[Received by the Editor, August 19, 1955.]

Abstract

This paper presents some results of a survey of vegetation and climate on the hills near Dunedin which provide evidence of vegetation changes shortly preceding those for which European settlement has been responsible, and reveal some difficulties in correlating rainfall, evaporation and temperature records at selected sites with the type of vegetation they carry to-day. A map is presented which shows the distribution of indigenous coniferous forest, silver-beech, and snow-tussock, and the areas over which log remains of Podocarpus hallii and Dacrydium biforme and surface dimpling provide evidence that the snow-tussock has recently replaced forest. Most existing silver-beech forests are shown to be regenerating vigorously and tending to encroach upon snow-tussock and coniferous forest, while regeneration rates of the conifers appear generally low. The snow-tussock is subject to invasion by shrubs and trees or fescue-tussock, particularly if heavily grazed after burning.

Rainfall and evaporation records for numerous temporary stations established over the area are also mapped The 45in isohyet is shown to lie to the west of a line joining Flagstaff and Silver Peak, while land above 2,000ft on Maungatua receives over 40in of rain annually. Temperatures fall with altitude and evaporation tends to do likewise. Frost temperatures are recorded which show that marked inversion patterns develop in grassland areas. Persistence of forest (and particularly of D. biforme) on the summit of Mt. Cargill is correlated with high rainfall (54in), low evaporation rate, and frequent sea-fogs, but the upper slopes of the tussock-covered Swampy Hill and Maungatua have not proved drier than their seaward faces on which forest persists. It is suggested that these areas of snow-tussock replaced forest that would have remained had it not been destroyed by fires of Maori origin. In the northern part of the district, however, there are indications that dry N-W winds may be directly responsible for treelessness.

In this paper an attempt is made to elucidate the relative importance of present and past climate and human intervention in determining the pattern of vegetation on the hills near Dunedin. Brief accounts of the flora and plant communities of this area have been published (Simpson and Scott-Thomson 1928, 1938; D.N.F.C., 1932) and fuller descriptions prepared in the course of the present survey have been deposited in the library of the Otago University (Wardle, 1953; Mark, 1954).

Methods

1.Estimation of rainfall

A few official rain-gauges have contributed to the records presented on Map I, but most of the data are from improvised apparatus (Fig. 1) read at intervals of 4–6 weeks, either over the period September, 1952 to August, 1953, or April, 1953 to March, 1954. A spirit level was used to adjust each funnel in the field, and the exact area of its mouth was measured with a planimeter. In the beginning each can was primed with that quantity of water which it retained when inverted with the funnel removed. The reliability of these instruments was tested in two ways.

(a) Gauges were maintained alongside official ones at Lake Mahinerangi and at Musselburgh. The annual total recorded by the improvised gauge at Lake Mahinerangi was 31.54ins, while the official record was 31.18ins. At Musselburgh, the mean of two improvised gauges was 29.40ins and the official record 31.62ins. This larger discrepancy was undoubtedly due to the cans being placed on the surface of the lawn where they were subject to heating by the sun. At Lake Mahinerangi and in the field, cans were set in the ground and a cairn built about the base of the funnel. This protected them from heating and from freezing.

(b) In four localities two gauges were sited close together. The differences over the year were .01, .01, .65, and 1.30ins. Such differences are unimportant.

Rainfall figures entered upon Map I are suggested annual means Over the northern hills, where the records were taken between September, 1952, and August, 1953, they are the actual measurements made since all official stations in the district had recorded a rainfall over that period very close to their mean for the last ten years. But the corresponding means for Lake Mahinerangi and Taieri Airport, the two official stations nearest to Maungatua, were 1.2 times

Measurements of rainfall and evaporation rate at numerous stations on the area are recorded on Map I. The rainfall figures are assumed annual means, and the atmometer readings are for the summer, of 1953–54, except those in square brackets, which were made in the previous year with different apparatus and are thus comparable only among themselves.

This map also shows the average inland limit of the sea-fogs on the northern hills. These fogs, which normally come from the north-east and have their base at about 1,000ft are important because during the warmer months they produce a saturated atmosphere and heavy fog-drip on the land they cover when relative humidity west of the fog limits may be comparatively low. For instance, over a period of 34 days beginning on December 17, 1952, sea-fogs, unaccompanied by rain, developed on fourteen days, and on five of these the relative humidity at Taieri Airport at noon was below 60%, while on all but two of the remainder it was below 80%. Maungatua is only occasionally affected by these sea-fogs, and here the normal fog-base is higher (about 2,000ft). During a period of 75 days beginning December 25, 1953, sea-fogs covered both Mt. Cargill and Swampy Hill on twenty occasions, while Maungatua was covered only five times. Southerly fogs also occur (nine during this period), but these cover the northern hills and Maungatua equally, and are normally accompanied by rain.

The effect of aspect differences on local distribution of rainfall was investigated. The figures in Table I show that in most cases, the type of site chosen for the gauge did not greatly affect its reading The fact that the rainfall figures recorded over the northern hills show only a gradually changing pattern accords with this. But the last two pairs of readings which are from Maungatua establish that aspect some-times substantially affects the rainfall on this mountain. In both instances, the two sites were within a quarter mile of one another.

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

Table 1.—Effect of Aspect in Rainfall.

Altitude (ft)	Site	Aspect	Rainfall	Greatest Diff. as % of lowest reading
2,400	Steep slope	N	48.55
2,350	Small basin	S	50.24	9.5%
2,150	Gully-side	W	51.62
1,950	Flattish spur	S	47.18
2,000	Moderate slope	SE	42.21
1,900	Moderate slope	W	44.09	7.2%
1,850	Moderate slope	S	41.13
1,450	Moderate slope	E	36.99
1,700	Moderate slope	N	39.02	5.5%
1,800	Gentle slope	S	36.64
1,800	Gentle slope	N	37.27	1.73%
300	Flat valley bottom	–	30.22
400	Flat spur	–	27.58	9.6%
1,000	Flattish summit	–	31.12
800	Flat ridge-top	–	29.30	14.7%
500	Moderate slope	N	33.60
2,100	Flat summit	–	29.84
2,050	Gentle slope	S	31.18	4.5%
850	Moderate slope	SE	32.39
850	Steep slope	N	30.50	6.2%
2,700	Steep slope	NW	35.60
2,750	Steep slope	SE	29.48	20.7%
1,600	Moderate slope	SE	30.10
1,800	Hill crest	–	20.60	46.1%

absence of southern rata (Meterosideros umbellata) from the Dunedin district illustrate other phases of the reduction of the regional tree flora.

Kowhai (Sophora microphylla) has entered two of the forests sampled (Bell's Gully and N. Waikouaiti R.), both where a low rainfall (approximately 30in) has thinned the subordinate strata, particularly the ground ferns (Blechnum capense and B. discolor are both absent), thus admitting much light to the forest floor when a break in the canopy develops. In places, kowhai is accompanied by kanuka (Leptospermum ericoides). Perhaps the fairly extensive stands of kowhai and kanuka in the middle reaches of the Silverstream Valley have originated from similar forest in which regeneration of the podocarps had failed completely.

Shrubland and Grassland.

In 1850 the grassland was probably almost wholly snow-tussock (Danthonia flavescens) since even now it only includes substantial amounts of low tussock (Poa caespitosa and Festuca novae-zealandiae) where it has been much burnt and grazed. It occurred under all climatic regimes of the district and shared margins with every forest association. This suggests that it was in large part an entrenched non-climax community, and its behaviour under the grazing and burning practices of the last century confirms this view. Large areas lying between 1,000–2,000ft have been invaded by the unpalatable and light-demanding seedlings of kanuka (Leptospermum ericoides) and particularly manuka (L. scoparium) whose spread in periodically burnt areas is probably favoured by the greater protection from heat that the larger capsule affords to the seed. Early records prove the derivation from tussock of some of the present Leptospermum communities, the most extensive being in the Silver-stream Valley, where ring counts show their age to be about 70 years. More recent extensions often reveal their origin through containing persistent specimens of the grassland flax (Phormium colensoi). On the other hand, some replacement of snow-tussock by kanuka appears to have pre-dated European settlement, since about 160 growth rings have been counted in the very large L. ericoides (about 40ft high and 18in in diameter) that sometimes intervenes between forest and grassland—e.g., between Misery Creek and Bendoran). Even above 2,000ft on Maungatua the snow-tussock is not stable, but the shrubs which tend to invade it at these altitudes are different principally Dracophyllum longifolium, Hebe buxifolia and Cassinia fulvida).

Leptospermum seedlings are so strongly light-demanding that regeneration does not occur until the ground is almost bare. Failing destruction by fire or felling there is thus normally only one generation upon a site, on which it provides protection for shade-tolerant seedlings of a more permanent community. The older Leptospermum stands now contain seedlings of the components of coniferous-dicotyledonous forest or silver beech to an extent depending mainly upon the proximity of parent trees. Young conifers have not been observed more than 2 chains from mature trees. Rimu seedlings and poles are plentiful along with silver beech in a succession occurring between the two main tributaries of the S. Waikouaiti R. at 1,200ft elevation and under a rainfall of 45in. Rimu seedlings were even found under Leptospermum in the N. Waikouaiti Valley with a rainfall of 30 inches and only stunted mature trees surviving. Holloway (1954, p. 391) draws attention to the fact that this species will establish itself in Leptospermum in areas where it fails to regenerate in forests. But the mountain totara promises to be the most conspicuous podocarp in these new forests.

extended by fire into areas climatically suited for forest, important recent changes are easily understood—namely, the marginal pole stands of beech at Silver Peak and Woodside; the rapid spread of scrub into tussock in the Silverstream Valley; and the very general tendency for this process to occur elsewhere. These successions have been able to begin, partly through cessation of burning (e.g., Silverstream Valley is now a Fire District) but mainly because browsing animals hitherto absent have retarded the recovery of burnt tussock and reduced the competitive advantage which its resilience after fire had hitherto conferred upon it. This is discussed more fully in another paper (Mark, 1955).

Certain forest species are often conspicuous among the snow-tussock, noticeably Cyathea colensoi, Polystichum vestitum, Blechnum minor and Fuchsia colensoi, and probably wherever they occur they indicate that the climate would support forest It is noteworthy that these, and invading manuka become inconspicuous or absent in the valleys of the north-west where the rainfall declines to about two-thirds of that on Silver Peak while judging by the single station established, the evaporation rate increases by about 70 per cent. In this country dry north-west winds are more frequent and more violent than they are nearer Dunedin city, and their responsibility for disappearance of the forest may be more direct.

Acknowledgments

We are indebted to Professor G. T. S. Baylis for much assistance and advice, and to the Research Fund of the University of New Zealand for a grant to cover travelling and equipment. The co-operation of many others in matters referred to them is also gratefully acknowledged, particularly Mr. H. R. Orman, of the Forest Research Institute, the Director and Mr. J. Finkelstein of the Meteorological Office, and many residents or former residents of the district.

References

Barker, A. P., 1953. An ecological study of tussock grassland, Hunters Hills, South Canterbury. N. Z. Dept. Sci. Ind. Res. Bull. 107, 58 pp.

Beilmann, A. P., and Brenner, L. G., 1951. The recent intrusion of forests in the Ozarks Ann. Mo Bot. Gard. 38, 283–291.

Braun-Blanquet, J., 1932. Plant Sociology. Trans. by G. D. Fuller and H. S. Conard, 439 pp., New York.

Dunedin Naturalists' Field Club, 1932. Catalogue of the indigenous, and introduced flowering plants, ferns, and seaweeds occurring in the Dunedin district, 41 pp., Dunedin.

Garratt, G. A., 1924. The identification of some New Zealand woods. Circular, N. Z. State Forest Service.

Holloway, J. T., 1954. Forests and climates in the South Island of New Zealand. Trans. Roy. Soc. N.Z., 82, 329–410.

Mark, A. F., 1954. A study of the vegetation of the Maungatua Range. Thesis, Otago University Library.

—— 1955. Shrubland and grassland of Maungatua, Otago. N.Z. Journ. Sc. Tech., A37, 349–366.

Orman, H. R. and Reid, J. S., 1946. Wood anatomy of New Zealand Dacrydium species. N. Z. Jour. For., 5, 215–219.

Phillips, E. W. J., 1941. The identification of coniferous woods by their microscopic structure. J. Linn. Soc. Bot., 259–320.

Richards, P. W., 1952. The Tropical Rain Forest. 450 pp., Cambridge.

Shaw, M. and Farrant, E. D., 1949. The Taieri Plain—tales of the years that have gone. 238 pp., Dunedin.

Simpson, G. and Scott-Thomson, J., 1929. On the occurrence of silver southern-beech in the neighbourhood of Dunedin. Trans. N.Z. Inst., 59, 326–342.

—— 1938. The Dunedin sub-district of the South Otago botanical district. Trans. Roy. Soc. N.Z., 67, 430–442.

Wardle, P., 1953. An ecological survey in the hill country north of Dunedin. Thesis, Otago University Library.

P. Wardle,

Botany School,
Downing Street,
Cambridge, England.

A. F. Mark,

Department of Botany,
Duke University,
North Carolina, U.S.A.