ArticlePDF Available

Vegetation of Subantarctic Marion and Prince Edward Islands

'Marion was a lovely picture. She rose, a jade jewel, out of the sea. Her lush green coat
was fringed with the black lace of the cliffs and her heights draped in scintillating
from J. H. Marsh, 1948. No pathway here. Howard B. Timmins, Cape Town. p. 71
Figure 15.1 Fellfield vegetation with Azorella selago on Marion Island. In the foreground
a narrow tongue of Holocene black lava extends over a Pleistocene grey lava flow. In the
background, 4 km away, is the meteorological station surrounded by a mosaic of mire,
slope and fellfield. The cinder cone on the right is Junior's Kop.
V.R. Smith
ST R E L I T Z I A 19 (200 6)
Vegetation of Subantarctic Marion
and Prince Edward Islands
Valdon R. Smith and Ladislav Mucina
Table of Contents
1 Introduction 701
2 The Subantarctic Region and Tundra 701
3 Geography of Marion and Prince Edward Islands 703
3.1 Climate 703
3.2 Topography 704
3.3 Geology 704
3.4 Soils 705
3.5 Palaeohistory 706
4 Flora and Major Vegetation Patterns 707
5 Vegetation Dynamics 708
5.1 Vegetation Succession 708
5.2. Primary Production, Decomposition and Nutrient Cycling 709
6 Current Conservation Status of the Islands 709
7 Threats to the Islands' Ecosystems 710
7.1 Implications of Climate Change 710
7.2 Alien Flora, Alien Animals and Human Influence 711
8 The Future 712
8.1 Future Research 712
8.2 Development and Conservation Threats 712
9 Definition of Biomes, Vegetation and Mapping Units 713
10 Description of Vegetation Units 714
11 Credits 721
12 References 721
List of Vegetation Units
Marine Macroalgal Vegetation 714
AZm 2 Subantarctic Kelp Beds 714
Subantarctic Tundra Biome 715
ST 1 Subantarctic Coastal Vegetation 715
ST 2 Subantarctic Biotic Herbfield and Grassland 715
ST 3 Subantarctic Mire 716
ST 4 Subantarctic Drainage Line and Spring Vegetation 717
ST 5 Subantarctic Fernbrake Vegetation 718
ST 6 Subantarctic Fellfield 719
ST 7 Subantarctic Cinder Cone Vegetation 719
Polar Desert Biome 720
PD 1 Subantarctic Polar Desert 720
700 Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
Fig 15 x Vegetation Map of the Prince Edward Islands
Prince Edward Islands
ST 1, 2 Subantarctic Coastal Vegetation
ST 3, 4, 5 Subantarctic Mire-Slope Vegetation
ST 6 Subantarctic Fellfield
ST 7 Subantarctic Cinder Cones
PD 1 Subantarctic Polar Desert
Scale 1:150 000
0 5 10 Km
Ship's Cove
Trypot Beach
Macaroni Bay
Bullard Beach
Kildalkey Bay
Cape Hooker
Crawford Bay
Rook's Bay
Mixed Pickle Cove
Trigaardt Bay
Cape Davis
Storm Petrel Bay
King Penguin Bay
Marion Island
RSA Point
Cave Bay
South Cape
McNish Bay
Prince Edward Island
Prince Edward Islands
Figure 15.2 Vegetation maps of the Prince Edward Islands.
Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
limnological or biological context. Even within these disciplines,
workers have considered different numbers of islands and island
groups as subantarctic. Oceanographers generally define the
subantarctic region as the zone of waters between the Antarctic
and Subtropical Convergences (Deacon 1960). The latitudinal
distributions of those organisms dependent entirely upon the
sea for their existence seem to follow these boundaries quite
well but the limits of the main types of terrestrial vegetation
(Wace 1965), insect fauna (Gressitt 1970) and avifauna (Barrat
& Mougin 1974) do not coincide with them so closely.
Botanists have generally preferred to define the subantarctic
region on plant physiognomic, vegetation structure or climatic
criteria. For instance, Wace (1960, 1965) proposed delimiting
the subantarctic region as the area south of the southern limit
of tree or shrub growth (distinguishing it from the cool temper-
ate zone) and north of the southern limit of closed phanero-
gamic vegetation (delimiting it from the maritime or low-ant-
1. Introduction
South Africa possesses two subantarctic islands—Marion Island
(46° 54´ S, 37° 45´ E) and Prince Edward Island (46° 38´ S, 37°
57´ E) (Figure 15.3). Together they are known as The Prince
Edward Islands. Marion Island was annexed by South Africa
on 29 December 1947 and Prince Edward Island on 4 January
1948. The annexations became effective with the passing by
Parliament of the Prince Edward Islands Act (43) of October
1948. Marion Island has been occupied permanently by South
African research and logistic personnel since February 1948.
There is no permanent occupation of Prince Edward Island.
2. The Subantarctic Region and Tundra
The term subantarctic refers to various regions of the Southern
Ocean, depending upon whether it is used in an oceanographic,
Figure 15.3 Position of the Prince Edward Islands in Southern Ocean, between Antarctica and Africa (courtesy of I. Meiklejohn).
702 Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
arctic to the south). In Wace’s scheme the following terrestrial
plant communities were considered to characterise the subant-
arctic vegetation zone:
Closed herbfield communities in which large-leaved peren-
nial herbs are conspicuous.
Communities of pedestal-forming tussock grasses, espe-
cially on the coast.
Soligenous mires in which the peat-forming plants are bry-
ophytes and sedges (not sphagna or cushion-forming vas-
cular plants).
Fellfield or wind-desert communities composed of flower-
ing plants with very compact mat or cushion growth
Although Wace (1965) considered this classification to be inde-
pendent of climatic, geological or other environmental data,
the delimitation boundaries correspond closely to those of the
climatological classification of Holdgate (1964) based on the
range of mean monthly temperatures. In that classification the
subantarctic zone is characterised by the absence of tempera-
tures warmer than 8.5°C, considered to preclude the occur-
rence of trees (Pearsall 1950, Holdgate 1964).
Holdgate (1970, 1977) proposed a more definitive classification
of the southern subpolar zone that, with some modification,
has been widely accepted (Block 1984, Pickard & Seppelt 1984,
Clark & Dingwall 1985, Bonner & Lewis Smith 1985, Smith &
Lewis Smith 1987). The geographic, climatic and biotic char-
acteristics of the various subdivisions (regions and provinces)
of the zone are outlined by Lewis Smith (1984). The suban-
tarctic region comprises six islands or island groups (South
Georgia, Macquarie, Heard and McDonald, Kerguélen, the
Crozets, Marion and Prince Edward), all situated within a few
degrees latitude of the Antarctic Convergence. Because of the
cold Bouvet Current that originates in the Weddell Sea, the
Convergence extends farther north in the Indian Ocean so that
relatively low-latitude islands such as Marion, Prince Edward
and Îles Crozet are included as subantarctic, even though they
are farther north than the islands of the New Zealand conti-
nental shelf (Campbell, Antipodes, Auckland, Snares). These
latter, because they possess a well-developed shrub and/or low
tree vegetation, are considered to belong to the cool temperate
zone, together with the Tristan da Cunha-Gough Island group,
Île Amsterdam, Île Saint-Paul, Falkland Islands and southern
Tierra de Fuego. At the other extreme, Bouvetoya and the
South Sandwich Islands are sometimes included in the suban-
tarctic region (Walton 1985), although they are more typically
maritime antarctic, like the South Orkney Islands and South
Shetland Islands (Lewis Smith 1984). Alternative classifications
of mid- to high-latitude regions of the southern hemisphere are
discussed by Smith & Lewis Smith (1987).
Tundra is usually defined as areas where the average annual
temperature is below 0°C and permafrost occurs. It has also
been used when referring to areas where the temperature is too
low, and precipitation and wind too great, for a natural forest
vegetation to develop. For example, oceanic moorland, some
mountainous regions and two subantarctic islands, all of which
have somewhat higher annual mean temperatures than tundra
sensu stricto, were included in the Tundra Biome Study of the
International Biological Programme (Rosswall & Heal 1975, Bliss
et al. 1981). In fact, subantarctic vegetation has traditionally
been decribed as tundra. Very few taxonomic affinities exist
between the biotas of subantarctic islands and the Arctic or
Subarctic, but there are several physiognomic similarities in their
vegetations. In order to highlight the unique nature of subant-
arctic ecosystems, it is useful to compare terrestrial areas of the
Subantarctic briefly with those of the Antarctic and also with
those of the Subarctic.
Owing to the overwhelming influence of the ocean, the sub-
antarctic region exhibits a very limited annual temperature
range, so that winters are warmer and summers cooler than
continental subpolar sites of the northern hemisphere (e.g.
see Figure 2 in French & Smith 1985). Thus, while the subarc-
tic environment may be considered to be a less severe version
of that of the Arctic, the difference between subantarctic and
antarctic conditions was described by Rudmose-Brown (1928)
as follows:
‘The term subantarctic is justified rather by proximity to the
Antarctic than by any real approximation to Antarctic conditions.
The truly Antarctic climate is typically continental, in contrast
to the climate of the subantarctic islands, which is essentially
oceanic, and in most respects cool-temperate, rather than polar.
This important difference between subantarctic and subarctic
regions is due to the unequal proportion of land and sea in
the two zones (Di Castri et al. 1970). While large continental
masses of North America, Asia and Europe belong to the sub-
arctic region, subantarctic terrestrial areas are very discontinu-
ous. This has a series of climatic, biogeographical and ecologi-
cal consequences, some of which have been demonstrated by
French (1974, 1981) and French & Smith (1985) in multivariate
comparisons of subantarctic and maritime antarctic terrestrial
ecosystems with northern hemisphere tundra and tundra-like
ecosystems (ranging from high arctic to cool-temperate and
temperate alpine). The comparisons, based on climate, soil
chemical composition and vegetation, emphasise the effects
of the extreme oceanity and the influence of sea spray and
animal manuring on the ecosystems of the southern hemisphere
subpolar sites. They clearly show that subarctic ecosystems are
generally less severe forms of arctic ones, and decreasing lati-
tude leads to increasingly milder environments with no great
changes overall in continentality. In contrast, the Subantarctic
combines elements from the extremes of the range of northern
hemisphere tundra (i.e. high-arctic and cool-temperate oce-
anic) with its own peculiar features, e.g. geographical isola-
tion, wind exposure, high rainfall, small seasonal temperature
range and the strong influence of the marine ecosystem (espe-
cially of seals, seabirds and salt spray) to produce ecosystems
that are qualitatively different from both subarctic systems and
the continental antarctic regions to which they are geographi-
cally closest.
The isolation of terrestrial subantarctic ecosystems and, in
most cases, their relatively recent origin causes their floras to
be species-poor, in contrast to most subarctic and low arctic
areas of the northern hemisphere. Despite this, the multivariate
analyses done by French & Smith (1985) show that the south-
ern hemisphere sites exhibit almost the full range of vegeta-
tion types found in the northern hemisphere. Cluster analy-
ses of climates and soils link the Southern Ocean islands with
northern hemisphere sites at higher latitudes, indicating the
effect of the Antarctic continent on the climate of the southern
islands. Climatic wetness, together with strong winds, applies
a significant chill factor so that the vegetation of the southern
islands resembles that of much higher latitudes in the northern
hemisphere. The latitudinal difference between northern and
southern sites with equivalent vegetation types is as much as
20 or even 30 degrees. The range of soil nutrient levels at the
southern hemisphere subpolar sites is approximately equal to
that at the northern hemisphere subpolar sites; however, in the
southern sites soil nutrient levels are inversely related to climatic
severity, while in the north the reverse is true. The two most
Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
Climate diagrammes
10-08-2005, 07:19, marionIslandClimateDiagramForPDF.xls
Marion Island
>2A 2327 ^^
2A4G 13
>2E 5.6 4
>75 48 U
; 7 > 2 > ; ; 2 D @ ? 5
Table 15.1 Average monthly and annual air temperature, pre-
cipitation and sunshine for Marion Island. Values are derived
from data supplied by the South African Weather Bureau and
were obtained from measurements made between 1949 and
2002 at the Meteorological Station which is about 25 metres
above sea level.
Temperature (˚C) Precipitation (mm) Sunshine
(hours day-1)
Jan 7.3 5.8 9.1 209 90 350 5.1 3.4 6.4
Feb 7.8 6.2 9.3 188 73 333 4.8 3.8 6.3
Mar 7.5 6.5 8.9 199 52 397 3.8 2.8 5.6
Apr 6.4 4.5 8.2 210 125 363 3.1 1.7 4.3
May 5.3 3.7 7.1 222 100 428 2.5 1.4 3.3
Jun 4.6 2.8 6.0 203 91 461 1.9 0.9 2.8
Jul 4.1 2.6 5.9 196 98 319 2.1 1.3 2.9
Aug 3.7 2.1 4.7 180 98 250 2.8 1.5 4.3
Sep 3.8 2.3 5.3 183 101 360 3.4 2.4 4.6
Oct 4.6 3.4 6.2 171 75 288 4.5 2.8 7.2
Nov 5.5 4.1 7.7 172 48 316 5.3 3.2 6.9
Dec 6.4 4.7 8.2 200 73 302 5.2 3.8 6.7
Year 5.6 4.8 6.6 2327 1861 2992 3.7 3.2 4.1
important causes of these opposed trends are the relative geo-
logical and pedological ages of the sites (younger sites tend to
be at higher latitudes in the north but at lower latitudes in the
south) and the much greater effects of sea-based vertebrates
on the southern islands.
Subantarctic islands are geologically comparatively young com-
pared with land masses of the northern hemisphere subpolar
region. The islands are also extremely remote, mostly more
than 1 500 km from the nearest continent. Their indigenous
biota has been established by the slow process of immigration
and colonisation over large expanses of ocean, so that the flo-
ras and faunas of the islands exhibit low diversity. For instance,
Lewis Smith (1984) listed 72 vascular plant species indigenous
to the Subantarctic, whereas the Canadian Arctic Archipelago,
a climatically far more extreme region and ca. 25° of latitude
nearer the Pole than the Subantarctic, has more than 340 indig-
enous vascular plant species (Porsild 1964). Several vascular
plant species are endemic to the Subantarctic, but occur on
more than one island or island group (Walton 1985). Only four
species are endemic to one island or island group (Lewis Smith
1984). The low degree of endemism in the vascular floras of
the subantarctic islands, compared with those of the older cool-
temperate islands to the north—for instance 31% of the 200
vascular species on the New Zealand shelf islands are endemic
(Devine 1984)—probably results from the relatively short time
available for the evolution of new species since the last gla-
ciations. However, despite their floristic poverty, subantarctic
islands exhibit almost the full range of vegetation types found
in tundra (sensu I.B.P.) of the northern hemisphere (French &
Smith 1985). Perhaps the most notable exception is that lichen-
dominated communities similar to those found in tundra of the
northern hemisphere do not occur.
3. Geography of Marion and Prince
Edward Islands
3.1 Climate
The Prince Edward Islands are amongst the warmest of the sub-
antarctic islands; only the Crozet Archipelago is as warm. Mean
monthly and mean annual temperature, precipitation and sun-
shine at Marion Island are shown in Table 1. Other climatic data
are depicted in Figure 15.4. The surrounding Southern Ocean
causes the islands’ climate to be thermally stable. The differ-
ence between mean temperature of the coldest and warmest
months is only 4.1°C and the mean diurnal temperature vari-
ation is only 1.9°C. Precipitation is high, more or less equally
distributed throughout the year and, although snow and ice
rain do occur, is mostly in the form of rain at the coast. The high
precipitation is associated with a high incidence of cloudiness
and infrequent direct sunshine.
Despite the thermal stability of the islands, annual mean tem-
perature has increased steadily in the past three decades and
six of the last seven years (1996 to 2002) were the warmest on
record. Since 1969 air temperature has increased, on average,
by 0.04°C per year (Smith 2002) and warming has occurred in
all months except June. The highest rate of warming has been
from late austral winter to midsummer (September to January),
and the lowest in late summer, autumn and winter (February
to August), with the notable exception of April which showed
the greatest warming of all the months. Recently, Mélice et
al. (2003) showed that annual mean sea temperature around
Marion Island has risen, on average, by 0.04°C per year since
the 1970s, the same as the mean increase in air temperature.
Similar air temperature increases have been reported for three
other subantarctic islands (for Macquarie Island by Adamson
et al. 1988, Kerguélen Island by Frenot et al. 1997 and Heard
Island by Budd 2000).
Annual precipitation has decreased since the mid-1960s, so that
the 1990s was the driest of the five decades that precipitation
has been measured on the island. All months except October
have become drier. Interannual variability in annual total sun-
shine hours is irregular, but a significant proportion of it can be
ascribed to an average increase of 3.3 hours per year between
1951 and 2002. Hours of sunshine increased for all months in
that period.
The radiation, air and sea temperature increases, and the
precipitation decrease, on Marion Island might be associated
with changing atmospheric circulation patterns (perhaps also
changing oceanic circulation). Smith & Steenkamp (1990) pro-
Figure 15.4 Climate diagram of the weather station Marion. The
graph shows mean monthly precipitation (blue bars), and mean of
daily minimum and maximum temperatures for each month (red lines).
MAP: Mean Annual Precipitation. APCV: Annual Precipitation Coeffi-
cient of Variation. MAT: Mean Annual Temperature. MFD: Mean Frost
Days per year.
704 Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
posed a model whereby interannual differences in radiation,
temperature and precipitation can be explained by the posi-
tions of cyclone tracks relative to the island. Approximately 100
cyclones pass the island in a year. In warmer years the cyclonic
centres pass, on average, further to the south of the island so
that it spends a longer period subjected to the rain-bearing
northwesterly winds of the warmer sector of the cyclones, and
less time to the icy southwesterly winds of the post-cold front
sector. Validation of this model awaits a detailed analysis of sea
level atmospheric pressure records for the South Indian Ocean
sector of the Southern Ocean. Such an analysis for the Pacific
region of the Southern Ocean showed that the direction of
atmospheric pressure gradients governs whether warm north-
easterly or cold southerly air flows over Macquarie Island and
3.3 Geology
The islands are located near the centre of
the West Indian Ocean Ridge and are the
surface expressions of a mantle plume
marking the present position of a long-
lived hotspot. They were considered to
be extinct volcanoes but in September
1980 there was a small eruption on the
west coast that resulted in a lava flow
covering about 9 ha (Verwoerd et al.
1981) and some soil profiles contain bur-
ied A-horizons under volcanic ash depos-
its near the surface, attesting to sporadic
volcanic activity in the recent past.
The lavas are typical oceanic island
basalts. McDougall et al. (2001) recog-
nised eight periods of volcanic activity,
from 450 ka up to the Holocene one
that started about 10 ka and extends up
to the present. Two ‘types’ of lava flow
can clearly be distinguished in the field
(Verwoerd 1971, McDougall et al. 2001;
Figure 15.6):
Figure 15.6 Marion Island (below First Red Hill cinder cone): Three different geologies repre-
sented by young black lava (on the left), old grey lava (on the right) and cinder-cone scoria (in the
foreground) converging and accommodating a small lava lake (tarn). The cushions on the scoria
are Azorella selago.
hence determines air temperature there
(Adamson et al. 1988).
3.2 Topography
Marion Island is 290 km2 in area and
the highest peak is 1 230 m above sea
level. Prince Edward is 45 km2 in area
and the highest peak is 672 m above sea
level. Marion Island consists of a central
highland area sloping down to a coastal
plain which on the northern and eastern
sides forms a 4 to 5 km wide area ris-
ing gently from sea level to the foot of
the mountainous interior at about 300
m altitude (Figure 15.5). The western
and southern coastal areas consist of a
narrow discontinuous plain of less than
100 m altitude, having been eroded by
wave action caused by the strong west-
erly winds. Much of the low-altitude
area supports closed vegetation. The
southeastern part of the Prince Edward
Island slopes up gently from the coast to
the top of a central plateau, a distance
of about 7 km. The plateau is separated
from the northwestern coastal lowlands
by a precipitous escarpment.
V.R. Smith
Figure 15.5 Landscapes of Marion Island: A view from the Meteorological Station across
straw-coloured mires (dry Agrostis magellanica), dark green belt of fernbrakes towards the
snow-clad slopes of the central mountains carrying Polar Desert on the summits.
L. Mucina
(1) Older grey lavas that have been glaciated (smooth striated
outcrops, roches moutonées, moraines) and can be regarded as
Pleistocene. They build much of the higher ground, as a radially
arranged series of horsts, or wide ridges that are separated by
depressions, or grabens, filled by younger flows. Glaciations
have stripped the surface of the older lavas of all volcanic sur-
face features. This horst and graben topography of alternating
high and low segments has been ascribed to radial faulting
(Verwoerd 1971). Although evidence of radial fracturing does
exist, slope failure and land-sliding under the influence of grav-
ity probably played a more important role in shaping the larger
valleys (Chevallier 1986, McDougall et al. 2001).
(2) Younger black lavas that are all of Holocene age, because
none of them show any glacial erosion effects and some
Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
encroach upon moraines. They comprise aa and (less commonly)
pahoehoe flows and are associated with cinder cones (approxi-
mately 130 on Marion Island and 14 on Prince Edward). In addi-
tion, phreatomagmatic activity along coastal plains resulted in
surtseyan tuff cones and hyaloclastite deposits (Verwoerd &
Chevallier 1987).
3.4 Soils
Low temperatures and waterlogging result in low rates of chem-
ical decomposition and clay mineral synthesis, but are optimal
for accumulation of organic matter. The soils thus do not have
well-differentiated profiles and those under vegetation at lower
altitudes are highly organic. On the younger black lava flows
the peat is shallow, generally less than 1 m deep but profiles up
to 2 m have been found. On the older grey lava flows the peat
can be deeper, 3 m being the deepest found (Schalke & Van
Zinderen Bakker 1971).
Azorella selago cushions are the cardinal agent of soil formation
on the island; decomposition of old leaves in the cushion inte-
rior results in an organic matrix mixed with volcanic ash. The
cushions tend to grow toward the wind, the leeward margin
dying off and exposing the organic/ash substrate which is then
colonised by other plants (in fact, several vascular and bryo-
phyte species establish directly on the cushions, rooting them-
selves in the decomposing matrix of the cushion interior). On
flat areas the build-up of the organic matrix leads to impeded
drainage and a higher water table, encouraging bryophytes
which become the main agent of peat formation so that mire or
bog vegetation develops. Typically, the water table in mire veg-
etation is near or at the surface and the peat is light brown and
amorphous throughout the whole depth. Where drainage is
better and the water table fluctuates, there is some horizon dif-
ferentiation (Figure 15.7a); generally a light brown surface hori-
zon up to 20 cm deep overlies a yellow-orange to light brown
layer that extends to bedrock and which Smith (1978a) termed
‘organic clay’. Indurated iron pans and/or organic pans some-
times occur in this clay layer, in which case gleying takes place
and the clay is a yellow-grey colour beneath the induration.
Soils under slope plant communities possess better-developed
profiles, especially where they are deep and where there is
free drainage. Shallower soils (10–30 cm deep) consist of an
organic horizon directly overlying mineral material which is gen-
erally scoria (volcanic ash). Deeper slope soils, especially those
under fernbrake (Figure 15.7b), consist of a dark coloured
A-horizon containing much litter and up to 30 cm deep. Below
this a well-developed B-horizon of a ‘yellow-brown loamy clay’
(Smith 1976a) occurs, usually stained in the upper regions by
the organic A-horizon. Red indurated plinthic layers sometimes
occur in this B-horizon. In poorly drained soils on less steep
slopes a gleyed horizon occurs beneath the plinthic layer and
extends to bedrock.
Figure 15.7 Profiles of soil under (a) mire, (b) fernbrake slope and (c) fellfield vegetations.
Subhorizons are indicated thus: l = litter, f = fermentation, h = humus-enriched, s = sesquioxide-
enriched, Fe pan = iron pan, G = gley horizon (from Huntley 1971; courtesy of A.A. Balkema,
Cape Town).
(a) (b) (c)
706 Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
Fellfield soils (Figure 15.7c) consist of skeletal, shallow gravelly
‘loams’ (Smith 1976a, 1978a). An A-horizon occurs beneath
Azorella cushions but, where no plants are growing, the soil
surface is either bare or covered by a layer of small pebble-
sized scoria. Beneath the surface is a brown gravelly loam that
becomes orange-brown with depth and contains much scoria.
Bedrock is mostly within 10–15 cm of the surface, but some
fellfield soils are deeper.
The polar deserts of the islands do not have soils, merely coarse
scoria where any trapped moisture undergoes frequent freeze-
thaw cycles. Vegetation (mosses, also with Azorella selago at
lower altitudes) occurs mainly where seepage allows some
moisture accumulation or where rock surfaces are exposed to
moisture-laden winds.
True ornithogenic soils (soils essentially devoid of organic mat-
ter other than guano which overlies, but does not mix with,
weathered rock) as described from Antarctica (Syroechkovsky
1959, Ugolini 1972) do not occur on the islands. Even in pen-
guin rookeries the high rainfall prevents the build-up of guano
deposits. However, many soils in the coastal zone are heavily
influenced by birds and their profiles often contain evidence
that birds were present at the site during their development
(squid beaks, feathers; Smith 1977a). Following Tatur (1989),
these may also be considered as ornithogenic soils. Rates of
peat accumulation, especially around penguin rookeries, are
strongly stimulated (Lindeboom 1979) and the soils are much
darker (dark brown to black, compared with the light brown
peat in areas not influenced by birds; Smith 1978b).
Since the soils of the islands developed from weathering of vol-
canic ash under cool humid conditions, they are suspected of
containing allophanes as the dominant clay mineral, and the
horizons of lowland soil profiles have been described as clays,
organic clays or loamy clays (Huntley 1971, Smith 1976a, 1978a).
However, the few mineralogical analyses carried out so far have
not revealed crystalline or amorphous clay minerals (Gribnitz et
al. 1986). There is a large concentration of glass fragments in
the volcanic ash component of the soils. In many cases hydroly-
sis of the glass has lead to a yellow, brown or orange palagonite
(the precursor of allophane) and Gribnitz et al. (1986) consider
that what have been described as clays, organic clays or loamy
clays are actually palagonitised fine-grained volcanic ashes, the
cold climate and young age of the soils having precluded the
formation of clay minerals. According to Gribnitz et al. (1986)
the soils of the islands are formed from Holocene ash falls
as follows:
Accumulation of vegetable matter and raw humus on the
surface of ash by colonising plants.
Percolating humic acids stain the immediately underlying
ash and mobilise some iron by attacking the palagonitised
glass fragments.
Precipitation of the iron at lower levels as hydroxides, the
depth at which this occurs depending on the depth to the
water table.
Three features make it difficult to classify the island soils using
international systems: their immaturity, the negligible influence
of parent material on the profiles, and the marked effect of
slight variations in topography and wind exposure on these pro-
files. Taylor (1955) and Huntley (1971) classified Macquarie and
Marion Island soils, respectively, according to the most impor-
tant factors influencing their development. They regard soils
under slope plant communities as high-moor peat and those
of mire and bog areas as low-moor peat or fen and bog peat.
The red-brown loamy soils of slopes on Marion Island resem-
ble those of well-drained slopes and level ground beneath dry
grassland on South Georgia, which have been termed brown
soils by Lewis Smith & Walton (1975) since they approximate
arctic brown soils.
Conditions such as waterlogging, low temperatures and high
acidity lead to substantial amounts of organic acids in the soils.
This, with the intensive leaching effect of the high rainfall, will
favour podsolisation (Jenny 1941, Kononova 1966) and follow-
ing most climatic systems of soil classification, the island soils
should be predominantly podsolised. However, bleached eluvial
horizons are not found in any lowland soil profiles, although
Huntley (1971) and Smith (1976a) presented chemical data indi-
cating the presence of an eluviated horizon below, and stained
by, the A-horizon in some slope soils. Frequent or permanent
waterlogging at or below the soil surface favours gleying and
many of the lowland soils are gley (all horizons gleyed or semi-
gley) or gleyed (only the lower horizons are so affected). In this
respect they are similar to moorland soils of the southwestern
Chilean islands which were classified by Papadakis (1969) as
gley podsols (eluvial horizons masked by high organic matter
content), humic gley soils and peaty soils. Tentatively, slope
complex soils on Marion Island may be regarded as gley podsols
and those of the mire complex as humic gley soils or, where the
organic horizon lies directly on rock, as peat.
Soils under fellfield vegetation on Marion Island are approxi-
mately similar to fellfield soils of the northern hemisphere.
Noting this similarity, Huntley (1971) referred to the island
fellfield soils as rawmark. Taylor (1955) used the terms tundra
soils and dry tundra soils when referring to fellfield soils on
Macquarie Island. However, although subantarctic fellfield is
the vegetation type that most closely resembles tundra vegeta-
tion, the well-drained soils under subantarctic island fellfield
are actually very different from fellfield soils of northern hemi-
sphere tundra, which are mostly poorly drained, strongly gleyed,
mineral soils that overly permafrost.
3.5 Palaeohistory
All studies of geochronology to date have been carried out
on Marion Island. Little is known of Prince Edward Island. The
oldest K-Ar dates available for the lavas indicate an age of
450 000 ±10 000 years (McDougall et al. 2001). Since the island
is only 2° of latitude north of the Antarctic Convergence, past
falls in world temperature would have had the effect of shifting
the Convergence closer to, and even northwards of, the island
(Van Zinderen Bakker 1971). The glaciated surfaces of the old
grey lavas and intercalated sediments with glacial characteristics
testify to earlier glaciations. Hall (1978, 1981) described three
distinct tills of different ages and concluded that during the last
300 000 years the island was subjected to three glacial episodes,
each comprising a series of stades and interstades. At many
localities the tills are separated by thick sequences of interca-
lated basaltic lavas and pyroclasts, suggested to have erupted
during the interglacials as isostatic response to the removal of
the weight of the ice (Hall 1981, 1982). However, although
the two most recent periods of effusivity are clearly interglacial,
some of the earlier ones seem to have coincided with glacial
stages (McDougall et al. 2001).
A temperature drop of at least 3.5°C during the glaciations is
indicated from the periglacial evidence (Hall 1978); this sup-
ports previous conclusions derived from palynological (Van
Zinderen Bakker 1973) and ocean-floor sediment studies (Hays
et al. 1976). During the interglacials the temperatures were
as high as they are at present (Van Zinderen Bakker 1971). At
some localities palaeosols are thought to have formed in the
interglacial deposits (Hall 1978). Scott & Hall (1983) found that
Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
organic-rich basal sediments associated with the interglacial
deposits contain pollen spectra indicating a vegetation assem-
blage similar to that found today. Zonation in the pollen profile
suggests that during the interglacial phase the ocean shore was
close to its present position and that it moved some distance
away at the onset of the glacial stage.
Other workers (Kent & Gribnitz 1983, Gribnitz et al. 1986) con-
sider that Hall’s palaeosols and tills are tuffs and tuff breccias,
i.e. both are volcanogenic rather than glacigenic. They contend
that the interpreted stratigraphic successions and the palaeocli-
matological deductions of Hall (1978, 1981) are not substanti-
ated. However, morainic deposits occur at many sites on Marion
Island and it is not doubted that the island was glaciated in the
Quaternary. The ice cover of the last glacial began to disappear
rapidly about 12 000 years ago (Van Zinderen Bakker 1973)
and the youngest lavas that have been dated (15 000 ±8 000
years BP; McDougall 1971) may have erupted in response to
this disappearance. Radiocarbon dating of some mire peats on
the island indicates minimum ages of 3 180 ±120 to 4 020 ±65
years BP (Schalke & Van Zinderen Bakker 1971). However, it is
certain that many of the lava flows and overlying peat deposits
are much younger (Scott 1985, Gribnitz et al. 1986).
4. Flora and Major Vegetation Patterns
The vascular flora of Marion Island consists of 22 native species,
18 introduced species and three species` of unknown status
(N.J.M. Gremmen, pers. comm.). Of the alien plant taxa, 12 are
still present on the island, but the other six have disappeared.
Prince Edward has 21 native and three alien vascular species. A
large proportion of the indigenous species has a wide ecologi-
cal amplitude, and occurs over much of the range of habitats
on the islands. There is a much higher diversity of cryptogams;
for both islands ca. 100 moss, 42 liverwort and 100 lichen spe-
cies have been recorded. Mosses and liverworts dominate many
of the plant communities. Comprehensive descriptions of the
island plant communities have been provided by Huntley (1971)
and Gremmen (1981).
Because of the lack of trees and tall shrubs, the island presents a
bleak or barren appearance when viewed from offshore. Closer
inspection, however, reveals that the vegetation is not so sparse,
or even as uniform, as it appears. Huntley (1971) recognised
13 plant communities (he termed them ‘noda’) based on flo-
ristic composition and autecological characteristics of the spe-
cies. These were grouped into five complexes according to the
most important factor controlling their distribution, such as salt
spray, manuring, trampling, exposure and drainage. Gremmen’s
seminal work resulted in a phytosociological classification of
the islands’ vegetation that includes 41 plant communities.
Environmental information (soil depth, moisture content, pH,
loss-on-ignition, depth of groundwater, severity of manuring
and trampling by animals and of salt spray) was related to flo-
ristic composition of the plant communities in order to group
them into six community complexes. Subsequently, Smith &
Steenkamp (2001) added one more complex.
The complexes are:
The salt-spray (Crassula moschata) complex, restricted to
shore-zone areas strongly affected by wind-blown sea
spray. On the west coast the belt of salt-spray vegetation
extends up to 300 m inland but on the east coast it is
restricted to a narrower zone along the tops of coastal
The biotic (Callitriche antarctica–Poa cookii) complex, influ-
enced by trampling and manuring by animals. This com-
plex consists of a wide variety of communities, most of
which occur on the coastal zone near colonies of seals and
penguins. Inland, the influence of surface-nesters and bur-
rowing species is also manifested by the presence of com-
munities belonging to this complex.
The Blechnum penna-marina complex (fernbrake commu-
nities) that dominates the vegetation of well-drained low-
land slopes.
The Acaena magellanica–Brachythecium complex, which
forms at mire and lowland slope sites where there is pro-
nounced lateral subsurface water movement. Communities
of springs, flushes and drainage lines belong to this
The Juncus scheuchzerioides–Blepharidophyllum densifo-
lium mire complex, dominated by bryophytes and grami-
noids and occurring on wet peat.
The fellfield (Andreaea–Racomitrium crispulum) complex
which forms in rocky habitats strongly exposed to wind.
This complex, consisting of communities of the cushion-
forming Azorella selago, bryophytes and lichens, domi-
nates the vegetation above 300 m altitude. Fellfield com-
munities also occur at lower altitudes where they exhibit
fairly high (up to 60%) aerial vegetation covers.
Polar desert. An important vegetation type (it covers about
120 km2 of the 290 km2 total area of Marion Island, but a
much smaller proportion on Prince Edward Island) that was
not included in the phytosociological study of Gremmen
(1981) and for which there is no published ecological infor-
mation. It was recognised as a cardinal habitat type on
Marion Island by Smith & Steenkamp (2001; see also Smith
et al. 2001) who placed it in a polar desert complex. It is the
only vegetation type above 550 m altitude on the islands.
Soil moisture and exposure to wind are the most important
factors determining variation between plant communities. The
wet-sheltered dry, exposed gradient is also associated with a
change from organic peat at sheltered sites to mineral soils at
exposed sites. The type of plant community which develops at
any particular point on the wet-dry gradient is mainly deter-
mined by manuring and trampling by seabirds and seals and/or
by wind-blown salt spray.
Huntley’s (1971) ecological classification and Gremmen’s (1981)
phytosociological classification implicitly define the terrestrial
habitat types that occur on the island according to the main
patterns of abiotic and biotic variation in the ecosystem, since
both consider the role of the main environmental and biological
forcing variables (moisture, exposure, parent soil material, salt
spray and manuring and trampling by seals and seabirds) on the
island. Canonical correspondence analysis and cluster analysis
of soil chemistry and botanical variables for 176 sites on the
island recognised 23 habitats in seven habitat complexes in a
‘habitat classification’ (Smith & Steenkamp 2001) that supports
all of Huntley’s and Gremmen’s conclusions and that closely
reflects the between-habitat variation in the relative magni-
tudes of the main abiotic and biotic forcing variables.
The habitat classification gives a broader grouping (21 habi-
tats) than the phytosociological one (41 communities, plus
variants of some communities). However, because it was based
on plant types and soil chemistry considerations, rather than
just plant species, the habitat groupings can be considered as
nodal assemblages of plant communities having vegetation and
edaphic affinities. They can thus be considered analogous to
the six community complexes of the phytosociological classifi-
cation since those complexes were recognised on the basis of
708 Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
ecological and edaphic factors as well as floristic considerations.
In this respect, the habitat classification gives a more detailed
grouping of sites than the phytosociological scheme. It is also
more detailed than Huntley’s five nodal ecological complexes.
Biotically influenced sites, in particular, are floristically (and
edaphically) the most diverse on the island and these are distin-
guished as seven habitats in three separate complexes, rather
than in one complex as in previous schemes. In contrast, the
various types of drainage lines (also a floristically and edaphi-
cally diverse group) on the island are less finely categorised
within the habitat framework than they are by the phytoso-
ciological classification. In some respects, the habitat analysis
yields a similar dispensation of the drainage lines as does the
ecological complex scheme of Huntley (1971) where they are
included as two noda, together with mires and bogs, as part
of the swamp complex. The main reason for establishing the
habitat classification was to provide a framework of structur-
ally and functionally well-defined units against which to detect
and evaluate the effects of climate change and human-induced
perturbations. Smith et al. (2001) describe the altitudinal distri-
butions of the habitats and provide a scenario of how climate
change is expected to affect them.
5. Vegetation Dynamics
5.1 Vegetation Succession
The very small number and wide ecological amplitude of the
vascular plant species prevent the development of well-defined,
floristically distinctive seral communities. However, a simplistic
scheme of vegetation succession is as follows:
Rocky plateaus are colonised by lichens and cushion-forming
mosses and dicotyledons (mainly Azorella selago). These initi-
ate peat formation and also act as traps for fine, wind-blown
volcanic ash. The peaty material fills the crevices and porous
structure of the lava and impedes drainage. If the rate of water
acquisition by the area is greater than the rate of drainage,
diverts the fellfield-fernbrake succession to the one toward mire
or bog vegetation.
The pathway of primary succession on cinder cones follows a
simple scenario; the lower slopes are less exposed to wind, more
stable and climatically less extreme than the upper reaches of
the cones and this allows the establishment of Azorella selago.
Several vascular and cryptogam species grow as epiphytes on
the cushions and organic matter built up under the cushions
consolidates the scoria. This pattern of colonisation and succes-
sion proceeds up the slope (Figure 15.8), and the recent warm-
ing of the island is suspected to be increasing the rate at which
this is occurring.
At coastal rocky sites subjected to salt spray, the fellfield is dom-
inated by Azorella selago and Crassula moschata. Peat accu-
mulation in these areas results in a well-developed salt-spray
vegetation dominated by C. moschata and Cotula plumosa.
Impeded drainage results in a bog salt-spray community in
which C. moschata and a few liverwort species occur.
Strong lateral flow of water may modify the vegetation at any
level of succession, leading to a group of plant communities
characterised, on slopes and stream banks, by the abundance
of Acaena magellanica in the herb layer and Brachythecium
subplicatum or B. rutabulum in the bryophyte stratum. These
communities are associated with drainage lines, water tracks,
flushes and springs in slope areas. Drainage lines in mires are
characterised by a dominance of Breutelia integrifolia and
Bryum laevigatum.
Animal activities may have a marked influence at any stage of
succession. Other than Salvin‘s Prions (Pachyptila salvini), ani-
mals seldom establish nests or otherwise exploit fellfield areas,
preferring mires and well-vegetated slopes. In the salt-spray
zone, sites influenced by animals support luxuriant Cotula
plumosa–Poa cookii vegetation that is especially characteristic
of areas surrounding seal and penguin colonies (Figure 15.9). At
mire and bog sites the influence of animals causes the develop-
ment of wet biotic communities in which coprophilous, tram-
pling-resistant species become established (mainly Callitriche
Figure 15.8 Progressing primary vegetation succession in upslope trend on the northern slopes
of Junior’s Kop. The green patches on the slope are cushions of Azorella selago, with white
patches of dry culms of Agrostis magellanica.
bryophyte-dominated bogs develop.
Peat accumulation raises the bog surface
above the water table and mire-grassland
vegetation dominated by graminoids and
bryophytes results. Further peat accumu-
lation may make the surface even drier
and fernbrake vegetation then devel-
ops. However, closed fernbrake on flat,
level areas is rare. Dry mire communities
dominated by Blechnum penna-marina,
Racomitrium lanuginosum and Uncinia
compacta generally represent the culmi-
nation of the wet-dry succession on level
Where drainage from the fellfield area is
unimpeded, the cushion plants develop
a stable peat for fernbrake elements to
develop. At first these elements may
arise epiphytically on, or in the shelter
of, the cushions, but eventually extend
across the peat to form an open type of
fernbrake that is often succeeded by a
closed carpet of Blechnum penna-marina.
This sequence from fellfield to fernbrake
occurs predominantly on slopes. On flat
plateaus, impeded drainage (caused by
peat blocking the drainage channels)
L. Mucina
Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
antarctica and Montia fontana). Where fernbrake communi-
ties are colonised by burrowing birds, closed tussock grassland
dominated by Poa cookii develops (Smith 1976b). However,
the occurrence and extent of tussock grassland (especially the
inland type, the Leptodontio proliferi–Poetum cookii brachythe-
cietosum rutabuli of Gremmen 1981) has declined markedly
since the early 1970s. For example, tussock grassland occu-
pied 7.8 ha of a 1 040 hectare inland study site on the island’s
eastern coastal plain in 1971–1972 (Smith 1976a, 1977b). This
shrank to 0.9 ha by 1991 (Smith et al. 2001). This observa-
tion was made on only ca. 8% of the lowland area supporting
closed vegetation but the same trend has been observed over
the whole island since 1975 (N.J.M. Gremmen, unpubl. obser-
vations of permanent quadrates). Mostly, the inland tussock
grasslands have been replaced by closed fernbrake or dwarf-
shrub fernbrake. The reason for the shrinkage of tussock grass-
lands is almost certainly related to the influence of feral domes-
tic cats (Felis catus) on burrowing petrel and prion populations
at the island during the same period. Cats were introduced to
the island in 1949 and their population grew rapidly, by 23%
per year; by the mid 1970s it was estimated at ca. 2 100 adults
(Van Aarde 1979). Undoubtedly, the cats had a devastating
effect on the petrel species of the islands; in the mid-1970s
they were eating a minimum of 635 000 petrels and prions
each year (Van Aarde 1980). Where petrels and prions burrow
into slopes, their guano enriches the soils (Smith 1976b) and
this results in a succession to tussock grassland. When the birds
disappear, the tussock grassland reverts surprisingly quickly to
one of the non-biotically influenced slope communities. In the
case of the tussock grassland studied by Smith (1976b), there
were 1.2 burrow entrances per square metre in 1972; by 1987
there were no burrows and the area was occupied by closed
fernbrake (in the wetter areas by dwarf-shrub fernbrake). With
the elimination of cats (the last was killed in 1991; Bester et
al. 2000) there are indications that the petrel and prion popu-
lations are recovering (Cooper et al. 1995) and it remains to
be seen whether this will reverse the shrinkage of the tussock
grasslands. Tussock grassland is very common and widespread
on Prince Edward Island, where cats were never present.
Another biotic factor that is increasingly affecting the veg-
etation of the islands, particularly the coastal zone, is the ris-
ing population of fur seals. Trampling by these animals has
degraded extensive areas of tussock grassland. A similar situ-
ation occurred on Bird Island, South Georgia, where fur seals
destroyed most of the tussock grassland (Bonner 1985).
5.2. Primary Production, Decomposition and
Nutrient Cycling
The subantarctic climate is oceanic and in many respects cool-
temperate rather than polar. Hence, the islands do not experi-
ence bitterly cold or dry periods and the vegetation growing
season is long. Additionally, the plants are all C3 photosynthetic
types that are able to maintain fairly high rates of photosyn-
thesis and growth under the consistently low temperatures.
Annual primary production is consequently high, comparable
to even the most productive temperate herbaceous plant com-
munities (Smith, 1987a, b). The high annual production results
in a substantial annual requirement for nutrients by the vegeta-
tion (Smith 1987c, 1988a). There is a paucity of grazers and
predators (there are no indigenous herbivorous mammals and
insects play only a small role in herbivory), so most of the energy
and nutrients incorporated in primary production goes through
a detritus chain, rather than a grazing chain. Decomposition,
with the concomitant release of nutrients, is overwhelmingly
the main bottleneck in nutrient cycling and primary produc-
tion for most of the plant communities (Smith 1988b, Smith &
Steenkamp 1992a). The consistently cool climate, high cloud
cover and very high rainfall result in low soil temperatures and
excessive soil moisture, both of which restrict microbial activity.
That is the reason why peat accumulation (a major determinant
of vegetation succession, Section 5.1) is such a conspicuous
part of the ecosystems of the islands, at least at lower altitudes
where plant growth is possible. Decomposition mediated by soil
micro-organisms alone is simply too slow to satisfy the large
annual requirement for nutrients by the vegetation. There are
large concentrations of soil macro-invertebrates (earthworms,
moth larvae, weevils, snails etc.) that, by feeding on plant litter,
are responsible for the bulk of energy flow and nutrient cycling
on the island (Crafford 1990a, Smith & Steenkamp 1992b). The
activities of these animals are strongly temperature-dependent
(Klok & Chown 1997, Chown et al. 1997) so, providing noth-
ing else threatens their populations, increasing temperature will
result in enhanced rates of litter consumption and hence of
nutrient release, allowing the potential for increased primary
production due to elevated temperature to be realised. Threats
to the soil macro-invertebrates of the islands, and hence to eco-
system functioning, are discussed in Section 7.2.
6. Current Conservation Status of the
The National Parks Act (Act 42 of 1962) does not list Marion or
Prince Edward Islands in its schedule of National Parks; neither
do the islands enjoy Provincial Nature Reserve status although
juristically they are considered as part of the Western Cape.
Statutory environmental protection for the islands (to the low
water mark) is provided by the Environmental Conservation Act
(Act 73 of 1989), in terms of which both islands were proclaimed
as Special Nature Reserves on 3 November 1995. The Maritime
Zones Act (Act 15 of 1994), Article 54 of the Sea Fishery Act
(Act 12 of 1988) and The Fishing Industry Development Act (Act
86 of 1978) afford some control over exploitation of the mari-
time zone around the islands. Additionally, the Seabirds and
Seals Protection Act (Act 46 of 1973) controls the capture and
killing of most species of seabirds and seals on the islands and
South Africa is an original signatory of the Convention on the
Conservation of Antarctic Marine Living Resources (CCAMLR),
which, while not impinging on national sovereignty rights,
focuses on conservation and management of marine living
resources on and around the islands.
Proclamation of the islands as Special Nature Reserves required
the formulation of a Management Plan for the islands, which
was published in 1996 (Department of Environmental Affairs
and Tourism 1996). The primary aim of the management plan
is the conservation and sustained preservation of the islands’
unique ecosystems for all people of South Africa and for the sci-
entific community at large. The plan recognises that the Prince
Edward Islands are an integral part of South Africa’s national
heritage and territorial integrity and that a rational and rigor-
ous management of the islands is in keeping with an emerging
international ethos and political order that recognises a strong
environmental ethic.
A Prince Edward Islands Management Committee (PEIMC) was
established to implement the management plan, which contains
many provisions and regulations to ensure that activities on the
islands take place with the minimum of environmental distur-
bance. The Management Committee advises the Department of
Environmental Affairs & Tourism (DEAT) on planned and exist-
ing activities on the islands. A Conservation Officer and Team
Leader are appointed to each relief team to the islands and
710 Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
together they are tasked with the day to
day ‘management’ of the islands in terms
of the management plan.
Under the management plan four zones
have been designated on the islands.
Zone 1 is a service zone that includes the
Marion Island meteorological station and
a small area surrounding it. Zone 2 is a
buffer zone around the station which
extends in roughly triangular shape
between Trypot Beach, Junior’s Kop
and Ships Cove. Areas around the field
huts are also considered as Zone 2 areas
because of limited human impact. The
rest of Marion Island is a Zone 3 or wil-
derness area, with the exception of areas
where there are Southern Giant Petrel
colonies, Gentoo Penguin colonies, three
Wandering Albatross study colonies, and
Greyheaded Albatross colonies. These
colonies are Zone 4 or protected zone
areas. The whole of Prince Edward Island
is Zone 4.
Entry to these zones is controlled by a
permitting system. The permits are issued
by DEAT on the advice of the Prince
Edward Islands Management Committee.
the thalli remain hydrated. Incorporating hydration/desicca-
tion cycles into the model resulted in a very substantial low-
ering of annual net carbon exchange. However, attempts to
include the increase in aridity known to have occurred at the
island since 1971 yielded conflicting scenarios for the effect
on annual carbon acquisition, depending on whether atmos-
pheric drying or thallus drying was considered.
Similar studies relating physiological performances to microcli-
mate factors have been carried out on the insects of the islands.
For each of the six weevil species of the islands, upper lethal
temperature corresponds closely to the maximum microclimate
temperature in the habitat of the particular species (Van der
Merwe et al. 1997), suggesting that climatic warming might
be deleterious to the weevils’ survival. Klok & Chown (1997)
examined the thermal tolerance and desiccation resistance of
larvae of a flightless moth (Pringleophaga marioni, the most
ubiquitous insect on the islands) and concluded that should the
warming trend on the islands continue, it will have a profound
negative effect on the moth population, especially if the warm-
ing is accompanied by increasing aridity.
It was suggested in Section 3.1 that the climatic change occur-
ring in the subantarctic is associated with changes in sea level
(atmospheric and oceanic) circulation patterns, changes that
themselves have implications for the biota and ecosystems
of the islands. For instance, approximately one million pairs
of burrowing petrels and prions occur on Marion Island, and
possibly even more on Prince Edward Island. By feeding in the
sea and depositing guano on land they represent an important
source of marine-derived nutrients and energy (Smith 1976b)
and are therefore a major driving force in vegetation succession
(Smith & Steenkamp 2001). The birds use atmospheric frontal
systems to move between the islands and their feeding areas
(Mendelsohn 1981), which may be several hundred kilometres
distant, in the surrounding ocean. If atmospheric circulation
patterns are changing, then the birds may not be able to reach
their usual feeding grounds; alternatively, associated changes
in oceanic circulation may move the feeding grounds to other
localities. Hence, changing sea level circulation patterns may
Figure 15.9 King Penguin (Aptenodytes patagonica) rookery on Marion Island, surrounded
by tussock grassland (Poa cookii) vegetation. Light green plants are Cotula plumosa.
Generally, all visitors to the islands receive permits for Zones 1
and 2. Personnel involved in field research are permitted for
Zone 3 and some for work in bird colonies that have Zone 4
status. Other team members may apply for permits to Zone
3 or may accompany research personnel doing field work on
the permits of the research personnel involved. Visits to Prince
Edward Island take place only with special permission and under
strict conditions. These include intensive efforts to reduce the
risk of alien introductions to the island.
Collection of any material from the islands, irrespective of the
zone in which the collections are carried out, is also regulated
by a permit system.
7. Threats to the Islands' Ecosystems
7.1 Implications of Climate Change
A changing climate has implications for the indigenous biota
of the islands, especially the terrestrial species, which have
all evolved under the cool, humid conditions typical of sub-
antarctic islands. Smith & Steenkamp (1990) proposed some
scenarios of the direct effects of warming (and drying) for the
biota of Marion Island and there have been a few case stud-
ies on particular taxa. Most have concentrated on the eco-
physiological responses of particular organisms, or groups of
organisms, to abiotic factors such as moisture, light and tem-
perature. For example, Smith & Gremmen (2001) showed that
the light/temperature response of photosynthesis in the lichen
Turgidosculum complicatulum on Marion Island is such that
under the prevailing climatic regime the lichen would exhibit
near maximal photosynthesis rates for 75% of the photope-
riod over the year, if sufficiently hydrated. A model of the pho-
tosynthetic response predicts that changes in temperature and
radiation by the amounts known to have occurred in the past
few decades, and even more drastic changes (temperature
up by a further 2°C, radiation up by 10%), would negligibly
affect the annual amount of carbon acquired, again provided
V.R. Smith
Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
and adults), flies, spiders, earthworms and snails. The mouse
population is strongly temperature-limited and is increasing,
probably as a result of ameliorating temperatures. Several stud-
ies have unequivocally established that house mice are cardinal
determinants of the population dynamics of terrestrial macro-
invertebrates on the island. Crafford & Scholtz (1987) showed
that there are striking differences between Marion Island and
Prince Edward Island (which is only 22 km away) in the size,
structure and composition of their macro-invertebrate popula-
tions (and also in the maximum body size attained by the vari-
ous species) and ascribed this to the fact that mice do not occur
on Prince Edward. The severity of the impact that mice have
on their invertebrate prey on Marion Island is illustrated by the
fact that mice daily consume between 30 and 107 g of moth
(P. marioni) larvae per hectare (depending on habitat; Smith et
al. 2002), or an average consumption for the island’s eastern
coastal plain of 65 g per day (Crafford 1990b). This is equivalent
to 0.7% of the annual mean biomass of larvae, which take a
long time to mature (2 to 10 years; Crafford 1990a, b). Such
a high daily consumption rate must have severe implications
for an insect with such a long life cycle, especially since almost
the whole of it comprises the non-reproductive phase. In the
case of weevils, sympatric speciation associated with assortive
mating owing to differences in body size is occurring in the
Ectemnorrhinus similis species complex on the island (Chown
1990), and size-selective predation by mice might be consider-
ably enhancing the rate at which this is taking place (Chown &
Smith 1993).
Predation by mice on soil invertebrates directly influences other
components of the biota of the island. For example, the Lesser
Sheathbill (Chionis minor) is the only non-migratory bird species
on the island and relies on soil macro-invertebrates as food in
winter. Smith & Steenkamp (1990) proposed that escalating
predation by mice on soil invertebrates could adversely affect
the sheathbill population. Huyser et al. (2000) subsequently
reported that between the mid-1970s and mid-1990s the
island’s sheathbill population decreased by 23%, whereas the
one on Prince Edward Island did not change.
More insidious, but probably even more profound than these
direct effects of house mice on the biota of the islands, is
their influence on ecosystem functioning (primary productivity,
decomposition and nutrient cycling). As was stated in Section
5.2, primary production on the island is high, which means that
the vegetation has a large annual requirement for nutrients,
most of which is met by litter-feeding macro-invertebrates.
Increasing temperature is expected to increase productivity
and nutrient demand even further. Since soil microbiological
processes on the islands appear to be more strongly limited
by waterlogging than by temperature (Smith et al. 1993), the
increased temperature might not translate into substantially
greater rates of nutrient release by soil microbes alone. However,
the activities of detritivorous macro-invertebrates, which are
responsible for the bulk of energy flow and nutrient cycling
on the islands, are strongly temperature-dependent (Klok &
Chown 1997, Chown et al. 1997) and increasing temperature
will result in enhanced rates of litter consumption and hence
of nutrient release, which will allow the potential for increased
primary production due to elevated temperature to be realised.
An increasing mouse population, through enhanced predation
pressure on soil invertebrates, will decrease rates of nutrient
cycling and cause imbalances between primary production and
decomposition and hence affect, amongst other things, rates of
peat accumulation. This, along with more direct effects of mice
(e.g. granivory; Chown & Smith 1993) has important implica-
tions for vegetation succession and ecosystem structure and
functioning on the island.
be expected to influence the breeding success and population
densities of these animals, with obvious implications for nutri-
ent cycling and vegetation succession on the islands. Similar
considerations probably apply to the large populations of seals,
penguins and albatrosses.
Even the identity of Marion Island (with its sparse, low-growing
vegetation devoid of trees, or even shrubs, and an avifauna con-
sisting entirely of seabirds) is challenged by the changes in its
climate over the past 30 years or so. In Section 2 it was shown
that the most widely accepted definition of ‘subantarctic’ (at
least its terrestrial connotation) is based on a hybrid scheme
derived from a vegetation-based classification by Wace (1965)
and a climatological classification by Holdgate (1964). The sub-
antarctic region has no months warmer than 8.5°C (considered
the lower temperature limit for tree growth) and occurs south
of the southern limit of tree or shrub growth (distinguishing it
from the cool-temperate zone to the north) and north of the
southern limit of closed phanerogamic vegetation (delimiting it
from the maritime or low Antarctic to the south). Prior to 1990
Marion Island fitted comfortably into the subantarctic category
but in the 1990s mean temperatures for January, February and
March were between 8 and 8.5°C. Only slight further warming
will put the island into the south cool-temperate zone, a region
where the oceanic islands are inhabited by tall shrubs, trees and
land birds (even parakeets!).
7.2 Alien Flora, Alien Animals and Human
Probably of greater importance than the direct effects of climate
change on the biota of the islands, is that a warmer climate
will increase the ease with which the island can be invaded by
alien species (Bergstrom & Chown 1999). Such invasions, espe-
cially by functional groups generally not found on subantarctic
islands (e.g. vertebrate and invertebrate predators, mamma-
lian herbivores, particular groups of pathogens), represent by
far the greatest threat to the indigenous biota of subantarctic
islands. An example of this is the deleterious effect that intro-
duced organisms are having on the Kerguélen cabbage (Pringlea
antiscorbutica), the only species in its genus, endemic to four
subantarctic island groups. The species is the last remaining
relict of a once extensive circum-antarctic flora. The distribu-
tion of P. antiscorbutica on Îles Kerguélen and the Îles Crozet
has decreased in the last 25 years due to grazing by introduced
rabbits Oryctolagus cuniculus (Chapuis 1995). On Marion Island
its distribution and abundance has declined alarmingly over
the past 25 years for several reasons, all to do with invasive
alien biota. The European Slug (Deroceras caruanae) was intro-
duced to the island in the mid-1960s (Smith 1992) and the
Kerguélen cabbage is one of its preferred food items there. The
Diamondback Cabbage Moth (Plutella xylostella), a major pest
of crucifers worldwide, is also a recent arrival on the island (first
discovered in 1986; Crafford & Chown 1990) and P. antiscorbu-
tica is its only host plant there. Many stands of the cabbage are
severely infected by the moth. Botryotinia fuckeliana (conidial
state: Botrytis cinerea), the causal organism of grey mould rot in
crucifers and other vegetable crops, has also reached the island
only quite recently (probably through vegetables sent as food
for the island personnel; Kloppers & Smith 1998). Many stands
of the cabbage have been infected by the fungus, whole plants
collapsing into a black slimy residue.
The introduced house mouse (Mus musculus) on Marion Island
offers a particularly striking example of how profoundly an
invasive alien organism can influence the biota and ecosystem
of a subantarctic island. House mice on the island feed mainly
on soil macro-invertebrates such as moths and weevils (larvae
712 Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
8. The Future
8.1 Future Research
Biological research on Marion and Prince Edward Islands is
financially supported by the South African Antarctic Research
Program (SANAP) of the Department of Environmental Affairs
and Tourism (DEAT), which until 2003 also co-ordinated and
defined the focus and objectives of the research. DEAT, with
the Department of Public Works, maintain research laborato-
ries at the meteorological station (Figure 15.10). Smith (1991)
provided a synopsis of the directions and priorities of ter-
restrial biology research on the Prince Edward Islands during
the 1970s and 1980s. Since 1990 the focus of the research
has been very largely directed toward increasing the ability to
rationally manage and conserve the biota and ecosystems of
the islands, including predicting the effects of the very marked
climate change that is occurring there. Special attention has
been given to monitoring and conservation projects, in accord-
ance with national (e.g. requirements for managing Special
Nature Reserves) and international (e.g. CCAMLR) obligations.
The directives identified by SANAP and against which terrestrial
biology research projects carried out between 2000 and 2004
were being evaluated and are being supported by DEAT, were:
Factors influencing spatial and temporal patterns in
Indigenous and introduced species: differential responses
to environmental change.
Adaptation, plasticity and response.
In 2003 the research function of SANAP was transferred to the
Department of Science and Technology (DST), and is co-ordi-
nated by the National Research Foundation (NRF). Logistics and
management of the islands remain the responsibility of DEAT.
From 2007 research at the islands will be aligned with the themes
of the International Council for Science's International Polar Year
2007–2008 program (htp://
8.2 Development and Conservation Threats
The DEAT is committed to the conservation and preservation
of the biota and ecosystems of the islands and in terms of the
increasing interest in tourism to the islands. In November 2002
the South African supply vessel, the S.A. Agulhas, was char-
tered to take a large contingent of birdwatchers to both islands,
although no one was allowed ashore. There is evidence that
two unauthorised tourist landings occurred at Prince Edward
in the late 1990s and since 1996 several tour operators have
applied to DEAT for permission to visit the Prince Edward Islands
and land on Marion Island.
Since the current Prince Edward Islands Management Plan does
not explicitly address tourism, the PEIMC recommended to the
Minister of Environmental Affairs and Tourism that permission
for those visits be denied. However, the PEIMC recognised
that pressure for tourism activities at the islands was likely to
increase, that such visits would be far more preferable if under-
taken under controlled, monitored circumstances, and also that
tourism had a positive spin-off in terms of generating income for
SANAP and increasing public awareness of the unique nature
of the islands and the research and conservation efforts being
carried out on them. Accordingly, the PEIMC recommended
to the Director General of DEAT that an environmental impact
assessment of tourism at Marion Island should be undertaken.
A subcommittee of the PEIMC was tasked with the EIA, with
the following terms of reference:
Investigate the possibilities for controlled, limited tourism
to Marion Island.
Undertake an EIA of the impact of tourism at certain sites
on the coast of Marion Island between Ships Cove and
Trypot Beach.
Investigate the possibility of construction facilities to house
tourists in the event of an emergency, and of erecting
structures such as catwalks that would mitigate the impact
of people on the biota of the island.
The recommendations of the subcommittee (Heydenrych &
Jackson 2000) were that large tours of more than 100 per-
sons not be allowed under any circumstance, but that smaller
groups of up to 100 persons may land on Marion Island under
a permitting system under an Impact Management Plan to be
drawn up by DEAT. Tour visits would be limited to Zones 1 and
2 on Marion Island and no tourism should be allowed on Prince
Edward Island. The rest of the recommendations relate to the
Figure 15.10 Meteorological Station and scientific base on the eastern lowland plain of Marion
Island. Large cinder cone behind the station is Junior's Kop.
proclamation of the islands as Special
Nature Reserves, the Management Plan
must be adhered to in all future research
and logistic activities. Two aspects have
arisen quite recently that will test that
The Antarctic is increasingly fascinating
tourists and antarctic and subantarc-
tic tourism is growing faster than the
average for tourism worldwide. Tourism
visits to the Antarctic Peninsula islands,
South Georgia and the Falkland Islands,
the New Zealand subantarctic islands,
Macquarie Island and the French suban-
tarctic islands have increased exponen-
tially over the past 25 years. The Prince
Edward Islands have largely escaped
this trend, mainly because in addition to
being isolated they are not near other
sites of tourist interest that can provide
alternative possibilities for the tour oper-
ator to satisfy clients should the weather
be bad during the time that the ship visits
them. However, of late there has been an
V.R. Smith
Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
timing and number of tour ship visits, the size of landing groups,
the infrastructure and manpower requirements needed before
tourism activities would have an acceptable minimum impact on
the ecosystem of the islands, the use of income from tourism
to pay for that infrastructure to and monitor the effects of tour
visits, and the need for tour operators to have comprehensive
insurance against unforeseen circumstances that would nega-
tively impact on the islands, such as oil spills that would require
clean-up, search and rescue operations and emergency board
and lodging. The subcommittee specifically recommended that
no additional accommodation facilities be erected on the island
to cater for tourists.
One recommendation of the subcommittee is especially of
note—that tourism increases the risk of introducing new alien
organisms to the island (see the analysis by Chown et al. 1998
showing very convincingly that increasing numbers of human
occupants increase the risk of propagule transfer to an island).
Most importantly, the recommendation specifically points out
that new introductions might lead to severe impacts on the
biota of an island, the costs of mitigation of which might be
extremely high. An example is the possible introduction of rats
to Marion Island. That would have an enormous impact and the
expense of eradication would likewise be enormous, judging
from the cost of successful eradications of rats from the much
smaller and more accessible islands off New Zealand. An idea
of the cost is given by the 1995 estimate made at a workshop
on the possibility of eradicating house mice on Marion Island
(Chown & Cooper 1995) that the poison bait alone would cost
R2.75 million per year, and that the eradication programme
would take several years. Added to that is the cost of spreading
the bait by air (estimated at 2 000 hours helicopter flying time)
and of putting measures in place to reduce the very real risk
of poisoning scavenging birds—about two thirds of the skua
population on Enderby Island, south of New Zealand, was killed
through taking bait put out for rabbits (Torr 1993). Assuming
that eradication of rats on Marion Island could be completed
in three years, the 2003 estimate of total cost (manpower, bait,
helicopter and ship costs) was ca. R37 million. The report of the
tourism subcommittee pertinently, and very correctly, points out
in its recommendations that such high mitigation costs cannot
realistically be claimed from tour operators. The reality is also
that South Africa does not have the resources to undertake
such an exercise, which means that the introduced rats will
become a part of the ecosystem of the island, an ecosystem
that will be very different and much less valuable as a laboratory
for furthering ecological knowledge and theory.
For this reason the possibility that tourism to the islands might
become a non-option is to be welcomed. The Draft National
Environmental Management Protected Areas Bill published in
2002 lists three main types of protected areas, Special Nature
Reserves, Nature Reserves, and National Parks. Tourism is explic-
itly stated as one purpose for declaring a protected area as a
Nature Reserve or a National Park, but the only stated purposes
for establishing Special Nature Reserves are ‘to protect highly
sensitive, outstanding ecosystems, species, geological or physi-
cal features in the area and to make the area primarily available
for scientific research or environmental monitoring’.
The second issue that will soon test DEAT’s commitment toward
the conservation and preservation of Marion Island, in particu-
lar, is that construction of a new base station on the island
commenced in September 2003. This will take four years and
will be followed by the dismantling and removal of the old sta-
tion. Such activities will involve large numbers of personnel and
the importation of a huge amount of material. DEAT and the
Department of Public Works (which is carrying out the con-
struction) have undertaken to put all possible measures in place
to ensure that the activities do not result in new introductions
of alien organisms and that they will be undertaken in a way
that reduces environmental disturbance to a minimum. The
acid test of the seriousness of their commitment will be the
situation in five years’ time. If at that time the island possesses a
new ultramodern base erected over a wasteland of bare eroded
peat and adjacent an urban dump (the site of the current base),
and a whole suite of new alien organisms, then DEAT will have
failed the test.
9. Definition of Biomes, Vegetation and
Mapping Units
Spatial variation of the vegetation of the islands occurs on a very
small scale, due to the influence of the hummocky topography
formed by the lava flows on aspects such as insolation, shelter
from wind, moisture deposition, retention and drainage, and
peat accumulation. Except for fellfield and polar desert, large
expanses of single vegetation types are quite rare; at low alti-
tudes (<200 m) even very small areas generally contain several
vegetation types. For instance, a vegetation map by Gremmen
(1981, p. 72) shows 18 plant communities sharing an area of
0.5 ha on Marion Island. Eleven of these are mire communities,
two are biotic, two are slopes and one is fellfield. Other than
the mires, the rest of the communities occupy patches mostly
smaller than 20 m2. Mapping at the scale of the maps pre-
sented here would show the vegetation as mire. Smith (1976c,
p. 17) provides a vegetation map for a 1.5 ha area 500 m north
of the one in Gremmen (1981), on the same lava flow. There,
the vegetation is mapped more coarsely, not according to phy-
tosociological associations. There are four slope vegetation
types, one mire type and one fellfield type. Most of the types
are represented in more than one patch but the total areas
occupied by each type are more or less equal. Together, the
four slope types are more extensive than mire or fellfield, so at
the scale of the maps here the vegetation would be indicated as
slope vegetation. However, between the two sites the lava flow
contains a similar-sized site dominated by fellfield and spatially
the three sites (mire-dominated, slope-dominated and fellfield-
dominated) cannot be depicted separately at the scale of the
maps presented here. The particular lava flow extends about 2
km south, 1 km north, and about 4 km inland of the two sites.
Overall, this pattern of either mire, slope or (less commonly)
fellfield communities being locally dominant is repeated over
and over across the lava flow, which is only one of many similar
flows that comprise the low altitude region of the island. All
that can be said is that the lavas are occupied by a mosaic of
slope, mire and fellfield communities.
The situation is even more complex on the coast, since salt-spray
plant communities and communities influenced by animals are
added to the mosaic. It is only at higher altitudes that more
uniform vegetation types (fellfield and polar desert) occur over
expanses large enough to be able to delineate them at the reso-
lution of the map. Even there, the transition between fellfield
and the mire-slope fellfield mosaic is not a sharp one, nor is the
transition from fellfield to polar desert. None of those units can
be distinguished on the aerial photographs of Marion Island
(there are no photographs for Prince Edward). To compile the
maps presented here, the mapping units were delineated on
basis of field research (vegetation samples, photographs taken
from known locations and other field observations done during
on-foot and helicopter-assisted expeditions).
Here we distinguish three groups of vegetation (and mapping)
714 Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
Marine Macroalgal Vegetation contains only one vegetation
(mapping) unit: Subantarctic Kelp Beds. Due to a lack of
appropriate field data we do not show this unit on the maps
of the islands.
Subantarctic Tundra. This group corresponds to the extent of
Subantarctic Tundra Biome and is found on the islands’ lower
altitude (<300 m) regions. Vegetation units ST 3, 4, 5 and 6,
form an intricate small-scale spatial mosaic. In terms of
Gremmen’s (1981) classification this mosaic encompasses,
respectively, the Juncus scheuchzerioides–Blepharidophyllum
densifolium complex (mire), the Acaena magellanica–
Brachythecium complex (slope drainage lines, springs and
stream banks), the Blechnum penna-marina complex (slope
fernbrake communities), and low-altitude representatives of
the Andreaea–Racomitrium crispulum complex (fellfield). It is
not feasible currently to disentangle this mosaic in the form
of discrete spatial units at conventional mapping scales.
Therefore, in spatial terms (on our map), we summarise veg-
etation units ST 3, 4 and 5 into a single mapping unit called
'Subantarctic Mire-Slope Vegetation' with the hope of being
able to depict them on a more detailed scale at a later stage.
The same applies for vegetation units ST 1 and 2, which
were mapped together as the mapping unit 'Subantarctic
Coastal Vegetation Vegetation'.
Subantarctic Polar Desert crowns the highest tops of the
islands, and on Marion Island descends as low as 100 m in
places. It might seem surprising, even paradoxical, to apply
the term polar desert to a habitat on a mid-latitude island
that receives such a high rainfall, and most vegetation ecolo-
gists have tended to avoid using it when describing vegeta-
tion formations on subantarctic islands. For instance, Huntley
(1971) referred to the vegetation of the habitat as Ditrichum–
Bartramia montane desert and included it with Azorella
selago fellfield in a wind desert complex. On other subant-
arctic islands, vegetation formations corresponding to
Marion Island polar desert have generally been considered as
fellfield (Taylor 1955, Hughes 1987, Bergström 1998) or
included with fellfield as a synusia in a Fellfield Formation
(Lewis Smith 1993). However, fellfield (Fjaeldmark,
Felsenfluren or Felsentundra) was originally applied to vege-
tation in which vascular plants, especially cushion-forming
ones, are a conspicuous (even important) component of the
vegetation. This is certainly not the case for the Marion
Island polar desert habitat, which has strong similarities (a
barren unstable surface caused by frost-sorting, supporting
extremely sparse vegetation dominated by mosses or, more
rarely, lichens) to high arctic polar deserts as defined by
Aleksandrova (1970) and Bliss (1981). In fact, it is floristically
even more polar desert-like than many of the high arctic
sites those authors cite as examples, where vascular plants
make up a much greater component of the vegetation than
is the case for the polar desert habitat on the island. For
example, in the polar desert uplands of Devon, Bathurst and
Cornwallis islands (all > 75° N in the Canadian North Western
Territory), four to six vascular plant species are typical, increas-
ing to eight to 12 at snow flush sites (Bliss 1981). Ten vascu-
lar species occur at 900 m altitude on Ellesmere Island
(Canadian N.W.T., 82° N), whereas none occur above 800 m
on Marion Island. The vegetation of the polar desert habitat
on Marion Island thus resembles the most extreme northern
hemisphere polar deserts; those dominated by mosses and
lichens have been termed ‘polar barrens’ (Benninghof 1974)
or simply ‘barrens’ (Longton 1998), to distinguish them from
fellfield where vascular plants are more important. One only
has to examine the photographs in Ochyra & Bednarek-
Ochyra (1999) and Figure 15.19 to be convinced of the
Figure 15.11 AZm 2 Subantractic Kelp Beds: Durvillea antarctica
in supra-littoral zone on coastal cliffs formed by head of young lava
stream (near the Meteorological Station on Marion Island).
L. Mucina
congruency in equating the Marion Island habitat with polar
barrens of theHigh Arctic.
The term ‘wind desert’ has also been used when referring to the
high-altitude vegetation of the subantarctic islands (Taylor 1955,
Wace 1965, Huntley 1971) but that is not necessarily a fitting
descriptor of the vegetation or the main factors affecting it.
This does not imply that wind is unimportant. The polar desert
surface on Marion Island, like the polar deserts of the Antarctic
and Arctic, is almost totally dominated by physical rather than
biological processes, and frost-heaving, frost-sorting and freeze-
thaw activity are strong and frequent (at high altitudes, freeze-
thaw cycles occur every day of the year; Jan Boelhouwers, pers.
comm.). The effect of wind is as much due to its contribution
to substrate instability and drying as to its more direct effects on
plant growth. In fact, within polar desert there is little correla-
tion between plant cover and shelter from wind, except where
the cause of the shelter also contributes to moisture deposition
and/or retention.
10. Description of Vegetation Units
Marine Macroalgal Vegetation
AZm 2 Subantarctic Kelp Beds
Distribution On the coast and offshore on the leeward (east-
ern) side of Marion Island and Prince Edward Island.
Vegetation & Seascape Features Kelp beds are an impor-
tant and spectacular element of the seascapes of the Prince
Edward Islands. They form two well-separated vegetation belts.
Durvillea antarctica forms a fringing macrophyte belt on the
Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
predominant wind direction on the islands is northwesterly and
southwesterly; on Marion Island ca. 1.5% of the area below an
altitude of 100 m, is occupied by salt-spray vegetation.
Vegetation & Landscape Features Herbfields dominated by
Crassula moschata (sometimes also by Cotula plumosa and at
very exposed localities, Azorella selago co-dominates in a type
of coastal fellfield vegetation).
Geology, Soils & Hydrology Mostly on black lava flows, on
fibrous, black or dark brown peat that is often saline, being
subjected to salt spray and occasional inundation by waves.
Climate The combination of low altitude and coastal location
makes the microclimate of vegetation of salt-spray habitats
extremely oceanic, with small diurnal and seasonal variation.
Important Taxa Herbs: Cotula plumosa (d), Crassula moschata
(d), Azorella selago, Callitriche antarctica, Colobanthus kergue-
lensis, Montia fontana, Ranunculus biternatus. Graminoids:
Agrostis magellanica, Poa cookii. Mosses: Bucklandiella crispula,
Calyptichaete apiculata, Campylopus clavatus, Muelleriella
crassifolia, Orthotheciella varia. Liverworts: Clasmatocolea ver-
micularis (d), Fossombronia australis, Plagiochila heterodonta.
Remarks This vegetation is frequently also influenced by birds,
in particular by Rockhopper Penguins (Eudyptes chrysocome)
and Fur Seals (Arctocephalus tropicalis).
References Huntley (1971), Gremmen (1981), Smith & Steenkamp (2001),
Smith et al. (2001).
ST 2 Subantarctic Biotic Herbfield and
Biotic Complex (Huntley 1971). 4.2. Callitriche antarctica–Poa cookii
Complex (Gremmen 1981). 4. Biotic Grassland Complex & 5. Biotic Herbfield
Complex (Smith & Steenkamp 2001, Smith et al. 2001).
Distribution Found almost exclusively in coastal areas influ-
enced by trampling and manuring by seabirds and seals. Also
inland on slopes supporting colonies of burrowing petrels and
on level areas surrounding Wandering Albatross and Giant Petrel
nests; on Marion Island ca. 7% of the area at an altitude below
Remark 3 This vegetation unit is a sub-
antarctic analogon to the AZm 1 Cape
Kelp Beds (see Chapter 14 Coastal
Vegetation), dominated by Ecklonia
maxima and occurring in cold waters
influenced by Benguela Current.
References De Villiers (1976), Gremmen (1981),
Hay (1986), Attwood et al. (1991), Pakhomov et
al. (2002), Mélice et al. (2003).
Subantarctic Tundra Biome
ST 1 Subantarctic Coastal
Salt Spray Complex (Huntley 1971). 4.1. Crassula
moschata Complex (Gremmen 1981). Coastal
Salt-spray Complex (Smith & Steenkamp 2001,
Smith et al. 2001).
Distribution On the eastern coasts of the
islands this vegetation is mostly restricted
to within 50 m of the shore but on the
northern, western and southwestern
coasts it is found further inland, since the
Figure 15.12 ST 1 Subantarctic Coastal Vegetation: Coastal salt-spray vegetation dominated by
Crassula moschata (yellow-green plant cover in the left foreground) and Cotula plumosa (light
green cover in the right foreground and midground) on Marion Island.
V.R. Smith
coastal rocks and cliffs. Holdfasts are attached to the rock in
the lower littoral zone (De Villiers 1976) and the fronds (up to
8 m long) float freely on the water surface. Macrocystis laevis
is found further off-shore, forming an interrupted belt up to
150 m broad.
Hydrology & Substrate The coastal cliffs and the sea shelf
are built of young Holocene basaltic lavas. The tidal range on
Marion Island is 71 cm at springs and 21 cm at neaps. Swell
heights on the east coast may reach up to 4 m and as they
travel unimpeded in deep water, they extend their energy
against the sea cliffs, especially on the west, to lesser extent on
the south and north coasts. Both kelp species prefer eastern lee
shores (more or less outside the direct influence of roaring west
winds and associated wave exposure), but D. antarctica occurs
all around the island, being seemingly adapted to the relentless
and large wave action that produces surf up to 30 m high in
places. M. laevis forms dense 'kelp forests' of individuals up to
20 m long in water 5 to 20 m deep. It has a uniform distribu-
tion of frond-length frequency, suggesting that only the old-
est fronds are lost by wave action or senescence. Strong wave
action in more shallow water limits the availability of suitable
habitat, whereas in deeper water light limits the kelp’s growth
(De Villers 1976, Attwood et al. 1991). Sea temperature at the
island is around 4°C to 5°C in winter and 6°C to 7°C in sum-
mer (Mélice et al. 2003). Salinity is 32 to 33 psu (Pakhomov et
al. 2002).
Important Taxa (both Macroalgae) Durvillea antarctica (d),
Macrocystis laevis (d).
Remark 1 Mean biomass of Macrocystis laevis was esti-
mated at 0.67 kg C m-2 and net production at 7.7 g C m-2 d-1
and 11.5 g C m-2 d-1 during April and August, respectively
(Attwood et al. 1991). Most of the production of M. laevis
is expected to be exported to the open ocean pelagic system
(Attwood et al. 1991).
Remark 2 Macrocystis laevis is endemic to the Prince Edward
Islands. It was described by Hay (1986); previously it was known
as Macrocystis pyrifera (De Villers 1976, Gremmen 1981), a
species that occurs around many islands in the Subantarctic and
cold temperate regions of the Southern Ocean.
716 Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
100 m, but less than 1% of that between 100 and 500 m alti-
tude is occupied by biotically influenced vegetation. The largest
patches of biotic vegetation are found around large penguin
rookeries (Kildalkey Bay, Bullard Beach and King Penguin Bay).
Vegetation & Landscape Features Coastal platforms with
numerous depressions supporting Cotula plumosa-dominated
herbfields and Callitriche antarctica-dominated organic mud as
well as steep slopes supporting Poa cookii-dominated tussock
grasslands. Saline habitats influenced by seabirds and seals;
overall, species diversity is low.
Geology & Soils The soils are mostly peat with high concen-
trations of organic and inorganic nitrogen and phosphorus and
they may be saline; where trampling is severe the soil surface
may be bare and where both trampling and manuring are
severe (e.g. seal wallows and penguin rookeries) an anaerobic,
almost liquid, organic mud occurs.
Climate The combination of low altitude and coastal location
makes the microclimate of salt-spray vegetation extremely oce-
anic, with small diurnal and seasonal variation. Soils denuded of
vegetation through trampling can reach high surface tempera-
tures (up to 35°C) on sunny days.
Important Taxa (*Aliens) Herbs: Callitriche antarctica (d),
Cotula plumosa (d), Montia fontana (d), Sagina procumbens*
(d), Azorella selago, Ranunculus biternatus. Graminoids:
Agrostis stolonifera* (d), Poa annua* (d), P. cookii (d), Agrostis
magellanica, Juncus scheuchzerioides. Mosses: Brachythecium
rutabulum (d), Leptodontium gemmascens, Schizymenium
campylocarpum. Liverworts: Clasmatocolea vermicularis (d),
Fossombronia australis, Lophocolea randii, Lophozia cylindri-
formis var. australis, Marchantia berteroana.
Remark 1 Twenty-nine bird and three seal species use the islands
to breed and moult. Most of the birds, and the Subantarctic
Fur Seal, occur in large numbers and have a significant impact
on the vegetation, through trampling and manuring. Tussock
grasslands occur around King Penguin
(Aptenodytes patagonicus) and Macaroni
Penguin (Eudyptes chrysolophus) rooker-
ies and Elephant Seal (Mirounga leonina)
wallows and are often pedestalled, the
tops of the 30–100 cm peat pedestals are
occupied by Poa cookii and their bases by
Cotula plumosa or Callitriche antarctica,
sometimes also Montia fontana and Poa
annua. Areas occupied by Rockhopper
Penguins (Eudyptes chrysocome), Gentoo
Penguins (Pygoscelis papua), Crozet
Cormorants (Phalacrocorax atriceps) or
Fur Seals (Arctocephalus tropicalis and
A. gazella) are generally occupied by veg-
etation dominated by C. plumosa and P.
cookii. Inland slopes where burrowing
petrels (10 species) or prions (two spe-
cies) occur are also dominated by P. cookii,
generally with Acaena magellanica and
Brachythecium rutabulum in the under-
storey. In mires, Wandering Albatross
nests are generally surrounded by veg-
etation dominated by P. cookii and also
often containing other nitrophilous plants
(Callitriche antarctica, Montia fontana, C.
plumosa and P. annua) that are not nor-
mal constituents of mire vegetation.
Remark 2 Biotic vegetation, because of
the disturbance through trampling and
probably also the nutrient-enriched soils, is readily invaded
by alien plant species, especially Agrostis stolonifera, Sagina
procumbens, Poa annua and P. pratensis.
Remark 3 Liverworts are infrequent components of vegeta-
tion influenced by seabirds or seals; the notable exceptions are
Marchantia berteroana and Clasmatocolea vermicularis.
References Huntley (1971), Smith (1978c), Gremmen (1981), Smith &
Steenkamp (2001), Smith et al. (2001).
ST 3 Subantarctic Mire
Swamp Complex p.p. (Huntley 1971). 4.4. Juncus scheuchzerioides–
Blepharidophyllum densifolium Complex (Gremmen 1981). 6. Mire Complex
(Smith & Steenkamp 2001, Smith et al. 2001).
Distribution Mire vegetation is found in most lowland areas,
being most extensive below 200 m, but found up to 400 m alti-
tude. On Marion Island approximately 30% of the area below
100 m and approximately 3% of that between 100 and 300
m is occupied by mire vegetation; the largest mires on Marion
Island are found on the coastal plain between Repetto’s Hill
and Long Ridge, inland of East Cape, Macaroni Bay and on the
western coastal plain between Kleinkoppie and Kampkoppie.
Vegetation & Landscape Features Flat or only slightly sloping
localities are dominated by conspicuous graminoids Agrostis
magellanica, Juncus scheuchzerioides, Uncinia compacta and an
abundant bryophyte flora (about 60 species). In wetter areas the
dominating liverworts and mosses include Blepharidophyllum
densifolium, Clasmotocolea humilis, Sanionia uncinata and
Distichophyllum fasciculatum, while in drier mires Jamesoniella
colorata and Racomitrium lanuginosum are prevalent and
Blechnum penna-marina also occurs. In mires where there
is pronounced lateral flow of water Bryum laevigatum and
Breutelia integrifolia (sometimes also Brachythecium subplica-
tum) dominate.
Figure 15.13 ST 2 Subantarctic Biotic Herbfield and Grassland: Vegetation heavily influenced
by penguin and seal manuring and trampling at Cave Bay, Prince Edward Island. Tussocks of Poa
cookii occur on peat pedestals, separated by wallows of a highly organic, eutrophic mud covered
with Callitriche antarctica. Cave Bay was the site of the annexation ceremony in 1948.
V.R. Smith
Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
Geology, Soils & Hydrology Mire vegetation occurs on both
the older grey and younger black lava flows and overlies highly
organic, generally waterlogged, peat which may be up to 2.5 m
deep; since most of the mires receive water from surrounding
areas as well as directly from rainfall, they are not ombrotrophic
in the strict sense. However, most lateral water movement
occurs at the bottom of the profile and the high rainfall leads
to downward leaching and acidic upper horizons; the vegeta-
tion is effectively out of contact with the lower horizons and
not influenced by mineral soil or water that has been in contact
with mineral soil of surrounding areas. The vegetation relies on
nutrients in rainfall (and to a lesser extent on biological nitrogen
fixation) and so the mires can be regarded as ombrotrophic. In
their hydrology and vegetation physiognomy subantarctic mires
closely resemble raised bogs and blanket bogs of the northern
Climate High moisture content of mire peat dampens the
already small diurnal soil temperature fluctuations. Freezing
of the top layer (1–2 cm deep) of peat or bryophyte mat is a
regular occurrence in autumn, winter and spring, even at low
Important Taxa (*Aliens, SSubmerged or floating macro-
phytes in lakes and tarns) Graminoids: Agrostis magellanica
(d), Juncus scheuchzerioides (d), Uncinia compacta (d), Agrostis
stolonifera*S (aquatic form). Herbs: Ranunculus biternatus (d),
Acaena magellanica, Azorella selago, Blechnum penna-marina,
Limosella australisS, Lycopodium magellanicum, Montia fontana,
Potamogeton nodosusS, Sagina apetala*. Mosses: Breutelia
integrifolia (d), Bryum laevigatum (d), Distichophyllum fascicula-
tum (d), Ptychomnion densifolium (d), Racomitrium lanuginosum
(d), Sanionia uncinata (d), Campylopus clavatus, Catagonium
nitens, Plagiothecium ovalifolium. Liverworts: Blepharophyllum
densifolium (d), Jamesoniella colorata (d), Clasmatocolea
humilis (d), Lepidozia laevifolia (d), Fossombronia australis,
Cryptochila grandiflora, Jensenia pisicolor, Leptoscyphus expan-
sus, Lophozia cylindriformis, Metzgeria decipiens, Plagiochila
heterodonta, Riccardia multifida, R. pinguis, Symphyogyna
marionensis. Lichen: Peltigera polydactyla.
Biogeographically Important Taxa (endemics of Kerguélen
Floristic Province) Herb: Ranunculus moseleyi (on Marion
only, but not recorded since 1987). Mosses: Campylopus
austrostramineus, C. subnites. Liverworts: Andrewsianthus
the edges, especially Sanionia uncinata. The wettest mires have
standing water on the surface and are probably more correctly
considered as pools. They are dominated by Agrostis magel-
lanica, Juncus scheuchzerioides, Ranunculus biternatus and
Sanionia uncinata. Small streams are common on the islands,
especially Marion Island. Agrostis stolonifera (submerged form)
is the only vascular plant, and Schistidium falcatum the main
bryophyte, that establish in stream beds.
Remark 2 The mires occur on both islands as an intricate,
small-scale mosaic with fernbrakes and low-altitude fellfield.
Ecotones between mire and fernbrake, and between mire and
fellfield, are a significant component of the slope-mire (and
some fellfield) mosaic.
References Huntley (1971), Gremmen (1981), Smith & Steenkamp (2001),
Smith et al. (2001).
ST 4 Subantarctic Drainage Line and Spring
Swamp Complex p.p. (Huntley 1971). 4.3. Acaena magellanica–
Brachythecium Complex (Gremmen 1981). 3. Slope Complex p.p., incl. 3.5.
Slope Drainage Line and Streambank Habitat & 3.6. Spring and Flush Habitat
(Smith & Steenkamp 2001, Smith et al. 2001).
Distribution This vegetation occurs on both islands up to
400 m altitude (rare above 300 m).
Vegetation & Landscape Features On slopes with impeded
drainage, in sheltered depressions and drainage lines of slopes,
also on the banks of streams—vegetation dominated by the suf-
frutex Acaena magellanica and mosses (mostly Brachythecium
rutabulum, B. subplicatum and Sanionia uncinata).
Geology, Soil & Hydrology Common on younger black lava
and also on the sides of horsts formed by older grey lava, e.g.
southern slopes of Skua Ridge. Soils are wet peat, sometimes
with substantial volcanic ash.
Climate The complete plant cover restricts diurnal soil tem-
perature fluctuations and the soils have never been observed
to freeze.
Important Taxa (*Aliens) Graminoids: Agrostis stolonifera*
(d), Poa cookii (d). Herbs: Acaena magellanica (d), Blechnum
penna-marina, Montia fontana, Pringlea antiscorbutica,
Figure 15.14 ST 3 Subantarctic Mire: Mire dominated by Agrostis magellanica and Breutelia
integrifolia formed in a basin inland of Hendrik Fisher Kop on Marion Island.
V.R. Smith
carinatus, A. lancistipus, A. marionensis,
Diplophyllum marionense, Metzgeria
grollei, M. marionensis.
Remark 1 Many ponds, crater lakes and
tarns occur on the islands. Most are small
(< 400 m2 in area) and, except for the cra-
ter lakes which are found in the craters
of some cinder cones, all are less than 1
m deep. The bottom is generally volcanic
ash mixed with organic matter. Benthic
algal felts occur in some, especially the
smaller ones which are protected from
the scouring effects of wind-driven
wave action. The most common plants
are Ranunculus biternatus, Agrostis
magellanica, A. stolonifera, and Juncus
scheuchzerioides. Rarer, and found only
on the bottom of ponds and tarns, are
Limosella australis and Ranunculus
moseleyi. Bryophytes are uncommon
on the bottoms of ponds and tarns
(Drepanocladus aduncus is sometimes
found there) but very abundant around
718 Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
Ranunculus biternatus. Mosses: Brachythecium rutabulum (d),
B. subplicatum (d), Sanionia uncinata (d), Breutelia integrifo-
lia, Leptodontium gemmascens, Orthotheciella varia, Philonotis
polymorpha. Liverwort: Lophocolea randii.
References Huntley (1971), Gremmen (1981), Smith & Steenkamp (2001),
Smith et al. (2001).
ST 5 Subantarctic Fernbrake Vegetation
Slope Complex (Huntley 1971). 4.5. Blechnum penna-marina Complex
(Gremmen 1981). 3. Slope Complex p.p. (incl. Habitats 3.1 to 3.4) (Smith &
Steenkamp 2001, Smith et al. 2001).
Figure 15.15 ST 4 Subantarctic Drainage Line and Spring Vegetation: Moss-rich (Breutelia
integrifolia, Bryum laevigatum) spring mire with abundant Agrostis magellanica at the Van den
Boogaard River below Tafelberg on Marion Island.
V.R. Smith
best developed and show the greatest
horizon differentiation of all the soils of
the islands, and show some evidence
of podsolisation; a dark brown or black
A-horizon made up mainly of plant litter
overlies a yellow-brown or yellow-grey
loamy clay that sometimes contains red
plinthic layers. The lower horizons of the
soils are often gleyed.
Climate The Blechnum penna-marina
fronds (10–18 cm long) are tightly
packed vertically to form a thick carpet
that protects the soil from insolation and
radiative heat loss at night. Diurnal tem-
perature variation is very small and the
soils have never been observed to freeze.
Important Taxa Graminoids: Agrostis
magellanica, Poa cookii, Uncinia com-
pacta. Herbs: Blechnum penna-marina
(d), Acaena magellanica, Azorella
selago, Grammitis poeppigiana. Mosses:
Brachythecium rutabulum, Campylopus
introflexus, Distichophyllum fascicu-
latum, D. imbricatum, Isopterygiopsis
pulchella, Leptodontium gemmascens,
Plagiothecium ovalifolium, Racomitrium
Figure 15.16 ST 5 Subantarctic Fernbrake Vegetation: Lowland area on the eastern side of
Marion Island showing the mosaic of fernbrake (dark green vegetation on slope) dominated by
Blechnum penna-marina and mires (light green and brown flat areas in between the fernbrakes)
dominated by graminoids Agrostis magellanica, Juncus scheuchzerioides and Uncinia compacta
and liverworts such as Jamesoniella colorata and Blepharidophyllum densifolium.
Distribution Extremely common on
both islands, occurs up to 300 m altitude
but rare above 200 m. On Marion Island
approximately 24% of the area below
100 m, and approximately 10% of that
between 100 and 300 m, is occupied by
fernbrake vegetation.
Vegetation & Landscape Features
Almost all slope communities are domi-
nated by the fern Blechnum penna-
marina. Closed fernbrake, the climax
slope community, consists of an almost
pure carpet of the fern; an open fern-
brake is found in more exposed or rocky
localities, where the fern is co-dominant
with Azorella selago and cushion-forming
mosses; in wetter habitats on less steep
slopes, graminoids (especially Uncinia
compacta, Poa cookii and Agrostis
magellanica) and bryophytes such as
Jamesoniella colorata and Racomitrium
lanuginosum occur, but the vegetation
is still overwhelmingly dominated by B.
Geology & Soils Mainly on black lava
flows, rare on grey lava. Soils are the
lanuginosum, Sanionia uncinata. Liverworts: Acrobolbus ochro-
phyllus, Blepharophyllum densiflorum, Clasmatocolea humilis,
Fossombronia australis, Jamesoniella colorata, Jensenia pisi-
color, Lepidozia laevifolia, Leptoscyphus expansus, Lophocolea
randii, Lophozia cylindriformis, Metzgeria decipiens, Plagiochila
Biogeographically Important Taxa (endemics of Kerguélen
Floristic Province) Herbs: Elaphoglossum randii, Polystichum
marionense. Liverwort: Symphyogyna marionensis.
References Huntley (1971), Gremmen (1981), Smith & Steenkamp (2001),
Smith et al. (2001).
L. Mucina
Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
ST 6 Subantarctic Fellfield
Wind Desert Complex p.p. (incl. Azorella selago fjaeldmark) (Huntley
1971). 4.6. Andreaea–Racomitrium crispulum Complex (Gremmen 1981). 2.
Fellfield Complex (Smith & Steenkamp 2001, Smith et al. 2001).
Distribution On both islands, dominant between 200 and 550
m altitude, although fellfield does occur at lower altitudes at
exposed areas. On Marion Island ca. 33% of the area below
100 m, 72% of that between 100 and 500 m, and 15% of that
above 500 m, is occupied by fellfield. Fellfield occupies around
46% of the total surface area of the island.
Vegetation & Landscape Features Dominated by Azorella
selago, and frequently also by cushion- and ball-forming
mosses. Several species grow on the A. selago cushions. By
far the most common is Agrostis magellanica which may be
co-dominant with A. selago in low-altitude fellfield. Crustose
lichens are common. Total vegetation cover is from 5% to 50%,
often more at lower altitudes.
Geology & Soils Fellfield occurs on both black and grey lava
flows but is more common on the latter. Soils are more mineral
than soils of any of the other vegetation types of the islands.
They are generally skeletal but some are quite deep and con-
sist of a matrix of organic matter (derived from A. selago) and
volcanic ash.
Climate Fellfield occurs on rocky ridges and slopes that are very
exposed to wind—this, with the sparse vegetation cover, means
that the surface soil is subjected to frequent freeze-thaw cycles.
On clear days soil surface temperature can rise to 15°C above
air temperature.
Important Taxa (AOnly in fellfield at altitudes below 250 m):
Herbs: Azorella selago (d), Acaena magellanicaA, Blechnum
penna-marina, Colobanthus kerguelensis, Grammitis poep-
pigiana, Hymenophyllum peltatum, Lycopodium saururus A.
Graminoid: Agrostis magellanica (d). Mosses: Andreaea acu-
minataA (d), A. acutifolia (d), Ditrichum strictum (d), Bartramia
patens, Bucklandiella crispula, Catagonium nitens, Guembelia
kidderi, Hypnum cupressiforme, Kiaeria pumila, Philonotis scabri-
folia, Racomitrium lanuginosum. Liverworts: Herzogobryum ver-
miculare, Jensenia pisicolor, Jungermannia coniflora, Lepidozia
laevifolia, Metzgeria decipiens, Plagiochila heterodonta.
Biogeographically Important Taxa (endemics of Kerguélen
Floristic Province) Moss: Valdonia microcarpa. Herb: Colobanthus
Remark 1 The terms ‘Fjaeldmark’ (Warming 1887), ‘Felsenfluren’
or ‘Felsentundra’ (Schimper & Von Faber 1935) were originally
applied to polar vegetation that is sparse and in which vas-
cular plants, especially cushion-forming ones, are a conspicu-
ous (even important) component of the vegetation. Here, the
English equivalent ‘Fellfield’ (Warming 1909) is used.
Remark 2 Overall, fellfield occupies a smaller proportion of
the total area of the mire-slope-fellfield mosaic than do mire or
slope communities but its importance increases with altitude up
to the upper limit of the mosaic, which is between about 200
and 250 m altitude depending on aspect and exposure. On grey
lava horsts that extend to, or almost to, the coasts (Long Ridge,
Skua Ridge, Stoney Ridge, Kerguélen Rise) these are dominated
by fellfield with patches of very wet mire vegetation. They are
here mapped as fellfield.
References Huntley (1971), Gremmen (1981), Smith & Steenkamp (2001),
Smith et al. (2001), Le Roux & McGeogh (2004), Le Roux et al. (2005).
ST 7 Subantarctic Cinder Cone Vegetation
Distribution On both Marion (approximately 130 cones) and
Prince Edward Islands (14 cones).
Vegetation & Landscape Features Cinder cones are a promi-
nent landscape feature. Most of them have no vegetation, but
where the scoria is relatively stable, some sublithic mosses
and hepatics (protected by the top layer of scoria fragments)
may occur.
Geology Holocene cones of red volcanic ash (scoria) are a
prominent feature on both islands, representing the centres of
Figure 15.17 ST 6 Subantarctic Fellfield Vegetation: Azorella selago (large, green cushions)
fellfield above Tafelberg on Marion Island, with Ditrichum strictum (small olive-green cushions)
and the grass Agrostis magellanica. Peat and volcanic ash deposits are built up under the Azo-
rella cushions.
explosive eruption of the younger black
lavas. The scoria consists of vesicular
fragments from cm to dm in size, mostly
loose but in some places welded into
Climate The macroclimate of cinder
cones is harsh due to wind exposure but
the sublithic microclimate (high humidity
and, on sunny days, warm temperatures)
is favourable for bryophyte growth.
Important Taxa (SSublithic) Mosses:
Catagonium nitens S, Ditrichum conicum,
D. strictum, Hypnum cupressiformeS.
Liverworts: Acrobolbus ochrophyllusS,
Herzogobryum vermiculare S, Lepidozia
laevifoliaS, Metzgeria decipiens S,
Plagiochila heterodontaS. Herbs: Azorella
selago (d), Colobanthus kerguelen-
sis, Hymenophyllum peltatum, Pringlea
Remarks Gremmen (1981) considers
this cryptic vegetation found on some
cinder cones as a part of the fellfield (the
Andreaeo acutifoliae–Racomitrietum
crispuli hypnetosum cupressiformis).
Reference Gremmen (1981).
V.R. Smith
720 Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
Polar Desert Biome
PD 1 Subantarctic Polar Desert
Wind Desert Complex p.p. (incl. Ditrichum–Bartramia montane desert)
(Huntley 1971). 7. Polar Desert Complex (Smith & Steenkamp 2001, Smith
et al. 2001).
Distribution On Marion Island polar desert occupies ca. 10%
of the area below 300 m altitude, 32% of that between 100
and 500 m, and 98% of that above 500 m.
Vegetation & Landscape Features Mosses are the domi-
nating plant group in this vegetation (Andreaea acutifolia, A.
regularis, Bartramia patens, Bucklandiella orthotrichacea, B. val-
don-smithii, Dicranella gremmenii, Ditrichum strictum, D. sub-
australe, Grimmia kidderi, Notoligotrichum australe, Philonotis
scabrifolia. Lichens: Pannaria dichroa, P. hookeri, Placopsis bicolor,
P. cf. cribellans. Herb: Azorella selago.
Biogeographically Important Taxa (endemics of Kerguélen
Floristic Province) Mosses: Bucklandiella valdon-smithii, Dicranella
gremmenii, Ditrichum subaustrale, Valdonia microcarpa.
Remarks In terms of area, fellfield and polar desert are the
dominant habitats on the two islands, perhaps less so in the
case of Prince Edward due to its much lower topography. Truly,
if any vegetation types can be said to be archetypical of Marion
and Prince Edward Islands, it is the two—fellfield and polar
desert—that most closely resemble northern hemisphere tun-
dra vegetation, and not the luxuriant coastal plant communities
which have so far enjoyed the lion’s share of biological investi-
gations at the islands.
References Huntley (1971), Gremmen (1981), Smith & Steenkamp (2001),
Smith et al. (2001).
Figure 15.18 ST 7 Subantarctic Cinder Cone Vegetation: Sparse fellfield (with Azorella selago
and Agrostis magellanica) and oligotrophic lake in the cinder - cone crater of Junior’s Kop on
Marion Island.
Figure 15.19 PD 1 Subantarctic Polar Desert on the high plateau of Marion Island with mostly
black scoria overlying permanent ice. Vegetation is extremely sparse, comprising a few cushion-
forming moss species.
scabrifolia and Valdonia microcarpa). The
moss cover is generally less than 2%; it
may reach up to 50% or more in locali-
ties where snow-melt accumulates but
such sites are very localised (seldom
more than 2 or 3 m2) and they are sur-
rounded by large expanses (sometimes
tens of hectares) of bare rock with only
an occasional cushion-forming moss or
crustose lichen. Azorella selago is the
only vascular plant in the habitat but is
infrequent, rarely attaining more than
2% cover and then only at altitudes
below 650 m. It is absent above 800 m;
lichens are mainly crustose forms and are
much less common than in the fellfield
habitats of lower altitudes (only a few
have been identified).
Geology & Soils Polar desert is found
mainly on the black lava flows since they
are the dominant flows in the higher
regions of the islands. There are no soils
in polar desert; sometimes a thin, grit-like
layer of volcanic ash occurs under and
between the rocks, or the habitat may
be found on large expanses of uncon-
L. Mucina
V.R. Smith
solidated deposits of pebble-sized scoria
with only the occasional rock or boulder
on the surface; the scoria show various
relict and active periglacial features and
polar desert is dominated by physical,
rather than biological, processes.
Climate Polar desert has the most
extreme microclimate on the islands. The
tops of the scoria layer are frozen almost
every morning and the almost total
absence of plants allows the same layer
to reach high temperatures during peri-
ods of extreme sunshine. The free drain-
age and wind-drying (see notes below)
make polar desert the driest habitat on
the islands.
Important Taxa Mosses: Ditrichum stric-
tum (d), Andreaea acutifolia, A. regu-
laris, Bartramia patens, Bucklandiella
or t h otrich a c ea, Gu e m belli a kid-
deri, Hymenoloma antarcticum,
Notoligotrichum australe, Philonotis
Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
11. Credits
The introductory text to the chapter as well as most of the
descriptions of the vegetation units were written by V.R. Smith,
who also produced the first version of the vegetation map. L.
Mucina contributed by shaping the concepts of some vegeta-
tion units to the lists of species and by describing vegetation
units (ST 3, PD 1 and AZm 2) and by introducing some changes
to the original map after a joint excursion with V.R. Smith to
Marion Island in April–May 2004. Most of the photographs
were supplied by V.R. Smith, some by L. Mucina.
The authors of this chapter are obliged to I. Meiklejohn for
kindly providing Figure 15.3. L.W. Powrie and M.C. Rutherford
provided the climate diagram (Figure 15.4). We thank the Chief
Director: Surveys & Mapping for digital and hard-copy maps
of the Marion Island topography. L.W. Powrie used these data
to create the base map onto which the vegetation units were
mapped. L.W. Powrie also digitised the original hard-copy veg-
etation map prepared by V.R. Smith. R. Ochyra provided an
unpublished checklist of mosses for the Prince Edward Islands
featuring the latest nomenclature.
12. References
Adamson, D.A., Whetton, P. & Selkirk, P.M. 1988. An analysis of air tem-
perature records for Macquarie Island: decadal warming, ENSO cooling
and Southern Hemisphere circulation patterns. Pap. Proc. R. Soc. Tasmania
122: 107–112.
Aleksandrova, V.D. 1970. The vegetation of the tundra zones in the
USSR and data about its productivity. In: Fuller, W.H. & Kevan, P.G. (eds),
Productivity and conservation in northern circumpolar lands, pp. 93–114.
IUCN Publ., Morges, CH.
Attwood, C.G., Lucas, M.I., Probyn, T.A., McQuaid, C.D. & Fielding, P.J. 1991.
Production and standing stocks of the kelp Macrocystis laevis Hay at the
Prince Edward Islands, Subantarctic. Polar Biol. 11: 129–133.
Barrat, A. & Mougin, J.L. 1974. Données numériques sur la zoogéographie
de l’avifauna antarctique et subantarctique. Comité National Français des
Recherches Antarctiques 33: 1–18.
Benninghof, W.S. 1974. Macrobiology and ecology in polar deserts. In:
Smiley, T.L. & Zumberge, J.H. (eds), Polar deserts and modern man, pp.
91–97. Univ. Arizona Press, Tucson, AZ.
Bergström, D. 1998. Plant ecology on Heard Island: a new vegetation scheme
and predictions for climate change. A.N.A.R.E. Res. Notes 101: 3, 4.
Bergström, D.M. & Chown, S.L. 1999. Life at the front: history, ecology and
change on Southern Ocean islands. Trends Ecol. Evol. 14: 472–477.
Bester, M.N., Bloomer, J.P., Bartlett, P.A., Muller, D.D., Van Rooyen, M. &
Buchner, H. 2000. Final eradication of feral cats from subantarctic Marion
Island, southern Indian Ocean. S. Afr. J. Wildl. Res. 30: 53–57.
Bliss, L.C. 1981. North American and Scandinavian tundras and polar deserts.
In: Bliss, L.C., Heal, O.W. & Moore, J.J. (eds), Tundra ecosystems: a com-
parative analysis, pp. 8–24. Cambridge Univ. Press, Cambridge.
Bliss, L.C., Heal, O.W. & Moore, J.J. (eds) 1981. Tundra ecosystem: a com-
parative analysis. Cambridge Univ. Press, Cambridge.
Block, W. 1984. Terrestrial microbiology, invertebrates and ecosystems. In:
Lewis, R.M. (ed.), Antarctic ecology, Vol. 1, pp. 163–236. Academic Press,
Bonner, W.N. 1985. Impact of fur seals on the terrestrial environment at
South Georgia. In: Siegfried, W.R., Condy, P.R. & Laws, R.M. (eds), Antarctic
nutrient cycles and food webs, pp. 641–646. Springer, Berlin.
Bonner, W.N. & Lewis Smith, R.I. 1985. Conservation areas in the Antarctic.
Scientific Committee on Antarctic Research, ICSU, Cambridge.
Budd, G.M. 2000. Changes in Heard Island glaciers, King Penguins and Fur
Seals since 1947. Pap. Proc. R. Soc. Tasmania 133: 47–60.
Chapuis, J.L. 1995. Restoration of two islands in the Kerguélen Archipelago
by eradication of the rabbit (Oryctolagus cuniculus). In: Dingwall, P.R. (ed.),
Progress in conservation of the subantarctic islands, pp. 157–160. IUCN
Publ., Cambridge.
Chevallier, L. 1986. Techtonics of Marion and Prince Edward volcanoes (Indian
Ocean): result of regional control and edifice dynamics. Tectonophysics
124: 155–175.
Chown, S.L. 1990. Speciation in the subantarctic weevil genus Dusmoecetes
Jeannel (Coleoptera: Curculionidae). Syst. Entomol. 15: 283–296.
Chown, S.L. & Cooper, J. 1995. The impact of feral house mice at Marion
Island and the desirability of eradication: report on a workshop held at the
University of Pretoria, 1617 February 1995. Dept of Environmental Affairs
and Tourism, Pretoria.
Chown, S.L., Gremmen, N.J.M. & Gaston, K.J. 1998. Ecological biogeogra-
phy of Southern Ocean islands: Species-area relationships, human impacts
and conservation. Am. Nat. 152: 562–575.
Chown, S.L. & Smith, V.R. 1993. Climate change and the short-term impact
of feral house mice at the subantarctic Prince Edward Islands. Oecologia
96: 508–516
Chown, S.L., Van der Merwe, M. & Smith, V.R. 1997. The influence of habi-
tat and altitude on oxygen uptake in subantarctic weevils. Physiol. Zool.
70: 116–124.
Clark, M.R. & Dingwall, P.R. 1985. Conservation of islands in the Southern
Ocean, a review of the protected areas of Insulantarctica. Cambridge Univ.
Press, Cambridge.
Cooper, J., Marais, A.V.N., Bloomer, J.P. & Bester, M.N. 1995. A success
story: breeding of burrowing petrels (Procellariidae) before and after the
extinction of feral cats Felis catus at subantarctic Marion Island. Marine
Ornithology 23: 33–37.
Crafford, J.E. 1990a. Patterns of energy flow in populations of the domi-
nant insect consumers on Marion Island. Ph.D. thesis, Dept. of Entomology,
Univ. of Pretoria.
Crafford, J.E. 1990b. The role of feral house mice in ecosystem functioning
on Marion Island. In: Kerry, K.R. & Hempel, G. (eds), Antarctic ecosystems:
ecological change and conservation, pp. 359–364. Springer, Berlin.
Crafford, J.E. & Chown, S.L. 1990. The introduction and establishment of
the diamondback moth (Plutella xylostella L., Plutellidae) on Marion Island.
In: Kerry, K.R. & Hempel, G. (eds), Antarctic ecosystems: ecological change
and conservation, pp. 354–358. Springer, Berlin.
Crafford, J.E. & Scholtz, C.H. 1987. Quantitative differences between the
insect faunas of subantarctic Marion and Prince Edward Islands: a result of
human intervention? Biol. Conserv. 40: 255–262
Deacon, G.E.R. 1960. The southern cold temperate zone. Proc. R. Soc.
London 152: 441–447.
Department of Environmental Affairs & Tourism 1996. Prince Edward Islands
Management Plan. Dept of Environmental Affairs and Tourism, Pretoria.
De Villiers, A.F. 1976. Littoral ecology of Marion and Prince Edward Islands
(Southern Ocean). S. Afr. J. Antarct. Res. Suppl. 1: 1–40.
Devine, W.T. 1984. Scientific research in New Zealand’s subantarctic nature
reserves. Dept of Lands and Survey, Wellington.
Di Castri, F., Covarrubias, R. & Hajek, E. 1970. Soil ecosystems in subantarctic
regions. Ecol. Conserv. 1: 207–222.
French, D.D. 1974. Classification of IBP Tundra Biome sites based on climate
and soil properties. In: Holding, A.J., Heal, O.W., MacLean, S.F. & Flanagan,
P.W. (eds), Soil organisms and decomposition in tundra, pp. 3–25. Tundra
Biome Steering Committee, Stockholm.
French, D.D. 1981. Multivariate comparisons of IBP Tundra Biome site char-
acteristics. In: Bliss, L.C., Heal, O.W. & Moore, J.J. (eds), Tundra ecosystem:
a comparative analysis, pp. 47–75. Cambridge Univ. Press, Cambridge.
French, D.D. & Smith, V.R. 1985. A comparison between northern and
southern hemisphere tundras and related ecosystems. Polar Biol. 5: 5–21.
Frenot, Y., Gloaguen, J.J. & Tréhen, P. 1997. Climate change in Kerguélen
Islands and colonization of recently deglaciated areas by Poa kerguelensis
and P. annua. In: Battaglia, B., Valencia, J. & Walton, D.W.H. (eds), Antarctic
communities: species, structure and survival, pp. 358–366. Cambridge
Univ. Press, Cambridge.
Gremmen, N.J.M. 1981. The vegetation of the subantarctic islands Marion
and Prince Edward. Dr W. Junk, The Hague.
Gressit, J.L. 1970. Subantarctic entomology and biogeography. Pacific
Insects Monogr. 23: 295–374.
Gribnitz, K.-H., Kent, L.E. & Dixon, R.R. 1986. Volcanic ash, ash soils and
the inferred Quaternary climate of subantarctic Marion Island. S. Afr. J. Sci.
82: 629–635.
Hall, K. 1978. Evidence for Quaternary glaciation of Marion Island (subant-
arctic) and some implications. In: Van Zinderen Bakker, E.M. (eds), Antarctic
glacial history and world palaeoenvironments, pp. 137–147. A.A. Balkema,
Hall, K. 1981. Geology as an aid to climatic-ecological reconstructions: an
example from Marion Island. Lamda 3: 18–22.
Hall, K. 1982. Rapid deglaciation as an initiator of volcanic activity: a hypoth-
esis. Earth Surf. Proc. 7: 45–51.
Hay, C.H. 1986. A new species of Macrocystis C. Ag. (Phaeophyta) from
Marion island, southern Indian Ocean. Phycologia 25: 241–252.
Hays, J.D., Lozano, J.A., Shackleton, N.J. & Irving, G. 1976. Reconstruction of
the Atlantic and Western Indian Ocean sectors of the 18-000 B.P. Antarctic
Ocean. Geol. Soc. Amer. Mem. 145: 337–372.
Heydenrych, R. & Jackson, S. 2000. Environmental impact assessment of
tourism on Marion Island. Dept for Environmental Affairs and Tourism,
Holdgate, M.W. 1964. Terrestrial ecology in the Maritime Antarctic. In:
722 Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
Carrick, R., Holdgate, M.W. & Prévost, J. (eds) Biologie antarctique, pp.
181–194. Herman Press, Paris.
Holdgate, M.W. 1970. Vegetation. In: Holdgate, M.W. (ed.) Antarctic ecol-
ogy. Vol. 2, pp. 729–732. Academic Press, London.
Holdgate, M.W. 1977. Terrestrial ecosystems in the Antarctic. Phil. Trans. R.
Soc. London 279: 5–25.
Hughes, J.M.R. 1987. The distribution and composition of vascular plant
communities on Heard Island. Polar Biol. 7: 153–162.
Huntley, B.J. 1971. Vegetation. In: Van Zinderen Bakker, E.M., Winterbottom,
J.M. & Dyer, R.A. (eds), Marion and Prince Edward Islands, pp. 98–160.
A.A. Balkema, Cape Town.
Huyser, O., Ryan, P.G. & Cooper, J. 2000. Changes in population size, habitat
use and breeding biology of lesser sheathbills at Marion Island: impact of
cats, mice and climate change? Biol. Conserv. 92: 299–310.
Jenny, H. 1941. Factors of soil formation. A system of quantitative pedology.
McGraw-Hill, London.
Kent, L.E. & Gribnitz, K.-H. 1983. Problematical Quaternary successions
on Marion Island: volcanogenic or glacigenic? S. Afr. J. Antarct. Res. 13:
Klok, C.J. & Chown, S.L. 1997. Critical thermal limits, temperature toler-
ance and water balance of a subantarctic caterpillar, Pringleophaga mari-
oni (Lepidoptera: Tineidae). J. Insect Phys. 43: 685–694.
Kloppers, F.J. & Smith, V.R. 1998. First report of Botryotinia fuckeliana on
Kerguélen Cabbage on the sub-Antactic Marion Island. Plant Disease
82: 710.
Kononova, M.M. 1966. Soil organic matter: its nature, its role in soil forma-
tion and in soil fertility. Pergamon, Oxford.
Le Roux, P.C. & McGeogh, M.A. 2004. The use of size as an estimator of age
in the subantarctic cushion plant, Azorella selago (Apiaceae). Arct. Antarct.
Alp. Res 36: 509–517.
Le Roux, P.C., McGeoch, M.A., Nyakatya, M.J. & Chown, S.L. 2005. Effect of
a short-term climate change experiment on a sub-Antarctic keystone plant
species. Glob. Change Biol. 11: 1–12.
Lewis Smith, R.I. 1984. Terrestrial plant biology of the subantarctic and
Antarctic. In: Lewis, R.M. (ed.), Antarctic ecology. Vol. 1, pp. 61, 62.
Academic Press, London.
Lewis Smith, R.I. 1993. Dry coastal ecosystems on subantarctic islands. In:
Van der Maarel, E. (ed.), Ecosystems of the world 2A. Dry coastal ecosys-
tems, polar regions and Europe, pp. 73–93. Elsevier, Amsterdam.
Lewis Smith, R.I. & Walton, D.W.H. 1975. South Georgia, Subantarctic. Ecol.
Bull. 20: 399–423.
Lindeboom, H.J. 1979. Chemical and microbiological aspects of the nitrogen
cycle on Marion Island (subantarctic). Ph.D. thesis, Univ. of Groningen.
Longton, R.E. 1998. Biology of polar bryophytes and lichens. Cambridge
Univ. Press, Cambridge.
McDougall, J. 1971. Geochronology. In: Van Zinderen Bakker, E.M.,
Winterbottom, J.M. & Dyer, R.A. (eds), Marion and Prince Edward Islands,
pp. 72–77. A.A. Balkema, Cape Town.
McDougall, I., Verwoerd, W. & Chevallier, L. 2001. K-Ar geochronology of
Marion Island, Southern Ocean. Geol. Mag. 138: 1–17.
Mendelsohn, J. 1981. Movements of prions Pachyptila spp. and low pres-
sure systems at Marion Island. In: Cooper, J. (ed.), Proceedings of the
Symposium on Birds of the Sea and Shore, 1979, pp. 223–231. African
Seabird Group, Cape Town.
Mélice, J.-L., Lutjeharms, J.R.E., Rouault, M. & Ansorge, I.J. 2003. Sea-sur-
face temperatures at the subantarctic islands Marion and Gough during
the past 50 years. S. Afr. J. Sci. 99: 363–366.
Ochyra, R. & Bednarek-Ochyra, H. 1999. Racomitrium valdon-smithii (Musci,
Grimmiaceae) sp. nov. from Subantarctic Marion Island. Fragm. Flor.
Geobot. 44: 209–217.
Pakhomov, E.A., Kaehler, S. & McQuaid, C.D. 2002. Zooplankton community
structure in the kelp beds of the subantarctic Prince Edward Archipelago:
are they a refuge for larval stages? Polar Biol. 25: 778–788.
Papadakis, J. 1969. Soils of the world. Elsevier, Amsterdam.
Pearsall, W.H. 1950. Mountains and moorlands. Collins, London.
Pickard, J. & Seppelt, R.D. 1984. Phytogeography of Antarctica. J. Biogeogr.
11: 83–102.
Porsild, A.E. 1964. Illustrated flora of the Canadian Arctic Archipelago. Natl.
Mus. Can. Bull. 146.
Rosswall, T. & Heal, O.W. 1975. Structure and function of tundra ecosystem
Ecol. Bull. 20: 1–450.
Rudmose-Brown, R.N. 1928. Antarctic and subantarctic plant life and some
of its problems. Amer. Geogr. Soc. Spec. Publ. 7: 343–352.
Schalke, H.J.W.G. & Van Zinderen Bakker, E.M. 1971. History of the vegeta-
tion. In: Van Zinderen Bakker, E.M., Winterbottom, J.M. & Dyer, R.A. (eds),
Marion and Prince Edward Islands, pp. 89–97. A.A. Balkema, Cape Town.
Schimper, A.F.W. & Von Faber, F.C. 1935. Pflanzen-Geographie auf physiolo-
gischer Grundlage, edn. 3. G. Fischer, Jena.
Scott, L. 1985. Palynological indications of the Quaternary vegetation history
of Marion Island (subantarctic). J. Biogeogr. 12: 413–431.
Scott, L. & Hall, K.J. 1983. Palynological evidence for interglacial vegeta-
tion cover on Marion Island, Subantarctic. Palaeogeogr. Palaeoclimatol.
Palaeoecol. 41: 35–43.
Smith, V.R. 1976a. Standing crop and nutrient status of Marion Island (sub-
antarctic) vegetation. S. Afr. J. Bot. 42: 231–263.
Smith, V.R. 1976b. The effect of burrowing species of Procellariidae on the
nutrient status of inland tussock grasslands on Marion Island. S. Afr. J. Bot.
42: 265–272.
Smith, V.R. 1976c. The nutrient statuses of Marion Island plants and
soils. M.Sc. thesis, Dept. of Botany, Univ. of the Orange Free State,
Smith, V.R. 1977a. The chemical composition of Marion Island soils, plants
and vegetation. S. Afr. J. Antarct. Res. 7: 28–39.
Smith, V.R. 1977b. Vegetation standing crop of the grey lava flows and of
the eastern coastal plain on Marion Island. S. Afr. J. Bot. 43: 105–114.
Smith, V.R. 1978a. Plant ecology of Marion Island—a review. S. Afr. J.
Antarct. Res. 8: 21–30.
Smith, V.R. 1978b. Soil chemistry of Marion Island (subantarctic). S. Afr. J.
Sci. 74: 174, 175.
Smith, V.R. 1978c. Animal-plant-soil nutrient relationships on Marion Island
(Subantarctic). Oecologia 32: 239–253.
Smith, V.R. 1987a. Production and nutrient dynamics of plant communities
on a subantarctic Island. 1. Standing crop and primary production of mire-
grasslands. Polar Biol. 7: 57–75.
Smith, V.R. 1987b. Production and nutrient dynamics of plant communities
on a subantarctic Island. 2. Standing crop and primary production of fjaeld-
mark and fernbrakes. Polar Biol. 7: 125–144.
Smith, V.R. 1987c. Production and nutrient dynamics of plant communities
on a subantarctic Island. 3. Standing stocks, uptake and loss of nutrients in
mire-grasslands. Polar Biol. 8: 135–153.
Smith, V.R. 1988a. Production and nutrient dynamics of plant communities
on a subantarctic Island. 4. Standing stocks, uptake and loss of nutrients in
fjaeldmark and fernbrakes. Polar Biol. 8: 191–211.
Smith, V.R. 1988b. Production and nutrient dynamics of plant communities
on a subantarctic Island. 5. Nutrient budgets and turnover times for mire-
grasslands, fjaeldmark and fernbrakes. Polar Biol. 8: 255–269.
Smith, V.R. 1991. Terrestrial biological research at the Prince Edward Islands.
S. Afr. J. Antarct. Res. 21: 118–123.
Smith, V.R. 1992. Terrestrial slug recorded from subantarctic Marion Island.
J. Mollusc. Stud. 58: 80, 81.
Smith, V.R. 2002. Climate change in the subantarctic: an illustration from
Marion Island. Climatic Change 52: 345–357.
Smith, V.R., Avenant, N.L. & Chown, S.L. 2002. The diet and impact of house
mice on a subantarctic island. Polar Biol. 25: 703–715.
Smith, V.R. & Gremmen, N.J.M. 2001. Turgidosculum complicatulum on
subantarctic Marion Island: carbon acquisition response to climate change.
Polar Biol. 24: 455–459.
Smith, V.R. & Lewis Smith, R.I. 1987. The biota and conservation status of
subantarctic islands. Environ. Int. 13: 95–104.
Smith, V.R. & Steenkamp, M. 1990. Climatic change and its ecological impli-
cations at a subantarctic Island. Oecologia 85: 14–24.
Smith, V.R. & Steenkamp M. 1992a. Soil nitrogen transformations on a sub-
antarctic island. Antarct. Sci. 4: 41–50.
Smith, V.R. & Steenkamp, M. 1992b. Soil macrofauna and nitrogen on a
subantarctic island. Oecologia 92: 201–206.
Smith, V.R. & Steenkamp, M. 2001. Classification of the terrestrial habitats
on subantarctic Marion Island based on vegetation and soil chemistry. J.
Veg. Sci. 12: 181–198.
Smith, V.R., Steenkamp, M. & French, D.D. 1993. Soil decomposition poten-
tial in relation to environmental factors on Marion Island (subantarctic). Soil
Biol. Biochem. 25: 1619–1633.
Smith, V.R., Steenkamp, M. & Gremmen, N.J.M. 2001. Terrestrial habitats on
subantarctic Marion Island: their vegetation, edaphic attributes, distribu-
tion and response to climate change. S. Afr. J. Bot. 67: 641–654.
Syroechkovski, E.E. 1959. The role of animals in primary soil formation in
polar regions of the globe (examples from the Antarctic). Zool. Zh. Ukr. 38:
1770–1775. (In Russian.)
Tatur, A. 1989. Ornithogenic soils of the maritime Antarctic. Pol. Polar Res.
10: 481–532.
Taylor, B.W. 1955. The flora, vegetation and soils of Macquarie Island.
A.N.A.R.E. Rep. Ser. B. 2: 1–192.
Torr, N. 1993. An interim report on Enderby and Rose Island Rabbit eradica-
tion program following expedition 9 February–8 May 1993. Report, New
Zealand Department of Conservation, Wellington.
Ugolini, F.C. 1972. Ornithogenic soils of Antarctica. Antarct. Res. Ser. 20:
Van Aarde, R.J. 1979. Distribution and density of the feral house cat Felis
catus on Marion Island. S. Afr. J. Antarct. Res. 9: 14–19.
Van Aarde, R.J. 1980. The diet and feeding behaviour of feral cats, Felis
catus at Marion Island. S. Afr. J. Wildl. Res. 101: 123–128.
Vegetation of Subantarctic Marion and Prince Edward Islands
ST R E L I T Z I A 19 (200 6)
Van der Merwe, M., Chown, S.L. & Smith, V.R. 1997. Thermal tolerance
limits in six weevil species (Coleoptera, Curculionidae) from subantarctic
Marion Island. Polar Biol. 18: 331–336.
Van Zinderen Bakker, E.M. 1971. Introduction. In: Van Zinderen Bakker, E.M.,
Winterbottom, J.M. & Dyer, R.A. (eds), Marion and Prince Edward Islands,
pp. 1–15. A.A. Balkema, Cape Town.
Van Zinderen Bakker, E.M. 1973. The glaciation(s) of Marion Island (suban-
tarctic), In: Van Zinderen Bakker, E.M. (ed.), Palaeoecology of Africa, the
surrounding islands and Antarctica, Vol. 8, pp. 163–178. A.A. Balkema,
Cape Town.
Verwoerd, W.J. 1971. Geology. In: Van Zinderen Bakker, E.M., Winterbottom,
J.M. & Dyer, R.A. (eds), Marion and Prince Edward Islands, pp. 40–62. A.A.
Balkema, Cape Town.
Verwoerd, W. & Chevallier, L. 1987. Contrasting types of surtseyan tuff
cones on Marion and Prince Edward Islands, southwest Indian Ocean. Bull.
Volcanol. 49: 399–417.
Verwoerd, W.J., Russell, S. & Berruti, A. 1981. 1981 volcanic eruption
reported on Marion Island. Earth Planet. Sci. Lett. 54: 153–156.
Wace, N.M. 1960. The botany of the southern oceanic islands. Proc. R. Soc.
London 152: 475–490.
Wace, N.M. 1965. Vascular plants. In: Van Miegham, J. & Van Oye, P. (eds),
Biogeography and ecology in Antarctica, pp. 201–266. Dr W. Junk, The
Walton, D.W.H. 1985. The subantarctic islands. In: Bonner, W.N. & Walton,
D.W.H. (eds), Key environments: Antarctica, pp. 293–317. Pergamon Press,
Warming, E. 1887. On the vegetation of Greenland. Meddelelser om
Grønland 12. Kjobenhavn. (In Danish.)
Warming, E. 1909. Oecology of plants. Clarendon Press, Oxford.
... 2000 mm during the same period (le Roux & McGeoch, 2008b). The vegetation is closely coupled with abiotic conditions and consists of 23 native vascular plant species and ranges from near continuous short-statured plant cover in sub-Antarctic Tundra to barren Polar Desert (Smith & Mucina, 2006). The island is remote, has a relatively recent origin-only emerging above sea level for the first time c. ...
... Five vegetation units have been mapped previously based on field research, photographs and field observations, and informed by expert opinion (Smith & Mucina, 2006; Figure 1). Smith and Mucina (2006) recognized that mapping at the scale of plant community identified in previous studies (Gremmen, 1981;, in vegetation that changes within a few meters, would not be possible, and thus mapped five units ( Figure 1) at a broader scale. ...
... Five vegetation units have been mapped previously based on field research, photographs and field observations, and informed by expert opinion (Smith & Mucina, 2006; Figure 1). Smith and Mucina (2006) recognized that mapping at the scale of plant community identified in previous studies (Gremmen, 1981;, in vegetation that changes within a few meters, would not be possible, and thus mapped five units ( Figure 1) at a broader scale. Polar Desert was indicated by the absence of vascular plant species and by the presence of bryophytes . ...
Full-text available
The updating and rethinking of vegetation classifications is important for ecosystem monitoring in a rapidly changing world, where the distribution of vegetation is changing. The general assumption that discrete and persistent plant communities exist that can be monitored efficiently, is rarely tested before undertaking a classification. Marion Island (MI) is comprised of species‐poor vegetation undergoing rapid environmental change. It presents a unique opportunity to test the ability to discretely classify species‐poor vegetation with recently developed objective classification techniques and relate it to previous classifications. We classified vascular species data of 476 plots sampled across MI, using Ward hierarchical clustering, divisive analysis clustering, non‐hierarchical kmeans and partitioning around medoids. Internal cluster validation was performed using silhouette widths, Dunn index, connectivity of clusters and gap statistic. Indicator species analyses were also conducted on the best performing clustering methods. We evaluated the outputs against previously classified units. Ward clustering performed the best, with the highest average silhouette width and Dunn index, as well as the lowest connectivity. The number of clusters differed amongst the clustering methods, but most validation measures, including for Ward clustering, indicated that two and three clusters are the best fit for the data. However, all classification methods produced weakly separated, highly connected clusters with low compactness and low fidelity and specificity to clusters. There was no particularly robust and effective classification outcome that could group plots into previously suggested vegetation units based on species composition alone. The relatively recent age (c. 450,000 years B.P.), glaciation history (last glacial maximum 34,500 years B.P.) and isolation of the sub‐Antarctic islands may have hindered the development of strong vascular plant species assemblages with discrete boundaries. Discrete classification at the community‐level using species composition may not be suitable in such species‐poor environments. Species‐level, rather than community‐level, monitoring may thus be more appropriate in species‐poor environments, aligning with continuum theory rather than community theory. We reexamined the classification of vegetation on sub‐Antarctic Marion Island and found that there were no strong floristic grounds for a discrete classification. A community‐level classification using species composition may not be suitable in such species‐poor environments. Species‐level, rather than community‐level, monitoring may be more appropriate in species‐poor environments.
... Plant communities on the PEIs fall into one of two biomes: Polar Desert or sub-Antarctic Tundra. These are grouped at a broad scale into six habitat complexes or eight vegetation types (Smith and Mucina 2006), and at a finer scale have been grouped into 24 unique habitat types . The harsh climate, together with biotic factors such as trampling and manuring, result in a complex mosaic of plant communities ( Figure 6) that develop along gradients of nutrients, wind exposure, temperature and moisture . ...
... then classified the plant communities on Marion Island using ordination and clustering methods based on the vegetation and soil chemistry, with six habitat complexes that now included the Polar Desert. After this, Smith and Mucina (2006) classified and described the vegetation at a broader spatial scale into eight vegetation types (Table 2) according to their distribution, important taxa, geology, soils, climate and landscape patterns based on the previous classifications. Lastly, Gremmen and Smith (2008) used a combination of the above classifications and descriptions to create a hybridized classification to describe the vegetation units, which they termed "habitats", as ecological and soil conditions were also included together with floristics (Table 2). ...
... Lastly, Gremmen and Smith (2008) used a combination of the above classifications and descriptions to create a hybridized classification to describe the vegetation units, which they termed "habitats", as ecological and soil conditions were also included together with floristics (Table 2). These habitats were also grouped into habitat complexes, which are equivalent or similar to the broad scale vegetation types classified by Smith and Mucina (2006). "Vegetation type (map)" (Figure 7) and "Vegetation type" were classified according to Smith and Mucina, (2006), whereas "Habitat complex" and "Habitat type" were classified according to Gremmen and Smith (2008). ...
Technical Report
Full-text available
The National Biodiversity Assessment (NBA) 2018 is a collaborative effort to synthesise the best available science on South Africa’s biodiversity. For the first time, South Africa’s southernmost territory, the sub-Antarctic Prince Edward Islands (PEIs), have been included in the NBA. A map of 34 ecosystem types was developed, supporting a greater understanding of biodiversity patterns and providing a foundation for a systematic assessment of all marine and terrestrial ecosystem types. The Threat Status of ecosystems and some species is discussed and Ecosystem Protection Levels reported for the first time.
... None of the vascular species are endemic. The vegetation changes dramatically with distance from the coast (Smith & Mucina, 2006) where salt-spray and grassland are associated with The island is of recent volcanic origin (initiated c. 450,000 before present; McDougall et al., 2001). Older previously glaciated grey lava is fine-grained, compact and has few vesicles with relatively uniform topography, whereas the highly dissected and irregular vesicular black lava has not been glaciated (Verwoerd, 1971). ...
... Another important and changeable driver is nutrient input to the ecosystem. Allochthonous faunal influences (seals and seabirds) and marine inputs are important in the low-elevation 'biotic sites' (Smith & Mucina, 2006) and explains the strong relationship between plant cover and elevation. This is particularly true on the grey lava transect with greater access to the sea on a gentle slope along the glacial moraine, whereas the transect through black lava terminated above 10-20 m cliffs. ...