model blows up using different advection scheme

Report or discuss software problems and other woes

Moderators: arango, robertson

Post Reply
Message
Author
adundunis
Posts: 15
Joined: Mon Jul 04, 2016 1:31 pm
Location: Inha University

model blows up using different advection scheme

#1 Unread post by adundunis »

Hi all,

My model (with river, tide and initial conditions inputs) finished its run using TS_A4HADVECTION and TS_A4VADVECTION. However, the results produced unrealistic salinity and temperature. I got maximum salinities of up to almost 200, and my minimum salinity is negative.

I have been advised to use TS_MPDATA instead. However, my model blows up after several hours of model run, not even reaching the time of the first discharge.

My discharge runs from north to south and my river positions are located in a 4 x 5 grid (meaning, river = 20) in the river mouth, this means that some of the river positions must be surrounded by water cells. Can ROMS handle such configuration? I read somewhere that a river position must have a 0 mask of u or v, and must be located between a land and water cell. Anyway I tried to run such configuration with TS_A4HADVECTION and TS_A4VADVECTION. It didn't blow up but that's where I got the weird salinity values.

So I suspected maybe I should change the configuration to have just river = 5 (in the outermost v's) so that all my river positions are between a land and a water cell. I also made sure that my mask_v's in these river positions are 0's. My discharge are from north to south so my river_direction are all ones and my river_transport are all negative values. But still, my model blows up even before the first discharge when I use TS_MPDATA as an advection scheme.

Has anyone encountered this problem before?

I hope someone can give me some insights.

Thank you!

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

Re: model blows up using different advection scheme

#2 Unread post by kate »

Yes, I've had rivers blow up. Maybe not for the same reason as yours though.

Yes, you need to have just the five river sources at the land/water interface.

Yes, you need to provide river tracers - explicit temperature and salinity for the inflows. Otherwise you are using an unstable downwind condition on those tracers.

I'm not using MPDATA, so can't speak to that issue.

Post Reply