ROMS prestep3d , Crack-Nicolson method, lambda

General scientific issues regarding ROMS

Moderators: arango, robertson

Post Reply
Message
Author
Honolulu_Wang
Posts: 17
Joined: Mon Sep 08, 2014 4:17 pm
Location: Hohai University

ROMS prestep3d , Crack-Nicolson method, lambda

#1 Unread post by Honolulu_Wang »

Hello All,
I am going through the ROMS code. What I found is in the ROMS user manual, it said it will use Crank_niskson for calculating the vertical diffusion term time-stepping.

And it will consistent with ROMS has a pre3d_t.F routine, then it finally gets to step3d_uv.F and step3d_t.F.

But when I check it, I found in MODULE mod_scalars, lambda = 1.0_r8
So it tells lambda is always 1.0

! Weighting coefficient for the newest (implicit) time step derivatives
! in the Crack-Nicolson implicit scheme (usually, lambda=0.5).
!
real(r8) :: lambda = 1.0_r8 ! backward implicit


And in pre_step3d_mod, it comments :

! It also computes the time "n" vertical viscosity and diffusion !
! contributions of the Crank-Nicholson implicit scheme because the !
! thicknesses "Hz" will be overwriten at the end of the 2D engine !
! (barotropic mode) computations.

It tells me ROMS only and always use backward implicit other then Crank-Nicholson implicit scheme. Is this same from your understanding?

Thanks

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

Re: ROMS prestep3d , Crack-Nicolson method, lambda

#2 Unread post by arango »

Yes, we need to update the documentation.

Post Reply