Earthquake lights (EQL)

Earthquake lights fall into the general category of "earthquake precursors", phenomena believed by geophysicists to be symptomatic of the building and discharging stresses that sometimes lead to earthquakes. These phenomena include electrical, visual, infrared and other possible signals, but the field is in general poorly understood and certain reported phenomena are controversial. Reports of unidentified lights have been associated anecdotally with earthquakes for centuries and today many geophysicists (although not all) accept that the association is real. But even now the exact causes and mechanisms remain obscure.

Most observations of EQL are "white to bluish flashes or glows lasting several seconds associated with moderate to large earthquakes" n1Derr, J., 'What are Earthquake Lights? Are they real?', US Geological Survey FAQ.. The largest study of modern EQL sightings dealt with more than 40 reported examples during the 1988-89 quakes in Saguenay, Quebec n2St-Laurent, France, 'The Saguenay, Quebec, Earthquake Lights of November 1988 - January 1989.' Seismological Research Letters, 71, no. 2, p. 160-174, 2000. According to Derr they fell into 6 types of luminous phenomenon: (1) seismic lightning, (2) atmospheric luminous bands, (3) globular incandescent masses, (4) fire tongues, (5) seismic flames, and a newly-recognized category, (6) coronal or point discharges. The 1995 Kobe, Japan earthquake (Mag 6.9) produced 23 sightings within 50km of the epicenter n3Tsukuda, Tameshinge, 'Sizes and some features of luminous sources associated with the 1995 Hyogo-ken Nanbu earthquake', Journal of Physics of the Earth, 45, no.2, p. 73-82, 1997.. They were a white, blue, or orange light all with an upper height of 200 meters and a linear dimension of 1 to 8 km. The types of phosphorescent phenomena were classified as: lightning with zig-zag lines, swelling shield-shaped sources, upward-extending fan-shaped sources, or a belt of lights (including arcshaped sources) n4Derr, J., op.cit..

On the night of July 27-28, 1976, many people in the area of Tangshan, northeast China, reported strange multicoloured lights and loud sounds. Some people reported flashes of light, others saw fireballs flying across the sky. These were followed by loud roaring sounds. Workers at Tangshan airport said the noises were louder than aircraft engines. Pets and farm animals were behaving very abnormally. At 3:42 am on the morning of Jul 28, a magnitude 7.8 quake struck Tangshan and devastated the city, killing over 240,000 people. It was the deadliest earthquake of the century n5Chen Yong, et al., The Great Tangshan Earthquake of 1976: An Anatomy of Disaster (New York: Pergamon Press, 1988) 53..

Mechanisms proposed for EQL include piezoelectricity, heat of friction, sonoluminescence, phosphine gas emissions and more. Problems with the favoured electrical charge migration theories (such as the piezoelectric theory) have centred around getting a sufficient negative electron density to the surface through the rocks. A recent promising theory developed by Friedemann Freund of NASA suggests that EQLs are instead caused by positive hole charge carriers that turn rocks momentarily into p-type semiconductors n6143 Freund, Friedemann T., 'Rocks that Crackle and Sparkle and Glow: Strange Pre-Earthquake Phenomena' Journal of Scientific Exploration, 17, no. 1, p. 37-71,2003. n7144 St-Laurent, France, and Freund, Friedemann T, 'Earthquake Lights and the Stress Activation of Positive Hole Charge Carriers in Rocks'. International Workshop on Seismo Electromagnetics (IWSE), 2005. Coronal or point discharges, such as were observed at Saguenay, are believed to be strong support for this positive hole theory. We consulted Prof. Freund during our investigation of the Channel Islands phenomena.

The Channel area is not especially geologically active. After a notable quake below the Kent coast in 2007 (the strongest in the northeast Channel area since 1950 and the strongest inside the UK for 100 years) the British Geological Survey n8 listed only a dozen quakes (of all magnitudes) within 50km of the new epicentre since 1328AD. It is therefore the more interesting that this 2007 earthquake occurred on April 27, only four days after the UAP sighting over the Channel Islands.

Bearing in mind that the processes that eventuate in earthquakes develop with lead times of perhaps weeks or months, and that the deep geology of the English Channel might connect Alderney and the Kent coast, we considered that a related EQL four days earlier might be possible. With this in mind we looked at the geology and seismicity of the Channel Islands area.

Fig.38. Approximate triangulated locations of UAPs (yellow circles) in relation to the Alderney- Ushant fault system.
Fig.38. Approximate triangulated locations of UAPs (yellow circles) in relation to the Alderney- Ushant    fault system.

We found from Jersey Meteorological Dept/BGS records for the years 1996-2006 n9 that the area is subject to a few earthquakes every year, mostly minor tremors of around Mag 1.0 or less, and mostly with epicentres near Jersey. The geological feature of most interest, however, was a fault or complex of faults known as the Alderney-Ushant fault system, passing down the Channel a few miles N of Alderney and extending in a NE-SW direction to the island of Ushant. And inside the boundaries of Alderney-Ushant system just NW of Alderney lies the deepest seafloor structure in the Channel, the Hurd Deep (Fig.38).

We found ambiguous views about the geological origin of the Deep, recent opinion n10 G. Lericolais, P. Guennoc, J.-P. Auffret, J.-F. Bourillet and S. Berne, 'Detailed survey of the western end of the Hurd Deep (English Channel): new facts for a tectonic origin', Geological Society, London, Special Publications 1996; v. 117; p. 203-215 seeming to favour a tectonic origin (i.e., a faulting origin) over earlier theories involving ancient tidal scouring. But a 1985 theory by Smith n11 Smith, A. J. 1985. 'A catastrophic origin for the paleovalley system of the eastern English Channel', Marine Geology, 64, 65-75. proposing a catastrophic origin in a massive North Sea breakthrough event that created the Channel by flood scouring seems newly relevant in the light of recent sonar studies that have identified characteristic scars of a "megaflood" on the sea floor n12 S. Gupta et al., Nature 448 (2007), pp 342-345.

The geology of the underlying faulted Channel bedrock is that of an ancient rift valley whose contours it appears determined the profile of the overlying chalk landbridge. Consequently there was relatively low ground here, and when the Rhine outflow caused the level of the ice-choked North Sea to rise a river system probably developed through the lower-lying parts of the chalk, beginning an erosion which accelerated until eventually (perhaps aided by an earthquake trigger) the sea broke catastrophically through the land bridge to scour the Channel. It appears that the Hurd Deep is also a pre-existing fault trough that may have been deepened by scouring and then choked by up to 140m of sedimentary infill to its present depth of 75m.

The Kent quake at 0718Z, April 28, 2007, registered Richter Magnitude 4.2, and had an epicentre near Folkestone. It was followed by a series of nine Mag ~1.0 aftershocks until June 5. It occurred at a shallow depth (<5km) below the coast. The BGS moment tensor solution indicated nodal planes of the fault movement oriented either SSW-NNE or SE-NW. No tremors appear to have been reported in or near the Channel Islands, either then or around April 23, but it seemed possible to us that the Channel Islands area might be connected to the Kent coast area by this system of Channel faults and that associated tectonic stress might have built up in the underlying rocks near the Hurd Deep prior to the Folkestone earthquake.

We also found varying opinions about this. We asked the opinion of Dr Roger Musson, British Geological Survey, who replied that he was not aware of EQLs occurring at such a great distance (330km) from a quake hypcentre and doubted the connection because of the small fractured volume in this case (order of 1km3) n13 Email from Roger Musson to Jean-Francois Baure, 25.06.07. However Freund told us that stress can accumulate at 300km from the fracture and that it is very hard to place boundaries on the distribution of underground stress n14 Email from Friedemann Freund to Jean-Francois Baure, 22.05.07. A study of the Saguenay EQL reports indicates 4 observations at ranges>160km, 2 at >200km n15 St-Laurent, France, 'The Saguenay, Quebec, Earthquake Lights of November 1988 - January 1989.' Seismological Research Letters, 71, no. 2, p. 160-174, 2000. Freund regarded a tectonic origin as plausible in terms of his p-hole process, but "whether or not this situation is applicable to the reported sightings . . . of course I don't know" n16 Email from Friedemann Freund to Martin Shough, 25.05.07. John Derr discussed the sighting report with Freund and with Canadian EQL experts St-Laurent and Theriault, and was more optimistic: "I think we agree that the sighting is highly likely to be precursory EQL," he told us, "and may lead to new insights into the mechanism of generation of these lights" n17 Email from John Derr to Jean-Francois Baure, 08.06.07.

We also looked for historical anecdotal evidence that local tectonic conditions might be favourable for EQLs. We could only find one EQ-related story n18 Although other reports of strange lights exist of course. For example, a 3-minute 1975 pilot sighting of "orange lights" moving at ~500ft off the coast of Guernesey was reported in an MoD signal sent to the Defence Intelligence office DI55 on 09 Oct 1975:
A 3 mins 090632A
B 4 separate orange lights
C both pilots of Herald aircraft on approach - land runway 09 Guernesey approach
D naked eyes
E between Guernesey and Jersey moving south or SW
F objects at all times appeared to be lower than the Herald aircraft which was at 2000 feet. They were possibly at 500 ft. The objects, which gave the impression of being in a line were not more than 5 or 6 miles away and tracking down the east coast of Guernesey.
K Guernesey a/f
L Guernesey ATC
M -- ----------- Teddington MX 15 years experienced pilot
O -- -------------
P 090730
Q Yes
Distribution: S4(Air) action; DI55; DI13d; Ops(GE)2; Science 3
Not followed up

The signal was discovered recently by David Clarke during research at the National Archives, Kew (TNA ref DEFE 31/171, file ref D/DI55/1/15/1 Pt 9). We were unable to find a record of any seismic event in the Channel area near this date.
, recorded in 1843.

At 7:30 pm on Dec 20, 1843, a "very remarkable meteor" was seen in the sky of Guernsey, a luminous body "like a clouded moon" moving slowly for 10-15 minutes. Two days later at 3:00 in the afternoon the hitherto bright sky filled with clouds that were strangely coloured with tints of green, red and purple. At 3:50 an earthquake struck, shaking buildings, ringing church bells and causing minor damage, whilst a loud undulating rumble was heard all over the island n19 Falla, G., The Remarkable Guernsey 'Meteor' and Earthquake of 1843, BUFORA, Aug 2007..

One issue of possible relevance to the EQL theory is that our UAPs would apparently have been located over the ocean. In terms of some EQL theories this might be problematic. For example, friction heating or piezoelectric corona discharge would seem to require a rock-air interface to generate luminous bodies in the atmosphere. In any electron charge migration theory there is already a difficulty with the transport of sufficient charge to the surface of exposed rocks. Conceivably sparking stress fractures in rocks under the sea bed could emit radio waves, but these would be blocked or drastically attenuated n20 The extremely low (order of 10Hz) electromagnetic frequencies that are not quite effectively blocked by water carry very little energy and to get enough of it through the conductive water blanket seems difficult to say the least. Electrical breakdown causing corona discharge glows needs hundreds of thousands of V/cm at rock edges and corners. Such gradients seem unlikely to arise from million-metre ELF waves weakly reaching the atmosphere over a huge surface area of ocean. by the overlying volume of water.

The recent p-hole theory of Freund has attracted attention largely because it offers a promising alternative mechanism of charge migration in rock n21 The "holes" are defects in the negative electron structure of the crystal which can be treated as positive charges.. This mechanism is, as he explained to us,"relatively trivial" n22 Email to Martin Shough from Friedemann Freund, 25.05.2007 "If positive hole charge carriers are activated in an otherwise insulating dielectric medium, they throng to the surface. They build up a surface potential from which we can calculate the electric field. If the concentration of positive holes is sufficiently high, the field can easily exceed 500,000 V/cm, even on a flat surface, a value that invites speculations that air molecules should become ionized (field-ionization, loosing an electron to the surface and become airborne as positive ions). About ten years later [c.1994] I was able to experimentally show that this effect takes place and leads to luminous effects emerging from edges and corners of rocks due to corona discharges in the air." and probably could not circumvent the problems of more conventional corona discharge theories in the case of an EQL over water. However there is a second and more intriguing possibility "based on observations which suggest that the wavefunction associated with positive hole charge carriers is not localized on any one oxygen anion but spreads out over many oxygen anion positions, maybe as many as several hundred. If the number density of pholes reaches a threshold . . . the wavefunctions will begin to overlap and the system will undergo a transition from a weakly doped semiconductor state to a highly doped (quasi-metallic) plasma state. I have 'seen' this transition in a number of experiments . . . ." n23 Ibid.

The relevance of this mechanism to the Channel Islands UAPs is the possibility that a rapid build-up of tectonic stress in a small source-volume of the Earth's crust may lead to such a plasma state which becomes unstable and "bursts" outward through the surface. Freund speculates that some EQLs are such p-hole plasmas, and in answer to our questions he opined that "a shallow body of water would not be an impediment (I think) to the outburst of the plasma bubble". However he cautions that it is only "a plausible physical process (untested yet) that could explain part of the story" and there is as yet neither a theoretical nor an experimental basis for saying that large, stable, luminous shapes at altitude can be caused by such plasma outbursts - if indeed they occur n24Ibid..

This theory raised some further questions in our minds. An unstable plasma propagating rapidly outward through the rock from the source volume would, one imagines, prefer to travel as an expanding wavefront of p-type plasmons radiating from this "hypcenter" to the "epicenter" (of the wave, not the quake), leading to a diminishing surface potential spreading outward from the epicenter. In this case the energy density at any region of the surface must be a tiny fraction of that at the source, in a short pulse, with a tendency to dissipate further in the air. We wondered how the stability and somewhat high energy density implied by the idea of a "plasma bubble" EQL might arise from the unstable dissipative process of a p-hole wave generated deep underground. What secondary mechanism(s) might be at work? Is it possible that the process recyles sufficiently fast to keep the EQL pumped, and if so is it possible to guess at the frequency? And/or is some sort of focusing mechanism possible owing to the underground configuration of charge carrying rocks?

Prof. Freund replied that the answers to such questions are largely in the realm of speculation at the moment, but offered this very interesting suggestion n25Crediting the original idea to Robert Theriaux in discussion with himself, France St-Laurent and John Derr.:

I know from my lab experiments that pholes can (i) be activated in every igneous and high-grade metamorphic rock that I have had a chance to study in at least some detail, and (ii) propagate through such rocks. I also know that pholes cannot be activated in regular technical glass (window glass), but they can propagate through glass. However, pholes can neither be activated in marble nor propagate through marble. The reason (for which I have rather convincing crystallographic arguments) is that the structure of Ca-carbonate, the mineral calcite in marble, does not support the formation of the parent defects out of which pholes can be activated. What is true for Ca-carbonate must be true for other carbonates as well, in particular for Mg-carbonate magnesite, which is often (I have been told) associated with magmatic lamprophyres. Therefore the idea came up that the presence of lamprophyric dykes in the deep underground could constrain or 'focus' the flow of pholes in such a way as to more easily achieve the critical density necessary for entering into a plasma state. It would certainly be a worthwhile task to inspect geological maps of those places where EQLs have been observed whether there are any magmatic or sedimentary carbonates in the neighborhood n26 Email to Martin Shough from Friedemann Freund, 26.05.2007 .

We found out that lamprophyres are indeed present throughout the geologically-connected area (northern edge of the Massif Armoricain) containing north Brittany, Contentin, and the Channel Islands. All of the islands (with the exception of little Sark) contain lamprophyres n27C. J. D. Adams, 'Geochronology of the Channel Islands and adjacent French mainland', Journal of the Geological Society 1976; v. 132; p. 233-250; esp. Table 1, P.235. These are relatively young rocks dating to the period of "Variscan plutonism" about 280-345MY ago and presumably overly many of the older igneous and ancient basement rocks, but we have not found a precise map of the distributions.

It is possible that at least one of the triangulated UAP locations could have been close to the Casquets Rock, a small islet west of Alderney bearing the Casquets Lighthouse. What, we wondered, might be the effect of a p-hole wave reaching a rock/ocean boundary which is penetrated by an isolated rock/air discharge point in the form of an island - or nearly penetrated by a submarine seamount? How does the wave of positive holes behave when reaching the water, i.e. what possible ion transfer effects are there involving dissolved salts at the interface, and what happens a) in the deep region and b) at the rising seafloor around the island? Is it possible that the discharge-to-air route is a preferred minimum energy path and that the seabed potential is shunted towards the island, so that the island acts to "collect" and concentrate the current from a wide area?

We were unable to find clear answers to such questions, although it seems likely that "the dielectric properties of the medium will play an enormous, if not controlling role in the 'bursting' of the bubble" n28 Email to Martin Shough from Friedemann Freund, 26.05.07: "Depending upon the dielectric constants of the rocks in the deep underground, a propagating wavefront can be expected to be diffracted and/or focused in a way that I don't even begin to understand. The fact that liquid water has a high dielectric constant of 81 (versus epsilon of rocks being more in the 6-10 range), will certainly also have to enter into consideration.". The p-hole theory is not yet well enough developed to predict where and in what form EQLs might be observed, even if the local geomorphology and seismic conditions were known in detail. But it is an intriguing possibility, and the proximity to the Alderney-Ushant fault, the probable presence of lamprophyre dykes in the vicinity, and the significant Kent earthquake just a few days later, are all at least circumstantial evidence. It also seems possible that an aerial plasma of this type would have a small radar cross-section at the centimetre wavelengths of ATC and weather radars (see Section 4).

It is worth listing some of the negative indications: The persistence of UAP#1 for at least 12min; stability of form for 12min; "very sharply defined" binocular outline; extreme brightness in strong daylight (typical EQLs are not perceived as brilliant even at night, but as aurora-like glows); anisotropic luminous output (pale "yellow/beige" from the south; "very brilliant" from the north n29Note that Capt Patterson was initially unable to see anything, even when some 6nmi closer to the triangulated location of UAP#1. At this time the depression angle from the Jetstream to UAP#1 would have been ~5.0°, reducing at the sighting time to about 1.1° (UAP #2 would initially have been almost directly below the aircraft, latterly falling almost directly behind.). It is not known whether the visibility of some type of EQL emission might be sensitive to the elevation angle (and/or bearing angle) of viewing. Capt Bowyer's viewing angle remained much less than 5.0° (visually estimated max. -2.0° depression; calculated max. [Sect.3, Fig.8], 1.1° below geometrical horizon, reducing through 0° to about +0.5°).); yellowish or yellow-orange colour (blue or purplish colours - the colours of corona discharge - seem statistically favoured); apparent immobility for the whole observation period; all of the above duplicated in an identical (apart from angular size) UAP#2 for at least 6min; the "graphite grey" bands at the same position on each object; distance from dry land (in at least one case) by several nmi and 1500-2000ft of altitude.

We are aware of no well-authenticated observation of EQL that reproduces even a few of these features. But given that the nature and mechanism of EQL is at best obscure it does not seem possible to do more than define the class of EQL phenomenologically, rather than physically. This being the case one does not know a priori whether a given observation should be excluded from the class or included to redefine the class. It is certain that reports of many types of aerial phenomena exist that find no comfortable home in any classification today but may do so in future.

On this basis, and with a view to the striking coincidence of the Kent earthquake as well as the local geology, we cannot rule out novel EQL-related effects in this case.

Plausibility (0-5): 3