Met'Map : Your new online tool for METARs & TAFs... Powerful & Free !!!!!!
Absolutely first class .. I shall certainly add this to my opening pages on my browser.
Couple of questions though ....
It would be terribly useful to have a zoom feature .. when bimbling about from Fairoaks EGTF (doesn't have it's own met but about 10 miles SW of EGLL and 5 miles NW of EGLF), I am not overly interested in what is happening over the whole of Wales, Scotland and Eire (other than an initial "see at glance" check on the different colours to see where the weather is.) and would like to just see the airfields within say 50 miles of either my base or destination.
Would it be possible for the "base chart" to be based on one of the aeronautical topographical charts ??
That is the way it works in the USA ..... http://skyvector.com/ granted they don't have the plain language feature. But you can get that from DUATS if you need it.
Please don't think I am being picky .. great job well done.
Dave
Couple of questions though ....
It would be terribly useful to have a zoom feature .. when bimbling about from Fairoaks EGTF (doesn't have it's own met but about 10 miles SW of EGLL and 5 miles NW of EGLF), I am not overly interested in what is happening over the whole of Wales, Scotland and Eire (other than an initial "see at glance" check on the different colours to see where the weather is.) and would like to just see the airfields within say 50 miles of either my base or destination.
Would it be possible for the "base chart" to be based on one of the aeronautical topographical charts ??
That is the way it works in the USA ..... http://skyvector.com/ granted they don't have the plain language feature. But you can get that from DUATS if you need it.
Please don't think I am being picky .. great job well done.
Dave
Thread Starter
Join Date: Jan 2008
Location: Europe
Posts: 34
Likes: 0
Received 0 Likes
on
0 Posts
Dear Dave
Thanks for your positive feedback.
If you click on one airfield (or if you request a specific ICAO code on the left formular) you should get a resized map, centered on the airfield you requested.
Example for EGKA :
http://www.orbifly.com/ORBITEST/metm...ao_choose=EGKA
You can even bookmark that direct link with your favourite airfield.
In that case, is the zoom factor enough or would you like a bigger zoom factor ? This is something feasable I think, if needed.
Best regards,
Alexandra
ORBIFLY
Thanks for your positive feedback.
If you click on one airfield (or if you request a specific ICAO code on the left formular) you should get a resized map, centered on the airfield you requested.
Example for EGKA :
http://www.orbifly.com/ORBITEST/metm...ao_choose=EGKA
You can even bookmark that direct link with your favourite airfield.
In that case, is the zoom factor enough or would you like a bigger zoom factor ? This is something feasable I think, if needed.
Best regards,
Alexandra
ORBIFLY
Thanks Alexandra, I obviously hadn't experimented enough.
Personally I think that zoom factor is fine, it is enough to "unclutter" overlapping airfield symbols and so is all that is needed.
Dave
Personally I think that zoom factor is fine, it is enough to "unclutter" overlapping airfield symbols and so is all that is needed.
Dave
Join Date: Dec 2006
Location: Amsterdam - EHLE
Posts: 58
Likes: 0
Received 0 Likes
on
0 Posts
Any chance of expanding the map a bit further East
Hi Alexandra,
Is there any chance of getting airfields added for Ukraine? I do a lot of flying out of there and would really find this helpful.
Thanks,
Blue Albatross
Is there any chance of getting airfields added for Ukraine? I do a lot of flying out of there and would really find this helpful.
Thanks,
Blue Albatross
Join Date: Feb 2007
Location: Amsterdam
Posts: 4,598
Likes: 0
Received 0 Likes
on
0 Posts
There are some errors in the METAR decode code. For instance this one, right now:
2008/01/21 15:57 EGVA 211557Z AUTO 9999 FEW012 OVC022 10/10 A2971 RMK AO2 PK WND 27026/1556 RAE1557 CIG 016V040 SLP061 $
Makes it into "very intense IFR conditions" due to a visibility of 12 meters....
Because the wind component in the METAR is missing, wind is reported as 999/99 knots and the FEW012, I think, is seen as visibility.
Otherwise, very useful.
2008/01/21 15:57 EGVA 211557Z AUTO 9999 FEW012 OVC022 10/10 A2971 RMK AO2 PK WND 27026/1556 RAE1557 CIG 016V040 SLP061 $
Makes it into "very intense IFR conditions" due to a visibility of 12 meters....
Because the wind component in the METAR is missing, wind is reported as 999/99 knots and the FEW012, I think, is seen as visibility.
Otherwise, very useful.
Join Date: Oct 2001
Location: Ready to Depart
Age: 45
Posts: 291
Likes: 0
Received 0 Likes
on
0 Posts
Good spot!
May I suggest, if I've guessed your method correctly, that you use Regular Expressions to match code groups, rather than splitting on spaces and assuming a given order.
Vis will always be nnnn, cloud will always be xxxnnn, if no matches then look for CAVOK.
May I suggest, if I've guessed your method correctly, that you use Regular Expressions to match code groups, rather than splitting on spaces and assuming a given order.
Vis will always be nnnn, cloud will always be xxxnnn, if no matches then look for CAVOK.
Thread Starter
Join Date: Jan 2008
Location: Europe
Posts: 34
Likes: 0
Received 0 Likes
on
0 Posts
Dear all,
Thanks again for feedbacks !
The decoding method is quite dense, because we had to take into account the fact that sometimes there are mistakes in reports.
For example, when no visibility known, there should be //// , no wind should be //////KT , no wind direction should be ////15KT, but sometimes they omit completely the block, or mix them up.
Some pressures are in inches (ETAR Ramstein air base, for example).
Some visibilities are in SM (US weather), other are in meters. Sometimes even kilometers ! (have seen "70KM" in austria).
Some winds are in KT, some other in MPS (Sofia, for example).
Temperatures were also a deep problem. because sometimes missing, sometimes not at the right position...
Sometimes clouds are BKN050, but sometimes they are BKN050TCU or BKN050CB or BKN050/// !
Automatic METARs have also some "special features". :-)
etc.
Some deeper mistakes sometimes : I've seen RA (rain) given before wind !
And even once, the 2000 meters visibility had been written with the last character : letter O instead of number 0 ! (two - zero - zero - O letter)
Which makes no difference for your eyes, but was not decoded properly by the computer.
So we had to take into account, not only the blocks in order, but also to analyze each block.
So first we split the message : all until weather, and decode that first part.
Includes RVR decoding if needed, which is a long sequence to interpret correctly because there are quite a lot of possible suffixes (until now, max 2 different RVR are decoded per airfield, but soon I'll modify that and there will be no limit thanks to a loop... I've seen one single metar with 5 different RVR once !).
Then look for CAVOK.
Then extract special blocks (pressure, temp, NOSIG...) and analyze.
Then a weather analysis thanks to a weather codes database (Rain, Sand, Ice crystals, Fog...).
Then a progressive loop to analyze layers and find ceiling, except if CAVOK, SKC, CLR, NCD.
Then add suffixes (some BECMG and special features not usual in METARS)
So we haded some scripts to make it able to also interpret codes which are usually not allowed by the strict METAR rules, but that are used by some meteorologists.
Each time we find a bug (99% due to an errorwhen the METAR was edited, not respecting the legal form) we modify the code to make it able to handle such errors and give a correct interpretation, that's why the code becomes more and more dense as time goes by
, and interpretations are more and more reliable (those few errors usaually make the script to considere weather stronger than it is actually).
But we still provide the original METAR, because there is no machine nor computer able to handle all possible human errors :-)
And we also have to pay attention to the fact that the code mustn't become too heavy, otherwise the page loading will be excessively long (especially for the whole European chart, there are nearly 500 airfields depicted).
Thanks BackPacker to have reported the error with the missing wind, you are completely right that is the reason for the wrong decoding of vis. I'll check that next week end to improve the code !
Alexandra
ORBIFLY IFR SCHOOL
PS : For Ukraine : next time we improve the chart I'll remember to extend it more east ! Right now there are only 2 fields available on the map : LRTR Timisoara and LRSM Satu Mare, with winds in meters per second :-)
I assume you would have liked Kiev, Minsk, Sebastopol and Chisinau ?
Thanks again for feedbacks !
The decoding method is quite dense, because we had to take into account the fact that sometimes there are mistakes in reports.
For example, when no visibility known, there should be //// , no wind should be //////KT , no wind direction should be ////15KT, but sometimes they omit completely the block, or mix them up.
Some pressures are in inches (ETAR Ramstein air base, for example).
Some visibilities are in SM (US weather), other are in meters. Sometimes even kilometers ! (have seen "70KM" in austria).
Some winds are in KT, some other in MPS (Sofia, for example).
Temperatures were also a deep problem. because sometimes missing, sometimes not at the right position...
Sometimes clouds are BKN050, but sometimes they are BKN050TCU or BKN050CB or BKN050/// !
Automatic METARs have also some "special features". :-)
etc.
Some deeper mistakes sometimes : I've seen RA (rain) given before wind !
And even once, the 2000 meters visibility had been written with the last character : letter O instead of number 0 ! (two - zero - zero - O letter)
Which makes no difference for your eyes, but was not decoded properly by the computer.
So we had to take into account, not only the blocks in order, but also to analyze each block.
So first we split the message : all until weather, and decode that first part.
Includes RVR decoding if needed, which is a long sequence to interpret correctly because there are quite a lot of possible suffixes (until now, max 2 different RVR are decoded per airfield, but soon I'll modify that and there will be no limit thanks to a loop... I've seen one single metar with 5 different RVR once !).
Then look for CAVOK.
Then extract special blocks (pressure, temp, NOSIG...) and analyze.
Then a weather analysis thanks to a weather codes database (Rain, Sand, Ice crystals, Fog...).
Then a progressive loop to analyze layers and find ceiling, except if CAVOK, SKC, CLR, NCD.
Then add suffixes (some BECMG and special features not usual in METARS)
So we haded some scripts to make it able to also interpret codes which are usually not allowed by the strict METAR rules, but that are used by some meteorologists.
Each time we find a bug (99% due to an errorwhen the METAR was edited, not respecting the legal form) we modify the code to make it able to handle such errors and give a correct interpretation, that's why the code becomes more and more dense as time goes by
, and interpretations are more and more reliable (those few errors usaually make the script to considere weather stronger than it is actually). But we still provide the original METAR, because there is no machine nor computer able to handle all possible human errors :-)
And we also have to pay attention to the fact that the code mustn't become too heavy, otherwise the page loading will be excessively long (especially for the whole European chart, there are nearly 500 airfields depicted).
Thanks BackPacker to have reported the error with the missing wind, you are completely right that is the reason for the wrong decoding of vis. I'll check that next week end to improve the code !
Alexandra
ORBIFLY IFR SCHOOL
PS : For Ukraine : next time we improve the chart I'll remember to extend it more east ! Right now there are only 2 fields available on the map : LRTR Timisoara and LRSM Satu Mare, with winds in meters per second :-)
I assume you would have liked Kiev, Minsk, Sebastopol and Chisinau ?
Last edited by Orbifly; 23rd Jan 2008 at 05:56.
Join Date: Feb 2007
Location: Amsterdam
Posts: 4,598
Likes: 0
Received 0 Likes
on
0 Posts
I understand that decoding a METAR is pretty hard, particularly if the met observers out there don't observe the standards. Providing the raw METAR is the obvious solution and to be honest, that's the one I read anyway.
But this error caught my eye because all of a sudden this field was reporting Intense IFR, while all fields around it were VFR. Odd.
As a computer technician who's dealt with this kind of stuff before (not METARs though): maybe you should add another series of scripts which do not decode the metar, but only extract the visibility and cloudbase to determine VFR/IFR conditions and use it for color coding. This would prevent using the first cloud layer as visibility in any case.
Still would not provide a solution for a METAR with the viz group missing though: does that mean all the 9's, or no viz at all?
============
As an aside, do you have a test bank of "difficult" METARs and their proper decode so that you can run regression tests once your scripts change? And obviously can add difficult METARs to once new errors are found?
But this error caught my eye because all of a sudden this field was reporting Intense IFR, while all fields around it were VFR. Odd.
As a computer technician who's dealt with this kind of stuff before (not METARs though): maybe you should add another series of scripts which do not decode the metar, but only extract the visibility and cloudbase to determine VFR/IFR conditions and use it for color coding. This would prevent using the first cloud layer as visibility in any case.
Still would not provide a solution for a METAR with the viz group missing though: does that mean all the 9's, or no viz at all?
============
As an aside, do you have a test bank of "difficult" METARs and their proper decode so that you can run regression tests once your scripts change? And obviously can add difficult METARs to once new errors are found?
Join Date: Jan 2008
Location: London
Age: 68
Posts: 1,269
Likes: 0
Received 0 Likes
on
0 Posts
Widgets to add
It looks great.
for fast reference those with a Mac (there is also a PC option) can add the following widget to their desktop where they can quickly see their 4 preferred TAF's METAR's
but i bookmarked the site too! Well done!
Bert
for fast reference those with a Mac (there is also a PC option) can add the following widget to their desktop where they can quickly see their 4 preferred TAF's METAR's
but i bookmarked the site too! Well done!
Bert
Join Date: May 2005
Location: Abroad
Posts: 1,172
Likes: 0
Received 0 Likes
on
0 Posts
Decoding METARs
Have you looked into phpweather?
That's what I use for my METAR/TAF decoding needs and works very well. It is pretty good at taking care of all the different units and regional flavours these things come in--it also retrieves the latest metars automatically from the NOAA website (caching to a MySQL DB by default). Really good piece of PHP code
That's what I use for my METAR/TAF decoding needs and works very well. It is pretty good at taking care of all the different units and regional flavours these things come in--it also retrieves the latest metars automatically from the NOAA website (caching to a MySQL DB by default). Really good piece of PHP code
Red On, Green On
Join Date: May 2004
Location: Between the woods and the water
Age: 24
Posts: 6,487
Likes: 0
Received 2 Likes
on
2 Posts
Or there's WeatherPixie
Thread Starter
Join Date: Jan 2008
Location: Europe
Posts: 34
Likes: 0
Received 0 Likes
on
0 Posts
Hello
Phpweather is a good decoder. As I was lazy I first checked this tool for Met'Map.
But 2 problems :
1) the code is too heavy when you need a great number of decoded Metars, the page loads too slowly (nearly 500 at once for the European page, plus the script that draws the coloured spot)
2) it does not always interpret properly "exotics" metars
Moreover, I wanted to color also the important part of the decoded Metar (low ceiling, strong wind, low vis....).
It appeared that programming everything by myself from scratch, and having it tested by my companion, was the best and most efficient option to get exactly what we wanted!
Alexandra
ORBIFLY
PS : Dear Vee-Tail : Wot about Pays de Galle ? Yes it is quite empty...
Just send us on [email protected] some ICAO codes of airports with weather services, and we'll add them on the chart !
(with a chart for position it is even better)
Phpweather is a good decoder. As I was lazy I first checked this tool for Met'Map.
But 2 problems :
1) the code is too heavy when you need a great number of decoded Metars, the page loads too slowly (nearly 500 at once for the European page, plus the script that draws the coloured spot)
2) it does not always interpret properly "exotics" metars
Moreover, I wanted to color also the important part of the decoded Metar (low ceiling, strong wind, low vis....).
It appeared that programming everything by myself from scratch, and having it tested by my companion, was the best and most efficient option to get exactly what we wanted!
Alexandra
ORBIFLY
PS : Dear Vee-Tail : Wot about Pays de Galle ? Yes it is quite empty...
Just send us on [email protected] some ICAO codes of airports with weather services, and we'll add them on the chart !
(with a chart for position it is even better)
Join Date: May 2005
Location: Abroad
Posts: 1,172
Likes: 0
Received 0 Likes
on
0 Posts
Hi,
Try caching the results. Consider using a cronjob for that too
I am interested on that. Do you have any examples handy? If you come across any, could you please post them?
Yep, that bit would need to be done by hand. So, to revise my original advice, allow me to suggest you may wish to use phpweather to borrow from or crosscheck your own solutions (which you are probably doing anyway, but thought I would mention it )
1) the code is too heavy when you need a great number of decoded Metars, the page loads too slowly (nearly 500 at once for the European page, plus the script that draws the coloured spot)
2) it does not always interpret properly "exotics" metars
Moreover, I wanted to color also the important part of the decoded Metar (low ceiling, strong wind, low vis....).
)
Thread Starter
Join Date: Jan 2008
Location: Europe
Posts: 34
Likes: 0
Received 0 Likes
on
0 Posts
Thanks LH2 for your advice, we really appreciate all help and ideas.
We are now training students for CPL-IFR-ATP checkrides for some weeks, so I'm not programming very often, and right now I'm developping an online US logbook for our students and instructors that we need as soon as possible.
But as I planned to develop the TAF part of Met'Map to make it more friendly, I'll be back on my weather code within some weeks, and I will make some tests on the PHPWeather code with some exotics Metars at the same time, and send you my results.
PHPWeather is a really great piece of code, user-friendly, easy to install, easy to understand and adapt, and I especially love their weather icons !
Alexandra
PS1 : I do not know how to "stick" this post, sorry !
PS2 : I added some airfields in Pays de Galle (Wales for French pilots :-) ) but some of them seem not to have local weather services, or not transmitted to the NOAA.
We are now training students for CPL-IFR-ATP checkrides for some weeks, so I'm not programming very often, and right now I'm developping an online US logbook for our students and instructors that we need as soon as possible.
But as I planned to develop the TAF part of Met'Map to make it more friendly, I'll be back on my weather code within some weeks, and I will make some tests on the PHPWeather code with some exotics Metars at the same time, and send you my results.
PHPWeather is a really great piece of code, user-friendly, easy to install, easy to understand and adapt, and I especially love their weather icons !
Alexandra
PS1 : I do not know how to "stick" this post, sorry !
PS2 : I added some airfields in Pays de Galle (Wales for French pilots :-) ) but some of them seem not to have local weather services, or not transmitted to the NOAA.