Custom Query (964 matches)

I continue my slow updating of all the Matlab scripts that use the NetCDF interfaces. As mentioned in ​WikiROMS, we are moving to the native NetCDF interface. However, we still support the MEXNC and SNCTOOLS interfaces.

I added a new function nc_contant.m to return the numeric value associated with the name of a NetCDF library parameter. It is an umbrella on top of the native function or older NetCDF interfaces. For example,

> value = nc_constant('NC_FLOAT')
  or
> value = nc_constant('nc_float')

It turns out that we have several scripts in this repository that use these named string parameters. They are defined if the MEXNC interface is installed... Some of you were having problems when using the the native interface alone without having the MEXNC interface installed.

I also reinstate old function nc_vname.m to get the names of all variables in a NetCDF file. This function is OBSOLETE but it is kept for backward compatibility. Use nc_vnames.m in the future.

I still have more work ahead to port several old scripts to use also the native Matlab interface.

I continue revising the Matlab scripts:

• Several matlab scripts were corrected to use nc_constant('nc_ulimited') instead of undefined parameter ncunlim. Many thanks to Arthur E. Ramos for reporting this ​problem

• I cleaned and retested initial/d_climatology.m and initial/d_initial.m template drivers and everything is working correctly.

• Corrected several typos in grid/contact.m. We need to use isempty instead of empty native funtions. Many thanks to John Wilkin for reporting these typos.

• I fixed a typo in initial/roms2roms.m and boundary/obc_roms2roms.m. We need to use nc_vinfo instead of nc_varinfo. Many thanks to John Wilkin for bringing this to my attention.

• Added nc_unlimited parameter to nc_constant.m. Also corrected a typo in the error message.

Several Matlab scripts where updated. The processing scripts roms2roms.m and obc_roms2roms.m have the incorrect calls to the nc_vinfo function. Many thanks to John Wilkin to remind me of these typos.

In addition several new scripts were introduced:

• coastline/get_coast.m: Extracts coastline data from the Global Self-consistent, Hierarchical, High-resolution Shoreline (GSHHS) database for the box corners bounded by (Llon,Blat) and (Rlon,Tlat).

• forcing/d_core2_frc.m: Driver template script showing how to create ROMS forcing NetCDF file(s) using ROMS metadata structure. The data source is the CORE 2 Global Air-Sea Flux Dataset. Notice that the original data set is sampled for the Gulf of Mexico (GOM).

• netcdf/check_metadata.m: Checks ROMS metadata structure for consistency and fills unassigned fields. This structure will be used elsewhere to create NetCDF files in a compact way.

• netcdf/nc_append.m: Defines/appends new variables to an existing NetCDF file. The variable(s) metadata is provided in input structure, which includes information about variable type, dimensions and attributes.

• roms_metadata.m: Sets metadata structure for requested ROMS NetCDF variable. The structure contains the same fields that are returned by nc_inq or native interface ncinfo. This can be used elsewhere to create a NetCDF file in a compact way.

• plot_field.m: Plots requested ROMS variable from input history NetCDF file. This function is very useful when debugging a ROMS application. The plotting is not that fancy but it provides enough information for browsing ROMS variables very quickly. The location of the minimum is marked with a filled magenta circle whereas the maximum is marked with a filled magenta square. It also draws the coastline, if the data is available in the grid file or structure.

Notice that the roms_metadata function builds a structure with similar information as in ROMS varinfo.dat. The strategy here is to create a structure that can be used with nc_create to generate ROMS input NetCDF files in a compact way. The driver template forcing/d_core2_frc.m provides an example of how the ROMS metadata structure is used to create several atmospheric forcing NetCDF files. It creates a single NetCDF file per forcing field. This is kind of nice since creating NetCDF file is too verbose and requirse hundreds of lines of coding. Here this is done in a very compact and generic way and with few lines. In addition, it provides the correct CF compliance.

I assume the the function roms_metadata will continue evolving in the future as we add more fields.