New Zealand's alpine region is populated by many (∼613) species of vascular plants with a high endemism (∼93%). To investigate the potential impact of climate warming, we used species-area relations to estimate current and projected vascular plant floras and tested model sensitivity scaling from the whole world to small alpine regions.
Within their limitations, these models show that if the present mean temperature of ∼0.6°C higher than in 1900 were maintained, together with a large pool of exotic species, 40–70 species of native plants could become at risk. With a rise of 3°C, an approximate expectation for the following 100 yr, the total New Zealand alpine vascular flora could reach ∼550–685 species and lose 200–300 indigenous alpine species, the rest being exotic. Fragmentation of alpine areas could, over millennia, favor speciation, but in the short term, the loss of ∼80% of existing alpine islands will severely increase extinction risks.
These model projections will be modified by downward extension of species through unplanned vegetation destruction, or following deliberate vegetation clearance to create habitats favorable to alpine species, as well as through a number of other as yet unquantified factors. These projections are not predictions of extinctions but rather broad probabilities of risk to a whole flora.
Introduction
Global climate change has affected biotas through the ages, but the present century is predicted to show a rate of temperature rise amongst the highest in the world's history. The 1997 Intergovernmental Panel on Climate Change (IPCC) scenarios of future global warming suggested an increase in temperature of 1 to 3.5°C by the year 2100 (Watson et al., 1997); the 2001 report increased these estimates to 1.4 to 5.8°C (IPCC, 2001). It has been claimed that many species of plants and animals will be at increased risk of extinction due to warming climates and that mountaintop communities are particularly at risk given their isolation and lack of an “escape route” (Grabherr et al., 1995).
New Zealand is unique in being simultaneously one of the most isolated, ancient, and topographically diverse continental islands. Its form and size have fluctuated markedly over millions of years (Lee et al., 2001), significantly affecting the geographic distribution of alpine species (McGlone et al., 2001). Although the land is ancient, the present mountains developed only in the last 5 million to 7 million yr (Walcott, 1998). The high species richness of alpine vascular plants (>600 species; Mark and Adams, 1995) and invertebrates (Watt, 1975) in the alpine region may seem extraordinary in view of an evolutionary time span of only ∼2 million yr (Wardle, 1978), although a longer time span may have been available, according to other studies (Fleming, 1963; Wardle, 1968; McGlone, 1985; McGlone et al., 2001). Overall endemism of these plants is around 93%, but many can exist below treeline in a variety of open environments (Mark and Adams, 1995; McGlone et al., 2001).
Climate change and increasing rates of human disturbance will affect the alpine biota concurrently. Human impact since first arrival around 1,000 yr ago has been pervasive. Forest destruction in many areas has allowed alpine and subalpine plants to significantly extend their ranges downslope, while high-alpine plants have extended locally into low-alpine tussock grasslands opened by burning and grazing by introduced herbivores (e.g., McGlone et al., 1997). To a large degree, these relatively recent changes override potential present climate-change signals for rising vegetation lines.
Several approaches can be used to study the effects of climate change on mountain biodiversity. Long-term monitoring through photo-points or permanent plots has a long history in New Zealand (Evans, 1973; Mark, 1977; Wills and Begg, 1986; Scott et al., 1988; Dickinson et al., 1992; Bellingham et al., 1999; Bellingham et al., 2000, Duncan et al., 2001). A claim for a small recent advance in southern beech treelines (Wardle and Coleman, 1992) was disputed by a later study designed specifically to determine the response to global warming. The latter study found occasional storm disturbances to have obscured any possible climatic response (Cullen et al., 2001). On a wider scale, an alpine vegetation transect on the Southern Alps in south Westland is one of the most detailed such studies in New Zealand (Mark et al., 2000), and its repetition in a number of years should provide key information on change. Although many of the locations and designs of these studies were meant to determine the effects of introduced herbivores, fire, and other more or less direct human impacts on vegetation, some of this information will be of use for assessment of effects of climate change as well. Careful analysis of historical records may also provide valuable insights.
Additional research approaches including spatial explicit modeling and experimental manipulation at a small scale would provide new insights (Gottfried et al., 1998; Guisan et al., 1998). A large scale approach, used here for the first time in New Zealand, is to model nonlinear species-area relations at the appropriate scales to provide scenarios of climate-change effect as vegetation zones migrate. This approach provides hypotheses that can help design more rigorous long-term monitoring as well as developing scenarios to guide management responses. Calculations of species-area relationships have been used to estimate threats from climate change in a detailed study on mammals in isolated mountains of the U.S. Southwest (McDonald and Brown, 1992) and have also been used on a very much coarser global scale (Malcolm and Markham, 2000). We used recent 1:50,000 topographic maps with 20-m contour intervals, incorporating latitudinal changes in altitudinal zonation of treeline to provide a credible range of estimates of the risk of vascular plant species loss under various climate-change scenarios.
Methods
New Zealand was divided into 16 rectangular horizontal blocks (Fig. 1) from the Land Information NZ (LINZ, http://www.linz.govt.nz) database of 1:50,000 contour maps. We used ArcInfo to create triangular irregular networks (TINs), then calculated areas above each 100-m contour line with ArcView (37 vertical slices). Treeline (or timberline; they are similar in New Zealand and represent the lower limit of the alpine vegetation) was set at 900 m asl for the southernmost regions, gradually increasing to 1400 m for the northernmost mountains. The increase in treeline from coastal to inland mountains was not considered, given the scale of the study.
The species-area relationship was modeled by the nonlinear Arrhenius equation y = a + bx, where y is log number of species, x is log of area, and a and b are constants derived from the empirical data (Preston, 1962). Different sets of empirical data will lead to different values of a and b, and hence estimates of species numbers. The model is better known in island biogeography and conservation biology for estimates of species loss with area reduction (MacArthur and Wilson, 1967; Simberloff, 1974; May, 1981; Soulé and Wilcox, 1980). Although such uses have their limitations (Lugo et al., 1993; Brokaw, 1998) they are underpinned by the theoretical basis of lognormal species abundance distributions (Preston, 1962; Halloy, 1998; Hubbell, 2001). The species-area relation is a very general yet simple rule, but if the parameters are not correctly chosen, the calculations are of little value. We address some of these practical limitations by testing the sensitivity of the species-area function with six different data sets (Table 1), scaling down from global estimates to smaller sets increasingly relevant to the New Zealand alpine vascular flora (full list of sites available on request):
A = 67 areas around the world
B = 12 islands up to the size of Madagascar
C = 30 alpine regions around the world
D = 6 alpine regions in New Zealand + total alpine New Zealand + total New Zealand
E = 6 alpine regions in New Zealand
F = 2 points: total alpine New Zealand + total New Zealand
Estimations of actual species numbers tend to increase with time as biotas become better known. However, given the exponential nature of species-area relationships, small changes in estimates will not affect the general shape of the relation based on many sites.
The number of indigenous species at saturation after climate change is estimated parsimoniously as a proportion of the total-saturation flora at the given scenario level equal to the proportion of the original natives to the original total-saturation flora (saturation is used here interchangeably with the term equilibrium as used in original biogeographic theory from MacArthur and Wilson, 1963 onward). The assumption of neutral species-area models is that “all species are equal” in a statistical sense (e.g., Hubbell, 2001). Hence, as the alpine area contracts, exotic and native species are equally at risk (although see discussion on types of species more at risk).
The reduction in area available is determined assuming a 0.6°C/100-m altitudinal lapse rate (Körner, 1999; Mark et al., 2000; Wardle, 1998). Thus, if, for example, treeline was at 1000 m, in the 3°C warming scenario it will be raised to around 1500 m, and the new area for alpine plants will have been reduced accordingly (Fig. 2). The scenario of rising forests assumes that the New Zealand treeline is determined to a major extent by temperature (Körner, 1999; Tranquillini, 1979; Wardle, 1974). The increase in available area above the present upper limit of alpine plants is insignificant at this scale, Only 0.03% of the total area above treeline is higher than the highest recorded vascular plants.
In addition to area reduction, we studied the effect of a rising treeline on the fragmentation and elimination of alpine “islands” (every separate mountain area emerging above the assumed treeline altitude and separated from other such areas by passes below that altitude) by quantifying the transformation as contour lines are raised from 1000 to 1500 m on a 1:1,000,000 topographic map of New Zealand (Department of Lands and Survey, 1984). Such an analysis is a rough approximation because it disregards the change in altitude of the treeline with latitude as well as the smallest fragments not apparent at the mapping scale.
Results
First we describe the prehuman species levels and how they compare to the species-area models. Then we move to the present situation as affected by habitat degradation, exotic species, and developing climate warming. Finally we examine the effect of expected warming in the coming 100 yr, in terms of both total area reduction and its effects on the loss of alpine islands (fragments).
PREHUMAN BENCHMARK
In the Landcare Research database, 2039 native and 2065 adventive vascular plant species are recorded for the contemporary vascular plant flora of New Zealand (E. Nicol, 1997, in Halloy et al., 2000). Groombridge (1992) estimated the total indigenous vascular flora at 2371 species. To avoid overestimating extinctions, we use a slightly higher estimate of 2450 species for all of New Zealand. The total New Zealand alpine vascular flora has been estimated at about 613 species (Mark and Adams, 1995). Alpine plants are distributed with a maximum species richness around 1250-m elevation, with numbers declining rapidly below treeline elevation and around 50 species when elevation reaches sea level (Fig. 3). Although occasional isolated plants have been recorded up to 2930 m (Mark and Adams, 1995), we used a conservative assumption of 0 plants above 2400 m because extremely few vascular plants grow above this altitude. These numbers of vascular plants are good representations of the prehuman benchmark as only around 3 (Groombridge, 1992) to 5 (de Lange et al., 1999) vascular plant species may have become extinct since human settlement, and none of them are alpine.
As one might expect from an island isolated for 80 million yr, both the total vascular flora and its alpine component are well below species numbers for areas of similar size that are in contact with larger species pools, as shown in Scenarios A (estimates close to 5000 species) and B (∼3600 species) for the total flora and Scenario C (∼860 species) for the alpine flora (Table 1). Although Scenario C underestimates the total flora (giving ∼1630 species), it provides the overall best estimate for the alpine flora at different scales. Scenarios D and E provide the closest estimates to the very small Mt. Burns high-alpine area, but at the cost of strongly underestimating the total New Zealand flora. Scenario F is useful as a comparative benchmark but less so for small-scale alpine regions. Later we discuss only the most likely scenario (C), together with two extremes presented by B and E, to show the range (sensitivity) of potential loss of species. To verify the precision of each model at different scales, Table 1 provides three specific alpine areas differing in size and documented diversity (from Halloy and Mark, 1996).
PRESENT SITUATION—DEGRADATION AND EXOTIC SPECIES
After ∼800–1000 years of Polynesian and 160 years of European occupation, around 90% of indigenous vegetation has been degraded to some degree (Halloy et al., 2000). At the same time, a large pool of vascular plant species has been introduced into New Zealand, with estimates ranging from 14,195 counted in the Landcare Research database (E. Nicol, 1997, in Halloy et al., 2000) to 26,000 reported from the Ministry of Agriculture and Fisheries database of plants in New Zealand (M. Dickson, 2001 pers. com.) and 40,000 estimated by an expert (W. Sykes, 2001 pers. com.). Already 2065 exotic vascular plant species, more than the total number of indigenous vascular species recorded, have become established in the wild (E. Nicol, 1997, in Halloy et al., 2000). However, most of these species are only in and around urban areas and farmland. Very few exotic vascular species have become established in alpine regions, many of which regions show a relatively low level of degradation, particularly in the higher-rainfall regions (Mark et al., 2000).
Given the larger pool of exotic vascular species now available in New Zealand, it has been estimated that the total New Zealand vascular flora could rise from the present ∼2450 to ∼5000 species (Halloy, 1996). Estimates A and B in Table 1 are in reasonable agreement with this projection, as is the rapid human-mediated establishment of exotic vascular plants up to now. At the same time, as a result of both habitat degradation and exotic competition, the probabilities are that indigenous vascular species could decline to less than half their current number, or close to 1000 species, with the remaining 4000 species being exotic. These changes would occur regardless of climate change in a “biogeographic equilibrium” situation (sensu MacArthur and Wilson, 1963).
For the alpine zone the &ldquo=uilibrium”, or saturated flora, assuming that a large pool of appropriately adapted exotic vascular species remains available for a long time, could rise to around 860 (Scenario C) to 1350 (B) species. As long as habitat degradation remains low, such an equilibrium alpine flora would still include close to the present 600 native vascular plant species. However, present invasion by exotic alpine vascular species into New Zealand's alpine zone seems to be delayed by the following factors: (1) Exotic species in New Zealand are predominantly of Northern Hemisphere origin. Northern Hemisphere alpine species are from continental climates with short, warm summers [in contrast to Southern Hemisphere alpine longer but cooler summers and milder winters, e.g. Halloy and Mark (1996), Mark et al. (2000), Troll (1948), Troll (1960)], and few seem to adapt to New Zealand's oceanic alpine climate. (2) The pool of potentially adapted alpine vascular species is small compared to the total number of exotics. (3) Most true alpine species have low dispersal capability, and most are grown in lowland gardens far from the alpine regions.
PRESENT SITUATION—CLIMATE WARMING
If the present mean temperature of around 0.6°C higher than in previous centuries were to be maintained for a long time, the equilibrium vascular flora could still rise to around 800 (Scenario C) to 1200 (Scenario B) species, including some 530 (Scenarios A and B) to 560 (Scenario D) native species. Hence, some 40 to 70 indigenous vascular species are already at risk of extinction as a result of climatic warming alone.
FUTURE PROJECTIONS
Given the range of temperature-change scenarios presented by the IPCC for 2100 (increase of global temperature between 1.4 to 5.8°C, IPCC 2001), we calculated the sensitivity of the different species-area models and show a conservative middle estimate as well as the range of scenarios in Figures 4 and 5. With a midpoint rise of 3°C the total present alpine area estimate of 30,000 km2 will reduce to 15,400 km2. The total equilibrium New Zealand alpine vascular flora (exotic and indigenous) could drop to around 460 (D) to 685 species (B, Fig. 4), which would include 300 (B) to 410 (D) indigenous species. The number of indigenous alpine species remaining ranges from 37 to 68% of the original flora with all scenarios, with 51 (B) to 64% (C) the more parsimonious values (Fig. 5). Thus, with a 3°C temperature rise, the equilibrium New Zealand alpine vascular flora may lose 200–300 species.
FRAGMENTATION
We counted 441 “alpine islands” above 1000 m and 364 above 1500 m in New Zealand (Table 2). A comparison between the North and South Island's provides insights on how topographic differences may have a major effect on biogeographic fragmentation and evolution. With its alpine area dominated by relatively smooth volcanic cones, the North Island has 87 separate areas above 1000 m. If vegetation zones rose 500 m (with a 3°C temperature rise), 81 of these islands would be lost (93%, in addition to major area reductions). In turn, the 6 remaining islands would fragment into 20 smaller islands. In contrast, the more rugged South Island has 354 separated areas above 1000 m, and 344 would still exist above 1500 m. However, 77% of the original 354 would have disappeared, the remaining ones being new fragments from the larger remaining blocks.
Discussion
DYNAMIC EXPANSION AND CONTRACTION
Palynological, sedimentological, and other paleovegetation and paleoclimatic evidence shows that New Zealand underwent major vegetation changes during and after the last glaciation (e.g., Fleming, 1963; McGlone, 1985, 1989; Stevens et al., 1988). During the last glaciation, suitable habitat for alpine plants would have been substantially larger than at present. Yet the number of species presently below biogeographic equilibrium, even for the smaller area today, suggests that the former larger area did not persist long enough for the evolution of a saturated flora or that some major extinction event intervened. Major human-induced habitat modification in New Zealand in the last millennium has created a range of situations from rapid recovery to ongoing degradation; changes that in many cases could override potential effects of climate change until the present. However, as the magnitude of climate change increases, the effects should become increasingly apparent over and above local degradation.
The fragmentation analysis highlights the dynamic fluctuations in area, size distribution of those areas, and distances between them that are fundamentals of the evolutionary biogeography of extinction and speciation. Clearly, the temporal variations in past millennia of fluctuating vegetation lines would have created conditions of fragmentation and refugia formation, followed by fusion and amalgamation, over and over again. Such conditions have been postulated as causes for high diversity and evolutionary dynamics in areas such as the Amazon (Brokaw, 1998) and may relate to the high speciation rate observed in several endemic or mostly endemic alpine vascular genera of New Zealand (e.g., Celmisia and Aciphylla; see for example Wardle, 1978; Lee et al., 2001; McGlone et al., 2001). Yet, paradoxically given this potential increase in species richness, area reduction alone leads to reduction of species. Total submersion of alpine islands beneath forest would clearly lead to extinction of at least some low-dispersal local endemics. The reason for this paradox is that the time scale of extinction may be decades and centuries, while that for speciation is more like hundreds of thousands to millions of years.
An additional component of threat to population viability occurs through reduction in the number of populations of any one species. For example, a species with a total available area of x, distributed in 50 populations over 50 mountain peaks, could have its total area reduced to x/2, but the number of populations could conceivably reduce to only five or even one. Such populations can no longer benefit from metapopulation dynamic exchanges and are at increased risk of extinction (Eriksson and Jakobsson, 1998; Tilman, 1996).
MODIFYING FACTORS
Many factors could affect models of extinction predictions:
1. downward extension of alpine plants through forest and shrubland destruction;
2. varying rates of upward invasion of exotic plants, depending on the nature of the species pool;
3. the fact that the flora is below saturation;
4. in addition to temperatures rising, the increase in carbon dioxide partial pressure could enhance growth rates and competitive ability (Halloy, 1981), favoring forests and shrublands in competition with alpine plants;
5. climatic and topographic heterogeneity, including pockets of drier, wetter, windier, cooler, steeper, or other less common conditions, some of which may restrict the advance of more competitive vegetation and allow the survival of alpine species;
6. biotic interactions, including disease and predation, in addition to the competition factor already considered.
RATES OF CHANGE
The calculations of probabilities are based on an achieved equilibrium, possibly after hundreds or thousands of years. However, the conservative scenario of a 3°C mean temperature rise by 2100 implies a rise of isotherms of around 50 m per decade. It is unlikely that most vascular plant species will be able to move at such rates (including the rising forests). A long-term study in the European Alps suggests that the most common alpine species ascended 1–2 m per decade, with a maximum of 4 m during the last half century in which climatic warming equated to an 8–10-m rise or more per decade (Grabherr et al., 1995). Although there is still uncertainty (Cullen et al., 2001), comparable rates of ascent have been noted for the New Zealand treeline (Wardle and Coleman, 1992). The speed of change is thus likely to place pressure on vascular plants near vegetation limits over and above the pressures due to area reduction and fragmentation.
In addition, extinction may be subject to long lag times. A temperature rise of 3°C may place a certain number of species at risk, but extinction may not follow for many decades, and the species may linger on as an “extinction debt” (Tilman, 1996) or “living dead” (Revkin, 2000). It is possible to attempt to identify the species most at risk through a combination of ecological and biogeographic information (e.g., dispersal capability, habitat specialization, population sizes and distribution, recruitment rates). For example, McGlone et al. (2001) estimated that out of 571 New Zealand alpine vascular plants, 244 (43%) had low dispersal capabilities. These species (which are also mostly specialists with smaller altitudinal ranges) are at the highest risk from climate change. In contrast, some widespread species with good dispersal mechanisms may not suffer such a fate because they may be able to disperse from one mountain to a higher one.
UNCERTAINTY AND GENERALITY
Because the total number of New Zealand alpine vascular plant species after warming could be above 600 (Scenarios A and B), there may be no extinction pressure from reductions in area per se for the ∼600 existing indigenous species. In other words, because the alpine flora is below saturation, it could conceivably tolerate considerable compression without extinctions. However, regardless of area, the main pressures in those scenarios would be due to invading species, total loss of endemics on alpine islands, and the relative rapidity of change requiring upward migration into areas where suitable soils may not be available.
The range of uncertainty of extinction predictions for alpine vascular plants at the scale of New Zealand should not detract from the basic statistical generality and certainty of the message formalized in species-area models: if climatic warming leads to species of taller vegetation establishing at higher elevations on mountains, the consequent reductions in available habitat area for alpine species (often to zero) may lead to the extinction of a considerable number of species unless countermeasures are taken.
Given the generality of the species-area relationship, other large biotic groups, such as insects, could suffer similar proportional degrees of threat. However, these generalizations may not hold for smaller taxonomic groups where specific taxonomic-ecological-scaling relations may apply, potentially leading to higher or lower risks.
Other vegetation types and ecosystems at lower altitudes will also be forced upward by rising temperatures, reducing their available area, with similar degrees of threat applicable in accordance with Figures 4 and 5. At sea level, the areas vacated by these rising ecosystem boundaries is likely to be filled by generalist, high-dispersal species, both indigenous and exotic, with a high degree of plasticity.
Conclusion
Long-term monitoring sites in New Zealand were mostly set up to follow recovery from grazing and burning within representative sites (Dickinson et al., 1992). Even in relatively undisturbed sites, workers have concluded that large-scale, occasional disturbances were overriding climate-change signals in relation to treeline dynamics (Cullen et al., 2001). Monitoring to interpret climate-change signals in New Zealand will require focused research approaches and experimental designs that are aimed at distinguishing such signals from the background noise (e.g., Mark et al., 2000; Pauli et al., 2001). Such an approach has recently been initiated at the Mt. Burns site in Fiordland.
Species-area models suggest that a 3°C rise in temperature could lead to a loss of 33–50% of indigenous alpine vascular taxa in the longterm. Additional losses would occur from the submersion of alpine islands, with ∼80% of those potentially disappearing. At the same time, the resulting flora could be enriched by over 100 exotic vascular species. The resulting flora would include a much larger proportion of high-dispersal, genetically plastic species. Local endemics and low-dispersal and specialist species will be the first at-risk species to disappear.
As climate changes, rainfall, wind patterns, snow-lie, human disturbance, and invasive species behavior will also change in interdependent ways. Temperature rise is only one aspect of a multifaceted changing environment. Alpine plant species generally require open habitats and many have ruderal, colonizing abilities. Some are able to extend downward when forest cover below treeline is removed. Such species could actually enlarge their ranges due to extensive deforestation, irrespective of climate change. The scenarios presented here must be seen within these contexts. The value of these scenarios is that they are based on transparent assumptions and explore the sensitivity of the empirical datasets used to calculate the equation parameters.
Species-area, fragmentation, and metapopulation models can help suggest testable hypotheses and guide the design of large-scale monitoring projects as well as quantify potential threats particularly to smaller-scale mountains and low-dispersal endemic species. Overall, the present models suggest that maintaining all existing alpine species can be done only at a cost. Mitigating measures could include reducing the effect of competition from taller plants (trees and shrubs), as is already done today to maintain some areas of low-elevation induced tall-tussock grasslands (Calder et al., 1992), and the transfer of species trapped on mountaintops to higher summits in the same region, if available. Such decisions run into ethical dilemmas regarding how much society is willing to “engineer” species conservation as well as society's will to constrain the unequal resource distribution that drives the excessive production of greenhouse gases (e.g., Costanza et al., 1997).
This study also identifies important knowledge gaps. Our level of ignorance, even in a relatively well-studied country such as New Zealand, is sobering. Out of the 441 alpine islands counted here (which may be much less than the real number), how many have biological inventories? How many have even been visited? Without inventories, we have no knowledge of regional endemisms, and our taxonomic information is patchy at best. Accurate predictions of biotic impacts require this information.
Acknowledgments
We would like to thank all the people who contributed information on long-term monitoring sites, particularly Louise Cullen, Malcolm Douglas, Peter Espie, Alan Nordmeyer, and Greg Byron, who also provided access and guidance for using the Land Information New Zealand database. Matt Scott summarized the data for Figure 3. We are grateful to Katharine Dickinson, Bill Lee, Peter Wardle, Tracy Williams, two anonymous referees and the journal editor for commenting on earlier versions of the text.
References Cited
FIGURE 2.
Compound topographic profile of New Zealand, representing the sum of areas in every 100-m contour interval of terrain (bold line). Present treeline and summer snowline are shown (median value for around 42°S), together with the equilibrium treeline at a +3°C scenario, to show the resultant reduction in land area
FIGURE 3.
Species richness of New Zealand alpine vascular plants in relation to elevation. The values are based on the 655 taxa (613 species) and their altitudinal ranges as described in Mark and Adams (1995)
FIGURE 4.
Projections of numbers of species of New Zealand alpine vascular plants (native and exotic) at equilibrium after a given temperature rise. Scenarios B (solid line), C (dashes), and E (dashes and dots) calculated from parameters in Table 1, as explained in Methods
FIGURE 5.
Predicted native species decline in relation to air-temperature increases given the species-area Scenarios B (solid line), C (dashes,) and E (dashes and dots) of Table 1. Leptinella dendyi flower (upper right) and Raoulia eximia cushion (lower left) as examples of alpine species
TABLE 1
Vascular plant species numbers and species-area constants for different scenarios obtained through different expirical data sets. A = 67 areas around the world; B = 12 islands up to the size of Madagascar; C = 30 alpine regions around the world; D = 6 alpine regions in New Zealand + total alpine New Zealand + total New Zealand; E = 6 alpine regions in New Zealand; F = 2 points: total alpine New Zealand + total New Zealand. Probabilities (t-test and F-test) for Scenarios A to D < 0.001, for E < 0.05, for F n/a (only 2 points)
TABLE 2
Number of alpine islands in New Zealand before and after a 3°C warming
