Partridge 92
Biological Conservation 1992, 61, 59-71
Vegetation recovery following sand mining on
coastal dunes at Kaitorete Spit, Canterbury,
New Zealand
T. R. Partridge
Botany Institute, DSIR Land Resources, Private Bag, Christchurch, New Zealand
(Received l l April 1991; revised version received 21 A u g u s t 1991; accepted 6 September 1991)
A section of the extensive sand dunes at Kaitorete Spit, Canterbury, New
Zealand, has been mined for sand over a period of 40 years. Unmined dunes are
dominated by dense stands of the otherwise now restricted indigenous sand
binder Desmoschoenus spiralis, making them an area of great conservation
value. Plant communities on mined surfaces of various age and on unmined
dunes were examined by using classification and ordination. Classification
clearly distinguished communities of unmined and mined dunes respectively.
The principal ordination gradients represent the typical landward dune sequence
and the mined/unmined differences. Although there are sites on unmined dunes
that carry vegetation of the mined group, there is no evidence that mined sites
have recovered communities typical of the unmined dunes. The conclusion is
that there is no sign of recovery of the original dune communities despite partial
colonisation by Desmoschoenus. Two explanations are offered. Adventive Am-
mophila arenaria has invaded the older mined dunes, displacing Desmoschoenus
or excluding it from re-invading, while the remaining mined area has developed
a sparse sand-plain vegetation, the result of lateral sand movement. Implications
for conservation management are discussed.
INTRODUCTION sand-binding plants. Even if full recovery takes a
long time, it might be expected that trends to-
Opencast mining typically causes such massive wards recovery would be detectable.
damage to landscapes and biological features One sand-dune system that has been affected by
(Down & Stocks, 1977) that considerable restora- mining is that at Kaitorete Spit, Canterbury, New
tion work needs to be undertaken to establish veg- Zealand. A section of these dunes has been mined
etation which is generally unrelated to that origi- since 1952 for its coarse-grained, angular, well-
nally present (Johnson & Bradshaw, 1979). In sorted sand. The dunes are important for conser-
certain situations, especially where substrate re- vation values, being one of the few remaining sys-
mains, there is a reasonable expectation that the tems in New Zealand where the indigenous
original plant communities will at least show some sand-binding sedge, pingao Desmoschoenus spi-
natural recovery despite the damage. One such ralis, still dominates. This species has declined
situ- ation is the mining of sand from coastal dune considerably, initially through destruction of the
systems (Brooks, 1976; Lewis, 1976). Certainly, indigenous dune cover by burning and grazing
mining severely damages the original dune struc- that accompanied the arrival of Europeans in
ture, but the processes that build dunes are likely New Zealand 150 years ago. Most dune systems
to remain: a supply of sand, wind to move it, and were subsequently stabilised by extensive planting
of the introduced sand binder marram Ammophila
Biological Conservation 0006-3207/92/$05.00 © 1992 Elsevier arenaria, a species that has further displaced much
Science Publishers Ltd, England. Printed in Great Britain of the remaining Desmoschoenus where the two
59
60 T. R. Partridge
come into competition. Kaitorete Spit therefore aided by a known history of mining operations, at
rates as one of New Zealand's most botanically least in relative terms, as mining has taken place
valuable dune systems. As well as the high pro- in a systematic west to east direction over a period
portion of communities dominated by indigenous of 40 years. Even where areas have been remined
species, there are endemic invertebrates and plants this is known.
(Carmichaelia appressa, an unnamed species of As- Nomenclature follows Webb et al. (1988) and
perula, and possibly others), otherwise localised Connor and Edgar (1987) and references therein.
plants (e.g. Austrofestuca littoralis), and important
archaeological sites; the dune system is also pecu-
liar in being of a dry coarse sand instead of the STUDY SITE
more usual fine sand. It is desirable therefore that
the mined section, although relatively small, re- Kaitorete Spit is a sand/gravel barrier complex
covers as much as possible its valuable botanical that separates brackish Lake Ellesmere from the
features. Some attempts have been made to assist Pacific Ocean (Fig. 1). The complex is only about
recovery, but have either failed completely (re- 5,000-7,000 years old, and the product of rapid
planting of Desmoschoenus spiralis cuttings) or are growth, as it is now 28 km long and up to 3.2 km
of uncertain success (spreading of Desmoschoenus wide at the eastern end (Kirk, 1969; Armon,
spiralis seed following mining), because mining 1974). Continuous sand dunes occur on the sea-
and conservation interests dispute the effectiveness ward side at an average width of 220 m. An older,
of the methods employed. Such claims are a fea- inactive, discontinuous series of sand dunes occurs
ture of emotionally charged conservation debates. a further 100 m inland, but is not included in the
Certainly there is no difference between areas study. The eastern half of the spit's dunes has a
where restoration has been claimed to have been stable to slowly accreting seaward margin, while
carried out and where it has not. the western half is eroding under wave attack, and
Little has been published on the ecology of New a number of parabolic dune blowouts have oc-
Zealand sand dunes and their species. Early stud- curred. The area examined in this study is on the
ies are particularly rare (Cockayne, 1911, 1928; eastern half and has a stable dune margin. Inland
Pegg, 1914; Logan & Holloway, 1934), although of the dunes is dry grassland dominated by intro-
Esler (1969, 1970, 1974, 1975) produced detailed duced species and especially the grass Stipa no-
descriptions of some North Island dune systems. dosa. The whole dune system is subjected to fre-
Recent interest has been greater, especially in the quent winds from both the south (moist, cool,
South Island (Johnson, 1982; Simpson & Mason, especially in winter) and northwest (dry, warm,
1984; Smith et al., 1985; Sykes & Wilson, 1987, mostly in summer). Mean annual rainfall at Lin-
1988, 1989), many studies being detailed examina- coln University, some 20 km to the north, is 590
tions of plant/environment interactions. Although mm. This, coupled with the coarse base material,
there have been a number of conservation reports makes the area very drought-prone.
(e.g. Palmer, 1980) and theses (e.g. Peace, 1975; Sand-mining has taken place in a confined area
Holland, 1981; Courtney, 1983) on botanical as- of the central section of the dunes and to the east
pects of the Kaitorete Spit dunes, little has been of Kaitorete Scientific Reserve, beginning on a
published in the scientific literature, the exceptions small scale in 1952. Records were started in 1964
being the brief description of Burrows (1969) and with the granting of the first licence, and at that
inclusion of this area in a larger study of vegeta- time some 6,000 m 3 were removed per annum. Re-
tion/environment relations in dunes by Sykes and moval of sand peaked in 1974 when 33,000 m 3
Wilson (1988). The effects of mining have not were extracted (Palmer, 1980). Restrictions on
been specifically addressed. amounts and areas available considerably reduced
The aims of this study were twofold: to deter- the extraction of sand in the 1980s. The total
mine exactly how sand mining has affected the length of dunes mined by 1990 was approximately
plant communities and to know whether there has 1,300 m. Sand has been mostly removed from the
been complete or partial recovery of the original central section, but at the peak extended from
communities after mining. If recovery has taken close to the front dune almost to the grass flats.
place, the processes by which this occurs are of The effects of sand-mining on dune structure are
great importance, and if not, the reasons for non- demonstrated by six dune profiles of the areas
recovery become significant. The study has been sampled (Fig. 2). Of the two unmined sections, A
Vegetation recovery on coastal dunes 61
South Island
New Zealand
l'~aitor~et e / Paeific~
Spit / Ocean
500 Ill
' 1D
Fig.1. Location ofthe KaitoreteSpitdunesshowingthepositionofthesixsampleareas.
map
has a single, and F a double ridge system. Section divided into 20 × 20 m squares, making a 10 × 11
18 was mined in the 1950s-1960s, but not to a great sampling shape. Within each square a 5 x 5 m
depth, while in the most recently mined section E quadrat was randomly placed and species rooted
(late 1980s) mining has been restricted to only a frequency measured in the 25 1 x 1 m sub-
narrow part of the dune sequence. The most ex- quadrats at this site.
lensively mined dunes are in the central sections C Sites and species were classified by Cluster Anal-
and D, where almost the entire dune sequence was ysis. The association measure was Canberra Metric
removed during the 1970s, often including the and the sorting strategy Flexible with beta set at
whole front dune, and to below mean sea level. -0.25 to create groups of comparatively even size
(Clifford & Stephenson, 1975). Gradients in the
data were examined by Detrended Correspondence
METHODS Analysis (Hill & Gauch, 1980). It was necessary to
exclude two species - - Carex pumila and Austro-
Each of the six sections (A to F) covers a 200 x festuca littoralis - - because of their excessive
220 m area with the longer axis perpendicular to scores on the first ordination axis. This resulted in
the shore. The wide dunes at section B necessi- the removal of one site. Analyses were performed
tated increasing this to 240 m. Each section was using the P A T N software package (Belbin, 1989).
62 T. R. Partridge
A D
8 7 7 5 3 3 1513 19 15 11 910121012131317132017
7 5 7 6 4 3 315111 1 91010121313131517 - 2
8 5 5 7 7 7 2 3 2 2 2 9101012121212 - 18217
"o 8 6 7 7 8 7 3 1 2 2 9109101212119 1719
8 8 6 7 7 1 3 4 4 2 1 1 o 910101012121717171717
.= 8 8 7 7 7 5 5 4 1616 16 91010101012131717 - 19
8 7 7 7 7 5 3 3 2 16 11 910111213131317171917
8 8 8 8 6 7 3 4 2 2 1 0 1 0 1 0 11 1 2 1 2 1 2 1 3 1 7 1 8 1 7
6 7 7 7 7 8 5 5141 4 912121313121311719
76 7 7 8 5 1 1 2 4 2 9101215131313 - - 19
10m
5m
0
B E
8 5 5 7 71412121820419 8 5 6 5 1111 5 3 2
8 8 8 7 7 1612132019 2 1 9 5 5 18 - 1 9 1 4 14 2 11
8 8 7 16 16 1 2 1 2 1 5 2 0 2 0 - 20 5 515 5 - 15111414 417
8 8 7161612172018202020 9 3 5 5 5 ~'"/15 1 217
8 7 7 51611 - 2 0 1 5 1 7 - 16 9 5 5 5 5 111115 2 317
7 5 7 71616161715 - 2016 9 5 5 5 11 5171 15 2 1
8 7 5 8 7 15161417 - - 9 9 5 5 5 5 2 1 4 1 4 2
9 5 7 71616171616 3 4 9 5 6 5 5 5 515 2 1 15
7 7 7 7 716201515181917 9 510 5141 517 2 1 17
7 7 7 7 16 1 6 1 5 - 19 - 15 9 5 10 4 5 1-414 -~- 2 2 15
C F
101012 . . . . 1 3 2 15 9 5 5 5 5 5 4 1 3 4 1 8
1010121212121215 2 415 9 5 5 6 5 5 5 5 4 4 2
9 9101212171913 - 2 20 9 5 5 5 3 5 4 315316
9 91012121 1514 - - 19 9 7 5 4 5 5 3 3 4 2
10101012 713151714 - 2 9 9 5 31411131 3 318
910101--21215151513 - 19i 8 75 6 41414244 2
91010111215141517 - 2 5 6 4 54 1 1 3152 2
:9 6 6 6 4 - 1 1 1 5 1 3 - 17 9 6 5 3 4 4 1 4 152
8 7 6 6 514151515 19 9 6 6 5 3 3 4 1 322
8 8 515141"415151520 20 9 65 5 5 341515215
Fig. 2. Community maps of the six sample areas, A - F . The numbers refer to the 20 communities described (1-10 = mostly
unmined, 11-20 = mostly mined). Underlining indicates sites that have been mined; dash indicates mined with no vegetation;
blank is unmined with no vegetation. Dune profiles are included for the central part of each area, with the marked points
indicating the extent of mining.
V e g e t a t i o n r e c o v e r y on c o a s t a l d u n e s 63
RESULTS Communities characteristic of unmined dunes
(U1-UIO)
UI : Moderately vegetated Desmoschoenus spiralis
Community descriptions (25 sites). Dominated by Desmoschoenus spiralis,
but at a lower abundance than any of the other
The twenty communities discriminated by Cluster unmined communities, and tending to occur
Analysis are described below. The term frequency mostly on open areas on the steep back face of the
is used as a within-community characteristic, foredune. Characteristic species: Desmoschoenus
while sample site frequency is termed abundance. spiralis, Lagurus ovatus, Hypochoeris radicata and
Interpretation is aided by Table 1, which sum- Calystegia soldanella.
marises frequency and mean site abundance of the U2: Dense Desmoschoenus spiralis on front dune
major species within the communities, and Fig, 2, (41 sites). The typical unmined front dune com-
which consists of the community maps of the six munity. Characteristic species: Desmoschoenus spi-
sections. Table 2 gives the number of unmined ralis, Lagurus ovatus, Hypochoeris radicata and
and mined sites within each community. Calystegia soldanella.
Table 1. Frequency and abundance within the 20 communities
Community
Species 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
N o . o f sites 25 41 30 31 78 20 55 31 36 35 18 40 24 21 44 24 34 8 16 16
Acaenaagnipila + 9-1 4-1 8-1 7-2 9-4 F-5 5-2 + 3-1
Trifoliumarvense + + 8-3 9-7 6-2 3-1 + + + 5-1
Zoysia pungens 2-+ L 9-6 F-8 3-1 L + L
Rytidosperma clavatum + 5-2 9-5 +
Stipa nodosa + + + 7-3 9-6 +
Aira caryophyllea + + L 5-1 + +
Silene gallica + + + 2-1 3-+ 4-1
Trifolium glomeratum 3-1 5-1 +
Bromus diandrus L L 7-1 6-2 9-5 9-5 7-2 4-2 8-4 L 8-3 + + + +
Rumexacetosella L 8-2 8-2 9-4 9-3 F-7 F-8 9-8 9-4 6-3 3-1 9-2 +
Desmoschoenus F-5 F~ F~8 F-8 F~i 9-4 9-4 + 2-1 + 3-1 8-2 F-2 F-3 9-3 8-3 9-3 F-1 F-4 +
spiralis
Hypochoerisradicata F-2 F-5 F-3 9-5 9-2 7-2 9-5 F~ 9-3 9-3 4-1 F-2 9-1 F-4
Lagurusovatus F-4 F~6 F-8 F-8 F~ F-8 F-9 F-9 9-8 9-6 7-3 7-3 + 9-2 5-1 F-5 4-1 +
Calystegiasoldanella 7-2 9-5 9-4 9-4 5-1 9-4 3-1 + + 6-2 + + 3-1 9-3 7-1
Ammophilaarenaria + L + L 3-1 5-1 L 3-1 5-2 F-7 5-1 6-2 F--6
Raouliaaustralis 4-1 L 1-+ + 5-1 9-6 6-2 + + 6-2 F-2 7-3 5-2 + +
Erodium cicutarium L + + + L + 9-4 6-2 + + L +
Carmichaelia appressa 1-1 5-2 L L L L
Muehlenbeckia L L L F- 7 L + + L L
complexa
Melicytus alpinus L L +
Holcus lanatus + + + L L + + +
Poa cita 5-1 + 9-2 5-1 L + +
Pteridium esculentum L L L L + 4-3 F-8 L L
Carex breviculmis L
Pseudognaphalium + 3-+ + + + + 3-+ + + +
luteoalbum
Anagallis arvensis + L
Foeniculum vulgate L
Lolium perenne + L + + +
Cirsium arvense L L +
Carex pumila L L
Elymus rectisetus L L
O n l y the 31 m o s t a b u n d a n t species are included. F o r each species the first n u m b e r is f r e q u e n c y within the c o m m u n i t y ( F = 100%,
9 -- 90-99%, 8 = 80-89°/,,, etc.) a n d the second is m e a n site a b u n d a n c e (expressed similarly as frequency in the 25 s u b q u a d r a t s ) . L indicates
locally a b u n d a n t , a n d + present b u t rare.
64 T. R. Partridge
Table 2. Number of unmined and mined sites within each species: Desmoschoenus spiralis, Rumex acetosella,
community
Hypochoeris radicata, Lagurus ovatus, Calystegia
Community Unmined Mined soldanella, Zoysia pungens and Trifolium arvense.
This is one o f the richest communities.
1 21 4 U8." Zoysia pungens sand flats (31 sites). Similar
2 40 1
3 30 -- to community U7, but on flatter surfaces at the
4 31 -- westernmost two sections. Characteristic species:
5 72 6 Lagurus ovatus, Hypochoeris radicata, Rumex ace-
6 20 --
7 55 -- tosella. Zoysia pungens, Raoulia australis, Tri-
8 31 -- folium arvense, Stipa nodosa and Acaena agnipila.
9 36 -- The small sedge Carex breviculmis is locally abun-
10 35 -- dant in a few sites and the sub-shrub Pimelea
1-10 371 11 urvilleana is occasionally present.
11 7 11 U9: Grass flats (36 sites). Behind the dunes are
12 I 39
13 1 23 extensive grass fiats that extend beyond the sam-
14 13 8 piing area. Characteristic species: Stipa nodosa,
15 12 32 Rytidosperma clavatum, Lagurus ovatus, Poa cita,
16 13 11
17 8 26 Acaena agnipila, Trifolium arvense, Trifolium glom-
18 2 6 eratum, Raoulia australis, Hypochoeris radicata.
19 5 11 This is the richest of the communities.
20 5 11 UIO: Pteridium esculentum (35 sites). Dense
11-20 67 178 stands of Pteridium esculentum occur on a discrete
Bare 5 38 area o f the back dunes o f the two central sections,
the otherwise typical dune species all being less
U3: Dense Desmoschoenus spiralis in central part common. Characteristic species: Pteridium esculen-
of dunes (30 sites). Occurs only in the zone be- tum, Lagurus ovatus, Hypochoeris radicata, Rumex
tween the front dune and back dune, or on the acetosella and Bromus diandrus.
single dune crest where present. Characteristic
species: as for community U2 plus Rumex ace- Communities characteristic of mined dunes
tosella and Bromus diandrus. (M11-M20)
U4: Dense Desmoschoenus spiralis with Acaena M l l : Weedy sites (18 sites). Sparsely vegetated
agnipila (31 sites). As for c o m m u n i t y U3, plus and poorly defined with low species faithfulness
Acaena agnipila, although this species is never and abundance, but a large complement of
abundant. species. Scattered (a) on recently mined sites with
U5." Dense Desmoschoenus spiralis on back dunes little sand remaining, (b) as a weedy strand, (c) as
(78 sites). Characteristic c o m m u n i t y o f the gently isolated sites on roads and parking areas. Charac-
sloping faces o f back dunes, except where these teristic species: Rumex acetosella, Lagurus ovatus
are of low stature, mined, or covered in Pteridium and Calystegia soldanella.
esculentum. Characteristic species: Desmoschoenus M12. Recently mined sites with sand (40 sites).
spiralis, Lagurus ovatus, Hypochoeris radicata, Recently mined areas o f low sand activity nearer
Rumex acetosella and Bromus diandrus. the back of the mined dunes and rather poorly
U6: Muehlenbeckia complexa on back face of back vegetated except by Erodium cicutarium. Charac-
dune (20 sites). A distinct c o m m u n i t y o f low- teristic species: Erodium cicutarium, Bromus dian-
growing woody species mixed amongst community drus and, with low abundance, Desmoschoenus spi-
U5. Characteristic species: shrubs - - Muehlen- ralis, Lagurus ovatus and Raoulia australis.
beckia complexa, Carmichaelia appressa; herbs - - M13: Sparsely vegetated pavement (24 sites). Pave-
Desmoschoenus spiralis, Lagurus ovatus, Bromus ment occurs where wind deflation leaves a stony
diandrus, Rumex acetosella, Hypochoeris radicata, surface, this process favouring the cushion species
Acaena agnipila and Poa cita. Raoulia australis. Characteristic species: Raoulia
U7: Desmoschoenus spiralis/Zoysia pungens defla- australis, Erodium cicutarium and Desmoschoenus
tion dunes (55 sites). A c o m m u n i t y of low hum- spiralis (low cover).
mocky deflation dunes at the westernmost two M14: Sparsely vegetated pavement~dune with
sections where back dune is absent. Characteristic Rumex acetosella (21 sites). With a different com-
Vegetation recovery on coastal dunes 65
position to community M13, this occurs more often Summary of communities
on unmined areas. Characteristic species: Desmo-
schoenus spiralis, Hypochoeris radicata, Lagurus Communities U1 to U10 form the natural un-
ovatus, Rumex acetosella and Raoulia australis. mined vegetation, although the part of the dune
Notable associated rarer species include Austrofes- pavement community M14 is also unmined.
tuca littoralis, Asperula sp. and Scleranthus biflorus. Desmoschoenus spiralis dominates the dunes from
M15: New low dunes on mined areas of intermedi- the front of the front dune to the back of the back
ate age (44 sites). New, low, hummocky dunes dune. On the front face of the front dune it typi-
have formed in the seaward half of the mined sec- cally forms community U2 with U3 in the central
tions where sand is most active. Characteristic section. The back dune, where present, has U5
species: Desmoschoenus spiralis, Hypochoeris radi- mixed in with the shrub-dominated U6. Where the
cata, Raoulia australis, Lagurus ovatus and Am- back dune is absent the shrubby vegetation occurs
mophila arenaria (all low cover). A small number on the back face of the single dune and there are
of sites form a narrow unmined strand dominated extensive hummocky deflation dunes and sand
by Carex pumila. fiats (U7, U8) further inland. Beyond the dunes
M16: New dunes on older mined areas (24 sites). are grass fiats (U9). The back dunes dominated by
Mostly in areas where dense Ammophila arenaria Pteridium esculentum (U10) are an important vari-
is displacing Desmoschoenus spiralis on the oldest ant.
mined dunes. Also includes unmined front dune The mined dunes of communities M11 to M20
where Ammophila arenaria has spread up the front have a sparser vegetation. Desmoschoenus spiralis
face from its establishment point at the strand. is often present, varying from tiny isolated clumps
Characteristic species: Ammophila arenaria (dense), (M18) to relatively dense (M19), but never as
Desmoschoenus spiralis, Lagurus ovatus, Hypocho- dense as in the unmined communities. Ammophila
eris radieata and Calystegia soldanella. arenaria is important, especially in M16 and M20,
M17: Areas mined on to foredune front (34 sites). where it usually dominates as a dense stand. Asso-
Occurs where mining has both directly and indi- ciated species in these communities are variable
rectly (through over-steepening) destroyed the and unpredictable, and there are a number of
front dune, resulting in new active dunes with a weedy species that seldom occur on unmined sites,
sparse cover of Desmoschoenus spiralis. Character- the most notable being Erodium cicutarium. Where
istic species: Desmoschoenus spiralis, Calystegia wind deflation occurs a Raoulia australis stone
soldanella and Ammophila arenaria (all low cover). pavement (M 14, M 15) is established.
M18: Extremely sparse Desmoschoenus spiralis (8 The separation of the mined and unmined
sites). Almost bare, actively moving sand near the groups is very clear and highly significant (X2 =
front dune on mined areas carries Desmoschoenus 451, p < 0.0001). The greatest number of excep-
spiralis as isolated small plants, not yet building tions are of sites in the typically mined group (M)
dunes. Ammophila arenaria sometimes occurs simi- that occur on unmined dunes. Many of these are
larly. Characteristic species: Desmoschoenus spi- on the strand (M11), natural pavement (M14), ac-
ralis and Ammophila arenaria (low cover). tive front faces of the front dunes (M 17), or areas
M19." Pure Desmoschoenus spiralis on foredune invaded by marram (M16). Mined sites that carry
(16 sites). Mostly occurring on mined front dunes, vegetation of typically unmined (U) communities
but also in active blowouts, with a moderate cover are less common. Bare sites are clearly mostly re-
of Desmoschoenus spiralis, but with no associated lated to mining.
species. Characteristic species: Desmoschoenus
spiralis. Distribution of the communities
M20: Dense Ammophila arenaria (16 sites). Virtu-
ally all sites occur at the oldest mined area. There At the unmined Section A (Fig. 2), 90% of the
are few associated species that are never abun- sites carry vegetation classified as part of the un-
dant. Characteristic species: Ammophila arenaria. mined group of communities (U1-U10). Here
Bare sand (43 sites). Extensive bare areas are pre- there is only a single, but wide, dune ridge in the
sent where sand movement is too great to allow central part of the dune system. All the Desmo-
establishment, and on over-steepened dune faces, schoenus spiralis-dominated communities are well-
especially the eroding rear face of the foredune. represented (U1-U5), with some Muehlenbeckia
They are generally the result of sand-mining. complexa (U6) on the backslope. Behind this are
66 T. R. Partridge
the deflation dunes and sand flats characterised by M12, this being characteristic of recently mined
Zoysia pungens (U7, U8) and not the grass fiats of sites where sand is accumulating, and community
the central and eastern sections. Sites carrying M 13, which is pavement where sand is eroding. Be-
vegetation classified as part of the mined group hind, the unmined area is dominated by Pteridium
(M) include some weedy strand areas (Mll), a esculentum (U10), with grass flats (U9) beyond.
large patch of dense Ammophila arenaria on the Mining at Section E is restricted to a narrow
front dune face (M16), and a few sites in the east part between the foredune and backdune. The
that have been destabilised by adjacent sand min- foredune is made up of many communities includ-
ing (M 15). ing active foredunes otherwise characteristic of
At Area B, sand-mining has removed most of mined areas (M17), and both moderate (U1) and
the front and central sections of the dunes. Of the dense (U2) Desmoschoenus spiralis. There are two
Desmoschoenus spiralis-dominated unmined com- areas of community M14, sparsely vegetated
munities, only the common U5 is at all well-repre- dune/pavement, both on unmined dunes. In the
sented. Most notable of the mined communities west a small area has been left unmined to protect
are M16 and M20, both characterised by dense a population of the rare Asperula sp., while an
Ammophila arenaria, this also occupying unmined area in the east has Austrofestuca littoralis. Much
foredune remnants and some of the unmined of the mined area is bare or has weeds ( M l l ) or
backdunes as well. Behind the foredune remnants establishing dunes (M15). A number of mined
are extensive bare areas where sand is extremely sites have the otherwise typically unmined com-
mobile. The grouping of sites of the typically munity U5 - - these have been invaded by many
recently mined community M 12 at the eastern end Desmoschoenus spiralis seedlings and associated
represents an area that was remined in the 1980s. species common in unmined areas on sand of low
Beyond the mined area, the deflation dune and activity. More typical community U5 occurs be-
sand flat communities (U7, U8) typical of Section hind the mined area, although in the west there is
A are again well-represented. some Pteridium esculentum (U10). Grass fiats be-
The mining at Section C has been far more exten- hind are mostly of community U9.
sive. Sites representing typically unmined commu- At Area E, unmined communities dominate, ex-
nities are present only on foredune remnants (U2), cept near the foredune front where some active
small unmined 'islands' such as that in the east areas have community M 15. Unlike Area A, there
(U2-U4), and backdunes (U5, U6). Pteridium es- are two distinct dune ridges and no sand flats.
culentum (U10) dominates most of the backdunes Dense Desmoschoenus spiralis occurs throughout,
and spreads onto the grass fiats (U9) beyond, al- with communities U2-U5 all being well-repre-
though the western end is beyond the range of this sented variants. In the west the backslope of the
species and has sand flats (U8). The mined area is backdune has abundant woody vegetation, espe-
dominated by sparse vegetation, especially of cially of Muehlenbeckia complexa (U6), while be-
community M15, which is characteristic of mined yond are grass fiats (U9).
areas of intermediate age, most sites of this com-
munity occurring here, Many of these sites con- Ordination
tain Ammophila arenaria. Also well-represented is
community M12, which dominates a large, deep The first two axes of the DCA site ordination are
hollow in the east, this being an area that was presented using separate diagrams for each of two
remined in the 1980s. On some of the older mined sets of site attributes. Separate diagrams were used
areas, erosion of sand has resulted in Raoulia aus- to avoid clutter, as there were many sampling
tralis-dominated dune pavement (M 16). sites. Figure 3 divides the sites into mined versus
Foredune damage has been a feature of the unmined. There is, in general, a good separation
mining at Section D. This has resulted from over- of the sites both on axis 1 and especially axis 2.
steepening of the backslope which has been fol- The many unmined sites on the lower left are
lowed by blowouts, to leave large active areas be- densely clumped whereas the mined sites on the
tween small, isolated foredune turrets. The upper right are more scattered, those with the
distinctive community M17, with its presence of highest values on axis 2 being vegetation domi-
Calystegia soldanella, now occupies most of the nated by non-dune weeds and Erodium cicutarium
former foredune area along with many bare areas (Fig. 4). There is, however, considerable overlap
of moving sand. The central part has community of mined and unmined sites at the lower right.
Vegetation recovery on coastal dunes 67
350 Unmined 350 Mined
300 300
• a
- : , ,
250 250 ... s
¢q • o, ,,.
200 o = 200
• . ' . : . • ".'%, .
<
. . . . -: =~.~..~. ~.:
150
,......: . :~.':-~. ,.;; . 150
" ••
• :,, • .. , .%$...~'v. ..t~.~.. .~
• ...: ".~. ¢.:.,. .:'¢,.'... ,
100
• :. ~ ; 100
•. • • ,.. , ,°
• qk~l., . • •
21 -:., t. .
• . " . • s ..
50 • e" • " " . •° 50
0 0
50 100 150 200 250 '300 350 400 50 100 150 200 250 300 350 400
Axis | Axis 1
(a) (b)
Fig. 3. Site ordination divided into (a) unmined and (b) mined sites.
These are dominated by Ammophila arenaria and 1, and from bottom to top on axis 2, with increas-
occur on the unmined foredune where this species ing distance inland. The trend is more pronounced,
has displaced Desmoschoenus spiralis, and on the however, on axis 1. Unmined sites show the trend
oldest mined areas. very clearly while it is also present, but less well-
Figure 5 separates the sites by distance from the defined, in the mined areas. Mined sites, as shown
sea, each diagram representing a 40 m step. The in Fig. 1, are mostly in the central sections, leaving
trend is from coastal (right) to inland (left) on axis the extremes the mostly clearly defined groupings.
eLolium eAnagalis
400' eErodium
Ciraiume
300 Po8
• "'e ° l c u s
n Bromus
Pteridium • Raoulia
Elymua
¢-4
T.glomeratum •MeliCytus
~< 200 Rumex
< • Muehlenbeckia
Carmichaelia e •
eCarex ekagurus
Acaena
100 Stipa eH y p ° c h ° e r | $ Desmoschoenus
• eSilene •
Zoysia
~ytidosperma •
• Ammophila e
T.arvense
Aira
eCalystegla
-loo 0 loo- 200 300 4o0
Axis 1
Fig. 4. Species ordination showing only the most important species (Table 1). Foeniculum vulgare and Elymus rectisetus are not
shown as they had large values on axis 2.
68 T. R. Partridge
350 0-40 m 350 120-160 m
3OO 300
250 250
eq
20(1 200 .: ..=."
• "" "*;" "k
.. : .':: :-...
150 150 -.
. ..! ." ~. :
° ••
°.° .~ :!-
100 •~ .'.'... 100
*l*gt . ".- • •
• ~'*• 0 0
50 J l 0
50
0
0 50 100 150 200 250' 300 350 400 o 5'0 xJo 3;o
Axis 1
350 40-80 m 350 160-200 m
°
300 300
250 250 " • |
200 200 ~° ••
t
° , ~
I • " •° :
150 150
° •
• •l~a • . , •
..° . ...... ..:.~
100 •* "~ - .o . °"
•d ¢ ° • • "°
100 °••° ° • •
° •° °
• ° •• . •
50 "-. 50
0 i i i
0 50 150 200 250 300 350 50 100 150 200 250 300 350 400
350 80-120 m 350 200-240 m
300 300
250 °~ 250
•• • .*
° °
200 t° •
°
200 •°
• :-.. •
• •• ~°°°, ° °~ ••
150 ° ,.° •
150 .;. • •
°-~ ° •"
":.." : : - • •D
..:~'." °, °
100 • .. • • 100
0 m 0
50 50
0 0
50 100 150 200 250 300 350 400 Jr) 100 150 200 250 300 350 400
F i g . 5. Site ordination divided according to distance from the shoreline. Each diagram r e p r e s e n t s a s t e p o f 40 m.
The ordination clearly shows two important and mined vegetation. Axis 2 might also be inter-
trends within the vegetation. The typical sea-to- preted as representing a stability gradient, this
land gradient is most obvious on axis l, while axis being very much the product of mining, but with a
2 more represents the difference between unmined coast to inland component as well.
Vegetation recovery on coastal dunes 69
DISCUSSION systems in other parts of New Zealand, such as
the Manawatu (Esler, 1969).
Sand-mining has caused major changes to both Ammophila arenaria is clearly having an impor-
the physical structure and the plant communities tant impact upon the dunes. On unmined areas
of the Kaitorete dunes. The central parts of the this grass establishes probably mostly from rhi-
dunes have been hollowed out to varying depths, zome fragments along the strand and spreads
and those areas mined in the 1960s and 1970s slowly up the front of the foredune by vegetative
have suffered the greatest, with sand having been growth (Gemmell et al., 1953). Its establishment in
removed from virtually the entire dune system. this zone demonstrates a distinct preference for
Furthermore, the effects of mining have caused in- areas of maximum sand activity. It is this prefer-
stability on some adjacent areas by over-steepen- ence for active sites that has caused it to become
ing of faces, and through mining sand from below so dominant in parts of the mined dunes. In such
mean sea level. areas, mining has cut into the stands of Am-
The results clearly demonstrate that the major mophila that had previously colonised the front
effect of mining on the plant communities is the face of the foredune, thus supplying the active
establishment of a suite of communities that are areas behind with an abundance of stem frag-
less common on unmined dunes, including those ments. In such areas initial establishment follow-
that are characterised by ephemeral weeds, estab- ing mining involves a large number of
lishment areas of Desmoschoenus spiralis, recently Desmoschoenus spiralis seedlings and a smaller
developed Raoulia australis pavement and stands number of Ammophila clumps from stem frag-
of Ammophila arenaria. Where these communities ments (community M12, for instance). However,
do occur on unmined dunes, they mostly tend to the ratio changes rapidly as (1) the Ammophila
be on disturbed sites of active dunes including the grows and spreads at a much faster rate, (2) the
strand at the base of the foredune. The major ex- Desmoschoenus is heavily grazed by rabbits and
ception is the stable unmined dune pavement that hares, and (3) Desmoschoenus disappears when the
clusters as one of the mined communities. two species come into competition, resulting even-
Mining has occurred over a period of some 40 tually in community M20. The result is a dense
years. If the plant communities are recovering, Ammophila stand with only little Desmoschoenus
then it is expected that there would be mined sites remaining. From the back of the mined area, Am-
that carry communities otherwise characteristic of mophila slowly spreads vegetatively on to the
unmined areas, especially on the older mined dunes behind as well, in the same way that it in-
dunes. The results clearly demonstrate that this vades unmined foredunes.
has not taken place. There is indeed no indication The more recently mined areas have not been
that mined areas are developing vegetation similar invaded by Ammophila arenaria. Here the initial
to that on unmined dunes. Such failure was noted establishment of Desmoschoenus spiralis proceeds
by Woodhouse (1982) on the east coast of North in two possible directions. Where sand accumula-
America. Although Desmoschoenus spiralis has tion occurs low dune hummocks of Desmo-
colonised mined areas it has formed novel, sparse schoenus establish, but where erosion takes place,
communities. The failure of the dunes to recover a stone pavement dominated by Raoulia australis
their original communities can be attributed to two develops. There is no real sign, however, of a ten-
factors. The first is the invasion of much of the dency for Desmoschoenus to build higher dunes,
older mined area by Ammophila arenaria. This probably because the sand supply is limited. There
:species has effectively halted any potential recov- is little prospect for such sand becoming available.
ery by the development of a dense sward that Mining has introduced a number of other weeds
competitively excludes Desmoschoenus spiralis. to the area. These are presently uncommon, but
The second is the development of a continuous some, including Lupinus arboreus and Cytisus
sand plain that runs parallel to the beach front, scoparius, have the ability to spread and become
through the mined area. Within this sand plain problems. All these species occur close to vehicle
lateral winds move sand through the mined dunes tracks and their presence can be attributed to
to form a sparsely vegetated hummocky low dune introduction by vehicles.
system dominated by Desmoschoenus spiralis The back dunes of the central mined area have
along with some erosion pavement. Such sand- a dense covering of Pteridium esculentum fern in-
plain vegetation is characteristic of some dune stead of the Desmoschoenus spiralis and Muehlen-
70 T. R. Partridge
beckia complexa typical of other areas. The ACKNOWLEDGEMENTS
success of Pteridium cannot be attributed to min-
ing, but may be the result of fires that have spread The author wishes to thank Richard Cross for
onto the dunes from the grassland. Indeed, to the help with field work and Brain Molloy, Peter
east, Pteridium covers the entire dune system in an Johnson, H a b i b a Gitay, Colin M e u r k and Bastow
area that was previously a military firing range Wilson for comments. The study was funded by
and has a history of burning. Desmoschoenus spi- the N e w Zealand Department o f Conservation.
ralis seems unable to compete with Pteridium, but
it is not known whether the fern is increasing its
range. REFERENCES
The non-recovery of plant communities and the
Armon, J. W. (1974). Late quaternary shorelines near Lake
invasion by A m m o p h i l a arenaria have important Ellesmere, Canterbury, New Zealand. N . Z . J . Geol., 17,
implications for both mining and the management 63-73.
of the dune botanical resources. It is clear that Belbin, L. (1989) PATN Technical Reference. CSIRO Divi-
sion of Wildlife & Ecology, Canberra.
sand mining creates conditions in which commu- Bradshaw, A. D. & Chadwick, M. J. (1980). The Restoration
nity recovery is unlikely. Restorative planting and of Land. The Ecology and Reclamation of Derelict and De-
replacement of a layer o f sand seem to be having graded Land. Blackwell Scientific Publications, Oxford.
no effect. These plantings of pot-grown material Brooks, D. R. (1976). Rehabilitation following mineral sand
mining on North Stradbrooke Island, Queensland. In
have been more successful than the cuttings used Landscaping and Land Use Planning as Related to Mining
earlier, but there is so far no evidence that sur- Operations. Australian Institute of Mineralogy and Metal-
vival is better than in areas left unplanted. Al- lurgy, Adelaide, pp. 93-104.
Burrows, C. J. (1969). Flora and vegetation of Kaitorete
though some of the communities that establish are (Ellesmere) Spit. Cant. Bot. Soc. J., 2, 5-10.
dominated by native plants, they are not the same Clifford, H. T. & Stephenson, W. (1975). An Introduction to
as those removed during mining. The mining, Numerical Classification. Academic Press, New York.
Cockayne, L. (1911). Report on the Dune-areas of New
however, is due to cease soon, although much of Zealand, Their Geology, Botany and Reclamation. Appendix
the central part o f Section F is planned for sand to the Journal of the NZ House of Representatives, C13.
extraction. This pattern is in contrast to the situ- Government Printer Wellington, pp. !-76.
ation in Queensland, Australia (Brooks, 1976; Cockayne, L. (1982). The Vegetation of New Zealand, 2nd
Edn. Englemann Press, Leipzig.
Bradshaw & Chadwick, 1980), where restorative Connor, H. E. & Edgar, E. (1987). Name changes in the in-
planting on mined dunes has been successful. digenous New Zealand Flora 1960-1986 and Nomina
However, there most of the sand has been re- Nova IV, 1983-1986. N.Z.J. Bot., 25, 115-70.
Courtney, S. P. (1983). Aspects of the ecology of
turned to the system following removal of rare Desmoschoenus spiralis (A. Rich) Hook f. MSc thesis, Uni-
minerals, unlike the situation at Kaitorete where versity of Canterbury.
virtually all the sand is removed and virtually Down, C. G. & Stocks, J. (1977). Environmental Impact of
Mining. Applied Science Publishers, London.
none returned. Esler, A. E. (1969). Manawatu sand plain vegetation. Proc.
If left unchecked, it seems that A m m o p h i l a are- N.Z. Ecol. Soc., 16, 32-5.
naria will spread over the entire mined area and Esler, A. E. (1970). Manawatu sand dune vegetation. Proc.
N.Z. Ecol. Soc., 17, 41-6.
onto adjacent dunes as well. In the absence of Esler, A. E. (1974). Vegetation of the sand country bordering
mining, this species is restricted to along the the Waitakere Range, Auckland: the southern beaches.
strand, and has only a slow, largely manageable Proc. N.Z. Ecol. Soc., 21, 72-7.
spread up the foredune. In only a small number of Esler, A. E. (1975). Vegetation of the sand country bordering
the Waitakere Range, Auckland: Piha Beach. Proc. N.Z.
locations has it reached a situation where its man- Ecol. Soc., 22, 52-6.
agement requires extensive and expensive treat- Gemmell, A. R., Greig-Smith, P. & Gimingham, C. H.
ment. If allowed to spread throughout the mined (1953). A note on the behaviour of Ammophila arenaria
(L.) Link in relation to sand-dune formation. Trans. Bot.
area, it will become a vast management problem. Soc. Edinb., 36, 132-6.
The oldest mined dunes have already reached that Hill, M. O. & Gauch, H. G. (1980). Detrended correspon-
stage, but the remaining areas where A m m o p h i l a dence analysis: an improved ordination technique. Vegeta-
tio, 42, 47-58.
arenaria is either sparse or absent have the poten-
Holland, L. D. (1981). Plants and sand dune development;
tial to be saved from the invasion o f this species. Ammophila arenaria versus Desmoschoenus spiralis on
To preserve the valuable botanical features of the Kaitorete Barrier, Canterbury. MSc thesis, University of
Kaitorete Spit dunes, it is important that sand- Canterbury.
Johnson, M. S. & Bradshaw, A. D. (1979) Ecological princi-
mining should cease, and A m m o p h i l a arenaria ples for the restoration of disturbed and degraded land.
spread be controlled. Appl. Biol., 4, 141-200.
Vegetation recovery on coastal dunes 71
Johnson, P. N. (1982). Naturalised plants in south-west Smith, S. M., Allen, R. B. & Daly, B. K. (1985). Soil-vegeta-
South Island, New Zealand. N.Z.J. Bot., 20, 131~,2. tion relationships on a sequence of sand dunes, Tautuku
Kirk, R. M. (1969). Beach erosion and coastal development Beach, South-east Otago, New Zealand. J. R. Soc. N.Z.,
in the Canterbury Bight. N.Z. Geogr., 25, 23-35. 15, 295-312.
Lewis, J W. (1976). Regeneration of coastal ecosystems Sykes, M. T. & Wilson, J. B. (1987). The vegetation of a New
after mineral sand mining. Australian Mining, July 1976, Zealand dune slack. Vegetatio, 71, 13-19.
22-9. Sykes, M T. & Wilson, J. B. (1988). An experimental investi-
Logan, M. C. & Holloway, J. E. (1934). Plant succession on gation into the response of some New Zealand sand dune
the Oreti River sand dunes. Trans. R. Soc. N.Z., 64, species to salt spray. Ann. Bot., 62, 159-66.
122-39. Sykes, M. T. & Wilson, J. B. (1989). The effect of salinity on
Palmer, J. D. (1980). The Kaitorete Dunes. Department of the growth of some New Zealand sand dune species. Acta
Lands and Survey, Christchurch. Bot. Neerl., 38, 173-82.
Peace, M. (1975). The plant ecology of the dune systems of Webb, C. J., Sykes, W. R. & Garnock-Jones, P. J. (1988).
Kaitorete Spit. MSc thesis, University of Canterbury. Flora of New Zealand, Vol. IV. Botany Division, DSIR,
Pegg, E. J. (1914). An ecological study of some New Zealand Christchurch.
sand-dune plants. Trans. Proc. N.Z. Inst., 46, 150-77. Woodhouse, W. W. Jr (1982). Coastal sand dunes of the US.
Simpson, M. J. A. & Mason, R. (1984). The flora of the Can- In Creation and Restoration of Coastal Plant Communities,
terbury dune system (New Brighton-Waipara) in the 1980s ed. R. R. Lewis III. CRC Press, Botan Roca, Florida, pp.
Mauri Ora, 11, 99-111. 1-44.
Vegetation recovery following sand mining on
coastal dunes at Kaitorete Spit, Canterbury,
New Zealand
T. R. Partridge
Botany Institute, DSIR Land Resources, Private Bag, Christchurch, New Zealand
(Received l l April 1991; revised version received 21 A u g u s t 1991; accepted 6 September 1991)
A section of the extensive sand dunes at Kaitorete Spit, Canterbury, New
Zealand, has been mined for sand over a period of 40 years. Unmined dunes are
dominated by dense stands of the otherwise now restricted indigenous sand
binder Desmoschoenus spiralis, making them an area of great conservation
value. Plant communities on mined surfaces of various age and on unmined
dunes were examined by using classification and ordination. Classification
clearly distinguished communities of unmined and mined dunes respectively.
The principal ordination gradients represent the typical landward dune sequence
and the mined/unmined differences. Although there are sites on unmined dunes
that carry vegetation of the mined group, there is no evidence that mined sites
have recovered communities typical of the unmined dunes. The conclusion is
that there is no sign of recovery of the original dune communities despite partial
colonisation by Desmoschoenus. Two explanations are offered. Adventive Am-
mophila arenaria has invaded the older mined dunes, displacing Desmoschoenus
or excluding it from re-invading, while the remaining mined area has developed
a sparse sand-plain vegetation, the result of lateral sand movement. Implications
for conservation management are discussed.
INTRODUCTION sand-binding plants. Even if full recovery takes a
long time, it might be expected that trends to-
Opencast mining typically causes such massive wards recovery would be detectable.
damage to landscapes and biological features One sand-dune system that has been affected by
(Down & Stocks, 1977) that considerable restora- mining is that at Kaitorete Spit, Canterbury, New
tion work needs to be undertaken to establish veg- Zealand. A section of these dunes has been mined
etation which is generally unrelated to that origi- since 1952 for its coarse-grained, angular, well-
nally present (Johnson & Bradshaw, 1979). In sorted sand. The dunes are important for conser-
certain situations, especially where substrate re- vation values, being one of the few remaining sys-
mains, there is a reasonable expectation that the tems in New Zealand where the indigenous
original plant communities will at least show some sand-binding sedge, pingao Desmoschoenus spi-
natural recovery despite the damage. One such ralis, still dominates. This species has declined
situ- ation is the mining of sand from coastal dune considerably, initially through destruction of the
systems (Brooks, 1976; Lewis, 1976). Certainly, indigenous dune cover by burning and grazing
mining severely damages the original dune struc- that accompanied the arrival of Europeans in
ture, but the processes that build dunes are likely New Zealand 150 years ago. Most dune systems
to remain: a supply of sand, wind to move it, and were subsequently stabilised by extensive planting
of the introduced sand binder marram Ammophila
Biological Conservation 0006-3207/92/$05.00 © 1992 Elsevier arenaria, a species that has further displaced much
Science Publishers Ltd, England. Printed in Great Britain of the remaining Desmoschoenus where the two
59
60 T. R. Partridge
come into competition. Kaitorete Spit therefore aided by a known history of mining operations, at
rates as one of New Zealand's most botanically least in relative terms, as mining has taken place
valuable dune systems. As well as the high pro- in a systematic west to east direction over a period
portion of communities dominated by indigenous of 40 years. Even where areas have been remined
species, there are endemic invertebrates and plants this is known.
(Carmichaelia appressa, an unnamed species of As- Nomenclature follows Webb et al. (1988) and
perula, and possibly others), otherwise localised Connor and Edgar (1987) and references therein.
plants (e.g. Austrofestuca littoralis), and important
archaeological sites; the dune system is also pecu-
liar in being of a dry coarse sand instead of the STUDY SITE
more usual fine sand. It is desirable therefore that
the mined section, although relatively small, re- Kaitorete Spit is a sand/gravel barrier complex
covers as much as possible its valuable botanical that separates brackish Lake Ellesmere from the
features. Some attempts have been made to assist Pacific Ocean (Fig. 1). The complex is only about
recovery, but have either failed completely (re- 5,000-7,000 years old, and the product of rapid
planting of Desmoschoenus spiralis cuttings) or are growth, as it is now 28 km long and up to 3.2 km
of uncertain success (spreading of Desmoschoenus wide at the eastern end (Kirk, 1969; Armon,
spiralis seed following mining), because mining 1974). Continuous sand dunes occur on the sea-
and conservation interests dispute the effectiveness ward side at an average width of 220 m. An older,
of the methods employed. Such claims are a fea- inactive, discontinuous series of sand dunes occurs
ture of emotionally charged conservation debates. a further 100 m inland, but is not included in the
Certainly there is no difference between areas study. The eastern half of the spit's dunes has a
where restoration has been claimed to have been stable to slowly accreting seaward margin, while
carried out and where it has not. the western half is eroding under wave attack, and
Little has been published on the ecology of New a number of parabolic dune blowouts have oc-
Zealand sand dunes and their species. Early stud- curred. The area examined in this study is on the
ies are particularly rare (Cockayne, 1911, 1928; eastern half and has a stable dune margin. Inland
Pegg, 1914; Logan & Holloway, 1934), although of the dunes is dry grassland dominated by intro-
Esler (1969, 1970, 1974, 1975) produced detailed duced species and especially the grass Stipa no-
descriptions of some North Island dune systems. dosa. The whole dune system is subjected to fre-
Recent interest has been greater, especially in the quent winds from both the south (moist, cool,
South Island (Johnson, 1982; Simpson & Mason, especially in winter) and northwest (dry, warm,
1984; Smith et al., 1985; Sykes & Wilson, 1987, mostly in summer). Mean annual rainfall at Lin-
1988, 1989), many studies being detailed examina- coln University, some 20 km to the north, is 590
tions of plant/environment interactions. Although mm. This, coupled with the coarse base material,
there have been a number of conservation reports makes the area very drought-prone.
(e.g. Palmer, 1980) and theses (e.g. Peace, 1975; Sand-mining has taken place in a confined area
Holland, 1981; Courtney, 1983) on botanical as- of the central section of the dunes and to the east
pects of the Kaitorete Spit dunes, little has been of Kaitorete Scientific Reserve, beginning on a
published in the scientific literature, the exceptions small scale in 1952. Records were started in 1964
being the brief description of Burrows (1969) and with the granting of the first licence, and at that
inclusion of this area in a larger study of vegeta- time some 6,000 m 3 were removed per annum. Re-
tion/environment relations in dunes by Sykes and moval of sand peaked in 1974 when 33,000 m 3
Wilson (1988). The effects of mining have not were extracted (Palmer, 1980). Restrictions on
been specifically addressed. amounts and areas available considerably reduced
The aims of this study were twofold: to deter- the extraction of sand in the 1980s. The total
mine exactly how sand mining has affected the length of dunes mined by 1990 was approximately
plant communities and to know whether there has 1,300 m. Sand has been mostly removed from the
been complete or partial recovery of the original central section, but at the peak extended from
communities after mining. If recovery has taken close to the front dune almost to the grass flats.
place, the processes by which this occurs are of The effects of sand-mining on dune structure are
great importance, and if not, the reasons for non- demonstrated by six dune profiles of the areas
recovery become significant. The study has been sampled (Fig. 2). Of the two unmined sections, A
Vegetation recovery on coastal dunes 61
South Island
New Zealand
l'~aitor~et e / Paeific~
Spit / Ocean
500 Ill
' 1D
Fig.1. Location ofthe KaitoreteSpitdunesshowingthepositionofthesixsampleareas.
map
has a single, and F a double ridge system. Section divided into 20 × 20 m squares, making a 10 × 11
18 was mined in the 1950s-1960s, but not to a great sampling shape. Within each square a 5 x 5 m
depth, while in the most recently mined section E quadrat was randomly placed and species rooted
(late 1980s) mining has been restricted to only a frequency measured in the 25 1 x 1 m sub-
narrow part of the dune sequence. The most ex- quadrats at this site.
lensively mined dunes are in the central sections C Sites and species were classified by Cluster Anal-
and D, where almost the entire dune sequence was ysis. The association measure was Canberra Metric
removed during the 1970s, often including the and the sorting strategy Flexible with beta set at
whole front dune, and to below mean sea level. -0.25 to create groups of comparatively even size
(Clifford & Stephenson, 1975). Gradients in the
data were examined by Detrended Correspondence
METHODS Analysis (Hill & Gauch, 1980). It was necessary to
exclude two species - - Carex pumila and Austro-
Each of the six sections (A to F) covers a 200 x festuca littoralis - - because of their excessive
220 m area with the longer axis perpendicular to scores on the first ordination axis. This resulted in
the shore. The wide dunes at section B necessi- the removal of one site. Analyses were performed
tated increasing this to 240 m. Each section was using the P A T N software package (Belbin, 1989).
62 T. R. Partridge
A D
8 7 7 5 3 3 1513 19 15 11 910121012131317132017
7 5 7 6 4 3 315111 1 91010121313131517 - 2
8 5 5 7 7 7 2 3 2 2 2 9101012121212 - 18217
"o 8 6 7 7 8 7 3 1 2 2 9109101212119 1719
8 8 6 7 7 1 3 4 4 2 1 1 o 910101012121717171717
.= 8 8 7 7 7 5 5 4 1616 16 91010101012131717 - 19
8 7 7 7 7 5 3 3 2 16 11 910111213131317171917
8 8 8 8 6 7 3 4 2 2 1 0 1 0 1 0 11 1 2 1 2 1 2 1 3 1 7 1 8 1 7
6 7 7 7 7 8 5 5141 4 912121313121311719
76 7 7 8 5 1 1 2 4 2 9101215131313 - - 19
10m
5m
0
B E
8 5 5 7 71412121820419 8 5 6 5 1111 5 3 2
8 8 8 7 7 1612132019 2 1 9 5 5 18 - 1 9 1 4 14 2 11
8 8 7 16 16 1 2 1 2 1 5 2 0 2 0 - 20 5 515 5 - 15111414 417
8 8 7161612172018202020 9 3 5 5 5 ~'"/15 1 217
8 7 7 51611 - 2 0 1 5 1 7 - 16 9 5 5 5 5 111115 2 317
7 5 7 71616161715 - 2016 9 5 5 5 11 5171 15 2 1
8 7 5 8 7 15161417 - - 9 9 5 5 5 5 2 1 4 1 4 2
9 5 7 71616171616 3 4 9 5 6 5 5 5 515 2 1 15
7 7 7 7 716201515181917 9 510 5141 517 2 1 17
7 7 7 7 16 1 6 1 5 - 19 - 15 9 5 10 4 5 1-414 -~- 2 2 15
C F
101012 . . . . 1 3 2 15 9 5 5 5 5 5 4 1 3 4 1 8
1010121212121215 2 415 9 5 5 6 5 5 5 5 4 4 2
9 9101212171913 - 2 20 9 5 5 5 3 5 4 315316
9 91012121 1514 - - 19 9 7 5 4 5 5 3 3 4 2
10101012 713151714 - 2 9 9 5 31411131 3 318
910101--21215151513 - 19i 8 75 6 41414244 2
91010111215141517 - 2 5 6 4 54 1 1 3152 2
:9 6 6 6 4 - 1 1 1 5 1 3 - 17 9 6 5 3 4 4 1 4 152
8 7 6 6 514151515 19 9 6 6 5 3 3 4 1 322
8 8 515141"415151520 20 9 65 5 5 341515215
Fig. 2. Community maps of the six sample areas, A - F . The numbers refer to the 20 communities described (1-10 = mostly
unmined, 11-20 = mostly mined). Underlining indicates sites that have been mined; dash indicates mined with no vegetation;
blank is unmined with no vegetation. Dune profiles are included for the central part of each area, with the marked points
indicating the extent of mining.
V e g e t a t i o n r e c o v e r y on c o a s t a l d u n e s 63
RESULTS Communities characteristic of unmined dunes
(U1-UIO)
UI : Moderately vegetated Desmoschoenus spiralis
Community descriptions (25 sites). Dominated by Desmoschoenus spiralis,
but at a lower abundance than any of the other
The twenty communities discriminated by Cluster unmined communities, and tending to occur
Analysis are described below. The term frequency mostly on open areas on the steep back face of the
is used as a within-community characteristic, foredune. Characteristic species: Desmoschoenus
while sample site frequency is termed abundance. spiralis, Lagurus ovatus, Hypochoeris radicata and
Interpretation is aided by Table 1, which sum- Calystegia soldanella.
marises frequency and mean site abundance of the U2: Dense Desmoschoenus spiralis on front dune
major species within the communities, and Fig, 2, (41 sites). The typical unmined front dune com-
which consists of the community maps of the six munity. Characteristic species: Desmoschoenus spi-
sections. Table 2 gives the number of unmined ralis, Lagurus ovatus, Hypochoeris radicata and
and mined sites within each community. Calystegia soldanella.
Table 1. Frequency and abundance within the 20 communities
Community
Species 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
N o . o f sites 25 41 30 31 78 20 55 31 36 35 18 40 24 21 44 24 34 8 16 16
Acaenaagnipila + 9-1 4-1 8-1 7-2 9-4 F-5 5-2 + 3-1
Trifoliumarvense + + 8-3 9-7 6-2 3-1 + + + 5-1
Zoysia pungens 2-+ L 9-6 F-8 3-1 L + L
Rytidosperma clavatum + 5-2 9-5 +
Stipa nodosa + + + 7-3 9-6 +
Aira caryophyllea + + L 5-1 + +
Silene gallica + + + 2-1 3-+ 4-1
Trifolium glomeratum 3-1 5-1 +
Bromus diandrus L L 7-1 6-2 9-5 9-5 7-2 4-2 8-4 L 8-3 + + + +
Rumexacetosella L 8-2 8-2 9-4 9-3 F-7 F-8 9-8 9-4 6-3 3-1 9-2 +
Desmoschoenus F-5 F~ F~8 F-8 F~i 9-4 9-4 + 2-1 + 3-1 8-2 F-2 F-3 9-3 8-3 9-3 F-1 F-4 +
spiralis
Hypochoerisradicata F-2 F-5 F-3 9-5 9-2 7-2 9-5 F~ 9-3 9-3 4-1 F-2 9-1 F-4
Lagurusovatus F-4 F~6 F-8 F-8 F~ F-8 F-9 F-9 9-8 9-6 7-3 7-3 + 9-2 5-1 F-5 4-1 +
Calystegiasoldanella 7-2 9-5 9-4 9-4 5-1 9-4 3-1 + + 6-2 + + 3-1 9-3 7-1
Ammophilaarenaria + L + L 3-1 5-1 L 3-1 5-2 F-7 5-1 6-2 F--6
Raouliaaustralis 4-1 L 1-+ + 5-1 9-6 6-2 + + 6-2 F-2 7-3 5-2 + +
Erodium cicutarium L + + + L + 9-4 6-2 + + L +
Carmichaelia appressa 1-1 5-2 L L L L
Muehlenbeckia L L L F- 7 L + + L L
complexa
Melicytus alpinus L L +
Holcus lanatus + + + L L + + +
Poa cita 5-1 + 9-2 5-1 L + +
Pteridium esculentum L L L L + 4-3 F-8 L L
Carex breviculmis L
Pseudognaphalium + 3-+ + + + + 3-+ + + +
luteoalbum
Anagallis arvensis + L
Foeniculum vulgate L
Lolium perenne + L + + +
Cirsium arvense L L +
Carex pumila L L
Elymus rectisetus L L
O n l y the 31 m o s t a b u n d a n t species are included. F o r each species the first n u m b e r is f r e q u e n c y within the c o m m u n i t y ( F = 100%,
9 -- 90-99%, 8 = 80-89°/,,, etc.) a n d the second is m e a n site a b u n d a n c e (expressed similarly as frequency in the 25 s u b q u a d r a t s ) . L indicates
locally a b u n d a n t , a n d + present b u t rare.
64 T. R. Partridge
Table 2. Number of unmined and mined sites within each species: Desmoschoenus spiralis, Rumex acetosella,
community
Hypochoeris radicata, Lagurus ovatus, Calystegia
Community Unmined Mined soldanella, Zoysia pungens and Trifolium arvense.
This is one o f the richest communities.
1 21 4 U8." Zoysia pungens sand flats (31 sites). Similar
2 40 1
3 30 -- to community U7, but on flatter surfaces at the
4 31 -- westernmost two sections. Characteristic species:
5 72 6 Lagurus ovatus, Hypochoeris radicata, Rumex ace-
6 20 --
7 55 -- tosella. Zoysia pungens, Raoulia australis, Tri-
8 31 -- folium arvense, Stipa nodosa and Acaena agnipila.
9 36 -- The small sedge Carex breviculmis is locally abun-
10 35 -- dant in a few sites and the sub-shrub Pimelea
1-10 371 11 urvilleana is occasionally present.
11 7 11 U9: Grass flats (36 sites). Behind the dunes are
12 I 39
13 1 23 extensive grass fiats that extend beyond the sam-
14 13 8 piing area. Characteristic species: Stipa nodosa,
15 12 32 Rytidosperma clavatum, Lagurus ovatus, Poa cita,
16 13 11
17 8 26 Acaena agnipila, Trifolium arvense, Trifolium glom-
18 2 6 eratum, Raoulia australis, Hypochoeris radicata.
19 5 11 This is the richest of the communities.
20 5 11 UIO: Pteridium esculentum (35 sites). Dense
11-20 67 178 stands of Pteridium esculentum occur on a discrete
Bare 5 38 area o f the back dunes o f the two central sections,
the otherwise typical dune species all being less
U3: Dense Desmoschoenus spiralis in central part common. Characteristic species: Pteridium esculen-
of dunes (30 sites). Occurs only in the zone be- tum, Lagurus ovatus, Hypochoeris radicata, Rumex
tween the front dune and back dune, or on the acetosella and Bromus diandrus.
single dune crest where present. Characteristic
species: as for community U2 plus Rumex ace- Communities characteristic of mined dunes
tosella and Bromus diandrus. (M11-M20)
U4: Dense Desmoschoenus spiralis with Acaena M l l : Weedy sites (18 sites). Sparsely vegetated
agnipila (31 sites). As for c o m m u n i t y U3, plus and poorly defined with low species faithfulness
Acaena agnipila, although this species is never and abundance, but a large complement of
abundant. species. Scattered (a) on recently mined sites with
U5." Dense Desmoschoenus spiralis on back dunes little sand remaining, (b) as a weedy strand, (c) as
(78 sites). Characteristic c o m m u n i t y o f the gently isolated sites on roads and parking areas. Charac-
sloping faces o f back dunes, except where these teristic species: Rumex acetosella, Lagurus ovatus
are of low stature, mined, or covered in Pteridium and Calystegia soldanella.
esculentum. Characteristic species: Desmoschoenus M12. Recently mined sites with sand (40 sites).
spiralis, Lagurus ovatus, Hypochoeris radicata, Recently mined areas o f low sand activity nearer
Rumex acetosella and Bromus diandrus. the back of the mined dunes and rather poorly
U6: Muehlenbeckia complexa on back face of back vegetated except by Erodium cicutarium. Charac-
dune (20 sites). A distinct c o m m u n i t y o f low- teristic species: Erodium cicutarium, Bromus dian-
growing woody species mixed amongst community drus and, with low abundance, Desmoschoenus spi-
U5. Characteristic species: shrubs - - Muehlen- ralis, Lagurus ovatus and Raoulia australis.
beckia complexa, Carmichaelia appressa; herbs - - M13: Sparsely vegetated pavement (24 sites). Pave-
Desmoschoenus spiralis, Lagurus ovatus, Bromus ment occurs where wind deflation leaves a stony
diandrus, Rumex acetosella, Hypochoeris radicata, surface, this process favouring the cushion species
Acaena agnipila and Poa cita. Raoulia australis. Characteristic species: Raoulia
U7: Desmoschoenus spiralis/Zoysia pungens defla- australis, Erodium cicutarium and Desmoschoenus
tion dunes (55 sites). A c o m m u n i t y of low hum- spiralis (low cover).
mocky deflation dunes at the westernmost two M14: Sparsely vegetated pavement~dune with
sections where back dune is absent. Characteristic Rumex acetosella (21 sites). With a different com-
Vegetation recovery on coastal dunes 65
position to community M13, this occurs more often Summary of communities
on unmined areas. Characteristic species: Desmo-
schoenus spiralis, Hypochoeris radicata, Lagurus Communities U1 to U10 form the natural un-
ovatus, Rumex acetosella and Raoulia australis. mined vegetation, although the part of the dune
Notable associated rarer species include Austrofes- pavement community M14 is also unmined.
tuca littoralis, Asperula sp. and Scleranthus biflorus. Desmoschoenus spiralis dominates the dunes from
M15: New low dunes on mined areas of intermedi- the front of the front dune to the back of the back
ate age (44 sites). New, low, hummocky dunes dune. On the front face of the front dune it typi-
have formed in the seaward half of the mined sec- cally forms community U2 with U3 in the central
tions where sand is most active. Characteristic section. The back dune, where present, has U5
species: Desmoschoenus spiralis, Hypochoeris radi- mixed in with the shrub-dominated U6. Where the
cata, Raoulia australis, Lagurus ovatus and Am- back dune is absent the shrubby vegetation occurs
mophila arenaria (all low cover). A small number on the back face of the single dune and there are
of sites form a narrow unmined strand dominated extensive hummocky deflation dunes and sand
by Carex pumila. fiats (U7, U8) further inland. Beyond the dunes
M16: New dunes on older mined areas (24 sites). are grass fiats (U9). The back dunes dominated by
Mostly in areas where dense Ammophila arenaria Pteridium esculentum (U10) are an important vari-
is displacing Desmoschoenus spiralis on the oldest ant.
mined dunes. Also includes unmined front dune The mined dunes of communities M11 to M20
where Ammophila arenaria has spread up the front have a sparser vegetation. Desmoschoenus spiralis
face from its establishment point at the strand. is often present, varying from tiny isolated clumps
Characteristic species: Ammophila arenaria (dense), (M18) to relatively dense (M19), but never as
Desmoschoenus spiralis, Lagurus ovatus, Hypocho- dense as in the unmined communities. Ammophila
eris radieata and Calystegia soldanella. arenaria is important, especially in M16 and M20,
M17: Areas mined on to foredune front (34 sites). where it usually dominates as a dense stand. Asso-
Occurs where mining has both directly and indi- ciated species in these communities are variable
rectly (through over-steepening) destroyed the and unpredictable, and there are a number of
front dune, resulting in new active dunes with a weedy species that seldom occur on unmined sites,
sparse cover of Desmoschoenus spiralis. Character- the most notable being Erodium cicutarium. Where
istic species: Desmoschoenus spiralis, Calystegia wind deflation occurs a Raoulia australis stone
soldanella and Ammophila arenaria (all low cover). pavement (M 14, M 15) is established.
M18: Extremely sparse Desmoschoenus spiralis (8 The separation of the mined and unmined
sites). Almost bare, actively moving sand near the groups is very clear and highly significant (X2 =
front dune on mined areas carries Desmoschoenus 451, p < 0.0001). The greatest number of excep-
spiralis as isolated small plants, not yet building tions are of sites in the typically mined group (M)
dunes. Ammophila arenaria sometimes occurs simi- that occur on unmined dunes. Many of these are
larly. Characteristic species: Desmoschoenus spi- on the strand (M11), natural pavement (M14), ac-
ralis and Ammophila arenaria (low cover). tive front faces of the front dunes (M 17), or areas
M19." Pure Desmoschoenus spiralis on foredune invaded by marram (M16). Mined sites that carry
(16 sites). Mostly occurring on mined front dunes, vegetation of typically unmined (U) communities
but also in active blowouts, with a moderate cover are less common. Bare sites are clearly mostly re-
of Desmoschoenus spiralis, but with no associated lated to mining.
species. Characteristic species: Desmoschoenus
spiralis. Distribution of the communities
M20: Dense Ammophila arenaria (16 sites). Virtu-
ally all sites occur at the oldest mined area. There At the unmined Section A (Fig. 2), 90% of the
are few associated species that are never abun- sites carry vegetation classified as part of the un-
dant. Characteristic species: Ammophila arenaria. mined group of communities (U1-U10). Here
Bare sand (43 sites). Extensive bare areas are pre- there is only a single, but wide, dune ridge in the
sent where sand movement is too great to allow central part of the dune system. All the Desmo-
establishment, and on over-steepened dune faces, schoenus spiralis-dominated communities are well-
especially the eroding rear face of the foredune. represented (U1-U5), with some Muehlenbeckia
They are generally the result of sand-mining. complexa (U6) on the backslope. Behind this are
66 T. R. Partridge
the deflation dunes and sand flats characterised by M12, this being characteristic of recently mined
Zoysia pungens (U7, U8) and not the grass fiats of sites where sand is accumulating, and community
the central and eastern sections. Sites carrying M 13, which is pavement where sand is eroding. Be-
vegetation classified as part of the mined group hind, the unmined area is dominated by Pteridium
(M) include some weedy strand areas (Mll), a esculentum (U10), with grass flats (U9) beyond.
large patch of dense Ammophila arenaria on the Mining at Section E is restricted to a narrow
front dune face (M16), and a few sites in the east part between the foredune and backdune. The
that have been destabilised by adjacent sand min- foredune is made up of many communities includ-
ing (M 15). ing active foredunes otherwise characteristic of
At Area B, sand-mining has removed most of mined areas (M17), and both moderate (U1) and
the front and central sections of the dunes. Of the dense (U2) Desmoschoenus spiralis. There are two
Desmoschoenus spiralis-dominated unmined com- areas of community M14, sparsely vegetated
munities, only the common U5 is at all well-repre- dune/pavement, both on unmined dunes. In the
sented. Most notable of the mined communities west a small area has been left unmined to protect
are M16 and M20, both characterised by dense a population of the rare Asperula sp., while an
Ammophila arenaria, this also occupying unmined area in the east has Austrofestuca littoralis. Much
foredune remnants and some of the unmined of the mined area is bare or has weeds ( M l l ) or
backdunes as well. Behind the foredune remnants establishing dunes (M15). A number of mined
are extensive bare areas where sand is extremely sites have the otherwise typically unmined com-
mobile. The grouping of sites of the typically munity U5 - - these have been invaded by many
recently mined community M 12 at the eastern end Desmoschoenus spiralis seedlings and associated
represents an area that was remined in the 1980s. species common in unmined areas on sand of low
Beyond the mined area, the deflation dune and activity. More typical community U5 occurs be-
sand flat communities (U7, U8) typical of Section hind the mined area, although in the west there is
A are again well-represented. some Pteridium esculentum (U10). Grass fiats be-
The mining at Section C has been far more exten- hind are mostly of community U9.
sive. Sites representing typically unmined commu- At Area E, unmined communities dominate, ex-
nities are present only on foredune remnants (U2), cept near the foredune front where some active
small unmined 'islands' such as that in the east areas have community M 15. Unlike Area A, there
(U2-U4), and backdunes (U5, U6). Pteridium es- are two distinct dune ridges and no sand flats.
culentum (U10) dominates most of the backdunes Dense Desmoschoenus spiralis occurs throughout,
and spreads onto the grass fiats (U9) beyond, al- with communities U2-U5 all being well-repre-
though the western end is beyond the range of this sented variants. In the west the backslope of the
species and has sand flats (U8). The mined area is backdune has abundant woody vegetation, espe-
dominated by sparse vegetation, especially of cially of Muehlenbeckia complexa (U6), while be-
community M15, which is characteristic of mined yond are grass fiats (U9).
areas of intermediate age, most sites of this com-
munity occurring here, Many of these sites con- Ordination
tain Ammophila arenaria. Also well-represented is
community M12, which dominates a large, deep The first two axes of the DCA site ordination are
hollow in the east, this being an area that was presented using separate diagrams for each of two
remined in the 1980s. On some of the older mined sets of site attributes. Separate diagrams were used
areas, erosion of sand has resulted in Raoulia aus- to avoid clutter, as there were many sampling
tralis-dominated dune pavement (M 16). sites. Figure 3 divides the sites into mined versus
Foredune damage has been a feature of the unmined. There is, in general, a good separation
mining at Section D. This has resulted from over- of the sites both on axis 1 and especially axis 2.
steepening of the backslope which has been fol- The many unmined sites on the lower left are
lowed by blowouts, to leave large active areas be- densely clumped whereas the mined sites on the
tween small, isolated foredune turrets. The upper right are more scattered, those with the
distinctive community M17, with its presence of highest values on axis 2 being vegetation domi-
Calystegia soldanella, now occupies most of the nated by non-dune weeds and Erodium cicutarium
former foredune area along with many bare areas (Fig. 4). There is, however, considerable overlap
of moving sand. The central part has community of mined and unmined sites at the lower right.
Vegetation recovery on coastal dunes 67
350 Unmined 350 Mined
300 300
• a
- : , ,
250 250 ... s
¢q • o, ,,.
200 o = 200
• . ' . : . • ".'%, .
<
. . . . -: =~.~..~. ~.:
150
,......: . :~.':-~. ,.;; . 150
" ••
• :,, • .. , .%$...~'v. ..t~.~.. .~
• ...: ".~. ¢.:.,. .:'¢,.'... ,
100
• :. ~ ; 100
•. • • ,.. , ,°
• qk~l., . • •
21 -:., t. .
• . " . • s ..
50 • e" • " " . •° 50
0 0
50 100 150 200 250 '300 350 400 50 100 150 200 250 300 350 400
Axis | Axis 1
(a) (b)
Fig. 3. Site ordination divided into (a) unmined and (b) mined sites.
These are dominated by Ammophila arenaria and 1, and from bottom to top on axis 2, with increas-
occur on the unmined foredune where this species ing distance inland. The trend is more pronounced,
has displaced Desmoschoenus spiralis, and on the however, on axis 1. Unmined sites show the trend
oldest mined areas. very clearly while it is also present, but less well-
Figure 5 separates the sites by distance from the defined, in the mined areas. Mined sites, as shown
sea, each diagram representing a 40 m step. The in Fig. 1, are mostly in the central sections, leaving
trend is from coastal (right) to inland (left) on axis the extremes the mostly clearly defined groupings.
eLolium eAnagalis
400' eErodium
Ciraiume
300 Po8
• "'e ° l c u s
n Bromus
Pteridium • Raoulia
Elymua
¢-4
T.glomeratum •MeliCytus
~< 200 Rumex
< • Muehlenbeckia
Carmichaelia e •
eCarex ekagurus
Acaena
100 Stipa eH y p ° c h ° e r | $ Desmoschoenus
• eSilene •
Zoysia
~ytidosperma •
• Ammophila e
T.arvense
Aira
eCalystegla
-loo 0 loo- 200 300 4o0
Axis 1
Fig. 4. Species ordination showing only the most important species (Table 1). Foeniculum vulgare and Elymus rectisetus are not
shown as they had large values on axis 2.
68 T. R. Partridge
350 0-40 m 350 120-160 m
3OO 300
250 250
eq
20(1 200 .: ..=."
• "" "*;" "k
.. : .':: :-...
150 150 -.
. ..! ." ~. :
° ••
°.° .~ :!-
100 •~ .'.'... 100
*l*gt . ".- • •
• ~'*• 0 0
50 J l 0
50
0
0 50 100 150 200 250' 300 350 400 o 5'0 xJo 3;o
Axis 1
350 40-80 m 350 160-200 m
°
300 300
250 250 " • |
200 200 ~° ••
t
° , ~
I • " •° :
150 150
° •
• •l~a • . , •
..° . ...... ..:.~
100 •* "~ - .o . °"
•d ¢ ° • • "°
100 °••° ° • •
° •° °
• ° •• . •
50 "-. 50
0 i i i
0 50 150 200 250 300 350 50 100 150 200 250 300 350 400
350 80-120 m 350 200-240 m
300 300
250 °~ 250
•• • .*
° °
200 t° •
°
200 •°
• :-.. •
• •• ~°°°, ° °~ ••
150 ° ,.° •
150 .;. • •
°-~ ° •"
":.." : : - • •D
..:~'." °, °
100 • .. • • 100
0 m 0
50 50
0 0
50 100 150 200 250 300 350 400 Jr) 100 150 200 250 300 350 400
F i g . 5. Site ordination divided according to distance from the shoreline. Each diagram r e p r e s e n t s a s t e p o f 40 m.
The ordination clearly shows two important and mined vegetation. Axis 2 might also be inter-
trends within the vegetation. The typical sea-to- preted as representing a stability gradient, this
land gradient is most obvious on axis l, while axis being very much the product of mining, but with a
2 more represents the difference between unmined coast to inland component as well.
Vegetation recovery on coastal dunes 69
DISCUSSION systems in other parts of New Zealand, such as
the Manawatu (Esler, 1969).
Sand-mining has caused major changes to both Ammophila arenaria is clearly having an impor-
the physical structure and the plant communities tant impact upon the dunes. On unmined areas
of the Kaitorete dunes. The central parts of the this grass establishes probably mostly from rhi-
dunes have been hollowed out to varying depths, zome fragments along the strand and spreads
and those areas mined in the 1960s and 1970s slowly up the front of the foredune by vegetative
have suffered the greatest, with sand having been growth (Gemmell et al., 1953). Its establishment in
removed from virtually the entire dune system. this zone demonstrates a distinct preference for
Furthermore, the effects of mining have caused in- areas of maximum sand activity. It is this prefer-
stability on some adjacent areas by over-steepen- ence for active sites that has caused it to become
ing of faces, and through mining sand from below so dominant in parts of the mined dunes. In such
mean sea level. areas, mining has cut into the stands of Am-
The results clearly demonstrate that the major mophila that had previously colonised the front
effect of mining on the plant communities is the face of the foredune, thus supplying the active
establishment of a suite of communities that are areas behind with an abundance of stem frag-
less common on unmined dunes, including those ments. In such areas initial establishment follow-
that are characterised by ephemeral weeds, estab- ing mining involves a large number of
lishment areas of Desmoschoenus spiralis, recently Desmoschoenus spiralis seedlings and a smaller
developed Raoulia australis pavement and stands number of Ammophila clumps from stem frag-
of Ammophila arenaria. Where these communities ments (community M12, for instance). However,
do occur on unmined dunes, they mostly tend to the ratio changes rapidly as (1) the Ammophila
be on disturbed sites of active dunes including the grows and spreads at a much faster rate, (2) the
strand at the base of the foredune. The major ex- Desmoschoenus is heavily grazed by rabbits and
ception is the stable unmined dune pavement that hares, and (3) Desmoschoenus disappears when the
clusters as one of the mined communities. two species come into competition, resulting even-
Mining has occurred over a period of some 40 tually in community M20. The result is a dense
years. If the plant communities are recovering, Ammophila stand with only little Desmoschoenus
then it is expected that there would be mined sites remaining. From the back of the mined area, Am-
that carry communities otherwise characteristic of mophila slowly spreads vegetatively on to the
unmined areas, especially on the older mined dunes behind as well, in the same way that it in-
dunes. The results clearly demonstrate that this vades unmined foredunes.
has not taken place. There is indeed no indication The more recently mined areas have not been
that mined areas are developing vegetation similar invaded by Ammophila arenaria. Here the initial
to that on unmined dunes. Such failure was noted establishment of Desmoschoenus spiralis proceeds
by Woodhouse (1982) on the east coast of North in two possible directions. Where sand accumula-
America. Although Desmoschoenus spiralis has tion occurs low dune hummocks of Desmo-
colonised mined areas it has formed novel, sparse schoenus establish, but where erosion takes place,
communities. The failure of the dunes to recover a stone pavement dominated by Raoulia australis
their original communities can be attributed to two develops. There is no real sign, however, of a ten-
factors. The first is the invasion of much of the dency for Desmoschoenus to build higher dunes,
older mined area by Ammophila arenaria. This probably because the sand supply is limited. There
:species has effectively halted any potential recov- is little prospect for such sand becoming available.
ery by the development of a dense sward that Mining has introduced a number of other weeds
competitively excludes Desmoschoenus spiralis. to the area. These are presently uncommon, but
The second is the development of a continuous some, including Lupinus arboreus and Cytisus
sand plain that runs parallel to the beach front, scoparius, have the ability to spread and become
through the mined area. Within this sand plain problems. All these species occur close to vehicle
lateral winds move sand through the mined dunes tracks and their presence can be attributed to
to form a sparsely vegetated hummocky low dune introduction by vehicles.
system dominated by Desmoschoenus spiralis The back dunes of the central mined area have
along with some erosion pavement. Such sand- a dense covering of Pteridium esculentum fern in-
plain vegetation is characteristic of some dune stead of the Desmoschoenus spiralis and Muehlen-
70 T. R. Partridge
beckia complexa typical of other areas. The ACKNOWLEDGEMENTS
success of Pteridium cannot be attributed to min-
ing, but may be the result of fires that have spread The author wishes to thank Richard Cross for
onto the dunes from the grassland. Indeed, to the help with field work and Brain Molloy, Peter
east, Pteridium covers the entire dune system in an Johnson, H a b i b a Gitay, Colin M e u r k and Bastow
area that was previously a military firing range Wilson for comments. The study was funded by
and has a history of burning. Desmoschoenus spi- the N e w Zealand Department o f Conservation.
ralis seems unable to compete with Pteridium, but
it is not known whether the fern is increasing its
range. REFERENCES
The non-recovery of plant communities and the
Armon, J. W. (1974). Late quaternary shorelines near Lake
invasion by A m m o p h i l a arenaria have important Ellesmere, Canterbury, New Zealand. N . Z . J . Geol., 17,
implications for both mining and the management 63-73.
of the dune botanical resources. It is clear that Belbin, L. (1989) PATN Technical Reference. CSIRO Divi-
sion of Wildlife & Ecology, Canberra.
sand mining creates conditions in which commu- Bradshaw, A. D. & Chadwick, M. J. (1980). The Restoration
nity recovery is unlikely. Restorative planting and of Land. The Ecology and Reclamation of Derelict and De-
replacement of a layer o f sand seem to be having graded Land. Blackwell Scientific Publications, Oxford.
no effect. These plantings of pot-grown material Brooks, D. R. (1976). Rehabilitation following mineral sand
mining on North Stradbrooke Island, Queensland. In
have been more successful than the cuttings used Landscaping and Land Use Planning as Related to Mining
earlier, but there is so far no evidence that sur- Operations. Australian Institute of Mineralogy and Metal-
vival is better than in areas left unplanted. Al- lurgy, Adelaide, pp. 93-104.
Burrows, C. J. (1969). Flora and vegetation of Kaitorete
though some of the communities that establish are (Ellesmere) Spit. Cant. Bot. Soc. J., 2, 5-10.
dominated by native plants, they are not the same Clifford, H. T. & Stephenson, W. (1975). An Introduction to
as those removed during mining. The mining, Numerical Classification. Academic Press, New York.
Cockayne, L. (1911). Report on the Dune-areas of New
however, is due to cease soon, although much of Zealand, Their Geology, Botany and Reclamation. Appendix
the central part o f Section F is planned for sand to the Journal of the NZ House of Representatives, C13.
extraction. This pattern is in contrast to the situ- Government Printer Wellington, pp. !-76.
ation in Queensland, Australia (Brooks, 1976; Cockayne, L. (1982). The Vegetation of New Zealand, 2nd
Edn. Englemann Press, Leipzig.
Bradshaw & Chadwick, 1980), where restorative Connor, H. E. & Edgar, E. (1987). Name changes in the in-
planting on mined dunes has been successful. digenous New Zealand Flora 1960-1986 and Nomina
However, there most of the sand has been re- Nova IV, 1983-1986. N.Z.J. Bot., 25, 115-70.
Courtney, S. P. (1983). Aspects of the ecology of
turned to the system following removal of rare Desmoschoenus spiralis (A. Rich) Hook f. MSc thesis, Uni-
minerals, unlike the situation at Kaitorete where versity of Canterbury.
virtually all the sand is removed and virtually Down, C. G. & Stocks, J. (1977). Environmental Impact of
Mining. Applied Science Publishers, London.
none returned. Esler, A. E. (1969). Manawatu sand plain vegetation. Proc.
If left unchecked, it seems that A m m o p h i l a are- N.Z. Ecol. Soc., 16, 32-5.
naria will spread over the entire mined area and Esler, A. E. (1970). Manawatu sand dune vegetation. Proc.
N.Z. Ecol. Soc., 17, 41-6.
onto adjacent dunes as well. In the absence of Esler, A. E. (1974). Vegetation of the sand country bordering
mining, this species is restricted to along the the Waitakere Range, Auckland: the southern beaches.
strand, and has only a slow, largely manageable Proc. N.Z. Ecol. Soc., 21, 72-7.
spread up the foredune. In only a small number of Esler, A. E. (1975). Vegetation of the sand country bordering
the Waitakere Range, Auckland: Piha Beach. Proc. N.Z.
locations has it reached a situation where its man- Ecol. Soc., 22, 52-6.
agement requires extensive and expensive treat- Gemmell, A. R., Greig-Smith, P. & Gimingham, C. H.
ment. If allowed to spread throughout the mined (1953). A note on the behaviour of Ammophila arenaria
(L.) Link in relation to sand-dune formation. Trans. Bot.
area, it will become a vast management problem. Soc. Edinb., 36, 132-6.
The oldest mined dunes have already reached that Hill, M. O. & Gauch, H. G. (1980). Detrended correspon-
stage, but the remaining areas where A m m o p h i l a dence analysis: an improved ordination technique. Vegeta-
tio, 42, 47-58.
arenaria is either sparse or absent have the poten-
Holland, L. D. (1981). Plants and sand dune development;
tial to be saved from the invasion o f this species. Ammophila arenaria versus Desmoschoenus spiralis on
To preserve the valuable botanical features of the Kaitorete Barrier, Canterbury. MSc thesis, University of
Kaitorete Spit dunes, it is important that sand- Canterbury.
Johnson, M. S. & Bradshaw, A. D. (1979) Ecological princi-
mining should cease, and A m m o p h i l a arenaria ples for the restoration of disturbed and degraded land.
spread be controlled. Appl. Biol., 4, 141-200.
Vegetation recovery on coastal dunes 71
Johnson, P. N. (1982). Naturalised plants in south-west Smith, S. M., Allen, R. B. & Daly, B. K. (1985). Soil-vegeta-
South Island, New Zealand. N.Z.J. Bot., 20, 131~,2. tion relationships on a sequence of sand dunes, Tautuku
Kirk, R. M. (1969). Beach erosion and coastal development Beach, South-east Otago, New Zealand. J. R. Soc. N.Z.,
in the Canterbury Bight. N.Z. Geogr., 25, 23-35. 15, 295-312.
Lewis, J W. (1976). Regeneration of coastal ecosystems Sykes, M. T. & Wilson, J. B. (1987). The vegetation of a New
after mineral sand mining. Australian Mining, July 1976, Zealand dune slack. Vegetatio, 71, 13-19.
22-9. Sykes, M T. & Wilson, J. B. (1988). An experimental investi-
Logan, M. C. & Holloway, J. E. (1934). Plant succession on gation into the response of some New Zealand sand dune
the Oreti River sand dunes. Trans. R. Soc. N.Z., 64, species to salt spray. Ann. Bot., 62, 159-66.
122-39. Sykes, M. T. & Wilson, J. B. (1989). The effect of salinity on
Palmer, J. D. (1980). The Kaitorete Dunes. Department of the growth of some New Zealand sand dune species. Acta
Lands and Survey, Christchurch. Bot. Neerl., 38, 173-82.
Peace, M. (1975). The plant ecology of the dune systems of Webb, C. J., Sykes, W. R. & Garnock-Jones, P. J. (1988).
Kaitorete Spit. MSc thesis, University of Canterbury. Flora of New Zealand, Vol. IV. Botany Division, DSIR,
Pegg, E. J. (1914). An ecological study of some New Zealand Christchurch.
sand-dune plants. Trans. Proc. N.Z. Inst., 46, 150-77. Woodhouse, W. W. Jr (1982). Coastal sand dunes of the US.
Simpson, M. J. A. & Mason, R. (1984). The flora of the Can- In Creation and Restoration of Coastal Plant Communities,
terbury dune system (New Brighton-Waipara) in the 1980s ed. R. R. Lewis III. CRC Press, Botan Roca, Florida, pp.
Mauri Ora, 11, 99-111. 1-44.