Saw an unexpected spike this night

nrk94 wrote:

are you sure with your time estimates?

 

I have sudo server/database access - essentially, everything. I looked into this a bit further and went digging around in logs. It seems the power didn't go off between those times, but nothing hit the server between those times, and another time frame. I suspect either a part of Dan's network got turned off (circuit tripped) or the internet connection went down as there were no hits on the server at all. The outage times are as follows (all UTC):
2025-10-03 00:18:45 Last submission before outage.
2025-10-03 07:10:48 First submission after outage.
(So down between those two times)
Then again:
2025-10-03 07:18:58 Last submission before outage.
2025-10-03 16:11:40 First submission after outage.
Again down between the times. So it was two outages and was back up for around 9 minutes before going offline again. When I looked briefly earlier I was looking in the database and I stopped looking after I found the most recent outage and thought that was all. So it can be taken as two outages with a small recovery in the middle, or a larger outage between 00:18:45 and 16:11:40.

nrk94 wrote:

The data logs on radmon show a last measurement around....
 

Which logs are you looking at? The submitted time isn't necessarily that of the server. Users can submit from their devices using their local time if they are programmed to. Sadly, they tend not to mention that in the station description. I was looking at data from the failed submission logs (people submitting with the wrong password/usernames) and those times are logged by the server in UTC.

It is possible the other user in Denmark experienced what Mercator experienced, but in reality the spikes are tiny. I suspect something like a solar flare, or some space weather caused it. There were 2 or 3 M1.2 class flares on the 2nd and 3rd, although they weren't that large. So, I don't know what caused the spikes. I have had similar myself in the past, as have other users. It could be real radiation in the air, released form somewhere, but that kind of thing is really few and far between.

As for the Windy radiation data - the reddit post - I do think it was an exercise and the data was released by accident. They went public about it way too quick for it to be a real release. If it were real, we wouldn't hear about it for at least a couple of days I suspect, unless it was something catastrophic. Also, if you look at the Windy.com picture in the OP, you will see that the wind is blowing north, yet the radiation is travelling south. That kinda breaks the laws of physics. 

Good digging Simomax, yes radmon runs unless there is a power or internet outage. I have very recently installed a solar power system which is intended to bring radmon,org off power grid. A local power outage here does not affect the internet, so this should give maximum uptime.

However recently, there have been some experiments with the building containing radmon.org and running off the off-grid system. There has been at least one low-battery outage and a failure to start the database after a power loss outage.

Unfortunately, as I have been quiet for some time I haven't logged this directly in news. The outage is likely to be one of these events from about 2-3 weeks ago.

I'll be more active soon, as the list of important life and work tasks that have been keeping me quiet are being ticked off.

It is a shame that the outage may have nearly missed a detection event and I'll try to keep disruption to a minimum as I prove the power system. The end goal is of course to make radmon.org as reliable as possible

Mercator wrote:

Ok Spaceweather never crossed my mind. However, wouldn't  you agree that spaceweather events should be visible by all stations at the same time?
 

No, I wouldn't agree at all. Space weather, solar flares, CME's, planets going super-nova and so on are mostly blocked by the magnetosphere. That that gets through is then subject to the earth's atmosphere along with all the moisture up in the air, clouds etc. This blocks some from reaching the earth too. During this process a thin beam of particles may get through, enough to hit a small area. The magnetosphere can also change the path of the emission, thus instead of it hitting the earth dead on, it may get deflected off to one side, and simply skim the edge of the earth, giving it a much lesser detection area. I believe even the wind itself (think jet stream travelling at 110 MPH) can cause particles, electrons etc to physically be blown of course. I'm not so sure that applies to muons though too. Another thing is when the particles are in space, they are in a vacuum, so they have no resistance, when they hit the magnetosphere, and then the atmosphere, they slow down. Only some will get through, at localized places where either the magnetosphere is thinner, the atmosphere is thinner, or simply the weather doesn't slow them down as much, so they eventually hit the earth, but not one whole half of the hemisphere. 

Weather patterns on earth are localized, so in the event of some emission hitting the earth over Europe, if Germany has bad weather, but Italy is glorious sunshine without a cloud in sight, then Italy is likely to receive more of the emission as there is less to stop it. Another thing to note is that these emissions are often more of a beam of particles as opposed to a shower of particles, so when they come in and hit the earth, they appear localized.

We have had a few bad space weather/solar flare events that have actually damaged electricity transformers - unbalanced magnetic flux hitting the iron cores of the transformers causes eddy currents to go mental, and can cause a phase or two to heat up unproportionally to the other phases - makes the whole thing out of whack, and they heat up, fast, in an irregular way causing the transformer to break down, fast, and they blow up. If the space weather/solar flares covered half of the hemisphere, then it wouldn't be one or two transformers that blew up, it would be entire continents of transformers that blew up. But it isn't like that, as the magnetic flux is generally localized, so one, maybe two transformers get hit. If it were the case that half the hemisphere was hit by emissions, then every Geiger counter (on radmon.org) would show exactly the same, every time. It would also be historic, published many moons ago that space weather etc covered half the hemisphere when they hit, if that was the case, but it isn't so literature doesn't reflect this.

In a nutshell, space weather isn't uniform, orderly or predictable (in movement - where it hits), just like our weather here on earth. Different places get more or less than others. Oh, altitude has a big affect too. The higher up you are the more susceptible you would be of receiving it, the lower down, the more atmosphere above, slows or stops it, and as the atmosphere isn't even, more gets through in some areas than others.

As for AI I think all it would ever say is 'Yep, you got a spike.' It would have no idea what caused the spike other than you got one. Looking at your graphs, it looks like you already have a bunch of data, so you could try it for yourself, but I don't think it would be any help, possibly more of a hindrance in some cases, getting you to follow a red herring or something. As for adding to to radmon.org, I don't think it would server any useful purpose. Everyone has access to AI, whether paid or free, and adding a feature on here would simply just be replicating what is already out there, and through an API which someone would have to pay for - AI API's aren't free, at least OpenAI's API isn't free. I suppose one thing it might be able to do quickly is cross reference a dozen stations to see if they all picked up something at the same time, but that would consume thousands of AI tokens (the way in which the LLM works) and the more tokens consumed, the higher the cost for the API process. Anyone can download another station's data so if someone wanted to compare say, 5 stations close by each other, it's really not a big job to download the CSV's from each station page and upload them to an AI of your choice for analysis.

The only real benefit I have seen using AI with radiation is for isotope analysis, but then you need a gamma spectrometer. Aside that I haven't really used AI for anything more when it comes to processing radiation data.

Something that just came to mind regarding your spike, what was the weather like during that small spike period? Was it raining? It is possible something came down with rain as I have seen a few times before, and that can cause quite localized small spikes. Any weather underground (wunderground) weather stations close by you? That may give you a localized historic record of the weather back then.