CHAPTER 5 Habitat Fragmentation and Landscape Change

Home , Habitat fragmentation, Landscape ecology, Relict (biology)

Sodhi and Ehrlich: Conservation Biology for All. http://ukcatalogue.oup.com/product/9780199554249.do

CHAPTER 5 Habitat fragmentation and landscape change

Andrew F. Bennett and Denis A. Saunders

Broad-scale destruction and fragmentation of fragmentation” for issues directly associated with native vegetation is a highly visible result the subdivision of vegetation and its ecological of human land-use throughout the world (Chap- consequences. ter 4). From the Atlantic Forests of South America This chapter begins by summarizing the con- to the tropical forests of Southeast Asia, and in ceptual approaches used to understand conserva- many other regions on Earth, much of the original tion in fragmented landscapes. We then examine vegetation now remains only as fragments the biophysical aspects of landscape change, and amidst expanses of land committed to feeding how such change affects species and commu- and housing human beings. Destruction and nities, posing two main questions: (i) what are fragmentation of habitats are major factors in the implications for the patterns of occurrence of the global decline of populations and species species and communities?; and (ii) how does (Chapter 10), the modification of native plant and landscape change affect processes that influence animal communities and the alteration of ecosys- the distribution and viability of species and com- tem processes (Chapter 3). Dealing with these munities? The chapter concludes by identifying changes is among the greatest challenges facing the kinds of actions that will enhance the conser- the “mission-orientated crisis discipline” of conservation of biota in fragmented landscapes. vation biology (Soulé 1986; see Chapter 1). Habitat fragmentation, by definition, is the “breaking apart” of continuous habitat, such as 5.1 Understanding the effects tropical forest or semi-arid shrubland, into dis- of landscape change tinct pieces. When this occurs, three interrelated 5.1.1 Conceptual approaches processes take place: a reduction in the total amount of the original vegetation (i.e. habitat The theory of island biogeography (MacArthur loss); subdivision of the remaining vegetation and Wilson 1967) had a seminal influence in sti- into fragments, remnants or patches (i.e. habitat mulating ecological and conservation interest in fragmentation); and introduction of new forms of fragmented landscapes. This simple, elegant land-use to replace vegetation that is lost. These model highlighted the relationship between the three processes are closely intertwined such that number of species on an island and the island’s it is often difficult to separate the relative effect of area and isolation. It predicted that species rich- each on the species or community of concern. ness on an island represents a dynamic balance Indeed, many studies have not distinguished be- between the rate of colonization of new species to tween these components, leading to concerns that the island and the rate of extinction of species “habitat fragmentation” is an ambiguous, or even already present. It was quickly perceived that meaningless, concept (Lindenmayer and Fischer habitat isolates, such as forest fragments, could 2006). Consequently, we use “landscape change” also be considered as “islands” in a “sea” of de- to refer to these combined processes and “habitat veloped land and that this theory provided a

HABITAT FRAGMENTATION AND LANDSCAPE CHANGE 89 quantitative approach for studying their biota. live fences, and pastures with dispersed trees This stimulated many studies in which species each support diverse assemblages of birds, bats, richness in fragments was related to the area dung beetles and butterflies (Harvey et al. 2006). and isolation of the fragment, the primary factors To these species, the landscape represents a mo- in island biogeographic theory. saic of land uses of differing quality, rather than a The development of landscape ecology contrib- contrast between “habitat” and “non-habitat”. uted new ways of thinking about habitat Recognizing landscapes as mosaics emphasizes fragments and landscape change. The concept the need to appreciate all types of elements in of patches and connecting corridors set within the landscape. This perspective is particularly rel- a matrix (i.e. the background ecosystem or evant in regions where cultural habitats, derived land-use type) became an influential paradigm from centuries of human land-use, have impor- (Forman and Godron 1986). It recognized the tant conservation values. importance of the spatial context of fragments. Different species have different ecological The environment surrounding fragments is great- attributes, such as their scale of movement, life- ly modified during landscape changes associated history stages, longevity, and what constitutes with fragmentation. Thus, in contrast to islands, habitat. These each influence how a species “per- fragments and their biota are strongly influenced ceives” a landscape, as well as its ability to by physical and biological processes in the wider survive in a modified landscape. Consequently, landscape, and the isolation of fragments de- the same landscape may be perceived by pends not only on their distance from a similar different taxa as having a different structure and habitat but also on their position in the landscape, different suitability, and quite differently from the types of surrounding land-uses and how they the way that humans describe the landscape. influence the movements of organisms (Saunders A “species-centered” view of a landscape can be et al. 1991; Ricketts 2001). obtained by mapping contours of habitat suitabil- The influence of physical processes and distur- ity for any given species (Fischer et al. 2004). bance regimes on fragments means that following habitat destruction and fragmentation, habitat 5.1.2 Fragment vs landscape perspective modification also occurs. Mcintyre and Hobbs (1999) incorporated this complexity into a con- Habitat fragmentation is a landscape-level pro- ceptual model by outlining four stages along a cess. Fragmented landscapes differ in the size trajectory of landscape change. These were: and shape of fragments and in their spatial con- (i) intact landscapes, in which most original veg- figuration. Most “habitat fragmentation” studies etation remains with little or no modification; have been undertaken at the fragment level, with (ii) variegated landscapes, dominated by the orig- individual fragments as the unit of study. How- inal vegetation, but with marked gradients of ever, to draw inferences about the consequences habitat modification; (iii) fragmented landscapes, of landscape change and habitat fragmentation, it in which fragments are a minor component in is necessary to compare “whole” landscapes a landscape dominated by other land uses; and that differ in their patterns of fragmentation (iv) relict landscapes with little (<10%) cover of (McGarigal and Cushman 2002). Comparisons original vegetation, set within highly modified of landscapes are also important because: (i) land- surroundings. This framework emphasizes the scapes have properties that differ from those dynamics of landscape change. Different stages of fragments (Figure 5.1); (ii) many species along the trajectory pose different kinds of chal- move between and use multiple patches in the lenges for conservation management. landscape; and (iii) conservation managers must Many species are not confined solely to frag- manage entire landscapes (not just individual ments, but also occur in other land uses in mod- fragments) and therefore require an understand- ified landscapes. In Nicaragua, for example, ing of the desirable properties of whole land- riparian forests, secondary forests, forest fallows, scapes. Consequently, it is valuable to consider

© Oxford University Press 2010. All rights reserved. For permissions please email: [email protected] Sodhi and Ehrlich: Conservation Biology for All. http://ukcatalogue.oup.com/product/9780199554249.do

90 CONSERVATION BIOLOGY FOR ALL

a) b) many of these are intercorrelated, especially with the total amount of habitat remaining in the landscape (Fahrig 2003). Several aspects of the spatial configuration of fragments that usefully distinguish between different landscapes include: (i) the degree of subdivision (i.e. number of fragments), (ii) the aggregation of habitat, and (iii) the complexity of fragment shapes (Figure 5.3). Some kinds of changes are not necessarily evi- dent from a time-series sequence. Landscape change is not random: rather, disproportionate Individual fragments Whole landscapes size compositional gradients change occurs in certain areas. Clearing of vegeta- shape diversity of land-uses tion is more common in flatter areas at lower eleva- core area number of fragments tions and on the more-productive soils. Such areas vegetation type aggregation are likely to retain fewer, smaller fragments of orig- isolation structural connectivity inal vegetation, whereas larger fragments are more Figure 5.1 Comparison of the types of attributes of a) individual likely to persist in areas less suitable for agricultural fragments and b) whole landscapes. orurbandevelopment,suchasonsteepslopes, poorer soils, or regularly inundated floodplains. the consequences of landscape change at both the This has important implications for conservation fragment and landscape levels. because sites associated with different soil types and elevations typically support different sets of species. Thus, fragments usually represent a biased sample of the former biota of a region. There also is 5.2 Biophysical aspects of landscape a strong historical influence on landscape change change because many fragments, and the disturbance re- 5.2.1 Change in landscape pattern gimes they experience, are a legacy of past land settlement and land-use (Lunt and Spooner 2005). Landscape change is a dynamic process. A series of Land-use history can be an effective predictor of the “snapshots” at intervals through time (Figure 5.2) present distribution of fragments and ecosystem illustrates the pattern of change to the original condition within fragments. It is necessary to un- vegetation. Characteristic changes along a time derstand ecological processes and changes in the trajectory include: (i) a decline in the total area of past in order to manage for the future. fragments; (ii) a decrease in the size of many fragments (large tracts become scarce, small fragments 5.2.2 Changes to ecosystem processes predominate); (iii) increased isolation of fragments from similar habitat; and (iv) the shapes of frag- Removal of large tracts of native vegetation ments increasingly become dominated by straight changes physical processes, such as those relating edges compared with the curvilinear boundaries of to solar radiation and the fluxes of wind and water natural features such as rivers. For small fragments (Saunders et al. 1991). The greatest impact on frag- and linear features such as fencerows, roadside ments occurs at their boundaries; small remnants vegetation, and riparian strips, the ratio of perime- and those with complex shapes experience the ter length to area is high, resulting in a large pro- strongest “edge effects”. For example, the micro- portion of “edge” habitat. An increase in the climate at a forest edge adjacent to cleared land overall proportion of edge habitat is a highly influ- differs from that of the forest interior in attributes ential consequence of habitat fragmentation. such as incident light, humidity, ground and air At the landscape level, a variety of indices have temperature, and wind speed. In turn, these phys- been developed to quantify spatial patterns, but ical changes affect biological processes such as

HABITAT FRAGMENTATION AND LANDSCAPE CHANGE 91

Figure 5.2 Changes in the extent and pattern of native vegetation in the Kellerberrin area, Western Australia, from 1920 to 1984, illustrating the process of habitat loss and fragmentation. Reprinted from Saunders et al. (1993). litter decomposition and nutrient cycling, and the of disturbance-adapted butterflies and beetles structure and composition of vegetation. and elevated tree mortality extend 200 m or Changes to biophysical processes from land use more from the forest edge (Laurance 2008). In in the surrounding environment, such as the use of most situations, changes at edges are generally fertilizers on farmland, alterations to drainage pat- detrimental to conservation values because they terns and water flows, and the presence of exotic modify formerly intact habitats. However, in plants and animals, also have spill-over effects in some circumstances edges are deliberately man- fragments. Many native vegetation communities aged to achieve specific outcomes. Manipula- are resistant to invasion by exotic plant species tion of edges is used to enhance the abundance unless they are disturbed. Grazing by domestic of game species such as deer, pheasants and stock and altered nutrient levels can facilitate the grouse (see Box 1.1). In England, open linear invasion of exotic species of plants, which mark- “rides” in woods may be actively managed to edly alters the vegetation in fragments (Hobbs and increase incident light and early successional habi- Yates 2003) and habitats for animals. tat for butterflies and other wildlife (Ferris-Kaan The intensity of edge effects in fragments and 1995). the distance over which they act varies between Changes to biophysical processes frequently processes and between ecosystems. In tropical have profound effects for entire landscapes. In forests in the Brazilian Amazon, for example, highly fragmented landscapes in which most changes in soil moisture content, vapor pressure fragments are small or have linear shapes, there deficit, and the number of treefall gaps extend may be little interior habitat that is buffered from about 50 m into the forest, whereas the invasion edge effects. Changes that occur to individual

92 CONSERVATION BIOLOGY FOR ALL

fragments accumulate across the landscape. 5.3 Effects of landscape change on species Changes to biophysical processes such as hydro- Species show many kinds of responses to habitat logical regimes can also affect entire landscapes. fragmentation: some are advantaged and in- In the Western Australian wheatbelt (Figure 5.2), crease in abundance, while others decline and massive loss of native vegetation has resulted in a become locally extinct (see Chapter 10). Under- rise in the level of groundwater, bringing stored standing these diverse patterns, and the processes salt (NaCl) to the surface where it accumulates underlying them, is an essential foundation for and reduces agricultural productivity and trans- conservation. Those managing fragmented forms native vegetation (Hobbs 1993).

Box 5.1 Time lags and extinction debt in fragmented landscapes Andrew F. Bennett and Denis A. Saunders

Habitat destruction and fragmentation result 7 in immediately visible and striking changes to the pattern of habitat in the landscape. 6 However, the effects of these changes on the 5 biota take many years to be expressed: there is a time‐lag in experiencing the full 4 consequences of such habitat changes. Long‐ lived organisms such as trees may persist for 3 many decades before disappearing without Species richness 2 replacement; small local populations of animals gradually decline before being lost; and 1 ecological processes in fragments are sensitive ‐ 0 to long term changes in the surroundings. 1 10 100 1000 Conservation managers cannot assume that Area (ha) species currently present in fragmented landscapes will persist there. Many fragments Box 5.1 Figure A change in the species‐area relationship for and landscapes face impending extinctions, mammals in rainforest fragments in Queensland, Australia, between fi ‐ even though there may be no further change in 1986 ( lled circles) and 2006 (open circles) illustrates a time lag in the loss of species following fragmentation. Data from Laurance fragment size or the amount of habitat in the et al. (2008). landscape. We are still to pay the ‘extinction debt’ for the consequences of past actions. declined further (see Box 5.1 Figure), with most – Identifying the duration of time‐lags and declines in the smaller fragments. By 2006 07, forecasting the size of the extinction debt for one species, the lemuroid ringtail possum Hemibelideus lemuroides fragmented landscapes is difficult. The clearest ( ), was almost totally insights come from long‐term studies that absent from fragments and regrowth forests document changes in communities. For example, along streams and its abundance in these large nocturnal marsupials were surveyed in habitats was only 0.02% of that in intact forest et al. rainforest fragments in Queensland, Australia, in (Laurance 2008). 1986–87 and again 20 years later in 2006–07 (Laurance et al. 2008). At the time of the first REFERENCES surveys, when fragments had been isolated for 20–50 years, the fauna differed markedly from Laurance, W. F., Laurance, S. G., and Hilbert, D. W. (2008). that in extensive rainforest. Over the subsequent Long‐term dynamics of a fragmented 20 years, even further changes occurred. rainforest mammal assemblage. Conservation Biology, Notably, the species richness in fragments had 22, 1154–1164.

HABITAT FRAGMENTATION AND LANDSCAPE CHANGE 93

required for a single individual or breeding unit, or for a self-sustaining population. Some species persist in fragmented landscapes by incorporating multiple fragments in their territory or daily foraging movements. In England, the tawny owl (Strix aluco) occupies territories of about 26 ha (hectares) in large deciduous woods, but individuals also persist in highly fragmented areas by including several small woods in their territory (Redpath 1995). There is a cost, however: individuals using multiple woods have lower breeding success and there is a higher turnover of territories between years. Species that require different kinds of habitats to meet regular needs (e.g. for foraging and breeding) can be greatly disadvantaged if these habitats become isolated. Individuals may then experience difficulty in moving between different parts of the landscape to obtain their required resources. Amphibians that move between a breeding pond and other habitat, such as overwintering sites in forest, are an example. Other attributes (in addition to fragment size) that influence the occurrence of species include the type and quality of habitat, fragment Figure 5.3 Variation in the spatial configuration of habitat in shape, land use adjacent to the fragment, and landscapes with similar cover of native vegetation: a) subdivision (many the extent to which the wider landscape isolates versus few patches); b) aggregated vs dispersed habitat; and c) compact populations. In the Iberian region of Spain, for vs complex shapes. All landscapes have 20% cover (shaded). example, the relative abundance of the Eurasian badger (Meles meles) in large forest fragments is landscapes need to know which species are most vulnerable to these processes. 1.0 0.9 0.8 5.3.1 Patterns of species occurrence 0.7 in fragmented landscapes 0.6 0.5 Many studies have described the occurrence of 0.4 species in fragments of different sizes, shapes, 0.3 composition, land-use and context in the land- Frequency of occurrence 0.2 scape. For species that primarily depend on 0.1 fragmented habitat, particularly animals, frag- 0.0 ment size is a key influence on the likelihood of 2–5 6–10 11–20 21–50 51–100 >100 occurrence (Figure 5.4). As fragment size de- Woodland area (ha) creases, the frequency of occurrence declines Figure 5.4 Frequency of occurrence of the common dormouse and the species may be absent from many (Muscardinus avellanarius) in ancient semi‐natural woods in small fragments. Such absences may be because Herefordshire, England, in relation to increasing size‐class of woods. the fragment is smaller than the minimum area Data from Bright et al. (1994).

94 CONSERVATION BIOLOGY FOR ALL significantly influenced by habitat quality and consequence of habitat fragmentation, but arise forest cover in the wider landscape (Virgos from land uses typically associated with subdivi- 2001). In areas with less than 20% forest cover, sion. Populations may decline due to deaths of badger abundance in forests was most influenced individuals from the use of pesticides, insecti- by isolation (i.e. distance to a potential source cides or other chemicals; hunting by humans; area >10 000 ha), whereas in areas with 20–50% harvesting and removal of plants; and construc- cover, badgers were most influenced by the qual- tion of roads with ensuing road kills of animals. ity of habitat in the forest fragments. For example, in Amazonian forests, subsistence A key issue for conservation is the relative hunting by people compounds the effects of for- importance of habitat loss versus habitat frag- est fragmentation for large vertebrates such as the mentation (Fahrig 2003). That is, what is the rela- lowland tapir (Tapir terrestris) and white-lipped tive importance of how much habitat remains in peccary (Tayassu pecari), and contributes to their the landscape versus how fragmented it is? Studies local extinction (Peres 2001). of forest birds in landscapes in Canada and Aus- Commonly, populations are also affected by tralia suggest that habitat loss and habitat frag- factors such as logging, grazing by domestic mentation are both significant influences, stock, or altered disturbance regimes that modify although habitat loss generally is a stronger influ- the quality of habitats and affect population ence for a greater proportion of species (Trczinski growth. For example, in Kibale National Park, an et al. 1999; Radford and Bennett 2007). Important- isolated forest in Uganda, logging has resulted in ly, species respond to landscape pattern in differ- long-term reduction in the density of groups of the ent ways. In southern Australia, the main blue monkey (Cercopithecus mitza)inheavily influence for the eastern yellow robin (Eopsaltria logged areas: in contrast, populations of black australis) was the total amount of wooded cover and white colobus (Colobus guereza) are higher in in the landscape; for the grey shrike-thrush (Col- regrowth forests than in unlogged forest (Chap- luricincla harmonica) it was wooded cover togeth- man et al. 2000). Deterministic processes are partic- er with its configuration (favoring aggregated ularly important influences on the status of plant habitat); while for the musk lorikeet (Glossopsitta species in fragments (Hobbs and Yates 2003). concinna) the influential factor was not wooded cover, but the configuration of habitat and diver- Isolation sity of vegetation types (Radford and Bennett Isolation of populations is a fundamental conse- 2007). quence of habitat fragmentation: it affects local populations by restricting immigration and emigration. Isolation is influenced not only by the 5.3.2 Processes that affect species in fragmented distance between habitats but also by the effects landscapes ofhumanland-useontheabilityoforganismsto The size of any population is determined by the move (or for seeds and spores to be dispersed) balance between four parameters: births, deaths, through the landscape. Highways, railway lines, immigration, and emigration. Population size is and water channels impose barriers to move- increased by births and immigration of indivi- ment, while extensive croplands or urban devel- duals, while deaths and emigration of individuals opment create hostile environments for many reduce population size. In fragmented land- organisms to move through. Species differ in scapes, these population parameters are influ- sensitivity to isolation depending on their type enced by several categories of processes. of movement, scale of movement, whether they are nocturnal or diurnal, and their response to Deterministic processes landscape change. Populations of one species Many factors that affect populations in fragmen- may be highly isolated, while in the same land- ted landscapes are relatively predictable in their scape individuals of another species can move effect. These factors are not necessarily a direct freely.

HABITAT FRAGMENTATION AND LANDSCAPE CHANGE 95

Isolation affects several types of movements, small initial population size. A decline in genetic including: (i) regular movements of individuals diversity may make a population more vulnerable between parts of the landscape to obtain to recessive lethal alleles or to changing environ- different requirements (food, shelter, breeding mental conditions. sites); (ii) seasonal or migratory movements of Fluctuations in the environment, such as varia- species at regional, continental or inter-continen- tion· in rainfall and food sources, which affect birth tal scales; and (iii) dispersal movements (immi- and death rates in populations. gration, emigration) between fragments, which Small isolated populations are particularly vul- may supplement population numbers, increase nerable· to catastrophic events such as flood, fire, the exchange of genes, or assist recolonization if drought or hurricanes. A wildfire, for example, a local population has disappeared. In Western may eliminate a small local population whereas in Australia, dispersal movements of the blue- extensive habitats some individuals survive and breasted fairy-wren (Malurus pulcherrimus) are provide a source for recolonization. affected by the isolation of fragments (Brooker and Brooker 2002). There is greater mortality of 5.3.3 Metapopulations and the conservation individuals during dispersal in poorly connected of subdivided populations areas than in well-connected areas, with this difference in survival during dispersal being a key Small populations are vulnerable to local extinc- factor determining the persistence of the species tion, but a species has a greater likelihood of in local areas. persistence where there are a number of local For many organisms, detrimental effects of populations interconnected by occasional move- isolation are reduced, at least in part, by habitat ments of individuals among them. Such a set components that enhance connectivity in the of subdivided populations is often termed a “me- landscape (Saunders and Hobbs 1991; Bennett tapopulation” (Hanski 1999). Two main kinds of 1999). These include continuous “corridors” or metapopulation have been described (Figure 5.5). “stepping stones” of habitat that assist move- A mainland-island model is where a large main- ments (Haddad et al. 2003), or human land-uses land population (such as a conservation reserve) (such as coffee-plantations, scattered trees in pas- provides a source of emigrants that disperse ture) that may be relatively benign environments to nearby small populations. The mainland pop- for many species (Daily et al. 2003). In tropical ulation has a low likelihood of extinction, where- regions, one of the strongest influences on the as the small populations become extinct relatively persistence of species in forest fragments is their frequently. Emigration from the mainland ability to live in, or move through, modified supplements the small populations, introduces “countryside” habitats (Gascon et al. 1999; Seker- new genetic material and allows recolonization cioglu et al. 2002). should local extinction occur. A second kind of metapopulation is where the set of interacting populations are relatively similar in size and all Stochastic processes have a likelihood of experiencing extinction (Fig- When populations become small and isolated, ure 5.5b). Although colonization and extinction they become vulnerable to a number of stochastic may occur regularly, the overall population per- (or chance) processes that may pose little threat to sists through time. larger populations. Stochastic processes include The silver-spotted skipper (Hesperia comma), a the following. rare butterfly in the UK, appears to function as Stochastic variation in demographic parameters a metapopulation (Hill et al. 1996). In 1982, but- ·such as birth rate, death rate and the sex ratio of terflies occupied 48 of 69 patches of suitable offspring. grassland on the North Downs, Surrey. Over the Loss of genetic variation, which may occur due to next 9 years, 12 patches were colonized and seven ·inbreeding, genetic drift, or a founder effect from a populations went extinct. Those more susceptible

96 CONSERVATION BIOLOGY FOR ALL

Figure 5.5 Diagrammatic representation of two main types of metapopulation models: a) a mainland‐island metapopulation and b) metapopulation with similar‐sized populations. Habitats occupied by a species are shaded, unoccupied habitat fragments are clear, and the arrows indicate typical movements. Reprinted from Bennett (1999).

to extinction were small isolated populations, 5.4 Effects of landscape change whereas the patches more likely to be colonized on communities were relatively large and close to other large 5.4.1 Patterns of community structure occupied patches. in fragmented landscapes The conservation management of patchily- distributed species is likely to be more effective by For many taxa—birds, butterflies, rodents, rep- taking a metapopulation approach than by focus- tiles, vascular plants, and more—species richness ing on individual populations. However, “real in habitat fragments is positively correlated with world” populations differ from theoretical models. fragment size. This is widely known as the Factors such as the quality of habitat patches and species-area relationship (Figure 5.6a). Thus, the nature of the land mosaic through which move- when habitats are fragmented into smaller pieces, ments occur are seldom considered in theoretical species are lost; and the likely extent of this loss models, which emphasize spatial attributes (patch can be predicted from the species-area relation- area, isolation). For example, in a metapopulation ship. Further, species richness in a fragment typi- of the Bay checkerspot butterfly(Euphydryas editha cally is less than in an area of similar size within bayensis) in California, USA, populations in topo- continuous habitat, evidence that the fragmenta- graphically heterogeneous fragments were less tion process itself is a cause of local extinction. likely to go extinct than those that were in topo- However, the species-area relationship does not graphically uniform ones. The heterogeneity reveal which particular species will be lost. provided some areas of suitable topoclimate each Three explanations given for the species-area year over a wide range of local climates (Ehrlich relationship (Connor and McCoy 1979) are that and Hanski 2004). small areas: (i) have a lower diversity of habitats; There also is much variation in the structure (ii) support smaller population sizes and therefore of subdivided populations depending on the fewer species can maintain viable populations; frequency of movements between them. At and (iii) represent a smaller sample of the original one end of a gradient is a dysfunctional meta- habitat and so by chance are likely to have fewer population where little or no movement oc- species than a larger sample. While it is difficult to curs; while at the other extreme, movements distinguish between these mechanisms, the mes- aresofrequentthatitisessentiallyasingle sage is clear: when habitats are fragmented into patchy population. smaller pieces, species are lost.

HABITAT FRAGMENTATION AND LANDSCAPE CHANGE 97

50 specialized ecological requirements are those lost from communities in fragments. In several

40 tropical regions, birds that follow trails of army ants and feed on insects flushed by the ants include specialized ant-following species and 30 others that forage opportunistically in this way. In rainforest in Kenya, comparisons of flocks of 20 ant-following birds between a main forest and Number of species forest fragments revealed marked differences 10 (Peters et al. 2008). The species richness and number of individuals in ant-following flocks were 0 0 10 20 30 40 50 60 lower in fragments, and the composition of flocks Area (ha) more variable in small fragments and degraded forest, than in the main forest. This was a conse- 60 quence of a strong decline in abundance of five species of specialized ant-followers in fragments, 50 whereas the many opportunistic followers (51 species) were little affected by fragmentation 40 (Peters et al. 2008). 30 The way in which fragments are managed is a particularly important influence on the composi- 20 tion of plant communities. In eastern Australia, Number of species for example, grassy woodlands dominated by 10 white box (Eucalyptus albens) formerly covered 0 several million hectares, but now occur as small 0 10 20 30 40 50 60 fragments surrounded by cropland or agricultur- Tree cover (%) al pastures. The species richness of native under- Figure 5.6 Species‐area relationships for forest birds: a) in forest story plants increases with fragment size, as fragments of different sizes in eastern Victoria, Australia (data from Loyn expected, but tree clearing and grazing by domes- 2 1997); b) in 24 landscapes (each 100 km ) with differing extent of tic stock are also strong influences (Prober and remnant wooded vegetation, in central Victoria, Australia (data from Radford et al. 2005). The piecewise regression highlights a threshold Thiele 1995). The history of stock grazing has the fl fl response of species richness to total extent of wooded cover. strongest in uence on the oristic composition in woodland fragments: grazed sites have a greater invasion by weeds and a more depauperate na- Factors other than area, such as the spatial and tive flora. temporal isolation of fragments, land management The composition of animal communities in or habitat quality may also be significant predic- fragments commonly shows systematic changes tors of the richness of communities in fragments. in relation to fragment size. Species-poor commu- In Tanzania, for example, the number of forest- nities in small fragments usually support a subset understory bird species in forest fragments (from of the species present in larger, richer fragments 0.1 to 30 ha in size) was strongly related to frag- (Table 5.1). That is, there is a relatively predict- ment size, as predicted by the species-area rela- able change in composition with species tionship (Newmark 1991). After taking fragment “dropping out” in an ordered sequence in succes- size into account, further variation in species rich- sively smaller fragments (Patterson and Atmar ness was explained by the isolation distance of 1986). Typically, rare and less common species each fragment from a large source area of forest. occur in larger fragments, whereas those present Species show differential vulnerability to in smaller fragments are mainly widespread and fragmentation. Frequently, species with more- common. This kind of “nested subset” pattern

98 CONSERVATION BIOLOGY FOR ALL

Table 5.1 A diagrammatic example of a nested subset pattern of of communities. The loss of a species or a change distribution of species (A–J) within habitat fragments (1–9). in its abundance, particularly for species that in- Species Fragments teract with many others, can have a marked effect on ecological processes throughout fragmented 123456789 landscapes. A ++++ ++++ Changes to predator-prey relationships, for ex- B +++ ++++ C + +++++ + ample, have been revealed by studies of the level of D +++++ + predationonbirds’ nests in fragmented landscapes E ++++++ (Wilcove 1985). An increase in the amount of forest F ++ edge, a direct consequence of fragmentation, in- G +++ creases the opportunity for generalist predators H ++ fi I +++ associatedwithedgesormodied land-uses to J + prey on birds that nest in forest fragments. In Swe- den, elevated levels of nest predation (on artificial eggs in experimental nests) were recorded in agricultural land and at forest edges compared with has been widely observed: for example, in butter- the interior of forests (Andrén and Angelstam fly communities in fragments of lowland rainfor- 1988). Approximately 45% of nests at the forest est in Borneo (Benedick et al. 2006). edge were preyed upon compared with less than At the landscape level, species richness has fre- 10% at distances >200 m into the forest. At the quently been correlated with heterogeneity in the landscape scale, nest predation occurred at a great- landscape. This relationship is particularly rele- er rate in agricultural and fragmented forest land- vant in regions, such as Europe, where human scapes than in largely forested landscapes (Andrén land-use has contributed to cultural habitats that 1992). The relative abundance of different corvid complement fragmented natural or semi-natural species, the main nest predators, varied in relation habitats. In the Madrid region of Spain, the overall to landscape composition. The hooded crow richness of assemblages of birds, amphibians, rep- (Corvus corone cornix) occurred in greatest abun- tiles and butterflies in 100 km2 landscapes is dance in heavily cleared landscapes and was pri- strongly correlated with the number of different marily responsible for the greater predation land-uses in the landscape (Atauri and De Lucio pressure recorded at forest edges. 2001). However, where the focus is on the com- Many mutualisms involve interactions be- munity associated with a particular habitat type tween plants and animals, such as occurs in the (e.g. rainforest butterflies) rather than the entire pollination of flowering plants by invertebrates, assemblage of that taxon, the strongest influence birds or mammals. A change in the occurrence or on richness is the total amount of habitat in the abundance of animal vectors, as a consequence of landscape. For example, the richness of woodland- fragmentation, can disrupt this process. For many dependent birds in fragmented landscapes in plant species, habitat fragmentation has a nega- southern Australia was most strongly influenced tive effect on reproductive success, measured in by the total extent of wooded cover in each 100 terms of seed or fruit production, although the km2 landscape, with a marked threshold around relative impact varies among species (Aguilar 10% cover below which species richness declined et al. 2006). Plants that are self-incompatible rapidly (Figure 5.6b) (Radford et al. 2005). (i.e. that depend on pollen transfer from another plant) are more susceptible to reduced reproduc- 5.4.2 Processes that affect community structure tive success than are self-compatible species. This difference is consistent with an expectation that Interactions between species, such as predation, pollination by animals will be less effective in competition, parasitism, and an array of mutual- small and isolated fragments. However, pollina- isms, have a profound influence on the structure tors are a diverse group and they respond to

HABITAT FRAGMENTATION AND LANDSCAPE CHANGE 99 fragmentation in a variety of ways (Hobbs and smaller fragments, a likely consequence of the Yates 2003). smaller population sizes of species and the great- Changes in ecological processes in fragments er vulnerability of such fragments to external dis- and throughout fragmented landscapes are com- turbances. For example, based on a sequence of plex and poorly understood. Disrupted interac- surveys of understory birds in tropical forest frag- tions between species may have flow-on effects ments at Manaus, Brazil, an estimate of the time to many other species at other trophic levels. How- taken for fragments to lose half their species was ever, the kinds of changes to species interactions approximately 5 years for 1 ha fragments, 8 years and ecological processes vary between ecosystems for 10 ha fragments, and 12 years for a 100 ha and regions because they depend on the particular fragment (Ferraz et al. 2003). sets of species that occur. In parts of North Ameri- Ecological processes within fragments also ex- ca, nest parasitism by the brown-headed cowbird perience ongoing changes in the years after isola- (Molothrus ater) has a marked effect on bird com- tion because of altered species interactions and munities in fragments (Brittingham and Temple incremental responses to biophysical changes. 1983); while in eastern Australia, bird commu- One example comes from small fragments of nities in small fragments may be greatly affected tropical dry forest that were isolated by rising by aggressive competition from the noisy miner water in a large hydroelectric impoundment (Manorina melanocephala)(Greyet al. 1997). Both of in Venezuela (Terborgh et al. 2001). On small these examples are idiosyncratic to their region. (< 1 ha) and medium (8–12 ha) fragments, isola- They illustrate the difficulty of generalizing the tion resulted in a loss of large predators typical of effects of habitat fragmentation, and highlight the extensive forest. Seed predators (small rodents) importance of understanding the consequences of and herbivores (howler monkeys Alouatta senicu- landscape change in relation to the environment, lus, iguanas Iguana iguana, and leaf-cutter ants) context and biota of a particular region. became hyperabundant in these fragments, with cascading effects on the vegetation. Compared with extensive forest, fragments experienced reduced recruitment of forest trees, changes in veg- 5.5 Temporal change in fragmented fi landscapes etation composition, and dramatically modi ed faunal communities, collectively termed an “eco- Habitat loss and fragmentation do not occur in a logical meltdown” (Terborgh et al. 2001). single event, but typically extend over many decades. Incremental changes occur year by year as remaining habitats are destroyed, reduced in size, 5.6 Conservation in fragmented or further fragmented (Figure 5.2). Landscapes landscapes are also modified through time as the human population increases, associated settlements ex- Conservation of biota in fragmented landscapes pand, and new forms of land use are introduced. is critical to the future success of biodiversity In addition to such changes in spatial pattern, conservation and to the well-being of humans. habitat fragmentation sets in motion ongoing National parks and dedicated conservation re- changes within fragments and in the interactions serves are of great value, but on their own are between fragments and their surroundings. too few, too small, and not sufficiently represen- When a fragment is first isolated, species richness tative to conserve all species. The future status of does not immediately fall to a level commensu- a large portion of Earth’s biota depends on rate with its long-term carrying capacity; rather, a how effectively plants and animals can be main- gradual loss of species occurs over time—termed tained in fragmented landscapes dominated by “species relaxation”. That is, there is a time-lag in agricultural and urban land-uses. Further, the experiencing the full effects of fragmentation persistence of many species of plants and animals (see Box 5.1). The rate of change is most rapid in in these landscapes is central to maintaining

100 CONSERVATION BIOLOGY FOR ALL ecosystem services that sustain food production, Provide speciﬁc habitat features required by par- clean water, and a sustainable living environment ·ticular species (e.g. tree hollows, rock crevices, for humans. Outlined below are six kinds of ac- “specimen” rainforest trees used by rainforest tions necessary for a strategic approach to conser- birds in agricultural countryside). vation in fragmented landscapes. 5.6.3 Manage across entire landscapes

Managing individual fragments is rarely effective 5.6.1 Protect and expand the amount of habitat because even well managed habitats can be de- Many indicators of conservation status, such as graded by land uses in the surrounding environ- population sizes, species richness, and the occur- ment. Further, many species use resources from rence of rare species, are positively correlated different parts of the landscape; and the pattern with the size of individual fragments or the total and composition of land uses affect the capacity amount of habitat in the landscape. Consequent- of species to move throughout the landscape. ly, activities that protect and expand natural and Two broad kinds of actions relating to the wider semi-natural habitats are critical priorities in landscape are required: maintaining plant and animal assemblages (see Manage specific issues that have degrading im- also Chapter 11). These include measures that: ·pacts across the boundaries of fragments, such as Prevent further destruction and fragmentation of pest plants or animals, soil erosion, sources of pollu- ·habitats. tion or nutrient addition, and human recreational Increase the size of existing fragments and the pressure. ·total amount of habitat in the landscape. Address issues that affect the physical environ- Increase the area specifically managed for conser- ·ment and composition of the land mosaic across ·vation. broad scales, such as altered hydrological regimes · Give priority to protecting large fragments. and the density of roads and other barriers. All fragments contribute to the overall amount and pattern of habitat in a landscape; consequent- 5.6.4 Increase landscape connectivity ly, incremental loss, even of small fragments, has Measures that enhance connectivity and create a wider impact. linked networks of habitat will benefit the conservation of biota in fragmented landscapes. Con- nectivity can be increased by providing specific 5.6.2 Enhance the quality of habitats linkages, such as continuous corridors or stepping stones, or by managing the entire mosaic to Measures that enhance the quality of existing allow movements of organisms. Actions that en- habitats and maintain or restore ecological pro- hance connectivity include: cesses are beneficial. Such management actions must be directed toward specific goals relevant · Protecting connecting features already present, to the ecosystems and biota of concern. These such as streamside vegetation, hedges and live include actions that: fences. Filling gaps in links or restoring missing connec- Control degrading processes, such as the inva- ·tions. ·sion of exotic plants and animals. Maintaining stepping-stone habitats for mobile Manage the extent and impact of harvesting nat- ·species (such as migratory species). fi ·ural resources (e.g. timber, rewood, bushmeat). Retaining broad habitat links between conserva- Maintain natural disturbance regimes and the ·tion reserves. ·conditions suitable for regeneration and establish- Developing regional and continental networks of ment of plants. ·habitat (see Boxes 5.2 and 5.3).

HABITAT FRAGMENTATION AND LANDSCAPE CHANGE 101

Box 5.2 Gondwana Link: a major landscape reconnection project Denis A. Saunders and Andrew F. Bennett

In many locations throughout the world, The southwest region of Australia is one of conservation organizations and community the world’s 34 biodiversity “hotspots”.Itis groups are working together to protect and particularly rich in endemic plant species. The restore habitats as ecological links between region has undergone massive changes over otherwise‐isolated areas. These actions are a the past 150 years as a result of development practical response to the threats posed by for broadscale agricultural cropping and habitat destruction and fragmentation and are raising of livestock. Over 70% of the area of undertaken at a range of scales, from local to native vegetation has been removed. The continental. Gondwana Link, in south‐western remaining native vegetation consists of Australia, is one such example of an ambitious thousands of fragments, most of which are less plan to restore ecological connectivity and than 100 ha. Many areas within the region enhance nature conservation across a large have less than 5–10% of their original geographic region. vegetation remaining. continues

Box 5.2 Figure Diagrammatic representation of the Gondwana Link in south‐west Western Australia. Shaded areas indicate remnant native vegetation. continues

102 CONSERVATION BIOLOGY FOR ALL

Box 5.2 (Continued)

This massive removal of native vegetation and Plate 6). It also involves protecting and has led to a series of changes to ecological managing the fragments of native vegetation processes, producing a wide range of problems that they are reconnecting. that must be addressed. Without some form of The groups believe that by increasing remedial action, over 6 million hectares of land connectivity and restoring key habitats they (30% of the region’s cleared land) will become will enable more mobile species that are salinized over the next 50 years, over 50% of dependent on native vegetation to move safely vegetation on nature reserves will be between isolated populations. This should destroyed, around 450 endemic species of plant reduce the localized extinctions of species from will become extinct, over half of all bird species isolated fragments of native vegetation that is from the region will be adversely affected, and happening at present. Gondwana Link should no potable surface water will be available in also allow species to move as climatic the region because of water pollution by salt. conditions change over time. The revegetation Addressing the detrimental ecological should also have an impact on the hydrological consequences involves the revegetation, with regime by decreasing the amount of water deep‐rooted trees and shrubs, of up to 40% of entering the ground water, and reduce the cleared land in the region. Gondwana Link is an quantity of sediment and pollution from ambitious conservation project involving agriculture entering the river and estuarine individuals, local, regional and national groups systems. addressing these detrimental ecological In addition to addressing environmental consequences. The objective of Gondwana Link issues the project is speeding up the is to restore ecological connectivity across development of new cultural and economic south‐western Australia. The aim is to provide ways for the region’s human population to ecological connections from the tall wet forests exist sustainably. of the southwest corner of the state to the dry woodland in the arid interior. This will involve protecting and replanting native vegetation Relevant website along a “living link” that stretches over 1000 km from the wettest corner of Western • Gondwana Link: http://gondwanalink.org/ Australia into the arid zone (see Box 5.2 Figure index.html.

Box 5.3 Rewilding Paul R. Ehrlich

Some conservation scientists believe that the facilitating the recovery of strongly interactive ultimate cure for habitat loss and species, including predators. Rewilding is the fragmentation that is now spreading like goal of the “Wildlands Network,” an effort led ecological smallpox over Earth is a radical form by Michael Soulé and Dave Foreman (Foreman of restoration, called rewilding in North 2004). The plan is to re-connect relatively America. The objective of rewilding is to restore undisturbed, but isolated areas of North resilience and biodiversity by re‐connecting America, into extensive networks in which severed habitats over large scales and by large mammals such as bears, mountain lions, continues

HABITAT FRAGMENTATION AND LANDSCAPE CHANGE 103

Box 5.3 (Continued)

wolves, elk, and even horses and elephants states more like those that prevailed before (which disappeared from North America only industrial society accelerated the 11 000 years ago) can roam free and resume transformation of the continent. Similar their important ecological roles in ecosystems rewilding projects on other continents are where conflict with humans would be minimal. now in the implementation stage—as in the Rewilding would restore landscape linkages— “Gondwana Link” in Australia (see Box 5.2). employing devices from vegetated overpasses The possible downsides to rewilding include over highways to broad habitat corridors— the spread of some diseases, invasive species, allowing the free movement of fauna and flora and fires and the social and economic and accommodation to climate change. The consequences of increased livestock cooperation of government agencies and depredation caused by large, keystone willing landowners would eventually create predators (as have accompanied wolf four continental scale wildways (formerly reintroduction programs) (Maehr et al. called MegaLinkages): 2001). Careful thought also is needed about thesizeoftheseWildways;tobesurethey Pacific Wildway: From southern Alaska are large enough for these species to again through the Coast Range of British Col- persist in their “old homes”. Nonetheless, it umbia, the Cascades, and the Sierra Ne- seems clear that such potential costs of vada to the high mountains of northern rewilding would be overwhelmed by the Baja California. ecological and economic-cultural benefits that Spine of the Continent Wildway: From well designed and monitored reintroductions the Brooks Range of Alaska through the could provide. Rocky Mountains to the uplands of West- ern Mexico. Atlantic Wildway: From the Canadian Maritime south, mostly through the REFERENCES AND SUGGESTED READING Appalachians to Okefenokee and the Everglades. Donlan, J. C., Berger, J., Bock, C. E., et al. (2006). Pleisto- fi Arctic-Boreal Wildway: Northern North cene rewilding: an optimistic agenda for twenty- rst century conservation. American Naturalist, , America from Alaska through the Canadi- 168 – an arctic/subarctic to Labrador with an ex- 660 681. tension into the Upper Great Lakes. Foreman, D. (2004). Rewilding North America: a vision for conservation in the 21st Century. Island Press, Many critical ecological processes are Washington, DC. mediated by larger animals and plants, and Maehr, D. S., Noss, R. F., and Larkin, J. L., eds (2001). Large the recovery, dispersal, and migration of these mammal restoration: ecological and sociological chal- keystone and foundation species (species that lenges in the 21st centuary. Island Press, Washington, are critical in maintaining the structure of DC. communities disproportionately more than Soulé, M. E. and Terborgh, J. (1999). Continental conser- their relative abundance) is essential if nature vation: scientific foundations of regional reserve net- is to adapt to stresses such a climate change works. Island Press, Washington, DC. and habitat loss caused by energy Soulé, M. E., Estes, J. A., Miller, B., and Honnold, D. L. development, sprawl, and the proliferation of (2005). Highly interactive species: conservation roads. Rewilding will help restore ecosystems policy, management, and ethics. BioScience, 55, in the Wildways to structural and functional 168–176.

104 CONSERVATION BIOLOGY FOR ALL

5.6.5 Plan for the long term profound influence on fragments and their biota, particularly at fragment edges. Landscape change is ongoing. Over the long- Different species have different ecological attri- term, incremental destruction and fragmenta- ·butes (such as scale of movement, life-history stages, tion of habitats have profound consequences what constitutes habitat) which influence how a for conservation. Long-term planning is re- species perceives a landscape and its ability to sur- quired to sustain present conservation values vive in modified landscapes. and prevent foreclosure of future options. Differences in the vulnerability of species to land- Actions include: ·scape change alter the structure of communities and Using current knowledge to forecast the likely modify interactions between species (e.g. pollina- ·consequences if ongoing landscape change occurs. tion, parasitism). Developing scenarios as a means to consider al- Changes within fragments, and between frag- ·ternative future options. ·ments and their surroundings, involve time-lags be- Developing a long-term vision, shared by the fore the full consequences of landscape change are ·wider community, of land use and conservation experienced. goals for a particular region. Conservation in fragmented landscapes can be ·enhanced by: protecting and increasing the amount of habitat, improving habitat quality, increasing 5.6.6 Learn from conservation actions connectivity, managing disturbance processes in the wider landscape, planning for the long term, Effective conservation in fragmented landscapes and learning from conservation actions undertaken. demands that we learn from current management in order to improve future actions. Several issues include: Suggested reading

Integrating management and research to more Forman, R. T. T. (1995). Land mosaics. The ecology of land- · fi effectively evaluate and re ne conservation mea- scapes and regions. Cambridge University Press, Cam- sures. bridge, UK. Monitoring the status of selected species Hobbs, R. J. and Yates, C. J. (2003). Turner Review No. 7. ·and ecological processes to evaluate the longer- Impacts of ecosystem fragmentation on plant popula- term outcomes and effectiveness of conservation tions: generalising the idiosyncratic. Australian Journal actions. of Botany, 51, 471–488. Laurance, W. F. and Bierregard, R. O., eds (1997). Tropical forest remnants: ecology, management, and conservation of Summary fragmented communities. University of Chicago Press, Chicago, Illinois. Destruction and fragmentation of habitats are · Lindenmayer, D. B. and Fischer, J. (2006). Habitat fragmen- major factors in the global decline of species, the tation and landscape change. An ecological and conservation fi modi cation of native plant and animal commu- synthesis. CSIRO Publishing, Melbourne, Australia. nities and the alteration of ecosystem processes. · Habitat destruction, habitat fragmentation (or subdivision) and new forms of land use are closely Relevant websites intertwined in an overall process of landscape change. Sustainable forest partnerships: http://sfp.cas.psu. Landscape change is not random: dispropor- · edu/fragmentation/fragmentation.html. · fl tionate change typically occurs in atter areas, Smithsonian National Zoological Park, Migratory at lower elevations and on more-productive soils. · Bird Center: http://nationalzoo.si.edu/Conservation Altered physical processes (e.g. wind and water AndScience/ MigratoryBirds/Research/Forest_ ·flows) and the impacts of human land-use have a Fragmentation/default.cfm.

HABITAT FRAGMENTATION AND LANDSCAPE CHANGE 105

United States Department of Agriculture, Forest ricultural landscapes of Costa Rica. Conservation Biology, · Service: http://nationalzoo.si.edu/Conservation 17, 1814–1826. AndScience/MigratoryBirds/Research/Forest_ Ehrlich, P. and Hanski, I., eds (2004). On the wings of check- Fragmentation/default.cfm. erspots: A model system for population biology. Oxford Uni- Mongabay: http://www.mongabay.com. versity Press, Oxford, UK. · Fahrig, L. (2003). Effects of habitat fragmentation on biodiversity. Annual Review of Ecology and Systematics, 34, 487– 515. REFERENCES Ferraz, G., Russell, G. J., Stouffer, P. C., Bierregaard, R. O., Pimm, S. L., and Lovejoy, T. E. (2003). Rates of species Aguilar, R., Ashworth, L., Galetto, L., and Aizen, M. A. loss from Amazonian forest fragments. Proceedings (2006). Plant reproductive susceptibility to habitat frag- of the National Academy of Sciences, United States of mentation: review and synthesis through a meta-analy- America, 100, 14069–14073. – sis. Ecology Letters, 9, 968 980. Ferris-Kaan, R. (1995). Management of linear habitats for Andrén, H. (1992). Corvid density and nest predation in wildlife in British forests. In D. A. Saunders, J. L. Craig, relation to forest fragmentation: a landscape perspective. and E. M. Mattiske, eds Nature conservation 4: the role of – Ecology, 73, 794 804. networks, pp. 67–77. Surrey Beatty and Sons, Chipping Andrén, H. and Angelstam, P. (1988). Elevated predation Norton, Australia. rates as an edge effect in habitat islands: experimental Fischer, J., Lindenmayer, D. B., and Fazey, I. (2004). Ap- – evidence. Ecology, 69, 544 547. preciating ecological complexity: habitat contours as a Atauri, J. A. and De Lucio, J. V. (2001). The role of land- conceptual landscape model. Conservation Biology, 18, scape structure in species richness distribution of birds, 1245–1253. amphibians, reptiles and lepidopterans in Mediterra- Forman, R. T. T. and Godron, M. (1986). Landscape ecology. – nean landscapes. Landscape Ecology, 16, 147 159. John Wiley and Sons, New York. Benedick, S., Hill, J. K., Mustaffa, N., et al. (2006). Impacts of Gascon, C., Lovejoy, T. E., Bierregaard, R. O., Malcolm, ﬂ rainforest fragmentation on butter ies in northern Bor- J. R., Stouffer, P. C., Vasconcelos, H. L., Laurance, neo: species richness, turnover and the value of small W. F., Zimmerman, B., Tocher, M., and Borges, – fragments. Journal of Applied Ecology, 43, 967 977. S. (1999). Matrix habitat and species richness in tropical Bennett, A. F. (1999). Linkages in the landscape. The role of forest remnants. Biological Conservation, 91, 223–229. corridors and connectivity in wildlife conservation. IUCN- Grey, M. J., Clarke, M. F., and Loyn, R. H. (1997). Initial The World Conservation Union, Gland, Switzerland. changes in the avian communities of remnant eucalypt Bright, P. W., Mitchell, P., and Morris, P. A. (1994). Dor- woodlands following reduction in the abundance of mouse distribution: survey techniques, insular ecology noisy miners, Manorina melanocephala. Wildlife Research, and selection of sites for conservation. Journal of Applied 24, 631–648. – Ecology, 31, 329 339. Haddad, N. M., Bowne, D. R., Cunningham, A., Danielson, Brittingham, M. and Temple, S. (1983). Have cowbirds B. J., Levey, D. J., Sargent, S., and Spira, T. (2003). Corri- – caused forest songbirds to decline? BioScience, 33,31 35. dors use by diverse taxa. Ecology, 84, 609–615. Brooker, L. and Brooker, M. (2002). Dispersal and popula- Hanski, I. (1999). Metapopulation ecology. Oxford University tion dynamics of the blue-breasted fairy-wren, Malurus Press, Oxford, UK. pulcherrimus, in fragmented habitat in the Western Aus- Harvey, C. A., Medina, A., Sanchez, D., et al. (2006). Pat- – tralian wheatbelt. Wildlife Research, 29, 225 233. terns of animal diversity in different forms of tree cover Chapman, C. A., Balcomb, S. R., Gillespie, T. R., Skorupa, in agricultural landscapes. Ecological Applications, 16, J. P., and Struhsaker, T. T. (2000). Long-term effects of 1986–1999. logging on African primate communities: a 28 year com- Hill, J. K., Thomas, C. D., and Lewis, O. T. (1996). Effects of parison from Kibale National Park, Uganda. Conserva- habitat patch size and isolation on dispersal by Hesperia – tion Biology, 14, 207 217. comma butterﬂies: implications for metapopulation Connor, E. F. and Mccoy, E. D. (1979). The statistics and structure. Journal of Animal Ecology, 65, 725–735. biology of the species-area relationship. American Natu- Hobbs, R. J. (1993). Effects of landscape fragmentation on – ralist, 113, 791 833. ecosystem processes in the Western Australian wheat- Daily, G.C., Ceballos, G., Pacheco, J., Suzán, G., and Sán- belt. Biological Conservation, 64, 193–201. chez-Azofeifa, A. (2003). Countryside biogeography of Hobbs, R. J. and Yates, C. J. (2003). Turner Review No. 7. neotropical mammals: conservation opportunities in ag- Impacts of ecosystem fragmentation on plant

106 CONSERVATION BIOLOGY FOR ALL

populations: generalising the idiosyncratic. Australian Radford, J. Q. and Bennett, A. F. (2007). The relative impor- Journal of Botany, 51, 471–488. tance of landscape properties for woodland birds in Laurance, W. F. (2008). Theory meets reality: how habitat agricultural environments. Journal of Applied Ecology, 44, fragmentation research has transcended island biogeo- 737–747. graphic theory. Biological Conservation, 141, 1731–1744. Radford, J. Q., Bennett, A. F., and Cheers, G. J. (2005). Lindenmayer, D. B. and Fischer, J. (2006). Habitat fragmen- Landscape-level thresholds of habitat cover for wood- tation and landscape change. An ecological and conservation land-dependent birds. Biological Conservation, 124, synthesis. CSIRO Publishing, Melbourne, Australia. 317–337. Loyn, R. H. (1987). Effects of patch area and habitat on bird Redpath, S. M. (1995). Habitat fragmentation and the indi- abundances, species numbers and tree health in fragmen- vidual: tawny owls Strix aluco in woodland patches. ted Victorian forests. In D. A. Saunders, G. W. Arnold, A. Journal of Animal Ecology, 64, 652–661. A. Burbidge, and A. J. M. Hopkins, eds Nature conserva- Ricketts, T. H. (2001). The matrix matters: effective isolation: the role of remnants of native vegetation,pp.65–77. tion in fragmented landscapes. The American Naturalist, Surrey Beatty and Sons, Chipping Norton, Australia. 158,87–99. Lunt, I. D. and Spooner, P. G. (2005). Using historical Saunders, D. A. and Hobbs, R. J., eds (1991). Nature conser- ecology to understand patterns of biodiversity in frag- vation 2: The role of corridors. Surrey Beatty & Sons, Chip- mented agricultural landscapes. Journal of Biogeography, ping Norton, New South Wales. 32, 1859–1873. Saunders, D. A., Hobbs, R. J., and Arnold, G. W. (1993). The MacArthur, R. H. and Wilson, E. O. (1967). The theory of Kellerberrin project on fragmented landscapes: a review of island biogeography. Princeton University Press, Prince- current information. Biological Conservation, 64,185–192. ton, New Jersey. Saunders, D. A., Hobbs, R. J., and Margules, C. R. (1991). McGarigal, K. and Cushman, S. A. (2002). Comparative Biological consequences of ecosystem fragmentation: a evaluation of experimental approaches to the study of review. Conservation Biology, 5,18–32. habitat fragmentation effects. Ecological Applications, 12, Sekercioglu, C. H., Ehrlich, P. R., Daily, G. C., Aygen, D., 335–345. Goehring, D., and Sandi, R. F. (2002). Disappearance of McIntyre, S. and Hobbs, R. (1999). A framework for con- insectivorous birds from tropical forest fragments. Pro- ceptualizing human effects on landscapes and its rele- ceedings of the National Academy of Sciences of the United vance to management and research models. Conservation States of America, 99, 263–267. Biology, 13, 1282–1292. Soulé, M. E. (1986). Conservation biology and the “real Newmark, W. D. (1991). Tropical forest fragmentation and world”. In M. E. Soule, ed. Conservation biology. The the local extinction of understory birds in the Eastern science of scarcity and diversity, pp. 1–12. Sinauer Associ- Usambara Mountains, Tanzania. Conservation Biology, 5, ates, Sunderland, Massachusetts. 67–78. Terborgh, J., Lopez, L., Nunez V. P., et al. (2001). Ecological Patterson, B. D. and Atmar, W. (1986). Nested subsets and meltdown in predator-free forest fragments. Science, 294, the structure of insular mammalian faunas and archipe- 1923–1926. lagoes. Biological Journal of the Linnean Society, 28,65–82. Trzcinski, M. K., Fahrig, L., and Merriam, G. (1999). Inde- Peres, C. A. (2001). Synergistic effects of subsistence hunt- pendent effects of forest cover and fragmentation on the ing and habitat fragmentation on Amazonian forest ver- distribution of forest breeding birds. Ecological Applica- tebrates. Conservation Biology, 15, 1490–1505. tions, 9, 586–593. Peters, M. K., Likare, S., and Kraemar, M. (2008). Effects of Virgos, E. (2001). Role of isolation and habitat quality in habitat fragmentation and degradation on ﬂocks of African shaping species abundance: a test with badgers (Meles ant-following birds. Ecological Applications, 18,847–858. meles L.) in a gradient of forest fragmentation. Journal of Prober, S. M. and Thiele, K. R. (1995). Conservation of the Biogeography, 28, 381–389. grassy white box woodlands: relative contributions of Wilcove, D. S. (1985). Nest predation in forest tracts size and disturbance to ﬂoristic composition and diver- and the decline of migratory songbirds. Ecology, 66, sity of remnants. Australian Journal of Botany, 43, 349–366. 1211–1214.