WRF is performing forecasting or simulation?

Any issues with the actual running of the WRF.

WRF is performing forecasting or simulation?

Postby Xiaohei.John » Sun Nov 13, 2016 8:23 am

How can I know whether WRF is performing forecasting or simulation ?
Is there any option specific for it ?

As everyone knows, forecasting is using the data at one moment to forecast a few hours head time while simulation is using the data at start and end time to interpolate the hours between start and end.

So I did an experiment with WRF-ARW using ERA-Interim as forcing data and configured the start time at 2008-09-30_18:00:00 while end time at 2008-10-01_06:00:00.
1. Without giving data of October, ungrib output error, indicating that no way to proceed.
2. With data of September and October, I set the end_time as 2008-10-01_12:00:00. As a result, WPS had successfully run while WRF output error as there was no ".nc" file of 2008-10-01_12:00:00.
3. With data of September and October, I set the start_time and end_time according to the namelist.wps in WPS while set the run_hours as 18 which means running hours is longer than the subtraction of end_time minus start_time(12). As a result, I still got an error "run out of valid boundary conditions in wrfbdy_d01".

I am still confused of whether WRF is performing forecasting or simulation.
Anyone know that there is option specific for it ?
Xiaohei.John
 
Posts: 3
Joined: Sun Nov 13, 2016 12:05 am

Re: WRF is performing forecasting or simulation?

Postby kwthomas » Mon Nov 14, 2016 5:46 pm

When you run "wrf.exe", you'll get files "rsl.out.0000" and "rsl.error.0000". If you are running MPI, you'll see additional files that begin rsl.out and rsl.error.

In the case of (3), you need to make sure you aren't running a forecast that goes past your
lateral boundary conditions, which is the contents of "wrfbdy.d01". Doing that is illegal.
Kevin W. Thomas
Center for Analysis and Prediction of Storms
University of Oklahoma
kwthomas
 
Posts: 168
Joined: Thu Aug 07, 2008 6:53 pm

Re: kwthomas

Postby Xiaohei.John » Tue Nov 15, 2016 7:47 am

Hi, Thomas. Thanks for your reply.
As the result of three experiments showed, it couldn't perform forecasting without the data at end_time. Note that, it is kind of simulation, interpolating the value with data at start_time and end_time. However, WRF is supposed to have the capability to perform forecasting. I just want to know how it works
Xiaohei.John
 
Posts: 3
Joined: Sun Nov 13, 2016 12:05 am

Re: WRF is performing forecasting or simulation?

Postby kwthomas » Tue Nov 15, 2016 7:44 pm

You need initial conditions and boundary conditions for up to the end of your run time. Your boundary condition data can extend beyond your forecast length, but it can't be shorter, as this is a fatal error.

A model without boundary condition data will do nothing useful at the boundaries. Eventually, this could contaminate the forecast.

Let me give you an example. Suppose your initial conditions are entirely in a warm sector with a strong arctic front just to the north of the boundary at t=0. Given a winter time scenerio, you would expect that cold air to plunge south in your domain. Without the use of the boundary data from another model, this would never happen, as your run wouldn't know there is a strong front near the domain.
Kevin W. Thomas
Center for Analysis and Prediction of Storms
University of Oklahoma
kwthomas
 
Posts: 168
Joined: Thu Aug 07, 2008 6:53 pm

Re: kwthomas

Postby Xiaohei.John » Thu Nov 17, 2016 10:37 pm

Hi, Thomas. Thanks for explaining the essential of initial conditions and boundary conditions for me.

I don't know whether it is right that I am taking the ERA-Interim data as my initial conditions data and I have set up two coarser domains as the boundary conditions for the inner-most domain, totally three doamins.

The purpose of my attempt to figure out that WRF is doing forecasting or simulation is to do real-time forecast. However, I have no idea about how to configure namelist.input to do forecast.

Suppose I have the latest real-time data, I want to forecast precipitation of the future six hours.

Any idea?
Xiaohei.John
 
Posts: 3
Joined: Sun Nov 13, 2016 12:05 am

Re: WRF is performing forecasting or simulation?

Postby kwthomas » Fri Nov 18, 2016 6:34 pm

Check the file run/README.namelist in the WRF distribution. It tells you a lot.

One general rule to keep in mind. Your timestep must be no larger than 6x the grid spacing in km. If larger, it can go unstable. Even at 6x, it sometimes can go unstable, depending on your initial conditions. Runs initialized with severe convection sometimes go unstable very quickly.

The best thing to do is try it out, and look for complaints in the rsl* files. You can always post what you see here.
Kevin W. Thomas
Center for Analysis and Prediction of Storms
University of Oklahoma
kwthomas
 
Posts: 168
Joined: Thu Aug 07, 2008 6:53 pm

Re: WRF is performing forecasting or simulation?

Postby Abhishek sharma » Sun Dec 11, 2016 6:32 pm

hello Xiaohei.John,
I am also facing the same confusion and the question is still not answered. I have also done the same test but all showing me the same error that is (out of boundary condition wbdy01).
Anyone please help..!!
Abhishek sharma
 
Posts: 3
Joined: Wed Dec 07, 2016 4:25 pm


Return to Runtime Problems

Who is online

Users browsing this forum: No registered users and 3 guests

cron