Chantal, GFS Bogusing and How You Can Monitor TC Initialization with the GFS


GFS 24 forecast of 500 MB height (white), PMSL (black), and 6 hour QPf valued at 12 UTC Friday 12 July 2013.

Yesterday provided another opportunity to examine and see how bogusing can impact the GFS performance with tropical cylones (TCs) and how through a little work, forecasters can identify whether bogusing is occurring.  We discussed the relocation and bogusing used in the GFS to some degree last year in the blog post – TD Nine (Isaac) GFS Initialization and Newly Available GFS Relocation & Bogusing Information from EMC while this post provides some background on bogusing and vortex relocation with the GFS and a link to a presentation on “How are tropical cyclones represented in operational model initial conditions.”

Forecasters on Thursday afternoon (7/11) noticed that the GFS was more aggressive in redeveloping the remnants of Chantal compared to their expectations and some of the other NWP guidance. Later in the day, Mike Brennan from NHC coordinated via a quick chat message to the southeastern U.S. WFOs alerting them that the forecast was likely overdone and suspect (his chat message is shown below).




So what happened here?  When the NHC runs guidance on a system regardless of whether it is an active TC or just an invest, a “TC vitals” file is created that is sent to the NCEP supercomputer.  This file has information on the location, intensity, and structure of the TC.  Here is an example from 1800 UTC:

NHC  03L CHANTAL   20130711 1800 224N 0780W 315 098 1013 1014 0278 13 074 -999 -999 -999 -999 S -999 -999 -999 -999 -9 -99N -999W -999 -999 -999 -999

For numbered TCs (such as “03L” for Chantal, the third storm in the Atlantic basin), the NCEP GSI data assimilation (DA) scheme reads these files to determine if a vortex relocation or bogusing has to be is needed for that TC for that GFS cycle.  The vortex relocation is used when the GFS generates a strong enough circulation but it needs to be moved to the NHC analyzed position.  Bogusing usually only occurs for systems where there is a very weak representation of the system in the model analysis and a synthetic circulation is added to ensure that the model initializes with some reasonable representation of the system.

If the system is weak such as an invest with a “90s” number, then this process does not occur and there is no vortex relocation or bogusing.  However, for former TCs like Chantal, the NHC will often hold onto the active TC number (03L) for a couple of reasons, most notably in case the storm regenerates. Currently, the GFS GSI DA scheme only looks at the system number to determine if it’s an active TC or not, and doesn’t look at other information such as the “status” that NHC assigns. So on Thursday, the NHC still had Chantal referenced as “03L” and the GFS GSI at 12Z and 18Z thought Chantal was a real TC.

To prevent this from happening again with Chantal, the NHC renumbered Chantal back to AL96 on Thursday afternoon in time for the 00 UTC cycle. The NHC is communicating with EMC about how to address this issue in the future.

So how can you look this up with a little detective work?

•    First go to the Nomads web site –
•    Next, click on the folder for the particular GFS run you are interested in (“gfs.2013071112″ for example)
•    Then look for the relocate file named in this format “gfs.t[HH]z.inform.relocate.tm00” (such as “gfs.t12z.inform.relocate.tm00” for the 12 UTC cycle)

The “gfs.t12z.inform.relocate.tm00” file will look something like the example below for Chantal from 12 UTC on 07/11. Note how it says “bogused” near the bottom. That means that a synthetic circulation was inserted into the GFS and located at the NHC position.






The “gfs.t18z.inform.relocate.tm00” file from 6 hours later at 18 UTC on 7/11 is shown below. Note how this file does not mention “bogused” near the bottom. That is because by the 18 UTC run, the GFS had a decent enough circulation present, likely resulting from the first guess of the 6-hour forecast from the 12 UTC run, that it did not need an artificial vortex inserted but it was relocated to the NHC position.






I need to acknowledge the assistance of Dr Michael Brennan (NHC) with this post who provided most of the information and his assistance is greatly appreciated.

This entry was posted in TC Improved Initial Conditions. Bookmark the permalink.

2 Responses to Chantal, GFS Bogusing and How You Can Monitor TC Initialization with the GFS

  1. skeighton says:

    This is a great summary/explanation Jonathan; thanks! Based on NHC-NCEP coordination, why couldn’t this be added to the NCEP model status messages sent by the SDM when this is known to be a factor? That would avoid having forecasters go through the detective work steps you outlined.

  2. Jonathan Blaes @ WFO RAH says:

    Hey Steve,

    We have recently reached out to the WPC and provided them this blog post as a training item, another post from last year and Dr Lackmann’s training presentation. I think they plan on including it in the model diagnostic discussion.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s