search

JMMP-Group / SEVERN-SWOT

Severn estuary 500m ocean model
MIT License
1 stars 2 forks source link

tide boundary generation #20

Closed mpayopayo closed 3 years ago

mpayopayo commented 3 years ago

From @micdom: @jpolton @mpayopayo I'm back to this: what are the files you're listing in namelist*.bdy for sn_src_hgr, sn_src_zgr, sn_src_msk? I don't have access to the /projectsa/accord folder (and probably I shouldn't to run a different configuration).

I think those files contain the domain from which the tides are picked up to be generated into our domain, maybe @jpolton can confim? Maybe there could be a copy of those files in projectsa/SWOT and pick them from there?

micdom commented 3 years ago

@jpolton @mpayopayo yes, I don't have access to accord, so I can't see how a file inputs_src_zgr.ncml looks like. Having a local copy of the SWOT files that I can access would be just perfect.

jpolton commented 3 years ago

@mpayopayo @micdom Good point. The IMMERSE build of PyNEMO that we use did not expect people to not actually want T and S data, so when it runs it expects certain grid files to exist, even though it doesn't use them. (This is a reported 'bug' that we need to work around for now). So the various files that you can not read are grid files to (for the Indian Ocean). My thinking was that we would replace them with Severn relevant ones when we got to using T and S at the boundaries. In the meanwhile I will put these "accord" files on /scratch

jpolton commented 3 years ago

@mpayopayo @micdom I am hoping that these are the only files you need...

/scratch/jelt
 ls -l SEVERN-SWOT_accord_files
total 449120
-rw-r-x---. 1 jelt msm   9024716 Jul 20 13:47 CMEMS_subdomain_coordinates.nc
-rw-r-x---. 1 jelt msm 450803004 Jul 20 13:46 CMEMS_subdomain_mask.nc
-rwxr-x---. 1 jelt msm       310 Jul 20 13:45 inputs_src_zgr.ncml
micdom commented 3 years ago

@jpolton still missing files... in namelist_FES14.bdy there is sn_src_dir = './NCML/CMEMS_2005.ncml' and in that file there is 

<aggregation type="union" >
     <netcdf xmlns="http://www.unidata.ucar.edu/namespaces/netcdf/ncml-2.2">
        <aggregation type="joinExisting" dimName="time" >
           <scan location="/projectsa/accord/SANH_jenjar/CMEMS_SANH/CMEMS_data/2005/"  regExp=".*download.nc$" />
        </aggregation>
     </netcdf>
     <netcdf xmlns="http://www.unidata.ucar.edu/namespaces/netcdf/ncml-2.2">
        <aggregation type="joinExisting" dimName="time" >
           <scan location="/projectsa/accord/SANH_jenjar/CMEMS_SANH/UV_components/2005/"  regExp=".*download_UV.nc$" />
        </aggregation>
     </netcdf>
  </aggregation>

I don't have access to /projectsa/accord/SANH_jenjar/CMEMS_SANH/

micdom commented 3 years ago

@jpolton with the access to the accord folder it works. I've played a bit to see what are the files that the scripts really wants (relevant for porting to archer2):

micdom commented 3 years ago

hello @jpolton @mpayopayo, I think I have got the model running with tides for 24 hours!

Screenshot 2021-07-23 at 13 12 42
jpolton commented 3 years ago

@jpolton with the access to the accord folder it works. I've played a bit to see what are the files that the scripts really wants (relevant for porting to archer2):

  • the link to ln -s /projectsa/accord/SANH_jenjar/CMEMS_SANH/CMEMS_subdomain_coordinates.nc and it has to be that file, if I link another file with the same name it doesn't work.
  • access to /projectsa/accord/SANH_jenjar/CMEMS_SANH/
  • in the namelist_FES14.bdy I've changed sn_src_hgr and sn_src_msk (see below) and it seems to work fine. The sn_src_zgr is from accord and I've copied it locally.
sn_src_hgr = '/projectsa/NEMO/nibrun/ARCHER_DATA/INPUTS/AMM15/coordinates.nc'   !  parent /grid/
   sn_src_zgr = 'inputs_src_zgr.ncml'   !  parent
   sn_dst_hgr = '../DOMAIN/domain_cfg.nc'  ! child
   sn_dst_zgr = 'inputs_dst.ncml' ! rename output variables
   sn_src_msk = '/projectsa/NEMO/nibrun/ARCHER_DATA/INPUTS/AMM15/amm15_mask.nc'      ! parent
   sn_bathy   = '../DOMAIN/bathy_meter.nc'        ! child

@jpolton with the access to the accord folder it works. I've played a bit to see what are the files that the scripts really wants (relevant for porting to archer2):

  • the link to ln -s /projectsa/accord/SANH_jenjar/CMEMS_SANH/CMEMS_subdomain_coordinates.nc and it has to be that file, if I link another file with the same name it doesn't work.
  • access to /projectsa/accord/SANH_jenjar/CMEMS_SANH/
  • in the namelist_FES14.bdy I've changed sn_src_hgr and sn_src_msk (see below) and it seems to work fine. The sn_src_zgr is from accord and I've copied it locally.
sn_src_hgr = '/projectsa/NEMO/nibrun/ARCHER_DATA/INPUTS/AMM15/coordinates.nc'   !  parent /grid/
   sn_src_zgr = 'inputs_src_zgr.ncml'   !  parent
   sn_dst_hgr = '../DOMAIN/domain_cfg.nc'  ! child
   sn_dst_zgr = 'inputs_dst.ncml' ! rename output variables
   sn_src_msk = '/projectsa/NEMO/nibrun/ARCHER_DATA/INPUTS/AMM15/amm15_mask.nc'      ! parent
   sn_bathy   = '../DOMAIN/bathy_meter.nc'        ! child

Good job @micdom. These accord file issues will be resolved when the boundaries are replaced with suitable T,S,u,v ones for this domain. A decision needs to be made about which parent model to use. @mpayopayo has done some thinking on this.

So I think that this ticket "tide boundary generation" is closed too, unless there are further mods to the wiki and scripts required to get this joy. (Please re-open if you feel this is not closed)