Theory of Human Response to the Effects of Tectonic Stress

[Alan Watson]

I would hope that this may be an appropriate Forum for consideration of a Cross-Disciplinary Theory.

Having published a summary of my research into the Theory of a Human Response to Tectonic Stress there was very little response from the Geoscientific Community. This may be because work has only recently developed on the interactions between the lithosphere and the biosphere. This is not a 'pure' Geological theory and neither does it fit comfortably in the Humanities.

Can I urge this Forum to consider the possibility that the effects of tectonic stress may influence human behaviour. A summary of my research can be found on the Geological Society of London's website at the following link :

 

http://www.geolsoc.org.uk/Geoscientist/Archive/November-2013/Gravity-and-mind

 

Briefly, having undertaken a comparative study of the timing of riots and civil disturbance relative to the timing of earthquakes in England and Wales between 1980 and 2012, my research has found a compelling statistical significance in a higher than expected number of riots in the 14 days prior to earthquakes of 2.5ML or greater. The probability of the tested hypotheses being wrong have been shown to be less than 1%.

 

This is not just a new theory; it may even be a New Science!

 

Alan Watson

UK Ground Investigation Specialist and Civil Engineer

 

[studiot]

Alan, thank you for sharing this.

 

You seem to have negotiated the proper side of the line in providing enough information in your opening post to justify a link to a much larger paper IMHO.

 

I don't know about the correctness of the hypothesis, I have heard studies of correlations between weather conditions and human (and other life) behaviour and also studies correlating weather conditions with quake activity.

So perhaps there are more links to be drawn?

 

An authority who might well be interested is Brian Fagan

 

https://www.google.co.uk/search?q=brian+fagan&hl=en-GB&gbv=2&oq=brian+fagan&gs_l=heirloom-serp.3..0i67j0i7i30l8j0.11781.11781.0.12015.1.1.0.0.0.0.141.141.0j1.1.0....0...1ac.1.34.heirloom-serp..0.1.141.1CHWHcSRVUM

[Alan Watson]

Thank you for your interest Studiot

 

As you mentioned the weather, one aspect of my research was to look at the distribution of riots as they occurred in months of the year.

 

There was a clear correlation showing a higher incidence of riots in the summer months. A Plot published in my work is attached which is % riots by month plotted against average hours of sunshine by month for the majority of the study period in England & Wales.

 

Very many disparate influences contribute to the onset of rioting, including sociological, economic, racial, religious, sporting, confinement & weather.

 

This shows that when there are so many influences, the possibility of further contributory factors not previously considered may indeed play a part in the onset of violence, including, in the case of my study, the effects of tectonic stress.

 

 

Chart 3-1 for Science Forum.pdf

[studiot]

 

There was a clear correlation showing a higher incidence of riots in the summer months

 

Whilst I'm not suprised to hear this, I was more interested in its corollary.

 

There should therefore be a lower than average incidence during periods of inclement weather.

 

You often find freak (inclement) weather associated with/following quake activity.

 

This suggests a pattern (to be investigated)

 

quake [math] \to [/math] freak weather [math] \to [/math] increased/decreased riot activity

 

or perhaps all three can stem from a common cause.

 

[Alan Watson]

Indeed, the underlying common cause is of most interest as the main aim of the research is to identify recognizable behaviour patterns which may be indicative of an increased risk of earthquakes before they occur.

 

Your quake followed by freak weather followed by increased / decreased riot activity sequence does not lend itself to the advance warning of an increased risk of quake.

 

The hypotheses I tested were :

 

‘There is a significantly higher incidence of riots and disorder in the 14 days immediately before earthquakes of 2.5M_L or greater [in England and Wales 1980 to 2012] than would be expected by chance.’

 

and conversely

 

‘There is a significantly lower incidence of riots and disorder after more than 140 days has passed since the last most recent earthquake and when more than 14 days remain before the following earthquake of 2.5M_L or greater [in England and Wales 1980 to 2012], than would be expected by chance.’

 

The probabilities of the hypotheses being wrong have been shown to be substantially less than 1% in both cases.

[Klaynos]

How many riots and earthquakes?

 

Correlation doesn't necessarily means causation. Do you have a proposed mechanism? Is the mechanism measurable?

 

The correlation is certainly interesting.

[Alan Watson]

Indeed correlation does not necessarily mean causation. Earthquakes do not cause riots.

 

Rather, it is the same underlying influence ( i.e. the effects of Tectonic Stress ) which is a contributory factor in both the onset of violence and quakes.

 

In England and Wales between 1980 and 2012 I have found 64 riots.

 

In England and Wales between 1980 and 2012 the British Geological Survey report 151 earthquakes of 2.5M_L or greater.

 

After allowing for some earthquakes occurring within 14 days of each other, the number of riots which would be expected in the 14 day periods before the earthquakes is :

 

9

 

After allowing for some earthquakes occurring within 14 days of each other, the number of riots which would be expected in the 14 day periods after the earthquakes is :

 

9

 

In reality there were 16 riots in the 14 days before and only 5 riots in the 14 days after the 151 earthquakes in England and Wales between 1980 and 2012.

 

i.e. there are more than three times as many before compared with after the earthquakes even though the same number would be expected by chance.

 

Not only does this show that the effects of Tectonic Stress increase the likelihood of riots before earthquakes, but equally, there is a substantially lesser risk of riots following shorlty after earthquakes in England and Wales.

[studiot]

Another methodology query, Alan.

 

I can be reasonably sure that the BGS report of quakes is accurate.

 

For 'riots' I am less sure about the reporting accuracy.

 

Who reported them to whom

Who assessed what the definition of a riot is and whether each report met this definition.

How confident can you be that before quake when there may have been little to fill the news space riots were not over reported and after a newsworthy event they were not under-reported.T

 

These factors may not account for the near 3:1 ratio you have presented, but have you considered them?

 

Can I commend to you

 

Standard Deviations by Gary Smith?

 

In particular the discovery of the causes of cholera by methods such as you are employing

[swansont]

Indeed correlation does not necessarily mean causation. Earthquakes do not cause riots.

 

Rather, it is the same underlying influence ( i.e. the effects of Tectonic Stress ) which is a contributory factor in both the onset of violence and quakes.

 

In England and Wales between 1980 and 2012 I have found 64 riots.

 

In England and Wales between 1980 and 2012 the British Geological Survey report 151 earthquakes of 2.5M_L or greater.

 

After allowing for some earthquakes occurring within 14 days of each other, the number of riots which would be expected in the 14 day periods before the earthquakes is :

 

9

 

After allowing for some earthquakes occurring within 14 days of each other, the number of riots which would be expected in the 14 day periods after the earthquakes is :

 

9

 

In reality there were 16 riots in the 14 days before and only 5 riots in the 14 days after the 151 earthquakes in England and Wales between 1980 and 2012.

 

 

 

If things were randomly distributed. But the aftermath of an earthquake might occupy peoples' attention, making them less likely to engage in acts of civil unrest.

[Alan Watson]

Studiot,

 

The good thing about studying riots in England and Wales between 1980 and 2012 is that you can be sure that the vast majority of 'significant' riots have been reported.

 

The media have been highly interested in riots since the 1980s due to the socio-economic implications / causes of riots.

 

Think about it. If a significant riot occurred today in England and Wales do you think it would go unnoticed?

 

You can be sure that my riot listing has been thoroughly researched as my e-book's Chapter 3 was entitled 'Riots in England and Wales 1980 to 2012' and in itself is a comprehensive study of public and prison disorder in that period.

 

In order to develop my theory I have started a website ( www.alwriting.co.uk ) which lists all the riots between 1980 and 2012 and it affords the opportunity for contributions from interested readers where they can notify the website link if they believe a particular riot they are aware of is missing from the list.

 

Despite a 6-page summary of my theory appearing on the pages of Geoscientist Magazine in 2013 the theory has never been effectively opposed.

 

Thank you for the reference.

[Klaynos]

Have you considered a control where you select random days (the same number of days as you have earthquake events) and apply the same methodology?

 

It's quite a small data set you're working with, which is to be expected, larger than I feared it might be though.

 

Might be interesting to look at France over a similar period? Similar seismic stability. I think picking somewhere with lots of quakes might make it too difficult.

[Alan Watson]

Hi swansont

 

I can see you are based in Washington DC region. Your perception of earthquakes seems to be that all earthquakes have an 'aftermath'.

 

Please note that my statistical study has focussed on England and Wales between 1980 and 2012 where the greatest magnitude earthquake was 5.8M_L.

 

The Tectonic Conditions in England and Wales are characterised by low magnitude quakes. There is generally no aftermath to England and Wales quakes.

 

By selecting a relatively benign tectonic region I have placed my study in an environment away from many contaminating effects present in active regions.

[studiot]

Hello Alan,

 

I meant to add in tha last post.

 

I hope you take my comments as genuine testing of the methodology, not attempts to discredit the study.

The subject is genuinely intriguing.

[Alan Watson]

Hi Klaynos

 

I understand your question. You should find that the attached Chart goes some way towards providing the clarification you seek.

 

The chart plots the ratio of number of riots before, to number of riots after quakes throughout the entire riot-quake cycle of 32 years.

 

The reference to 'Tails' is Tails of Clusters - I undertook a dual study, firstly including all the riots, and then removing the Tails of Clusters so that the unwanted influence of 'copy cat' rioting can be allowed for. It made no substantial difference; the theory can be proven statistically in both cases.

 

Chart 4-3 for Science Forum.pdf

[swansont]

Hi swansont

 

I can see you are based in Washington DC region. Your perception of earthquakes seems to be that all earthquakes have an 'aftermath'.

 

Please note that my statistical study has focussed on England and Wales between 1980 and 2012 where the greatest magnitude earthquake was 5.8M_L.

 

The Tectonic Conditions in England and Wales are characterised by low magnitude quakes. There is generally no aftermath to England and Wales quakes.

 

By selecting a relatively benign tectonic region I have placed my study in an environment away from many contaminating effects present in active regions.

 

 

As a resident of DC I can observe that benign quakes have an outsized effect, since buildings in quiet areas are generally not built with quakes in mind. We had a 5.8 quake here a few years ago. Lots of damage. We got laughed at by the southern California crowd, because for them that quake is not unusual. (Turnabout being fair play, they freaked out when the temperature dropped to freezing a couple of years later)

 

 

How close are the earthquakes to the riots?

 

Klaynos mentioned small data sets, which looks to be a big issue. Are your results statistically significant? If so, at what level?

[Alan Watson]

And Klaynos, regarding France, I would need to obtain a verified and complete list of earthquakes ( and riots for that matter ). The BGS state that they believe their list of England and Wales earthquakes is complete for quakes >2.5M_L since 1979. Does the same apply to a record of France quakes? If I could be provided with a complete France list then that would potentially be a next step for developing the theory.

[studiot]

 

swansont

How close are the earthquakes to the riots?

 

Excellent question +1

[Klaynos]

Hello Alan,

 

I meant to add in tha last post.

 

I hope you take my comments as genuine testing of the methodology, not attempts to discredit the study.

The subject is genuinely intriguing.

I'd second these sentiments.

[Alan Watson]

Swansont

 

Regarding statistical significance I have based this on a Chi-Squared Test looking at the expected and actual incidence of riots in three periods, namely up to 14 days, 14 days to 140 days, and more than 140 days before and after quakes. The probability of the hypothesis (a relation between the occurrence of riots and the occurrence of earthquakes exists) being wrong is substantially less than 1%.

 

The geographical distribution of the riots relative to the earthquakes requires some explanation.

 

The Role of the New Geophysics

Conventional ‘Sub-critical’ Geophysics has never been able to satisfactorily explain how tectonic stress can accumulate over large lateral distances before high magnitude earthquakes without breaking out into multiple low magnitude shocks. The ‘New Geophysics’, described by Stuart Crampin et al, has suggested a mechanism where stress-energy prior to earthquakes may be spread out thinly over large volumes of rock without initiating low magnitude earthquakes.

 

The idea that tectonic stress or other non-seismic pre-earthquake effects can influence the onset of rioting and disorder, may be supported by these relatively recent developments which can allow comparisons between lists of riots and earthquakes over large geographical areas subject to widespread increases in stress.

 

The statistical significance I have found, is consistent with the proposals of the New Geophysics.

Studiot, Klaynos - That's OK, and no problem - please drill down as much as you like !

 

This is a theory which needs to be bottomed out.

 

I attempted to provoke a discussion on the letters pages of Geoscientist Magazine, but the Earth Science community don't seem to be taking up the challenge !

 

The Editor, Ted Neild suggested it may be due to the 'Cross-Dressy' nature of the study - i.e. Multi-Disciplinary.

 

The British Psychological Society are aware.

 

Please note I am a Civil / Geotechnical Engineer / Ground Investigation Specialist so something of an 'Outsider' in all respects.

[John Cuthber]

 

 

In England and Wales between 1980 and 2012 I have found 64 riots.

 

 

A riot in the UK twice a year seems rather unlikely.

Wiki only seems to have found rather fewer

 

https://en.wikipedia.org/wiki/List_of_riots#1980s

 

What mechanism are you using to find these "riots"?

Edited May 9, 2016 by John Cuthber

[Alan Watson]

John, I understand your skepticism, but I can assure you that I found 64 significant riots in England and Wales between 1980 and 2012.

 

The full list can be found on my website at : http://www.alwriting.co.uk/id4.html

 

I would welcome any comments from you if you doubt the accuracy of any entries on the list.

 

Some years have no riots, particularly during seismically quiet periods.

 

The list will be periodically updated based on comments received. To date the list is as accurate as possible and the theory remains valid and effectively unopposed.

[swansont]

Swansont

 

Regarding statistical significance I have based this on a Chi-Squared Test looking at the expected and actual incidence of riots in three periods, namely up to 14 days, 14 days to 140 days, and more than 140 days before and after quakes. The probability of the hypothesis (a relation between the occurrence of riots and the occurrence of earthquakes exists) being wrong is substantially less than 1%.

 

 

 

I don't believe that to be true. The sampling error is the square root of the sample size if it's white noise. Seeing as your data points are around 2 deviations from the average (9±3, so 2 deviations is 15 and 3), so being substantially less than 1% doesn't seem like a reasonable conclusion. And that excludes any bias which would give you a systematic shift.

 

Are there any other 14-day windows with such large and small counts that aren't associated with an earthquake?

———

 

How do you get 9 earthquake per two-week period, anyway? 151 quakes in 33 years. There are 26 two-week spans per year, so there are 858 bins in the data set. How do you get 9 quakes per period from that?

 

How do you get 21 riots per earthquake (16 before, 5 after) with 151 quakes and only 64 riots?

[John Cuthber]

John, I understand your skepticism, but I can assure you that I found 64 significant riots in England and Wales between 1980 and 2012.

 

The full list can be found on my website at : http://www.alwriting.co.uk/id4.html

 

I would welcome any comments from you if you doubt the accuracy of any entries on the list.

 

Some years have no riots, particularly during seismically quiet periods.

 

The list will be periodically updated based on comments received. To date the list is as accurate as possible and the theory remains valid and effectively unopposed.

It's the other part of my question that's more important.

How did you find them?

If you are doing this sort of research it's all too easy to fall into the Texas sharp shooter trap.

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

There's also a concern that some of them are clearly driven by external events- the miners' strike protests were carefully coordinated (by the police or the miners depending on whom you believe but they were not "spontaneous").

Some of the events are "copycat" riots (anything Handsworth can do, Nottingham can do better). that amounts to "double counting" of one event.

The London protests about the G20 summit were not a riot.

It's my understanding that so called "football violence" which has little to do with the game is often stage managed in advance by groups who enjoy that sort of thing.

Anti-capitalist events on May day are more likely to be influenced by the date than the earthquakes.

 

So- how do you define a riot?

 

Incidentally, a rough tally (I only looked at years in office, rather than month- feel free to do it properly) says that 52 of the events took place under a Tory government as opposed to 17 under New Labour.

The terms in office were about 17 and 13 years for the two parties respectively.

The Tories had about 57% of the time in office, but 75% of the riots/ events.

 

To me that looks like a rather stronger (and more plausible) effect than seismology (Unless Vulcan hates Tories).

Edited May 9, 2016 by John Cuthber

[Alan Watson]

Swansont

 

Thank you for your continued interest.

 

Firstly, it is 9 riots to be expected in the 14 days before the earthquakes and 9 riots in the 14 days after the earthquakes.

 

64 riots taken in the study period, which to be precise now, starts on 18th December 1979 fourteen days before the first earthquake considered and ends on 11th March 2012, fourteen days after the last earthquake considered. ( I was writing my e-book in mid-2012 so that is where I drew the line ). This study period therefore covers 11,773 days - that is eleven thousand, seven hundred and seventy three days.

 

You have arrived at 858 bins by multiplying 33 years and 26 2-wk periods per year. So you mean there are 858 2-wk periods in 33 years. But that is not the number of 'bins'. I think by your 'bins' you actually mean the 14-day/2-wk periods before the 151 earthquakes. That might change your understanding of bins for your purposes to 151? But please consider the number of days in these 151 x 2-wk periods - it is not 151 x 14 because some earthquakes occur within 14 days of each other. By inspecting all the 151 earthquakes and reducing the total number of days to allow for the over-lap in the 'bins' where the earthquakes occur within 14 days of each other, the total number of days in the bins is 1,700. ( i.e. exactly 1,700 days - that is not an approximate figure ).

 

So the calculation is : 1,700 days in the bins divided by 11,733 days in the study period x 64 riots = 9.24, rounded down to 9 riots would be expected in the 14-day periods immediately before and after the earthquakes.

 

Is this starting to change your view of how you have treated the Standard Deviation in the first part of your query? The bins are actually quite narrow aren't they? What I am saying is if the 64 riots were randomly distributed across the 11,733 days you would expect 9 of those 64 to occur within the 14 days before and after any one of all the earthquakes by chance alone.

 

Continuing with your query, the 21 riots ( 16 before and 5 after ) are about 33% ( i.e. 32.8%) of the total of 64 riots in the period. So the total 'bin' for those 21 riots is 1,700 x 2 = 3,400 which is about 29% (i.e. 28.9%) of the 11,733 days in the study period.

 

Hopefully, this will maintain your interest? I really appreciate your thoughts on this and wish to get to the bottom of your concerns about the SD. Can I suggest you may be interested to know the statistics for the occurence of riots in any 14-day period across the earthquakes cycle. You may be puzzled by your 858 bins and why we seem to have such a small target. Please note the following additional explanation.

 

The relative timing of earthquakes has been studied in order to determine the number of riots which would be expected by chance. The expected average frequency of riots will increase anyway in the period leading up to earthquakes due to the effect of ‘overlap’ in the time periods. The total number of days in the T day periods immediately prior to N earthquakes is somewhat less than T x N because some earthquakes occur less than T days apart. This effect is lessened nearer to earthquakes as T approaches zero and overlap is less likely meaning that the chance of a riot in any given 14 day period increases nearer to the shocks. The 9 expected has allowed for this effect. There are long periods away from earthquakes where no riots occur. My study has also looked at the reduced risk of riots when more than 140 days has passed since the last earthquake. The effect is equally significant, as I have proven statistical significance in unexpectedly few riots >140 days prior to the earthquakes.

 

Kind regards

 

Alan

 

 

John

 

The Oxford Dictionary definition in this context is ' a disturbance of the peace by a crowd; an occurence of public disorder '.

 

I believe every entry in my list qualifies under that broad definition.

 

Equally, I fully understand your concerns about the definition, but please bear with me whilst I alert you to one or two facts.

 

The list of significant rioting and disorder shows that a substantial number of cases appear in clusters with a common initial cause. Dual studies have therefore included both a full appraisal of the cases including the tails of clusters as well as excluding the tails. One would imagine that copy-cat rioting in the tails of riot clusters would be influenced to a lesser extent by seismicity than might be the case for the initial onset of violence. The dual study therefore removes the uncertainties resulting from these potential effects. The outcome is that the case has been proven for both with tails and without tails of clusters.

 

I fully accept your concerns about football violence being organised. So the question needing an answer is : ' How many of the 16 riots potentially influenced by the effects of tectonic stress were football-related ? '

 

Answer : 2 (12.5% of the 16) potentially influenced by the effects of tectonic stress and 11 football-related riots in total ( 18% of the 64 ).

 

However, the 16 riots potentially influenced by tectonic stress are 25% of the total of 64.

 

This shows that football-related disorder is less likely to be influenced by the effects of tectonic stress than other types of riot. So your observations about football-related trouble are entirely consistent with the findings and conclusions of my study.

 

There was significant violence and 63 arrests at the G20 Summit disorder in April 2009 and I stand by my classification of that as a riot.

 

With regard to May Day violence I am sure you are aware that there have been May Day protests / demonstrations every year since 2000 but there aren't riots every year. Neither are there earthquakes of >2.5M_L every year within 14 days after May Day. Please consider that if there are a number of peaceful protests and the rare occurrences of violence happen to be prior to an equally rare earthquake that may suggest a tectonic influence.

 

Thank you for the Texas Sharp Shooter reference - if you can allow me to look into that I will get back to you with some comments. In the meantime look at my last post in answer to Swansont in case that sheds any more light for you.

 

Also, interested in your observations about Tories and New Labour! I'll look at that!

John

 

Following on from earlier, your Tory / New Labour analysis is broadly correct. However, this is a Science Forum and I don't wish this to degenerate into a political spat, so my suggestion would be to leave this to one side for now. The onset of public disorder is influenced by a number of factors as I have commented in earlier posts. This multiple contribution reality merely shows that there may be contributions from other factors not normally recognised, such as the effects of tectonic stress. Please be assured, the underlying tectonic influence is statistically recognisable irrespective of the multi-layered factors at other levels.

 

The coverage of rioting in England and Wales in my e-book looked at the relation between unemployment levels and riots in broad socio-economic terms. I chose to leave out the right / left divisions for the same reason I would rather leave this alone now. Basically, it is political dynamite.

Regarding the Texas Sharpshooter Fallacy, and in answer to both John and Klaynos, I have a mechanism and I hypothesised before gathering the data.

 

A TSF doesn't seem plausible. A suggestion that the 64 riots are an inadvertent sub-set of all the multitude of somethings out of everything that could happen shortly before earthquakes?

 

I could take Klaynos up on his challenge and repeat the test for a different geographical area. Not sure about France - will check it out.

 

Compiling the riot list may take a while.....................time for bed now though. And I have ground investigations to attend to later this week so there may be a lull in the posts over the next few days. Watch this space !

[swansont]

Yes, my mistake, I had that backwards. 9 riots in a two week period.

 

That's even worse, since you only have 64 of them. You claim 21 riots per earthquake actually happened, before and after. That only requires 3 earthquakes. Something is not adding up.

 

The bins I'm referring to is just any random two week period. Surely you analyzed a null hypothesis that there's no correlation at all. The expectation of a riot in any period is less than 0.1 and yet you have 16 of them before an earthquake. Again, something doesn't add up.