T

750 km u 1

Figure 7.5 Minimum radius of curvature of trench that can be defined by a plane dipping at 45°.

Figure 7.5 Minimum radius of curvature of trench that can be defined by a plane dipping at 45°.

Figure 7.6 The 'ping-pong ball' mechanism.

trench itself is a well-defined, near semicircular, feature which occurs from about 55°-62°S at 25°W, and forms the eastern boundary of the relatively small Scotia Plate. This plate is bounded to the north and south by strike-slip faults; the northern fault connects the northern limit of the trench to the tip of Tierra del Fuego, while the other inferred boundary fault runs from the southern limit of the trench almost to the eastern end of the South Shetland Islands. The final boundary to this small plate is defined by the Shackleton Fracture Zone.

It will be seen (Figure 7.7) that to the west of the Scotia Plate lies a deep abyssal plain at a depth of over 5000 m. To the east of the Scotia Trench is the S Atlantic, which, in this area, reaches a depth of 4-5000 m. The Scotia plate itself contains extensive, and also smaller, areas, where the depth of ocean ranges from 3000 m to as little as 1000 m. The structure of the Scotia Plate (Figure 7.7) gives some indication of why the ocean floor is relatively shallow. This diagram shows relatively short lengths of floor of different ages, indicating that ridge-like spreading has taken place in both a mainly east-west, and also a north-south, direction.

As regards the initiation and development of the Scotia plate and Arc, the problems that must be faced are (1) when and (2) why this feature developed. As a starting point, we take the position of the Antarctic Peninsula relative to the southern portion of S America, prior to the initiation of the southern-most, S Atlantic, 135 Ma ago (presented in Chapter 6, Figure 6.23, and repeated here as Figure 7.8a). Following this event, what was to become the Antarctic Peninsula drifted to the south, but probably continued to overlap the 'straight' tip of S. America for several tens of millions of years.

It will be recalled that Sir John Barrow (1830-31), followed by Arctowski (1895) and Suess (1909), suggested that 'the Andes are seen again in Graham Land'. Two attempts to explain the development of the Scotia Arc have been predicated upon the original linear alignment of the Andes and Graham Land (i.e. the northern tip of the Antarctic Peninsula). The first of these was by Hawkes (1962) (Figure 7.9a), and a similar initial configuration and subsequent migration of elements in the chain was presented by Dalziel and Elliot (1973), (Figure 7.9b). The outstanding feature of these constructions is that it was assumed that the swing of the Andes from N-S to E-W at Cape Horn did not initially exist, and, moreover, that the Andes were aligned with an uncurved Graham Land.

Sense of shear movement ^ M orientation of on N and SScotis Ridge ^ >— bi>K * dyke swarms

Sense of shear movement ^ M orientation of on N and SScotis Ridge ^ >— bi>K * dyke swarms

Figure 7.7 Schematic map showing the Scotia Arc and associated features (after Barker et al., 1984).

However, few recent workers have had the temerity to straighten out the southernmost extent of the Andes. Indeed, Dalziel changed his mind (as a co-author) and supported a construction presented by Barker et al. (1984), which comprises a different configuration of elements, but shows them as changing from a cuspate point which, these authors suggested, existed 30 Ma ago (Figure 7.10). Although this reconstruction, by Barker et al., shows more detailed geological information than that presented by Hawkes, the approach is intrinsically the same; namely, they both indicate the inferred track of various elements of the Scotia Arc from its inception (Hawkes) or over the last 30 Ma (Barker).

The pictorial development of the elements which 'frame' the Scotia Arc (Figures 7.9 and 7.10) is equivalent to the Bewegungsbild, or movement picture, that plays an important role in the thinking of some structural geologists. However, it must be emphasised that such movement patterns, desirable though they may sometimes be, are NOT mechanistic explanations regarding the geological structures or the development of the Scotia Arc. Indeed, as far as we are aware, no mechanistic model exists, which permits one to understand how the Scotia Arc developed.

The fine lines in Figure 7.8a (derived by the Atlas 3.3 program) show the northern tip of Graham Land, at 135 Ma, superimposed upon Patagonia. The straightened Andes and its relationship with the micro-plate of Graham Land are shown by thicker lines in Figure 7.8a. How the geometrical straightening of Patagonia in Figure 7.8a was derived is shown in Figure 7.8b.

It is important to note that Lawvers et al. (1985) argue that, at this time (i.e. 135 Ma ago), Graham Land was not attached to Antarctica (Figure 7.11); indeed they state that Graham Land did not become attached to Antarctica until 119 Ma.

The question that now poses itself is, at what time did the development of the Patagonian Orocline and hence, the curvature of the southern part of the Andes, take place?

Following Carey's (1955) suggestion regarding the development of the Patagonian Orocline, Dalziel and Elliot (1973) carried out a preliminary palaeomagnetic study of the area, which indicated that a degree of

Figure 7.8 (a) Overlap of the Patagonian Orocline and the northern tip of Graham Land, as shown by the Atlas 3.3 Program 135 Ma ago showing the impossible superposition of these two land masses. (b) The geometrical construction used to straighten the orocline.

such bending had taken place. This was confirmed by Burns et al. (1980) who stated that bending in the range 40° to 27° had occurred between 80 Ma and 21 Ma ago.

Figure 7.8 (a) Overlap of the Patagonian Orocline and the northern tip of Graham Land, as shown by the Atlas 3.3 Program 135 Ma ago showing the impossible superposition of these two land masses. (b) The geometrical construction used to straighten the orocline.

such bending had taken place. This was confirmed by Burns et al. (1980) who stated that bending in the range 40° to 27° had occurred between 80 Ma and 21 Ma ago.

Figure 7.9a Development of the Scotia Arc (after Hawkes, 1962).

Figure 7.9b Development of the Scotia Arc (after Dalziel and Elliot, 1973).

Cunningham et al. (1991) who presented the results of a more recent study, stated that neither of the earlier studies (by Dalziel et al. and Burns et al.) provided data that pass modern confidence 'filters' for data reliability. The new data presented by Cunningham et al. enabled them to conclude that this orocline is the result of approximately 90° of anticlockwise rotation of the crust, and that this has taken place since the mid-to late-Cretaceous (i.e. about 100-65 Ma ago).

If the early, intuitive interpretation regarding the original 'straightness' of the Andes has proved to be correct, then it is not unreasonable to question whether Graham Land too was originally straight. In fact, Grunow et al. (1987) have shown, from palaeomagnetic data, that Graham Land has rotated 30° in a clockwise direction since about 190 Ma. ago. However, it is interesting to note that Grunow et al., in a construction which purports to represent the distribution of S America and the Antarctic Peninsula, for 90 Ma, show the S American Orocline at its full development. We would reiterate Cunningham's statement that, as regards this particular point, Grunow's construction for the 90 Ma relationship is incorrect.

Figure 7.10 Initiation of the Scotia Arc from a cuspate form at 30 Ma (after Barker et al., 1984).

Consequently, we repeat the conviction of the early explorers in this region, namely that the 'Andes are seen again in Graham Land' and suggest that the most obvious conclusion is that, 140 Ma ago, the 'Antarctic Peninsula' was, in all probability, physically attached to S America as a direct continuation of the southernmost 'straightened' Andes.

It was further noted (as argued in Chapter 6) that the S Atlantic Ocean was initiated at its southernmost tip of (straightened) S America at around 135 Ma. Hence, it is suggested, that the 'Antarctica Peninsula' was separated from S America at this time.

From 135 Ma to about 65 Ma, Graham Land migrated southward toward Antarctica. The precise position of specific reference points on the S American and the 'Antarctic Peninsula' can only be given within a circle of error at the 95 per cent level of certainty, where the diameter of the circle is at least 100 km. Consequently, one can only state that, at some time between 80 Ma and 65 Ma, the northern tip of what by now was truly the Antarctic Peninsula, may have cleared or had only a modest overlap with, the still-straight portion of S America. This position is not too dissimilar from that assumed by Hawkes (1962) and Dalziel and Elliot (1973), shown in Figures 7.9a and b, and 7.10 a, b and c.

7.4.1 Mechanisms

Let us now consider the mechanisms which could give rise to a curved trench. Three questions must be answered: (1) how or why did the plate in the S Pacific break through on a relatively narrow front? (2) why

Figure 7.11 The motions of Africa and Antarctica relative to S America from the late Jurassic to the early Tertiary. The current Antarctic Peninsula did not become attached to the rest of Antarctica until 119 Ma ago. is the subduction trough almost exactly semicircular? and (3) how or why did the subduction sense become reversed and develop in the Atlantic lithosphere?

A linear trench can be given apparent curvature (dashed lines in Figure 7.12) if the trench is cut by a sequence of transform faults which exhibit a consistent amount of strike-slip displacement. To obtain a greater angle of arcuate development, it would be necessary to have a mirror image of the faults shown in Figure 7.12. The different sense of shear along the transform faults could, presumably, be attributed to variations in the speed of the ocean plate between the transforms, where the lowest rate of movement is in the centre. However, such structures would be readily observed. Such hypothetical structures could develop if an area of well-rounded, plateau basalt or ridge on the Pacific side failed to subduct beneath S America (cf. the Nazca-Sala-y-Gomez Ridge). However, the force of the Pacific lithosphere would still be opposed and balanced by that of the Atlantic lithosphere, so that there is no specific reason for breakthrough to take place.

The conventional mechanisms discussed in Chapter 3 do not provide a satisfactory explanation of how a breakthrough could have taken place. The circular form of the trench can only be derived by special pleading, and the question of how the subduction reversal took place must be taken on trust. However, such questions can readily be answered in terms of what was set out in Chapter 5, where aspects of impact tectonics were discussed.

A schematic representation of the ridge-push forces (P1-4), which are likely to have obtained when the tip of the Antarctic Peninsula was adjacent to, or slightly overlapped, the southern tip of S America, is given in Figure 7.13. We suggest that the geometry of the Scotia Arc itself is evidence regarding the probable mechanism by which it formed, and are aware of only one mechanism capable of developing this geometry on such a scale. That mechanism is a major impact (Figure 7.14).

At the time we first considered this problem, the only faults with a circular plan which we remembered to be associated with a major impact were the normal fractures which defined the peripheral graben (Figure 7.15). Because we are concerned with a situation in which the maximum principal stresses in the ocean lithosphere some 100 km from the ridge are horizontal and compressive, it follows that we were faced with an analytical problem.

Figure 7.12 Schematic representation of how a linear trench can be changed to a pseudo-arc by shear along transform faults.

Figure 7.13 The balance of forces to the east and west of S America and Graham Land prior to the initiation of the Scotia Arc.

Figure 7.14 Sketch map showing location of a major meteoritic impact relative to S America and Graham Land, together with the circumferential thrust(s) and radiating tensile fractures.

It can be shown (Price and Cosgrove, 1990) that, for normal faults with an angle of dip of 60° or more, it is usually impossible for a horizontal maximum, principal stress to cause re-shear, in the reverse mode, on such faults. Hence, the stresses will behave as though the normal faults did not exist. What was required by our hypothesis was the existence of arcuate thrusts which crop out around the crater. We went back to the data supplied by Jones (1995). Originally, we were unaware of the existence of such thrusts (see Price, 1975). In this paper I presented a sketch of a major impact, described verbally by Jones, which shows concentric normal faults and radiating vertical fractures, but no thrusts. It will be seen in Chapter 5 that such arcuate thrusts did, in fact, develop in the Snowball explosive experiment. A hypothetical model had been

300 km

Was this article helpful?

0 0

Post a comment