Scheduled Downtime
On Friday 21 April 2023 @ 5pm MT, this website will be down for maintenance and expected to return online the morning of 24 April 2023 at the latest

WRF em_hill2d_x not working for finer resolutions

Dear WRF experts/ users,
I am trying to setup the WRF em_hill2d_x test case with dx=dy=dz=500m. However, at this resolution, i either see that there is a segmentation fault or only the output for the first time step, after that, there is a segmentation fault. I have compiled wrf using smpar and expected that increasing the processers for the task should help solve it fast. However, that is not the case. Any idea why this is happening?
Best,
Hrishi
 
Hrishi,
One principle in numerical modeling is that dz <<dx. In your case, you set dx=dy=dz, which will cause numerical instability and the model will crash.
Another issue is that, for 2D ideal case, we recommend to compile the code in serial mode. This is based on the consideration of domain decomposition.
 
HI Ming,
This statement “One principle in numerical modeling is that dz <<dx”. Can you elaborate what you are trying to say here?
Why should dx be >> dz? Is it because - d/dx and d/dy are much smaller than d/dz, especially near the surface? Can you please elaborate about this?
Best,
Hrishi
 
Last edited:
Hrishi,
I am sorry to answer late but this is because we are busy with mpas and mpas-jedi tutorial. To answer your question why the vertical resolution should be smaller than horizontal resolution, I believe it is because (1) the ratio of the horizontal and vertical resolutions must be consistent with the slope of the weather phenomena of interest, and (2) the atmosphere is thin compared to the size of the sphere.
 
Top