Welcome, Guest
Username: Password: Remember me
26 Sep 2020
No bulk downloads - PLEASE! THIS IS OVERLOADING OUR SERVER AND DENYING ACCESS TO COMMUNITY USERS
SVP - Pas de téléchargement en masse ! Cela surcharge le serveur et empêche la communauté d'utilisateurs d'accéder aux données
_______________________________
WE WILL BLOCK IP'S THAT ARE BULK DOWNLOADING! Nous bloquerons les IPs des personnes qui agissent de la sorte.
PLEASE GO DIRECTLY TO THE GRIB PRODUCERS (NOAA, DWD...) FOR BULK DOWNLOADS. Allez chercher les données directement à la source (NOAA, DWD ...)
Read More...
  • Page:
  • 1
  • 2

TOPIC:

WRF boundary conditions - GFS v16 8 months 1 week ago #877

  • ameerrh
  • ameerrh's Avatar Topic Author
  • Offline
  • Junior Member
  • Junior Member
  • Posts: 25
  • Thank you received: 0
Hi,
The Global Forecast System (GFS) was upgraded to version 16 on March 22nd. The main change to the model includes Increase in the number of model vertical layers from 64 to 127.
I know that the Boundary conditions for WRF 12km & 4km are provided by GFS forecast data.
ِAre Boundary conditions now provided by the new version of GFS model?

Thanks

Please Log in or Create an account to join the conversation.

WRF boundary conditions - GFS v16 8 months 1 week ago #878

Hi,

Yes, all the openWRF forecasts started using GFS v16 for both initialization boundary conditions.

The increase of vertical layers to 127 in internal to the GFS simulation. NOAA only publishes about 40+ vertical layers as before, so WRF is using the same number of input layers. However, the quality of the GFS forecast is now better.

David

David

Please Log in or Create an account to join the conversation.

WRF boundary conditions - GFS v16 8 months 1 week ago #879

  • ameerrh
  • ameerrh's Avatar Topic Author
  • Offline
  • Junior Member
  • Junior Member
  • Posts: 25
  • Thank you received: 0
Hi,

Does WRF use about 40 vertical layers or 64 layers?

Please Log in or Create an account to join the conversation.

WRF boundary conditions - GFS v16 8 months 1 week ago #880

In WRF with the ARW core it is possible to set any number of vertical levels and since version 4 of WRF there are possibilities to set the distribution of the layers for different levels of the atmosphere.
In openWRF we are using the default of 45 layers with automatically determined distribution of the layers.

Please Log in or Create an account to join the conversation.

WRF boundary conditions - GFS v16 8 months 1 week ago #881

  • ameerrh
  • ameerrh's Avatar Topic Author
  • Offline
  • Junior Member
  • Junior Member
  • Posts: 25
  • Thank you received: 0
Are there any plans to increase the number of vertical layers of WRF for example doubling it from 45 layers to 90 layers?
I know that the increase of vertical layers greatly improves the performance and accuracy of the model.

Please Log in or Create an account to join the conversation.

WRF boundary conditions - GFS v16 8 months 1 week ago #882

I know that the increase of vertical layers greatly improves the performance and accuracy of the model.

What is your basis for this claim?

There is no magic. Research into the valued of increasing WRF number of vertical layers shows very limited improvements. This has some value in the area of tropical storms and hurricanes. Even then there is no improvement in forecasting the track of the storm but only slightly better data on upper levels.

A ratio needs to be maintained between spatial resolution and vertical resolution and this means exponential increase in run time for increasing vertical resolution.

For WRF the default of 45 layers is considered optimum. The balance of the forecast domain, the choice of spatial resolution, the choice of physics and dynamic schemes, the choice of calculation time-steps and other setting all have significantly more impact on the quality of a forecast than extra vertical resolution.

Please Log in or Create an account to join the conversation.

  • Page:
  • 1
  • 2
Time to create page: 0.031 seconds

Resources on this site will always be free and open with no commercialization. However, operating costs of the servers to pre-process and deliver GRIB files to XyGrib are not trivial.

If you found the site useful, please share in the effort and help keep us up and running

 

Please Support

Go to top