18. Spatial Distribution of Plants and Plant Communities

18.2.1 Traits and Vectors

Dispersal vectors may change over time depending on environmental conditions. Species that can use multiple dispersal vectors are considered “polychor”. Many plants do not use only one vector but possess different morphological adaptations to exploit several vectors. The ability to use several strategies improves a plant’s chances of finding an ideal site for establishment. In special cases, however, a “secondary” dispersal strategy could prove more important. For example, an African elephant eats its favourite fruit, containing seeds, which are passed later in the day, 20 km from where they were eaten (primary dispersal). The surviving seeds are then further dispersed by dung beetles, which transport the seeds over short distances before burying them (secondary dispersal) (Fig. 18.1c) (Engel 2000). The importance of “secondary” dispersal is often underestimated. The example of dung beetles shows that the most favourable microsites are often only found in the secondary transport phase. Secondary transport may also lead to a wider range of dispersal.

18.2.2 Effectiveness of Dispersal Mechanisms

Dispersal can be limited by physical barriers, such as large differences in altitude of mountainous regions, oceans and large arid regions. The absence of certain environmental conditions or requirements (i.e. temperature, water, light, soil nutrients, symbiotic patterns) may significantly limit successful establishment following dispersal. Keep in mind that these barriers or requirements can be manipulated in plants favoured by the human vector, where people can improve growing conditions by removing competition, improving soil conditions through fertilisation or aeration.

A clear understanding of how far propagules can disperse is especially important in the field of ecosystem restoration, which could provide key information about species’ ability to persist in fragmented habitats. As this fragmentation is quite common in human-altered landscapes, better knowledge about successful seed dispersal is indispensable. This is especially true for ballistic and other autochoric species, including ant- and bat-dispersed vectors, where dispersal distances travel is rarely beyond a few metres. In contrast, many vertebrates, wind and water are able to disperse propagules for distances exceeding 10 km. In some of the best-case dispersal distance scenarios, orchid seeds and spores from ferns have been known to travel distances of several hundred kilometres.

When flora becomes established on young islands or within post-disturbance landscapes, it will occur primarily by anemochoric species, unless a human vector facilitates its introduction. Once established, primary dispersal mechanisms tend to shift from primarily wind to water or birds, as illustrated in a fallow field (Fig. 18.2). During the first few years, wind-dispersed species become established predominantly via long-distance transport; however, as time passes, allochoric species become more dominant after the first two decades and, finally, typically within a century, replace wind-dispersed species (Sect. 17.​2). Other quantitative aspects of seed dispersal to consider include (1) the number of seeds produced and dispersed, (2) the number of visits of dispersers (e.g. birds), and (3) the continuity of favourable weather conditions (e.g. duration of air flow). The efficiency of seed dispersal also has qualitative aspects. The quality of a seed may determine whether or not a bird will decide to disperse it, how the bird manipulates it and, finally, how it is transported, all of which will affect the probability of seed germination.

The dispersal of propagules is usually deemed successful if they reach a site that is favourable for their germination and establishment and ensures growth till flowering and seed production (safe site). Following germination and establishment it is extremely important to the long-term survival of the species that it exploit its new site, which is at some distance from the parent plant, contributing to the expansion of the species range or area. An expanding range or shift provides a few benefits, especially if it occurs at fast enough rates. A fast shift can allow a plant species to escape certain pathogens and enhance the population’s chances of survival. To a large degree this depends on the available mechanisms for dispersal and their effectiveness regarding a targeted (suitable site for growth) and far-reaching (gain space) transport of propagules. However, such substrate “targets” are typically not achieved through a single form of dispersal. To ensure successful colonisation, a series of different dispersal vectors and their associated interactions can be used to improve seed dispersal (heterospory). Multiple dispersal mechanisms can supplement each other to ensure successful dispersal. Under unfavourable external conditions (i.e. a short vegetation period in high mountains or long dry periods), reproductive dispersal is combined with vegetative propagation. For example, the alpine species Polygonum viviparum produces both seeds and bulbils on the same stalk, with the bulbil fraction increasing with altitude.

These particular advantages associated with specific regions typically occur when the species adapts to the local conditions over time (local adaptation). For instance, wind-dispersed propagules in arid regions are often transported over long distances because only minimal obstacles occur over land. Another example of plant-dispersal local adaption is the production of flashy or tasty fruit-bearing plants, which increase the likelihood of attracting animal dispersal vectors. Essentially, different strategies have evolved in which successful arrival at safe sites and spatial expansion provide a niche balance. Studies on seed dispersal in forests have shown that many trees in the humid tropics with rainy and dry seasons have created morphologically adapted seeds that facilitate transportation by animals. In temperate forests this also applies to shrubs, but here seeds from trees are predominantly transported by wind. This change in the form of dispersal is explained by regular winds occurring in temperate latitudes. Also in dry areas, where almost no obstacles impede air currents, wind dispersal dominates. All this is to say that there is no concrete evidence for certain dispersal vectors being exclusively used or associated with specific biomes (Tables 18.1 and 18.2).

Several hypotheses and models have been developed to help explain the link between certain plant species and their vectors. The low investment model is associated with plants’ investing little per propagule produced, resulting in a large number of small seeds. This model favours dispersal vectors associated with long-distance dispersal (wind-dispersed, attached to animals). Because little is invested in each seed, the recruitment rate is quiet low; however, this can be offset by volume. This strategy is typically used by early successional species. The second model, the high investment model, is essentially the opposite, where plants produce very few, large seeds. The large seed size provides ample amounts of energy and nutrients to ensure good rates of germination, assuming it attracts the appropriate dispersing agent (bird or bat) and falls upon a favourable substrate. The seeds in this strategy are only distributed over short distances, but they usually find relatively safe conditions for germination. In animal-dispersed species, both strategies are used: seeds externally attached to animals (with low investment) and those transported in the gut (with high investment). The directed dispersal hypothesis first proposed by Howe and Smallwood (1982), regarding targeted, relatively safe dispersal, also aligns with the high investment model. This contrasts with the colonisation hypothesis, where opportunists exploit opportunities for rapid dispersal over large areas. The premise for the escape hypothesis is that the chances for establishment and germination depend on the low density of conspecifics, so the mother plant cannot be nearby (e.g. Janzen-Connell effect) (Sect. 19.​3).

18.2.3 Propagule Bank and Seedling Establishment

Seed dispersal is deemed successful only if the transported propagules germinate at a potential site or are incorporated into the soil (in which case it will become established at a later time). The stock of all dispersal units in the soil is called the propagule or seed bank, also a seed pool. Seeds may remain viable for several years, in some cases up to a few centuries, until more favourable conditions for development occur. Seeds from Medicago lupulina have been found to be viable in the soil after 26 years, significantly lower than the seed of Spergula arvensis, which has been shown to maintain viability up to 1700 years (Urbanska 1992). Many propagules reduce metabolic activity before germination. Germination-inhibiting chemical substances may prevent further development. Sometimes temperatures must fall below a certain minimum value (vernalisation) before the seedling can develop. Such periods are called dormancy and may be determined genetically or by external conditions. Many seeds are protected by a thick pericarp, which can be impermeable to water. Dormancy acts as an environmental checklist for the seed, where until all environmental condition (typically light and temperature) cues are checked off, the seed will remain in the soil until conditions are met. The amount of time a seed will remain viable in a seed bank will vary from species to species.

Baskin and Baskin (2001) identified the differences between dormancy types; the first, endogenous dormancy, is where the properties associated with the seed embryo inhibit germination (e.g. physiological inhibiting mechanisms or an undeveloped embryo), whereas exogenous dormancy is described as where the properties of the endosperm or any other tissues of the seed or fruit inhibit germination (e.g. physical, chemical or mechanical constraints on germination and embryo growth).

The period of dormancy is different for different types of propagules. In tropical rainforests, seeds of shade-tolerant trees will tend to germinate immediately if they are found within a suitable growing site. Seeds of shade-tolerant trees often remain dormant until favourable conditions emerge, which will then allow for germination to occur, for example, after a fallen tree opens a gap in the canopy. In dry regions with very variable wet periods, species differ considerably in dormancy, even within individuals of the same species. In general, there is a clear favourability for seeds to be non-dormant in wet tropical rain forests (60% of species are non-dormant), mirroring an environmental gradient of decreasing precipitation and increasing seasonality or uncertainty of rainfall events. Generally, it may be assumed that the dormancy of propagules serves to tune germination to growth conditions that will provide a suitability bridge for a cold winter or a dry season.

A seed bank is usually located in the uppermost 10 cm of soil. Seed banks are made up of two types: temporary and long-term. Temporary propagule banks consist of the accumulation of seeds that will germinate in the short-term (next year). Long-term propagule banks are particularly important for the regeneration of plant communities. Propagules in temporary banks have only limited ability to germinate—as for tropical shade trees—and the regeneration of forests after a severe disturbance is unlikely. This also applies to rainforests in southern Chile outside the tropics. A few years after clearing of the almost natural stands and reforestations of the area with Pinus radiata, the local seed banks contained hardly any propagules of indigenous species (Scherer and Deil 1997). This underlines the importance of seed banks for the protection of species and biotopes.

The phase of seedling establishment is particularly sensitive in the life cycle of plants. Pathogens, herbivores and competitors, but also climatic abnormalities, may lead to large losses. Some seedlings require the protection of neighbouring species (e.g. Arabis hirsuta, Primula viridis), but for others germination is reduced by neighbouring species (Plantago lanceolata, Sanguisorba minor). A third group of species was able to germinate under all experimental conditions (e.g. Medicago lupulina). Large, time-dependent fluctuations have also been observed. There are close relations between rates of germination and interannual fluctuations of climatic conditions, particularly of temperatures. Most species germinate more successfully at higher temperatures, while others germinate better at lower temperatures (Espigares and Peco 1993). In some species of Acer, seed germination can commence only if it experiences a certain number of cool, moist days (known as stratification) near the freezing point to break dormancy and allow for germination to ensue (Solarik et al. 2016).

While the molecular and biochemical bases for dormancy and germination are well known for certain model species under controlled environmental conditions, we still lack a clear understanding of some of the fundamental processes under natural conditions that inhibit or promote seed germination and establishment. The successful establishment of a plant population is only secured by a permanent input of propagules, even if edaphic and climatic conditions at the growing site are suitable, pollinators (when necessary) are present, and the species is able to successfully compete with co-occurring individuals and protect itself against pathogens.

18.2.4 Distribution Patterns

As a result of different dispersal mechanisms, plant species develop distinct spatial patterns of distribution. These patterns are not permanently fixed in space because they depend largely on abiotic and biotic factors, which continuously fluctuate over time (daily, weekly, monthly, seasonally and yearly). Among these factors especially those related to climate, soil and mechanical impacts (flooding, thunderstorms) and—in cultural landscapes—agrochemical and agro-technical influences are important. Accurately identifying the propagule distribution pattern will depend on the spatial scale, where variability from the local-scale small patches to large-scale landscapes must be considered. A several-square-kilometre area with small forest islands, grasslands and bogs will show different patterns when compared with dispersal at finer scales (cm², m²) in the same area, which typically will yield much different results. Today, a species’ distribution is typically assessed by the use of a grid system in analysis. Although results will depend on the size of the grid chosen for analysis, the spatial scale (e.g. “patch”, “local” or “landscape”) becomes essential when considering dispersal patterns.

Distribution patterns will typically fall under three main types: clumped, regular and random distributions. The clumped distribution is the most common type in natural vegetation, mainly because of a close relationship between abiotic and biotic resources associated with plant establishment and growth. As these resources often present a mosaic structure, species and plant communities often become clumped. The introductory photo of Part IV shows an example of clumped distribution in a tropical high-mountain belt with abiotic site factors changing on a large scale. On a smaller scale, spatially aggregated clumped distribution can be the result of a restricted seed shadow, where seed is either dispersed or animals fail to travel far before excreting or dropping the seed. Fangliang et al. (1997) found in the highly diverse tropical rainforests of Malaysia a 4:1 relation of clumped to random distributed tree species, which suggests a strong influence of animal-dispersed seeds in these systems.

A regular distribution occurs when the propagule is dispersed evenly over an area, where a comparable amount of seeds are found at long and short distances from the parent tree. This distribution pattern is rarely found in nature. The regular distribution can occur when the environmental conditions across the dispersal distance change in a regular way: the zonation of different plant community changes with a significant change in the landscape. Vegetation along a riverside is an example. The mechanical force of the water during the rainy season and the different water storage capacities of the sediments on the low fluvial terraces with different grain sizes can result in the formation of a regular sequence of different species and communities (Fig. 17.​25). Regular patterns are also sometimes observed in tropical dry woodlands (“leopard skin”) if the distance between plant individuals is the result of competition for water, the limiting resource. The introductory photo of Chap. 18 shows an example of regular vegetation patterns at the landscape scale, with an evergreen gallery riverine forest and a seasonally moist dry woodland. Most examples of regular patterns found in nature are actually artificial, for example, in fruit tree plantations or even crop cultivation (introductory photo in Chap. 17).

Although rare, regular spatial patterns are typically determined by animals and wind vectors, they are most important for random distributions. One precondition is that the plant species concerned must be generalists, that is, they have a wide resource-based spatial niche (area where species-specific requirements are met), and the spatial distribution of the environmental resources is rather homogeneous. Abiotic and biotic factors are thus spaced in an unpredictable way, as are the individuals of plant species.

18.3.1 Characterisation and Interpretation of Areas

There are various ways to map areas. Frequently, when considering species distributions, only borders are drawn without providing any additional information concerning the distribution and abundance of species within the area. This can be avoided by creating dot maps, which can provide species locations. The most popular way to represent species distributions is through the use of species range maps. However, these maps often show only a portion of the entire range and rarely contain any information about species abundance. Such maps also show the climatic differentiation of areas often better than those based on meteorological data. Regardless of their shortcomings, these maps are indispensable for applied tasks, for example, nature protection.

The species range or area is essentially the spatial distribution in which a biological taxon is found, that is, where the environment is typically favourable for that species. The range limits are typically controlled by a number of historical and ecological factors, for example, climate, interspecific (other species) and intraspecific (same species) competition, site quality, food resources, water and landscape. Isolating the contribution of each of these factors individually is extremely difficult, especially since they can be co-dependent on each other. One attempt at determining range limits has been done through the use of isotherms (average temperatures); however, plants rarely react to average conditions. For example, the eastern species range of beech (Fagus sylvatica) is constrained by the lack of precipitation; however, another constraint is the severe cold winter (> −30 °C), which is known to cause extensive damage to bud development. At the species’ southern range limit, the opposite occurs, where a combination of a lack of summer precipitation and high temperatures causes significant drought events. Finally, beech’s northern limit has been determined by a combination of prolonged winters and late-season frost events. However, it has been speculated that beech reached its northern limit in the north-west of the British Isles after the last glaciation. These observations have relied on the current distribution patterns of adult trees while ignoring environmental constraints on seed germination and seedling establishment, key processes in determining any long-term presence of a species. This will be of special importance in studies about climate-change-driven dynamics as species attempt to maintain their climatic niches by expanding their ranges in altitude and latitude. As decisive for alpine plant communities and their upper distribution boundary is summer frost resistance (Taschler and Neuner 2004).

Identifying the constraining factors influencing range limit is particularly important for olive trees in the Mediterranean region (Fig. 18.3). In the mountainous portion of these regions, prolonged frost, coupled with a shorter growing season, minimises biomass accumulation in the stem. In southern Europe, early summer is often too moist, inhibiting pollination and making fruit setting difficult. In North Africa, towards the south, increasing drought is limiting. At these limits, Walter’s law of the relative constancy of habitats can be applied to understand the occurrence of plants. In the southern species range of Quercus pubescens, for example, its occurrence indicates moist outposts of Mediterranean habitats, while at the northern range, in the upper Rhine Valley, it is found only on the warmest and driest, south-facing slopes. If the climate within the area of a plant species is changing in a particular direction, this species moves into a habitat that compensates for this climate change, so that the environmental conditions remain rather constant, assuming that the demands of a specific plant species remain constant (Walter 1986).

It should be mentioned that species range limits are dynamic, under constant flux, and can change quite readily owing to climatic factors influencing the biological system. The postglacial retreat of ice caused by climate warming led to significant species range shifts, which made it possible to make inferences on the rates of possible migration. For some important Central European tree species a range expansion of approximately 100 m/year was calculated, a considerable rate when one considers that trees are sessile organisms that typically require decades to reach reproductive maturity. Furthermore, quiet often many tree species will produce ample seed crops only once every several years (masting).

Considering the changes associated with climate change, species invasion, natural disturbances and new land-use practices, inventorying species ranges will continue. Complicating matters, the presence of synanthropic species or species that benefit from the presence of humans makes mapping species’ ranges and range limits extremely difficult. Ecologists have thus introduced the concept of a potential or fundamental range and a realised range. A potential range is an area where the range of environmental conditions is suitable for existence without the influence of competition or predation from interspecific species, whereas a realised range is the part of the potential range that is actually occupied by the species. A realised range is typically limited by various site conditions. Figure 18.4 shows schematically how differently the spatio-temporal formation of areas may occur. Areas may grow or shrink; they can also divide (disjunction), and new taxa may evolve by allopatric differentiation (metapopulations of the same species live in clearly separated geographical areas).

Species ranges are also characterised by their (1) size, (2) form and (3) geographical location. The differentiation between cosmopolitan and endemic species is based on the size of distribution areas. Cosmopolitan species typically occur over large areas, commonly extending over continents and different climatic zones. These species have extremely effective dispersal mechanisms, are highly competitive and can be categorised as phylogenetically old. Some common examples of cosmopolitan species include bracken (Pteridium aquilinum) and annual meadow grass (Poa annua). Cosmopolitans are differentiated from species that are generalists in terms of their site requirements (e.g. Pinus sylvestris).

Species areas are also differentiated by their form, particularly whether they are closed areas, that is, the species is established in a single, clearly delimited space, or whether the area consists of several partial areas and is thus disjunct. In closed areas, gaps between individual growing sites are so small that they may be bridged easily and quickly by the transport of propagules or pollen. In disjunct areas, this is no longer possible. It may be assumed that polyphyletic origin of the same species does not occur and therefore other explanations for the genesis of disjunct areas must be found. One explanation might be an extremely rare distribution event, for example, by migrating birds or through some anthropogenic influence. In many cases, however, it is known that present disjunct areas were once closed and were separated by tectonic events (e.g. continental drift), the formation of mountains or climatic changes (e.g. change in cold and warm periods). A good example of the latter are arcto-alpine species, which were widely distributed in Europe during the Pleistocene, growing in the lowland tundra habitats between the Scandinavian and alpine glaciers. Today, under a warmer period, they are restricted to high alpine and arctic habitats, as well as to isolated patches in lower mountain ranges in Central Europe (e.g. Dryas octopetala).

With the separation of once closed areas, so-called vicarious areas may develop. In these partial areas, the populations of once uniform species developed further in different ways. This would be called geographical vicariance with examples of the genera Kleinia and Aeonium (Fig. 18.5). The centres of diversity of these species in southern and eastern tropical Africa as well as in Macaronesia (including the Cape Verde islands) are separated by the North African dry regions. If subspecies or other species of a genus are established in the same space at different sites, this is called ecological vicariance; examples are the two alpine rhododendrons Rhododendron hirsutum and R. ferrugineum, respectively, on soils of limestone and silicate rocks.

The multitude of factors affecting a site shows that there is not a single factor that explains differences in vegetation. Most plants growing on siliceous substrates do not grow on limestone because of the soil chemistry (Fig. 18.6). Conversely, competition in the root layer stops the immigration of limestone species into siliceous sites, although they could grow there without competition. While 34 species of the siliceous matgrass meadows (Nardetum) are not able to colonise calcareous bluegrass meadows (Seslerietum) because of abiotic factors such as Fe deficiency, a large supply of phosphate or Al toxicity, it is mainly competition by roots of the matgrass species that stops invasion by bluegrass species. These abiotic factors, on the other hand, only affect nine species (Gigon 1987).

If phylogenetically related species are distributed closely together, it may be concluded that they have developed within the same region and that the area is a centre of diversity (genetic centre, central zone of related groups), or at least a maintenance centre of this genus. Plant breeders aim to find such centres for economically important plants because they hope to find important gene reserves.

18.3.2 Area Types-Floristic Elements-Plant Kingdoms

Areas of similar basic structures, size and geographical positions are categorised into area types. Various species do not occupy absolutely identical areas, but all species of the same area type are considered to be geo-elements if spatial aspects are considered more and floristic elements if floristic aspects are considered. Nevertheless, these terms may be regarded as synonyms.

The spatial distribution of area types is linked to climatic characteristics: (1) zonal temperature dependence with latitude, (2) dependence on the duration of the growing season with altitude above sea level, (3) dependence on the influence of oceanic or continental climates that affect temperature ranges. Groups of geo-elements for Europe are shown on the map in Fig. 18.7. The following abbreviations are generally used for the zonal sequence from north to south, which can be further subdivided based on clearly defined floristic contrasts between the observed areas (e.g. sub-Mediterranean).

Vegetation may thus be regarded as the combination of species of a certain area or geo-element. The causes of boundaries between areas and types remain poorly understood; however, their evaluation and interpretation can be based on ecological characteristics of the species. For example, Müller-Hohenstein (1988) characterised vegetation units on the Arabian peninsula using an area typological approach (Fig. 18.8). The figures show the different interpretations of the boundaries, which could then be used by climatologists and plant geographers. It is easy to see from this figure that there is a lot of variability between boundaries, suggesting that the north-eastern Arabian coastal regions should not be classified as tropical, according to neither the climate nor the vegetation.

Area types and the corresponding geo-elements are the basis of a hierarchical division of the Earth according to floral relationships. At the highest level of this hierarchy are the six plant kingdoms. This division is made on the best possible homogeneity within and the greatest contrast between neighbouring units. The differentiation is mainly based on the geographical positions in which these species have developed for different periods in isolation. To define the plant kingdoms, mainly higher taxonomic units (families) are used, which are based on not only their exclusive distribution but also the spatial centres of their distributions. Plant kingdoms can be further subdivided into floral regions and floral provinces (Fig. 18.9). The early separation of the southern continents with the corresponding formation of three plant kingdoms, Antarctic, Capensis and Australis, is particularly important. The sole plant kingdom in the northern hemisphere, which covers the middle and higher latitudes, is the Holarctic. In the lower latitudes of the tropics, the Palaeotropics and Neotropics are differentiated.

The Holarctic is the largest plant kingdom with distribution centres of many plant families (Apiaceae, Betulaceae, Brassicaceae, Caryophyllaceae, Fagaceae, Primulaceae, Ranunculaceae, Rosaceae, Salicaceae). Current differences, for example, between holarctic regions in North America and Eurasia, are caused by recent geological events (ice ages). However, in more recent geological time periods, no insurmountable barriers have arisen.

This does not apply to the southern, tropical regions. The separation of the African continent from South America led to the subdivision of two tropical plant kingdoms; this subdivision is justified, despite existing parallels (pantropical species and families, e.g. Annonaceae). Particularly characteristic families in the Palaeotropics, including the African continent and the South-East Asian archipelagos, are Combretaceae, Dipterocarpaceae, Euphorbiaceae, Moraceae (with over 1000 species of the genus Ficus), Nepenthaceae, Pandanaceae and Zingiberaceae. For the Neotropics, including most parts of Central and South America, species of Araceae, Bromeliaceae, Cactaceae and Solanaceae are particularly characteristic. Tropaeolaceae are entirely limited to this plant kingdom.

The Capensis in the south of the African continent is the smallest plant kingdom, but it is a particularly autonomous realm, where some families have developed into many species. This applies particularly to Ericaceae and Mesembryanthemaceae. In these two families relations to the Holarctic and Antarctic become obvious. The Bruniaceae is an endemic family of the cape, and representatives of the Proteaceae and Restionaceae are dominant.

The Southern Hemisphere plant kingdom of Australis comprises only the Australian continent and Tasmania; at the level of genera and species it is rich in endemic species. Eucalyptus species (Myrtaceae) are particularly important and are now distributed worldwide. The neighbouring New Zealand belongs in part to the Palaeotropics, but in the south partly to the Southern Hemisphere, the Antarctic, which has its largest area almost completely inhabited by plants. However, in the southern tip of South America, southern beeches (genus Nothofagus) developed.

18.4.1 Equilibrium Theory of Island Biogeography

Relations between the distribution and establishment of organisms and the size of areas have been analysed in particular for islands. Islands are clearly bound with a limited number of different habitats, often under relatively uniform climatic conditions. They occur in various sizes and are situated at different distances from the mainland. Therefore, they have been chosen as examples to clarify basic relations between number of species and size of area.

The first question is easily answered. Dispersal to islands situated far from the mainland, and thus from the closest source of propagules, can only be achieved by a form of long-distance dispersal. Transport by sea birds may be a method of dispersal (particularly epizoochoric), but also abiotic vectors, such as air and sea currents, or exceptional events, for example, tropical cyclones, must be considered. On islands in closer proximity to the mainland, birds remain important; however, at greater distances dispersal by wind and water is more prevalent. Propagules can be transported to islands from several directions. Almost half of all species on Hawaii originate from the Indo-Pacific region, more than a quarter from the Holarctic and 17% from cold temperate Southern Hemisphere areas (Fenner 1985). A simple equation was developed to explain how plants can become established on islands. Arrhenius (1921) showed that the relationship between the size of an island and the number of species it maintains can be described by

(18.1)

or

(18.2)

where S is the number of species of a taxon on the island, A denotes the area, z is a constant that changes little worldwide (i.e. the slope of the linear regression, when log S is plotted against log A, has values between 0.17 and 0.4), and c is a constant of proportionality and is dependent on the dimensions in which A was measured, in terms of the biogeographical area and taxonomic group.

The relationship between species and area is shown in Eqs. 18.1 and 18.2, where the number of species of a given group is halved if the area is reduced by a factor of 10. This attempt to provide a mathematically comprehensible theoretical expression relating the number of species to the size of islands is one of the important bases of the equilibrium theory of island biogeography by MacArthur and Wilson (1967). This has been confirmed, for instance, for plant species of the Scottish islands (Fig. 18.10). These relationships are explained by two hypotheses: (1) the habitat diversity hypothesis (Gorman 1979), which suggests that on larger islands there is a greater number of diverse habitats, and (2) the area-alone hypothesis (Kohn and Walsh 1994), which assumes a direct relationship between island size and number of species.

In addition, the researchers assumed that the number of species on islands depends on the diversity of habitat, and that if other islands exist between the island and the mainland, those will play an important role as a stepping stone. It was also concluded that islands possess a limited capacity for harbouring new species, which results in a sensitive equilibrium between the rate of colonisation and extinction. It should be noted that some islands may be “oceanic” islands (e.g. formed by volcanic eruptions) and as a result have never had direct contact with the mainland, or “continental” islands, where species contact has existed as a proportion of established species originating from some period of contact.

The course of colonisation and extinction is shown schematically in Fig. 18.11a, where the intersection of both curves shows the equilibrium state, where similar rates of colonisation and extinction lead to a constant number of species, that is, the loss of species is compensated by the arrival of new species. While the number of species at equilibrium is constant, the identity of species and species composition may change, that is, there is a turnover of species. The rate of colonisation (establishment on the island) is dependent on the distance from the mainland. The greater the distance, the more difficult it is for a species to become established there. The quality, type and quantity of the “source” of propagules that might arrive on the island must also be noted. The rate of extinction, however, is determined by the size of the island and will be much higher on small islands than on large ones because small islands can only support smaller population sizes, which have a higher risk of extinction. These conclusions have led to additional modification (Fig. 18.11b), where the following points may be deduced:

• Small islands have smaller numbers of species and higher turnover rates than large islands.

• Islands near the mainland have more species and a higher turnover rate than those further away.

• An island near the mainland returns to equilibrium after disturbance more quickly than one further away.

The equilibrium theory (or steady-state theory) was not generally well accepted at first. Some of the predictions of the model were verified empirically by Simberloff and Wilson (1969), who observed four islands off the coast of Florida following complete defaunation (sterilisation) and found re-establishment with the expected number of species and turnover rates in relation to size and position of the islands. Bush and Whittaker (1991) reconstructed rates of colonisation and extinction for spermatophytes on the volcano Rakata (Krakatau islands) on the basis of expedition reports and confirmed this theory as well (Fig. 18.12). On the other hand, different observations suggested that the extinction rate was dependent not only on the size of the island but also on the position of the island relative to the pool of propagules. The turnover of species on islands near the pool of propagules is rather low because of the continuous supply of propagules, which is of greater significance than the size of the island, which is known as the rescue effect, that is, a decreased rate of extinction due to recolonisation and immigration (Fig. 18.11c).

Criticism of the MacArthur and Wilson (1967) model is also directed at the type of predictive mathematical models used. Barkman (1990) believed that to describe complex functional relations in ecosystems, only descriptive models with limited validity should be considered. The characteristics of individual plant species and their interactions among each other have not been taken into account at all in these models. It is also assumed that the increased number of species is exclusive due to colonisation; genetic evolution is not considered, nor are the aspects of species saturation.

Lomolino and Weiser (2001) suggested that species richness may vary independently of island area, especially on small islands, and they termed this the small island effect. Responsible for a higher and more predictable speciation rate on larger islands is the internal geographical isolation of very different habitats such as large river basins and high mountain ranges, which are necessary for in situ speciation. Within-island speciation rates can exceed immigration as a source of species richness, at least on islands larger than 3000 km².

It has also been argued that phylogenetic diversification takes place on the same time scale as immigration and extinction. In this case, species richness on islands seems to be almost independent of all discussed factors. Finally, in a number of recent publications it has been highlighted more frequently that habitat heterogeneity, spatial scale and time scales (geological) are not sufficiently taken into consideration. Whittaker et al. (2008) proposed a general dynamic model (GDM) that provides explanations of biodiversity patterns by describing the relations between speciation, immigration and extinction, taking into account the life cycle of islands because climatic changes and geological tectonic events question the equilibrium model as well. In a proposed GDM, the most important variables considered remain—speciation, immigration and extinction—but are now combined with the evolutionary history of single islands and complete island archipelagos. The age of islands is also now taken into account; in particular, the geological history and the major tectonic events help to extend the thus far discussed conceptions of the equilibrium theory of island biogeography.

The theory of island biogeography also demonstrates that ecological equilibrium does not imply that ecosystems are constant and unchanging. The theory is based on a stochastic dynamic equilibrium, with a constant change in the rates of colonisation and extinction, which results in interannual variability for the actual number of species on islands and does not include speciation. Today it is becoming even more difficult to test empirically the simple assumptions of the theory of island biogeography simply due to the increasing anthropogenic influences. Although this theory has been criticised, it will be increasingly difficult to dispute, as empirical studies become increasingly harder to conduct, where the theory’s application can go beyond oceanic islands.

18.4.2 “Oceanic” and “Mainland” Islands

The results of island biogeography were formulated for islands and island groups surrounded by seawater and thus lack application for terrestrial island-like habitats. Nevertheless, attempts have been made to transfer this knowledge to the mainland and island-like habitats. These habitats are islands in lakes, mountain peaks in mountainous areas, “Inselberge” in the tropics and very small, well-isolated systems such as individual deciduous trees in a coniferous forest, caves or flower heads. Nowadays, all anthropogenic forests or biotope fragments in our managed landscapes could be interpreted to some extent as islands.

However, there are a number of commonalities to justify the application of island biogeography models to mainland islands. They have defined areas, relatively sharp borders with neighbouring habitats, are smaller than the surrounding area, and are often situated in a hostile surrounding area, at least for the taxa occurring within them. However, differences should not be overlooked, such as isolations (genetic separation) and separations (spatial separation). Usually, the distance to neighbouring islands (e.g. other forest fragments) is not very large, and the surrounding area may be hostile, but it allows at least some short-term bridging. Most of all, species turnover is faster because immigration rates are higher (e.g. from species avoiding intensive agriculture and fleeing into residual forests), as are emigration rates (e.g. because of the relative proximity of comparable neighbouring islands or sudden external disturbances).

The disappearance of a species in a habitat does not necessarily result in the extinction of the species at all. If the tree line in mountains gets lower because of a climate cooling, species will probably find refuge in the valley. Ultimately, the risk of extinction is low. With subsequent warming certain species will become established again in higher altitudes. The same principle can be applied to the diversity of species in regions that became mainland species during prolonged cold periods and then, owing to certain climatic constraints within the species, became islands again in warm periods. In this case the equilibrium theory was contrasted with the relict theory, which states that the present-day occurrence of species and communities is the result of changes in the past.

For intensively used landscapes with many small habitats a mosaic concept was developed (Duelli 1993). The number of species is explained as individual “stones” in a mosaic; within the habitat the number of species increases with the number of habitat types (even those created by humans). In a mosaic landscape there are several transitional stages (ecotones) that may be colonised by specialists. Among animals habitat diversity favours those that are dependent on a seasonal change of habitats. This concept underlines the importance of transitions and edges between neighbouring habitats.

For small forest islands, species-rich edge zones can be observed where light-demanding species occur, but not typical forest species. In near-natural ecosystems, edge zones often act as buffer zones that reduce the immigration of external species to the inside of the island, which provides the habitat for obligate forest species. In intensively used agricultural landscapes such buffer zones are often lacking and the environmental gradients to the island edges are steeper. The relationship between area and number of species may thus differ according to the proportion of the edge and the core zone. In small forest patches species composition is determined by the species of the edge zone. With increasing area, species density decreases and species typical of continuous forest zones appear. The highest diversity is achieved within this transition zone between the edge and the core. In special cases this apparently simple relation is complicated by the differences in the quality and range of influences within this zone. A distinction must also be made between natural (wind, radiation) and anthropogenic (emissions, fertilisers, mechanical disturbances) environmental factors.

The function of habitat fragment corridors or stepping stones between larger mainland islands should also be noted. These may stimulate connectivity between habitats by assisting movement across the landscape, that is, by allowing transient residence of taxa without providing a permanent habitat. At the same time they may act as refuges. It is assumed that species are able to be more successful in expanding their range when stepping stones are present. This has been empirically shown for the assisted movement of animal species in fragmented landscapes, for example, for birds (Fischer and Lindenmayer 2002), but evidence on plants remains unclear.

Plants must move within landscapes as vegetative parts (ramets), pollen or seeds, assisted by wind, water or animals (Sect. 18.​2.​1). Therefore, the characteristics of the spatial pattern within the mainland might be equally or even more important than that of the stepping stones or corridors. However, because plants are sessile organisms, they are much more limited in their ability to decide whether a habitat is “suitable” or “hostile”, where mobile species can respond more readily to gradients of resource availability. Thus, more recently, the simple corridor–matrix model of metapopulations and landscape ecology has been replaced by an integrative functional mosaic model, where the landscape is composed of patches of different movement and flow characteristics, which provides a much more natural interpretation of the relationships (Murphy and Lovett-Doust 2004).

There are examples showing that mainland islands serve as solid proof of the validity of species–area relationships according to the theory of island biogeography if the influences of humans on the island are not too significant (Fig. 18.13a). These were confirmed, for example, for island mountains on the Ivory Coast, where the number of species increases steadily with increasing area (Porembski et al. 1995). However, in the tropical rainforests of Malaysia, increasing areas led to a significant saturation of species (in the reported case with an area of around 50 ha) (Fig. 18.13b). This contradiction has far-reaching consequences, particularly since nowadays extinction rates of plant and animal species are on the rise. For example, owing to the increasing losses of tropical rainforests, species extinction rates are calculated on the basis of species–area curves. The course of the curves in the figures shown here therefore allows a (too) broad spread of extinction rates. The saturation rate in tropical rainforests for an area of about 50 ha also indicates that plant sociological research examining minimum areas is possible theoretically but not practically.

Because most of us today live in managed landscapes that typically have a high degree of fragmentation, it becomes more relevant to apply the theory of island biogeography to current problems, most prominently for questions surrounding protected areas. The relation of core to edge zones in small habitats becomes extremely important. With a spatial decrease in such habitats the diversity of conditions at a site decreases linearly, but the quality decreases exponentially (Mader 1983). Species that are locally adapted to specific site conditions may go locally extinct if abrupt changes in these environments occur. In contrast, more widely distributed species, which have broader habitat requirements, may suffer limited consequences in these edge environments, unless habitat is lost due to a large or frequent disturbance. Therefore, application of the theory of island biogeography is, at the moment, only limited. Akatov (2012) discusses in more detail the shortcomings of this theory, its possible uses and recommendations for practical planning of nature conservation projects.

In spite of these known shortcomings, conservation planning is still mainly based on island theory, although established relations between the size of an area and distance (for mainland islands, the distance to the next suitable habitat) have stimulated further discussions on the minimum size of areas that should be sustained. Of course, the size of a protected area should not be calculated according to the assumptions of island biogeography exclusively. Habitats of the same or similar quality should be maintained not too far away as (perhaps only intermittently required) refuge areas, in connection with the previously mentioned stepping stones. These stepping stones are an important component of the concept of biotope connectivity for protected areas. In our agriculturally managed landscapes, such stepping stones could be small forest islands (refuge) as well as hedges, edges of fields, long-term fallow land, abandoned stone quarries or railway lines. Concepts of nature protection must incorporate, along with the protected areas, buffer zones and smaller areas of habitat fragments around the protected area, and these areas should be maximised.

Preserving a given number of species should not be the sole aim of protection measures. Large areas of mosaics of optimal and suboptimal habitats should be the primary goal of such an area. As such, there is no one-size-fits-all approach when considering the size and spatial patterns for a protected area and stepping stones (particularly the distances between stepping stones and the protected area). Each of these aspects depends on the communities one is interested in protecting and can vary substantially. Very few empirical studies have been conducted to determine the minimum distances and areas that should be used for individual groups and organisms. Currently, urgent attempts to create protected areas are based on local conditions, the plausibility of plans and the availability of land. Increasingly, the more dynamic metapopulation concept and especially the functional mosaic model (Murphy and Lovett-Doust 2004) are becoming more important to help determine these issues of protected areas. The latter not only considers the dichotomy between unsuitable and suitable habitat patches but underlines the nature of the composite landscape mosaic as a key determinant of the fate of plant populations.

18.4.3 More Models of Island Biogeography Related to the Number of Species and Area

The theory of island biogeography was taken as a paradigm. Since the 1970s and 1980s various alternative interpretations have been advanced and discussed. Those who emphasise stochastic processes explain the establishment of plant species on islands differently to those who favour determinism. All theoretical considerations are based more or less on the same factors, which are regarded as decisive for the establishment of plants; however, they are interpreted differently. As the state of knowledge currently stands, the following factors have been identified as important in this theory: the pool of propagules on the mainland, mechanisms for dispersal, distance to the island and its size, habitat characteristics, phases of succession and the conditions of competition on the island.

Connor and Simberloff (1979) assumed in their simulation model that the establishment of plants occurs stochastically without competition. They tested this model with actual data and concluded that competition does not play a role. In their interpretation, it was the parameters of dispersal that determined the success of a coloniser. Gilpin and Diamond (1982), regarding competition as the decisive factor, criticised this interpretation. They mentioned that closely neighbouring islands possess different species compositions, although an exchange of species would be possible. However, this exchange does not take place because the appropriate species niches are already occupied, where invading species will not displace species already established. In this more empirical model the explanation of diversity on islands is not possible without considering competition, so the species first present will be a contributing factor to succession and the plant community.

These and other models use only those parameters that are considered important. Often, the dimension of the temporal scale has not been considered. Time has been shown to be an important component for the establishment of different species, as pointed out by the models developed by Grime (1979) and Tilman (1988, Sect. 17.​3). For the early establishment phase, a stochastic interpretation is more relevant, whereas for the later phases, a deterministic one is possible.

Models developed by Whittaker and Jones (1994), which criticise purely stochastic models, show how much the immigration of species depends on the dispersing vector and on the stage of succession, as exemplified by their data from the island of Krakatau during the course of colonisation over 100 years (Fig. 18.14). They found that during the initial phase, colonising species reaching the island became established on undeveloped substrates. During the second phase, anemochoric species became more important. However, it was only during the third phase, after substrates and the first succession stages of vegetation had developed, that zoochory played an increasingly important role. These results also suggest an initial stochastic but later deterministic establishment of species. It is overlooked in many interpretations that the turnover of species is not an inherent feature of a system but is largely regulated by changes in the environment. In this analysis and calculations, a minimum turnover rate is considered, but actual speciation rates are not available.