FORECASTING  EARTHQUAKES
BY
   PATTERN   RECOGNITION

Report written by E.D.G.  Latest Update: December 20, 2001
E-mail: edgrsprj@ix.netcom.com   Return to E.D.G. homepage


       This report proposes that it might be possible to forecast some earthquakes by running pattern recognition tests on certain types of earthquake and earthquake precursor data.

       It could also be that there are variations in those types of data which are so large that this would not be possible.


DETECTABLE PATTERNS IN EARTHQUAKE
AND EARTHQUAKE PRECURSOR DATA

Seismic Activity and Storm Data
http://www.freewebz.com/eq-forecasting/131.html

and
http://www.freewebz.com/eq-forecasting/132.html

       It is my present belief that many of the signals listed on that Web page represent electromagnetic energy field fluctuations (possibly geomagnetic storm related) which are associated with events taking place in earthquake fault zones around the world in the days and weeks before some of our earthquakes are occurring.

       Data from that Web page being discussed in this present report show that multiple, similar signals are being detected before at least some earthquakes. And that indicates to me that pattern recognition tests on the appropriate types of precursor data might enable us to detect the approach of a certain percentage of our earthquakes.

       Data in the GIF chart discussed in this report also show, I believe, that many earthquakes are themselves demonstrating the existence of those types of potentially useful data patterns.


BACKGROUND THEORY INFORMATION

       I am presently attempting to determine which of the two theories outlined here (if either) explains why those multiple precursor signal events are occurring before some earthquakes.

       My present theories propose that during the days and weeks before many powerful earthquakes occur some type of global electromagnetic energy field is being created. It is especially strong in the vicinity of the fault zone where the earthquake is about to occur. And its appearance and disappearance times are controlled in part by interactions between the gravitational pulls of the sun and the moon and the Earth's crust. At times, more than one approaching earthquake might be involved with the creation of that energy field.

       The following reports discuss how those sun and moon gravity - Earth crust interactions may be taking place. The first two reports are mine. The second one presents a more recent theory picture. The third report was prepared by researchers in Greece.

How Earthquakes Are Being Triggered
http://home.netcom.com/~edgrsprj/121.html

How To Develop An Earthquake Forecasting Program
http://www.freewebz.com/eq-forecasting/127.html

Tidal forces trigger electrical signals generation and large EQs occurrence
http://users.otenet.gr/~thandin/tidalreport/tidalweb.htm

       At some point that global electromagnetic energy field collapses or there is a spike in it. And at that time a detectable precursor signal may be generated. Additionally, while that energy field is active the sun and moon gravities may cause geologic structures around the world such as other earthquake fault zones or the borders of tectonic plates to somehow interact with it and temporarily "short-circuit" it or cause a spike to appear in it. And as some of my Seismic Activity and Storm Data Web page data show, that can happen repeatedly during a relatively short period of time in the days before an earthquake occurs.

QUESTION:    Why are multiple precursor signals being generated like that before some earthquakes?

THEORY 1.    Interactions between the gravitational pulls of the sun and the moon and geologic structures around the world such as earthquake fault zones and tectonic plate borders cause those structures to temporarily and repeatedly drain that global electromagnetic energy field and cause it to collapse for short periods of time. Or, they produce spikes in it.

THEORY 2.    Just as a floating piece of cork is moved forward by an ocean wave, crustal strain related to shifts in those geologic structures is carried along in a westward direction by the "Solid Earth Tide." And when those abrupt increases or reductions in strain reach the fault zone responsible for that energy field they cause the energy field to temporarily and repeatedly collapse. Or, they produce spikes in it.


PRECURSOR AND EARTHQUAKE DATA

       The following are some of my Seismic Activity and Storm Data Web page precursor and earthquake data for the destructive February 4 and May 30, 1998 earthquakes in Afghanistan, and the January 13 and January 26, 2001 earthquakes in El Salvador and India. I believe that they demonstrate that reproducible patterns can be observed in some earthquake precursor data. The two earthquakes in Afghanistan and the one in India occurred around 70E longitude.

       To make them easier to compare with one another the longitudes listed here have had 0, 90, 180, or 270 degrees added to them so that they fall into the range 0W to 90W longitude degrees. An explanation for why I believe such longitude adjustments are permissible can be found in the following Web page:

Longitude Symmetry Explained
http://www.freewebz.com/eq-forecasting/151.html#longitude-symmetry

       The first table listed here contains original data from that 132.html Seismic Activity and Storm Data Web page. In the second table some of those data have been organized into groups of similar signals.

       See the following Web page for more information regarding what the numbers in these tables mean:

Key To The Table Abbreviations
http://www.freewebz.com/eq-forecasting/151.html#key

TABLE 1

Earthquake or precursor signal UTC date + reference number
  |      Signal characteristics
  |        |     Earthquake longitude (adjusted)
  |        |     |    Combined sun and moon gravity longitude
  |        |     |    (adjusted)
  |        |     |    |    Moon gravity longitude (adjusted)
  |        |     |    |    |    Sun gravity longitude (adjusted)
  |        |     |    |    |    |  Ocean tide crest or trough
  |        |     |    |    |    |  | longitude (adjusted - its
  |        |     |    |    |    |  | accuracy is unknown)
  |        |     |    |    |    |  |    Sun - Earth - Moon angle
  |        |     |    |    |    |  |     |
UTC-#   Signal  A-lon  C   M   S   T   Angle
010126-1        20    33  27  46  54     17  INDIA 01/01/26
010125-7   AG         36  32  44  57     12
010125-6   AG         29  25  38  51     12
010125-5   BG         19  15  27  40     11
010125-4   BG         10   6  18  32     11
010125-3   CG         79  76  87  11     11
010125-2   BG         69  66  77  91     10
010125-1   BG         58  55  65  80     10
980530-1        20    51  35   6  74     58  AFGHANISTAN 98/05/30
980526-3   HG         51  47  60  76     13
980526-2   HG         38  34  47  63     12
980526-1   HG         18  14  26  43     12
980204-1        20    52  30  34  71     98  AFGHANISTAN 98/02/04
980130-2   HG         41  32  64  64     33
980130-1   HG         24  15  46  47     32
971215-1   E          36  44  58  73   -166
010113-2        89     3  26  81  63   -125  El SALVADOR 01/01/13
010111-3   BG          9  24  51  56   -154
010111-2   BG         60  75  10  17   -156
010111-1   AG         38  52  77  84   -157


OBSERVATIONS AND THEORIES:

***  With these types of earthquakes where more than one precursor signal is detected, all of signals appear to be generated within perhaps a single 6 hour period of time. And I suspect that since that proposed global electromagnetic energy field is controlled to a large degree by the gravitational pulls of the sun and the moon its shortest cycle might be a little over 6 hours long.

***  As those India, El Salvador, and May 30 Afghanistan earthquakes show, the longitude numbers for the final signal in those groups often match those of the earthquake itself. I suspect that when that final signal is generated it indicates that the global energy field has collapsed for the last time. And the fault zone has now entered into a metastable condition which often ends within a few days when the earthquake occurs.

***  The first three signals detected before the India earthquake were about 45 minutes apart. Then there were no signals for about 85 minutes. Then the second group of 3 signals was detected, again, about 45 minutes apart. And the final, seventh signal was detected about 30 minutes after the sixth one.

       With both of those two groups of 3 signals, the sun longitude for the first signal became the moon longitude for the second one. And the sun longitude for the second signal became the moon longitude for the third one. That suggests to me that the signals were linked with something which was oscillating at a fixed frequency. And a theory which I am presently considering proposes that those signals might have been linked with something taking place in the vicinity of the Juan de Fuca tectonic plate. A powerful and destructive earthquake occurred about a month later in the Seattle, Washington, USA area.

Tectonic Plate Map (might no longer be available)
http://k12.rutgers.edu/curriculum/musicalplates/platemap.html

***  Because of that previously discussed possible interaction between the earthquake fault zone and other geologic structures, when I and probably some other researchers detect a single precursor signal before an earthquake, the signal might point to the wrong fault zone and make it difficult or impossible to determine where the earthquake is going to occur. I suspect that this could be especially true for earthquakes which occur in Japan and Central America locations where there are multiple fault zone borders in the same general area and a number of different ways in which an earthquake can be triggered.

TABLE 2

UTC-#   Signal  A-lon  C   M   S   T   Angle
Group 4
980530-1        20    51  35   6  74     58  AFGHANISTAN 98/05/30
980204-1        20    52  30  34  71     98  AFGHANISTAN 98/02/04
980526-3   HG         51  47  60  76     13
971215-1   E          36  44  58  73   -166
Group 3
010126-1        20    33  27  46  54     17  INDIA 01/01/26
010125-7   AG         36  32  44  57     12
980526-2   HG         38  34  47  63     12
980130-2   HG         41  32  64  64     33
Group 2
010125-6   AG         29  25  38  51     12
980130-1   HG         24  15  46  47     32
Group 1
010125-5   BG         19  15  27  40     11
980526-1   HG         18  14  26  43     12


OBSERVATIONS AND THEORIES:

***  Group 1 signals were detected first in those sequences followed by Groups 2, 3, and 4. Data for the earthquakes themselves matched Groups 3 and 4.

       Those four groups of data points indicate to me that before some earthquakes which are about to occur in fault zones having the same longitude, precursor signals may be generated when the sun and the moon are in certain longitude positions in the sky relative to the fault zones. Fault zone latitude and sun and moon position latitudes could be of less importance to that phenomenon. If reliable fault zone longitude dependent signal patterns are in fact being generated before some earthquakes then perhaps pattern recognition tests run on the appropriate types of precursor data could be used to tell when and where those earthquakes were going to occur.

***  Variations in the signal and earthquake longitudes within each of those groups might be associated with the facts that values for the "Theta" angle were different for each of the three earthquakes and their precursors as were the latitudes of the positions of the sun and the moon in the sky. In my reports Theta is the angle formed by lines drawn between the centers of the sun and the Earth, and the centers of the Earth and the moon. Positive values for Theta indicate that the sun position in the sky was 0 to 180 longitude degrees to the west of that of the moon.


OBSERVABLE PATTERNS IN EARTHQUAKE DATA

       Data points in the following GIF chart and the How Earthquakes Are Being Triggered report which discusses it indicate to me that some earthquakes are themselves demonstrating the existence of these types of data patterns.

       In this chart the longitudes (horizontal axis) of about 80 widely felt or destructive earthquakes occurring around the world plus dozens of those earthquakes occurring in the Japan area and dozens more occurring in the People's Republic of China were plotted against Theta, the sun - Earth - moon angle (vertical axis). Positive values for Theta indicate that the sun position in the sky was 0 to 180 longitude degrees to the west of that of the moon.

Click here for a larger and clearer version of that GIF chart
http://home.netcom.com/~edgrsprj/121b.gif

***  The fact that the Theta values in that chart for earthquakes occurring in western Pacific Ocean areas merged into those four circled, somewhat symmetrically arranged groups indicates to me that many earthquakes in those areas are following distinct triggering patterns.

***  Data points along the line drawn to the left of the circled group in that chart which is centered around +85 Theta degrees indicate to me that a gradual decrease in Theta values for that group can be observed as earthquakes are examined which have longitudes farther and farther to the west of the circled group (across Asia and into Europe). That indicates to me that some earthquake triggering patterns are fault zone longitude dependent.


RECOMMENDATIONS

(1)   It is my recommendation that geophysicists, seismologists, geologists, and professional and amateur earthquake forecasters study the data listed on those two Seismic Activity and Storm Data Web pages and in that GIF chart and attempt to determine if some earthquakes might in fact be forecast through the use of pattern recognition type tests. Signals listed on those Seismic Activity and Storm Data Web pages which have a "G" present in their "type of signal" column could be the most useful.

(2)   Earthquake forecasters should compare the times when signals listed on that Seismic Activity and Storm Data Web page are being detected with the times when their own precursor signals are being detected. If matches can be found between them then that would probably indicate that the signals are being controlled by the same processes.


COMMENTS

(1)   It is my expectation that even more accurate earthquake forecasts might be made if triangulation based detection procedures were used at the same time as the proposed pattern recognition tests discussed in this report.

(2)   I suspect that "Solid Earth Tide" data for the signals and earthquakes discussed here might produce better results than the sun and moon position and ocean tide crest and trough data with which I am presently working.


Homepage For This E.D.G. Web Site
This first version of this Web page was stored at this site
on June 12, 2001. Comments regarding this Web site are
welcome. Please send them to:
edgrsprj@ix.netcom.com
Information on this Web page represents expressions
of personal opinion. The address -URL- for this Web
page is:
http://home.netcom.com/~edgrsprj/125.html