Board index FlightGear Development Weather

Discontinuity in the weather

Everything related to weather simulation, visuals should be discussed in the shader subforum.

Re: Discontinuity in the weather

Postby Thorsten » Mon Mar 19, 2018 6:22 am

I've identified the file and relevant functions here such that you can look into the code yourself if you believe something is wrong - given that you improve the Concorde, you appear to be quite able to deal with Nasal code, so you can just hop in, follow what the routines do and go looking for bugs.

I've never seen discontinuities in my tests - doesn't mean they're not there, they can have been accidentially introduced, or be triggered by something exotic, but I plainly don't have the time to go looking myself whether I manage to trigger the same situation.
Thorsten
 
Posts: 9997
Joined: Mon Nov 02, 2009 8:33 am

Re: Discontinuity in the weather

Postby wkitty42 » Mon Mar 19, 2018 5:42 pm

could these discontinuities be related to the load on the system and how long updates take at different times?

would having a (eg) list of METARs from say 10 airports within a 50nm range and averaging them at the "border" with each other work to prevent these rapid changes? maybe more than 10 would be needed? possibly could eliminate some that have identical weather?
"You get more air close to the ground," said Angalo. "I read that in a book. You get lots of air low down, and not much when you go up."
"Why not?" said Gurder.
"Dunno. It's frightened of heights, I guess."
User avatar
wkitty42
 
Posts: 4762
Joined: Fri Feb 20, 2015 3:46 pm
Location: central NC, USA
Callsign: wk42
Version: git next
OS: Kubuntu 14.04.5

Re: Discontinuity in the weather

Postby Thorsten » Mon Mar 19, 2018 6:19 pm

I've tested continuity of parameters rather extensively when writing AW and I've never seen the wind jump if the chosen wind model didn't have jumps - so I can't really comment what causes the things I've never seen...
Thorsten
 
Posts: 9997
Joined: Mon Nov 02, 2009 8:33 am

Re: Discontinuity in the weather

Postby BEL ISAAK » Wed Dec 05, 2018 5:42 pm

Sorry for reviving this old topic, but I just wanted to add that I 've seen these sudden jumps very often on many flights, some short haul, some long haul. It mostly occurs when crossing an ocean, where there is a big difference in wind direction and/or speed, but it does also happen over the mainland with plenty of METAR stations nearby. I also use the aloft waypoints wind model with real weather normally. I don't think this adds something to the discussion other than that it's not only happening on one machine.
User avatar
BEL ISAAK
 
Posts: 208
Joined: Sat Jun 04, 2011 2:52 pm
Location: Leuven, Belgium
Callsign: OO-ISA
Version: 2018.2.2
OS: Windows 10

Re: Discontinuity in the weather

Postby Thorsten » Wed Dec 05, 2018 6:13 pm

Just to clarify once again - weather (winds, also pressure,...) can change *rapidly* when a new METAR comes in (we need to accomodate the station before you actually reach it even if you move with Mach 2 and we have no control over whether reports are consistent) but it can't change *instantaneously* as the new station is not only spatially interpolated but also faded in over time.

Since these rapid changes are fundamentally caused by the need to treat inconsistent and incomplete 'real' weather reports - use offline weather and see them gone.
Thorsten
 
Posts: 9997
Joined: Mon Nov 02, 2009 8:33 am

Re: Discontinuity in the weather

Postby legoboyvdlp » Wed Dec 05, 2018 6:41 pm

We really can't get past this without some form of aloft winds data provider - I wonder whether a service such as 'windy' could be used as a data provider... If the plugin to get the data is GPL, does the data that it fetches need to be GPL? I suppose this is one of those gray areas...?
User avatar
legoboyvdlp
 
Posts: 5834
Joined: Sat Jul 26, 2014 1:28 am
Callsign: YV-LEGO
Version: 2018.3.1
OS: Windows 10 HP

Re: Discontinuity in the weather

Postby Thorsten » Thu Dec 06, 2018 8:04 am

If the plugin to get the data is GPL, does the data that it fetches need to be GPL?


No, data is data whether you process it with GPL software or not - what matters is the terms of use of your provider - if what we do is okay with the provider, then we're good.
Thorsten
 
Posts: 9997
Joined: Mon Nov 02, 2009 8:33 am

Re: Discontinuity in the weather

Postby Mihajlo » Thu Dec 06, 2018 4:59 pm

Regarding aloft winds X-plane uses NOAA aloft wind data for the whole world. The link to NOAA GFS winds aloft data can be found on this page: https://aviation.stackexchange.com/ques ... plications
but I am not sure this is the exact source X-plane uses.
Mihajlo
 
Posts: 34
Joined: Tue Aug 28, 2018 5:47 pm

Re: Discontinuity in the weather

Postby Thorsten » Thu Dec 06, 2018 5:48 pm

Whatever the source is - someone has to

1) write code that fetches the data as needed
2) represents it in the property tree
3) inserts it into AW

I'm entirely willing to do 3), but I'm not going to do anything about 1) or 2) because I'm just not interested at all in METAR-driven weather (personally I think offline weather simulation has a lot more potential because of all the details that can be consistently generated) so someone else needs to do it.

All else is nice, but won't produce any change to FG.
Thorsten
 
Posts: 9997
Joined: Mon Nov 02, 2009 8:33 am

Re: Discontinuity in the weather

Postby V12 » Fri Dec 07, 2018 5:58 am

Is possible to feeds weather data from external app to the FG's property tree ? I tested FSX with free real weather utility FSXWX, this piece of software periodicaly downloads NOAA data, make interpolated grid of fake METAR stations and calculated data sends to FSX. On my test flights acros Europe from Olbia (LIEO) to Bratislava (LZIB) with flyover LFMN, LSGS and LOWI (https://skyvector.com/?ll=44.6483347187 ... OWI%20LZIB) I did not observed any discontinuity in baro pressure, or suddenly wind changes. Weather smoothly changed from clear sky over Olbia to full overcast and rain over Bratislava. There was very nice clouds formation over Alps. Same result in early morning flight from EGLL to LZIB.
Fly high, fly fast - fly Concorde !
User avatar
V12
 
Posts: 674
Joined: Thu Jan 12, 2017 4:27 pm
Location: LZIB
Callsign: BAWV12

Re: Discontinuity in the weather

Postby Thorsten » Fri Dec 07, 2018 6:01 am

Sure - it's OpenSource, near anything is possible - you just need to write the code.

(And deal with all the complaints of people who don't like your code of course, I won't hide that disadvantage from you...)

***

Edit: AW has had the capability to interpolate in space and time for an irregular grid of weather observations and wind data basically since the beginning. The capability was added by yours truly because people interested in real weather (which I'm most definitely not) proposed that FG at some point would be able to fetch such data.

In actual fact, weather-engine side all is there but the data (as you can easily check by inserting your own data manually, that capability also exists). What's mystifying to me is that among all the people who supposedly are interested in real weather, over the years no one could be bothered to make use of that capability.

Which, in a nutshell, is why we don't have a better representation of aloft winds (and likely won't have it for the coming decade).
Thorsten
 
Posts: 9997
Joined: Mon Nov 02, 2009 8:33 am

Re: Discontinuity in the weather

Postby V12 » Fri Dec 07, 2018 7:56 am

Aloft wind was in FG some years ago based on Jepessen weather service. But Jepessen changed licencing policy and today, this feature doesn't work.
Check http://wiki.flightgear.org/Howto:Fetch_live_aloft_data
Fly high, fly fast - fly Concorde !
User avatar
V12
 
Posts: 674
Joined: Thu Jan 12, 2017 4:27 pm
Location: LZIB
Callsign: BAWV12

Re: Discontinuity in the weather

Postby Thorsten » Fri Dec 07, 2018 8:09 am

Aloft wind was in FG some years ago based on Jepessen weather service.


I'm not aware of anyone ever bothering to feed that data to AW and I'm not sure whether the technology ever got beyond patching and rigging external tools.

Which, in a nutshell, is why we don't have a better representation of aloft winds (and likely won't have it for the coming decade).
Thorsten
 
Posts: 9997
Joined: Mon Nov 02, 2009 8:33 am

Re: Discontinuity in the weather

Postby V12 » Fri Dec 07, 2018 9:29 am

Thorsten wrote in Fri Dec 07, 2018 8:09 am:I'm not aware of anyone ever bothering to feed that data to AW...

Minimum 2 peoples - me and author of that wiki page :)
Fly high, fly fast - fly Concorde !
User avatar
V12
 
Posts: 674
Joined: Thu Jan 12, 2017 4:27 pm
Location: LZIB
Callsign: BAWV12

Re: Discontinuity in the weather

Postby Thorsten » Fri Dec 07, 2018 11:24 am

Well, I've never seen any patch designed to feed the Jepperson data to AW, nor see any patch described on the Wiki page, nor have I ever been contacted to implement such, nor did you seem aware of the fact that AW has that capability - so I hold on to my statement.
Thorsten
 
Posts: 9997
Joined: Mon Nov 02, 2009 8:33 am

PreviousNext

Return to Weather

Who is online

Users browsing this forum: No registered users and 3 guests