Manually changing Channel parameters and REFKDT

166 views
Skip to first unread message

Md Murad Hossain Khondaker

unread,
Dec 3, 2021, 11:49:59 AM12/3/21
to wrf-hydro_users
Hi,

I am trying to do a sensitivity test based on REFKDT and mannings roughness. I changed REFKDT (1~5) in GENPARM.TBL and mannings roughness (0.01~1) in ROUTELINK.nc (using python). However, there is no significant change in streamflow is found. 

Do I need to change anything additionally to change these parameters? Am I missing something? I am changing the values for domain that I got using CUAHSI subsetter.

Thanks in advance.

Murad.


Figure 2021-12-03 104830.png

willrudisill

unread,
Dec 3, 2021, 11:58:30 AM12/3/21
to wrf-hydro_users, Md Murad Hossain Khondaker
Have you complied WRF-Hydro with the "spatial soil" option? If so then you would need to change that parameter in the soil_properties.nc file, since the .tbl file is not being read. 

(I'm fairly certain) 


Md Murad Hossain Khondaker

unread,
Dec 6, 2021, 11:57:58 AM12/6/21
to wrf-hydro_users, willrudisill, Md Murad Hossain Khondaker
Hi,

Thanks a lot for the reply. I have changed soil_properties.nc. However, the sensitivity is still low. Is there any additional file that I need to modify for REFKDT?

And for Manning's roughness, is editing only 'n' in the Route_link.nc file is enough? Do I need to change anything else?

Thanks in advance,
Murad.

Juan Carlos Tufino

unread,
Jan 9, 2025, 11:28:18 AMJan 9
to wrf-hydro_users, Md Murad Hossain Khondaker

Hi Murad,

I hope you're doing well. I wanted to share that I have experienced a similar issue when modifying the Route_Link.nc file. Despite making changes to the Manning’s roughness values, I did not observe any significant change in streamflow either.

To further investigate, I even disabled the reading of the Route_Link.nc file, and surprisingly, the results remained the same—as if the file wasn’t being read at all, despite not having explicitly commented it out in the configuration.

This makes me wonder if there might be another process or internal setting that overrides the values or bypasses the use of Route_Link.nc. If you have found any workaround or additional insights, I would appreciate your feedback.

Thank you for raising this point—it's helpful to know others are encountering the same challenge.

Best regards,
Juan Carlos Tufino Bernuy

Arezoo RafieeiNasab

unread,
Jan 9, 2025, 1:21:23 PMJan 9
to wrf-hyd...@ucar.edu, Md Murad Hossain Khondaker
Hi, 

May I ask which channel option you are using? The option specified in the hydro.namelist?

! Specify channel routing option: 1=Muskingam-reach, 2=Musk.-Cunge-reach, 3=Diff.Wave-gridded
channel_option = 2

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/cb3be55e-9df3-4e1f-ad33-caf76f181449n%40ucar.edu.


--
Arezoo Rafieei Nasab, Ph.D.
NCAR/RAL Project Scientist II
office: 303-497-2888

Juan Carlos Tufino

unread,
Jan 9, 2025, 3:35:54 PMJan 9
to wrf-hydro_users, Arezoo RafieeiNasab, Md Murad Hossain Khondaker

Dear Dr. Arezoo,

Thank you for your message!

I am currently using channel option 3 (Diffusive Wave – gridded routing). However, when I used channel option 2 (Musk.-Cunge-reach), I encountered the following issue during the execution:

========================================================
(base) juantuf@JuanTuf:/home/soil_no_mod-5.2.0/trunk/NDHMS/Run$ ./wrf_hydro_NoahMP.exe  
Calling config noahlsm_offline  
wrfinput_flnm: './DOMAIN/wrfinput_d04.nc'  
xstart,xend, ystart, yend  1  102  1  102  
Found restart file:  './RESTART/RESTART.2024010100_DOMAIN4'  
...  
read file to get NLINKSL from ./DOMAIN/Route_Link.nc  
Number of Segments or Links on sparse network: 2839  
...  
**WARNING:** `get2d_real`: failed to find the variables: `CHAN_DEPTH` and `CHAN_DEPTH`  
...  
**Error:** double free or corruption (!prev)  
**Program received signal SIGABRT: Process abort signal.**  
========================================================

This error seems to be related to missing variables (CHAN_DEPTH) in the Route_Link.nc or an issue with memory allocation during channel routing initialization.

If you have any insights on how to handle this or have faced a similar problem, I would appreciate any suggestions!

Kind regards,
Juan Carlos

Arezoo RafieeiNasab

unread,
Jan 9, 2025, 3:46:36 PMJan 9
to Juan Carlos Tufino, wrf-hydro_users, Md Murad Hossain Khondaker
Hi Juan, 

If you use channel option 3, the routing would be Diffusive Wave and the manning n values will be read from the CHANPARM.TBL. The Routlink file is not used in this configuration, and that explains why you did not get a different answer when you commented out this file. If you want to look into the calibration of mannin in the gridded routing, you want to change the n in the CHANPARM.TBL file. Note that the n is a function of stream order in this configuration, and therefore you could have only ONE value per stream order per your domain. 

Regarding the error you got when using option 3, I have not seen that before. Did you prepare the domain using the CUASHI subsetter? 

Thanks!
Arezoo

Juan Carlos Tufino

unread,
Jan 9, 2025, 3:58:39 PMJan 9
to wrf-hydro_users, Arezoo RafieeiNasab, wrf-hydro_users, Md Murad Hossain Khondaker, Juan Carlos Tufino

Thank you very much for your prompt response. In this case, I used option 2, and I encountered that error. Additionally, I did not use the CUAHSI Subsetter. I am working in the Peruvian Andes.

Best regards,
Juan Tufino

Arezoo RafieeiNasab

unread,
Jan 9, 2025, 4:52:05 PMJan 9
to Juan Carlos Tufino, wrf-hydro_users, Md Murad Hossain Khondaker
Unfortunately, I cannot tell what might be the problem from the error message for using option 2. I assume you used the GIS pre-processing to generate the Routelink, right? 
Thanks!
Arezoo

Juan Carlos Tufino

unread,
Jan 9, 2025, 10:02:23 PMJan 9
to wrf-hydro_users, Arezoo RafieeiNasab, wrf-hydro_users, Md Murad Hossain Khondaker, Juan Carlos Tufino

Thank you for your response. Yes, I confirm that I used the GIS pre-processing tool compatible with WRF-Hydro version 5.2.0 to generate the Routelink file. I would really like to continue using the Routelink file, so I appreciate any guidance you can provide to resolve this issue.

Best regards,

Juan

Arezoo RafieeiNasab

unread,
Jan 10, 2025, 11:40:30 AMJan 10
to Juan Carlos Tufino, wrf-hydro_users, Md Murad Hossain Khondaker
Hi Juan, 

Could you share the files of your domain, I could try it on my end. Not sure if I could find the problem, but could give it a try. 

Thanks!
Arezoo

aubrey

unread,
Jan 10, 2025, 11:41:45 AMJan 10
to wrf-hydro_users, Juan Carlos Tufino, Arezoo RafieeiNasab, wrf-hydro_users, Md Murad Hossain Khondaker
Hi Juan Carlos:
Would you be able to share your namelists? Not all options are compatible with each other and not all domain files and configurations are compatible, so that might help us diagnose your issue.

I'll also note that refkdt is I believe only used in Noah-MP's RUNOFF_OPTION 3, so if you are using another option it will not change your results.

Thanks!
Aubrey

Message has been deleted

aubrey

unread,
Jan 14, 2025, 7:27:00 PMJan 14
to wrf-hydro_users, Juan Carlos Tufino
Hi Juan Carlos:
A couple of notes about your domain/setup:

1) Your hydro.namelist list a dxrt = 500, but I think your domain files show this should be 50 (this is your routing grid resolution). You also might want to check the documentation to make sure your specified timesteps (e.g., for terrain routing) are appropriate for your 50-m grid spacing.
2) Your Fulldom file is showing CHANNELGRID values of -1 for channel cells. This is causing a crash for the reach-based routing. Was this file created for reach-based routing or gridded channel routing? I was testing with the latest code and needed to change the channel cell values to 0 to get the reach-based channel routing to run. I'm checking with the GIS Preprocessing toolkit expert to understand the difference in how channel cells are specified.
3) Since you are using the MODIS land cover classification, make sure you are running the model using the HYDRO_MODIS.TBL file (located in the template/HYDRO directory with the code - you will need to copy this file into your run directory and rename it HYDRO.TBL). It looks like you were running with the HYDRO.TBL that works with the USGS land cover classification system.

In general I'm not sure if it is OK to run the diffusive wave gridded channel routing scheme using files that were created for the reach-based routing method, or mix-and-match files between the two setups. It is probably safer to use files created for the gridded channel routing configuration for that physics setting.  

At any rate - I was able to run the model successfully using Muskingum-Cunge channel routing with the changes mentioned above.

Hope that helps.

Thanks!
Aubrey




On Saturday, January 11, 2025 at 3:41:35 PM UTC-5 Juan Carlos Tufino wrote:

Hi Dr. Aubrey,

Thank you again for your support. I’ve attached my namelists for you to review and help identify any compatibility issues.

Currently, I have channel_option = 3 set in my hydro.namelist. However, when I changed it to channel_option = 2 (Musk.-Cunge-reach) to use the Route_Link.nc file generated with the WRF-Hydro GIS toolkit, keeping everything else constant, I encountered an error during the execution. I’ve attached the error log for you to review.

Additionally, you mentioned RUNOFF_OPTION = 3, and I’d like to review the related documentation to better understand which variables and processes are activated when using the following configuration:

  • DYNAMIC_VEG_OPTION = 4
  • CANOPY_STOMATAL_RESISTANCE_OPTION = 1
  • BTR_OPTION = 1
  • RUNOFF_OPTION = 3
  • SURFACE_DRAG_OPTION = 1
  • FROZEN_SOIL_OPTION = 1
  • SUPERCOOLED_WATER_OPTION = 1
  • RADIATIVE_TRANSFER_OPTION = 3
  • SNOW_ALBEDO_OPTION = 2
  • PCP_PARTITION_OPTION = 1
  • TBOT_OPTION = 2
  • TEMP_TIME_SCHEME_OPTION = 3
  • GLACIER_OPTION = 2
  • SURFACE_RESISTANCE_OPTION = 4

If you have any specific documentation or recommendations regarding how these options influence the variables and simulation results, I’d really appreciate it.

Thank you again for your time and guidance. I look forward to your comments and suggestions.

Best regards,
Juan Carlos

Kevin Sampson

unread,
Jan 15, 2025, 10:43:26 AMJan 15
to wrf-hyd...@ucar.edu, Juan Carlos Tufino
Aubrey,

I wanted to comment that the value of -1 in the CHANNELGRID means an inactive channel cell. The GIS Preprocessing tools allow users to "deactivate" channel cells outside of forecast basins in order to reduce computation and thin the outputs (CHRTOUT*). However, the default option is to create active channels (value=0) for the full domain.

Thanks,

Kevin

aubrey

unread,
Jan 15, 2025, 10:50:59 AMJan 15
to wrf-hydro_users, Kevin Sampson, Juan Carlos Tufino
Thank you, Kevin! That explains why it wasn't working properly. I'll also note that the basin mask in the domain is showing the basin to only be a small section of the domain in the upper corner, so my guess is these issues are related.

So Juan Carlos - you may want to check how you are defining the basin in the GIS pre-processing tool.

Thanks!
Aubrey

To unsubscribe from this group and stop receiving emails from it, send an email to wrf-hydro_users+unsubscribe@ucar.edu.

Juan Carlos Tufino

unread,
Jan 16, 2025, 12:08:05 PMJan 16
to wrf-hydro_users, aubrey, Juan Carlos Tufino

Dear Dr. Aubrey,

I want to express my sincerest gratitude for your observations and recommendations. I had been stuck at this stage for quite some time, and thanks to your valuable guidance, I have been able to significantly improve my model configuration and make progress in my project.

Currently, I am evaluating a single forecast point located in the upper corner of the domain, which, as you pointed out, might have limited the activation of channel cells. I am considering changing the CHANNELGRID values to 0 to activate the Route_Link and enable further calibrations. However, I am also reflecting on whether it might be more appropriate to keep the current setup, opt for option 3 = Diffusive Wave-gridded, and perform calibrations using table files like CHANPARM.TBL.

I deeply appreciate your guidance, which has been crucial in helping me take this important step. If you have any additional recommendations regarding these alternatives, they would be of great help as I continue to refine the project.

Best regards,
Juan Carlos


Reply all
Reply to author
Forward
0 new messages