about forcing

General scientific issues regarding ROMS

Moderators: arango, robertson

Post Reply
Message
Author
px123
Posts: 7
Joined: Wed Mar 03, 2010 8:15 pm
Location: OSI

about forcing

#1 Unread post by px123 »

Hi, I am new to ROMS or any other numerical models. I am a little confused about the model forcings. Please help!

(1)From an initial condition, when I add like 1999 forcing, then the result will be 1999. Next from 1999 restart file, I add 2000 forcing, then the result will be 2000. Is this correct?

(2)What if I use 2000 restart file and I use 1999 forcing file again, then what's result will be?

(3)how to spin-up the model? Run the model with climatology file several years, then the result already reaches equilibrium and can be used as an initial file. Later from this initial file, you add your own forcing, you don't need to spin-up the model any more. Is this correct?

Thanks all.

User avatar
kate
Posts: 4091
Joined: Wed Jul 02, 2003 5:29 pm
Location: CFOS/UAF, USA

Re: about forcing

#2 Unread post by kate »

What exactly are your goals? Do you want to run a specific year? Are you looking at short-term processes or long-term variability?

I tend to be a bit careless about spin-up, assuming everything is close enough after a few months - but the time depends on the size of your basin and the processes you care about. The things we are looking for are the seasonal cycle in the upper water column and interannual variability. Much of this comes from the forcing, so the forcing has to be from the right years. If you are looking at the deep overturning circulation, your spin-up time might be longer. If it's tides, the spin-up time is shorter.

I no longer spin up with climatology - better to just jump into real years and throw away the first one.

px123
Posts: 7
Joined: Wed Mar 03, 2010 8:15 pm
Location: OSI

Re: about forcing

#3 Unread post by px123 »

kate,

I am trying to make a short-term run, like one or two years. Here comes my first two questions

(1)From an initial condition, when I add like 1999 forcing, then the result will be 1999. Next from 1999 restart file, I add 2000 forcing, then the result will be 2000. Is this correct?

(2)What if I use 2000 restart file and I use 1999 forcing file again, then what's result will be?

about spin-up, what do you mean by throwing away the first one. If I run the model from 1999 to 2001, then which one to throw away?

px123
Posts: 7
Joined: Wed Mar 03, 2010 8:15 pm
Location: OSI

Re: about forcing

#4 Unread post by px123 »

anybody help?

User avatar
kate
Posts: 4091
Joined: Wed Jul 02, 2003 5:29 pm
Location: CFOS/UAF, USA

Re: about forcing

#5 Unread post by kate »

(2)What if I use 2000 restart file and I use 1999 forcing file again, then what's result will be?
The result is a hybrid, looking more like 2000 at the beginning, then becoming more like 1999 as the forcing effects come to dominate over the effects of the initial conditions. What is that timescale for your problem?

magaesco
Posts: 10
Joined: Wed Sep 01, 2010 8:01 pm
Location: ESPOL

Re: about forcing

#6 Unread post by magaesco »

Hi roms community :P ,

It's a quick question, I've realized that for long periods, the files in ROMS_FILES are so heavy.
For my interest, I only need all septembers, octobers and novembers since 1999 to 2009, so I've configured romstoolsparam.m in TEMPORAL PARAMETERS
% 6 Temporal parameters (used for make_tides, make_NCEP, make_OGCM)
%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%
Yorig = 1900; % reference time for vector time
% in roms initial and forcing files
%
Ymin = 1999; % first forcing year
Ymax = 2009; % last forcing year
Mmin = 9; % first forcing month
Mmax = 11; % last forcing month
%
Dmin = 1; % Day of initialization

But, in ROMS_FILES and DATA, I got all months (january-december) that quit space in my disk. My question is:
Is it necessary to configure like above (and get so much files) or should I configure per year (first 1999, after 2000, ...) for computing (start, make_grid, make_NCEP, make_OGCM) since september to november (so I'll have few files)?

Thanks for advances :D :D :D ,

Maga

seashellingolds
Posts: 39
Joined: Wed Jun 25, 2008 2:49 am
Location: Georgia Institute of Technology

Re: about forcing

#7 Unread post by seashellingolds »

kate wrote:I no longer spin up with climatology - better to just jump into real years and throw away the first one.
Hi Kate

This is interesting to me. Can you tell me the reason (related paper?) or your experience on that.
Is this particular for some cases but not for others? It really saves time

Y

User avatar
kate
Posts: 4091
Joined: Wed Jul 02, 2003 5:29 pm
Location: CFOS/UAF, USA

Re: about forcing

#8 Unread post by kate »

Sorry, I can't point to this in a paper. However, I know that our NEP domain was quite badly behaved when using monthly Levitus boundary conditions and just settled down nicely when given five-daily boundary conditions from SODA. I'm sure it has to do with the velocity and tracer fields being consistent with each other in SODA, also consistent with the timescales of the winds we're using. A climatology is a mish-mash from all available years and could be consistent with a rather unrealistic velocity field.

Post Reply