domain_tests - UK-FVCOM-Usergroup/uk-fvcom GitHub Wiki

Approach to test new domains with information on problems and potential solutions.

=== Western UK domain (ROSA domain) ===

This domain is being used for the ROSA project on aquaculture. The project requires simulations with FVCOM-FABM-ERSEM-ShellSIM. The domain covers the UK shelf excluding the northern and central North Sea. At the moment it is being forced with HYCOM at the 3 open boundaries and WRF output initialised with NCEP data.

=== Location of resources ===

The raw coastline can be found in ~pica/Data/QGIS/

Raw bathymetry can be found in ~pica/Data/POLCOMS/ or ~pica/Data/UKHO/nc/250m(500m and 1000m)

Alternatively, the scatter data as it is included in the SMS project files are in ~pica/Models/FVCOM/models/pml-aquaculture/grids/aqua_v09/scatter/

The data consists of a combination of sources including (....) but it does not include the UKHO data which is referenced to local datum for each of their scenes. It will best to reserve the UKHO data for high resolution subdomains and process only the required scenes rather than trying to process all of their raw data.

No attempt has been made at careful merging the different bathymetry sources.

=== Initial processing of raw data ===

The raw coastline has been automatically adjusted to have a max resolution of 500m with matlab scripts (XXXX).

not quite true at this state, I am actually converting an SMS grid (aqua_v09.2dm ) into a 14 grid format with

read_sms_mesh and then [Mobj] = write_admesh_mesh(Mobj,varargin)

The bathymetry was initially interpolated (if file too large it would need to be done by section) to 1000m resolution (/users/modellers/rito/Models/git/fvcom/rosa-implementation/grid/process_scatter_bathy.m). Each section was processed with adapted ROMS matlab scripts to improve the stability of the model (it uses GRID_LinProgHeuristic_rx0 from ROMS with rx0=0.08, with a bathymetry matrix x=1309 y=1278 it can take up to 3 hours to process).... Then you have to convert it to a format ADMESH can digest (xyz or ESRI/ArcGIS asc file). ASC files are apparently expected in spherical coordinates...

The coastline and bathymetry were then used to create the mesh in ADMESH.

Input files can be generated using fvcom-toolbox or PyFVCOM

=== Generating input files ===

For generating the input files we use fvcom-toolbox. The make_input.m is located in /users/modellers/pica/Models/FVCOM/models/pml-aquaculture/matlab

BC and restart data are currently located in the ./mat directory. Tides are still calculated using the TMD model. Rivers use a subset of NEMO locations and climatology flux and nutrient values.

To force ERSEM we have generated offline model output at 15min time resolution. The model has been run with FVCOM4 beta code and uses the standard advection and turbulence model. The atmospheric forcing has been generated with WRF adding the SST as input as well as NCEP simulations for the downscaling. We have tried running FVCOM with GOTM but the coupling of both models is not currently finished or maintained and there is no advection of TKE and TEPS.

The simulations with ERSEM require BC as well as initial and river values for ERSEM variables. The rivers have all nutrients as well as O3_c, O3_bioalk and O2_o (this one set to a constant value of 250). The initial values are interpolated from POLCOMS AMM model simulations with assimilation (from Stefano Ciavatta). The code for generating initial values can be found in (/users/modellers/rito/Models/git/fvcom/rosa-implementation/matlab).

The BC