The complexities of the ATCF (Automated Tropical Cyclone Forecasting) system
Posted by Chris in Tampa on 7/24/2011, 10:00 pm
We're in an inbetween period right now. While it has not officially been deleted or deactivated, they are no longer updating the file. It was last updated at 12Z (8AM EDT) this morning. Sometimes this means they forgot to delete it or deactivate it and you simply have to drop it from your site after awhile of no position updates or it may mean that since there is some chance it could be tracked again, they are holding off on declaring that it won't be coming back. (They could always bring it back, but once you put out one of the delete or deactivate files, it seems like they are usually saying that the chance is very remote and they don't intend for it to come back.) I still don't know what the difference is between delete and deactivate.

Right now because of this inbetween period, some sites will carry it and some will not. The way I do my site in my new system, which is not an official method since I don't know of one, is after 12 hours of no position updates say: "A track update has not been made to this storm recently." I think I actually have it set at about 13 hours to try to make sure that two position updates have not been made. I remove the storm from being active if there is a delete or deactivate file for it. I keep it around for 24 hours if there is no update for the best track and no delete or deactivate file. That is just me being really, really cautious. Sometimes the ATCF system has some errors and you have to program your site to deal with it and the problem is no two sites probably do it alike.

Since they do not have best track positions right now, they missed two, it usually means it is not trackable at the moment. (the actual center) If it does come back and becomes trackable again, then they will make an estimate of where it was for the positions missed.

You get position updates every 6 hours. (0Z, 6Z, 12Z, 18Z) Sometimes for weaker storms they skip the 6Z (2AM EDT), or just do a best track update and no models. The actual position updates are actually posted to the file usually around 30 to 45 minutes after the time they are valid. So today they did not do a 18Z and they have yet to do a 0Z. The 0Z (8PM EDT) should have been posted usually around 0:30 to 0:45Z (sometime before 9PM EDT usually).

So right now it's hard to say what its official invest status is. But, at the moment it appears the center is not trackable. NRL developed the ATCF system. You can read more about it here:
http://www.nrlmry.navy.mil/atcf_web/
I think the NHC is the only one to actually make any real time adjustments to best track information for the Atlantic and East Pacific. CPHC for the Central Pacific. I'm not sure who tries to fix the errors. I have talked with the NHC sometimes when they have an error and have heard other people relay information as well from the NHC when they are trying to fix something, but NRL might help to fix errors as well. As for posting info though, I think that is probably just the NHC since they are the official agency for the Atlantic and East Pacific and then the CPHC for the CP. It would be weird if more than one agency was changing things. I think that is the case because I have seen pictures and maybe even video of the ATCF system software and it seems very much like the forecaster must be the one to actually do it. (And the system can allow for a single forecasters initials that can go into it as well.)

Keep in mind the ATCF system is raw. It has errors all the time. It's complicated trying to program around the errors and dealing with them is up to each site you view the data on. This inbetween period could be an error or just the normal period you have before declaring it very likely to be dead by deleting or deactivating it. We'll know tomorrow.
93
In this thread:
90L Deactivated - Doorman, 7/24/2011, 2:09 pm
< Return to the front page of the: message board | monthly archive this page is in
Post A Reply
This thread has been archived and can no longer receive replies.