[MITgcm-support] Problem with input data

Alistair Adcroft adcroft at MIT.EDU
Thu May 20 17:23:32 EDT 2004


My 2 cent bet would be on either seaice_init() or thsice_ini_vars().
 
A.

--
Dr Alistair Adcroft            http://www.mit.edu/~adcroft
MIT Climate Modeling Initiative        tel: (617) 253-5938
EAPS 54-1624,  77 Massachusetts Ave,  Cambridge,  MA,  USA 

-----Original Message-----
From: mitgcm-support-bounces at mitgcm.org
[mailto:mitgcm-support-bounces at mitgcm.org] On Behalf Of Matthew Mazloff
Sent: Thursday, May 20, 2004 4:25 PM
To: mitgcm-support at mitgcm.org
Subject: Re: [MITgcm-support] Problem with input data


Hi all,

Good news.  By commenting out

C--   Initialize variable data for packages
C--      CALL PACKAGES_INIT_VARIABLES( myThid )

The problem is fixed.  The model looks great...But...what did I mess up by
not initializing the packages.  I assume it was only one part of the package
file that was messing up the temperature field.  Any idea what this may be.

Thanks,
Matt

Alistair Adcroft wrote:


I was thinking that you might be going into a "freezing" or convection
routine but based on your data file that doesn't seem to be the case (unless
you've been adding parameters to PARAMS.h and there's a mismatch between
common blocks in some routines).
 
Have you tried a "make clean; make cleanlinks; makemakefile; make depend;
make" i.e. a complete rebuild?
 
BTW, when did this start happening? Did you change something around that
time?
 
A.
--
Dr Alistair Adcroft            http://www.mit.edu/~adcroft
<http://www.mit.edu/%7Eadcroft> 
MIT Climate Modeling Initiative        tel: (617) 253-5938
EAPS 54-1624,  77 Massachusetts Ave,  Cambridge,  MA,  USA 

-----Original Message-----
From: mitgcm-support-bounces at mitgcm.org
[mailto:mitgcm-support-bounces at mitgcm.org] On Behalf Of Matthew Mazloff
Sent: Thursday, May 20, 2004 12:34 PM
To: mitgcm-support at mitgcm.org
Subject: Re: [MITgcm-support] Problem with input data


Hi Alistair and everyone,

It didn't help, but thanks for catching those errors.  What is happening to
the model between when it reads in the temperature field and timestep 0.  I
can't figure out what step is setting that part of the temperature field to
zero.  Assuming we can figure out where this may be happening, can I just
override this step, because the model works fine after this point.  I have
usePickupBeforeC35 as the default value of .FALSE. so I dont think the model
is even going into a correction step.  

Thanks,
Matt

Alistair Adcroft wrote:


Matt,
 
These shouldn't change anything but with MITgcm you never know:
 i) sRef=46*35.
 ii) add a comma at end of Qstrat line
 
A.
--
Dr Alistair Adcroft            http://www.mit.edu/~adcroft
<http://www.mit.edu/%7Eadcroft> 
MIT Climate Modeling Initiative        tel: (617) 253-5938
EAPS 54-1624,  77 Massachusetts Ave,  Cambridge,  MA,  USA 

-----Original Message-----
From: mitgcm-support-bounces at mitgcm.org
[mailto:mitgcm-support-bounces at mitgcm.org] On Behalf Of Matthew Mazloff
Sent: Thursday, May 20, 2004 11:25 AM
To: mitgcm-support at mitgcm.org
Subject: Re: [MITgcm-support] Problem with input data


Hi Chris,

I have only tried it on a mac, but I don't believe this is the problem.  The
input data is read fine.  I don't know much about the masks, but I now
believe the temperature mask may be the problem.  I see that the dry area of
my model has the reference temperature after several time steps.  So
initially all dry area has zero temperature.  All but 3/4 of one wet layer
is good, but there is that one layer that is set to zero in the first
correction step.  After 1 time step this wet area acts fine again (But the
damage is done to the initial conditions)  and then at some point, im not
sure where yet, the all the dry area except for the same layer that was
messed up before, takes on the reference temperature! This is then
maintained throughout the run.  All other fields are zero in the dry region.
I have attached output with a single timestep.  I don't know if it will
help. Any ideas as to what I should try next?

Thanks so much,
Matt



Chris Hill wrote:


Matt,



 Is it only on the mac? Also, can you put a full STDOUT somewhere on the

web, or attached to a mail.



Thanks,



CHris 



  

-----Original Message-----

From: mitgcm-support-bounces at mitgcm.org 

[mailto:mitgcm-support-bounces at mitgcm.org] On Behalf Of 

Matthew Mazloff

Sent: Wednesday, May 19, 2004 7:57 PM

To: mitgcm-support at mitgcm.org

Subject: Re: [MITgcm-support] Problem with input data



Update:



The data is read in correctly as the output shows



 Field Initial Temperature at iteration          1

(PID.TID 0000.0001) // CMIN =          4.057084608000000E-04

(PID.TID 0000.0001) // CMAX =          4.093527510783521E-04

(PID.TID 0000.0001) // CINT =          1.349737140130389E-07



However at the first checkpoint, after nothing has happened 

but a calculation of the pressure (I am running hydrostatic) 

the output reads



(PID.TID 0000.0001) %MON dynstat_theta_max            =   

4.0935275107835E-04

(PID.TID 0000.0001) %MON dynstat_theta_min            =   

0.0000000000000E+00

(PID.TID 0000.0001) %MON dynstat_theta_mean           =   

2.0377073219221E-04



note theta_min has gone to zero!



What could be happening?



Thanks,

Matt





Matthew Mazloff wrote:



    

I am running the MITgcm on a mac.  With Matlab, using gendata.m,  I 

create input files such as



fid=fopen('Ulinear.rect','w','b'); fwrite(fid,U,'float64'); 

fclose(fid);



I am using readbinaryprec = 64

All fields input fine, except for my the temperature field.  This 

field has a chunk of the data in the middle of the file set 

      

equal to 

    

zero.

I have tried other machine formats ('s', 'l', 'a', 'c'), and 

precisions, ('double'), but the problem  has only been made worse.

The temperature field has a larger amplitude than the 

      

velocity fields.  

    

But this is the only difference.  The place where the field is set 

equal to zero is the same each time.  I have checked that Matlab is 

writing the field correctly using fread, and it is fine.



Any ideas?

Thanks

Matt



_______________________________________________

MITgcm-support mailing list

MITgcm-support at mitgcm.org

http://dev.mitgcm.org/mailman/listinfo/mitgcm-support

      

_______________________________________________

MITgcm-support mailing list

MITgcm-support at mitgcm.org

http://dev.mitgcm.org/mailman/listinfo/mitgcm-support



    



_______________________________________________

MITgcm-support mailing list

MITgcm-support at mitgcm.org

http://dev.mitgcm.org/mailman/listinfo/mitgcm-support

  



  _____  


_______________________________________________

MITgcm-support mailing list

MITgcm-support at mitgcm.org

http://dev.mitgcm.org/mailman/listinfo/mitgcm-support

  



  _____  


_______________________________________________

MITgcm-support mailing list

MITgcm-support at mitgcm.org

http://dev.mitgcm.org/mailman/listinfo/mitgcm-support

  

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mitgcm.org/pipermail/mitgcm-support/attachments/20040520/f6b6f903/attachment.htm>


More information about the MITgcm-support mailing list