Art. XVI.—The Orientation of the River-valleys of Canterbury.
[Read before the Philosophical Institute of Canterbury, 1st September, 1915.]
The remarkable orientation of the river-valleys of Canterbury has attracted the attention of numerous writers on physiographic subjects since the main features were first pointed out by Edward Dobson in his report for the year 1865 on “The Possibility of constructing a Road through the Otira Gorge” furnished to the Provincial Council of Canterbury. This report was fully noticed by Haast in his “Geology of Canterbury” (p. 174), and by McKay in his report on Marlborough and Amuri Counties (“Reports on Geological Explorations for 1890–91,” p. 26). Reference has also been made to it in Kitson and Thiele's paper on the Upper Waitaki basin (Geographical Journal, vol. 36, 1910, p. 547) and by Gregory in his book on “The Nature and Origin of Fiords,” 1913 (p. 366). In view of these references it may not be altogether irrelevant at the present time to re-state the case and to examine in the light of later observations the evidence for and against this hypothesis.
Before dealing with the origin of the valleys it is as well to consider their arrangement in order to understand how far the statement made by Dobson is correct. He says (loc. cit., p. 50), “In addition to the folding
above described, the rocks of the central chain have been subject to a variety of upheavals and dislocations which have resulted in the formation of a system of valleys, the direction of which is very remarkable; the principal valleys, from the Taramakau on the north to the Makarora on the south, radiating from a common centre situated about fifty miles to the north of Mount Darwin. It might naturally be imagined that these valleys would form passes through the dividing range, but such is not the case, as, with the exception of the Hurunui Valley, they do not extend through the western portion of the chain, but terminate in glaciers or are bounded by high rugged precipices, as is shown in the sketch of the range at the head of the Waimakariri.”
The fact that the passes through the range do not accommodate themselves as they should renders it extremely doubtful if the existence of a series of great earth-fractures can be postulated to explain the orientation of the valleys. The middle portion of the valleys occupied by the great lakes of Canterbury—viz., Sumner, Coleridge, Tekapo, Pukaki, and Ohau —do correspond with the arrangement indicated by Dobson. Again, though several of the valleys seem to accommodate themselves naturally
as far as their main portion is concerned, as we approach the main divide there appears to be a marked departure from the arrangement. This is specially noticeable in the case of the Upper Waimakariri and the Rangitata, and the direction of the line of fracture accounting for the Rakaia is made by Dobson to follow the line of the Mathias River, and not the main Rakaia, from the junction of the former with the latter, although the tributary is of comparatively small size when compared with the main stream, whose upper valley lies right athwart the direction of the suggested fracture in that region. Therefore, if we grant that the dominating influence which determined the course of the main valleys was a system of earth-fractures, some other cause has determined the direction of the head-waters.
Dobson was no doubt conscious of other determining causes, since he has noted very fully the direction of fold and joints of the rocks and their important effect on stream-directions. I think that if full note be taken of the direction of the joints and folds it will explain most of the anomalies in the directions of the passes through the main divide and the departure from the system of earth-fissures—if, indeed, they exist—which is admitted by both Dobson and Haast.
First of all, the general direction of folding in the neighbourhood of Arthur's Pass and farther north is N. 22° E. We have therefore the peculiar circumstance that the direction of folds crosses the range at an angle, and is not parallel with its general trend. This is also the case in the Kaikoura Mountains, as I have been informed by Drs. Allan Thomson and C. A. Cotton. The angle cited by Dobson is the average angle, and it will be found that there are marked divergencies from this value in particular places—for example, in Otira Gorge the line of strike is almost in a line with the direction of the tunnel, being practically due north. In some places the same bed can be seen, owing to slight local variations, following along the top of the inside of the tunnel for several chains—that is, in a direction N. 8° W. It is thus apparent that the line of the Otira and Bealey Rivers follows almost exactly the line of strike of the beds; and, as the dip is in contrary directions on opposite sides of the Otira Valley, so that the valleys of these rivers have been eroded along the axis of an anticline, the formation of the pass between the heads of the valleys is in all probability due to the capture of the eastward-flowing streams originally running through the pass by streams discharging west—a result to be attributed very largely to glacier erosion and the sapping-back of heads of glaciers in the manner first described by Matthes in the report on the “Glacial Sculpture of the Bighorn Mountains” (21st Annual Report U.S. Geol. Surv., 1899–1900, pp. 167–90). This idea has been amplified and extended by W. H. Hobbs in a paper entitled “The Cycle of Mountain Glaciation,” published in the Geographical Journal, vol. 35, 1910, p. 146. This piracy has no doubt been hastened by the steeper grade of the rivers and also of the ice on the western side of the main divide.
On considering the directions of the valleys at the headwaters of the Waimakariri the influence of the general direction of folding is quite apparent, especially in the upper waters of the White River, the southern tributary of the Waimakariri. This is one of the cases noted by Dobson as difficult to explain on the supposition that the valley-directions were determined by a series of earth-fractures, since the strongly defined mass of Mount Greenlaw completely blocks the end of the valley of the main river.
On the other side of the divide between this valley and that of the Rakaia, in the neighbourhood of Browning Pass, the strike of the beds lies right across the main valley leading up to the pass; but when the valley of the Wilberforce is traced up to the bluff which closes its upper end almost completely it turns to the north and follows the line of the strike exactly. The beds here dip north-west, and the scarp slope has formed a line of cliff, composed of loose unstable rock extending for some miles, and quite unscalable from the east except at the pass itself, where the western wall has been broken down to some extent. This line is continued down the valley of the Taipo on the western slope of the range in almost perfect alignment as far as the junction of that river with the Seven-mile Creek, when the river makes a right-angled bend across the strike before it flows through its gorge to join the Taramakau. The parallelism of the upper waters of the Arahura, which are in close proximity to those of the Taipo, is noteworthy in this connection.
In the Whitcombe Pass, between the head of the Rakaia and the Whitcombe River, a tributary of the Hokitika, the direction of the strike of the beds corresponds fairly accurately with the long narrow trench which runs N. 10° E. for a distance of twenty miles through the range, and is occupied in Westland by the Whitcombe River and in Canterbury by the Louper Stream (Bulletin No. 6 (n.s), N.Z. Geological Survey, 1908; maps). When the direction of the pass is followed south across the Rakaia there is no sign whatsoever of the persistence of the trench. This valley is as marked in the alignment of its valley-walls as that of the Upper Bealey and the Otira, and the agreement in both cases with the folding of the beds seems to indicate that the former of Dobson's causes is the dominating one in determining the directions of the valleys in the neighbourhood of the passes, though Morgan (loc. cit., p. 73) inclines to a fault origin for the Whitcombe Valley.
If we now consider the passes at the head of the Godley, Tasman, Hunter, and Makarora Rivers, they appear to conform more generally to the average strike of the beds in those regions, and this, and not the presence of a series of major earth-fractures, may be the explanation of their orientation.
The actual passes where they cross the main divide seem, therefore, to be controlled in direction by the strike of the folded greywackes. In some cases, however, other general causes must be present which have had a determining effect on the course of the upper streams—e.g., there is such a marked resemblance in the headwaters of the Waimakariri and Rakaia that it can hardly be the result of chance. The chief source of the Waimakariri is the White River, which rises in a glacier on Mount Davie, to the west of Mount Greenlaw, a mass which apparently blocks the valley, when looking up-stream from the Bealey Township. The course of the White River Valley is for about three miles in a north-easterly direction, parallel to the strike of the beds, with its south-eastern side a wall of precipitous rocks standing almost vertical in agreement with the high dip; but the river breaks through this barrier in a course cut at right angles past the north-eastern end of Mount Greenlaw, receives a tributary from the northeast, and then flows about south-west, almost parallel to its original direction, past the front of Mount Greenlaw, and then finally turns again almost at right angles to the south-east and follows that general direction past Bealey and the Cass towards the sea.
The case of the Rakaia is analogous. The Lyell Glacier discharges by a river which runs north-east along the strike of the beds behind the
spur of Mount Goethe stretching down to Mein's Knob, junctions with the Ramsay, which flows south-west along the strike behind the Butler Range, and the combined stream, forming the main Rakaia, cuts across the beds at right angles, past the northern end of Mein's Knob and south of Jim's Knob, the southern termination of the Butler Range, and thence onward has a general south-easterly trend.
It seems certain that the direction of the folding does exert a determining influence on the upper courses of the river-valleys; but when we examine the middle courses of the streams, before they leave the mountain tract and debouch on to the plains, we find that their orientation has apparently little relation to either. The direction of folding varies so much from point to point that it is almost impossible to arrive at any general direction which is in agreement with the lines of the stream-valleys. In most cases. they cut across the strike diagonally, with angles varying from 45 to 90 degrees. The Waimakariri Gorge may be an exception to this. Above the narrow part, near the northern extremity of the Torlesse Range, the flow of the stream is approximately parallel to the strike, as is exemplified by that part of its course along the western flank of the Puketeraki Range, and then it crosses the strike of the beds almost at right angles. This is perhaps an illustration of the features of a young-river gorge to be noted directly; but after a consideration of all the river-valleys and the directions of the fold of the greywackes I can find no agreement of the orientation with even an average direction, and therefore some other cause must in all probability be considered as responsible for the agreement.
When, however, we examine the gorges of the rivers which are of recent origin, we find that in numerous cases their directions either as a whole or in part are directly controlled by both the directions of strike and of jointing. The determination of the latter is at times a matter of some difficulty, since its direction is by no means regular. In general a jointing at right angles to the stratification predominates, but besides this there are additional associated joint planes which intersect the dominating ones at an angle approaching 45 degrees. This is perhaps what we should expect if quadrangular blocks which had already been produced by the normal causes of jointing were subjected to a pressure at right angles to their faces, just as when the strength of a block of cement is tested under a crushing-machine it generally fails on lines which meet the faces of the block at an angle of 45 degrees. The persistence of the cross-joints and the relation of their angle to the dominant ones can probably be explained in this way. This is specially the case where the rock has shown symptoms of failure under earth-pressure; but where it is more resistant, especially where it consists of hard tough greywacke, the prevailing direction of jointing is at right angles to the stratification.
The right-angled bends at times characteristic of young-river gorges probably depend in general on the strike of the beds and the directions of the dominant joints, one reach corresponding to the strike and the next to the joint-direction, and the length of each of these reaches will bear some relation to the importance of the relative effect of the two causes. In general, however, it will be found that the strike is the dominating influence. If they are of equal importance, then the reaches will be approximately the same length. Parallel reaches, such as occur near the heads of the Rakaia and Waimakariri, are due to the river flowing first along the strike, then turning at right angles in direction of the joints, and then continuing the turning in the same sense, and running again along the strike in the
reverse direction. As the river passes from the youthful to the mature stage these abrupt right-angled turns will get smoothed out, and the control of rock-structure on the direction may be difficult, if not impossible, to determine; and specially will this be the case in a country where the direction of the strike, and with it that of the joints, is continually changing, even if we leave out of consideration the disturbing effect of another set of joints crossing at an angle approaching half a right angle. Thus it is that any, attempt to correlate our river-directions with the smaller rock structures will lead to doubtful result, although they have really had a dominating effect in the beginning.
Although Dobson states that the directions of the river-valleys was probably due to a series of earth-fractures, he did not submit positive proof of their existence, and I am unaware that any direct evidence has been forthcoming since he wrote his report with regard to the main valleys. In the case of some of the subordinate valleys there is, however, positive evidence of the presence of structural features controlling their direction.
In the valley of the Potts River, a tributary of the Rangitata, there is an outlier of coal-measures whose strike runs north-west parallel to the direction of the valley, and the coal-measures have been faulted down on a line running in this direction. This is altogether different from that of the strike of the greywackes in the locality, which runs approximately northeast. It appears, therefore, that there is an agreement with the line of fault in the direction of the valley. The valley has been enlarged by glaciation till its upper portion has become a typical glacial trough, but it is blocked at the lower end by a bar of greywacke through which the river has cut a gorge 600 ft. deep, where the effects of joints on the direction of short reaches is very apparent. This gorge exhibits for a length of about three miles a series of right-angled zigzags, one set of directions corresponding to the strike of the beds and the other to the cross-joints, the general trend of the gorge being, however, nearly at right angles to the line of the upper glaciated valley. Similarly, in the high country between Lake Heron and the Upper Cameron River there is an analogous occurrence of coal-measures, and, although no large valley runs north-west in association with them, small tributaries of the Cameron, together with well-marked landscape features such as bluffs, which may well be fault-scarps, and a series of peculiar depressions, follow closely the line of dislocation. This direction is parallel to that of the Lake Heron Valley, and remnants of Tertiary beds which are left in it are folded in places on a N. W —S E line
When we cross the Rakaia and consider the country in the neighbourhood of Lake Coleridge we find a series of five parallel valleys lying with a N W —S.E. direction (see map). These are—(1) The main Rakaia, between the gorge and Double Hill, bounded on the west by Mount Hutt and the Palmer Range, (2) the Lake Coleridge Valley, with its extension up the Wilberforce, (3) the valley through which the road passes from the southern end of Lake Coleridge to the Harper River, in which lie numerous small lakes, such as Georgina, Eveleen, Selfe, and which is separated from the Lake Coleridge Valley by Kaka Hill and Cotton's Sheep Range, (4) a valley to the north-east of this, at a much higher level towards the southern end of Lake Coleridge but specially well defined towards the Harper River, in which lies Lake Catherine, the valley being continued across the Harper up the Avoca River; (5) another valley, of small size, lying right against the western base of Mount Enys and the ridge running south-east from it. The remnants of Tertiary beds which exist at Redcliff Gully, on
the Rakaia, at the mouth of the Acheron River, near the Harper River, and at Mount Algidus give little idea of structural peculiarities which may find expression in the shape and position of these river-valleys.
When we cross into the Waimakariri basin we find the same arrangement of parallel valleys in the country between Broken River and the Cass. Taking the most westerly first, there are the following: (1) The valley occupied by Winding Creek and Lakes Pearson and Grasmere; (2) the valley occupied by Sloven's Creek and its continuation following the line of railway over St. Bernard Saddle to Lake Sarah and the Cass; (3) the valley of the Waimakariri.
The directions of the first two of these are certainly determined by structural features—at all events, in their lower portions. In Winding Creek the Tertiary coal-measures form a strip along the floor towards the junction with Broken River, and these have been folded with an axis running along the line of the valley to the north-west. Again, in Sloven's Creek there is a strip of Tertiary sedimentaries and associated volcanics with a north-west strike and a south-west dip which has been tilted and faulted down, and has certainly determined the direction of the valley in its lower reaches. Whether in both these cases the fractures continue north-west past the last exposures of Tertiary beds is quite uncertain; but it is extremely probable that it is so, the soft strata having been removed from the higher parts of the valleys by the increased intensity of erosive agents. There are no Tertiary sedimentary beds in the valley of the main Waimakariri indicating any line of structural weakness which may have determined its direction; but just opposite the point where the Esk River joins it there is proof of a dislocation running up the valley of the tributary in a north-easterly direction, the Tertiary sediments being subjected to folds on that line with a fault running up the western flanks of the Puketeraki Range in that region, tilting the sedimentaries along its line, and determining the direction of the valley either by the channel being formed by the fault-line or by depressing a strip of soft and easily eroded beds. This is the only marked structural line of weakness running north-east, but it is probably continued to the south-west toward the Castle Hill basin, and in the valley of Coleridge Creek, which leads towards Coleridge Pass, in the extreme south-west of the basin, we have a well-marked fault-line where the dislocation can be positively seen in the markedly different levels of beds dipping in the same direction on opposite sides of the valley.
We see, therefore, that, while certain subordinate valleys do show indications of their directions being dependent on deformations of the strata, the main valleys have not furnished any positive evidence of such; but this does not negative the statement that valley-directions are dependent in some way on such a cause. They may be lines of earth-fracture. It will be noted, however, that those cases which I have cited have a general northwesterly trend, whether they are in the Rangitata, the Rakaia, or the Waimakariri basin, and do not correspond with the lines suggested by Dobson. This divergence is most marked in the Waimakariri basin, for, while the map of fractures suggested by that author would make the direction of the upper Waimakariri almost east-and-west, the valleys which are dependent on fractures in that region run N.N.W.-S.S.E.; but it is not impossible that both sets exist, only there is positive evidence of the latter and not of the former. The chief difficulty in detecting fault-lines in the mountain region of Canterbury is that the rocks are of such a monotonous nature, with no well-defined zones of distinctive lithological character or beds which can be
distinguished by their fossil-content. The detection of structural dislocations is therefore a matter of extreme difficulty. Further, they have been modified by glaciation, and that agent produces landscape features in valleys which simulate fault-scarps, so that this criterion is very dangerous to apply. However, the marked agreement in grade between the main rivers and their tributaries indicates that any faulting which affected the main-river valleys and not the tributaries must have dated from a time long antecedent to the glaciation, or that all discordance in grade has been rapidly removed. The apparent conformity of the grade of the tributary with that of the main river is a remarkable feature of the valleys of Canterbury, and differentiates them from those of the West Coast Sounds, hanging valleys being quite infrequent except towards the heads of the rivers. It will be apparent, therefore, that this line of investigation does not hold out at present any great hope of furnishing data by which the problem may be solved.
It may be pointed out that the fractures which really do exist, with the exception of the Esk River and Coleridge Creek occurrences, do not follow the lines as indicated by McKay, and they are divergent from those described for the Kaikoura, region by McKay, Hector, and Cotton, which continue into north-eastern Canterbury (see map), and from those recorded in the west of Nelson by Henderson and Morgan.
There is one further suggestion which may be put forward tentatively to explain the arrangement of the valleys—viz., that their directions were primarily determined by the shape of the land-surface as it emerged from the sea in late Tertiary times. I have shown elsewhere (Trans N Z Inst., vol. 47, 1915, p. 353) that in all probability in late Cretaceous or early Tertiary times the present mountain region of Canterbury had the form of a peneplain; that subsequently this surface was depressed and was covered by a discontinuous veneer of marine sediments; that it was raised at the close of the Tertiary era and during the Quaternary era with some amount of folding and faulting. Now, the streams which. cut their beds on this land as it emerged would be consequent on the slope of the covering beds, and as they were removed by erosion our present drainage would become a superimposed drainage. The main valley-directions might then be of great antiquity, while the subordinate ones might be due to fractures of comparatively recent date. They would, therefore, resemble what we should get were the central district of the Wellington Province, between Wanganui and the Ruahine Range, still more elevated than at present, the mantle of sediments removed, and the base of older rocks exposed on which they lie unconformably. The present almost parallel alignment of the rivers would then be perpetuated on the exposed surface, and would have little relation to the arrangement of the underlying beds Even this method of establishment would not negative the existence of fractures; but with the complete removal of the covering beds it would be very difficult indeed, unless the strata exposed were of considerable diversity in lithological character or of fossil-content, to detect the presence of these fractures.