Direct Effect of the 1855 Earthquake on the Vegetation of the Orongorongo Valley, Wellington*
[Received by the Editor, June 14, 1957.]
Over 100 years have passed since the great Wellington earthquake of January, 1855 This movement, resulting in block faulting as much as 9ft, violently affected the Rimutaka Range, causing large slips on the western escarpment of the Orongorongo river valley. There is evidence that thousands of tons of greywacke rock caused sudden and catastrophic damage to forest vegetation. The effect is still clearly seen in the present forest pattern and is traced in some detail for the Green's Stream flanking valley, from the scree slip down to the junction of the stream with the Orongorongo River.
At 9.30 p.m. on the evening of January 23, 1855, a violent earthquake shock was experienced in Wellington. It was felt over the whole of the North Island and by ships out at sea 150 miles away from the coast.
The shock covered some 360,000 square miles, but its effects were most violent in the immediate vicinity of Wellington. Ongley (1943) gives an account of the historical descriptions of the event. Movement and elevation occurred along a 125-mile fault trace running down the west side of the Wairarapa Valley. The greatest elevation was in the Rimutaka Range, a southern spur of the Tararua Mountains. This range was elevated nine feet, and the whole block of land eastwards through Port Nicholson was inclined. Elevation was progressively less, until, on the west coast, it measured about one foot.
A detailed account and map of the fault trace, as followed for 93 miles north, is given in Ongley's paper together with indirect evidence from landslides and cracked hills. In some cases these landslides are half a mile from the actual fault, although this does not indicate the measure of the shock so much as it indicates the readiness of the country to slip. Evidence is available of past movements along this same fault, and Ongley (1943b) shows further movement as recent as the Wairarapa earthquake of June 24, 1942.
An account by Roberts (1855) of the Royal Engineers, states: “The Rimutaka Range was very much shaken in its elevation, and a great many large slips occurred, laying bare the western side as well as on the eastern”.
The present paper is concerned with the large slips, undoubtedly caused by the earthquake on the western flanks of the Rimutaka Range, some ten miles from the coast and forming the eastern inclines of the Orongorongo river valley (Fig. 1). Here numerous landslides
[Footnote] * Portion of a Ph.D. study carried out under the tenure of a Senior Fellowship of the University of N. Z., for which grateful acknowledgment is made.
have formed scars, many of which are still visible from Wellington. Some are extensive slips 20 chains in length and 10 chains wide and have exposed shattered greywacke, fault pug and brecciated rock, with development of long and wide screes or fans and clay beneath the scarp. They appear to be due to sudden slipping rather than slow erosion processes. The vegetation is sparse as it has had to colonise the stable bedrock face or else the central unstable scree debris. A few grasses are present, but for the most part the plants are herbs—e.g., Raoulia, Taraxacum, Gnaphalium, Acaena, Epilobium, Wahlenbergia, Euphrasia, Pratia, the tall pampas grass Arundo, the ferns Polystichum and Pteris, together with small woody shrubs such as Coprosma, Cassinia and Helichrysum.
Stevenson (1945) has studied the recent erosion and weathering process of these scarps. She does not advance an opinion on the origin or age of the original slips. Some smaller shingle slips near the coast have apparently been the result of interference with vegetative cover by burning and overgrazing with exposure of the shattered greywacke which readily forms screes. Although rapid movement of small stones was demonstrated for the midzone of the screes, Stevenson found that there was a stable tongue at the foot composed of large lichen-encrusted boulders up to 4ft diameter. The size of the lichen colonies indicated an undisturbed period of ten years or more. However, she confines herself to the problems of weathering and development of current scree debris and does not enquire into the fate of the mass of original material, the slipping of which has formed the arcuate slump scarps.
Green's Stream System
One of the largest of the old slip systems is found at the head of the Green's Stream valley, the complete system of which is mapped in Fig. 2 (see also Fig. 8). Here, some 100 chains from the junction of the stream with the Orongorongo River and 2,000ft above it, the valley narrows to a steep slip face. The main slip, about 25 chains long and 5 to 10 chains across, is more than 100ft deep, and has many smaller contributary side slips.
At the lower end of the valley much of the floor is filled with three high terrace areas of boulders and rubble. Fig. 3 is a sketch map showing these three avalanche terraces (A, B, C). As will be seen from the side clevation plan, terrace C, the highest, is continuous in slope with the two smaller lower terraces. All are composed of stony material and are more or less elliptical in shape They abut on to the steep slopes of a loam hill side forming the valley flanks and supporting an over-mature Dacrydium/Metrosideros forest. Some 15ft lower than these terraces are found the true flood terraces where shoulders narrow the valley floor.
Terrace A at 200ft, which is approximately 9 chains long by 2 chains wide (200yds × 45yds), is inclined towards the stream mouth and drops some 6ft in elevation over its total length. On the same side of the stream at 500ft above sea level, and separated by a lower incutting flood terrace supporting a Leptospermum ericoides scrub 25–30ft high, is a second terrace, B, also inclined down the valley. The distal or lower end is formed by a 6ft bank, and although the terrace is not so long as A, it is as wide (150yds × 45yds). Still further upstream at about 650ft above sea level is a third len-shaped area of greywacke rubble some 25 chains long by 5 chains wide (550yds × 100yds). Its depth can be estimated by the height of the stream banks which are here cut down 25–30 feet into the debris material. Terraces B and C are known locally as “mahoe flats”, the entire tree vegetation being of uniform Melicytus ramiflorus with a closed canopy at about 40ft.
The Forest Pattern
The forests of the Orongorongo river valley are of mixed podocarp-ratabroadleaf forest on the lower and more favourable sites and Nothofagus beech confined to the dryer steep slopes and ridges (Fig. 2). The Green's stream terraces are adjacent to the podocarp forest. This forest was sampled by the forest profile method
as detailed for tropical rain forest, Richards (1952). A rectangular strip of forest 25ft wide by (preferably) 200ft in length is marked out, and all trees are drawn to scale on a graphed board while facing one side of the strip. Undergrowth species up to 10ft are ignored, and the results are recorded as a profile diagram showing a profile transect of the forest structure to scale. Each profile diagram is accompanied by full ecological notes.
In the Orongorongo valley the podocarp-broadleaf forest type is illustrated by Fig. 4. Structurally three tree strata may be recognised:-
An upper stratum of podocarp species showing a tendency towards grouping. The height is between 80–100 feet and the layer is almost continuous with the rimu (Dacrydium cupressinum) the most prevalent species.
A middle stratum forming a discontinuous layer between 40–80 feet and composed of broadleaf trees—Knightia excelsa, Elaeocarpus dentatus, Weinmannia racemosa.
A lower stratum forming a continuous stratum between 10–30 feet in height and composed of small trees and tall shrubs with greater variance in heights than those of the higher strata. The predominant species are Melicytus ramiflorus and Hedycarya arborea. In places there is a dense supporting growth of Pseudowintera axillaris, as shown in Fig. 4.
Epiphytes, lianes and tree ferns are conspicuous features. Below these three tree strata, recorded only on the ecological data sheets, is a layer of small woody shrubs and tree ferns and young seedling trees of the upper strata sparse in density and averaging between 3–4 feet in height. The field layer is composed of seedlings, ferns and graminoid herbs with mosses covering trunk bases, exposed roots and fallen logs.
However, there is little doubt that the lowland podocarp-broadleaf forests of the Orongorongo valley are in the process of a slow change towards northern rata (Metrosideros robusta) dominated forests. The rata is initially an epiphytic plant in the crown of the podocarps (most commonly the emergent Dacrydium cupressinum) and sends down long adventitious roots which entwine the host tree and finally reach the floor of the forest. In time the crown of the host tree is shaded out by the spreading growth of the more vigorous Metrosideros which eventually replaces the host completely. By this time the aerial roots of the Metrosideros have increased enormously in girth and encircle the trunk of the host and become self supporting.
Fig. 4—Three-tiered podocarp-broadleaf forest in the Orongorongo valley. Symbols: Dc, Dacrydium cupressinum; Wr, Weinmannia racemosa; Ed, Elaeocarpus dentatus; Ss, Suttonia salicina; Ha, Hedycarya arborea; Mr, Melicytus ramiflorus; Ps a, Pseudowinteria axillaris; Metr, Metrosideros robusta; Bt, Beilschmiedia tawa; Ke, Knightia excelsa.
Fig. 5—Rata (Metrosideros robusta) forest in the Orongorongo valley. For symbols see Fig. 4. This and the preceding diagram represents a strip of forest 100ft long and 25ft wide. Only trees 10ft and over are shown.
Such rata forest is shown in Fig. 5, but most of the forest in the Orongorongo valley is of mixed transitional forest between Figs. 4 and 5. In the Metrosideros forest where Dacrydium has been replaced in situ by the slow ascendancy of the rata, the lower strata are similar in composition to the original podocarp/broadleaf forest. This development of Metrosideros forest is linked to the general decline of the podocarp element in New Zealand mixed forests and is merely a prolonged phase in the trend towards a replacement of podocarp forest by broadleaf forests at present evident in the New Zealand forest pattern, Robbins (1957).
A profile along terrace A shown in Fig. 6 is distinctly anomalous. This shows a canopy of tall podocarps between 60ft and 100ft, while the middle stratum is virtually lacking and the lower stratum represented by a continuous and very uniform layer at 25–30ft dominated by Melicytus ramiflorus (mahoe) and Hedycarya arborea. Small shrubs and ground ferns are well represented, but tree ferns are conspicuously absent. The podocarp trees are of small diameter and youthful form, being perhaps 200 to 300 years of age. Three species are represented but Dacrydium is absent.
As there is evidence in New Zealand for a seral relationship towards Dacrydium dominance this profile indicates a relatively youthful podocarp forest. It therefore is not comparable with the adjacent overmature podocarp/broadleaf forest. At the time of the earthquake it would be a young podocarp forest stage pioneering the original high river terrace on the shoulder of the Orongorongo River. Similar pioneer podocarps now 40ft high may be seen elsewhere in the valley. However, the entire absence of middle stratum species represented by Knightia (Ke), on the extreme right of the profile at the base of the loam soil hillside, even to lack of young stages in the lower strata needs further explanation. There was every indication that the original undertiers had been destroyed and that the uniform Melicytus layer was a later stage not contemporaneous with the podocarp canopy trees This was supported by the observation that the bases of the podocarp trees were buried in stony rubble.
Profile diagrams made on terraces B and C (Fig. 3) are seen in Fig. 7. It will be noted that these correspond remarkably with the uniform substratum found on terrace A but are some 10 feet higher, due no doubt to the lack of overstorey podocarp trees.
From the description of the valley system so far presented and the present correlation of the forest pattern it is suggested that the large slip areas at the head of the valley were of a sudden and catastrophic nature. Thousands of tons of greywacke rock must have poured down and levelled out in the lower reaches of the valley, destroying the stream bed vegetation and filling much of the valley floor. The present stream has cut down through this material, leaving the three avalanche terraces A, B and C. The geological evidence supports such a contention. All the debris consists of angular material, and there is no evidence of water-sorted grading or stratification. The vast quantity of debris, the decreasing size and continuity of slope and levels of the three terraces, together with the undulating form of the rubble fan to be seen in the largest terrace, would all support the view of simultaneous arrival of the debris rather than a slow accumulation by water transport. The decrease in average size of stones making up the debris is trom 8in by 12in on terrace C (which also has boulders up to 4ft × 6ft), down to 4in × 3in on terrace B and 2in × 3in on terrace A. The uniform stands of Melicytus covering the largest terrace and terrace B suggests that they represent a single unitorm generation of invading tree species pioneering the new stony flats. These stony terraces are distinct from the low flood terraces of river-washed gravel at present covered with dense Leptospermum ericoides scrub. These latter communities are well-established, and a ring count of over 50 years is average. There is no evidence of a seral trend towards Melicytus on these low flood terraces nor is there any evidence of a previous seral stage of Leptospermum within the Melicytus communities. Thus there is every indication that the Melicytus cover on eacn of the terraces represents the mature phase of direct colonisation after the original devastation.
If this is so then the anomalous lower stratum shown in Fig. 6 for the profile of terrace A can be explained. The unitorm Melicytus layer here is contemporary with that of the other terraces (Fig. 7) but has developed under a higher stratum of podocarps representing the survivors of the original forest. Examination of the podocarp trunks indeed revealed that they had been buried for a matter of several feet in stony rubble. If, as suggested, the terraces were formed from avalanche material, it would appear that here, near the junction of the stream with the Orongorongo River, the avalanche had spent most of its momentum and the whole mass was less devastating than higher up the valley where the stones were larger and the mass and momentum greater. Here, on terrace A, only the understorey trees and shrubs were destroyed, leaving the podocarps as survivors now out of phase with the present regenerating lower strata of Melicytus.
The highest terrace, some 50 chains in a direct line from the slip scarp (Fig. 8) and about 10 acres in area, forms the main mass of the avalanche, with some 30 feet depth of large greywacke rubble and boulders. Here the debris overwhelmed the original vegetation and filled in the valley floor to a level of 30 feet against the loam hillside flanking the valley. The aerial photo (Fig. 8) revealed two small groups of emergent trees on the sides of terrace C, which are marked a and b on Fig. 3. Ground reconnaissance showed that “a” was a higher spur belonging to the eastern bank but now cut off by the stream and forming a clay-loam “island” surrounded by stony rubble. It carries two or three badly formed podocarps. “b” is a similar spur belonging to the western flanks and extending out into the boulder-filled valley. These are remnants of the original forest which escaped destruction. Inside the forest on terrace C there is a sharp transition between the terrace floor vegetation and that on the flanking hill slopes. The dense stand of uniform Melicytus is confined to the boulder substratum and abuts onto a mature podocarp/broadleaf
Fig. 8—Aerial photo of the Green's Stream area as mapped in Fig. 2. The stream valley is in the centre originating from the slip scars near the summit of the Rimutaka Range (2,450ft) and descending within 100 chains to its junction with the Orongorongo River, part of which is seen in the top left-hand corner Terrace A is not shown, terrace B is at extreme top edge, and terrace C is outlined Scale 10 chains to 1 inch
Fig. 6.—Profile diagram of forest on terrace A, Fig. 3. Note the absence of a middle stratum and the uniformity of the lowermost stratum. Pf, Podocarpus ferrugineus; Pt, P. totara; Ps, P. spicatus. For other symbols refer to Fig. 4.
Fig. 7—Profiles of Melicytus communities on terrace C (left) and terrace B (right). See Fig. 3. Pe, Pittosporum eugenioides. For other symbols refer to Fig. 4.
forest on the hillside with the loam soil. This forest has scattered Dacrydium and Metrosideros under which Elaeocarpus and Knightia form a canopy and nikau palms, treeferns and various shrub species—e.g., Olearia rani, Macropiper excelsum, Hedycarya arborea, Coprosma australis, and Melicytus ramiflorus make up a subcanopy. Lianes are represented by Rhipogonum scandens and Rubus, while seedlings of Alectryon excelsum and Knightia excelsa are frequent. A limited ecotone is seen in this shrub layer with species extending reciprocally from both the mature forest and the Melicytus terrace communities.
Further evidence for the sudden slipping of the higher slopes is found further up the valley. Here one of the large subsidiary slips has a peripheral forest of large mature Weinmannia racemosa trees 40 feet in height. Many of these are now dead standing trees and their death has been attributed by wildlife officers to the depredations of opossums. However, the dead trees occur in a wide zone around the edge of the slip, and this suggests an abrupt change in soil drainage such as one brought about by a sudden slip. While some of these minor slips may have been brought about by cloudburst or other normal erosion factors, there is much to connect the forest catastrophe in the Green's Stream valley with the great Wellington earthquake of 1855.
In comparing the Melicytus stands on the three terraces some differences are apparent (Figs. 6 and 7). On both B and C terraces the Melicytus stands are dense with a closed canopy at 40 feet. On terrace C there is a greater abundance of massive asteliad epiphytes. The trunks are more thickly covered with corticolous mosses and a mixed understorey of shrubs and seedlings is becoming evident. Trees up to 12 inches diameter are common on terrace C, whereas the average diameter for terrace B is 9 inches.
On terrace A the Melicytus layer includes many Hedycarya and the trees are not so dense and their height is reduced to about 30ft. Asteliads here are confined to their true position above 50 feet in the podocarp stratum. Many plant species are represented in the lower shrub layers. Average diameter measurement for Melicytus is reduced further to 6 inches. These differences, however, may all be due to the difference in the size in area of the terraces and the grading of the stony rubble. The lower and sparser Melicytus layer on terrace A is no doubt due to its development under the podocarp layer. The differences in diameter of the trees on the three terraces is due to differences in growth rate as ring counts have not revealed any great difference in age, but just why this is so is not yet clear. Ring counts were made from a single average tree on each of the terraces and, assuming that the rings are annual, resulted in the following:–
Terrace A, 76 years; terrace B, 74 years; terrace C, 92 years. Such counts are difficult to make on Melicytus, and all are plus or minus 5–10 years. The average of these is 81 years, and this figure thus correlates reasonably with the 1855 earthquake period, allowing 10 to 20 years for Melicytus colonisation of the boulder terraces to commence. At first the area would be extremely dry and open, and the first plants would be the semi-woody Coriaria and Cassinia and the pampas grass Arundo, as occur on some of the present recent flood terraces in the area. However, none of these stages can be traced within the present Melicytus terraces.
The Green's Stream valley thus provides an interesting record of dated recolonisation on bare ground and of the regeneration of substorey species within a former forest.
The writer is indebted to officers of the Wildlife Division (now under New Zealand Forest Service), for facilities at the Opossum Research Station which forms part of the area under study.
Ongley, M., 1943. Surface trace of the 1855 earthquake. Trans. Roy. Soc. N. Z., 73. 2: 84–87.
—— 1943b. Wairarapa earthquake of June 24, 1942, together with map showing surface traces of faults recently active. N.Z. Jour. Sc. Tech, 25, 2: 67–78.
Richards, P. W., 1952. The Tropical Rain Fores.t Camb Univ. Press.
Robbins, R. G., 1957. The classification and status of New Zealand forest vegetation Ph.D. thesis. Univ. of N.Z.
Roberts, E., 1855. In R. Taylor's “Te Ika a Maui,” pp. 471–2.
Stevenson, G. B., 1945. Note on the movement of waste on screes in the Orongorongo district, near Wellington. Trans. Roy. Soc. N.Z., 74. 4: 315–319.
R. G. Robbins, Ph.D.,
Land Research and Regional Survey,