Increase Bottom Vertical Grid Resolution

General scientific issues regarding ROMS

Moderators: arango, robertson

Post Reply
Message
Author
osean
Posts: 47
Joined: Sat Jan 23, 2021 3:46 pm
Location: MIT-WHOI

Increase Bottom Vertical Grid Resolution

#1 Unread post by osean »

Hello ROMS experts

I am currently looking for a way to modify the vertical grid scheme so that I can have increased bottom vertical grid numbers to resolve the bottom mixed layer, for example. Ideally, I would like to have a grid that has high resolution for the surface mixed layer and the bottom mixed layer, with relatively sparse points in the interior.

So I thought of starting with a vertical grid scheme similar to that of the UPWELLING case, with a seemingly exponential increase in grid spacing from the surface to the bottom. Then I would like to modify it so that I can have an exponential decrease in grid spacing from the middle of the interior to the bottom. In that case, I can have a vertical grid that have high resolution near the surface and the bottom, but with a relatively low resolution in the interior.

Now I wonder how to proceed with such a change? I looked through the ROMS manual (2018 version) and files in trunk folder, but achieved little success. Perhaps I am missing out something obvious, or due to my poor ability to disentangle the information in the code. From the manual, I found the following initial set-up files are associated with the grid:

>> ana_grid Compute the grid(s) internally.
>> get_grid Read in the curvilinear coordinate arrays as well as f and h from one NetCDF file per grid.
>> mod_grid.F This provides the storage for the model grid fields.

These appear to be relevant. So I went into the \trunk file in the ROMS folder, and I went through ana_grid.h, it doesn’t show where I can modify the vertical grid spacing/scheme. So I feel mod_grid.F might be where the information is hidden, but I can't quite figure out where it is embedded within the code (if it is).

Appreciate any advice on how to proceed with this! many thanks

jcwarner
Posts: 1200
Joined: Wed Dec 31, 2003 6:16 pm
Location: USGS, USA

Re: Increase Bottom Vertical Grid Resolution

#2 Unread post by jcwarner »

here is a good start:
https://www.myroms.org/wiki/Vertical_S-coordinate

there is also an m file "roms_getnplot_grid_rutgers" that you can use to see what an adjustment to the various parameters will do to the vertical stretching.
-j

osean
Posts: 47
Joined: Sat Jan 23, 2021 3:46 pm
Location: MIT-WHOI

Re: Increase Bottom Vertical Grid Resolution

#3 Unread post by osean »

jcwarner wrote: Mon Apr 12, 2021 5:08 pm here is a good start:
https://www.myroms.org/wiki/Vertical_S-coordinate

there is also an m file "roms_getnplot_grid_rutgers" that you can use to see what an adjustment to the various parameters will do to the vertical stretching.
-j
Oh this is very helpful, thank you!

c.drinkorn
Posts: 110
Joined: Thu Mar 08, 2018 2:47 am
Location: German Research Centre for Geosciences

Re: Increase Bottom Vertical Grid Resolution

#4 Unread post by c.drinkorn »

HI Osean,

our application has a vertical grid like the one you are aiming at. We achieved this with the vertical transformation formulation 2 and the vertical stretching formulation 3 with theta_b = 1.5 and theta_s = 1.0. Our vertical resolution ranges from 0.3m in places at the bottom and surface to 240m in the interior. To "stabilize" this a bit we set Tcline = 100.
But I also encourage you to read about the s-coordinate like suggested by John. It's definitely worth understanding what's happening! :) Good luck!

Cheers,
Cate

osean
Posts: 47
Joined: Sat Jan 23, 2021 3:46 pm
Location: MIT-WHOI

Re: Increase Bottom Vertical Grid Resolution

#5 Unread post by osean »

c.drinkorn wrote: Tue Apr 13, 2021 8:38 am HI Osean,

our application has a vertical grid like the one you are aiming at. We achieved this with the vertical transformation formulation 2 and the vertical stretching formulation 3 with theta_b = 1.5 and theta_s = 1.0. Our vertical resolution ranges from 0.3m in places at the bottom and surface to 240m in the interior. To "stabilize" this a bit we set Tcline = 100.
But I also encourage you to read about the s-coordinate like suggested by John. It's definitely worth understanding what's happening! :) Good luck!

Cheers,
Cate
Hi Cate,

Thank you!! This is very helpful. I tried to configure with the default Vtransform == 2 and Vstretching == 4, but my best attempt so far could only get me down to approx. 12 m resolution at the bottom. I will attempt your values.

Post Reply