The Evolution of the Island Arc of Japan and the Formation of Granites in the Circum-Pacific Belt

JOURNALOF GEOMAGNETISMAND GEOELECTRICITY VOL. 23, No. 3, 4, 1971

Naoto KAWAI, Tadashi NAKAJIMA and Kimio HIROOKa+ Department of Material Physics, Faculty of Engineering Science, Osaka University, Osaka, Japan (Received April 27, 1971)

Abstract

From a palaeomagnetic study and radiometric investigation of Cretaceous intrusive rocks it was recently suggested that the Palaeozoic basin formed in the northeastern Japan was severely deformed at the end of Mesozoic era. This resulted in a narrowing and shortening of the entire length of the Japanese islands. The northeastern block moved southwestward by approximately 200km and southwestern northeastward by 150km, whereas the middle block remained relatively unmoved but was compressed between the two blocks. As the compressional forces increased, first the Palaeozoic sediments of the central block were uplifted, and subsequently the "median line" was formed. Along the latter line, the northern half of the southern block moved eastward relative to the southern half. A strong uniaxial stress superimposed upon a hydrostatic one occurred associated with the relative movement. The sediments as the results recrystallized to form the three major metamorphic belts. During and after these movements, the compressed zone in the middle of the island was pushed to the east to form a great warp in the island. Such a simple model of a bend accompanied by a tension crack as predicted by Kawai, Ito and Kume several years ago was reconsidered. The contraction of the basin was finally related to the eastward drift of Asian continent upon the Creta- ceous Pacific sea floor. The drift of the continents around the Pacific resulted in the narrowing of the ocean as well as the shortening of the coastal line. That of the Japanese islands is a part of the shortening that occurred around the Pacific zone. We suggest that the heat of friction was accumulated at the deep-seismic plane (Benioff zone) at which the down and west going ocean floor and the up-thrusting mantle have been conflicting. The mantle nearby the deep-seismic plane was, therefore, warmed up until it partially melted. An acidic and migmatitic magma initiated at the friction interface was intruded into the pre-existing fracture zone occupying the front. Production of frictional heat was, we believe, succeeded even in the Tertiary period when effective drift was no longer evident but the retardation of the motion of the continent was much stronger. Heat required to cause the Tertiary volcanic activity might be due to this friction.

1. Palaeomagnetic Results and Radiometric Dating of Igneous Rocks Sedimentary and igneous rocks ranging in age from Triassic to the Recent were collected from various localities in Japan. The directions of natural remanent magnetism of these rocks were measured. The results obtained show that the directions found

+ Present Address: Faculty of Education , Fukui University

267 268 N. KAWAI, T. NAKAJIMA and K. HIROOKA were nearly parallel or antiparallel in the late Tertiary rocks regardless of the localities from which the rock specimens were collected, although there exist local differences in the direction of the remanent magnetism in pre-Tertiary rocks. Rocks from the northeastern Japan are all found to have northwesterly directions. In contrast those from the southwestern Japan have northeasterly directions. In order to reconcile the discrepancy, Kawai et al. (1961) proposed an occurrence of a land mass motion with which it is possible to bend a comparatively straight pre- Cretaceous land to form the present bow-shaped island arc. They assumed an anti-

clockwise rotation of the northern Japan by about 50° relative to the southern Japan. This great bend was estimated to have been made at the end of Mesozoic era. Recently an extensive survey was carried out with the aim of determining the geological age of a number of igneous rocks intruded during the time of the deformation. Thus, it became possible not only to know when the bending was initiated and came to the end but also to clarify the relation between the plastic deformation and the igneous activity. Granites and granodiorites whose ages were already determined by Kawano and Ueda (1964, 1966) were collected from the northeastern Japan. After AC. and thermal cleaning the remanent magnetism was measured under an astatic magnetometer. In Table 1 are tabulated the results of the measurements, localities from which the rocks were collected, and the geologic ages determined. In Fig. 1 are shown the distribution of magnetic declinations and the localities of the collected rocks. The intrusive rocks in the northeastern Japan, according to the K-Ar dating of Kawano's results (1967), can be classified into the following three major groups; KITAKAMI MASSIF of the ages ranging from 120m.y. to 110m.y., ABUKUMA MASSIF of the ages ranging from 100m.y. to 90m.y. and ASAHI-TIDE MASSIF of the ages ranging from 70m.y. to 55m.y. The granitic rocks of Kitakami Massif were collected from the following localities, Sugo, Akabane, Ochiai, Hosozawa, Hitokabe, Senmaya, Tono, Yoshihama and Otomo in Iwate Pref. and also from Kinkazan in Miyagi Pref. The rocks of Abukuma Massif were collected from Abukuma, Shiobite, Tsushima and Hirusone in Fukushima Pref. On the other hand the rocks of Asahi-Tide Massif were collected from Nogawa Dam, Yakuwa Dam, Suganodai, Sekigawa and Ootorigawa in Yamagata Pref., respectively. In the Kitakami Massif, the oldest rock (121m.y.)

has the most westerly declination of 61°W, and the youngest (106m.y.) has declination

of 35°W. In Abukuma Massif, declinations of the oldest (101m.y.) and youngest (92

m.y.) rocks are 33°W and 7°E, respectively. In the above two districtss the rocks having intermediate age have intermediate declination as shown in the Schmidt's equal area projection (Fig. 2). On the other hand, Asahi-Tide intrusive rocks have almost easterly declinations. In Figs. 3 and 4 are shown the changes of magnetic declinations and incli- nations, respectively, with geologic age in the northeastern Japan. In Fig. 3, it can be seen that the change of declination occurred twice, once in a range extending from 120 to 115m.y. and then in a range from 100 to 85m.y. The Evolution of the Island Arc of Japan and the Formation of Granites in the Circum-Pacific Belt 269

Fig. 1 Directions of remanent magnetization of Cretaceous rocks in Japan A detailed palaeomagnetic research of the southwestern Japan had already been carried out by Kawai et al. (1961) and also followed by Sasajima et al. (1966, 1968). The results are summarized in Table 2. The localities from which the rocks were collected are all in the regions west of the geotectonic structure, the "Fossa Magna", and north of the "median line" as shown in Figs. 1 and 8. In Figs. 5 and 6 are shown the changes of declination and inclination, respectively, with geologic time in the southwestern Japan. As shown in Fig. 5, no abrupt change of declination occurred in the southwestern Japan.

The declination decreased gradually from 60°E to 40°E in a range from the Cretaceous to the Paleogene, and this gradual change has continued to the present. It seems 270 N. KAWAI, T. NAKAJIMA and K. HIROOKA

Fig. 2 Direction of N.R.M. of granitic rocks from the northeastern Japan (plotted on a Schmidt's equal area projection) AM, SB, YM, etc, are the abbreviations of the localities of the sampling sites tabulated in Table 1, and the numbers annexed to them in the brackets are the determined ages in million years.

Fig. 3 Change in declination of N.R.M, with geologic time in the northeastern Japan. The Evolution of the Island Acr of Japan and the Formation of Granites in the Circum-Pacific Belt 271

Fig. 4 Change in inclination of N.R.M, with geologic time in the northeastern Japan

Fig. 5 Change in declination of N.R.M, with geologic time in the southwestern Japan (Sasajima et al. 1966 and 1968) 272 N. KAWAI, T. NAKAJIMA and K. HIROOKA

Fig. 6 Change in inclination of N.R.M, with geologic time in the southwestern Japan (Sasajima et al. 1966 and 1968)

Fig. 7 Change in declination in Japan (summarized from Fig. 3 and Fig. 5) The Evolution of the Island Arc of Japan and the Formation of Granites in the Circum-Pacific Belt 273 reasonable to consider that this continuous curve is the result of the so-called "polar wandering". In Fig. 7 are summarized the changes of declination in both the northeastern and southwestern Japan. Mean declinations of the remanent vectors in Kitakami district differ by 85° and that in Abukuma district differ by 50° from the mean declination of the southwestern Japan. On the other hand, the mean declination of the rocks in Asahi- Iide does not show a large difference from the southwestern Japan. In contrast to the remarkable differences in declination in the above-mentioned districts, one cannot find any large difference in the inclination of the magnetic vectors over the entire island as shown in Figs. 4 and 6. In order to explain the facts mentioned above the present authors (1969) proposed an occurrence of an extreme plastic deformation in which are involved not only a great bend in the middle of the island but also other subsidiary bends at other places. At the present stage of our study in which the sequence of the Cretaceous intrusion was well clarified by means of radiometric age determination, it is possible to discuss the land mass deformation much more in detail as will be shown in the following para- graphs. Before going further into the land mass deformation it seems important to make clear the relation between the time when the rocks were magnetized in the direction of the past geomagnetic field and the time when potassium-argon clock was started to count the passage of geologic time. According to Neel's theory (1955) on the thermal fluctuation after-effect, effective fixing of magnetic domains occurs at the so-called blocking temperature of the rock- forming ferromagnetic mineral. This temperature, being slightly lower than the Curie

temperature of ferromagnetic material, is found in a range from 400° to 500℃ in actual intrusive rocks. One can, therefore, assume that the past geomagnetic field imprinted upon the rocks was fossilized therein when the rocks, after the solidification, was cooled down and reached the above-mentioned temperature range. It is also reasonable to suppose that in this temperature range the diffusion rate of the atoms in the crystalline lattice of the minerals had become so slow that the con- tained A40, even immediately after the disintegration from the K40, could not escape out of the minerals. Entire A40 atoms thus produced since that time have been well trap- ped and retained in the minerals, so long as the rocks were not subsequently reheated. The length of the time since the acquisition of the magnetization to the present, and that represented by the determined age of the rocks approximately coincide with each other in these particular igneous rocks. Next, with respect to the deformation, Kawai et al. (1969) proposed recently that auxiliary bends occurred at several places other than the middle of the island. Recent results show that the Kitakami block rotated at the beginning of the deformation. The

amount of this initial rotation was about 30° and the sense was anticlockwise relative 274 N. KAWAI, T, NAKAJIMA and K. HIROOKA

to the rest of the island mass. The date of this rotation lay in the range from 121 my. to 114m.y. In the period ranging from 114m.y. to 100m.y., however, no bending occurred although the active intrusion of granites and granodiorites continued in both the north and south of Japan. Then the second great bend took place in the middle of the island. The Abukuma

block rotated by 55° in the same anticlockwise sense relative to the southern Japan in the interval from 100m.y, to 92m.y. The Kitakami block which had already rotated prior to this movement continued with the Abukuma block. The total amount of the rotation in the northernmost block

slightly exceeded 85° as observed from the southern Japan. The bending of the northeastern Japan came to an end at about 85 m.y. Figure 11 shows how the initially almost parallel magnetic declination changed as time went on. The maximum rate of

the bend of the Kitakami and that of the Abukuma blocks can be estimated to be 10°/ m.y. and 6°/m.y. in the two places, respectively. It is interesting to point out that in the northeastern Japan the oldest intrusive Cre- taceous rock is found in the eastern part and the youngest in the western part (east of the Fossa Magna) whereas rocks having intermediate age in the middle part as shown in Fig. 8. Assuming that all of the older intrusive rocks have not been eroded and re- moved from only the western part of the northeastern Japan, one may safely as well as reasonably conclude that the igneous activity may have initiated in the eastern part of the northeastern Japan and migrated to the west. The deformation of the land mass has apparently followed this migration. As a result the Kitakami block was firstly rotated and then the rest of the land blocks was rotated. This conclusion, however, may not hold good about Miocene and latter time, because the youngest rocks just east of Fossa Magna are reasonably regarded as a part of plutonic rocks that are distributed all along the Kuril-Northeast Japan-Izu island arc system which seen cutting obliquely the chronological pattern in northeast Japan. Although one must look for a proper reason why the intrusion took place prior to rotation, the post-intrusion deformation was a fortunate geologic event. If it had been a pre-intrusion movement, the magnetic vectors could not indicate any sign of rotations that had occurred before. In the southwestern Japan, the migration of igneous activity does not show any such regularity as that in the northeastern Japan. It seems likely that the Cretaceous intrusions took place nearly simultaneously at several places in this district. In the southwestern Japan one cannot find any intrusive rocks older than 100m.y. except in the Hida district, whereas in the northeastern Japan many such older intrusive rocks have been found. However, most of the younger intrusive rocks are found in the southwestern Japan. In Fig. 8 is shown the distribution of the above-mentioned intrusive rocks. The Evolutionof the Island Arc of Japan and the Formationof Granitesin the Circum-PacificBelt 275

Fig, 8 Distribution of the granitic rocks (Kawano and Ueda 1967)

2. Shortening of Island Length

In the previous section the block rotation of Kitakami and that of Abukuma district were described. At the external zone of bending many regions should have been exposed to strong tensile forces, if the deformation had proceeded as in the case of the bending of a single bar. The deformation was so intense (as already described) and so restricted in area that plastic deformation of the land mass should have easily come to its yield point and be accompanied by subsequent ruptures. Breaks such as shown in Fig. 9, therefore, should be seen at, at least, two places along the Pacific coast, for example at Sendai bay area or in the region of Kanto uplift. However, neither the tensile gap nor any ruptures zone can be seen there. What 276 N. KAWAI, T. NAKAJIMA and K. HIROOKA

is more puzzling is that Kanto region is occupied by a series of metamorphic rocks which had been formed under the effect of compression necessarily accompanying a relatively high hydrostatic pressure. Furthermore, this metamorphic zone is found to be sandwiched between the central great uplift and Kanto mountain chains both involving compressional folding of sedimentary rocks.

Table 1. Data of Direction of N.R.M, and K-A Ages of Granitic Rocks from the Northeast- ern Japan

D=Declination. I=Inclination. S=Number of rock samples. α=Radius of 95 percent circle of confidence (Fisher, 1953), K=Fisher's precision factor. AK (b), SN(b), KN (b), NG, YK and SD.: After Kato and Muroi (1965). The Evolution of the Island Arc of Japan and the Formation of Granites in the Circum-Pacific Belt 277

Table 2. Data of Direction of I.R.M. and Geologic Ages of Rock Samples from the Southwest- ern Japan

D=Declination. I=Inclination. S=Number of rock samples, α=Radius of 95 percent circle of confidence (Fisher, 1953). K=Fisher's precision factor. 1-17, After Sasajima et al. (1966, 1968). 18-31, After Kawai et al. (1961). 278 N. KAWAI, T. NAKAJIMA and K. HIROOKA

Table 2. -continued-

From the recent results of a radiometric study it is clear that both the metamorphic zones and uplift were nearly simultaneous at the times when the central great bend took place. The bend that occurred twice in the main island in the past was by no means due to such a simple force with which a single bar can be bent in the middle. The land deformation has to be accompanied by the above-mentioned compressional stress distribution. To account for the compression and the simultaneous bend, a hypothetical shortening of entire length of island is now proposed. The shortening is assumed to have oc- The Evolution of the Island Arc of Japan and the Formation of Granites in the Gircum-Pacific Belt 279

Fig. 9 Tension cracks or gaps to have occurred, if Kawai, Ito and Kume's model (1961) were assumed.

curred when the northern block moved southwestward by approximately 150km and the southern block northeastward by approximately 100km, while central one remained unmoved but was compressed in-between the two blocks. Although it is difficult at the present stage to clarify the origin of this tectonic movement, this shortening can not only harmonize with many geological discrepancies discovered but also throw much light upon relevant problems remaining unsolved up to the present time. The first important implication that comes out is that the great bend itself is the result of this shortening. One may assume reasonably that the island, while it was shrinking, was so deformed that the middle block was squeezed too strongly to be able to remain in the original position. The block as the result was pushed towards the east and south accompanied by simultaneous rotation to form the island bend. Secondly the central mass comprising the high mountain chains came to its existence at the place where the southward moving northern block and the northward moving southern block collided. Pre-existing Palaeozoic sedimentary beds sandwiched between the two opposing land blocks were compressed and folded to form the basis of the mountain belts. The shortening can, thus, be related to the causation of the central uplift. Next the Mikabu metamorphic complexes as well as Sambagawa metamorphic belt lying east of this uplift may be assumed to have been recrystallized from Palaeozoic sediments in the huge natural pressure chamber subsequently formed by the colliding lands. 280 N. KAWAI, T. NAKAJIMA and K. HIROOKA

The mountain building and the metamorphism, therefore, took place almost simultaneously by the same mechanical force. Furthermore, the shortening can be considered to have played the most important role in the formation of the median line. The line separates southern block into two parts, one being the northern half and the other southern half as called respectively by the name "inner zone" and "outer zone" of the southwestern Japan. From the west- ernmost end of the line in Kyushu the median line extends almost due east-east-north, passing through the centre of the southern block and ends at Nagano hot spring area east of the central uplift showing a characteristic north bend. Regarding the origin of this prolonged tectonic line, especially with respect to the characteristic sharp bend, no comprehensive interpretation seems to have been established despite the number of arguments proposed and ideas exchanged among geologists. A mechanism with which this line can be made up simultaneously or subsequently with other tectonic events such as the bend, the mountain building etc., is indeed possible, and it is derived without difficulty from the length shortening hypothesis as will be shown in detail in the following. When the southern block was moving eastward during the shortening it may have met an obstacle preventing the movement. In front of the block there stood Palaeozoic sedimentary layers. The only northern part of them was squeezed and upheaved to form the mountain range but the southern part was intensely compressed as well as sheared and even metamorphosed in the final stage of the squeezing. Consequently, the resistance preventing the eastward land movement was not so extremely strong in front of the northern "inner zone". In contrast, it was much greater in front of the southern "outer zone" . The existence of resistance forces differing at two places north and south allowed a strong shear stress to have distributed along the present position of the median line, and its grade increased with further land movement until at last a rupture initiated and it extended over the entire region of the line. This great rupture made the inner zone more mobile and capable of traveling over a long distance, more than 100km, relative to the outer zone. Straight movement of the front, however, was blocked, since compression of Palaeozoic sediments in front of the land block became progressively stronger as the movement continued. The inner zone, therefore, preferring the place at which the minimum resistance existed, bent its head toward north and finally pushed itself out into the region of the present Japan Sea. We suggest that Noto peninsula projection was formed in this manner. Recrystallization of metamorphic minerals in both the Ryoke and Sambagawa belts possibly proceeded at moderate temperature and under high pressure produced in the zone along the median line. The Ryoke high grade metamorphic rocks were originated in the place nearest to the interface of the two zones and, next to it, the Sam- bagawa low grade metamorphic rocks were formed. To be mentioned in connection with this metamorphism is that the pressure causing recrystallization was extremely in- homogeneous and that it included dominant shear stresses superimposed upon a relatively high hydrostatic force. The Evolution of the Island Arc of Japan and the Formation of Granites in the Circum-Pacifc Belt 281

Friction that occurred at the boundary separating the two zones was strong enough to produce a great shear stress and was dissipated as the heat necessary for the mineral orientation in these metamorphic rocks. To Japanese geologists, however, it has become an almost fundamental concept that both the Ryoke and the Sambagawa metamorphic rocks were formed some time during the late Mesozoic epoch as a prolonged belt extending over a distance about 1000 km and that this simple structure, after the laps of geologic time, was deformed into its present configuration showing a remarkable north bend and a south bend in K_anto district. The north bend and the south bend were thought to join smoothly with each other in a region north of "Fossa Magna". Unfortunately, the join was entirely masked by Tertiary volcanic and sedimentary rocks and by a series of young intrusive rocks now exposed at the surface. The continuum of the median line to the Kanto mountain region through the Fossa Magna cannot be seen. In the length shortening hypothesis from which was derived a great shift of inner zone a discontinuous rather than the continuous join of the two bends is more acceptable. At the place where we now see the south bend of the metamorphic belt, squeezed palaeozoic sediments were recrystallized to form a series of metamorphic rocks whose main trend of developments was accidental- ly south-west in direction. This direction, when combined with that of the north bent, led many geologists to suppose a large concave upward folding of the originally straight belts. Rather than the post-metamorphic deformation, a syngenetic formation of the belts with the appearance of the median line, land rotations, and the central uplift is strongly supported, since all these movements seem to be involved in a major movement with the shortening of the entire island, hence they can be explained more reasonably by it. The southern front of the northern block pushed itself toward south into the Pacific Ocean where the existing resistance force was smallest while it was squeezing Palaeozoic sediment there. The southward bend of the front and the Boso off-shore projection have thus been made. This projection, together with the Noto projection, made the island widest in Kanto-Chubu district. Very strong shear stresses arose and were superimposed upon a high hydrostatic force at the time of the projection. Mikabu, Ryoke and Sambagawa metamorphic rocks having remarkable schistosity and characteristic lineation have been formed under these particular conditions along the south-east trend of the belts. According to the recent results obtained from isotopic analysis by ,Kenoet al. (1960), Miller and Shibata (1961)and Miller et al. (1961, 1962) the following ages were well established: Hida metamorphic belt 180-190m.y. (Triassic) Ryoke metamorphic belt 91-102m.y. (Cretaceous) Abukuma metamorphic belt 91-102m.y. (Cretaceous) Sambagawa metamorphic belt 84-93m.y. (Cretaceous) The Ryoke and Hida metamorphic belts lie in the "inner zone" of the southwestern Japan, and the Sambagawa belt in the "outer zone" of the southwestern Japan. The 282 N. KAWAI, T. NAKAJIMA and K. HIROOKA

Abukuma metamorphic belt however lies in the northeastern Japan. As the above table shows, the Ryoke, Abukuma and Sambagawa metamorphic belts were formed nearly simultaneously at the times when the central great bend occurred. The metamorphic belts mentioned above have been classified by geologists into the following two groups; 1) Low pressure-High temperature type (about 5Kb, 400-500℃)

2) High pressure-Low temperature type (about 10Kb, 200-300℃) The Ryoke, Abukuma and Hida metamorphic belts are the Low pressure-High temperature type, and the Sambagawa metamorphic belt is the High pressure-Low temperature type (Miyashiro, 1961). Although one needs look for a proper reason why two different types of the metamorphic belts have been formed, we believe that the origin of the metamorphism was closely related to the formation of the median line and other geotectonic structures. Although no radiometric information is available as to rocks in the Hokkaido island, the mean direction of the remanent magnetization of Cretaceous rocks differs greatly from that of the Kitakami district as is evident in Fig. 1. The vector in Hokkaido dif fers by approximately 45° from that in Kitakami. The striking feature mentioned above suggests the following two possibilities regarding the movement of Hokkaido block. The first case is that the Hokkaido block and Kitakami block were combined together at the beginning of the bending, at least until the initial stage of the second bend of the main island, then later the Hokkaido block rotated clockwise by approximately

45° relative to the Kitakami block as shown in Fig. 10-A. The second possibility is that the Hokkaido block, without significant rotation, shifted northwestwards whilst the Kitakami and the Abukuma blocks rotated anticlockwise relative to the Hokkaido as well as the southwestern Japan as shown in Fig. 10-B. Three diagrams in Fig. 11 schematically indicate the length shortening of the entire island. After the shortening two overlays appeared in between the neighbouring land blocks. Four diagrams in Fig. 12 demonstrate a story how the originally simply shaped island changed in geologic time until the end of the Cretaceous epoch. The origin of the local shortening in the vicinity of the Japanese islands should be considered. The drifts on the globe of the five continents from their original positions brought the opening of the Atlantic on one hand, enforced the narrowing of the Pacific on the other. This resulted in the shortening of the radius of the ocean and that of the circumference. The deformation and compression, therefore, occurred in the circum- Pacific zone. The shortening accompanying deformations in the Japanese islands is only a part of this mega-geotectonics. The Evolution of the Island Arc of Japan and the Formation of Granites in the Circum-Pacific Belt 283

Fig. 10 The two possible cases regarding the movement of Hokkaido block. A. The Hokkaido block and Kitakami block went on together in the beginning of the bending, then Hokkaido block rotated clockwise relative to the Kitakami block. B. The Hokkaido block, without no significant rotation, shifted north-west in direction.

3. Eastward Drift of Asian Continent and a Possible Folding of Palaeozoic Basin

In the previous section the hypothesis of the shortening of the Japanese islands was described but its real causation was assumed to be difficult to understand thoroughly. In this section, however, a possible Mesozoic drift of Asian continent relative to the Pacif- ic Ocean is first assumed to have occurred. The drift was related to the causation of the deformation of Palaeozoic sedimentary basin which originated at the continental margin to explain the local length shortening of the basin periphery. It is indeed quite reasonable to assume that in the past 100m.y. Palaeozoic rocks formed in the geosyncline were composed of relatively soft substances whose density and porosity resembled more or less to those of the present Mesozoic sediments. These past sediments had been formed in a past geosyncline whose bottom was successively deepened and was probably plunged into the past ocean floor with increasing depth of the basin. Therefore, if the continent had moved, the basin was not only pushed by the continent from the west but also compressed from the east by the pre-existing hard ocean- ic floor lying in front of it. It is thus possible that an intense plastic deformation in- 284 N. KAWAI, T. NAKAJIMA and K. HIROOKA

Fig. 11 The evolution of the Japanese islands S: Southern block A: Abukuma block K: Kitakami block H: Hokkaido block

cluding such length shortening as well as the metamorphism as already described was caused with no particular difficulty. On the other hand, it was only recently that Taylor-Wegener's famous continental drift hypothesis met an opportunity of the great reappreciation. Palaeomagnetic studies by Greer (1964), Runcorn (1962) and Blackett et al. (1965) offered a series of good information concerning relative movement of the five continents now separated by the present oceans. The radiometric-geochronological investigation directed by Hurley (1968) also demonstrated a beautiful age conformity as well as the famous geographical coincidence between the Atlantic coast of south America and that of Africa. Furthermore, Bullard's ingenious calculation (1965) has clearly indicated a remarkable join of the four continents around the Atlantic Ocean, i.e., North America, South America, Europe and Africa. The existence of the proto-continent called "Pangea" by Wegener and its Mesozoic splitting have become gradually accepted even by many conservative geoscientists. Sepa- The Evolution of the Island Arc of Japan and the Formation of Granites in the Gircum-Pacific Belt 285

Fig, 12 The evolution of the Japanese islands U: Uplift of the central massif K: Kamuikotan metamorphic belt KI: Kitakami intrusive rocks H: Hidaka metamorphic belt AB: Abukuma intrusive rocks N: Nato projection ML: Median line B: Boso projection R: Ryoke metamorphic belt h: Hida mountain range S: Sambagawa metamorphic belt k: Kanto mountain range M: Mikabu metamorphic belt a: Akaishi mountain range 286 N. KAWAI, T. NAKAJIMA and K. HIROOKA ration of Euro-Asian table land from North America with simultaneous formation of the Atlantic Ocean was assumed to have initiated in the middle of Mesozoic time. This separation implies eastward drift of the Euro-Asian table land over more than 2000km relative to North America upon Mesozoic Pacific Ocean. This drift might have reached the maximum some time in the late Cretaceous epoch when mountain building began around the circum-Pacific tectonic belts and subsequent intrusion of acidic rocks followed.

Fig. 13 The evolution of the palaeozoic basin

Shown in Fig. 13 there is a schematic representation of the palaeogeography in that active period of the drift, Palaeozoic sediments had deposited in a relatively wide basin occupying an area corresponding to the present Japan sea. The eastern edge of this basin was relatively straight, extending in a direction north-east, and facing directly to the Mesozoic Pacific Ocean. On the other hand the northern edge of the basin was represented by the Kamchatka-Hokkaido line. The eastern edge and the northern edge of the basin made approximately a right angle at the Hokkaido join. So long as the eastward drift of the entire continent continued, it is natural that soft sediments piled up at the both edges were squeezed intensely. The mountain ranges in direction parallel to the both edges were first formed. Due to the Mesozoic spread of Pacific Ocean floor the old Japan trench was also formed in a direction parallel to the eastern edge. The early mountain ranges now remain in Kitakami, Abukuma and Shikoku district and a stretch of the old Japan trench can now be seen remaining off the coast of Shikoku and Kyushu islands. When the drift increased one stage further, the Hokkaido join, after receiving a The Evolution of the Island Arc of Japan and the Formation of Granites in the Gircum-PacificBelt 287 very strong force from the Pacific Ocean, was brought back greatly toward the centre of the basin. The eastern and northern edges composed of deformed Palaeozoic sediments were also pushed back so significantly that relatively widespread old basin became gradually limited. The withdrawal of the basin front took place and this might reasonably be related to the local length shortening of island block that was discussed in detail in the previous section. The narrowing of the basin could also be connected to the subsidence of the internal part of the basin to form the present Japan sea and to the uplift of Aikhota mountain ranges in east Manchuria.

4. Origin of the Circum-Pacific Mesozoic Intrusive Rocks Although a story on the evolution of Japanese Arc and other Mesozoic geology seems to have partly been clarified as already described, there remain still many important questions on which much light should be thrown. They are, for example the questions as to whether the continental drift had occurred prior to the deformation of the land mass in front of the continent, whether the latter had preceded the former, whether the two movements occurred simultaneously, or why an enormous Cretaceous intrusion of acidic magma had been formed only in a very short interval of the geologic time. If the deformation was assumed to have preceded before, one should consider that the land mass been deformed by the time when the continent gained the drift velocity after the start or at least in the beginning whilst it was not transported significantly. On the contrary, if the drift had preceded, one should imagine that the deformation took place when the continent was decelerated or when it was receiving a strong brake as movement was arrested. The locality at which this post-drift deformation developed is, therefore, approximately 2000km further east of the place where the basin had originated. If, on the other hand, the deformation was assumed to have succeeded during the time of the drift, then it is clear that the island arc and other geologic structures were formed on a relatively wide span along the small circle of the earth's surface on which the continent had drifted. The force required to compress Palaeozoic sediment was possibly arisen from the resistance preventing drift on the ocean from time to time. To solve these questions the present authors summarized the palaeomagnetic data obtained from both North America and Eurasia. These data were compared with the similar data accumulated in Japan. Pole paths determined independently for Eurasia and for North America, if they were traced back in time, have to coincide with each other at least in a time span from the recent time back to the geologic time when the relative movement had come to the end. Splitting of pole path into two, however, would be expected in the data older than that time. Increasing split with decreasing geologic age is due to the relative movement, and can be seen in a geologic time interval whilst continents were moving. No further split should be seen, however, in the two pole paths whose geologic ages are older than that of the continental drift. The said palaeomagnet- 288 N. KAWAI, T. NAKAJIMA and K. HIROOKA is evidence is to be especially useful to know what process among the above-mentioned ones played the most important role in the past. As shown in Fig. 14, American and Eurasian Cretaceous poles occupy the eastern part of Siberia, and the Japanese pole lies in the North Pacific. These slight discrepancies seem to support the pre-drift compression on which was superimposed the major deformation contemporaneous with the drift.

Fig, 14 Mean virtual geomagnetic poles for Cretaceous rocks from North America, Eurasia and Southwest Japan

Let us recall an important characteristic of geosynclinal deposition of sediments that we have learned in introductory geology courses. A geosyncline in general has a bottom which is continuously deepened so long as a new layer is forming on its upper surface, so it has a more or less constant depth of the sea above the deposits. Due to this intrinsic property the geosyncline can produce unusually thick sedimentary basin. As the geosyncline originated along the eastern coast of the Asian continent had its root deeply buried below the ocean floor plate, it is reasonable to suppose that the compression and squeezing advanced in the earlier stage at some time before the upper layers of sediments were sheared and separated from the remaining layers in the root. On the other hand, a movement of sea floor by the help of the mantle convection has recently been accepted by many geophysicists and assumed as a major driving force of the continental drift. Recent discovery of a regular pattern of the geomagnetic anom- alies on the present oceans was reasonably related to the alternative switching of the geomagnetic dipole field and the continuous spread of the sea floor. The spread can be considered to have continued since the time of the initiation of the continental movements. The Asian continent was floating on the basaltic sea floor, and moved east wards with it as it spread and expanded in the same direction. In front of the moving The Evolution of the Island Arc of Japan and the Formation of Granites in the Circum-Pacific Belt 289 continent, there stood a Pacific sea floor plate which was moving from east in a direction approximately opposite to that of the continent. The continental front and the Palae- ozoic basin probably received an enormous forces. By this force such a down buckling bulge as one can now see in the present continental margin was formed, although it was much smaller in radius at the beginning of the drift. The bulge, on further sliding of the continent, developed progressively. The radius became so large that a strong shear stress arose at place where the west-going and sinking Pacific floor was conflicting with the counter-going bulge. The bulge played an important role in halting the further sliding of the continent. The increasing friction with further sliding prevented the drift, and the continent at last halted approximately in its present position at the end of the Mesozoic era or in the beginning of Tertiary epoch. At the bulge developed below the continental margin the deformation of the Palae- ozoic basin was particularly enhanced near the surface of the margin to form there the geotectonic structures and the metamorphic belts described in the previous section. The force of this strong friction should not be underestimated. It was so strong and it had existed such long time during the entire sliding of the continent over the distance of more than 2000km that the energy dissipated into heat was accumulated and stored beneath the continent. When temperature in the lower part of the bulge was elevated to about 700℃ the sediments began to be melted. At higher temperatures even a partial melting under the nearby continent may have occurred. An acidic and migmatitic magma was originated at depth and subsequently intruded into a fracture zone of the overlying basin. The intrusive granites and granodiorites were thus formed around the Pacific coast of the Asian continent. The partial melting and intrusion succeeded in the later stage of the sliding. The age of earliest intrusion now found in Japan was 180m.y. The Hida metamorphic granite seems to be a typical example. Then a series of intrusions followed in a period from 135m.y. to 70m.y. Although sliding almost stopped at the beginning of Tertiary epoch, the friction did not decrease, but instead increased. The heat of friction was accumulated further and the temperature of the bulge in depth was elevated higher. Partial melting of its substances was succeeded in this period and it took place at a higher temperature. A volcanic magma was possibly formed from which Tertiary volcanic rocks and pyroclastic sediments were derived. The palaeomagnetic and the chronological information obtained from these granites depends greatly upon this special melting due to the dissipated energy. It is interesting to compile the geological aspects ranging from the pre-drift epoch to the post-drift epoch comprehensively in a time table as shown by the items arranged in Table 3. It is possible to estimate the heat of the friction that had been accumulated in the vicinity of the boundary between the sinking Pacific floor table and the up-thrusting upper mantle beneath the Asiatic continental bulge. To obtain this heat one must first find the shear strews τx, viscosity of the material η, and the velacity gradient in the shear- 290 N. KAWAI, T. NAKAJIMA and K. HIROOKA

Table 3.

ing zone in direction normal to the material flow d Vx/dz. The three values are mutual- ly joined by a simple equation

Tx=-ηdυx/d z

The shear energy E, accumulated within a unit volume of the shearing zone can be expressed by an integration of the stress over an entire distance on which the sea floor

plate had traveled. E, which is equal to ∫τxdx per cc., can be worked out when the distance was again assumed to be 2000km. Next, if a deep-seismic plane (Benioffzone (1954)) beneath the island arc of Japan was assumed as the place where the shear stress occurs, and also if the shearing distance The Evolution of the Island Arc of Japan and the Formation of Granites in the Circum-Pacific Belt 291 per year of the Pacific floor table was 4cm, then it is possible to evaluate the velocity gradient to be,

dυx/dz=4cm/year/ 107cm where the thickness of the zone was also assumed as 100 km as one can now observe in the present continental periphery.

Next η=1021 poise which came out from the data of the uplift of Scandinavian was employed for the sake of convenience and assumed as the actual viscosity of the upper mantle in question. When the work done by the shear stress over the travel distance of the sea floor was calculated, and then transformed into the dissipated heat, we found that the heat Q amounted up to approximately 6×107 cal per 1cc. This heat is so large indeed that it is possible to warm the Palaeozoic sediments in the geosyncline and also a part of the upper mantle up to their melting points and further to supply the materials the latent heat of fusion to cause an acidic magma from which the enormous granodiorites now exposed around the Pacific belts were derived. The molten material was possibly elevated from the depth and easily intruded into the fracture zones previously caused by drift in the overlying mantle or in the continental substances. If the fusion had occurred in the deeper part of the friction surface, it is reasonable that a larger volume of magma was generated, since the material in the depth is so warm already that not much heat is required for the melting. So that it is possible to expect an intrusion of the granodiorite even from a deeper part of the upper mantle as Hurley et al. (1962), and Shirahase et al. (1969) have recently confirmed from their relatively smaller abundance ratio of Sr87 to the ordinary Sr86 observed in the Cretaceous granodiorites, or as Matsumoto's (1964, 1965, 1969) phase equilibrium diagram has indicated the deep origin of the magma. In Fig. 15 is shown schematically the origin of granodioritic magma.

Fig. 15 The origin of granodioritic magma 292 N. KAWAI. T. NAKAJIMA and K. HIROOKA

References

Benioff, H.: Orogenesis and deep crustal structure, Bull. Geol. Soc. Amer., 65, 385-400 (1954). Blackett, P.M.S.: Comparison of ancient climates with the ancient latitudes deduced from rock magnetic measurements, Proc. Roy. Soc. A, 263, 1-30 (1961). Blackett, P.M.S., E.C. Bullard and S.K. Runcorn: A symposium on continental drift, Phil. Trans. Roy. Soc., 258, 41-51 (1965). Creer, K.M.: A reconstruction of the continents, for the upper Palaeozoic from Palaeomagnetic data, Nature, 203, 1151-1120 (1964). Dietz, R.S.: Continent and ocean basin evolution by spreading of the sea floor, Nature, 190, 854-857 (1961). Fisher, R.A.: Dispersion on a sphere, Proc. Roy. Soc. London, A 217, 295-305 (1953). Gromme, CS.: Paleomagnetism of Jurassic and Cretaceous Plutonic Rocks in the Sierra Nevada, Cali- fornia, and its significance for polar wandering and continental drift, J. Geophys.Res., 72, 5661-5684 (1967). Holmes, A.: Principles of Physical Geology,2nd ed., pp. 1287, Thomas Nelson and Sons Ltd., London (1964). Hurley, P.M., H. Hughes, G. Faure, H.W. Fairbain and W.H. Pinson: Radiogenic Strontium 87 model of continent formation, J. Geophys. Res., 67, 5315-5334 (1962). Hurley, P.M.: The confirmation of continental drift, Scientfic American, 218, No. 4, 53-64 (1968). Irving, E.: Paleomagnetism,John Wiley and Sons, New York (1964). Japan Meteorological Agency: Catalogue of major earthquakes which occurred in and near Japan, 1926- 1956, (1958). Japan Meteorological Agency: Catalogue of major earthquakes which occurred in and near Japan, 1957- 1965, (1966). Kalashnikov, AC.: The history of the geomagnetic field, Izo. Bull. Acad. Sci. USSR Geopkts. Ser., 9, 819- 837 (1961). Kato, Y, and I. Muroi: Palaeomagnetic studies of the Cretaceous granitic rocks in northeastern Japan, Ann. Progr. Rept. Rock Magnetism Res. Group Japan, 1965, 179-187 (1965). Kawai, N., H. Ito and S. Kume: Deformation of the Japanese islands as inferred from rock magnetism, Geophys.J. Roy. Astro. Soc., 6, 124-129 (1961). Kawai, N., S. Kume and H. Ito: Study on the magnetization of the Japanese rocks, J. Geomag. Geoelect., 13, 150-153 (1962). Kawai, N., K. Hirooka and T. Nakajima: Palaeomagnetic and potassium-argon age informations support- ing Cretaceous-Tertiary hypothetic bend of the main island Japan, PalaeogeographyPalaeoclimatol. Palae- oecol., 6, 277-282 (1969).

Kawano, Y. and Y. ueda: K-A dating on the igneous rocks in Japan (Ⅰ), Sci. Rep. Tohoku Univ., ser. Ⅲ,

9, 99-122 (1964).

Kawano, Y. and Y. ueda: K-A dating on the igneous rocks in Japan (Ⅱ), Sci. Rep. Tohoku Univ., Ser. Ⅲ,

9, 199-215 (1965).

Kawano, Y. and Y. ueda: K-A dating on the igneous rocks in Japan (Ⅲ), Sci. Rep. Tohoku Univ., Ser. Ⅲ,

9, 513-523 (1966).

Kawano, Y. and Y. Ueda: K-A dating on the igneous rocks in Japan (Ⅳ), Sci. Rep. Tohoku Univ., Ser. Ⅲ,

9, 525-539 (1966).

Kawano, Y. and Y. ueda: K-A dating on the igneous rocks in Japan (Ⅴ), Jour. Japan. Assoc. Min. Petr. Econ.

Geol., 56, 191-211 (1966).

Kawano, Y. and Y. Ueda: K-A dating on the igneous rocks in Japan (Ⅵ), Sci. Rep. Tohoku Univ., Ser. Ⅲ, 10, 65-76 (1967). Kuno, H., H. Baadsgaard, S. Goldich and K. Shihata: Potassium-argon dating of the Hida metamorphic complex, Japan, Japan. J. Geol. Geog., 31, 273-278 (1964). Larochelle, A.: Paleomagnetism of the Monteregian Hills: New Results, J. Geophys. Res., 73, 3239-3246 The Evolution of the Island Arc of Japan and the Formation of Granites in the Circum-Pacific Belt 293

(1968). Lee, J.T., H.M. Lee, H.S. Liu, C. Lie and S.S. Yee: Preliminary results of paleomagnetic study of some Cenozoic and Mesozoic red beds, Southern China, Acta GeologicaSinica, Peiking, 43, No. 3, 241-245 (1963). Le Pichon, X.: Sea-floor spreading and continental drift, J. Geophys.Res., 73, 3661-3697 (1968). Matsumoto, T.: Upper mantle symposium monograph IUGS (1964). Matsumoto, T.: Upper mantle symposium monograph JUGS (1965). Matsumoto, T.: Some aspect of the formation of primary granitic magma in the upper mantle, Memo. Geol. Soc. Jap., (1969). Miller, J.A. and K. Shibata: Potassium-argon ages of Ryoke granite from Obatake, Yamaguchi Prefec- ture, Bull. Geol. Surv. J., 12, 653-654 (1961). Miller, J.A, and K. Shibata: Potassium-argon ages of granitic rocks from the Kitakami highlands, Bull. Geol. Surv. J., 13, 709-711 (1962). Miller, J.A., F. Shido, S. Banno and S. Uyeda: New data on the age of orogeny and metamorphism in Japan, Jap. J. Geol. Geog., 32, 145-151 (1961). Miyashiro, A.: Evolution of metamorphic belt, Jour. Petr., 2, 277-311 (1961). Nee, L.: Some theoretical aspect of rock-magnetism, Advanc. Phys., 4, 191-243 (1955). Runcorn, S.K. (ed): Continental drift, International Geophysics Ser. Vol. 3, Academic Press (1962). Sasajima, S, and M. Shimada : Paleomagnetic studies of the Cretaceous volcanic rocks in the southwestern Japan, J. Geol. Soc. Jap., 72, 503-514 (1966). Sasajima, S., M. Shimada and J. Nishida: Paleomagnetism of the Paleogene system in the inner zone of southwest Japan, with special reference to the paleomagnetochronology, J. Geol. Soc. Jap., 74, 597-606 (1968). Shirahase, T., H. Kagami and S. Izumi: Strontium and lead isotope studies on granitic rocks, Iso. Peter. Geochem.,No. 5, 69-78 (1969). Uyeda, S., and K. Horai: Terrestrial heat flow in Japan, J. Geophys.Res., 69, 2121-2141 (1964). Uyeda, S., and V. Vacquier: Geothermal and geomagnetic data in and around the Island Arc of Japan. In The Crust and Upper Mantle of the Pacific Area. edd by L. Knopofl', C. Drake and P. Hart, Geophys. Monograph 12, Amer. Geophys. Un., 349-366 (1968). Wells, J.M, and J. Verhoogen: Late Paleozoic paleomagnetic poles and the opening of the Atlantic Ocean, J. Geophys. Res., 72, 1777-1781 (1967).