possble a little bug in varinfo.dat

[leon]

In file "varinfo.dat",

Code: Select all

'river_dye_'                                       ! Input
  'river runoff dye concentration'
  'kilogram meter-3'                               ! [kg/m3]
  'river_dye, scalar, series'
  'ocean_time'
  'idRtrc(inert(i))'
  'nulvar'
  1.0d0

I think the

Code: Select all

"ocean_time"

should be

Code: Select all

"river_time"

. Am I right?

[kate]

Set it how you need it to be set for your input files. I've had files with both bio_time and ocean_time and had to change the bio variables accordingly. varinfo.dat is a tool for you to use and modify.

[wilkin]

The name of the time variable associated with any particular data variable in a ROMS netcdf file can be set in a number of ways. These are documented in WikiROMS at
https://www.myroms.org/wiki/index.php/varinfo.dat
and in my posting at the end of thread
viewtopic.php?f=17&t=1816&p=6528&hilit= ... .dat#p6528

So the Vinfo(5) value of "river_time" here is not a bug - simply a default. My personal recommendation is to always overrule the varinfo.dat default by assigning a "time" attribute for every variable in the netcdf file - this avoids any ambiguity.

[kate]

My personal recommendation is to always overrule the varinfo.dat default by assigning a "time" attribute for every variable in the netcdf file - this avoids any ambiguity.

This is the cleanest way to go. However, depending on how you write your attributes in the NetCDF file, you may end up with a null character at the end of the string. Fortran doesn't know to make that go away before doing the string comparison (or we haven't taught ROMS that yet).