wrfhelp........please...help me.....

Any problems/questions that might not fall into the other topics in this category.

wrfhelp........please...help me.....

Postby tj9456 » Tue Jul 08, 2008 11:10 am

morning....
I got some problem in my model..

when i running the model(12km、4km).... it's show

370359 point exceeded cfl=2 in domain 2 at time 17.03333
MAX AT i,j,k: 99 129 30 cfl,w,d(eta)= NaN NaN 2.1000000E-02

963583 point exceeded cfl=2 in domain 1 at time 17.10000
MAX AT i,j,k: 208 159 30 cfl,w,d(eta)= NaN 3.7302179E-05 2.1000000E-02

What’s wrong ? can anyone tell me?
Indebted forever tj
tj9456
 
Posts: 11
Joined: Sun Jul 06, 2008 9:06 pm

Re: wrfhelp........please...help me.....

Postby jimmyc » Tue Jul 08, 2008 12:22 pm

The model is blowing up.

The CFL condition for a stable time step is being violated. Since we have no information about your model set up (i.e. namelist options), my recommendation would be to lower the time steps on the 2 domains.

Please see the technical description of WRF on the model's home page.
The views expressed in this message do not necessarily reflect those of NOAA or the National Weather Service or the University of Oklahoma.
James Correia, Jr
jimmyc
 
Posts: 519
Joined: Tue Apr 15, 2008 1:10 am

Re: wrfhelp........please...help me.....

Postby tj9456 » Tue Jul 08, 2008 9:14 pm

morning jimmyc
Thank you very much for your replying.......
about your suggestion... i did it ...
from [version 2 modeling sysytem user's guide ] july 2007 ...
the time_step:recommended 6*dx in km for a typical case
so.... my d01 resolution is 12km .....i set up the time_step for 72 seconds
and ....Above-mentioned phenomena appear
may be the time_step should be set to 6*dx(d02=4km)=24 seconds ?
anyway....i`ll try it again....and Cause supreme gratitudes
have a good day
tj
tj9456
 
Posts: 11
Joined: Sun Jul 06, 2008 9:06 pm

Re: wrfhelp........please...help me.....

Postby jimmyc » Wed Jul 09, 2008 12:08 am

The key word is "typical". The time step is very sensitive to the dynamics you are trying to simulate. This means you have to consider model topography (steep mountains vs flat terrain) and also the phenomena of interest. If convection is being simulated you have to make adjustments for that too.

The error information you provided is basic. If you could explain where your domain is located that would help. You can also post your namelist options. Please give us more details on how far into the simulation the errors first start to occur and where this is happening (in the vicinity of convection or near mountains).
jimmyc
The views expressed in this message do not necessarily reflect those of NOAA or the National Weather Service or the University of Oklahoma.
James Correia, Jr
jimmyc
 
Posts: 519
Joined: Tue Apr 15, 2008 1:10 am

Re: wrfhelp........please...help me.....

Postby tj9456 » Wed Jul 09, 2008 1:51 am

Cause your puzzlement, it is really uneasy,But I really want to solve the problem
fig1、2 show the error time started and nesting grid...
and fig3 it namelist.......
and ..........thanks for your help
tj
Attachments
12x6.png
12x6.png (170.46 KiB) Viewed 7356 times
Last edited by tj9456 on Tue Jul 15, 2008 9:23 pm, edited 2 times in total.
tj9456
 
Posts: 11
Joined: Sun Jul 06, 2008 9:06 pm

Re: wrfhelp........please...help me.....

Postby jimmyc » Wed Jul 09, 2008 3:04 am

try using the 3*dx for the time step for each domain:
d01 should have a time step close to 36 seconds and d02 should be closer to 12.

I am guessing that the problem is with domain 2. Once the tendencies are computed from domain2 and applied to d01 the cfl error emerges. This may be totally time step related.

This is a Taiwan area simulation? There is some significant terrain so the time step should be smaller.

One other possibility is that the PBL scheme is being called rather infrequently. try having it called every time step.

FIRST: turn off domain 2 and run the model. make sure no problems occur. You can at least argue for a longer time step IF your current set up without domain 2 runs fine. THEN rerun with domain 2 on and see if you can correct the problem.
The views expressed in this message do not necessarily reflect those of NOAA or the National Weather Service or the University of Oklahoma.
James Correia, Jr
jimmyc
 
Posts: 519
Joined: Tue Apr 15, 2008 1:10 am

Re: wrfhelp........please...help me.....

Postby tj9456 » Thu Jul 10, 2008 10:05 pm

Thank you for your guidance.........
and i`ll feedback the achievement...
tj9456
tj9456
 
Posts: 11
Joined: Sun Jul 06, 2008 9:06 pm

Re: wrfhelp........please...help me.....

Postby tj9456 » Tue Jul 22, 2008 11:01 pm

I have tested... and... d02 is out of question, and It`s not a question of time_step(*3 or *4 is no problem)

The problem lies in the bldt of pbl.....
as your propose... when i reduce the bldt to 1 minute...The problem has been solved.
I think the main factor is still the topography of Taiwan ,
so ...Thank you for your assistance again...
good day
see you next time,.............I have a lot of questions :lol:
tj9456
tj9456
 
Posts: 11
Joined: Sun Jul 06, 2008 9:06 pm


Return to Miscellaneous

Who is online

Users browsing this forum: No registered users and 2 guests