The GFS model uses balloon ascents as input data which also usually happen twice a day.
I don't think GFS uses ascents any more than every other weather model uses baloon ascents.
Today, the twice-daily ascents are just a small part of the input. A LOT of data comes from satellite observations.
I think the OGIMET data is good. The cloud tops forecast was within 1000ft on the flight to La Rochelle for example, all the way there.
I think the OGIMET data is good. The cloud tops forecast was within 1000ft on the flight to La Rochelle for example, all the way there.
I start using OGIMET when I first heard about it here on this website a few months ago, and I find the forecasts very accurate.
Ogimet fetches the data as it becomes available.
I'm sure they do, but it'll take a lot of processing on the raw GFS data before it's usable. It would be nice if we could have someone from OGIMET explaining the process in more detail. From my observations it seems that 0000Z GFS data becomes visible around 0500Z, and 1200Z GFS data is there at 1700Z.
I'm sure they do, but it'll take a lot of processing on the raw GFS data before it's usable.
What "lot of processing" do you think is needed?
To me, it looks like everything plotted is contained directly in some GRIB2 layer in the GFS dataset.
What "lot of processing" do you think is needed?
Since the creation of the graph is quite fast, I'm under the impression that most data is preprocessed, and if you click "Submit", most processing time is spent on creating the image.
The prediction of Icing and Turbulence is not in the GFS data so that need to be modeled. I don't think that is done real-time, as it would impact performance.
They also correct the GFS data for altitude and land fraction. I guess that's also done during some kind of batch processing.
Ogimet is doing a very good job, and I'm sure it has cost a lot of time and dedication.
I didn't know this impressive site. Thanks for mentionning it. A 9 hours delay before the data becomes available is not that much. In France, TEMSI (significant wheather charts) and WINTEM charts are made for the whole day at the same time in the morning (usually at 00UTC). If you fly in the afternoon, the forecast is already 12 hours old or more.
The OGIMET website is great and based I think for a large part on the data from the GFS model.
OK. We don't have this nice chart (yet) that runs along your route to show graphically the expected weather enroute as OGIMET does, but we basically present the same weather in our AeroPlus Aviation Weather app using a combination of GFS, METAR and TAF data.
What I do know is that the data as it comes from NOAA (the raw GFS data) has to be processed before it can be presented. In any case, it takes quite a lot of storage.
The turbulence comes from the GFS model in the form of lifted index and CAPE values. On my blog I explain what they mean. They depict the expected convective energy in the clouds and can give you quite a good forecast of the expected type of clouds (CBs for example).
For forecasting the top-of-clouds (thanks to the great information that Peter H. provided on this on his site) all you need is to get the ambient and dewpoint temperatures from the GFS model for a specific location and time and plotted through the atmosphere from the ground to e.g. FL310. Where the 2 lines meet you have a 100% relative humidity and could expect clouds (ok, it is more complex that that). Add to this the other information available in the GFS model and you now get an idea of what TYPE of clouds you could expect there. Now show this visually and add METAR, TAF, SAT IR cloud-top images and significant weather charts to the same app and you have all the needed data at your fingertips.
See my blog item explaining this here. I can't think of anything important that I am missing out on, except for the rain-radar and live weather in the cockpit. On that we are working. Any tips on what is missing in the app would be more than welcome. I highly respect the expertise of the pilots of this forum.
GFS has contained the humidity data for at least 10 years.
I recall, maybe 10 years ago, conversing with somebody working for a commercial weather business, who mentioned this (i.e. the ability to plot vertical IMC extent) but he didn't want it widely known. In the end his business didn't do anything with it anyway - AFAIK.
A bit later, Meteoblue offered this kind of plot, and (as described in my writeups) I used them for a while. That site was spoilt by crappy Java programming; various features barely worked and some worked only if you pressed the buttons in the right sequence. I think the site has now died, or nearly so. I believe the site was the result of a PhD project of somebody in Switzerland or somewhere near there...
The OGIMET presentation is brilliantly done. I really like the blue/black day/night depiction for example.
Also they can plot a route which is not a straight line e.g. EGKALFLNLSZH.
I've used OGIMET for quite a few years now. I used to be based in Spain for a while and they are a 'local' site one knows about (it's done by the University of Granada). Of the European wx sites I know and use regularly, it is by far the best, IMHO.
I'm sure they do, but it'll take a lot of processing on the raw GFS data before it's usable.
I think you are correct, at least in the Spanish version of the site you get a 'still processing' message when trying to access the charts too soon after the GFS model run.
GFS data is delivered in numeric form so the website doing the presentation has to download it and plot it.
Some do some extra processing on it, but you can't extract data which isn't there to start with.
Once in 6 hours, the GFS data is changed, so we check for changes regularly and download the Grib data files as soon as they are changed. These are binary files consisting a 3d grid model of the entire world in Mercator scale. We then parse this data so we are ready to present results to the pilots requesting weather info. This all happens of course in the background on our servers and not in the client app.
