This repository is an incomplete Fortran version of the Python Constrained Spectral Approximation Method.

The Fortran implementation was meant to be the orographic gravity-wave source for the Multi-scale Gravity-wave Model that is coupled to the German Weather Service's Icosahedral Nonhydrostatic (ICON) model. This codebase builds, runs, and outputs the spectral approximation of topographic data distributed in a geodesic grid cell. However, not all of the latest algorithmic and scientific features have been implemented in the Fortran code.

• A simple plotter for the output data can be found in notebooks/plotter.ipynb.
• Namelist files are in scripts.

Note that this codebase uses features from the Fortran 2008 standard.

Requirements

• gfortran-11
• BLAS
• LAPACK
• fortran-stdlib
• OpenMP
• NetCDF-Fortran
• CMake

Refer to CMakeLists.txt for more details.

Build

cd trifourtopo
mkdir build
cd build
cmake ..
make

The make command will create two binaries in build/bin:

• linker
• four_topo

Run

This codebase uses the USGS GMTED 2010 topographic dataset and an ICON grid.

Before executing four_topo for the spectral analysis, you must link the topographic dataset to the corresponding grid by setting the path in linker.nml and executing build/bin/linker. The program extends the NetCDF grid data with a lookup table to the corresponding data slice in the topographic dataset.

The binary four_topo outputs a NetCDF4 file comprising the computed spectral approximation over all the cells provided in the grid file. The output may be visualised with notebooks/plotter.ipynb.

The notebook nc_compactifie.ipynb creates a regional grid and topographic dataset and may help with debugging and testing.