Scheduled Downtime
On Friday 21 April 2023 @ 5pm MT, this website will be down for maintenance and expected to return online the morning of 24 April 2023 at the latest

RE: Contrasting WRF output results at higher resolution from 1.33 km - 1 & 0.8 km

This post was from a previous version of the WRF&MPAS-A Support Forum. New replies have been disabled and if you have follow up questions related to this post, then please start a new thread from the forum home page.

K

kunaldayal

Member
Hi,

I have run simulations using the WRF model for two-way nested domains for a tropical country with two major islands for the purpose of mesoscale wind resource mapping.

The 3 domains initially used was D1 = 20 km x 20 km, D2 = 4 km x 4 km and D3 = 1.33 km x 1.33 km. With this simulation configuration, the 10 m wind speed, 2 m temperature and the diurnal cycle of wind speed and temperature on the windward side for domain 3 (1.33 km x 1.33 km) are simulated similar to those observed on the ground from automatic weather stations (AWSs).

But on the leeward side stations, there is a poor simulation of the diurnal cycle of wind speed and temperature in comparison with the observations from AWSs. Also, there is overnight over-estimation of wind speed at almost all locations.

To fix the problem, I ran two test simulations over one island only by increasing the grid resolution of domain 3 to 1 km x 1 km and 0.8 km x 0.8 km.

To my surprise, these simulations at 1 km x 1 km and 0.8 km x 0.8 km actually solved the problem on the leeward side. Now, I could see a big improvement in the simulated 10 m wind speed, 2 m temperature and the diurnal cycle of wind speed and temperature on the leeward side of the island.

But, few locations (3 out of 6) on the windward of the island now have a poor simulation of the diurnal cycle of wind speed and temperature. Also, there is overnight over-prediction of wind speed for these windward side stations.

I am not sure whether such contrasting behavior is normal for the WRF model at higher resolutions of 1 km x 1 km and 0.8 km by 0.8 km. But the results are consistent for 1 km x 1 km and 0.8 km x 0.8 km grid resolutions.

I seek assistance and advice as to whether such behavior is normal for the WRF model?

And whether the 1 km x 1 km and 0.8 km x 0.8 km result consistency can be used to say that both are optimal grid resolutions?

Appreciate your assistance and advice.

Regards
Kunal
 
Kunal,
Can you please attach the namelist.input files you are using for these 3 simulations (initial, test1, and test2), and please let us know which version of the model you are running? Did you make any modifications to the WRF code, or is it the pristine out of the box code? Thanks!
 
Hi,

Please find attached the namelist.input files for the 3 simulations (initial, test1, and test2).

I used WRF v3.9.1.1 and I did not make any changes to the WRF code.

Appreciate your assistance and advice.

Thanks and Regards
Kunal
 

Attachments

  • namelist_initial-1.33km.txt
    5.1 KB · Views: 73
  • namelist_test1-1km.txt
    5.1 KB · Views: 74
  • namelist_test2-0.8km.txt
    5.1 KB · Views: 79
Hi,
I apologize for the delay in response. We got a bit backed-up during the WRF tutorial, and are now getting caught up. Can you try adding
Code: 
topo_wind = 1, 1, 1
in the physics section of the namelist and see if that makes a difference?
 
Hi,

I tried the option of using Topo_wind = 1,1,1 under the physics settings of the namelist.input file but there is no improvement in the results.

Is there anything else that I can do?

Also, is it normal for the WRF model to behave differently on either side (windward and leeward) of the islands using different domain sizes as I asked earlier?

Appreciate your advice.

Regards
Kunal
 
Hi Kunal,
I discussed this with some colleagues. They believe they have attempted to help you with this in the past, and that the underlying problem has something to do with discrepancies in your land/water masking. Was it you that they worked with? If so, is this the same problem you were previously having?
 
Hi,

I don't think I am the one your colleagues are mentioning.

Previously, I had some problems with the "mismatch between observed vs simulated diurnal cycle (https://forum.mmm.ucar.edu/phpBB3/viewtopic.php?f=48&t=5649) and the problem was solved by increasing the resolution from 1.33 km to 1 km and 0.8 km as mentioned in my first post.

Appreciate your advice.

Regards
Kunal
 
Hi,
I apologize for the delay. I wanted to try to get some more insight from my colleagues. Unfortunately sometimes it's just the nature of modeling that you may not be able to get the behavior you expect. We don't have a clear, easy solution to the problem. We can offer some suggestions that may or may not help, though:
1) You could try to turn off the cumulus scheme on the second domain (4 km) and increase the number of vertical levels. You have many levels in the lowest km, but very few aloft.
2) It may also be better to use mp_physics option 16 (the double moment version of 6 for microphysics) if you're seeing too much rain produced in the area.
3) Your domain 1 and 2 are not very large, but it is not clear whether enlarging the domains would help. It may.
 
You must log in or register to reply here.
Top