Difference between revisions of "Nesting Suite"

(Imported from Wikispaces)
 
Line 1: Line 1:
 +
{{Needs Update}}
 +
[[Category: Unified Model]]
 +
 
Up-to-date information and news about the nested suite can be found on the RMED wiki and Rose issue tracker:
 
Up-to-date information and news about the nested suite can be found on the RMED wiki and Rose issue tracker:
 
* https://code.metoffice.gov.uk/trac/rmed/wiki/suites/nesting
 
* https://code.metoffice.gov.uk/trac/rmed/wiki/suites/nesting

Revision as of 02:01, 18 March 2019

Template:Needs Update This page needs updating

Up-to-date information and news about the nested suite can be found on the RMED wiki and Rose issue tracker:

Nesting inside ERA-Interim

The suite can be driven from ERA-Interim analyses with a small amount of setup.

At NCI we have 6 hourly ERA-Interim analyses in GRIB format under the directory /g/data1/ub4/erai/grib. As ERA-Interim is under licence, make sure you comply with requirements here before starting

Currently code changes required to run from ERA-Interim are in the job u-at264, which is based on the 2017 nesting suite.

Make a copy of this job with

rosie copy u-at264

Open the job in `rose edit` and make the following changes to suite conf to enable ERA-Interim driving:

  • In 'General run options' make sure SITE is NCI
  • In 'Driving model setup', set:
    • 'dm_name' to 'ECMWF forecast fields'
    • 'dm_ec_ic_lbc_dir' to '\$ROSE_DATA/erai'
    • 'dm_ec_cb_freq' to '21600' (6 hours, the ERA-Interim analysis frequency)

To run from ERA interim the suite needs to generate lateral boundary conditions from the ERA-interim GRIB files. To enable this you need to create a new 'Resolution setup' that is a bit larger than your coarsest LAM domain. This resolution won't be used to run the model, instead the ERA interim data will be interpolated to it so that the boundary conditions can be cut out for the LAM. For instance if your coarsest LAM is a 200x200 grid with resolution 0.1 x 0.1 degrees, a 220x220 grid at the same resolution would be fine - this gives a border of 10 grid points around each edge to use for the boundary conditions. You could also use less grid points at a coarser resolution, though be cautious as this adds an extra interpolation step.

You will need some extra ancillary files when using ERA-Interim, set the following in your new resolution, as well as your coarsest LAM resolution:

  • vegsoil_anc: True
  • lai_anc: True
  • alb_anc: False
  • orog_strm: False
  • 3D_ancils: True
  • sstsmc_anc: True

Set the `nmods` to 0, since we just want the ancillaries at this resolution (this will print a warning in Rose, which can be ignored)

To make the ERA interim driver use this resolution you just set up, in the 'Driving model setup' page set 'dm_ec_lam_ancil_dir}}' to the new resolution's ancillary directory. For instance, if the new resolution is `resn_2` of region `Regn1`, set `Template:Dm ec lam ancil dir` to `{{\$ROSE_DATA/ancils/Regn1/resn_2`

Note that the resolution of the ERA Interim source dataset is ~ 0.75 x 0.75 degrees, keep this in mind when choosing the resolution of your coarsest LAM (i.e. make sure your LAM isn't inside a single ERA Interim grid point!).

https://code.metoffice.gov.uk/trac/rmed/wiki/suites/nesting/ECdriver

include component="page" wikiName="climate-cms" page="Index UM" editable="1"