GFS - MSLET vs PRMSL issue. Can UEMS initialize on MSLET?

Questions and other topics related to UEMS 18.
Post Reply
norulz
Posts: 92
Joined: Thu Mar 12, 2015 1:43 pm

GFS - MSLET vs PRMSL issue. Can UEMS initialize on MSLET?

Post by norulz » Sun Sep 02, 2018 12:20 pm

There appears to be a relevant debate going on regarding the accuracy of GFS PRMSL after smoothing is applied. It is claimed that the smoothing of PRMSL reduces resolution of the isobars significantly and this can lead to inaccurate forecasting.

MSLET is also pressure reduced to MSL but without smoothing applied.

Links with more information on this: https://ams.confex.com/ams/33HURRICANE/ ... 40155.html
https://ams.confex.com/ams/33HURRICANE/ ... 018-V4.pdf
http://luckgrib.com/blog/2018/08/28/gfs ... ssure.html

My understanding is that UEMS initializes with PRMSL when using GFS data. The personal tile server (gfsp25pt) includes MSLET in the initialization gribs but is it being used?
How can one force prep into using MSLET and to disregard PRMSL?

David
Last edited by norulz on Mon Sep 03, 2018 11:46 am, edited 2 times in total.

lcana
Posts: 70
Joined: Wed Nov 30, 2011 4:34 pm

Re: GFS - MSLET vs PRMSL issue. Can UEMS initialize on MSLET?

Post by lcana » Mon Sep 03, 2018 10:16 am

Very interesting topic, Norulz!

I'd like to know it too. If anyone knows about this, please share it with us

Best,

Luis

dcvz
Posts: 1
Joined: Tue Oct 23, 2018 7:24 am

Re: GFS - MSLET vs PRMSL issue. Can UEMS initialize on MSLET?

Post by dcvz » Tue Oct 23, 2018 7:45 am

Thanks for raising this issue. It's easy to use MSLET. 1) Make sure MSLET is in your GFS grib2 file and 2) edit the Vtable.GFS by finding the line for Sea-level pressure (PMSL) and change the GRIB2 param (next to last column) from 1 to 192. That's it.

I've tested it on a hurricane case using ARW and it produces only minor changes in the forecast. The reason for that is because while PRMSL is smoothed, all the other fields (U, V, T, Q, and GHT) are not. The model quickly adjusts to the state given by U, V, T, Q, and GHT.

In the poster, the authors are talking about forecasts made from plots of the GFS PRMSL. Not about numerical forecasts initialized from PRMSL. Nevertheless, it's probably best to use MSLET now that NCEP supplies it.

lcana
Posts: 70
Joined: Wed Nov 30, 2011 4:34 pm

Re: GFS - MSLET vs PRMSL issue. Can UEMS initialize on MSLET?

Post by lcana » Tue Oct 23, 2018 10:50 am

Great!
Thanks for the info dcvz!

Regards,

Luis

norulz
Posts: 92
Joined: Thu Mar 12, 2015 1:43 pm

Re: GFS - MSLET vs PRMSL issue. Can UEMS initialize on MSLET?

Post by norulz » Tue Oct 23, 2018 1:12 pm

Thanks for your explanation @dcvz.
I''ll give it a try but as you say it will most likely be hard to determine a difference.
David

Post Reply