When reading netcdf data, we often use a temporary variable and then copy into 3d_fields. This allows us to collapse dimensions, and convert data types (e.g. float to int). However the netdf interface can do this so, with the correct options, we can read data directly into 3d fields. This should be more efficient, and simplifies the code.
Do this, and merge into MPI branch, before proceeding with parallelisation of reads.
When reading netcdf data, we often use a temporary variable and then copy into 3d_fields. This allows us to collapse dimensions, and convert data types (e.g. float to int). However the netdf interface can do this so, with the correct options, we can read data directly into 3d fields. This should be more efficient, and simplifies the code.
Do this, and merge into MPI branch, before proceeding with parallelisation of reads.