Some of the options define the behaviour of the downloader, others define locations to save files and others detail models
and grid files to download. Finally the spatial extent to download is also required.
I/O and NCML file
-------------------------
The location of the NCML file is listed a string defining the source directory or "sn_src_dir". The output folder is also
defined here as "sn_dst_dir", **NOTE** if this directory does not exist it will need to be created and permissoned correctly
for PyNEMO to run properly. The NCML file details the input files to agregate and what the variable names are. This file
can be generated using the ncml_generator, with variable names found using the CMEMS catalogue. https://resources.marine.copernicus.eu/?option=com_csw&task=results
For more information please read the ncml generator page. The example CMEMS.ncml file includes: temperature, SSH and U and V components of ocean currents.
Firstly, the string "sn_fn" defines the prefix for the output files. The number "nn_fv" defines the fill value, and the number
"nn_src_time_adj" defines the source time adjustment. The rest of the boxes are CMEMS specific.
Data Source Configuration
--------------------------
The first section defines the CMESM data source configuration. The boolean "ln_use_cmems" when set to true will use the
CMEMS downloader function to download the requested data, this is defined in the ncml file which can be generated using the
NCML generator. Among other things this file defines what data variables to download. This term also changes the variable
names to CMEMS specific ones e.g. thetao for temperature and so for salinity. This is in contrast to the NEMO specific ones
such as Votemper and Vosaline. When set to false no download occurs and variable names are kept to NEMO specific.
MOTU Configuration
-------------------
In the next section when set to true "ln_download_cmems" will download the boundary tracer data, e.g. time series of temperature and saliniy.
When set to false PyNEMO will skip this download. The string "sn_cmems_dir" defines where to save these downloaded files.
PyNEMO requires grid data, this isn't possible to download using the same method as the tracer data which uses the MOTU
python client. To get the grid data, an ftp request is made to download the global grids which are then subset to the relevent
size. The booleans "ln_downlad_static" and "ln_subset_static" determine this behavior. Finally there is an int named
"nn_num_retry" this defines the number of times to retry downloading the CMEMS data. The data connections are prone to failure
so if a non critical error occurs the function will automatically try to redownload. This int defines how many times it will
try to do this. Typically this static data and subsetting are only required once so these can be set to true for first download
and then set to false when more time series data is required.
As mentioned previously, the time series boundary data is downloaded using MOTU, this is an efficent and robust web server that
handles, extracts and transforms oceanographic data. By populating a configuration file, this can be sent to the MOTU server
which will return the requested data in the requested format. The section CMEMS MOTU configuration sets this up. Most of these
options should not need changing. The location of the MOTU server for CMEMS is defined here, and the location of the config
template file and also the location of the config file to submit. The only options that should require changing are the model,
product and prefix options. These define which CMEMS model and product to download and the prefix is a user defined string to prefix
the downloads. A catalogue of the CMEMS model and products can be found at https://resources.marine.copernicus.eu/?option=com_csw&task=results
Currently PyNEMO has only been tested using the physical global forecast model although the downloader should be able to download
other models and products, it has not been tested and their are known issues with other products that restrict seamless download.
e.g. the NorthWest Atlantic model is not currently compatable due to differences in how the model variables are stored.
FTP Configuration for Static and Grid files
--------------------------------------------
The next section CMEMS FTP configuration, defines which FTP server, remote directory and files to download. This should require
modification unless CMEMS changes the file structure or names. Note it is important that the filenames are separated by a space
as this is what PyNEMO is expecting. Finally the location of CDO executable which should be installed to enable subsetting to occur.
This can be found by running::
$ where cdo
Extent configuration
---------------------
Finally the last box, this is where the extent to download is configured, it is up to the user to decide but it is suggested this
is at least 1 degree wider than the destination configuration. The depth range to request is also defined here. This information can
be extracted from the CMEMS catalogue. Once set for a given configuration this will not need to be edited.
By providing a global tidal model dataset (TPXO and FES are currently supported) PyNEMO can generate boundary conditions for the
NEMO configuration supplied using the namelist file.
Namelist options
----------------
To use the namelist needs to be configured with the required options. These are listed below::
ln_tide = .true. ! =T : produce bdy tidal conditions
...
...
@@ -22,12 +25,20 @@ To use the namelist needs to be configured with the required options. These are
sn_tide_fes = './FES/'
these options define the location of the tidal model datasets, note this differs depending on model as TPXO has all harmonic
constants in one netcdf file whereas FES has three separate netcdf files for each constant. Extra harmonics can be appended
to the clname(n) list FES supports 34 constants and TPXO7.2 has 13 to choose from. Other versions of TPXO should work with PyNEMO
but have not been yet been tested.
constants in one netcdf file whereas FES has three separate netcdf files (one for amplitude two for currents) for each constant. Extra harmonics can be appended
to the clname(n) list. FES supports 34 constants and TPXO7.2 has 13 to choose from. Other versions of TPXO should work with PyNEMO
but have not been yet been tested. **NOTE** FES dataset filenames must have be in the format of constituent then type. e.g.::
M2_Z.nc (for amplitude)
M2_U.nc (for U component of velocity)
M2_V.nc (for V component of velocity)
If this is not undertaken the PyNEMO will not recognise the files. TPXO data files are specified directly so these can be anyname although it is best to stick with the default
names as shown above. So far the tidal model datasets have been downloaded and used locally but could also be stored on a THREDDS server although this has
not been tested with the global tide models.
So far the tidal model datasets have been downloaded and used locally but could also be stored on a TREDDS server although this has
not been tested to generate tidal boundaries
Other options include "ln_tide" a boolean that when set to true will generate tidal boundaries. "sn_tide_model" is a string that defines the model to use, currently only
"fes" or "tpxo" are supported. "ln_trans" is a boolean that when set to true will interpolate transport rather than velocities.
