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

Size choice of the first and for the inner WRF domains

This post was from a previous version of the WRF&MPAS-A Support Forum. New replies have been disabled and if you have follow up questions related to this post, then please start a new thread from the forum home page.

Dear all,

I would like to use the WRF model with GFS 0.25 input to obtain forecasts from 0 to 120 hours with 1 km of spatial resolution over a small Region of southern Italy (Apulia).
I thought about using 2 domains with 5 and 1 km resolution respectively (5 scaling ratio).

The question is:
How big does the first domain have to be to avoid getting forecast errors especially beyond 36 hours?
Can I select southern Italy as the first domain or do I have to consider the whole country? Are there any rules about it?

Thank you very much for your support.

Hi Andrea,
Our "rule" is that each domain should be at least 100x100 (e_we and e_sn values). Take a look at this page for best practice suggestions on setting up your domain.
Dear all,

the suggested page for best practice on setting up the domains concerns the relationship between the gross domain and the higher resolution internal one.

In my run I consider 3 domains nested (15000-5000-1000).
Is there a rule that must be respected about the distance between the points of interest and the edge of the third domain that guarantees the reliability of the forecast? otherwise is it possible to consider the forecast for each point of the inner domain including the border ones as reliable?

Thanks a lot,

In the best practice web page link I sent above, there should be some information regarding the recommended distance of nests from their parent grid. We say that you should have about 1/3 of the parent domain surrounding all sides of the child. More information can be seen in the link above.
Hi Kwerner,
thank you very much for your answer and for the clarity.
I am sorry I was not so clear in the question I try to rephrase.
If the rule of 1/3 distance from the edge of the parent domain is respected as in the attached image, is the reliability of the prediction at points A and B the same? Or is that of point A better because it is further away from the edge?


  • WRF_domine.pdf
    12.4 KB · Views: 55
It's hard to say for sure without some testing and verification with observations, but I would assume point A would be a little better since it won't have any boundary interference. You should always make sure to have enough space around the area of interest, and not try to put the inner domain as close as possible to the interest point(s).