Page 1 of 1

conc2cdf bug

Posted: August 12th, 2013, 8:40 pm
by cory
Hi there,
I am using Hysplit (from the subversion repository) on a linux machine.
When I run hysplit with 5 pollutants conc2cdf works fine, but when I use 6 pollutants, the concentration array for the final pollutant in the netcdf file is full of 9e36. This looks to me like a memory allocation problem. I expect the problem also depends on the size of the concentration grid. I know that the original cdump file is fine, because the concplot output is fine.
Cheers,
Cory.

Re: conc2cdf bug

Posted: August 19th, 2013, 1:26 pm
by ariel.stein
Cory,

You might be reaching a limit in the NETCDF libraries? How big are your arrays?

Re: conc2cdf bug

Posted: August 20th, 2013, 9:56 pm
by cory
Thanks Ariel,

the array dimensions are:

dimensions:
longitude = 361 ;
latitude = 182 ;
levels = 9 ;
time = 12 ;

I compiled hysplit against netcdf library version 3.6.1 (64 bit). I might try a newer version.

Cheers,
Cory.

Re: conc2cdf bug

Posted: August 23rd, 2013, 9:52 am
by ariel.stein
Did you try a different netcdf version?