Accuracy

Hi folks,

I did some research and found an interesting example to show our accuracy. I found a high lightning strike density at the longwave radio transmitter "Sender Donebach" (363m). It is the second highest "building" in Germany, after the television tower in Berlin. Here is an overview of the area, with lightning strikes since September 2011:

You can see two hotspots with high lightning density in the center of the picture.

Zoom to the left one, and the two towers appear (you can see the shadows pointing to the north):

Most of our located strikes are only some hundred meters away from the towers! That's awesome.

But what`s about the right hotspot in the first picture? I didn't found any towers or high buildings on the satellite map. After some research I discovered, that three new wind engines were built there in 2010. They have a height of about 100m (+ diameter of rotor 92m). Some info here. Perfect

There is another high building in this area, the Reisenbach Telecommunication Tower. I put a marker on the next picture, which additionally shows all discovered lightning strikes from 2011 to now. You can see the two hotspots mentioned above, but nothing special around the marker:

I don't know, why there is no significant lightning density around the television tower. Maybe it's too low?

Another consideration: Notice, that most of the strikes that hit the radio transmitters and wind mills are displayed yellow in the last pic. That means, they are younger. They are from thunderstorms in November 2011 to January 2012. The other ones (red, orange) occurred some months before in autumn and summer. Thunderstorms in winter have a lower cloud base, so a lightning strike does not need so much power to reach the earth surface. With a high building on the ground, it is even more easier to produce a lightning strike. This should explain, why there yellow strikes are only around the radio transmitters and the wind mills - the energy in the cloud was to weak to reach the earth surface but high enough to reach very high buildings.

Feel free to publish this information to others, especially those who think Blitzortung.org is not accurate .

Best regards
Tobias

Hi Tobias,
This is a great idea!
Can you collect similar data related to our antenna towers?
Our highest tower is here, it is 314 m high:
Lakihegy tower:
47.37307,19.004838
But practically there are 3 towers, the two smaller are here: (around 100 m high)
small tower:
47.368594,19.006684
third tower:
47.38003,19.003696

And the "second", newer antenna tower is here:
Solt tower:
46.834244,19.03145
I would like to check the lightning events nearby.

t.janos

Hi Janos,

I've tested some of your coordinates, but I didn't find anything. As I've mentioned above, I had luck with the winter thunderstorms in my example.

Every participant has the possibility to search the lightning data on MyBlitzortung.org, like I've done it . First you have to login there (see this message). Then go to Archive -> Strike search in the top menu. Place the marker somewhere in the map, enter additional data below and press "Search". Sometimes it can take a long time, but it should go faster the following searches. Please do not search for too much strike (field "Count"). Maximum should be some thousands.

I've added another interesting statistic: A top-list of density counts, currently only for Germany and some surrounding area. See this page:
http://www.myblitzortung.org/myblitzortung/special/toplist*

If you are logged in, you can click on the positions and you will see something like this:

This example is the TOP 1 of our strike densities in a 2x2km grid since 2011 (radio antenna on a mountain - 2500m). See Säntis.

* Important hind: Unfortunately our TOA-calculations had a bug, so that a lot of strikes where calculated on almost the same positions (like a grid). Sometimes you will see only a couple of dots, because most of them are placed on top of each other. Egon fixed that some days ago.

Best regards
Tobias

Hai Toby,

I'm going to participate @ the blitzortung program because of the accuracy. Keep on going, because this is a great EU project with high possibilities.

Regards

Hi,
this morning in the news here in Belgium: a train station on fire (in the city Lier) after it probably got struck by lightning yesterday evening around 17h30.
I found it strange, because there hardly was any storm yesterday. So I decided to look it up on my blitzortung strike search.
It was very easy to find, there only was one strike So they must have been very unlucky.
I marked the train station with the pin. See attachement.

I find the deviation of 1.5km already very acceptable, to be honest. There are more blitzortung stations popping up in Belgium, so we hope this continues to increase accuracy in that part of europe

Hi,

very interesting! The locating was almost perfect: Participated stations in every direction, not too far and not too close. Most signals were good:

Our mean deviation is about ~4km. This lightning strike is a bit above this value.

Some more general notes on the deviation: We have to consider, that this value is the distance from the mean position of intersections of all used hyperbolic curves to the furthermost intersection. However, in most cases, our real deviation may be better. There's only one real option to estimate this real value: Watch thunderstorms, note positions and exact time of strikes. You can use this thread for observations. Professional providers of lightning data do that with video cameras. They also use the method in my first post, but this practise does not give 100% reliability. By the way: For higher accuracy, we would need better controllers on the evaluation-boards to record the beginning of the signal.

Here's another value, we have to consider: The yield. The only way to prove this value, is to compare our data with that of commercial providers. Here's an example for the area of Belgium:

We already detected about 50% (dotted line) of lightning strokes compared to a commercial provider. Unfortunately, with more strokes, that value got much lower in the past. Maybe that will be better this year, because Egon made some improvements in the TOA-calculation. We also have to consider, that our antenna alignment is not intended to detect intra-cloud or cloud-to-cloud strokes.

Also have a look on my coverage map, especially the text below that map.

Best regards
Tobias

Hi Robo,

here are the results:

Maybe the marked stroke is the right one. But we cannot be sure! For example, there could be other strokes that were not detected by our system. We should concentrate on cases, where we know position and exact time! The only real alternative is to search for high buildings/towers with high strike rate.

I found another example (Eifel Transmitter). It's only one tower, not two like my example in the first post. The yellow dots are strokes that occurred two days ago. There are no other yellow dots in the surrounding area. I think we can be sure, that those yellow dots, which are very close to the transmitter, did hit the tower. The marked strike is only ~200m away from the tower, but has a deviation of 4.3km.

As I've mentioned in my third post, every participant can do his own research on Myblitzortung.org !

Regards
Tobias

It's maybe not a good example of accuracy... but today 2 houses near to where I live got hit by lightning (I've witnessed the lightnings myself) and caught fire. Strangly only 1 strike is recorded and not at the location of the burning houses. I've witnessed 3 strikes in total, and I think only the first one is the one detected. So the 2 other strikes would be missed?

I have a similar occurence of a lightning strike on a house. A strike 34 mins before the reported time was recorded by blitzortung and was also not at the location of the burned out house.

I'm investigating a third event atm...

Hi all

I am fairly new here but wanted to add my findings for a recorded strike that struck a garage which caused a small fire and is within approximately 5 minutes of the actual recorded time and very close to the detected area approximately 1.4km with all other detected strikes well out of range and time.

Link to the article;

I hope that you find this useful and helpful too

Kind regards

Andy