NOAA Wind Updates

[climbator]

Thanks to Omar for sharing this excellent pice of software and thanks to everyone for contributing to the Wiki! While reading through the NOAAWinds section in the Wiki (http://www.tecepe.com.br/wiki/index.php?title=NOAAWinds) I was wondering about two issues:

1) Game vs. VRTool: It appears that in the current VVOR14 game the forcasts are not as static as described in the article. According to my observations the forcast values in the game window (12H..72H) are updated at 6:00/10:00/16:00/22:00 UTC, i.e. exactly at the same time instances as the VRTool download behaviour changes. Is that correct?

2) NOAA Server vs. VRTool: According to my observations new forecast values on the NOAA server are available quite a bit earlier than described in the article (e.g. the t06z folder is created around 8:00 UTC, forcasts up to f114 are available around 9:00 UTC, all forcasts up to f384 are available around 9:40 UTC). VRTool does not use these new values before 10:00 UTC, but downloads the old ones instead (t00z in that case). My question is therefore: Is it possible to adjust the VRTool "switching points" (10:00 UTC etc.) manually or are they hard coded in the software?

[mclean]

Hi Climbator

I helped Omar with that part of the wiki, and a lot of it was written for the last VVOR. You may be right about the game forecasts updating earlier and more often this time. I'll check it out and update it. The way VR did the change to CET obviously alters it as well. Also not sure about the NOAA runs starting earlier - seems unlikely but I'll check that as well. Thanks for the tips.

[mclean]

This from the NOAA server production status page at http://www.nco.ncep.noaa.gov/pmb/nwprod/prodstat/

Code: Select all

Cycle  Start  End
t00z   03:18  04:38
t06z   09:17  10:37
t12z   15:17  16:36
t18z   21:20  22:39

All UTC, average start and end times for the runs.

BTW, you should be able to trick VRTool into downloading the new cycle earlier by tweaking the time zone setting in the NOAA module.

[omar]

climbator,

As you see, the whole publication cycle takes about 1:20.
If you are interested only on the first few forecasts,
you can download earlier, by doing that time zone tweak...

You are right about game winds: forecasts 12, 24 etc
are not static, but updated during the day.
I have just seen a forecast change at 17:00 UTC
This must be the T12

[climbator]

Hi omar and mclean,

thank you both for your answers.

- VVOR winds updates: Looks like VRTool inspired the game designers to implement a new feature

- NOAA Server: With the "availability times" in my original post I was referring to the drop-down list in the respective subfolder at http://nomads.ncep.noaa.gov/cgi-bin/filter_gfs.pl. I have no idea how that list relates to the overall produciton status and whether that list really represents the data which is requested by VRTool, but it grows rather quickly. That was my original motivation for trying to download new forcasts earlier, especially for short-term forecast and routing.

- Time zone tweaking: Thanks for the tip, I'm curious to check it at the next download cycle.

[climbator]

I just made a few interesting observations that might be helpful to clarify the issue (please correct me in case I'm wrong):

- Both the VVOR game time and the NOAA server updates are affected by the daylight-saving time, i.e. the offset w.r.t. to UTC is +1 in winter time and +2 in summer time. The consequences are:

- NOAA server production times are (in case of the original t06z example): 09:17 - 10:37 UTC in winter, 08:17 - 9:37 UTC in summer.

- IMHO, the very instructive timescale figure in the Wiki is valid only in summer (game time = UTC+2).

- The time delay between NOAA file availability and their actual usage in the game is constant in winter/summer

- The recommended time tweak (thanks for the tip!) works excellent! What worked for me: Tweaking the time zone by +2, allowed me to download the first smoking hot t18 forcasts already at 22:30 CET / 21:30 UTC. Roughly speaking, an extra forcast advantage of 1.5 hours is possible for short-term forcasts (up to 24/36H), 1 hour for mid-term forcasts (120H).

[mclean]

Hi climbator

You've raised some interesting points. The page on the wiki needs an update but there are a few things to pin down first:

- Game updates: You are right, the game is doing frequent updates of the forecast gribs. Today I've seem them at 23:00 and 05:00. Omar at 17:00. Yours at 06:00, 10:00, 16:00, 22:00. So that's all over the place. I wonder if the game is smart enough to check every hour and recognise if there is new data? Needs a bit more observation.

- NOAA server updates: I think you are saying that the cycle folder starts populating earlier than what the production status page says. I'll check that out. But I would be surprised if the updates were synchonised to a summer/winter time zone rather then UTC, because so many applications world-wide depend on them. Not the sort of thing the Americans would do I think.

- The timing table in the wiki: Correct, in the last VVOR everything was done to UTC. Not his time unfortunately. I'll probably re-do it for the current game with the CET timing - at least it will work for the next six months.

[climbator]

Hi mclean,

good point. I'm afraid what adds confusion here is that the numbers posted throughout this thread experienced a time shift as well.

- Game updates: From what I've seen so far, the VVOR14 game forcast always updates at 6/12/18/24 CE(S)T sharp. This translates to 4/10/16/22 UTC in summer (my original post, the "6" instead of "4" was a typo) and 5/11/17/23 UTC in winter (what you observe now).

- NOAA server updates: I absolutety agree with you. One would rather expect the production cycles to be tied to UTC. However what I've seen so far suggests the contrary (in case of the t06z production cycle): always 10:17 - 11:37 CE(S)T = 09:17 - 10:37 UTC in winter = 08:17 - 9:37 UTC in summer.

Anyway, I'm happy to have a working VRTool setting that allows me to do the weather routing 1..1,5 hours earlier that what I see in the game forecast (15..15,5 hours of 100% sure winds). Thank you!

[climbator]

Doh! Looks like I fooled myself, sorry for confusing everyone!

I just realized that all the observations I'm 100% sure of were made after the time shift, i.e. in CET (UTC+1). So please forget all my statements regarding game wind updates and NOAA server production times at CEST (summer time), I'm simply not sure of anymore. No other chance than to wait for 6 months for final clarification.

[mclean]

I've updated the page - I think it's about right now.