#17 problems and solutions for agionccmd and swn_outnc

John Warner

dear swan developers-
thanks for access and recent updates to the code.
Here are a few things that came up.

1) When using gfortran, there was an error in agionccmd.ftn90. the error was
dnames = (/ 'time', 'run' /)
Error: different CHARACTER lengths (4/3) in array constructor.
The variables dname and dnames are char*6, so why not just have dnames be:
dnames = (/ 'time ', 'run ' /)
Making it be six characters long seemed to make gfortan happy. This was not an issue in ifort or pgi.

2)in agionccmd.ftn90, pgi had a problem with
character*nf90_max_name :: lonname
and i had to change it to be
character(len=nf90_max_name) :: lonname
this now works in pgi, ifort, and gfortran.

3) gfortran did not like swn_outnc.ftn90, line 183 where irq is defined as a logical. It can not be a logical and then used as an array indicator
OUTP_FILES(irq) <----- irq needs to be an integer.
In the rest of the file, irq is an integer, so i changed line 183 from
logical :: irq
integer :: irq
and that seems to work.

4) a suggestion:
i only tested a few netcdf files, but for example, the variable name for wave height in the netcdf file was “hs”. Is there some way to make this compatible with the name of the SWAN output var, like HSIGN? It would be nice to have the names of the variables in the netcdf file to be consistent with the names listed in the user manual.

john warner


  • Marcel Zijlema
    Marcel Zijlema

    All bug fixes have been implemented.

  • Marcel Zijlema
    Marcel Zijlema

    • status: open --> closed