Blowing-up problem

General scientific issues regarding ROMS

Moderators: arango, robertson

Post Reply
Message
Author
hbzong
Posts: 36
Joined: Thu Oct 04, 2007 4:14 am
Location: Fathom Science
Contact:

Blowing-up problem

#1 Unread post by hbzong »

I'm trying to setup a tidal model with wetting-drying scheme. The minimum bathymetry in my grid is 0.1m, and the minimum depth for wetting and drying was set to 0.5m. Blowing-up occurs at the step 16979. Only tide forcing was added to the model.

DT == 60.0
NDTFAST == 10

Minimum barotropic Courant Number = 1.00127762E-02
Maximum barotropic Courant Number = 4.44498424E-01
Maximum Coriolis Courant Number = 4.82332138E-03

rx0 = 2.000000E-01 (Beckmann and Haidvogel)
rx1 = 1.800000E+00 (Haney)
Fig1.jpg
Fig1.jpg (22.52 KiB) Viewed 3953 times
I checked the restart file and found that the blowing-up was occurred by abnormal v velocity. (See Figure 2, in grid cell x = 185, y = 123),
Fig2.jpg
I also plotted the zeta and v one step before the blowing-up (Figure3).
(one step before)
(one step before)
I can’t identify what caused the abnormal v velocity.
Can anyone give some suggestions?

liyizhen837
Posts: 11
Joined: Tue Aug 21, 2007 5:44 pm
Location: North Carolina State University

Re: Blowing-up problem

#2 Unread post by liyizhen837 »

Haibo,
Is it caused by DT/DTFAST too large?
CFL condition satisfied yet?

nacholibre
Posts: 81
Joined: Thu Dec 07, 2006 3:14 pm
Location: USGS
Contact:

Re: Blowing-up problem

#3 Unread post by nacholibre »

Which version of ROMS are you using?
Did you try a larger NDTFAST? (30 maybe?)
Zafer

User avatar
arango
Site Admin
Posts: 1347
Joined: Wed Feb 26, 2003 4:41 pm
Location: DMCS, Rutgers University
Contact:

Re: Blowing-up problem

#4 Unread post by arango »

Yes, we have mentioned several times in this forum about of the time-stepping properties of ROMS. There are several publications about this, see any of the Shchepetkin et al. papers. They discuss this in great detail. User of ROMS need to read the basic core publications about ROMS before starting running your applications. User should be aware that ROMS has an extensive literature with hundreds of papers. There is a lot of great tips and information in the scientific literature that maybe relevant to any your applications. I encourage users to read this literature.

We actually recommend having a value for NDTFAST to be larger than 30. A value of 10 is too small. You need to fine tune DT accordingly to get the appropriate baroclinic time-step. Of course, you need to first set your barotropic time-step according to the fast gravity-wave dynamics which primarily controls the CFL in ROMS.

Post Reply