witold F. Krajewski, and Felipe Quintero. "Multi‐Scale Hydrologic Evaluation of the National Water Model Streamflow Data Assimilation." JAWRA Journal of the American Water Resources Association, vol. 57, no. 6, 2021, pp. 875–884. https://doi.org/10.1111/1752-1688.12955.

Thanks!

Arezoo

--
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/beb2d083-775f-48cb-9340-1f03639c3b63n%40ucar.edu.

--

Arezoo Rafieei Nasab, Ph.D.

NCAR/RAL Project Scientist II

[zed li]

Dear Arezoo,

Thank you for your kind and insightful reply—it was truly helpful and greatly appreciated. I have carefully read the model technical description and the relevant papers you suggested, but I still have a few questions that I hope you can kindly address.

First, in the example_case/NWM, the simulation results include data from four stations. The nudgingTimeSliceObs file also contains streamflow data for four stations. Three of these stations match the data in the USGS_obs.csv file, but there is an additional station with the stationId 1374598. Based on its latitude, longitude, and streamflow data, I suspect it is an upstream station within the test basin. Why does the NWM include one more station than the observed data? Is there a specific purpose behind this?

Second, according to the paper, the effectiveness of nudging is influenced by the number of upstream assimilated stations and the controlled area. If my goal is to optimize and evaluate the simulated streamflow at the basin outlet, should I only assimilate observed streamflow data from upstream stations into the model—and not include observed data from the outlet station itself? Additionally, does this mean the simulation results at the upstream stations will align almost perfectly with the observations? Is my understanding correct?

I look forward to your reply. Once again, thank you sincerely for your generosity and support. 

Best regards,
Zed Li

[Arezoo RafieeiNasab]

Hi Zed, 

My response is embedded in your text below: 

Thanks!

Arezoo

On Thu, Sep 11, 2025 at 6:24 AM zed li <ddlw...@gmail.com> wrote:

Hi Arezoo,

Thank you for your kind and insightful reply—it was truly helpful and greatly appreciated. I have carefully read the model technical description and the relevant papers you suggested, but I still have a few questions that I hope you can kindly address.

First, in the example_case/NWM, the simulation results include data from four stations. The nudgingTimeSliceObs file also contains streamflow data for four stations. Three of these stations match the data in the USGS_obs.csv file, but there is an additional station with the stationId 1374598. Based on its latitude, longitude, and streamflow data, I suspect it is an upstream station within the test basin. Why does the NWM include one more station than the observed data? Is there a specific purpose behind this?

If I am not mistaken that gauge is outside the contributing area to the gage 0137462010 and hence not shown in the map below. Note the domain is a rectangle, so outside the basin boundary shown below there are still river reaches that are not shown in this figure.  

Second, according to the paper, the effectiveness of nudging is influenced by the number of upstream assimilated stations and the controlled area. If my goal is to optimize and evaluate the simulated streamflow at the basin outlet, should I only assimilate observed streamflow data from upstream stations into the model—and not include observed data from the outlet station itself?

 Depends on the purpose. We do nudging with the goal of improved forecasts, so we perform nudging at all locations including the downstream location. In this case the verification should be on the forecasts. Another type of verification is in space similar to what paper discussed, in that case you could assimilate at the upper gages and verify the impact on the downstream location (so not assimilating the outlet location). It depends on you and how you would like to design your experiment. 

Additionally, does this mean the simulation results at the upstream stations will align almost perfectly with the observations? Is my understanding correct?

Yes, that is correct.  

I look forward to your reply. Once again, thank you sincerely for your generosity and support.

Best regards,
Zed Li

[Arezoo RafieeiNasab]

No worries Zed, I figured out it was a glitch! I am glad it was helpful!

Good luck with your experience!

Arezoo

On Thu, Sep 11, 2025 at 6:18 PM zed li <ddlw...@gmail.com> wrote:

Hi Arezoo,

Thank you so much for your detailed and helpful reply—it has provided me with much greater clarity.

I also want to apologize if you received duplicate emails from me earlier. It appeared on my end that two previous messages, which included attached images, were automatically blocked or deleted by the system. I’m sorry for any inconvenience this may have caused.

Your guidance has been incredibly valuable to my work and understanding.

Thanks!

Zed Li

[zed li]

Hi Arezoo,

Thank you so much for your detailed and helpful reply—it has provided me with much greater clarity.

I also want to apologize if you received duplicate emails from me earlier. It appeared on my end that two previous messages, which included attached images, were automatically blocked or deleted by the system. I’m sorry for any inconvenience this may have caused.

Your guidance has been incredibly valuable to my work and understanding.

Thanks!

Zed Li

[zed li]

Dear Arezoo,

Thank you for your kind and insightful reply—it was truly helpful and greatly appreciated. I have carefully read the model technical description and the relevant papers you suggested, but I still have a few questions that I hope you can kindly address.

First, in the example_case/NWM, the simulation results include data from four stations. The nudgingTimeSliceObs file also contains streamflow data for four stations. Three of these stations match the data in the USGS_obs.csv file, but there is an additional station with the stationId 1374598. Based on its latitude, longitude, and streamflow data, I suspect it is an upstream station within the test basin. Why does the NWM include one more station than the observed data? Is there a specific purpose behind this?

Second, according to the paper, the effectiveness of nudging is influenced by the number of upstream assimilated stations and the controlled area. If my goal is to optimize and evaluate the simulated streamflow at the basin outlet, should I only assimilate observed streamflow data from upstream stations into the model—and not include observed data from the outlet station itself? Additionally, does this mean the simulation results at the upstream stations will align almost perfectly with the observations? Is my understanding correct?

I look forward to your reply. Once again, thank you sincerely for your generosity and support. 

Best regards,
Zed Li

在2025年9月11日星期四 UTC+8 00:04:48<Arezoo RafieeiNasab> 写道：

[zed li]

Hi Arezoo,

Thank you for your kind and insightful reply—it was truly helpful and greatly appreciated. I have carefully read the model technical description and the relevant papers you suggested, but I still have a few questions that I hope you can kindly address.

First, in the example_case/NWM, the simulation results include data from four stations. The nudgingTimeSliceObs file also contains streamflow data for four stations. Three of these stations match the data in the USGS_obs.csv file, but there is an additional station with the stationId 1374598. Based on its latitude, longitude, and streamflow data, I suspect it is an upstream station within the test basin. Why does the NWM include one more station than the observed data? Is there a specific purpose behind this?

Second, according to the paper, the effectiveness of nudging is influenced by the number of upstream assimilated stations and the controlled area. If my goal is to optimize and evaluate the simulated streamflow at the basin outlet, should I only assimilate observed streamflow data from upstream stations into the model—and not include observed data from the outlet station itself? Additionally, does this mean the simulation results at the upstream stations will align almost perfectly with the observations? Is my understanding correct?

I look forward to your reply. Once again, thank you sincerely for your generosity and support.

Best regards,
Zed Li

在2025年9月11日星期四 UTC+8 00:04:48<Arezoo RafieeiNasab> 写道：