Mass Movement and Landform in New Zealand and Hong Kong

Transactions and Proceedings of the Royal Society of New Zealand, Volume 88, 1960-61, Page 623

Mass Movement and Landform in New Zealand and Hong Kong By Leonard Berry and Bryan P. Ruxton [Received by the Editor, June 3, 1960.] Abstract Hillslope deposits of unsorted weathered debris with angular fragments set in a matrix of silty clay or clayey sand rest on shaved weathered sedentary rock in similar topographic situations both in Wellington and in Hong Kong. These features are explained in New Zealand as the result of gelifluxion in a former periglacial climate and in Hong Kong as a normal result of chemical weathering and denudation in a humid tropical region. Dell-like hollows on the margins of upland surfaces also occur in both areas. Factors common to the two regions are (or have been) the thick weathering profiles and conditions favouring the frequent saturation of the superficial portion of the rock waste. Deposits resulting from mass movement are not always reliable indicators of past climates. Introduction A detailed description of landform features (by Cotton) and deposits (by Te Punga) resulting from gelifluxion at Wellington, New Zealand, was given in Cotton and Te Punga (1955) and in the Lower Hutt area by Stevens (1957). The nature of the deposits was used as evidence of a former periglacial climate, but similar deposits are found in the humid tropical region of Hong Kong and south China. Factors common to New Zealand and south China are the thick mantle of weathered rock and a frequently saturated surface layer. Considerable climatic contrasts exist at present between the two areas. Whereas Hong Kong is thought to have had a natural rain forest cover, the Wellington area appears to have been unforested when the deposits were accumulating (Cotton, in Cotton and Te Punga, 1955, p. 1019). Large-scale mass movements have played an important role in the retreat of hillslopes around the basins of south China and the resulting deposits (“fans”) were noted by Heim (1930), de Chardin (1935), and Li (1936). De Chardin (1935) thought that the deposits were formed by processes not at present in operation, but convincing arguments to the contrary were put forward by Li (1936). De Terra (1943, pp. 298 and 328) and Movius (1943) noted the formation of slope wash, creep, and landslide deposits in Burma, though most of these were assigned to former pluvial periods. Observations by the writers along the coasts between northern Taiwan and Malaya indicate that at present landslipping, earth flow and creep play a considerable part in the topographical evolution of these deeply weathered monsoonal and humid tropical lands. Hillslope Deposits in Hong Kong The hill and basin landform of south China is often related to areas of “weak” rock marginal to, or surrounded by, rocks more resistant to chemical weathering and subsequent mass-wasting and erosion. The granite cupola surrounded by volcanic rocks in the Hong Kong harbour area is a fine example and has been studied in detail (Berry and Ruxton, 1960, Ann. Geomorph., in the press). Evolution of the basin form was initiated by the incision of a stream along the major axis of this elliptical cupola and was followed by back wearing of the hill-slopes. At present preferential chemical weathering takes place around the hillfoot

due to the concentration of sub-surface water frequently recharged with weathering agents, and this effect is often accentuated by the presence of the granite-volcanic contact on the hillslope. Weathering takes place much more rapidly on the granite, and the resultant removal of the debris from the lower slopes causes oversteepening. Gravity collapse of the upslope weathered debris follows, usually triggered by periods of intensely heavy rainfall. This debris, slides, rolls, and later flows down to the hillfoot (Fig. 1). It comes to rest on slopes varying between 5° and 32°, and may be from a few feet to over 50 feet thick. Fig 1.—A, B and C, deposits formed on a hillslope below a granite-ignimbrite contact due to preferential weathering and erosion of the granite. D and E, deposits formed on a hillslope on granite due to preferential weathering and erosion at the hillfoot. The hillslope deposits consist of an unsorted mass of weathered and fresh rock fragments in a silty-clay (from weathered volcanic rocks) or clayey-sand (from weathered granite) matrix. The rock fragments are angular or rounded, varying from 1 inch up to over 30 feet in longest diameter. Even the larger blocks have travelled up to half-a-mile. Rounded fragments have been derived from weathered corestones and fragments of their concentric sheaths are sometimes found in the debris. Many angular fragments are little altered joint blocks from the volcanic rocks, but some are derived from corestones which have broken up during movement. Cross cracking of corestones enclosed in migratory debris or resting on the surface is a common feature in many tropical areas and the products are usually angular boulders with one curved face. The matrix of the hillslope deposits is usually composed dominantly of simple mineral grains (mostly quartz, some grains of which are splintery) and clay. Many sections of these deposits show two distinct layers, the upper less weathered with larger and fresher boulders. The colour of the matrix may he reddened throughout on well-drained sites and mottled red and

yellow on poorly drained sites, but numerous sections show grey or brown shades characteristic of local weathered volcanic rocks. Contacts between the hillslope deposits and sedentary weathered material are usually very sharp and well defined and they often rest on Zone IIa (Ruxton and Berry, 1957, p. 1269) of the weathered granite. No fossil soils have been found between the hillslope deposits and bedrock or between the two layers of such a hillslope deposit. Thus, if a soil had formed, sub-surface corrasion has been sufficient to remove it as well as an unknown amount of weathered bedrock. This outline account indicates the many points of resemblance between the hillslope deposits in Hong Kong and geliflual debris in New Zealand. The photographs used by Te Punga of “frost-riven fragments of greywacke set in a matrix of silty sand” are indistinguishable from many sections we have studied in Hong Kong. It is well known that deeply weathered igneous rocks and associated superficial deposits in humid tropical regions were first mistaken for boulder clays (Belt, quoted in Romanes, 1912, p. 133, in Costa Rica; Ansted, quoted in Kingsmill, 1862, in south China; Agassiz and Hartt, quoted in Derby, 1896, p. 530, in Brazil), because of their angular boulders, unsorted nature and clayey matrix. Moreover, as noted by Sharpe (1938, p. 56), “it is probable also that certain deposits resembling till may be in reality old mudflows, for both consist of heterogeneous or poorly sorted material and may contain striated pebbles or boulders.” Te Punga (1955, p. 1007) stated that “angularity of large fragments of rock and of many mineral grains cannot be explained other than as frost riving”. The hillslope deposits in Hong Kong often have these characteristics, however, and frost riving has not occurred. Dominance of simple mineral grains and their angularity are also common features of weathered debris in tropical semi-arid, savannah, and humid regions. Comminution of grains of quartz is intensely active in the migratory layer on the pediment slope at Jebel Qasim in the savannah region of the Sudan (Ruxton, 1958, p. 373). The proportion and grain size of the minerals in the hillslope deposits compared with that of the weathered and fresh source rock might be a significant factor in climatic correlation. Even then, with no additional evidence, a large proportion of unaltered minerals in a superficial deposit could be due to derivation in an arid climate or from a low weathering zone in a humid tropical climate. Most of the geliflual debris at Wellington was derived from deeply weathered greywacke. An earlier period of intense chemical weathering had already decomposed and disintegrated the rock, and it would be very difficult to recognise extra comminution and disruption caused by frost riving and heaving from that caused by the movement of a migratory layer. Mass Movement The slow downslope movement of rock waste saturated with water was termed solifluction by Andersson (1906). In the initial definition there was no climatic limitation to the term, but later it was restricted to movement occurring over a frozen subsoil. The latter, Baulig (1957, note 21, p. 926) now terms gelifluxion, and he implies that solifluxion should be used for all other conditions. Penck (1953, p. 107) claimed that flowing rock waste is widespread in frost-free regions, especially when the surface mantle is moistened. Peltier (1950) pointed out that mass movement of saturated or lubricated surface material should be common in several climatic zones, becoming a relatively important factor in denudation in periglacial, maritime, and selva (humid-tropical) landscapes. Saturation of surface layers because of impeded drainage is undoubtedly the greatest single factor causing gelifluxion in periglacial regions. Saturation of the surface layers is also common in areas receiving numerous short periods of very intense rainfall, when drainage to the subsoil cannot keep pace with precipitation.

Mohr and Van Baren (1954, pp. 41–42) showed that the amount of water precipitated by tropical cloudbursts is many times (about 40 times when comparing Indonesia and Bavaria) more than that in storms of temperate latitudes. In Hong Kong the following analyses have been made over a period of 55 years (Starbuck, 1950): Average during the year of 33 occasions of over ½ inch of rain falling in 1 hour; average of 8 occasions of over 1 inch; average of 3 occasions of over 1½ inches; and average of 1 occasion of over 2 inches. Falls of over 3 inches per hour occur on the average once in 5 years. Maximum rainfall intensity in one hour during 55 years, 1884–1938, 3. 965 inches. Maximum daily rainfall during the same 55 years, 21.025 inches. Mass movements in the tropics as a result of such violent storms will be facilitated by the presence of clayey slip planes, a high clay or silt content, and a poorly permeable substratum. The larger movements, slips, slumps, and flows, occur on moderate or steep slopes, and there is no doubt that in many cases sub-surface corrasion is caused by them. Very slow movement or creep is also widespread in tropical regions in various climates on slopes of even 1° or less.* We believe that a form of mass movement operates on clay-covered slopes of 1 in 200 in the Butana, Sudan. Patches of damp, saturated, and wet clay alternate along the the contours after a rainstorm and differential expansion of saturated montmorillonitic clay against the wet clay downslope is thought to cause movement (Ruxton and Berry, J. Soil Sci., 11 (1), March, 1960). This is shown by the widespread presence of a migratory layer of rock waste resting on sedentary weathered rock. The two are sometimes separated by a stone pavement (stone line in section, Sharpe, 1938, p. 24; though this by itself is not evidence of creep as claimed by Sharpe). Stheeman (1932, p. 5, Fig. 2) cited evidence in south-west Uganda for the relatively more rapid movement of his polymict layer (3–4 feet thick) over a slightly deformed monomict layer (sedentary weathered granite) on slopes of 4–6°. These findings would substantiate the “removal and renewal” doctrine of Penck (1953, pp. 62–65) which can be restated that surface removal of the soil and sub-surface renewal of the weathered debris by a lowering of the basal surface is a normal method of denudation which necessitates a surface layer of migratory material (Ruxton, 1958, p. 373). Ruxton has shown that creep occurs in the migratory layer over weathered granite at Jebel Qasim, Sudan (1958, Figs. 9, 10), to a depth of 2 to 3 feet on a slope of 1 in 50 in a semi-arid climate. The rate of creep in the migratory layer decelerates rapidly downwards, and it seems certain that no corrasion is caused by it. There can be no absolute distinction between (a) sub-surface corrasion† Sub-surface corrasion can be defined as the forcible incorporation of sedentary weathered rock into the base of a migratory layer—i.e., landslide or soliflual flow. The following are comparative rainfall data from Wellington (J. F. Gabites, 1960, “The Climate of Wellington” in “Science in Wellington”, pp. 33–35):—The amounts of rain in periods of 10 minutes, 1 hour, 12 hours and 24 hours liable to be exceeded only once in a year, once in 10 years, or once in 100 years: Amount of Rain (inches): Duration of Fall. Frequency 10 min. 1 hour 12 hours 24 hours Once in 1 year 0.3 0.6 2.1 2.6 Once in 10 years 0.5 0.9 3.5 4. 8 Once in 100 years 0.6 1.2 6.0 6.9 , and (b) sub-surface renewal of the migratory layer by permissive incorporation of sedentary weathered rocks following surface removal of material. Nevertheless each process may be related to a type of topographic environment. If a hillslope mantled with weathered debris is oversteepened at its base large scale mass movement and

sub-surface corrasion may occur. This leads to a renewal of exposure on the upper slopes. But on pediments, valley-side strips, and plains removal of surface material causes a downward extension of the surface influences which promote creep. Sedentary weathered rock is thus gradually incorporated into the migratory mantle. Removal from above is compensated by renewal from below. If corrasion is defined as a freeing of loosened rock fragments from their place of origin (cf. Penck, 1953, p. 112) then sub-surface corrasion can be said to occur under both active mass movement and slow creep. If corrasion is defined as the mechanical wearing of surfaces by rock waste in transit (cf. Malott, 1828, p. 158) then sub-surface corrasion will be practically confined to hillslopes. Convex Profiles in Headwater Streams Waters (1953, p. 72, fig. 2) noted that the breaks in slope on Dartmoor streams are composite and convex. An upper knick on unconsolidated gravel has migrated upstream more rapidly than a lower knick on the underlying granite. The convex knick thus owed its form to a break in the physical mobility (the ease of removal by mass-wasting or erosion) of the affected material. In a similar way, if a surface is mantled with a thick, complete, gradational weathering profile, then in general the physical mobility of the material will decrease steadily downwards, with a further sudden decrease at the basal surface (Ruxton and Berry, 1959). An old age surface developed in a humid climate usually bears an old age weathering profile with thick upper zones resting abruptly on the basal surface. After uplift revival of weathering below the former basal surface will thicken the lower zones, especially near the upland margins (Fig. 2). A thick mantle of upper weathering zones will then rest abruptly on a complex of lower zones (IIb, III and IV, Ruxton and Berry, 1957) and there will be a considerable break in the physical mobility between them. A convex valley profile may then develop around the upland margin, as the resistant lower zones hold up retreat, while the physically mobile upper zones allow the ready removal of debris and the development of a normal concave profile above the knick. On Dartmoor, Waters (1957) noted that, while the valley sides at the breaks of slope were steep and narrow, wide basins were formed above them. If the upland margin is much dissected, as in Wellington, New Zealand, headwater streams will rise on the margins of the upland remnants, starting their courses with a convex profile as there will be insufficient catchment above them to allow the development of the upper courses. In Hong Kong convex knick points and convex headwater streams are common on the weathered granite, volcanic, and Fig. 2. —The profile of a headwater stream and the axis of a dell-like hollow in relation to the zones of weathering on the margin of a deeply weathered upland surface. The diagram is an idealised version of relationships observed on weathered granite and weathered ignimbrite in Hong Kong.

sedimentary rocks. Above them the floor may broaden out into a basin form with a continuation of the stream, as in Dartmoor, or it may end with a dell-like hollow near a ridge top as in New Zealand. Dell-like Hollows Dell-like hollows in Hong Kong frequently have a swampy floor with no trace of a stream course. Artificial sections in these basins show an upper migratory layer resting on well-weathered sedentary debris sometimes with a stone line between them. The migratory layer is moving slowly downhill while its upper surface is being eroded by sheet and rill wash. Mass movements in Hong Kong, as elsewhere (cf. Cotton, 1958a, 1958b), tend to soften the relief and to coarsen the drainage texture. There is a sharp contrast between the morphology on the granite and on the volcanic rocks. The former weathers to a clayey sand which is remarkably stable, whereas the volcanic rocks yield a silty clay which is very unstable, and slips, slumps, and flows are common. The drainage texture on the volcanic rocks is much coarser and the surface contours much smoother than on the granite, though a layer of migrating rock waste occurs en both. There is a detailed adjustment to structure both in the weathering profile and in the surface form on both the granite and the volcanic rocks (cf. Cotton, 1943, p. 239). Cotton (1955, p. 1017) has described V-shaped gullies cut into deeply weathered rock which are now partly infilled by strips of peat swamp up to 10–20ft wide. The change from cutting to filling was assumed to be due not to deforestation but to some recent micro-climatological change. Similar features occur in Hong Kong. The floors of some side valleys are aggraded with a reddish-brown silty clay with a marsh vegetation, while the side slopes and nearby spurs are intensely and deeply gullied (Ruxton and Berry, 1957, Pl. 1, Fig. 3). The development of these features seems to be partly due to the advanced stage of the gully cycle and partly to deforestation. In the Kowloon Hills where gully development is extreme with only narrow walls of weathered debris between adjoining gullies, some of the larger gullies have aggraded floors with marshy bottoms. Here aggradation appears to be part of the normal cycle of gully development and destruction. Conclusion The products of mass movement in similar geological environments appear to be comparable in monsoonal and in periglacial conditions, and they must be used with caution in climatic deductions. Common factors in several areas are thick weathering profiles and conditions favouring saturation of the surface layers of rock waste. Geliflual corrasion by saturated thaw debris moving over the frozen ground in dell-like hollows in New Zealand is paralleled in Hong Kong by permissive incorporation of the sedentary weathered debris into the migratory layer by “removal and renewal”. But, whereas in Hong Kong intense chemical weathering at the base of scarp slopes is coupled with the progressive removal of the finer and more physically mobile material and collapse of the upslope debris to form hillslope deposits, it is difficult to understand how a slope can be undermined in such a way under periglacial conditions. It seems that deeply weathered areas under many climatic regimes are very susceptible to mass movements provided sufficient relief and concentration of water are available. The Massif Central (Beaujeu-Garnier, 1953); Devon and Cornwall (Waters, 1957); Wellington (Cotton and Te Punga, 1955); Portugal (Guilcher, 1949); Brazil (Tricart, 1958), and Hong Kong all have thick weathering profiles which are affected by mass movements. We are grateful to Professor Sir Charles Cotton, Professor W. J. McCallien, and Dr. L. Curtiss for their reading of the manuscript and their many helpful suggestions.

References Andersson, J. G., 1906. Solifluction, a component of subaerial denudation. J. Geol., 14, 91–112. Baulig, H., 1957. Peneplains and pediplains. Geol. Soc. America, Bull., 68, 913–930. Beaujeu-Garnier, J., 1953. Modelé periglaciaire dans le Massif Central Français Révue de Géomorphologie dynamique, 4, 251–281.Berry. L., and Ruxton, B. P., Evolution of the Hong Kong harbour basin. To be published in Ann. Geomorph. Cotton, C. A., 1943. Oahu Valley sculpture: a composite review. Geol. Mag., lxxx, 237–243. Cotton, C. A., 1958a. Alternating Pleistocene morphogenetic systems. Geol. Mag., 95, 125–136. — 1958b. Dissection and re-dissection of the Wellington landscape. Trans. roy. Soc. N. Z., 85, 409–425. — and Te Punga, M. T., 1955. Solifluxion and periglacially modified landforms at Wellington, New Zealand . Trans. Roy. Soc. N. Z., 82, 1001–1031. Derby, O. A., 1896. Decomposition of rocks in Brazil . J. Geol., 4, 529–540. de Chardin, P., Young, C. C., Pei, W. C., and Chang, H. C., 1935. On the Cenozoic formations of Kwangsi and Kwantung . Bull. Geol. Soc. China, 14, 179–205. de Terra, H., 1943. The Pleistocene of Burma, in “Research on early man in Burma ”. Trans. Amer. phil. Soc., 32, 271–340. Guilcher, A., 1949. Sur les vallons cotières suspendus des environs de Lisbonne. C. R. Acad. Sci., Paris, 9 mai 1949, 1512–1514. Heim, A. Krejci-graf, K., and Chan-san, Lee, 1930. Geology of Canton. Geol. Surv. Kwantung and Kwangsi, Canton Spec. Publ., No. 7. Kingsmill, T. W., 1862. Notes on the Geology of the east coast of China . Geol. Soc. Dublin, 10, 1–6. Li, Lien-Chieh, 1936. The Physiographic significance of the occurrence of red earths in the Nanning Basin . Bull. geol. Soc. China, 15, 529–552. Malott, C. A., 1928. An analysis of erosion. Proc. Indiana Acad. Sci., 37, 153–163. Movius, H. L., 1943. The Stone Age of Burma, in “Research on early man in Burma ”. Trans. Amer. phil. Soc., 32, 341–394. Peltier, L., 1950. The geographic cycle in periglacial regions as it is related to climatic geomorphology. Assoc. Amer. Geog. Annals, 40, 214–236. Penck, W. (translation by Czech and Boswell) 1953. Morphological Analysis of Landforms. London. Romanes, J., 1912. Geology of part of Costa Rica . Quart. J. geol. Soc., 68, 103–139. Ruxton, B. P., 1958. Weathering and subsurface erosion in granite at the piedmont angle, Balos, Sudan . Geol. Mag., 95, 353–377. — and Berry, L., 1957. Weathering of granite and associated erosional features in Hong Kong . Geol. Soc. America Bull, 68, 1263–1292. — 1959. The basal rock surface on weathered granitic rocks. Proc. Geol. Assoc., 70, 285–290. Sharpe, C. F. S., 1938. Landslides and Related Phenomena. New York: Columbia University Press. Starbuck, L., 1950. A Statistical Survey of Hong Kong Rainfall. Hong Kong Government Printer. Stevens, G. R., 1957. Solifluxion Phenomena in the Lower Hutt Area, N. Z. J. Sci. Tech., B 38, 279–296. Stheeman, H. A., 1932. The Geology of Southwestern Uganda. The Hague. Tricart, J., 1958. Division morphoclimatique du Brésil Atlantique Central. Rev. de Géomorph. Dynam. 9 (2). Waters, R. S., 1953. Aits and breaks of slope on Dartmoor streams. Geography, 38, 67–76. — 1957. Differential weathering and erosion on oldlands. Geog. J., 123, 503–509. L. Berry, M. Sc., F.G.S., Dept. of Geography, University of Khartoum. B. P. Ruxton, M.A., F.G.S., Dept. Geology, University College of Ghana .