Art. LVI.—Brief Notes on the Theory of New Zealand Earthquakes.
[Read before the]Wellington Philosophical Society, 4th October, 1905.]
Plates LIII and LIV.
From time to time papers have appeared in the Transactions of the New Zealand Institute and of the Australasian Association for the Advancement of Science dealing with individual earthquakes. Captain Hutton's monograph on the Amuri earthquake of the 1st September, 1888, marked a new epoch in seismological science in New Zealand, as it treated the facts of that disturbance in accordance with the ideas of modern seismology. I have published during the last sixteen or seventeen years several other papers in which I have endeavoured to follow the same general lines in regard to all the important shocks that have occurred. Although, perhaps, we cannot yet with any certainty elaborate a complete theory of the earth-movements in the New Zealand region, nevertheless it appears to me that the time has arrived when we may attempt to coordinate the facts in our possession by some general explanation of them so far as the earthquakes enable us to do so.
The most accurate observations made in the colony have been those afforded by the two Milne horizontal pendulums installed at Christchurch and Wellington respectively. Some of the inferences to be obtained from their records, or seismograms, may be most readily understood by means of a few general remarks; as illustrations I take the copies of four seismograms on the New Zealand instruments (Plate LIII, figs. 1–4).
Fig. 1 on Plate LIII is a copy of the Christchurch record of the great Guatemala earthquake of the 19th April, 1902. The time is Greenwich mean civil time.
Fig. 2 shows a severe earthquake, on the 20th November, 1902, from an origin probably near 21° S. lat., 172° W. long.— that is, north-east of Tonga (Wellington record, G.M.C.T.).
Fig. 3 is the Wellington record of the earthquake at Cheviot, New Zealand, on the 16th November,1901, beginning at 7·47 a.m., New Zealand mean time.
Fig. 4 is a copy of the seismogram taken at Wellington of the East Coast earthquake of the 9th August, 1904 (G.M.C.T.).
The small letters a, b-h, above fig.1 mark the beginning of the eight phases into which Professor Omori divides the
waves. The most important of these are: a, the beginning of the preliminary tremors; e, the chief of the large waves; h, the last marked phase of the series. The position of the corresponding waves on fig. 2 is marked by the letters a, e, h. In figs. 3 and 4 the downstroke on the margin shows the beginning of the preliminary tremors (a); the black line right across the seismogram gives the position of c, d, and e; the last phase (h) is shown in fig 3 by another downstroke, but cannot be distinguished in fig.4.*
As I have stated in a former paper, the times of arrival of these waves (a) at the various seismological stations of the world appear to show that they travel round the earth, and not through its central portion—that is along arcs parallel to the surface, and not along chords. That is certainly true of the other waves (b-h).
The following table shows the transit-velocities of the chief waves of the four earthquakes just referred to, in kilometers per second:—
[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]
|Earthquake.||a Waves.||e Waves.||h Waves.|
The high transit-speed of the a waves or preliminary tremors can be accounted for only on the assumptions—(1) that the vibrations originated in rocks under a maximum strain (it will be remarked that the speed is greatest for the earthquake of greatest intensity); (2) that their path through the earth's crust was the path of maximum velocity, and therefore through the rocks of highest rigidity and elasticity.
The velocities of the e waves and h waves for various large earthquakes seem to vary very little, the average being 3·3 and 2·2 kilometers per second respectively. Now, from theory based upon the experimental determinations of the rigidity, elasticity, and density of various rocks, the speed of large normal or longitudinal vibrations through such rocks as hard granite is estimated to be between 3·1 km. and 3·95 km. per second, and the speed of transverse waves two-thirds of that of normal waves proceeding from the same origin at the same time and along the same path.
[Footnote] * The fact that the interval between the arrival of the preliminary tremors and of the normal waves is very small when the origin is near, and increases with the distance from the origin, is explained on the assumption that they start at or about the same time from the origin, and gradually become separated owing to the difference in their transit-speeds.
It is therefore highly probable that e and h on the seismograms (of which those given are fair specimens) represent the chief normal and transverse waves, due respectively to elasticity of volume and to elasticity of form. Now, if the rocks were subjected to great pressure, as must be the case at a few miles below the surface, the strain would be at a maximum just before yielding, and the consequent vibrations (a) would be small and rapid; as the rocks yielded to the strain and were compressed, larger and slower vibrations (e) would be produced; distortion of the rocks would also generally take place, giving rise to transverse waves (h). The conditions would be satisfied if, under unequal vertical or lateral pressures in two adjoining portions of the earth's crust, there occurred bending or folding of the strata, accompanied or followed by fracture of the rocks such as give rise to faults; there would be earthquakes, in fact, whenever any sudden adjustment took place, whether such adjustment were rapid tilting, the formation of a fracture, the rapid sliding of one rock-face over another, or simply the crushing of rocks under great increase of pressure, or, what is most probable, several of these causes operating together.
Have we any evidence that these conditions have been satisfied in the case of New Zealand earthquakes? The great earthquake of 1855 affords evidence that they have been satisfied in at least one instance. The origin—that is, the moving portion of the earth's crust—was at least as large as is indicated by the oval drawn on the map(Plate LIV). The evidence is very clear that on the north-eastern side of this area the elevation was greatest; that it diminished towards the middle; that there was neither elevation nor depression in Porirua Harbour; and that on the south-west side of the focal area there was a depression of at least 5 ft. This tilt, or folding, as we may fairly call it, was also accompanied by fracture of the rocks, showing itself by surface rifts that ran for many miles north-east and south-west.
Again, reference to the map will show that the line joining Wellington and the epifocal area of the Cheviot earthquake of November, 1901, is nearly parallel to the general axis of New Zealand; a great rift (called by Mr. A. McKay, Government Geologist, “the Clarence fault”) runs nearly in the same direction, and it is quite probable that it indicates the existence of a deep fault. The significance of this will be seen if we turn to the Wellington seismogram of the Cheviot earthquake (fig. 3); it will be observed that the mean position of the central or zero line after the shock is nearer the lower edge of the paper—that is, nearer the west—than it was before the shock: this shows that the surface of the earth on which the column rests was titled through an angle of about 1·3 seconds towards the west or north-
west.* This titling took place suddenly; but it is interesting to note that for some months before the earthquake a gradual lowering of the level on the west side had been taking place, and that a similar gradual movement has been going on up to the present time (September, 1905). The 1901 earthquake therefore gives us an example of tilting, or folding, in which the original fracture of the rocks did not apparently extend to the surface.
It is well known that strata originally more or less horizontal have been folded by forces acting on the earth's crust, the folds often appearing at the surface so as to form wrinkles or mountain ranges and valleys. Several possible causes of this folding may be assigned. Two of the most important are—(1) the gradual cooling of the earth; (2) the loading of the ocean-bed through denudation of the land-surface.
(1.) As to the cooling of the earth: The temperature of the rocks near the surface, being the temperature of space modified by the heat from the sun's rays, is practically constant; so that those rocks will not contract. All other layers of the crust will cool and contract at rates varying with the depth. The volume of the layers in which cooling is greatest will, after contraction, be too small to fill the space into which the layers fall, and they will therefore be pressed out by the weight of the rocks above them. But the upper strata, which cool less quickly, will be too large to fill the space into which they fall, and their surface will accordingly be crumpled. The effect will be to produce a series of elevations and troughs (anticlinal and synclinal folds). The folding or crumpling is determined partly by previously existing lines of weakness, partly by inequality of vertical pressures due to differences of texture and density in the upper layers. Unequal vertical pressures on adjoining portions of the earth's crust will cause unequal lateral pressures, and a tendency for the rocks to “creep” or move horizontally in order to repack themselves in a more stable condition. This lateral thrust, again, may produce fractures, reversed faults, and elevation of those strata which are less dense or subject to a smaller vertical pressure. Obviously, earthquakes will occur whenever any of these movements are sudden in character.
(2.) “Loading”: Another cause of folding is the transference of material from the land-surface to the bottom of the sea by the agency of rivers: the pressure on the strata underlying the land will be thereby relieved, and the pressure on the floor of the ocean will be increased. There will thus be a
[Footnote] * In the 1855 earthquake the angle of tilt could not have been less than 4 seconds, and may have been as much as 10 seconds.
folding of the strata near the junction of the land-surface and the ocean-bed, which may continue until fracture takes place, the stratå on the land side moving up, and those on the ocean side moving down. The displacement of the strata will form a normal fault. After the fracture the two faces of the strata will continue to slip and to slide one over the other so as to increase the amount of displacement. The slipping may go on very gradually, but from time to time sudden slips will probably occur, and these will produce earthquakes. All the recent earthquakes in New Zealand have been followed by other shocks, showing that the slipping of the rocks has continued for some after the principal shock. Again, as in the case of folding due to cooling, the differences of vertical pressure will induce lateral thrusts, tending to cause “reversed faults”; but in this case the tendency will be most marked at a considerable depth below the surface, where the pressures are greatest. Subsequent action may raise the deeper rocks, and reveal faulting that has occurred ages ago.
The facts in regard to the folding of the crust are not, of course, so simple as they have been stated here; but, generally speaking all the movements resulting from unequal vertical and lateral pressures between the rocks may be summed up in the term “repacking”
The evidence already given is perhaps not sufficient to establish completely the theory that earthquakes generally are connected with fault-movements and similar processes of repacking of the strata, but it makes that theory highly probable. It must be remembered that the value of such evidence necessarily depends upon its cumulative weight, and it would be too tedious to give a large mass of evidence of this kind here. Moreover, our knowledge of the position of the geological faults in New Zealand is too limited as yet to prove the theory from New Zealand examples. Dr. Charles Davison, an able British seismologist, has shown in a very careful series of investigations the connection between many British and European earthquakes and known lines of fault.
A study of the map of the seismic origins in the New Zealand region (Plate LIV) will strengthen the evidence already given in favour of the theory. Before considering these, it may be as well to name the nearest known origin outside the New Zealand region.
On the 27th January, 1892, an earthquake of the intensity VII-VIII (Rossi-Forel scale)—probably × near the origin—was felt over almost the whole of Tasmania, in Victoria as far west as Melbourne, and in the south-east part of New South Wales, but not in New Zealand. The epicentrum was
about 353 miles east of Launceston; it lies on the western slope of the Thompson Basin, the great trough in the Tasman Sea. The same region contains the origins of at least three well-marked Tasmanian earthquakes—namely, those of 13th July and 19th September, 1884, and 13th May, 1885.
From April, 1883, to December, 1886, 2,540 shocks were recorded by Captain Shortt, R.N., of Hobart, and his assistants, nearly all being very slight. Probably the earthquakes just named above were due to the principal movements, and the numerous smaller after-shocks or tremors indicated the slight adjustments of the Tasmanian land-mass. We have yet to discover what connection, if any, there is between these and other movements in or near Australia and Oceania, and those of New Zealand. Most of the evidence available seems to point to the hypothesis of a general elevation of the floor of the Western Pacific; but the evidence so far is very meagre and disconnected.
In the case of the New Zealand earthquakes, on the other hand, I think that a careful study of the map and of the facts mentioned in the text will immediately suggest the theory that our earthquakes are incidents in the history of folding and similar movements that have been going on for ages, the axes of the folds being parallel to the general axis of the country.
The origins of the New Zealand seismic region will be seen to arrange themselves in groups as follows:—
Group I.—Earthquakes felt most strongly on south-east coast of North Island; epicentra form a strip 180 miles from the coast, parallel to the axis of New Zealand, and to axis of folding of older Cainozoic rocks in Hawke's Bay. Chief shocks: 17th August, 1868; 7th March, 1890; 23rd and 29th July, 1904; 9th August, 1904 (intensity IX on R.-F. scale); 8th September, 1904; prob. 23rd February, 1863 (IX, R.-F.); &c.
According to Captain F. W. Hutton, F.R.S., the geological evidence shows that New Zealand rose considerably in the older Pliocene period, and was then probably joined to the Chatham Islands. At a later period subsidence occurred, followed again by elevation in the Pleistocene period, with oscillations of level since. The seismic origins of this group are at the foot of a sloping submarine plateau, about two hundred miles wide (marked B on the map), which culminates to the east-south-east in the Chatham Islands. This elevation is separated from the New Zealand coast by a trough from 1,000 to 2,000 fathoms in depth, which is widest and deepest at A A—that is between these origins and the mainland.
Group II.— (a.) South-east of Otago Peninsula. Shocks: 20th November, 1872, &c.
(b.) A strip south-east of Oamaru. Shocks: February, 1876; April. 1876; &c.
(c.) Many short and jerky, but generally harmless, quakes felt in Caristchurch, Banks Peninsula, and mid Canterbury. Chief shocks: 31st August, 1870; 27th December, 1888 (VII,R.-F.); &c. Focus of 1888 shock, sixteen miles long, from west-south-west to east-north-east, twenty four to twenty-five miles below surface, being deepest ascertained origin in New Zealand region.
These origins form a line parallel to the general axis of the land. It is quite possible that the loading of the sea-floor by the detritus brought down by the rivers of Canterbury is a contributing cause of the earthquakes of this group.
Group III.—Wellington earthquakes of January, 1855, and Cheviot earthquakes of November, 1901.
Remark has already been made as to a possible relation between these origins. The great earthquakes of October, 1848, probably came from the same region as those of January, 1855. The chief shocks of both series did extensive damage to property, and caused the formation of large rifts in the earth's surface; they are the only seismic disturbances since the settlement of the colony that can be assigned to degree × on the Rossi-Forel scale.
Group IV.—(a.) Region about twenty-five to thirty miles in length, and, say, ten miles or less in width, running nearly north-north-east from middle of Lake Sumner, about twenty miles below the surface, whence proceed most of the severer shocks felt from Christchurch to the Amuri, and a large number of minor shocks. Chief earthquakes: 1st February, 1868; 27th August to 1st September, 1871; 14th September and 21st October, 1878; 11th April, 1884; 5th December, 1881 (VIII, R.-F.), when christchurch Cathedral spire was slightly injured; Ist September, 1888 (IX,R.-F.), when upper part of same spire fell, and still more severe damage was done in the Amuri district.
(b.) A small, shallow origin not more than five to ten miles below the surface, a few miles south of Nelson. Earthquake: 12th February, 1893 (VIII to IX, R.-F.); chimneys thrown down and buildings injured.
(c.) Origin in Cook Strait, north-north-east of Stephen Island, about ten miles wide, and apparently traceable with few interruptions nearly to mouth of Wanganui River; depth, fifteen miles or more. More than half the earthquakes recorded in New Zealand belong to this region; earthquake of 8th December, 1897 (VIII to IX, R.-F.), and other severer ones come from south-south-west end.
(d.) An origin near Mount Tarawera, with a large number of moderate or slight shocks, most, but not all, volcanic and local in character—eg., those of September, 1866, and those of June, 1886, which accompanied and followed the well-known eruption of Mount Tarawera.
These origins of Group IV, (a), (b), (c), (d), are nearly in a straight line on the map; on or near the same line are the origins of earthquakes felt in the Southern Lake district (15th December, 1883, &c.), the volcanoes Ruapehu, Ngauruhoe, Tongariro, Tarawera, and White Island. It is evident that this line, which, like the rest, is parallel or nearly so to the general axis, is a line of weakness or of unstable equilibrium. Hence the adjusting movements that have caused earthquakes
may have, from time to time, relieved the pressure of the rocks that restrained overheated steam and other volcanic agents from bursting out, and so may have led to volcanic eruptions; just as the series of earthquakes in Guatemala and in the Caribbean Sea in April and May, 1902, were the signs of movements in the great folds of that part of the earth's crust, in the course of which, the pressure in the Antillean Ridge being relieved, the volcanic forces below Mount Pelée in Martinique, and Mount Souffrière in St. Vincent, caused the disastrous eruptions of that year.
Group V.—Off the coast near Raglan and Kawhia. Chief shock: 24th June, 1891 (VII-VIII, R.-F.). The line joining this origin to that of the earthquake of 1st February, 1882, is parallel to the other lines of origins (Groups I to IV); but we have no data to establish any connection between them.