I want a sensor that predicts nearby thunder so that I know when to close my window at night, so that I won’t be woken up by it.
It seems to me like the Thunderstorm Probability sensor of eg AccuWeather isn’t that accurate for that, as on a night in which there was one audible occurrence of thunder it only showed 60% probability. But perhaps there’s another sensor that can predict if there’s at least a single thunder occurrence, rather than an entire storm? And perhaps that’s more accurate then?
Living in Germany, btw.
Found this real time lightning detector component but is it useful? I mean it’s for lightning and not thunder. How should I set it up?
The lightning detector should get you what you want, because it detects the cause of thunder.
To detect thunder itself you’d need a microphone and something to determine that the sound heard actually was thunder - not a car or a dog or…
Lightning is WAY less complicated.
A human can hear a lightning strike (as thunder) for even up to almost 50km in some cases, so you’d figure out how to get the detector to detect lightning for approximately that distance tk start and then tune it from there.
You can get an Ecowitt WH57 lighting detector. It gives you range not direction, but you can infer the proximity and direction as you see the lighting strikes it closer to you. Like 24 miles, then 20 miles, then 15 miles, etc. You can use HA to create an automation f when the notification is with a certain range.
There also data available on the net to tell you how fast sound travels so you can anticipate when it will be a bother for you when the ligthing strikes at X distance.
Thunder is the direct result of air vibration caused by lightning therefore all lightning makes thunder. (unless you don’t have air. But that’s an entirely different problem)
How does that sensor help me then? If I understand correctly, it only gives me information that I can already get using my ears, doesn’t it?
It only detects current and past lightning, not that in the future. And if there’s lightning resulting in loud thunder in the present, I’d be able to hear it myself anyway.
It will detect a lightning strike. Lightning detectors are accurate for miles/kilometers (long way) and way farther than you can actually hear thunder (inverse square law governs thunders pressure so it is only audible for a specific distance - we can use this)
You get that info at light speed (immediate at earth distances)
Then you have exactly the amount of time for the pressure wave running at the speed of sound to reach you to warn you before you’re going to hear a thunderclap.(the speed of sound in air at sea level can be used to calculate the lag before thunder given distance)
Wait till a thunderstorm when your lightning detector is going off and then start tuning down teh distance for the alerts until they start coinciding with actual thunder
Yes you can predict thunder. This is just a technical version of the old counting trick you did as a kid.
Flash - one mississipi, two Mississippi boom (oh it’s about two seconds it’s a couple miles away.). In this case we don’t ‘see’ a flash but the lightning detector is absolutely accurate about the strike, intensity and location (it’s radio waves this is easy)
That said thunder gonna thunder - not like any of this avoids is it just tells you hey it’s about to thunder a few seconds to a minute before it happens. If you tune it down below the thresholds when you actually hear the thunder it makes a nice early warning system. (lightning is reported within x kilometers. Thunder may be present within the next x min) Which is also inherently better than mics because they cam ONLY tell you when it HAS happened. This way you get a (brief) warning because physics.
So essentially there’s a ~150s delay between a thunder strike occuring 50km away and me hearing it.
That would be good if I had motorized windows. Unfortunately, I don’t.
The problem is if I lie in bed already and sleep I couldn’t react to the sensor.
Seems like what I need is really a thunder/lightning forecast sensor.
True in a way, there’s always sound, but not entirely accurate at ground level. Last month a huge thunderstorm made the news here in the Netherlands. The flashes were seen by many (mee too), and while I was counting to estimate the distance I heard … nothing.
The explanation came in the news the next day: the lightning was not cloud to ground, but cloud to cloud. And on top of that, it was immensely high up. So many could see it, but only very few could actually hear it because the sound had dissipated. This seems to happen more often than we know, but we cannot always see it either.
In my installation, I have notifications sent to my mobile phone when that number drops below three different distances: 40 km (i.e., storms are in the area), 16 km (storms are getting close), and 8 km (thunder and lightning are likely).
This will have a fair number of false positives (storm cells that approach but don’t reach you), and it’s always possible an isolated nearby strike will happen without any others in the area. But overall it works well.
I suppose it is a forecast in a sense but I’m looking for one that goes beyond a few minutes into the future. I want something that predicts the next 9h.
That’s called a weather report. Chance of thunderstorms within 50km tomorrow.
Seriously that’s not practical to do with any level of accuracy. Weather report + the lightning integration already mentioned is probably the best available to anyone with the state of the art right now.
Is any of you guys using a lightning detection integration + motorized window? Has that ever reduced the sound of thunder in time for you? I wanna know if it works in practice.
Another thing I’m worried about is that for the loudest thunder strikes, the ones within 1 km of you, this setup isn’t even gonna help due to the time the motor takes to close the window, even, if it’s only tilted. And those are the thunder strikes for which you need it the most though.
Lastly, I wonder if the motor of the window itself is gonna wake me up or not.
Close it at < 10 km and open it again at > 10 km. I provded the sensor needed for that above.
Again, this will have false positives (many) and false negatives (rarely, in my experience). You’re absolutely right that this won’t work to close the window on demand for individual nearby strikes.