Primarily for general aviation discussion, but other aviation topics are also welcome.
#1724213
PaulB wrote:There seems to be much sharing of weather data between providers as you can get worldwide weather on virtually any app.

On a related note, many people rave over “Windy” and yes its visualisations are fab, but what assurance do we have that they are accurate and can we rely on it?


There are several forecast models which sites like Windy (or RASP, TopMeteo etc) rely on. On the "Windy" PC screen you can see a tag in the bottom corner marked "Forecast Models" which allows you to switch between ECMWF, ICON, NEMS and AROME and get different views. There is also GFS, and some Met Office provided ones. The public web sites tend to use the free ones, forecasters like the Met Office have their own.

Computing these models is a massive task - so the guys with the biggest computers have the best accuracy, but in practice to get a really accurate view you need to add the local knowledge and experience of a human forecaster. Any computer model is just a predication, it's accuracy for rain forecasting depends on many factors (eg size of the grid used for modelling, terrain model, historical data etc) so is just a best efforts prediction.

Google "weather forecast models" if you are interested!
Last edited by ls8pilot on Fri Oct 11, 2019 4:49 pm, edited 1 time in total.
By Lefty
FLYER Club Member  FLYER Club Member
#1724346
Weatheronline.co.uk/ expert charts
This allows you to see all the raw data that the human forecasters use to derive the forecasts they deliver to the public. You can select data from each of the different weather forecasting computer models.
defcribed liked this
#1724531
+1 for the expert charts on weatheronline.co.uk

Also, reliability/accuracy is a case of which weather forecasting model, not which app or site. Many apps and sites allow you to switch models and thus see different forecasts. Over at a certain other forum, where I believe the level of meteorological knowledge is generally higher than here, current consensus seems to be that the ICON forecast model is the best one.
#1724551
But they all work from the same source data presumably? Which is gathered by national services and then sold on to the people who want to run their models?

Or are there independent, perhaps competing, collectors of primary measurements?