Hacker News new | past | comments | ask | show | jobs | submit login
Satellite data show electromagnetic anomalies 19 days before Turkey earthquake (phys.org)
146 points by anigbrowl 6 months ago | hide | past | favorite | 33 comments



This is junk. They claim "anomalies" and fail to state the statistical odds of this anomaly occurring globally and the odds for this particular location. They also examine precisely one large earthquake rather than even considering looking at earthquakes in general.

"My cat walked across the keyboard right before the large earthquake" is not statistically significant.


This magnetic anomalies have been noticed before an earthquakes on many occasions.

For example this study found 2 out of 3 major earthquakes studied had similar precursor events. https://www.sciencedirect.com/science/article/abs/pii/S14747...

It’s not consistent, but it’s more than a one off correlation.


As Taleb would say, “Fooled by randomness”


Yeah, it should lead to predictions being possible fairly quickly if there was something to this.


Don't underestimate how dogma of social consensus can stifle otherwise innovative techniques, regardless whether they are proven effective or not.


There's no "dogma" or "social consensus" here. The authors of the paper simply failed to do their homework adequately.

Earthquake Prediction is filled with cranks. To deal with the issue, there is a neutral set of criteria for "how to test a prediction theory" which involves making predictions in advance with a clear description of what types of events match the predictions and likelihood of event. Geologists can then compare a series of predictions and their outcomes to what would be expected based on historic patterns. Math then allows one to determine if the prediction theory is better than random guessing.

No predictions have yet been better than random guessing.


I find the idea pretty wild that we should be able to predict earthquakes within days. My understanding is that earthquakes can be predicted like bridge collapses can be predicted. You can collect signs of a developing fault, but the uncertainty of when it rips is measured in years. And if your model is lacking for the specific instance, you'll misjudge the chances by orders of magnitude.

So I read papers like this one as hopeless fishing expeditions. But if the fishing expedition turns up with a good catch anyway, people will look into it more. Unfortunately I'm not versed enough to asses their catch. Will wait for a geologist friend for context.


I am a layperson, but it seems to me that scientists are discussing these topics for at least a decade (for example since L'Aquila in 2009): See section 2.2: https://arxiv.org/pdf/2303.03652

or

https://arxiv.org/abs/0905.2782

https://arxiv.org/pdf/2205.06592

https://arxiv.org/abs/0910.0797


Longer than that. This guy,

https://en.wikipedia.org/wiki/Didier_Sornette

made a name for himself a couple of decades ago with an approach to predicting earthquakes, financial crashes and other catastrophes:

https://www.pnas.org/doi/full/10.1073/pnas.022581999


Why do I think that he is responsible for writing his own Wikipedia entry?


Oh he absolutely did: https://en.wikipedia.org/w/index.php?title=Didier_Sornette&a...

Any edits he made that aren't his own page are editing in references to his own work in other articles.


Fun fact : 'sornette' means 'poppycock' in french.


Looks like a total whack with loads of credentials


We can predict volcanos erupting within days.

And this would not be "predicting", but more "noticing that the process already started"


Yes, from what I know for a volcano we have a good geological model of what is going on, and reliable ways of measuring the relevant parameters. Still predicting an eruption is an achievement, and predicting the not eruption especially so.

For earthquakes we are lacking both in models and in measurements. We don't know whether there is "a process started" in the days before a quake. This paper is looking for one.


SAR interferometry has been used to measure increased strain in the Earth's crust and identify where earthquakes are more likely. There a lot of published work which is credible. I didn't read the paper here but if results can be confirmed using other approaches could be a valid approach.


Earthquakes short-term prediction within days or weeks (not forecasting years prior to the event) is a taboo subject for many geologists mainly US based experts (I am looking at you USGS) [1].

We have performed data analytics based on publicly available seismic data that confirmed there are indeed reliable short-term precursor (minor earthquakes around 3 Mw) in all the latest major earthquakes (more than 7 Mw) within one month prior to the event. The seismic data that we analyzed latest major earthquakes from five countries including Turkey (the same latest major earthquakes this paper is reporting), New Zealand, Japan (2024), Indonesia and Philippines. However all prominent relevant journals for seismology/earth/geology/etc, and general journals for examples Nature (main journal and its sub journal Nature Communications) and Science (main journal and its sub journal Scientific Reports) were rejecting the reporting paper before it's even reviewed by the proper reviewers, and rejected by the editors. Heck, even arXiv.org rejected the paper and it's suppose to be pre-print archive not a proper journal.

Enough ranting, my point is that based on our analysis they are really precursors for major earthquakes, and we should be monitoring these properly. In fact in the first 100 days of this year 2024, earth have experienced devastating 3 major quakes of more than 7 Mw in Japan, China and Taiwan, the most frequent in world's recorded history within the short period. The so called seismic/geology/earthquake experts, however, seems to have their mind closed that they are treating earthquakes prediction like the investment or lottery prediction (snake oils).

[1] Can you predict earthquakes?

https://www.usgs.gov/faqs/can-you-predict-earthquakes


As a geophysicist friend used to say, in complete seriousness, the earthquake prediction problem is really easy to show expertise in.

Give location, time and magnitude. Period.

Have at it. If successful, your results will be noticed.


That's what we are going to do, if the paper is accepted in reputable journal using public offline data we can probably get grant and people working on the proof of concept for real-time implementation. Believe me to get any grant only with ideas without initial results is close to impossible, I tried a few times before. The fact that earthquake detection and prediction is a public service thing not really commercial, it not possible to get non-government or company interested in their development. I think the best is to get the algorithm and technique being utilized in nationwide IoT bssed earthquake monitoring system for example in China [1].

[1] Chinese Nationwide Earthquake Early Warning System and Its Performance in the 2022 Lushan M6.1 Earthquake:

https://www.mdpi.com/2072-4292/14/17/4269


Imagine what health care would look like if scientists and doctors shared that attitude.

This discussion is not about if the current prediction methods are good or not, it's about if we should even try to develop these methods to begin with. How many cancer researchers would there be today if you had to cure one yourself, with no grants or outside ressources before you are even allowed to enter the field? Probably not a lot.

Your argument is a self fullfilling profecy. You justify preventing them to research how to predict earthquake by saying that they have not already predicted one.

Maybe in a couple of decades we will find out that many of the big ones could have been very easily predicted, maybe not. The way things are going, we will never know.


You raised a very interesting comparison with medical doctors (MD) and cancer, I'd say their attitudes are not much different from the seismologist and geologist, but at least they are more open towards research publication using offline data because we managed to publish several publications. To be honest our algoritm and technique was developed originally for the early detection and prediction of myocardial infarction or heart attack, and the results are very encouraging with high accuracy of more than 99% accuracy, specificity and sensitivity, the main difference is that the data was heart vs seismic . During the beginning of the research, we tried to get permission to get government hospital patients data to check against the well-known biomarker Troponin for several years but we cannot even get the reply until today (not even a negative reply).

Dr. Randall W. Jones with FirstScan cancer product was among the pioneer of cancer screening device, an engineering doctor not an MD, and he even wrote a book on the struggles and endeavours [1]. According to him "Healthcare is in crisis" and “The system is broken".

Based on our initial results using the latest major earthquakes in five earthquake error prone countries based on publicly available offline seismic data, definitely we really can predict the earthquake within several days prior to the main major shocks hopefully preventing many lost of lives, unless we screw up the analysis bigtime. But again that's what peer reviews of papers are very important and critical, but we cannot even get past the all the reputable journal editors that we sent to, in order to get any proper review, c'est la vie.

[1] The Healthcare Disruptor: How An Underdog Inventor And His Companies Are Changing Medicine And Saving Lives:

https://books.forbes.com/books/the-healthcare-disruptor/


How do you distinguish between minor earthquakes which are followed by major earthquakes, and minor earthquakes which aren't followed by major ones?

The rule-of-thumb is that there are in the order of 10.000 3 Mw earthquakes for every single 7 Mw earthquake. How do you decide which 9.900+ ones to discard? How are you going to distinguish a critical precursor from a 2.3 Mw Taylor Swift concert[0]?

[0]: https://edition.cnn.com/2023/07/27/entertainment/taylor-swif...


How do you expect someone to come up with a good answer to that question with proper sources to back them up if this subject is blacklisted and no one can get grants or ressources to do the research? Maybe there is a pattern that can predict it, but we won't know if the people who have access to all of the data and knowledge are being told it will kill their career if they try to find the answers.


A huge swath of the scientific publishing community - including several journals for which geology is only a fraction of the fields they cover - engaging in a conspiracy to is a pretty extraodinary claim.

You have not produced extraordinary evidence - or even evidence at all.

You also don't provide the motive for such a conspiracy.

The kicker: you claim the USGS is leading the charge in shutting down all prediction research, yet the only page you cite is to USGS page which includes a link to a group that engages in prediction experiments: https://www.scec.org/research/csep


I'm just saying we cannot even get past the editors of proper journal and pre-print server arXiv for a proper review of the paper after many submissions, is it not a red flag?

Have you seen any USGS sponsored journal with special issue on earthquake prediction?

Have you have heard any earthquake early detection and prediction competition in the US, not unlike the yearly Computing in Cardiology (CINC) competition for the improvement of CVDs early detection and prediction?

[1]George B. Moody PhysioNet Challenge:

https://moody-challenge.physionet.org/2024/

https://www.cinc2024.org


Did you also check all those minor earthquakes that happened and there was no big earthquake after? or the claim is that if we wait enough there will be a big one ?

As other said just correctly preddict 10 and you will get your Nobel prize.


Between this and other potential electromagnetic phenomena like earthquake lights (which might or might not be power lines swinging and shorting out, creating sparks, or transformers blowing up), there's a lot we don't understand about earthquakes yet.

Piezoelectricity is one of the proposed mechanisms that might explain both earthquake lights and animal behavior changes minutes or hours before the actual shocks start.


There’s also bulk triboelectric effects where the mechanical forces, and the subsequent physical movement of large blocks of material can build up static charges due to having different electron affinity, like a rubber balloon and a wool sweater… by rubbing one against the other a static charge builds up… one theory for earthquake lights is ground to sky discharge of triboelectric charge buildup due to different electron affinity between the two sides of a fault plane under stress… a large enough area can make a small enough difference add up, so even if it isn’t earthquake lights, the triboelectric effect could be contributing to what was measured in this case… fault “surfaces” are going to put square kilometres of differing materials against each other… a non zero effect per square meter, can add up quite a bit when applied to several square kilometres.


These electromagnetic anomalies are likely caused by stresses in rocks deep underground similar to the phenomenon that occurs when piezoelectric materials are put under stress. As stresses build up before a quake, similarly EM anomalies ought to be detectable beforehand, thus detecting them early could give people enough forewarning for them to seek safety.

I don't know whether satellites are ideal or are sensitive enough (because of their distance from the ground) for this kind of earthquake prediction but over the years there have been significant numbers of terrestrially-bound observations of electrical/electromagnetic disturbances occurring before a quake happens to suggest the phenomenon ought to be fully investigated.

These EM anomalies have been the subject of a news report in an earlier HN story and I made a suggestion then that until the science was fully established that the Amateur Radio Service could be of possible assistance.

Amateur radio operators could perhaps establish a worldwide network to monitor and or detect changes in the EM field on, say, an ultra low frequency band (below the AM broadcast band) as it's likely the most suitable. I'd envisage Amateurs operating a network of CW TXes and signature changes in their carrier signals would act as the early warnings.

This could be done as a separate Amateur operation (national operations—RSGB, ARRL, WIA, etc. and or through the IARU), or in conjunction with various government agencies, IPS (Ionospheric Prediction Services), etc.

The advantage of using the Amateur Service or including it with government operations (which at the outset would be small) is that many more monitoring points could be established quickly. Moreover, given the science is not well established, it's unlikely governments would fund a large network early on. That's where Amateurs come in, by contrast, the cost of establishing an Amateur network would be minimal and would grow organically (as Amateurs are inherently flexible).


Rolling this technology out in practice is quite challenging.

How do you trigger "The Alarm" for a given area? How much of a guard band do you put around the event and what happens when you have false positives? Do you call everyone back to the region and say sorry it was a false alarm, only for it to trigger a week later! Would not want to be the one calling the shots on this.

It also seems like a good technology to throw into a movie plot - Super Mega Quake... starring Dwyane Johnson.


I'm posting now, there will be a fairly large earthquake on west coast in the next ten days.


Good one. But to be statistically meaningful you’d need to tell us how big, and where on the west coast. The west coast is huge.


Near Vancouver, Seattle, 8+.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: