So I just started using the Windy app in my flight planning. "Just" meaning this morning. You guys that use it regularly how accurate do you find the various forecasts to be ? ie cloud bases & tops, precip, wind etc.
Last edited:
Windy app
- Thread starter Kitch
- Start date
AlleyCat67
Pre-takeoff checklist
- Joined
- Oct 5, 2017
- Messages
- 485
- Display Name
- AleyCat67
- Home Airfield
- 20N
Keep in mind that there isn't really a single "Windy forecast". The app reports out fields from multiple weather models, including GFS (US) and ECMWF (Europe) long-range forecasts, and the NAM (US) short-term, high-resolution forecasts. Definitely explore the buttons at the bottom where you can switch between models. In general, the ECMWF is considered the gold standard for longer range forecasts (say, 3-7 days) although the NAM is probably better for short-term forecasts, especially in areas with topography. As a rule of thumb, if I see multiple models giving the same general prediction, I tend to have more confidence.
Accuracy really depends on which fields you're looking at. In general the basic meteorological fields like pressure and temperature are easier to predict than derived fields like cloud cover, precip, and surface wind speed, which all depend on spatial gradients. For my area (East Coast of the US), the ECMWF does a decent job a few days in advance (maybe 3-4 days *most* of the time) of separating an easy VFR day from a cloudy/rainy IFR day. Summer is harder since the bad weather is convective in origin, and the models can only give a rough idea of the probability of convection based on CAPE index, wind shear, and other variables.
A couple of quirks I've found: First, the ECMWF overestimates surface wind gust factor. I believe the definition of wind gust in that model is several meters off the ground, so it's not really at the surface. I always use GFS or NAM for assessing gusts. Second, the cloud base forecast is only available from ECMWF for some reason, and it can be quite inaccurate at least for aviation purposes. There's a built-in tool to generate forecast sounding profiles (or Skew-T) and I prefer that to get an idea where the cloud bases might be. I'm sure other folks have tips and tricks for using Windy tools, and I'd be interested in learning them.
Accuracy really depends on which fields you're looking at. In general the basic meteorological fields like pressure and temperature are easier to predict than derived fields like cloud cover, precip, and surface wind speed, which all depend on spatial gradients. For my area (East Coast of the US), the ECMWF does a decent job a few days in advance (maybe 3-4 days *most* of the time) of separating an easy VFR day from a cloudy/rainy IFR day. Summer is harder since the bad weather is convective in origin, and the models can only give a rough idea of the probability of convection based on CAPE index, wind shear, and other variables.
A couple of quirks I've found: First, the ECMWF overestimates surface wind gust factor. I believe the definition of wind gust in that model is several meters off the ground, so it's not really at the surface. I always use GFS or NAM for assessing gusts. Second, the cloud base forecast is only available from ECMWF for some reason, and it can be quite inaccurate at least for aviation purposes. There's a built-in tool to generate forecast sounding profiles (or Skew-T) and I prefer that to get an idea where the cloud bases might be. I'm sure other folks have tips and tricks for using Windy tools, and I'd be interested in learning them.
Last edited:
Perfect. Thanks for the info! We are supposed to fly into ART tomorrow and back to PYM on Sunday. My wife doesn't love flying in the clouds so I'm just trying not to torcher her with 2 hours in the soup on the way home. LOL
Everything I'm seeing says if we bail out Sunday morning we should be mostly VFR
Hang 4
En-Route
- Joined
- Aug 18, 2017
- Messages
- 3,422
- Display Name
- Hang 4
Agree with @AlleyCat67 on how-to. ECMWF for long range stuff, except summer in the southeast. NAM is best for short term. I do find the cloudbase forecast to be reasonably accurate.
The cloudbase forecast is from ECMWF, so it works best for larger, more established weather systems.
The radar is high resolution and easy to zoom in and out.
I also find the ECMWF gust factor to be wildly overstated most of the time.
One nice feature is looking at a location forecast with all the models at once. It really shows how difficult the forecasting process is. You'll scan down a specific time interval and the precipitation, cloud cover will be all over the place, even just a few hours out.
I use it as my main tool for pre-flight leading up to a long XC. I'll start looking three to five days out for obvious no-go things like a strong cold front with convective activity. As it gets closer, I'll be looking for whether I can go VFR or need to flex the IFR muscles. I do go to the Aviation weather .gov site morning of a flight, but not sure I get much more from that than from Windy.
Lastly, paying the few extra $ for a premium subscription seems worth it to me.
The cloudbase forecast is from ECMWF, so it works best for larger, more established weather systems.
The radar is high resolution and easy to zoom in and out.
I also find the ECMWF gust factor to be wildly overstated most of the time.
One nice feature is looking at a location forecast with all the models at once. It really shows how difficult the forecasting process is. You'll scan down a specific time interval and the precipitation, cloud cover will be all over the place, even just a few hours out.
I use it as my main tool for pre-flight leading up to a long XC. I'll start looking three to five days out for obvious no-go things like a strong cold front with convective activity. As it gets closer, I'll be looking for whether I can go VFR or need to flex the IFR muscles. I do go to the Aviation weather .gov site morning of a flight, but not sure I get much more from that than from Windy.
Lastly, paying the few extra $ for a premium subscription seems worth it to me.
The forecast data using the NOAA data is the same as you’ll see on the FAA site
Personal I find the European model better for 3+ days, and the HRRR better for 0-3 days.
AA5Bman
Line Up and Wait
- Joined
- Jun 14, 2009
- Messages
- 967
- Display Name
- He who ironically no longer flies an AA5B
I have started using the Windy app for aviation planning, too.
On one hand it feels like cheating, like since it’s not an aviation-specific product it must not be real weather and I’m taking some sort of risk in using it (pretty sure this is ridiculous and it’s the same underlying models/sources as everywhere else).
On the other, it amazes me how much better and more intuitive it is than the weather forecasting products available on, for instance, ForeFlight. Once you get used to map-based, graphical moving displays it’s really hard to feel that good about ForeFlight’s static, 6-hour images.
If nothing else, for broad trends and understanding of how weather is likely to develop, it seems pretty great to me.
On one hand it feels like cheating, like since it’s not an aviation-specific product it must not be real weather and I’m taking some sort of risk in using it (pretty sure this is ridiculous and it’s the same underlying models/sources as everywhere else).
On the other, it amazes me how much better and more intuitive it is than the weather forecasting products available on, for instance, ForeFlight. Once you get used to map-based, graphical moving displays it’s really hard to feel that good about ForeFlight’s static, 6-hour images.
If nothing else, for broad trends and understanding of how weather is likely to develop, it seems pretty great to me.
Ok so it's not just me that thinks all of the above LOL
- Joined
- May 25, 2006
- Messages
- 21,997
- Location
- KTTA, North Carolina
- Display Name
- Who you kiddin?
- Home Airfield
- KTTA
Maybe it is cheating. Do you think looking out the window is also cheating since your eyes are not an aviation-specific product either?
Nick Geber
Pre-takeoff checklist
- Joined
- Oct 10, 2022
- Messages
- 220
- Display Name
- NIckG
- Home Airfield
- KHND
Any thoughts of MeteoblueAI on windy?
Hang 4
En-Route
- Joined
- Aug 18, 2017
- Messages
- 3,422
- Display Name
- Hang 4
I've not been impressed with what I've seen so far, but I haven't spent a ton of time on it either. It takes input from multiple models to create a point forecast.
The description of how it works is more technical than I have the aptitude to evaluate, but if I look at what it spits out vs what I expect will happen, I don't see a good fit, but what I haven't done is look at the forecast, then see if the actual matches or not.
Still too new to really comment. It’s suggested low ceiling tomorrow but none of the other models really suggest that. It will be interesting to see what actually happens