A rather intense line of thunderstorms followed by roughly an hour or so of moderate to light rain moved through the Madison, WI area this morning, Sunday, July 14, 2024. Looking at Windy's ECMWF forecast strip, I see that during that time there was no forecasted rain through the period. In fact, as far back as the forecast strip goes, the sky is shown to be clear and no chance of rain.

It would be one thing if this was a somewhat isolated incident, but that would be far from reality. It's more often wrong than right. Looking back, the Meteoblue model seems more accurate, but that may just be a fluke. Overall, I have found myself increasingly questioning the accuracy of this app on almost all forecast models.

So, what gives here? Has there been a global breakdown in all the forecasting models? I get that weather forecasting is an imprecise science, but I think near term forecasts should be highly accurate, or at least they used to be. What happened?

So far so good. Reinstalled the app and battery drain is only 0.3% for over half the day now, and was about that low for the remainder of yesterday after reinstallation.

Thank you!

@idefix37 said in Android Widget Storm Track Arrows Are Backwards:

@djMot
How do you get these storm track arrows? On the Weather radar layer of the Android app?

Windy Radar Widget on Android. They just appeared one day not long ago. Checking Settings, I see there is a "Vectors" toggle which turns them on/off.

I think they are way too large, for one thing, but the fact that they point the opposite direction to the actual movement of the precipitation is the more relevant bug in this new feature.

@idefix37 - Sorry if I misread your comment, but when noting that there were thunderstorms in the Chicago area, I thought you were confirming what I was saying. If you backed up the radar, those were the ones that blasted through the Madison, WI area but had no forcast data indicating they were coming through Madison. I believed @Gkikas-LGPZ's to confirm that only afternoon thunderstorms were predicted, implying that there was, in fact, no forcasted event for the morning during which severe thunderstorms and about an hour-long rain event occured.

Point is, most models completely missed the large weather system that came through, and those that did pretty grossly underestimated its scale in terms of deliverable precipitation.

New version that just dropped. 45.0.x? Today is 2025-04-02 so the one that just dropped.

On my Samsung Galaxy S25 Ultra, the Windy app is not respecting the OneUI Nav bar at the bottom of the screen, so the app is appearing partially below the Home, Back, and Recent Apps nav bar.

@idefix37 - No. You've shown the place to change the weather model. I asked where the setting is to make a different model the DEFAULT model so I don't have to change it every time.

So, although you've all confirmed this anomaly, no one has any thoughts on the breakdown of all but maybe one forecasting model? It makes no sense to me to have to look back and see which model predicted most accurately, or is currently correct, since the "winning" model might change every day. I don't want to play guessing games with weather forecasting. Over at least the coming 24 hours, and hopefully over the coming 48 to 72 hours, I believe the models should be pretty accurate. Further, when the situation on the ground differs from the model data, I would expect the model to be self-correcting, but that, too, seems far from the truth.

Again, I wish I could say this was just a one-off curiosity, but it's not. The models are more wrong than right these days, and it makes weather awareness more of a crap-shoot than a reliable expectation. The ECMWF model is reputed to be one of the most globally accurate and seems to be the Windy default. Not only does it get precipitation forecasting wrong, I frequently see temperature variances of up to 10°F in its forecasts.

I understand this isn't a failing of Windy. Windy is just depicting the weather as predicted by the various models. My overarching question is WHY are the models so far off these days? And by "these days" I actually mean over the past many months. I've had this sense that the weather information I've been getting from Windy is more wrong than right for at least the past six months.

@Gkikas-LGPZ - Sadly then, this is the equivalent of the automotive start/stop "feature" found on cars today. It's on by default, there is no way to configure the car to just permanently shut it off, so you have to do it each and every time. Nothing but annoying. But why? So here, too, apparently, you have to change the forecast model each and every time you go into the app or visit the site in a desktop browser. But why?

As an application developer myself, to me this seems such a drop-dead simple matter. In the most basic of terms, simply add a field to whatever settings data structure exists, then when the user chooses a different model, update that setting field. Add the logic to read and set that model when the app/website loads up. Done.

So, when can we get this done?