spec_bdy_width question

Topics specifically related to the real.exe program
Post Reply
stormchasegenie
Posts: 37
Joined: Wed Oct 10, 2018 5:42 pm

spec_bdy_width question

Post by stormchasegenie » Wed Jul 21, 2021 12:13 am

Greetings,

I am trying to dampen boundary artifacts in my inner nest that are originating due to the fact that my one-way nest's southern boundary intersects with complex orographic features (it cannot be helped).

As such, I changed my spec_bdy_width from 5 to 10 points, increasing my relax zone from 4 to 7 and my spec_zone from 1 to 3. This helped a lot in reducing boundary artifacts that were quite intrusive to the nest's interior.

To understand what is happening, I ran an additional test in which spec_bdy_width = 10, but now spec_zone and relax_zone = 1 and 9 respectively. I got the exact same result as the previous experiment (relax_zone = 7, spec_zone = 3). This should not be the case with linear damping of the LBCs following Davies and Turner (1977), as the weights should be changed between the experiments. I am using wrfinput* files in ndown, which uses spec_bdy_width = 5, relax_zone = 4, and spec_zone = 1, but this shouldn't matter.

Thoughts?

Thanks,
-Stefan

Ming Chen
Posts: 1656
Joined: Mon Apr 23, 2018 9:42 pm

Re: spec_bdy_width question

Post by Ming Chen » Fri Jul 23, 2021 9:22 pm

Stefan,
I would expect different results when the relax_zone and spec_zone are different in different experiments. I have no answer to your question at present. I will talk to ur expert and get back to you soon. Thanks for your patience.
WRF Help Desk

Ming Chen
Posts: 1656
Joined: Mon Apr 23, 2018 9:42 pm

Re: spec_bdy_width question

Post by Ming Chen » Mon Jul 26, 2021 5:23 pm

HI Stefan,
It turns out that WRF ignores the relax and spec widths, and simply takes the total. It says total-1 = spec width, and the rest is relax. This explains the behavior.
WRF Help Desk

Post Reply

Return to “real.exe”