Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 77, 1948-49
This text is also available in PDF
(2 MB) Opens in new window
– 153 –

Some Post-Glacial Climatic Changes in Canterbury and Their Effect on Soil Formation.

[Received by the Editor, April 21, 1947; issued separately, April, 1948.]





Introduction, including Summary of Literature.


Evidence bearing on Climatic Changes.


Recent Lowering of Uppermost Limit of Close Vegetative Cover.


Recent Lowering of Timber Line.


Pattern of Soils on Canterbury Plains.


Pattern of Soils on Canterbury Downlands.


Relationship between Soil Types and Present Climates.


Nature of Recent Changes in Climate.


Dating of Climatic Changes.






Anomalies between vegetation and soils on the one hand, and present-day climates on the other, are discussed in their application to the problem of climatic changes. The climates now prevailing over regions where pedalfers (soils from which the alkalis and alkaline earths have been leached) exist, do not resemble climates associated with similar soils abroad: they resemble rather climates associated with pedocalcic soils (soils with accumulation of lime). It is suggested that this may be explained by the hypothesis that the soils in question have suffered several climatic fluctuations and spent part of their life under a wetter climate.

The nature of the latest climatic changes is discussed and the hypothesis is advanced that forests flourished on the downlands of Canterbury and on parts of Otago now devoid of forest, between the seventh and thirteenth centuries. At this period, it is suggested, the temperature was at least 2°C. higher than it is at present.

The soils of Canterbury and Otago would therefore appear to be related to a complex climatic factor. The theory is therefore advanced that soil profiles in these regions should be regarded as complex and not in equilibrium with present climates.


The importance of climate in soil formation makes it desirable to know whether the present soils are related to climates which have remained constant in the past or to climates which have fluctuated between extremes of rainfall and temperature. The view that there have been significant fluctuations in post-glacial climates in New Zealand has been advanced by several writers, notably Hutton, Speight, Cockayne, and Cranwell and Post.

Hutton (1892) claimed that the moa remains could not have been washed into the swamps of Central Otago and the Glenmark swamp in North Canterbury under a climate as arid as that prevailing now. He suggested that the bones accumulated when the climate was wetter.

– 154 –

Hardcastle (1908) claimed that forest remnants on the South Canterbury downlands indicated a climate wetter than the present.

Speight (1910) pointed out that genera related to the moa (ostrich, rhea, and emu), now inhabit dry regions, and that the moas probably established themselves in New Zealand under a steppe-like climate, and that a change to a wetter one brought about their extinction. The same writer (loc. cit.) summarised previous literature and collected evidence for the existence and distribution of ancient forests in regions where forest does not now thrive, and after a careful study of the climate which supports the present totara forests, concluded that prehistoric forests established themselves under a climate much wetter than the present. This contention was based on the supposition that totara requires at least 40in. of rainfall annually. Speight, also, from a study of the remains of an ancient forest near Christchurch, where the fallen logs and stumps were found buried under upwards of 12ft. of alluvium, concluded that the destruction of this forest was the result of a widespread change of climate.

Cockayne (1905) from a study of the xerophytic plant Discaria toumatou claimed that at some stage of its previous history it had been adapted to a wetter climate and that it retained a latent ability to respond to a wet atmosphere. The same writer (1928), discussing the totara remains, concluded with Speight that “The climate where these trees lie is distinctly too dry for the natural occupation by totara forest, and I can only conclude with Speight that the forest came into existence during a much wetter period than the present.” Cockayne also considered that a previous cycle would explain the rarity of Dacrydium cupressinum on Banks Peninsula, where the wet post glacial period led to the replacement of a primitive Nothofagus forest by one with D. cupressinum dominant, and the latter during a subsequent drier period by a Podocarpus totara forest.

Cranwell and Post (1936), from a study of pollen diagrams from the South Island, divided the post-Pliocene climate into three divisions:

1. A severe climate, when vegetation was mainly grassland, such as that to-day confined to the Sub-antarctic or montane regions, probably representing the last epoch of glaciation.

2. A wet period when podocarp forest was widespread, mainly without Nothofagus, but sometimes with Metrosideros, a forest expressive to-day of a very wet and almost warm and equable climate.

3. A period beginning with a cold spell, and giving way to the present colder and drier climate.

Ferrar (1905 and 1925) reported old moraines high above the level of the Ross ice-sheet in South Victoria Land, proving that the present ice level represented a retreat from a previously higher level. The high moraines were recorded as far back as 1900, and there have been many records of recession since then. In the Ross Dependency, high moraines are stranded more than 1,000 ft. above present lowlying moraines. In addition, bare roches moutonnées, valleys free from ice, and valleys containing ice slabs or remnants of glaciers with

– 155 –
Picture icon

Diagram I.

– 156 –

no visible source of supply, are regarded by Ferrar as proving a general progressive retreat of ice. Ferrar argued that since most of the snow which falls in Antarctic regions falls into the sea, and most of the air which circulates south of the isotherm of 0°C. is comparatively dry, one way of increasing the snowfall and therefore thickening the ice covering in South Victoria Land would be a shift of the isotherm of 0°C. some 5° or 10° of Lat. farther south, either during successive summers or for a continuous period of longer duration. Among the reasons which he advances to explain this movement is a general increase in the mean temperature of the Southern Hemisphere. He conjectured that a recent desiccation of the Ross Dependency should be reflected in New Zealand recent deposits.

A changing climate leaves its mark on vegetation, soils, and the sedimentary record, and all three were examined for evidence of adjustment in the past to a different climate from the present. The evidence for climatic fluctuations has been considered under the following heads:


Recent lowering of uppermost limit of vegetation.


Recent lowering of timber line.


Pattern of the soils on the Canterbury Plains.


Pattern of the soils on the Canterbury downlands.


Relationship of soils to present climates.

Evidence Bearing on Climatic Changes.

Recent Lowering of Uppermost Limit of Close Vegetative Cover.

There is evidence that the uppermost limit of vegetation once stood at a higher level on the Canterbury and Otago mountains. Between latitudes 44° S. and 45° S. a elose cover of vegetation ceases at approximately 6,000ft. (Gibbs and Raeside, 1945). Even where accelerated erosion has extended the scree downslope, residual patches of the original tussock cover cease at this altitude. Above 6,000 ft., even where there has been little interference with the vegetation, only scattered alpine plants persist. In a few regions, as, for example, the northern and more inaccessible parts of the Two Thumb Range, where stocking is light and fires infrequent, some fingers of a close tussock cover reach 6,500 ft., but are clearly no longer in equilibrium with their environment and are steadily being driven back by encroaching scree. Regeneration above 6,000 ft. seems to be now insufficient to sustain the close cover against the ravages of active geological erosion. These facts seem best explained by the assumption that the remnants of close cover above 6,000 ft. are survivals from a previous climatic cycle when the limit of a close cover stood somewhat higher than it does to-day.

It has been observed also that from 6,000 ft. to 6,500 ft. traces of what appears to be a yellow subsoil similar to that at present covering the slopes from 5,500ft. to 6,000ft. still persist, mixed with the rocky debris. An analysis of this material collected from an altitude of 6,000 ft. on Michael's Peak in the St. Bathans Range in Central Otago is shown in Table I (analysis 2166). These figures are consistent with the material having once been a subsoil of a strongly leached Kaikoura soil (Gibbs and Raeside, 1945).

– 157 –
Table I.—Soil Analyses.
Description. Sample No. Depth In. 1% Acid P2O5 Citric Sol. K2O pH C N C/N Base Exch. Cap. Total Exch. Bases. % Base Satn. Replaceable Bases CaO MgO
Alexandra silt loam 2232A 0–5 .032 .032 6.9 2.1 .18 12 11.2 9.2 83 7.6 1.2
from Cromwell 2232B 6–9 .027 .012 6.5 10.8 8.0 75 7.1 1.1
Kalkoura silt loam 2222A 0–3 .012 .008 5.2 1.4 .10 14 12.3 1.4 11 0.6 .5
from Manorburn dam 2222B 3–6 .021 .017 5.1 4.4 .25 17 20.0 4.6 23 3.2 1.5
Forest soil from 2618A 0–6 .014 .024 5.6 4.3 .34 13 16.0 8.3 52 6.1 2.2
Geraldine Downs 2618B 8–13 .013 .012 5.3 12.5 4.1 33 2.8 1.3
Grassland soil from 2412A 0–6 .009 .012 5.8 4.1 .33 12 15.8 8.2 52 5.2 2.9
Geraldine downlands 2412B 9–12 .006 .006 5.9 10.2 3.7 36 2.4 1.4
Soil from 6,800ft. on St. Bathans Range 2166 0–3 .046 .006 5.5 .3 .04 8 7.4 2.2 30 1.8 .5
Podsol soil from 1617A 0–4 .057 .038 4.8 9.65 .61 16 29.9 6.8 23 5.1 1.4
Tasman Valley 1617B 5–8 .026 .022 4.9 18.7 1.2 6 .7 .25

The evidence presented by the remnants of the ancient plant cover is not particularly clear-cut, but it would seem to suggest an ancient rise of at least 500ft. and to agree well enough with the evidence from the timber line.

Recent Lowering of Timber Line.

The existence of totara logs above the present timber line also suggests that during the forest period the temperature was somewhat higher than it is at present. Burnett (1926) mentions that logs of totara were found at an elevation of 4,000ft. above sea level on the Tasman Downs, lying to the south-east of Mt. Cook Station in the north-western Mackenzie Country. The vegetation of these downlands is now tussock grasslands, and there is no trace of totara forest. The timber line for this latitude is about 3,500ft. The nearest mixed podocarp forest lies some 60 miles to the east and the forests of Westland lie to the west of the main divide of the Southern Alps, here over 10,000ft. high. On the Kirkliston Range, which bounds the Hakataramea Valley on the west, the writer has noted logs of totara approximately 4,000ft. above sea level. Park (1908) records totara remains on the Pisa Range near Cromwell, where rainfalls are now less than 20in. Forest remnants high above the timber line on Mt. Tapuaenuku in the Inland Kaikoura mountains of Marlborough were mentioned in an account of the expedition of Lieutenant Governor Eyre to this mountain in 1849 (N.Z. Alpine Journal [1895] quoting New Zealand Spectator for 24th November, 1849). The altitude stated for these remains is 6,000ft., but this may be an overestimate.

Forest remains above the present timber line can be best explained by the hypothesis that at some time in the past the temperature was high enough to raise the timber line at least 500ft. higher than it is now. Such a rise would require a temperature increase of between 1½° and 2°C. This harmonises with the evidence of a recent lowering of the uppermost limit of vegetation.

Pattern of Soils on South Canterbury Plains.

The terrace structure and soil pattern of the South Canterbury plains reflect the climatic and vegetational changes which have followed

– 158 –

the last ice retreat. The supply of sediments to run-off waters has not been constant during this period. Times of plenty alternated with times of scarcity, and rivers alternately aggraded and degraded their beds. An increase in the burden of waste indicates increased erosion and a decrease suggests a soil surface adequately protected from erosion. It is suggested that periods of accelerated erosion were the consequence of climatic changes and that erosion took place as the vegetation adjusted itself to a new climate. Thus a fall in average temperature might be expected to expose a considerable area to soil erosion as the uppermost limit of vegetation receded down the mountainside.

During periods of climatic stability therefore, erosion was at a minimum and streams spent their energy in degrading their beds. During periods of adjustment, when a new cycle of erosion was beginning and the supply of waste was copious, aggradation took place. Cotton (Landscape, p. 168) states that: “Reduced rainfall induces aggradation by increasing the proportion of waste to water in streams: the effect is twofold, for not only is there less water, but almost certainly there is more waste to be transported, as forests may die out, or, at least, become less luxuriant and so protect the ground less effectively under drier conditions.”

The terrace structure in South Canterbury is best seen in cross sections of the river courses some miles inland. Further inland still terrace structure may be related to tilting or uplift rather than to climatic changes. No exact correlation between terraces in Canterbury and erosion sequences abroad would be detected (Zeuner, 1946). Only the terrace pattern immediately upstream from the ponding reaches near the coast has therefore been considered and will be discussed in its relationship to the soil pattern developed on it.

The oldest soils of the plains (those of the Waitohi series) have developed on a fine homogeneous light yellow silt which resembles the loess of the downlands very closely, but which contains many small greywacke and quartz pebbles. It appears to have been spread out fairly rapidly while the loess was accumulating on what was then the partly dissected surface of an ancient piedmont gravel plain which is tilted eastwards a few degrees and slightly uplifted. The South Canterbury loess contains very little organic matter (Wilde, 1919), and in this respect differs from most other deposits of loess. It seems probable therefore, that the land on which the loess fell supported a scanty vegetation, that the climate was dry, and that the insufficiently protected surface was actively eroded with each rainstorm. In this way vast quantities of silt would be washed on to the surrounding plains, remnants of which now remain as the highest terrace levels on which the Waitohi soils have developed. Diagram Ia shows an ideal sequence of alluvial soils and sediments.

The high terrace at Waitohi, some three miles north of Pleasant Point, from which these soils take their name, may be taken as a good example of this high terrace level. Twelve miles inland from the mouth of the Opihi River the step between this terrace and that immediately below is approximately 40ft. in height,

– 159 –
Picture icon

Diagram Ia.

and suggests prolonged degrading, although there is no evidence to show at what period in post-Waitohi times this took place, for the soils on this lower terrace (designated the Orari series) are much younger. The Orari soils are only slightly leached and show a barely perceptible clay shift, and their youth suggests that a long period elapsed between Waitohi and Orari times, during which it is unlikely that degradation continued without interruption. It is more likely that the Orari sediments conceal evidence of a good deal of post-Waitohi climatic history, although much of the evidence may have been obliterated by subsequent degrading and lateral planation.

Some light is thrown on the climate immediately prior to Orari times by the structure of the Canterbury Plains between the mouths of the Orari and Opihi Rivers. This region consists of an expanse of shallow and discontinuous peat swamp containing forest remnants, invaded by gravel-cored levees of Orari age. Silts associated with the peats are uniformly fine, and appear to be derived from a land surface on which erosion was slow, i.e., a surface well protected by vegetation. The uniform texture suggests also that the circumstances governing the supply of the small regular amount of silt to the streams remained constant, while the forest remains link it with the forest period.

– 160 –
Picture icon

Diagram II.

– 161 –
Picture icon

Diagram III.

The end of this period of stability is marked by the invasion of the low-lying area, now covered with poorly drained soils, by a well-defined levee of coarser sediments (Diagrams II and III), which on field evidence have been placed in the Orari series, and which were apparently deposited by the stream formed by an ancient confluence of the Orari and Waihi Rivers (Diagram II). This levee reaches almost to the present coastline, and shows signs of having been built fairly rapidly. After the Orari had returned to its present course, it built a further levee (Diagram III), the margin of which can be seen to the south of the stream, where it has escaped burial by younger alluvial deposits.

Sediments younger than those of Orari times have accumulated mostly since 1850. At this date streams were deeply entrenched in Orari sediments and were engaged in degrading them further. There was no sign of any aggradation, and early settlers found the Orari River, for example, entrenched to a depth of upwards of 20ft. in “clay” banks. It may be inferred, therefore, that the period immediately preceding the middle of the nineteenth century, like the forest period, was a period of stability when the vegetative cover was able to protect the surface from accelerated erosion. This vegetative cover consisted mainly of tussock grassland. Aggradation as we know it to-day began only when the first tussock and bush fires were lit on the ranges. The Tripp-Acland holocaust (Burdon, 1938), which swept the catchment districts of the Orari and Waihi Rivers, and by the light of which Acland claimed that the finest print could be read some distance away on the plains, was followed by a flood of heroic dimensions (that of 1868), and a new cycle of erosion was started. Equilibrium might have been reached through the medium of a new plant cover, but regular fires and grazing steadily whittled away such cover as still remained. The rivers have therefore continued to aggrade until now the channel of the Orari River is almost filled with shingle, which has in some places spread beyond the actual confines of the river bed.

The sequence of sedimentary and vegetational changes may be summarised as follows:—


The degrading immediately prior to settlement took place during the period when the dominant plant association on the South Canterbury downlands was tussock grassland.

– 162 –

Prior to the degrading there was a period of active sedimentation during which the Orari sediments were laid down.


This sedimentation in its turn was preceded by a period when the supply of waste was small, constant, and fine in texture, the vegetation consisted of forest, and the climate was warm and moist.


Between the forest period and Waitohi times, the sedimentary record is obscure and unreliable.


The oldest deposits of the plains, the silts which appear to be closely related to the loess of the downlands, suggest that while they were accumulating the downlands were sparsely covered with vegetation and the climate was dry.

Pattern of Soils on South Canterbury Downlands.

The liberal sprinkling of totara remains (principally logs) over the South Canterbury and Eastern Otago downlands (Diagram I) suggested to the first settlers that these downlands had formerly supported a forest cover (Buchanan, 1868; Hardcastle, 1908; Speight, 1910). On slopes too steep to plough, dimpled surfaces still testify to this ancient forest, of which a small remnant still survives on a hill behind Geraldine township under a rainfall just under 30 inches annually. This patch of forest shows no sign of advance, even where it is adequately protected, and would seem to be no longer in equilibrium with its environment, if 40 inches of annual rainfall is accepted as the minimum requirement of totara (Speight, 1910).

A comparison of the soil under the forest with that under grass on the adjacent downlands shows a close physical and chemical resemblance between the two profiles (Table I, Analyses 2618 and 2412). The principal field difference is in the topsoil colour, which changes gradually from the light brownish-buff of the forest top-soil through darker modifications in soils lately cleared of forest to the characteristic medium greyish-brown of the downlands which were under native grassland when the first settlers arrived. The soils of the eastern margin of the downlands, mapped as Timaru silt loam, may therefore be regarded as closely allied to that now carrying native forest on Geraldine Hill. The change in topsoil colour appears to be due to the darkening of an older forest topsoil by the grassland humus.

The origin of the clay pans of Timaru silt loam may possibly be related to climatic and vegetation fluctuations. A clay pan has been observed to form in a forest brown loam in Southland after the clearing of forest by settlers and its replacement by grassland (A. C. S. Wright, personal communication).

From Distribution of Soil Types.

The soil-climate relationships of the intemontane basins of Otago and Canterbury, when compared with the corresponding relationship observed in Europe and America, show certain discrepancies

– 163 –

which can best be explained by the hypothesis that in the past the soils have suffered a climate substantially different from that now prevailing.

Over large tracts of the montane semi-arid regions of Canterbury and Otago, where soils are derived from either greywacke or mica-schist and vegetation consists of low-tussock grassland, the uniformity of these two soil-forming factors might be expected to throw climate into relief.

The effect of a given climate in forming a soil is the sum of the effect of rainfall, its intensity and distribution, together with temperature and humidity. In assessing the part played by water, it should be remembered that it is the amount of water that percolates through the soil and not the amount precipitated on the surface which governs the soil-forming process. Rainfall figures expressing the total annual rainfall are not sufficient to define the amount that percolates through the soil, which requires for its definition figures covering temperature as well as distribution of rainfall and humidity. Several expressions have been proposed, notably Meyer's N-S (Niederschlag) Quotient and Thornthwaite's Precipitation- and Temperature-effectiveness ratios (Jenny, 1941).

The above indices have been calculated for a number of stations in Canterbury and Otago (Table II), and are listed again in Table III, where the observed relationships between climate and soil in Europe and America are shown, and also the corresponding relationships in New Zealand. From the table it will be seen that there is exact correspondence only where rainfalls are high, that is, in the podsol zone. At the arid end of the climatic scale there is a discrepancy, and the New Zealand soils are more leached than those existing under the same climate in the Northern Hemisphere.

Table II.—Climatic Indices.
Altitude Lat. Rainfall Rain Days Mn. T. Mn. Hum. Meyer A B
Alexandra 520′ 45°22′S 13.11 97 50.4 70 112 25 55
Ophir 1,000′ 45°07′S 16.11 96 48.8 71 160 33 50
Waipiata 1,550′ 45°14′S 18.09 104 48.3 71 184 39 44
Lake Coleridge 1,220′ 43°20′S 31.45 105 51.1 62 221 68 57
Lake Tekapo 2,350′ 44°00′S 22.53 79 47.0 68 196 54 45
Waihopai 860′ 41°40′S 32.60 115 52.6 64 230 67 65
Manorburn Dam 2,448′ 45°22′S 20.57 129 43.5 72 262 51 35
Hanmer Springs 1,225′ 42°33′S 45.04 131 50.1 71 483 105 57
Queenstown 1,110′ 45°02′S 30.41 98 50.1 72 273 66 55
Hermitage 2,510′ 43°45′S 161.41 140 46.6 64 1,405 466 44
Fairlie 1,000′ 44°06′S 28.21 111 49.3 75 331 69 55
Waimate 200′ 44°44′S 25.22 130 51.5 74 262 51 59
Gore 245′ 46°06′S 34.75 180 49.6 83 711 75 50
Westport 23′ 41°45′S 77.46 185 53.5 88 1,500 173 64
Timaru 56′ 44°22′S 23.00 115 51.8 76 244 46 59

Lat. = Latitude.

Mn. = Mean.

T. = Temperature,

Thornthwaite A = Thornthwaite PE Index.

Thornthwaite B = Thornthwaite TE Index.

Hum. = Humidity.

– 164 –

Diagram IV shows the comprehensive soil-climate relationship according to Thornthwaite's two indices and Table III according to Meyer's NS Quotient. The same discrepancies between New Zealand and Europe and the United States of America show up, with correspondence only in the podsol region (Analysis 1617).

Unfortunately, data from which climatic indices can be calculated are scant, and are collected at only a few stations. Many discrepancies, therefore, which were noted in the field can only be roughly documented from annual-rainfall figures and the annual number of rain days.

Alexandra, the driest station for which adequate climatic records are available, with a rainfall of 13 in., a PE index of 25, a TE index of 55, and an NS Quotient of 112, falls into the middle of the zone of Chestnut soils, yet soils other than alluvials developed in the neighbourhood (Table I, Analysis 2232) have none of the characters of pedocals, but are unmistakably pedalfers, and as the rainfall increases, show a gradual transition with increased leaching to podsols. The corresponding soils of Waipiata and Ophir, whose climates fall in the middle of the Chernozem zone, are also pedalfers rather more leached than the soils of Alexandra. The Manorburn Dam (Analysis 2222), Lake Tekapo, and Christchurch climates are spread through the Prairie Earth zone, but soils other than young alluvials developed under these climates are also much more leached than the Prairie soils, although the discrepancy is not so marked as at the arid end of the scale.

Picture icon

Diagram IV.

– 165 –

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Table III.—Classification of Climates according to Meyer's NS Quotients.
N.S. Quotients, Annual Values.
World Soil Groups. Europe. U.S.A. N.Z. Soil-Climate Relationship.
Pedocals Deserts & desert steppes. Grey 0–100 30–110
Steppe Soils
Light brown steppe soils 60–120 Alexandra (112)1 Pedalfer
Chestnut-brown steppe soils 140–170 100–180 Ophir (160)
Alexandra (112)
Ophir (160) Pedalfers
Waipiata (184)
Chernozem soils 130–250 140–250 L. Takapo (196) Moderately leached pedalfers
L. Coleridge (221)
Limit of Arid Climate ca. 200 220–250
Pedalfers Waihopai (230) Moderately leached pedalfers
Manorburn Dam (262)
Degraded 250–350 Christchurch (262)
Chernozems Manorburn Dam (202)
Waimate (262)
Queenstown (273)
Fairlie (331)
American prairies 260–350
Brown earth 320–460
Brown forest soils 280–400
Hanmer Springs (483) Strongly leached pedalfers
Podsol soils 400–1,000 380–750 Golden Downs (504)
Gore (711)
Hermitage (1405) Strongly leached pedalfers Podsol

In Central Otago, notably in the Ida Valley, the Alexandra region and the Maniototo basin, accumulations of soluble salts have been observed, notably calcium sulphate, with minor quantities of sodium sulphate, calcium carbonate, sodium carbonate and sodium chloride, but these are caused by saline ground waters and not by a pedological process.

Discrepancies between soil and climate may be explained by:


Parent material lacking lime.


Insufficient climatic data.


Climatic fluctuations.

A parent material with little lime could not be expected to produce significant accumulations of lime under a pedocal soil process. However, from analyses of greywacke and schist (Williamson, 1939) it is clear that there is enough lime to furnish recognisable accumulations.

[Footnote] 1 Number in brackets refers to NS Quotient for station.

– 166 –

Figures for humidity and average temperature are not altogether satisfactory. Both are recorded at 9 a.m. and are therefore not true average values. The discrepancies, however, vanish under high rainfall and are most marked under low rainfall, where the normal higher morning humidity should move the climatic indices to the wet side, i.e., indices for the places in the semi-arid region indicate a greater humidity than actually prevails. True average humidities in the semi-arid region would therefore be expected to decrease the values of the Thornthwaite and Meyer indices and so increase the discrepancies between soils and climate.

A climatic change from a previously wetter cycle to the present drier one, or a fluctuating climate, with the mean climate semi-arid, best fits the observed facts. According to this hypothesis, although the present climate should produce a pedocal at Alexandra, the previous pedalfer climate would have already removed the small supply of lime, the accumulation of which characterises the pedocal. If the sequence is more complex, with periods of pedocal formation alternating with periods of pedalfer formation, then it is obvious that a small accumulation of lime during a pedocalcic climatic period would be removed under the pedalfer process. Under such a cycle it would be impossible for a soil with normal pedocal properties to develop. If this view is correct, discrepancies should be greatest near the pedocal-pedalfer boundary, and least where the climate lies well within a major climatic zone. Thus a small increase in rainfall should make a considerable difference to a soil under a rainfall of 13 in. but little difference to one under 150 in. Discrepancies, in fact, do disappear on the western and wetter side of the intermontane climatic district and increase towards the dry eastern margin. Thus, under a podsol climate at the Hermitage there is no discrepancy and under the semi-arid climate of Alexandra the discrepancy is clearly marked.

Recent work has thrown some light on soil-climate relationships. The study of composite or polygenetic soil profiles has been proposed as a method of investigating climatic changes. The optimum conditions for the development of such soils are considered to lie on the margins of the semi-arid belt, that is, the zone separating pedocals [soil with lime enrichment in the lower horizons] from pedalfers [soils with iron enrichments in the lower horizons] (Bryan and Allbritton, 1943). The soils of Central Otago and parts of inland Canterbury lie in this interzonal area.

Nature Of Climatic Changes.

The available evidence, although not perfectly conclusive, may be accepted as illuminating certain broad features of post-Glacial climates. Most of the evidence bears on the nature of the climatic pulsations which immediately preceded the present cycle. The present climate seems to have followed a somewhat warmer and much wetter one. Moa remains in swamps, high moraines in the Antarctic, remnants of soils and vegetation above the present uppermost limit of vegetation, and evidence of a once-higher level of the timber line, together with evidence from pollen diagrams, indicate a recent fall

– 167 –

in temperature of at least 2° C. and discrepancies in soil pattern and forest remains in grassland climates, a change to lower rainfall: This evidence is summarised in the following table.

Table IV.—Summary of Evidences for, and Nature of, Climatic Changes.
Author. Evidence. Conclusion.
Hutton (1892) Moa remains in swamps. Climate previously wetter.
Speight (1911) Totara remains where totara does not grow at present. Climate previously wetter.
Speight (1919) Distribution of genera related to moa. Climate first dry, and then wetter.
Cockayne (1905) Study of Discaria toumatou. Climate previously wetter.
Cranwell and Post (1936) Pollen diagrams. Climate first cold, then warm and wet, and finally colder and drier.
Ferrar (1905) High-level moraines in Antarctic. Climate previously warmer.
Speight (1916) Buried forest near Christchurch. Widespread regional climatic change.
Raeside (1946) Totara remnants above present timber line. Climate previously warmer.
Raeside (1946) Discrepancies between soils and present climate in Canterbury and Otago. Climate previously wetter.
Raeside (1946) Previously higher vegetation level. Climate previously warmer.
Raeside (1946) Structural features and soil sequence of South Canterbury plains. Alternate periods of aggrading and degrading, probably correlated with climatic changes and consequent vegetation changes.

It is significant that discrepancies between climates and soils are confined to central and eastern districts of the southern part of the South Island. Most of the discrepancies lie to the east of the central mountains, while at the Hermitage, where the climate is more related to that of the West Coast, the climate and soil seem to be in equilibrium (Tables I, II, and III). In eastern and central regions the effect of the cyclonic westerly winds, which were highly charged with moisture, is greatly modified, and easterly winds play an important part in determining climate.

As judged from the present climate of the east coast of the South Island, a small temperature rise would be expected to lead to a higher rainfall. On the Canterbury downlands and the uplands of Eastern Otago the prevailing summer wind is the easterly or north-easterly, a cool sea breeze which piles masses of cloud and fog against the first elevated land in its path. Accordingly, rainfall and humidity gradients increase westwards towards the first upland masses, where summer is the wetttest season. Thus the coastal rainfall is about 25 in. and the rainfall at Orari Gorge, at the base of the Four Peaks block of mountains, is 49 in. A simple temperature rise might be expected to increase the amount of evaporation over the Pacific, and therefore the burden of moisture carried by the summer winds. Rainfall would increase correspondingly.

Brooks (loc. cit.) has suggested that the key to post-Glacial climatic changes in the northern hemisphere lies in the retreat and advance of the polar ice cap; that is, that these changes are related principally to changes in temperature and this change decides all other climatic consequences. He claims that the presence or absence

– 168 –

of this polar ice cap is dependent on very slight fluctuations in average temperature. A drop of 5° below the freezing point of sea water (28° F.) is sufficient to account for the formation of an ice cap which can ultimately reduce temperature 50° C. “Instead of having to account for changes in temperature of about 50° we have only to account for initial changes of 5°… the floating ice makes up the odd 45°.” Brooks further calculates that a rise in the temperature of the earth of only 2° would be sufficient to clear the polar seas of all ice, and he suggests that such a condition has been proved from the warm period of the seventh century onwards.

Dating Of Climatic Changes.

There is only slender local evidence for the dating of climatic changes in New Zealand. Meteorological records which might throw light on historic changes are scanty and cover little more than a quarter of a century. Moreover, unlike older countries, New Zealand possesses no historical records covering the periods of historical climatic change in Europe and Asia.

Two general lines of approach to the problem suggest themselves: (1) dating from such local evidence as may be available, and (2) dating by analogy with accurately documented changes in the northern hemisphere.

Local evidence is scanty. Historical records of the forest which once covered the downlands of South Canterbury and Eastern Otago survive in Maori tradition, but the evidence is not very conclusive, and throws only a dim light on the transition from forest to grassland. There is a hint of an event which may have hastened the change in a Maori tradition of a forest fire which swept through the forest about the middle of the thirteenth century. There is a possibility, however, that this fire may have been a tradition imported by the Maoris and not an historical event (Roger Duff, personal communication).

A Raratongan tradition of voyages into the Antarctic about 650 A.D. by one Ui-te-Rangiora (Buck, 1925, quoting Smith, 1910) suggests a climate congenial to primitive navigation in these waters, and the wave of exploration in the Southern Pacific between the eighth and fourteenth centuries might be regarded as further confirming that this period was a warm one.

Evidence from the forest remains themselves is more conclusive, though still somewhat indefinite. Totara ranks with Silver Pine (Dacrydium colensoi) as one of the most durable timbers in the world (Entrican). Well-preserved buildings and native canoes made of this timber have been estimated to be several hundreds of years old. Cockayne (1928) records a fallen trunk of Black Pine (Podocarpus spicatus), regarded of as lesser durability than Totara, which still exists on the wet forest floor of Stewart Island, and which is embraced by the trunk, originally the root, of a Weinmannia racemosa estimated to be from 300 to 400 years old. It is possible, therefore, that the fallen logs which testify to the former existence of forest could have remained undecayed in the drier climate of Canterbury and Otago for 400 or even more years. Trees charred by fire might last even longer.

– 169 –

Many of the fallen logs were from mature trees, and it may be assumed that the forest had been standing for several centuries before its destruction, since a totara can reach an age of 600 years. From the evidence of the forest, therefore, it can only be concluded that the forest is likely to have stood for at least 600 years, and that it could have given place to grassland as much as 500 or 600 years ago.

Argument by analogy with the Northern Hemisphere rests on the assumption that climatic changes in the northern hemisphere had their counterparts in the southern.* In the northern hemisphere climatic sequences have been traced for some twenty centuries in detail. The record is based on the study of tree rings, varves, succession of plant types in peats, faunal successions, changes in lake levels, and on documentary evidence. The following broad climatic changes have been outlined for Europe (Brooks, 1926):

Atlantic 5000–3000 B.C. Moist and warm.
Sub-boreal 2500–1000 B.C. Dry and warm.
Sub-Atlantic 850–400 B.C. Moist and warm.
200–300 A.D. Resembling present.
300–700 A.D. Warm and dry with 800 moister.
900–1200 A.D. Warm and dry.
After 1200 A.D. Gradually growing wetter and colder.

Rainfall, is, of course, determined by local circumstances of topography, and rainfall changes are not therefore of very general applicability. Temperature changes, however, may be supposed to be more general and probably even to apply to most non-tropical regions. According to Brooks, they are related to the behaviour of the ice caps and would necessarily be general. The changes that might be expected to be widespread, therefore, are from present day temperatures in the first three centuries of this era to warmer temperatures from the fourth century to the thirteenth, and falling temperatures from then till now.

The warm period from the fourth to the thirteenth centuries saw the almost complete melting of the Arctic ice cap (Brooks, 1926) and an active retreat of the Greenland glaciers. During this period grapes were grown in England where the climate is now too cold to support the vine (Cooper, 1934) in any but the most sheltered positions. Although the Domesday Book contains frequent mention of vineyards, and William of Malmesbury (1080–1145) describes good wine from these vineyards, wine could now be made successfully only if the summers were about 5° warmer (Huntington, 1941). The same period (from the seventh century A.D. to about the thirteenth) saw the flourishing of Irish agriculture. (loc. cit.) and the clearing of ice from the Alpine passes of Europe. It was also a period of great exploratory activity in the Old World. Polar seas were largely free from ice and there were frequent Norse voyages to and from Iceland

[Footnote] * Since the above was written, the Report of the Committee on Glaciers for 1945 has been published (Trans. Amer. Geophys. Union, 27, p. 219, 1946; summary in Nature, Vol. 158, No. 4003, 1946. The committee, after considering the evidence, concluded that:—

[Footnote] 1. The causative climatic variations have affected both hemispheres simultaneously and not in alternation.

[Footnote] 2.It is therefore reasonable to suppose that the same pronounced Post-Pleistocene variations and the major Pleistocene variations were also synchronous in the two hemispheres.

– 170 –

and Greenland (loc. cit.) and thence to North America. Norse sailing routes of the time show a direct track from Iceland to the East Coast of Greenland in Lat. 66° N., thence down the coast to Cape Farewell and up the west coast. This route was followed until the end of the thirteenth century, and there is no mention of ice in historical accounts of voyages along it. At the same time two settlements were established in Greenland where the settlers grew grain and, raised cattle and sheep. At present the ground is frozezn throughout the year. Excavations have unearthed the coffins of Vikings into which the roots of trees have penetrated, but no trees will now grow on the frozen ground. The gradual advance of the colder climate can be traced in the living conditions of the settlers. The change of diet as grain became more difficult to grow is reflected in worn teeth and changed physique. About the fourteenth century the settlements, which until then had kept in close touch with Norway, lost all contact with Europe, and their subsequent history has been reconstructed from evidence lately excavated.

It may be taken therefore as a safe working hypothesis that the warm forest period in Canterbury coincided with the warm period in the northern hemisphere between the seventh and the fourteenth centuries. The transition to the cooler grassland period probably corresponded with the change from the warm period in the northern hemisphere to the colder period which began about the fourteenth century.


The foregoing considerations throw some doubt on the wisdom of arbitrarily correlating soil groups with present climates unless the climate is such that moderate variations will not alter the nature of the soil-forming process. This condition appears to apply to climates falling well within the podsol zone and may doubtless apply also to other wet climates as well. Under rainfalls of less than 40in., however, climatic variations can superimpose one soil-forming process on another. This has been demonstrated for the semi-arid soils and may also apply to the clay-pan soils of the East Coast of the South Island and similar soils in the Wairarapa Valley and Central Hawke's Bay.

The climate, therefore, should not be regarded as a constant or static soil-forming factor, but rather as a dynamic or constantly changing one, fluctuating between certain extremes of rainfall, humidity and temperature. If the fluctuations are sufficient to change the soil-forming process—as, for example, from a pedalfer process to a pedocal one—the profile will become a complex or abnormal one, and difficult to correlate exactly with soils developed elsewhere in different latitudes, and with different climatic backgrounds. Moreover, if climate varies significantly, changes in vegetation might be expected, too, and these would serve further to complicate the profile. It would seem undesirable, therefore, to strain too much after exact correlation of New Zealand soils with overseas soils, but rather to regard New Zealand soils, except where the correspondence in profile and background is clear cut (for example, in the podsol group), as complex soils related to a purely local set of soil-forming factors among which the climate and vegetation undergo local variations.

– 171 –


1895. New Zealand Alpine Journal, October.

1849. New Zealand Spectator, November 24. Account of Lieutenant-Governor Eyre's Expedition to Mount Tapuanuku in the Inland Kaikouras.

Brooks, C. E. P., 1926. Climate Through the Ages. Benn, London.

Bryan, J., and Albritton, C. C., 1943. Soil Phenomena as Evidence of Climatic Changes. Amer. Jour, of Sci., August, 1943.

Buchanan, J., 1868. Sketch of the Botany of Otago. Trans. N.Z. lust., vol. 2.

Buck, P. H., 1925. The Coming of the Maori. Cawthron Leotures, vol. 2, no. 2.

Burdon, R. M., 1938. High Country. Whitcombe2491 & Tombs, Wellington.

Burnett, T. D., 1926. In The Natural History of Canterbury. Simpson & Williams, Christchurch.

Cockayne, L., 1905. The New Phytologist, vol. 4.

—— 1928. The Vegetation of New Zealand. Die Vegetation der Erde. Leipzig.

Cooper, Chas., 1934. The English Table in History and Literature. Sampson, Lowe, Marsdon & Co., Ltd.

Cotton, C., 1944. Landscape. Cambridge University Press.

Cranwell, L. M., and Post, L., 1936. Post-Pleistocene Pollen Diagrams from the Southern Hemisphere. Geogr. Annale, H. 3–4.

Entrican, E. The Properties and Uses of Totara. New Zealand State Forest Leaflet, no. 22.

Ferrar, H. T., 1905. Notes on the Physical Geography of the Antarctic. Geographical Journal, vol. 21, no. 4.

—— 1925. The Geological History of Ross Dependency. N.Z. Journal of Science and Tech., vol. 7, no. 6.

—— 1926. The Soils of the Irrigation Areas in Central Otago. N.Z. Geol. Survey Bull., no. 33 (new series).

Gibbs, H. S., Raeside, J. D., and Dixon, J. K., 1945. Soil Erosion in the High Country of the South Island. D.S.I.R. Bull., no. 92.

Hardcastle, J., 1908. Notes on the Geology of South Canterbury. Timaru Herald Print, Timaru.

Huntington, Ellsworth, 1941. Climatic Pulsations, in Conservation of Renewable Resources. Univ. of Pennsylvania Press.

Jenny, H., 1941. Factors of Soil Formation. McGraw-Hill Book Co., N.Y.

Kidson, E., 1930. Average Annual Rainfall in N.Z. Govt. Printer, Wellington.

Park, J., 1908. Geology of the Cromwell Subdivision. N.Z. Geol. Survey Bulletin, no. 5.

Smith, P., 1910. Hawaiki, 3rd edition. Whitcombe & Tombs, Wellington.

Speight, R., 1910. The Post-Glacial Climate of Canterbury. Trans. N.Z. Inst., vol. 43.

—— 1916. An Ancient Buried Forest near Riccarton. Trans. N.Z. Inst., vol. 49.

Wilde, L. J., 1910. Note on the Mechanical Composition of the So-called Loess at Timaru. Trans. N.Z. Inst., vol. 51, p. 286.

Williamson, J. H., 1939. The Geology of the Naseby Subdivision, Central Otago. N.Z. Geol. Survey Bull., no. 39 (new series).

Zeuner, F. E., 1946. Dating the Past, an Introduction to Geochronalogy. Methuen & Co., London.