https://github.com/mgduda/convert_mpas

A general conversion utility for MPAS-Atmosphere output files
https://github.com/mgduda/convert_mpas

Last synced: about 2 months ago
JSON representation

A general conversion utility for MPAS-Atmosphere output files

• Host: GitHub
• URL: https://github.com/mgduda/convert_mpas
• Owner: mgduda
• Created: 2016-05-23T21:21:46.000Z (about 9 years ago)
• Default Branch: master
• Last Pushed: 2024-07-03T19:57:06.000Z (11 months ago)
• Last Synced: 2025-04-11T21:13:51.794Z (about 2 months ago)
• Language: Fortran
• Size: 97.7 KB
• Stars: 15
• Watchers: 3
• Forks: 20
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
convert_mpas

============

The 'convert_mpas' project aims to develop a general framework for mapping 

native MPAS output to other meshes.

## Compiling:

After selecting a Fortran compiler and associated compiler flags with the FC and FFLAGS

variables in the top-level Makefile, the 'convert_mpas' program may be compiled by 

simply running 'make'.

## Running:

The 'convert_mpas' program takes one or more command-line arguments:

- If only one argument is given, both the MPAS mesh information and 

  the fields will be read from the specified file.

- If two or more file arguments are given, the MPAS mesh information will 

  be read from the first file and fields to be remapped will be read from

  the subsequent files.

All time records from input files will be processed and appended to the output 

file, which is named 'latlon.nc'.

Running 'convert_mpas' with no command-line arguments will print a usage summary.

By default, the 'convert_mpas' will remap all integer, real, or double-precision

fields that it finds in the input data file. However, by creating a list of

fields in a file named 'include_fields' in the run directory, with one field name 

per line, the 'convert_mpas' program will remap only those fields listed in 

the file. Alternatively, one can create a list of fields to be excluded from 

the output file; this list should be written to a file named 'exclude_fields'.

If both an 'include_fields' file and an 'exclude_fields' file are present in 

the run directory, only fields listed in the 'include_fields' file will be 

remapped, and the contents of the 'exclude_fields' file are ignored.

The target domain defaults to a 0.5x0.5-degree global lat-lon grid. However, one

may specify an alternate target domain using a file named 'target_domain' in 

the run directory. This file may contain lines assigning values to keywords, i.e.,

keyword = value

The following are available keywords for describing the target domain:

 - nlat : the number of latitude points in the grid (default value 360)

 - nlon : the number of longitude points in the grid (default value 720)

 - startlat : the starting latitude (default value -90.0)

 - startlon : the starting longitude (default value -180.0)

 - endlat : the ending latitude (default value 90.0)

 - endlon : the ending longitude (default value 180.0)

The actual points to which fields are interpolated are determined by dividing

the latitude and longitude ranges into the specified number of intervals, then

locating the interpolation points at the center of these intervals. For example,

specifying startlat=0, endlat=10, and nlat=10 would result in target latitudes

of 0.5, 1.5, ..., 8.5, and 9.5.

## Interpolation methodology:

Integer fields are remapped to the target grid using a nearest-neighbor scheme.

For all real-valued (single- or double-precision) fields, the 'convert_mpas' program

employs a barycentric interpolation, the output of which is C0 continuous. Cell-based

fields in the MPAS mesh are sampled from the three cell centers that form the vertices

of the Delaunay triangle containing the target point. Vertex- and edge-based fields

are sampled from the corners or faces, respectively, of the Voronoi cell containing

the target point.

## To-do:

- Transfer 'xtime' variable from input files to output file

- Ensure that, for cell fields, the interpolation location lies within the triangle 

  used for interpolation

- Make sure that, when dealing with existing output files, the target mesh matches

  what is found in the output file

- Experiment with OpenMP directives to speed up interpolation

- Allow locations of 'include_fields', 'exclude_fields', and 'target_domain' files

  to be specified with environment variables

- Decide what to do if input file contains no unlimited dimension