Converting ROMS data to finite element grid

General scientific issues regarding ROMS

Moderators: arango, robertson

Post Reply
Message
Author
dwperkey
Posts: 2
Joined: Tue Feb 21, 2017 2:42 pm
Location: ERDC-CHL

Converting ROMS data to finite element grid

#1 Unread post by dwperkey »

Hello everyone. I am new to modeling so I apologize if this question is one that is generally known.

For my research I am in need of bringing bathy and hydro information from a ROMS model of the York River Estuary into the SMS supported Particle Tracking Model (PTM). PTM utilizes a finite element grid system, so I need a method to convert the ROMS curvilinear grid to finite element. In my search through the forum I noticed some people have mentioned utilizing ADCIRC data in their ROMS models (which I assume would require a conversion of finite element to curvilinear), but I was curious if commonly used methods/tools exist to go the opposite direction and convert the ROMS grid into finite element.

Any tips, hints, or resources would be greatly appreciated. This is my first venture into the modeling world so I'm not familiar with what might be a common practice within the community.

Thanks,

Dave

User avatar
kate
Posts: 4091
Joined: Wed Jul 02, 2003 5:29 pm
Location: CFOS/UAF, USA

Re: Converting ROMS data to finite element grid

#2 Unread post by kate »

This sort of problem is generally known as regridding or remapping. One remapping toolbox comes with ESMF which also has a Python interface. I'm sure there are Matlab tools for such things as well - we're always having to remap to get boundary conditions, for example.

User avatar
ngarg
Posts: 19
Joined: Fri Jul 12, 2013 5:46 pm
Location: NTU

Re: Converting ROMS data to finite element grid

#3 Unread post by ngarg »

i routinely do something similar, where i remap data from a structured (stencil) type grid to unstructured (finite element) mesh. I use scrip for that purpose, also as Kate mentioned you can use EMSF (ESMpy). Also, you can use ncl (NCAR) which has some functions for working with both rectilinear, and unstructured grid.

NCL under the hood uses EMSF regridgen for the same purpose. There are some tools available for doing similar things for FVCOM, you can get started by looking at how they do it. I think they use MATLAB.

https://github.com/pwcazenave/fvcom-toolbox

and there is one more library, based in python from people at Stanford, where they use SUNTANS model, which is also FEM based.

https://github.com/mrayson/soda

dwperkey
Posts: 2
Joined: Tue Feb 21, 2017 2:42 pm
Location: ERDC-CHL

Re: Converting ROMS data to finite element grid

#4 Unread post by dwperkey »

Thanks for the tips. I'll look into both.

Post Reply