changing from old roms to new roms

Message

edith · #1 Unread post by **edith** » Tue Jan 16, 2007 3:30 pm

Hi,

I am trying to switch from an old version of ROMS (with “preprocessing tools” and “Run” and so on) to the new ROMS 2.2. that have a totally new structure and I am having some difficulties with that. My main problem is that I don’t understand what is replacing preprocessing part. How I am suppose to get NetCDF files if I have ascii files (it used to be make_grid, make_forcing and so on)?

Second problem I have is with compiling. I try to compile with g95. I get error:
g++: installation problem, cannot exec `f771´: No such file or directory
any ideas what could be a problem?

I apologise if my question is too trivial but I am still new to ROMS, fortran and linux and wiki ROMS (for stupid questions) still doesn't work.

jcwarner · #2 Unread post by **jcwarner** » Tue Jan 16, 2007 3:35 pm

as far as I know, the netcdf files that worked with previous versions of ROMS should work with ROMS v2.2. I am not familiar with those codes (make_*) to convert from ASCII to netcdf.

for the compiler:
sounds like a path issues. Try to find f771 (.exe??) on your system and see if it is in the path.

edith · #3 Unread post by **edith** » Tue Jan 16, 2007 3:42 pm

How do you get from your data files (usually ascii format?) netcdf files?
as much as i can understand from this program i need to provide some netcdf files in ocean.in file.

there are no f771 files in my computer not .exe or any other

jcwarner · #4 Unread post by **jcwarner** » Tue Jan 16, 2007 3:55 pm

i do not use g95, so maybe someone else can help with that.

For the forcing files- You can use the ROMS forcing packages as provided on the main web site. This would be a good place to start. We have many (not well documented but they work) application specific m files, as I am sure many other people have.

jacopo · #5 Unread post by **jacopo** » Tue Jan 16, 2007 4:06 pm

the only difference between roms-2.1 and roms 2.2 is that some variables have been renamed for CF compliancy. In particular, s_rho and s_w. But this is an easy fix. Not really relevant differences.
The preprocessing tool is somewhat outside ROMS, in the sense that it is not released within each ROMS-x.x tar.
To create your own NetCDF input files, you can use Hernan's packages
http://www.myroms.org/index.php?page=RomsPackages
or
http://www.myroms.org/index.php?page=Processing
or other freely available resources. However, you are supposed to change these scripts depending on your application anyway.

edith · #6 Unread post by **edith** » Tue Jan 16, 2007 4:20 pm

How i know what should be in my ocean_*.nc files? This packages seams to provide single field netcdf files. How do i know what exactly must be in them? And do they have to be already interpolated to the grid or can roms do it?

jacopo · #7 Unread post by **jacopo** » Tue Jan 16, 2007 4:32 pm

To answer to this question is a little bit a pain in the ass

check this out first:
viewtopic.php?t=63

Of course what you need to provide in the input NetCDF files strictly depends on your application.

You have to provide al vectors ALREADY rotated onto ROMS grid, tidal currents excluded (for some hystorical reasons unknown to me, tidal currents are rotated by ROMS when defining CURVGRID).

evan · #8 Unread post by **evan** » Wed Jan 17, 2007 5:43 pm

Edith, from your original post it sounds as if you have switched from the Roms Agrif version of ROMS to the Rutgers (2.2) version. This would explain why you see a different directory structure, and why you don't have the matlab scripts make_grid.m etc. If this is so then you haven't switched to a 'newer' version of ROMS, you've just switched to a different one.
If this was your intention that's ok but, particularly as you say you are new to ROMS, it means you are going to be a bit lost for a while with the new version.
I can't give you advice about which version to use; I use Agrif because that's the version that most of the people that I have contact with use.

Hope that helps, Evan

inga · #9 Unread post by **inga** » Thu Jan 18, 2007 8:28 am

Hi Edith,

I'm using roms-2.2 for which I have adopted the Pierrick Penven's Roms_tools (Agrif) toolbox. Now it evolved into my-own-a-very user-specific toolbox but if you want I can send you the roms-2.2 version of create_inifile.m to create an initial file (just write me on koszalka@isac.cnr.it).
Create_inifile.m is called by make_clim.m in the Preprocessing toolbox. I don't have an equivalent program for forcing or oa because I don't need it for my idealistic configuration.

edith · #10 Unread post by **edith** » Thu Jan 18, 2007 9:14 am

So there is two different versions of roms? I kind of understood that one had developed out from other?
But all questions and everything in forum seams to be about roms-2.2 and i can find hardly any info about roms_agrif version besides? i do have matlab scripts just that roms-2.2 do not seams to contain any that kind of tool and i could not understand why. So I am suppose to use roms_agrif prepossessing to great netcdf files for roms-2.2?

jacopo · #11 Unread post by **jacopo** » Thu Jan 18, 2007 10:08 am

There are nowadays 3 actually different version of ROMS -as far as I know-

1) ROMS "eastern coast" mainly managed by RUTGERS but with substantial contributions by several other groups worldwide.

http://www.myroms.org/

2) ROMS "western coast" (developed independently since ROMS-2.1?) by UCLA

http://www.atmos.ucla.edu/cesr/ROMS_page.html

3) ROMS-AGRIF, developed independently I believe since ROMS-2.0 or older

http://www.brest.ird.fr/Roms_tools/

The three versions share many common features but not everything.

I have to say that I believe you are supposed to develop your own preprocessing tool starting from the packages
http://www.myroms.org/index.php?page=RomsPackages
and
http://www.myroms.org/index.php?page=Processing
or from roms_agrif tool as well or whatever.

kate · #12 Unread post by **kate** » Thu Jan 18, 2007 6:47 pm

Yes, there are three ROMS branches and this forum is hosted by the Rutgers branch, though some readers are familiar with the others as well.

Perhaps the thing to do is to describe what it is you want to be doing in more detail so that we can help you with the preprocessing. For instance, if you want the CORE winds from the CCSM folks for the years 1958-2004, I know something about that. I also know a little about the global tidal model from OSU and something about the sea ice.