Skeptical Inquiries

Astrologers since ancient times have always asserted that the planets influence earthquakes. The timing of earthquakes has largely been determined by examining the timing of planetary aspects. That is, when the planets form precise geometric angles to one another earthquakes are supposed to occur. While the astrological community has generally accepted this theory the scientific establishment has largely ignored and ridiculed such assumptions. For the astronomers this would seem to imply effect at a distance, something that they have ruled out as possible. During the course of many years of astrological research it was noted that earthquakes did indeed seem to be associated with the aspects and many significant earthquakes seemed to be associated with major planetary aspects. Another astrological premise is that applying aspects, that is aspects between the planets that have not yet fully formed their perfect angles, are more active than separating aspects, i.e., ones that have already passed the exact angle. As an example the magnitude 5.7 earthquake that occurred in Jawa, Indonesia on March 12, 2001 at 23:35:08 UT, at 7S02 and 106E05 had aspects of Sun semi-sextile Uranus seven minutes applying, Sun semi-square Neptune, twenty-four minutes applying and Mercury square Saturn 58 minutes applying. Thus, this earthquake would be considered to be under an applying aspect as the Sun was applying by semi-sextile to Uranus at an angle of 30° 07’.   

 

Astrological theory states that the change in the level of activity initiated by the applying aspects should be apparent as an increase in the level of geoseismicity. The number of earthquakes that occur during applying aspects should be significantly higher during periods of applying aspects when compared to periods of separating aspects.   

 

Random earthquakes were observed over a period spanning many years in order to see if there was a correlation between the angular positions of the planets and the occurrence of quakes on the earth. What was found was that there appeared to be more earthquakes when there was an abundance of planetary aspects. From this beginning there began a program to try to predict when earthquakes would occur. The results were not very promising, but there was enough correlation to proceed with further investigations. An analysis of a series of randomly selected great earthquakes that occurred during the twentieth century was undertaken. There did seem to be more exact angular relationships between the planets than was to be expected, but this is a condition that is actually much more difficult to determine than it would at first appear. This difficulty arises due to the fact that the motions of the planets as observed from the earth follow a rather tortuous path and it cannot be easily determined what the probability of any given position of a planet will be. These apparent positions are referred to as the geocentric co-ordinates of the planets. The positions of the planets as viewed from the sun, their heliocentric positions, are much more easily predicted. These, like the geocentric positions, also seemed to indicate an excess of planetary aspects when the earthquakes were occurring, although what exactly was happening during these times was extremely difficult to determine.

 

Attempts at producing control charts to compare against the earthquake charts were initially largely unsuccessful. This was due to the fact that repeating cycles of planetary positions as viewed from the earth and the uneven apparent motions of the planets muddied all the results. It was somewhat easier to propagate a control group for the heliocentric positions of the planets. These comparisons did indicate that there were more aspects occurring during the earthquakes that during the randomly selected times, but the statistical relationships were not strong. One of the problems that arose was that the exact times of the earthquakes that were under study were not known precisely and this factor added too much uncertainty to the results. When the earthquake data collecting began in the 1970’s the circumstances surrounding the quakes such as their exact time and location were not easily obtained. This was before the era of the Internet and the posting on it by the U.S. Geological Survey and others of the exact times, positions and magnitudes of earthquakes. The Internet increased accessibility to the relevant data and this has facilitated the study thousands of quakes and has thus eased the attempt to find any correlation between the times that they occur and the corresponding positions of the planets in space.    First obtained was the data for the most significant earthquakes of 1998 as published by the U.S. Geological Survey. This set was comprised of eighty-four events. The astrological program "Solar Fire" was used to calculate the positions of the planets and their apparent positions in the sky at the exact moment and position on the earth to the nearest arc minute of longitude and latitude and to the nearest second in time of the quakes. Next was calculated the exact geocentric positions for the planets at the moment of each quake were calculated. From these times and dates there was generated a control group of charts that were made by using a random number generator to determine the dates for an equal number of random sky charts for 1998. This generated a data base of planetary positions and relationships that would give a very similar configuration of planetary positions as the earthquake data, but would be sufficiently different enough from the earthquake times to provide a completely dissimilar set of planetary configurations.  

 

 The initial results from the earthquake and random data bases using the angles of 60, 120, 180, and 0 degrees, the sextile, trine, opposition and conjunction aspects, showed that more aspects were present during the moments that earthquakes occurred than during the random periods. The difference between the two sets of data was not large, but it indicated that the initial presumptions regarding the observations that more aspects were occurring during earthquakes than at other times was supported. There were 781 aspects for the random group and 800 for the earthquake group.    The second set, utilizing the same aspect criteria, was generated from the heliocentric positions of the planets for both random times and the times that earthquakes occurred showed very similar results with 539 aspects occurring during the times that earthquakes occurred and 511 aspects for the random group. This was a slightly greater correlation than for the geocentric group and again supported the theory that more aspects were in force during times of aspects that during random times.    The standard orbs, the span during which the aspect is considered to be in force, were comparatively large and were actually in effect for several days before and after the earthquakes actually happened. To more closely examine the exact period of the actual earthquakes the "orb" was reduced to plus or minus one degree of arc. Two more aspect databases were generated and then compared. The earthquakes had 132 geocentric aspects and the random database produced 125 aspects. This is exactly the same ratio as for the first group of geocentric positions and standard orbs and further supported the theory that aspects were more abundant during periods when earthquakes occurred than during random times. The statistical strength of these ratios was not very pronounced and suggested that the results could be due to random error. To further reduce this possibility another set of events was generated which consisted of the ninety- six most significant earthquakes that happened during 1997 as determined by the USGS. The same type of analysis was carried out with these events as for the earthquakes of 1998. The random set was generated in a different manner, however. The random group was generated by taking the same time and place of the actual earthquake and subtracting three days from that moment and generating a new set of planetary positions for the new time. The results showed that there were again more aspects in force, using the standard orbs, than during the period three days before with 341 aspects in effect during the earthquakes and 307 during the random times. As a result of these findings the relationship between earthquakes and planetary positions appeared to be getting stronger.   

 

In order to determine if applying aspects were more predominant than separating aspects during earthquakes I looked at what aspect had the smallest orb when the earthquake occurred and noted whether, or not it was applying, or separating. For the 1996 group there were 42 earthquakes occurring during an applying aspect and 38 during a separating aspect. During 1997 there were 58 earthquakes with applying aspects and 38 with separating aspects. In 1998 there were 46 earthquakes that occurred when applying aspects dominated, while there were again 38 earthquakes with separating aspects. Once more this reinforced the notion that the aspects were generating higher levels of seismic activity than when compared to times when there were no aspects. Eventually 38 groups of earthquakes were generated from earthquakes that occurred from 1975 to 2000 of about 30 events each. (see Table 1.) Percentages of each group that were applying were compared to randomly generated percentages. These two sets of data were then compared and a t-test was performed upon the results. This t-test illustrated that the probability that the results were due to random errors was very remote. (see T-test 1.) Note: for all t-tests variable 1 is random. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

T-test 1 for all data bases examined  

 

 The most noticeable difference between these two groups of numbers is that the mean for the earthquake group is 61.8% of all earthquakes occurred during applying aspects, while for the random group they accounted for only 50.62% of the whole. We can also look at this study in a different manner to determine the probability that the results are due to random error. The probability that any given aspect is either applying, or separating is 50%. While there is a slight excess of applying aspects over separating aspects due to the idiosyncrasies of retrogradation, this effect is only seen with planets that appear to go through retrograde motion. The bulk of the total planetary aspects are with the sun and the moon, interactions with the planets and that are why there is a 0.62% excess of applying aspects in the random group. For all practical purposes we can assume that the relationship between applying and separating aspects is exactly the same as flipping a coin. The probability of a coin toss resulting in a non-random distribution for heads being up is 1 divided by two to the power of the number of times that you toss the coin. In this case we have only one set of earthquakes with a less than 50% distribution of applying aspects. Therefore, we have 2 divided by two to the 37th power as the probability that the results are due to random error. This is equal to approximately 275 billion to one odds against chance. It seems very unlikely that this would occur when flipping coins. This series however, is very fortuitous due to the fact that that the investigations were specifically looking for positive results. Some of the databases had inadvertently included overlapping data. If we look at all earthquakes and sequential databases that include all aspects the number of positive results reduces to about 70%. This would reduce the random chance in the above series to about 68 million to one odds. A t-test performed on a more critical subset of 30 databases demonstrates that there is a very low probability that the results are due to chance. These databases remove all overlapping data from the larger series that was previously analyzed.  T-test 2 for non-overlapping databases  t-Test: Two-Sample Assuming

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  During the next study the various aspects were examined to determine which were affecting the general seismicity of the earth. It was found that the conjunction, semi-square, sextile, square, trine, sesqui-quadrate and opposition were the only aspects that resulted in non-random effects when evaluated. When any other aspects, such as the septile, or novile, were added to the mix, then the results turned out to be random. One interesting thing to note is that when only aspects of the eighth harmonic used the results always tended to increase the number of applying versus separating aspects. For the final study 360 earthquakes were randomly selected and a control group of 360 non-earthquake horoscopes were constructed.

 

    The control group was constructed by taking earthquake times and subtracting three months and three days from the earthquake times and then erecting the resultant horoscopes. For each group the results were analyzed comparing the total number of applying and separating aspects that were dominant in each chart. The determination of the dominant aspect was determined by observing what aspect was closest to the point of perfection and whether, or not that aspect was applying, or separating. An aspect was considered to be any planetary combination involving angles of the twenty fourth harmonic that is traditionally considered to be an aspect. In other words the only "aspects" under consideration are the conjunction (0° ), semi-sextile (30° ), semi-square (45° ), sextile (60° ), square (90° ), trine (120° ), sesqui-quadrate (135° ), inconjunct (150° ) and opposition (180° .) Previous studies showed that any inclusion of other angles from the 24th, or any other harmonic, resulted in an increase in the randomization of the results. The only angles that resulted in a decrease in random results were the angles above. The 360 earthquakes were randomly selected from a random database of earthquakes that included events that occurred over the last approximately 100 years. Most of the events occurred during the 1990’s, a large number were from May, 1975 and a small number from 2000. All of the events were courtesy of the United State Geological Survey’s earthquake databases available on their web site. All events had the exact time to the nearest second recorded and the position to the nearest minute of arc. The random events were recorded to the same level of accuracy.The initial results of the trial showed that there were 61.8% of the earthquake charts that had applying aspects as the closest aspect, while the random group had 49% applying aspects and 49% separating aspects and 2% with no aspects less than 1 degree. In other words there were 61.8% of the earthquake charts with applying aspects and 38.2% with separating aspects. The random charts were virtually evenly split between applying and separating charts. Three hundred charts from each database were then divided into groups of 10. These 30 groups of earthquake and random charts were then analyzed by t-tests giving results that were very significant and virtually ruled out any possibility of the results being due to random error. The first t-test shows the probability that differences between the earthquake and random groups are not likely due to random error

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   If we hypothesize that the mean difference will be 3.3 for each group of thirty events then the results give a Virtually eliminates the probability that the results are due to chance. (The value of 3.3 is derived from the fact that 61% of 30 is 18.3 and 50% is 15, thus the mean difference should be 3.3.) 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

    A third t-test was performed to determine if the difference between the number of applying versus separating aspects for the earthquake charts was significant. The results for this test were that the probability that the results were due to chance was very low.  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  t-Test 3: Paired Two Sample for Means  Variable 1Variable 2Mean11.617.93333Variance19.6275919.23678Observations3030Pearson Correlation-0.8763 Hypothesized Mean Difference0 df29 t Stat-4.06229 P(T<=t) one-tail0.000169 t Critical one-tail1.699127 P(T<=t) two-tail0.000338 t Critical two-tail2.045231     If we assume that the difference between the number of earthquakes with applying aspects will be 6 then the probability against chance dramatically increases.    t-Test 4: Paired Two Sample for Means  Variable 1Variable