npzd_Powell

Discussion about coupled ecosystem models

Moderators: arango, robertson

Post Reply
Message
Author
hjsong

npzd_Powell

#1 Unread post by hjsong »

Hi,

I was trying to use npzd variables as passive tracers.
So I killed all biological interaction in /Nonlinear/npzd_Powell.h .
In order to test this, I ran just 10 days with zero values for npzd in the initial/boundary file.
I expected to see zero values throughout the run, but the output are not exactly zero( ~10^(-4)).
I made sure that there were no source and sink for npzd in Functionals/ana_stflux.h .

Why the values are not zero? Can I assume these as numerical errors?

Thank you!

Hajoon

User avatar
kate
Posts: 4088
Joined: Wed Jul 02, 2003 5:29 pm
Location: CFOS/UAF, USA

Re: npzd_Powell

#2 Unread post by kate »

You can also try this using a passive tracer which is spatially uniform. You will find that the advection-diffusion operators will render your tracer slightly non-uniform. If you read Sasha's papers on the ROMS timestepping, you will see that he tried his level best to maintain constancy preservation in the ROMS numerics, but there are still some sources/sinks due to the way the model has a split timestep.

Note that you can generate negative tracer values through the advection. My friends have found it useful to clamp the bio tracers >=0 before and after the bio timestepping. In my experience, the MPDATA option (which should prevent the advection from generating negative numbers) is (a) more costly and (b) worse at the constancy preservation in extremely shallow waters.

Post Reply