A Discussion on the Methodology of Earthquake Prediction

Taking a cue from a poster in a thread that had become very popular very quickly before it was classified as a hoax, I am beginning this thread based upon information that was uncovered which is unquestionably NOT a hoax.

To whit:

Originally posted by jadedANDcynical
Let's dig in to this and see if we can shake anything out.

...The computer model was generated under NASA's QuakeSim project, a computational framework for modeling and understanding earthquake and tectonic processes. QuakeSim focuses on deformation of Earth's crust, which can be measured using airborne and spaceborne technologies. The models and data can be used to better understand earthquake hazard, stress transfer between faults, and ground disturbance following earthquakes.

QuakeSim, a collaboration of JPL-Caltech, USC, UC Davis, UC Irvine, Indiana University, and NASA Ames, is sponsored by NASA's Advanced Information Systems Technology Program through the Earth Science Technology Office.

source

So we know for a fact that JPL works with quite a few groups studying earthquakes.

QuakeSim

We are forming a community-led InSAR Working Group dedicated to the advancement of radar remote sensing research. The potential of a robust InSAR observational capability has generated strong interest amongst the research and applications communities. The role of InSAR spans a broad spectrum of end uses including crustal deformation science related to earthquakes, volcanoes, hydrologic processes, ice sheet and glacier variability, vegetation structure, and disaster management. Long-term access to InSAR data will greatly advance our understanding of how these basic processes affect life on Earth. Consequently, the US scientific community should devise a long-term strategy for US InSAR activities, including the funding of dedicated US InSAR satellites, access to foreign SAR data, and continued education and advocacy for InSAR science.

More JPL stuff of interest

This report summarizes the major findings of a symposium attended by 260 scientists and engineers in an effort to guide U.S. efforts in Interferometric Synthetic Aperture Radar (InSAR), a critical tool for studying dynamic changes of the Earth’s surface and natural hazards associated with these changes. InSAR observations provide critical and otherwise unavailable data enabling comprehensive, global measurements to better understand and predict changes in the Earth system. The InSAR Workshop was funded jointly by NASA’s Earth Science program, the Geosciences Directorate of the National Science Foundation, and the U.S. Geological Survey. We hope that these and other agencies heed the call for a coordinated InSAR program to address these important research questions

From a report dated 2004

More:

Natural hazards. SAR interferometry has demonstrated valuable information for monitoring and predicting or forecasting a variety of hazards, from air, water, and earth. Large-scale hazards generated in the Earth include earthquakes and volcanic eruptions; each is driven by tectonic forces within the Earth’s crust. Observation of deformation from subsurface flow of magma and of the accumulation of strain within the crust is needed to be able to understand these great forces of nature. More localized, but often intense, hazards include landslides, mud flows, and land subsidence or collapse due to natural or human removal of subsurface material or fluids and permafrost melting. Flooding is the most damaging hazard in most areas, from rainfall, snow, and ice melting, and natural or human-made dam collapse. In coastal regions, hurricanes, intense local wind events, shore erosion, and oil spills are major hazards. Finally, fire in forests and other vegetation is a major hazard in many areas. For each of these hazards, InSAR has proven a help in assessing damage after the events and evaluating the risk of future events by understanding and monitoring the processes involved

Source

black bagged in:

5...
4...
3...
2...
1...

Love it. I'll be doing some more research and come back when I have something constructive to add...but yah, I had no idea these programs were even out there!

At least this came to light, so I see the whole thing as a big plus!!

Just to add the information I already found to this new thread, I will quote myself

Originally posted by Juggalette
Regarding M8... Out of curiosity, I searched this term in Google scholar: "Keilis-Borok's M8 prediction methodology" and came up with this paper Earthquake Prediction: State of the Art and Emerging Possibilities written by Vladimir Keilis-Borok. Looks like the paper is dated 2002.

An excerpt from the paper reveals this regarding the M8 earthquake prediction algorithm:

ALGORITHM M8

"This algorithm was designed by retrospective analysis of seismicity preceding the greatest (M 8) earthquakes worldwide, hence its name. We describe it using as an example the recent prediction of major earthquake in Southern Sumatera, Indonesia, on June 4, 2000, M D 8.0. That prediction was part of the Russian-American experiment in advance prediction of strongest earthquakes worldwide."

It goes on to give the specifics of the algorithm and how it "works" (sorry not a math person

), and how several other prediction methodologies that were used have worked. The paper also states that: "Algorithm M8 successfully predicted all six strong earthquakes that occurred from 1992–2000.", making it the most successful of the algorithms used.

Very interesting information!

reply to post by jadedANDcynical


I'm so glad you started this thread!! I was actually thinking the same thing after following the "thread that shall not be mentioned."

I have been so impressed with a some of our ATS earthquake aficionados, and the raw data they watch/analyze - I'm hoping this will wind up being a win. I'll be very interested to see their input and discussion.

S & F to the OP!!

I too am very interested in this thread and the information it may bring. Looking forward to learning more.

In this 2008 interview from Last Exit Magazine with Vladimir Keilis-Borok (who, from what I gather, was one of the original scientists to "work out" the M8 algorithm) I found this description of not only the M8 algorithm, but two other algorithms (SSE and RTP):

"Algorithm “M8”. Since 1992 it has captured into the alarms 11 out of 15 earthquakes with magnitude 8 or more worldwide. Altogether the alarms cover one third of the time-space considered. Among missed is the recent Wenchuan earthquake in China. A typical alarm covers time interval years and area hundreds to thousand km in diameter."

"Algorithm Second Strong Earthquake (“SSE”) is applied after a strong earthquake occurs. It predicts whether a next strong earthquake will or will not occur nearby during the 18 subsequent months. 30 predictions have been made since 1989. 24 of them were correct; [the errors include 2 failures-to-predict and 4 false alarms)."

"Algorithm “Reverse Tracing of Precursors” (“RTP”) gives shorter alarms, down to 9 months. It is tested since 2003 in California, Northern Pacific, Eastern Mediterranean, and Italy with adjacent areas. By now 5 out of 6 target earthquakes are captured by the alarms; and 9 alarms were false (two of them being near misses)."

So apparently, there is more than one tool they are using for modern earthquake prediction. However, it looks like the shortest time span for prediction is 9 months? I will look into this.

Fascinating!

ETA: I wonder if they applied the "SSE" algorithm to Japan after the 9.0??

I read through the USGS PDF in the OP....I'm still a little in shock:

Invite NASA to give a presentation on that agency's support of earthquake prediction research

Invite Tom Jordan to give a presentation on the Collaboratory for the Study of
Earthquake Predictability.

[ex ]USGS to provide NEPEC members with an appropriate selection of prediction literature.

"It was suggested that the Council should develop a set of generic statements that could be
applied to some of these situations such as unusual natural events, interesting scientific
approaches that have been blown out of perspective through media, and dismissals of
crackpot ideas."

"The council also discussed the timelines under which it would operate, with Chair Jim
Dieterich indicating that there would likely be situations where the Council would need
to deliberate quickly in response to a specific prediction that is generating attention."

linkearthquake.usgs.gov...

The only folks I recall predicting earthquakes would fall into USGS's "crackpot" category. Yet here they are admitting they can predict them!

Great minds think alike.

And fools seldom differ.

Let us demonstrate the former rather than the latter, shall we?



Further excerpts from the InSAR Workshop Summary report linked in my OP:

2. What stress transfer processes are important in triggering seismic activity? Are long-range interactions important

... Current research is very actively elucidating the nature of the earthquake/earthquake interactions, rigorously quantifying the statistical likelihood of linkages, and beginning to shed light on time-dependent processes (e.g., post-seismic relaxation, state/ rate fault friction) that influence triggered activity. However, emerging clues suggest longer-range interactions that are not mechanically understood. Any linkages should have deformation signatures, and synoptic InSAR imaging offers possibly the best means of detecting and elucidating the deformation causes and effects that may link regional earthquake events

emphasis mine

They seem to be saying that it is logical that the stresses and deformations preceding an earthquake should be noticeably different from those following said quake.

And here we come to the $64000 question:

3. Are there precursory deformation phenomena for either earthquakes or volcanoes and can they be detected with InSAR observations?

This is the Holy Grail for solid-Earth natural hazards research. Current earthquake hazard maps are at a coarse resolution in both time and geography. Such maps depict probability of exceeding a certain amount of shaking (generally that at which damage occurs) over the next 30 to 100 years, depending on the map. The spatial resolution is typically on the order of tens to hundreds of kilometers. These maps are based on information about past earthquakes observed in the geological or historical record. Measurement of crustal deformation, usually acquired using GPS, now provides information on strain rates; generally we find that earthquake rates are higher where strain rates are higher. The number of GPS stations that can be deployed on the ground limits the resolution of strain, and these stations can be expensive to install and maintain

They're making the case for their technology as being better suited for the study of stresses and deformations than that of using GPS. This certainly seems logical in that you can only place a finite number of GPS stations in a given area thus limiting the amount of precision measurements than can be made.

Think of trying to watch a HD program on an old-fashioned CRT television. Yeah, you might be able to see the general picture, but you're going to miss a LOT of detail.

Furthermore, future science studies of crustal deformation will yield insights into earthquake behavior, whether high strain rates indicate the initiation of failure on a fault or quiet release of stress, and how stress is transferred to other faults. These studies will lead to science findings for improvement of earthquake hazard maps both spatially and temporall

WC, this has implications for your slow-slip faulting, does it not?

The following is a screen grab from the .pdf
[atsimg]http://files.abovetopsecret.com/images/member/c696c23bcd51.jpg[/atsimg]

If you've ever seen stress pattern analysis done for machined parts, you will recognize the colored bands in the image. It is a way to determine failure points when stresses are introduced into a machined or cast part.The same principle applies here I would imagine.

Volcanoes aren't left out either:

Similar studies employing InSAR to map deformation on volcanic terrain can reveal subsurface transport of magma, an important factor affecting eruption probabilities. Detailed maps of the shape of the magma trail give clues as to where pressure may accumulate and also may help constrain the explosiveness of the potential eruption

And

[atsimg]http://files.abovetopsecret.com/images/member/6012e2b5459f.jpg[/atsimg]

As you can see from the images, there is an incredible amount of detail available with this method of scanning an area.

It remains to be seen if there is any connection between InSAR and M8, but we now have a thread in which to discuss such topics

reply to post by jadedANDcynical


Oh wow. I feel like a little kid that was just handed a HUGE loli pop! I don't know quite what to do with it, other than lick away and hope I get something out of it.

Wow...such a lot new information...and so many paths to go on. I know (now) that the Middlebrook reference is probably a red herring...but I am fascinated by the list of Thmas Middlebrook ASCM Awardees...and power company permitting (we lose electricity due to fog so who knows what havoc Ms. Irene will cause) I will continue reading about these scientists...and what they were noted for with their TM Award.

The M8 itself is fascinating like (real life) science fiction. Just a theory, but as Irene moves closer to the 37 parallel (the lat of the Mineral VA quake, but also that of New Madrid, Missouri; Trinidad, CO another recent 'hot spot') if the storm surge waves have any effect on the quakes. I wish I could understand (or even spell) alogrithms...

just marking this so i dont loose this thread.and i have figured out ats has some of the smartest cookies around(and fastest typers too) thanks for all your digging on this and on previous thread alot of good info to digest. VERY GOOD WORK GUYS AND DOLLS

I wanted to bring this post over from the thread-that-must-not-be-named as I felt it goes toward the intent of this tread very fittingly. I had hoped this poster would do so, but he has not

Megaboogie, I hope I am not stepping on your toes here you deserve full credit for the finding.

Originally posted by megabogie
]I found this looking up the Kellis-Borok connection, it's an abstract from a scientific paper

Premonitory activation of earthquake flow: algorithm M8

Thirty-nine out of the 44 strongest earthquakes which have recently occurred in different regions of the world are preceded by specific activation of the earthquake flow in the lower magnitude range. This activation is depicted by the algorithm M8, which was designed for diagnosis of times of increased probability (TIPs) of strong earthquakes.

The paper is authored by V. I. Keilis-Borok and V. G. Kossobokov aand it's dated from 1990

LINK:

www.sciencedirect.com...

reply to post by jadedANDcynical


Thank you Jaded for bringing that over. I wasn't sure if I was allowed to repost. I'll keep digging too...this topic
fascinates me! And by the way, I'm a she!

reply to post by megabogie


It figures I'd get that wrong...my most humble and abject apologies, m'dear.

A lot of people assume I'm female due to the continual shortening of the first part of my moniker to "jade, " which most often is a female's name rather than the full "jaded" which means worn out, tired (of the things that have made me cynical btw), but I am unquestionably male, or so say my 5 kids.

anywho...



A little more on M8:

The “M8” algorithm, originally developed for intermediate-term prediction of large events, uses a catalog of mainshocks to identify large scale seismicity patterns before large earthquakes in a given region (e.g. Gabrielov et al., 1986; Keilis-Borok et al. 1988; Keilis-Borok et al., 1990; Updyke et al., 1989; Healy et. al, 1992; Kossobokov et al., 1992, Keilis Borok and Rotwain, 1994; Kossobokov and Mazhkenov, 1994). We have developed an approach to measure the stability of the results, and to test specific hypotheses (e.g. Minster and Williams, 1992, 1994, 1995, 1996), and used it to assess the performance of M8 intermediate-term earthquake prediction algorithm.

Source

reply to post by jadedANDcynical


Thank you Jade. I have it downloaded. Hugs to you!

In looking through the link I got from Jade I found this dissertation.

[PDF] Methods for evaluating earthquake predictions
[PDF] from usc.eduJD Zechar - 2008 - earth.usc.edu
Page 1. METHODS FOR EVALUATING EARTHQUAKE PREDICTIONS by Jeremy
Douglas Zechar A Dissertation Presented to the FACULTY OF THE GRADUATE
SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial ...
Related articles - View as HTML - All 7 versions

reply to post by jadedANDcynical


Hey there! Just wanted to let you know I did a bit more searching regarding my thread that got moved to the hoax bin. Just made another post with some info I think you might be interested in.

It is somewhat relevant to this thread...but I don't want to cross-post for obvious reasons.

Great stuff! Love the learning process....

Well, the pot is stirring again. So I thought I would post a few more tidbits I've found in looking into this topic (earthquake prediction) and see if there is anything interesting or germane to the issue.

It would seem as though Italy is ahead of the game as far as actually utilizing these technologies:

The tests of predictions, performed on a global scale, allowed a first statistical assessment of the predictive capability of M8 and CN algorithms (Kossobokov et al., 1999; Rotwain and Novikova, 1999). Specifically, for the M8 algorithm the results obtained in real-time prediction mode since 1992 have already demonstrated the high confidence level (above 99%) of the prediction of the world’s largest earthquakes, in the magnitude range 8.0 – 8.5 (Keilis-Borok and Soloviev, 2003; Kossobokov et al., 1999). For the algorithm CN a preliminary estimate of the significance of the achieved prediction results, obtained for the period 1983-1998 in 22 regions of the world, gives a confidence level around 95% (Rotwain and Novikova, 1999).

Whoa!

99% & 95% confidence levels?! I don't know what to say here. To think that this type of predictive capability exists and has not been integrated into some sort of alert system (even a very preliminary one) reeks of intentional deception.

Maybe I am missing something here, please help me out on this someone.

Several experiments have been dedicated to assess the robustness of the methodology against the unavoidable uncertainties in the data (Peresan et al., 2000; 2002). With these results acquired, an experiment was launched in July 2003, aimed at the real-time test of M8S and CN prediction for earthquakes with magnitude larger than 5.4 in the Italian region. The results of the intermediate-term middle-range predictions in Italy are routinely updated and made accessible to a number of scientists. The goal of the experiment was to accumulate a collection of correct and wrong predictions (the latter include the false alarms and/or the failures to predict encountered in the test) permitting to verify and assess the predictive capability of the considered methodology.

Source .pdf < 2.7 MB

So they've been at it for at least 8 years. I wonder what they've been able to accomplish.

From .pdf found here:

Earthquake prediction is very probable and is being used today.

Some info on the M8 algorithm.

A cluster of "false alarm" TIP's in Japan lasted from the middle of 2001 to the end of 2010 gradually migrating from southwestern to the northern regions: It started with CI #64, expanded in 2006 to CIs #82 then #80; in 2007 #64 was called off and ##80-82 formed a new area of alarm, which has shrunk to a single #81 in January 2010. The alarm can be associated with the failures-to-predict the great 25/09/2003 M8.3, 15/11/2006 M8.3, and 13/01/2007 M8.2 earthquake, each of which were linked by the RTP chains of correlated quakes (Keilis-Borok et al., 2004; Shebalin et al., 2006) to the M8-MSc prediction areas, as well as to a series of earthquakes started with the 2002/06/28 M7.3 deep event (depth 566 km) near Priamurye-Northeastern China border in the back of and outside the alerted section of subduction zone, followed by twelve shallow magnitude 7.0 or larger earthquakes in the area alerted in 2002-2010, and ended with the 09/08/2009 M7.1 deep earthquake (depth 292 km) beneath Izu Islands. The Tokai silent earthquake initiated in 2001 and lasted for many years in the middle of this cluster of "false alarm" could be also physically related phenomenon. Finally, the mega-thrust on 11 March 2011 completed the peculiar history of the "false alarm" in Japan on the 70th day after it was formally called off (see figure above). Its first aftershocks (white dots) spread along the entire patch of the M8-MSc prediction outline. The M7.3 earthquake on 9 March 2011 preceded the mega-shock by 51 hours and, according to Takeshi Kudo (personal communication), was a suspect foreshock of the expected by Japanese seismologists "M7.5 class Miyagi-Ken-Oki earthquake". The premature termination of TIP in the 2011a Update of the Global Test of M8-MSc predictions happen to be due to the function Z1 (an inverse of the Zhurkov criterion, which is linear concentration of ruptures) anomaly threshold change from 2407 to 2440 bringing the voting scores from the required (and factual in the 2010b Update) 4:6/4:6 down to 4:6/4:5... A reasonable man would not even notice this change in one of the seven graphs involved in TIP diagnosis, while the prefixed in 1992 "black box" version of the M8 algorithm does. It is hard to disagree with Prof. Keilis-Borok saying: "The alarm, its premature termination notwithstanding, could have been used for damage reduction." The magnitude of the 2011 off the Pacific coast of Tohoku Earthquake might be larger than 8.9 (preliminary Global CMT estimate is 9.1 waiting for longer period records). On the other hand, unlike the 26 December 2004 mega-thrust, the rupture in Sendai earthquake is surprisingly limited to about 400 km. This compactness may explain why its precursory patterns were not recognized by M8 aimed at M9.0+, but were diagnosed in advance the 2010 Chile mega-thrust with about 600 km of rupture. The current alarm for M8.5+ in Japan was not terminated in January, however, the M9.0+ and M8.5+ ranges are outside the scope of Global Test of the M8-MSc predictions and could not be considered as documented in advance the 11 March 2011 mega-thrust.

Abbreviation: TIP, time of increased probability of a strong earthquake (an alarm).

And ,a warning to keep it from the public,hence the restricted access on my link.

Although the M8-MSc predictions are intermediate-term middle-range and by no means imply any "red alert", some colleagues have expressed a legitimate concern about maintaining necessary confidentiality. Therefore, the up-to-date predictions are not easily accessed, although available on the web-pages of restricted access provided to about 150 members of the Mailing List.

www.mitp.ru...

And,if you really want to learn more about how this prediction process works,here is a link to download a power point presentation.
www.lmd.ens.fr...

Scroll down to [email protected] and download it.