Customize velocity fields

The following can help you get started to implement your own analytic functions for velocities.

The procedure of loading velocities

There are usually many individual files used to save velocity data from numerical simulations. In order to read it, we need to map a list of filename to those data. The relavent fortran Character variables are defined in global.f90 as the following block.

!file names
INTEGER*8, ALLOCATABLE :: fn_ids(:,:)
INTEGER*8 :: fn_uvwtsg_ids(7),fn_xyz_tsg_mld_ids(3),fn_id_mld,FnPartiInitId

INTEGER*8 :: file_i0
INTEGER*8 :: filename_increment=1
CHARACTER (len=64) :: casename,path2uvw,path2grid,FnPartiInit,output_dir,fn_phihyd,fn_mld
CHARACTER (len=64) :: fn_UVEL,fn_VVEL,fn_WVEL,fn_THETA,fn_SALT,fn_GAMMA

CHARACTER (len=64), dimension(Nrecs, Nvar2read) :: filenames

If one-record-per-file, the program will use c_filenames.f90 to pre-calculate the filenames and save them to variable filenames(N,M), where N is the number of records and M is the number of variables.

The velocity data will be loaded into memeory by the subroutine load_uvw() in io.f90.

The velocity data will next be used by find_particle_uvw() in get_velocity.f90.

After the velocity data interpolated onto the particle positions, rk4() will be used to intergrate the trajectory position. Because rk4 uses several time steps, the subroutine find_particle_uvw() will be used four times in rk4, each time with an updated velocity.