You are on page 1of 23

3 • The ecological context: a landscape perspective

RICHARD J. HOBBS

INTRODUCTION are discernible, and in which units appear as con-


trasting, discrete states of physical or ecological phe-
Restoration ecology has developed from, and has been
nomena (Ostfeld et al., 1997). Patches may comprise
practised primarily on, a site-based approach. The
different ecosystems (e.g. lakes, rivers, forest, grass-
restoration of a well-defined area such as a minesite,
land), different land uses (e.g. urban, agricultural,
wetland or a degraded ecosystem of some descrip-
nature reserve), or different community types, suc-
tion is generally attempted. However, it is clear that
cessional stages or alternative states within a partic-
relatively large areas of the earth are in need of
ular ecosystem (e.g. post-fire, pole-stage and old-
some form of restoration, following degradation
growth forest stands).
through overuse or inappropriate management,
Landscape ecology considers three main aspects
which has impaired the functioning or altered the
of landscapes: structure (or pattern), function (or
structure of the landscape as a whole (MacMahon,
process) and change (Forman & Godron, 1986;
1998). Thus, there is a need to expand the scope of
Turner, 1989; Forman, 1995; Pickett & Cadenasso,
restoration ecology to embrace broader scales and
1995; Turner et al., 1995). The characteristics of indi-
tackle landscape-scale problems. While this is in-
vidual patches and their spatial relationship with
creasingly recognised, the science of landscape-scale
other patches determine landscape structure, while
restoration is still in a formative phase (Bell et al.,
landscape function is determined by physical, chem-
1997). In this chapter I present a summary of land-
ical and biotic transfers between patches. Landscape
scape structure and function, discuss the impacts of
change results from either changes in individual
human modification of landscapes, and present a
patches or changes in patch configurations and in-
series of options for developing guidelines for land-
terrelations. The three aspects of landscapes are
scape restoration. Some of the material in this chap-
closely interlinked, since structure strongly influ-
ter is modified from work presented elsewhere
ences function, which can feed back into structure,
(Hobbs, 1995, 1999; Hobbs & Harris, 2001; Hobbs &
and landscape change can affect both structure and
Lambeck, in press).
function.

LANDSCAPES: STRUCTURE Structure


AND FUNCTION Landscape mosaics are commonly complex entities
A landscape is defined as an area of land, at the scale consisting of numerous patches of varying types in a
of hectares to square kilometres, which consists of a variety of configurations. Our understanding of
collection of different, but interacting patches (also landscape patterns has increased greatly with the
called landscape elements). Patchiness focuses on advent of remote sensing and geographic informa-
the spatial matrix of ecological processes, and em- tion system (GIS) technologies. Remote sensing, us-
phasises the fluxes of materials and organisms ing airborne or satellite sensors, offers the possib-
within and between parts of the landscape. It is a ility to acquire large amounts of data on the
form of spatial heterogeneity in which boundaries characteristics of the earth’s surface in a relatively

24
Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
A landscape perspective 25

easy, repeatable and analysable way. Geographic add to the complexity of vegetation patterning. This
information systems allow us to develop spatially ex- patchiness can be as a result of chance dispersal
plicit databases on a wide range of landscape fea- events, response to localised disturbances or re-
tures, which again allows quantitative analysis of sponse to micro-environmental variation.
pattern (Haines-Young et al., 1993). The history of disturbances determines the distri-
An assessment of landscape pattern requires that bution of patches of different age and successional
the scale of investigation is defined, and this scale stage within a landscape, and may, in some cases, be
will be determined by the types of question being responsible for the development of mosaics of
asked and the types of organism or process being patches in alternative semi-permanent vegetation
studied. Many problems, misunderstandings, con- states (Turner, 1987; Baker, 1992; Hobbs, 1994). The
flicting results and misapplication of research find- components of the disturbance regime will deter-
ings arise from failures to determine the relevant mine the scale and pattern of variation observed,
scale of study or to ensure that studies are con- and each disturbance type may produce different
ducted at similar scales. While the landscape level is vegetation responses. Individual disturbance types
the primary focus here, it is important to recognise can also produce different responses depending on
that individual landscapes occur in a regional set- factors such as environmental variations within the
ting. Indeed, boundaries between landscapes are fre- disturbed areas, weather characteristics following
quently little more than convenient lines drawn by the disturbance, and interactions with other distur-
humans. A frequently used natural landscape unit is bances (e.g. Suffling, 1993). Important landscape-
the catchment or watershed (the area of land within scale disturbances include fire, drought, infrequent
which water drains into one watercourse), since this frosts or periods of higher than normal tempera-
has natural topographically determined boundaries. tures, severe storms, localised soil disturbance by an-
Also of importance is the patch level – i.e. the level imals, tree falls and insect outbreaks. In addition to
of the units which make up landscapes. Individual the natural disturbance regime, human disturbance
patches have a set of characteristics (e.g. size, com- is an important component of many ecosystems. In
position, age) which can influence processes both regions with long histories of human habitation, hu-
within the patch and at the landscape level. The man-induced disturbance has frequently been an
recognition that patterns and processes at one level important influence in shaping and maintaining
can be influenced by patterns and processes at other ecosystems and landscapes (e.g. le Houérou, 1981;
(higher and lower) levels is essential if we are to Thirgood, 1981). In many parts of the world, aborigi-
come to grips with the complexity of large-scale eco- nal inhabitants exerted a powerful influence on the
logical patterns and processes. landscape, but this was usually significantly modi-
Pattern in the landscape is the result of the in- fied following invasion by European settlers (e.g.
teraction of many influences. At a broad scale, pat- Flannery, 1994). Current human activities such as
terns of vegetation composition and structure can vegetation removal, timber or soil extraction and
be related to regional gradients in climatic vari- introduction of non-native plants and animals
ables such as temperature and rainfall, and to continue to produce a further overlay of variation
changes in soil and landform type and topography. on the landscape, as well as altering the natural or
Within any given region, climate, soil type and aboriginal regime.
landform generally determine the broad vegetation
patterning. Within that broad patterning, however,
numerous smaller scales of pattern may be present.
Heterogeneity
These may be determined by finer-scale variations Landscape heterogeneity is a complex multi-scale
in soil characteristics or microclimate, but may also phenomenon, involving the size, shape and composi-
be the result of other factors. Species turnover be- tion of different landscape units and the spatial (and
tween different parts of a landscape and patchy dis- temporal) relations between them. Adequate meth-
tribution of populations of individual species can ods for measuring and assessing its significance have

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
26 RICHARD HOBBS

yet to be developed. Landscape heterogeneity is fre- is process-dependent and hence can only be defined
quently examined at two levels, between landscapes in the context of a particular process (e.g. animal
(regional scale) and within landscapes (landscape movement, water flow) (Cale & Hobbs, 1994). The
scale). Various indices have been developed for use at measures of landscape heterogeneity currently
both these scales, and include various combinations available are frequently difficult to interpret and
of estimates of numbers of different landscape units should be used with caution in both research and
present, areas occupied by different landscape units management.
and lengths of landscape unit perimeters (Turner &
Gardner, 1991; Turner et al., 1991; Hargis et al., 1998).
Other approaches consider the underlying patterns
Ecotones and edges
of plant species richness and variations in evenness A component of landscape structure which has
(Scheiner, 1992). To date, most studies using these in- received increasing attention recently is the eco-
dices have investigated their ability to quantify land- tone, or edge between adjacent patches. Ecotones
scape pattern in space and with time. In general, the can be considered at a variety of scales, ranging from
indices have been considered useful if they have in- the biome down to the individual patch. Ecotones
dicated differences between landscapes already are considered important because they represent the
known to differ substantially, but there are few in- boundary between different patches through which
stances of indices providing new insights, allowing various landscape flows pass (Holland et al., 1991;
extrapolation to other situations. Similarly, the rela- Hansen & di Castri, 1992; Gosz, 1993). One compo-
tionship between the index and functional aspects of nent of this is that the ecotone between different
heterogeneity have rarely been explored (Cale & biome or vegetation types could be expected to be the
Hobbs, 1994). place where the first indications of responses to
Changing the scale of measurement affects the global climatic changes would be detected. Patch size
results obtained. Changing the grain (spatial resolu- and shape strongly affect the amount of edge that is
tion) and extent (total area of study) affects meas- present. A set of phenomena known as ‘edge effects’
ures of heterogeneity because they are sensitive to is associated with edges. These result from physical,
the number of patches detected, which changes chemical and biotic transfer into patches from adja-
with changes in scale (Turner et al., 1991). The spatial cent patches (see below).
configuration of patches influences the rate of
change in their number with changing scale.
An adequate description of landscape heterogene-
Landscape fluxes
ity must include not only a description of the num- Connectivity and movement of biota
ber, sizes and configurations of patches, but also Connectivity refers to the degree to which flows
some characterisation of the structure and composi- (for example, of animals or materials) are possible
tion within them. Study of landscape heterogeneity between different patches. For biota, populations
has tended to focus on one of these scales (i.e. con- in one area may be linked to other neighbouring
figuration of landscape units within the landscape populations through dispersal or source – sink re-
or configuration within landscape units), but sel- lations. More generally, biota often need to move
dom both together. across the landscape for a variety of reasons, includ-
The interpretation of differences in heterogeneity ing dispersal and resource acquisition, and that
between landscapes is difficult. If one landscape has movement is required to counter the potential ef-
a higher index than another, this may or may not fects of fragmenting populations into small, iso-
have any significance when particular processes or lated units. Movement can be thought of as mini-
functions are considered. The significance of the mising the impacts of demographic stochasticity
measured difference in heterogeneity depends on and inbreeding depression.
how well measured heterogeneity corresponds to A habitat network can be defined as an intercon-
functional heterogeneity. Functional heterogeneity nected set of habitat elements which together allow

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
A landscape perspective 27

movement, although recent studies of marked or ra-


Patch
dio-tagged animals have indicated that some species
do use corridors for movement, in preference to
moving across open ground. Other studies have
Matrix shown that corridors can act in complex ways, en-
hancing movement of some species in some cases,
while inhibiting that of others (e.g. Hill, 1995).
Corridor The universality of the patch–corridor–matrix
model has been questioned recently, particularly in
relation to the influence of the matrix, which may
not always be completely hostile to all elements of the
biota (McIntyre & Hobbs, 1999). Although corridors
Fig. 3.1. Simple patch–corridor–matrix model of have received widespread attention and are fre-
landscapes, which dichotomises landscape elements into quently part of conservation plans and activities,
habitat and non-habitat (matrix), with corridors linking their utility is often debated, and they are in reality
habitat patches. Derived from Forman (1995). only one part of a broader picture (Hobbs, 1992;
Dawson, 1994; Wilson & Lindenmayer, 1996; Hobbs &
for movement of biota and enhance population sur- Wilson, 1998; Bennett, 1999). Probably a more con-
vival probabilities. The basis of this concept is the structive approach is to consider overall landscape
assumption that patches of habitat are generally em- connectivity, the extent to which different elements
bedded in a matrix of ‘non-habitat’. A common per- of the landscape are functionally connected from the
ception among conservation biologists is that the viewpoint of particular biotic elements. Connectivity
matrix is hostile to the organisms within the rela- in a landscape depends on the relative isolation of
tively small fragments. This concept has been re- habitat elements from one another and the extent to
fined to more generally describe landscapes in terms which the matrix represents a barrier to movement of
of patch, corridor and matrix (Forman, 1995), with organisms. There have been attempts to derive some
corridors representing narrow strips of habitat or, if generalities concerning connectivity from modelling
not habitat, at least vegetation that allows biotic simple geometric relationships arising from the dis-
movement between patches (Fig. 3.1). The retention tribution of habitat patches in landscapes with differ-
or provision of corridors between fragments is often ent proportions of habitat and non-habitat (With &
seen as an important element of landscape manage- Crist, 1995; Pearson et al., 1996; Wiens, 1997; With,
ment and conservation (Harris, 1984; Hudson, 1991). 1997). These studies have indicated that there may be
Some argue, however, that the requirement for fau- thresholds where small changes in the proportion of
nal movement may have been overstated, and that habitat present result in large changes in connectiv-
corridors may not be required to foster it when it is ity. However, the problem remains that different
necessary (Simberloff & Cox, 1987; Simberloff et al., species will perceive the landscape differently, and
1992). While movement along corridors is fre- landscape connectivity will depend on the mobility
quently assumed to occur, there have been relatively and habitat specificity of the species involved (Cale &
few studies that have shown that corridors are actu- Hobbs, 1994; Pearson et al., 1996; Kolasa & Waltho,
ally required for movement (Hobbs, 1992). Studies 1998). Also, most modelling efforts consider straight-
that have been frequently cited as illustrating corri- forward habitat versus non-habitat dichotomies; they
dor use for faunal movement, do not, in fact, provide therefore do not deal with the possibility that the ma-
clear evidence. The types of study required to estab- trix may be more or less permeable, or conversely, re-
lish unequivocally that corridors are important for sistant to movement. Hence the role of all landscape
faunal movement are difficult and costly to design elements in facilitating or inhibiting movement
and implement and require intensive, long-term ob- needs to be considered (Taylor et al., 1993; Wiens,
servations. Few studies provide good data on animal 1997; McIntyre & Hobbs, 1999).

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
28 RICHARD HOBBS

An important function of animal movement is the landform and surface characteristics. Patch charac-
recolonisation of patches that have suffered local ex- teristics and configurations can influence intercep-
tinction. The concept of metapopulations has re- tion, infiltration, evapotranspiration and runoff
ceived increasing attention recently, especially in the patterns. Thus, for instance, perennial vegetation in
context of fragmented landscapes. This concept as- semi-arid areas can intercept and utilise more water
sumes that species’ populations can exist as a series than annual cropland, with the result that runoff
of interlinked sub- or metapopulations which persist and water input to groundwater will be greater
in separate habitat patches, but may be linked by in- under an annual cropping regime. Distribution of
terpatch movement (McCulloch, 1996; Hanski & water across the landscape clearly influences the
Gilpin, 1997). More explicitly, the metapopulation distribution of patch types – for instance, riparian
approach assumes that ‘populations are spatially forests and swamps can develop only where there
structured into assemblages of breeding populations are large amounts of water available close to the
and that migration among the local populations has surface.
some effect on local dynamics, including the possi- Nutrient and material fluxes across the land-
bility of population re-establishment following ex- scape result primarily from the processes of ero-
tinction’ (Hanski & Simberloff, 1997). Subpopula- sion, leaching and transport by wind and water.
tions characteristically undergo periodic extinctions Redistribution of nutrients over long time periods
and re-establish following recolonisation by dis- has resulted in areas of the landscape accumulating
persers from other elements of the metapopulation. nutrients which have been eroded from other areas.
Determining whether metapopulation dynamics ac- This in turn exerts a strong influence on the types
tually occur in fragmented landscapes is, however, of patch present in each area. Resource-rich patches
relatively difficult, and there is still some question as may be particularly important from many points of
to the general validity of the concept. view, both in terms of vegetation composition and
Metapopulation theory deals with the factors faunal assemblages, and also in terms of potential
that determine the likelihood of local extinctions human utilisation. Landscape patterning can in-
and subsequent recolonisation. In general, subpopu- fluence erosional processes, with some patch types
lations in small isolated patches are considered acting as interceptors of eroded material. For in-
more likely to go extinct while large well-connected stance, in managed landscapes windbreaks act
patches are more likely to be recolonised, resulting both to reduce the degree of wind erosion and to
in different probabilities of species occurring on intercept eroded material. Riparian strips also in-
patches of different sizes and connectivity. Clearly, tercept eroded material and nutrients transported
the validity of these relationships depends on the ex- in runoff, and hence affect nutrient and material
tent to which our perceptions of the influence of loadings in waterways. Riparian strips are fre-
size and connectivity match the actual influence of quently viewed as ‘buffer zones’, and this concept
these parameters on the species involved. It is also can be extended to any type of patch which pro-
possible that metapopulations exhibit non-linear tects adjacent patches from nutrient or material
dynamics and may collapse unexpectedly to extinc- inputs.
tion (Hanski et al., 1995), making it difficult to pre- Nutrients can be transferred across the landscape
dict the occurence or dynamics of populations in in a variety of other ways. Particularly important
fragmented landscapes. vectors are animals, which may feed in one patch
and defaecate in another, resulting in a transfer of
Water, nutrients and material nutrients between patches. In extreme cases, this
Fluxes of water, nutrients and material are often im- can significantly increase the nutrient input into
portant determinants of landscape patterning, and recipient patches, for instance in the case of seabird
can also be strongly affected by that patterning and colonies. Fire also redistributes nutrients in smoke
its dynamics (Hornung & Reynolds, 1995). Surface and ash, although it is difficult to quantify this
and subsurface hydrology is determined by geology, effect.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
A landscape perspective 29

Movement of pollutants and excess nutrients In this respect, episodic events can be considered as
across the landscape is important in determining particular types of disturbance. Any form of distur-
the nutrient and pollution loading in any given bance, such as fire, flood, storm or landslide can
location. It is becoming increasingly recognised that also act in this way. As discussed earlier, the distri-
landscape and regional processes have to be con- bution of patches within a landscape is determined
sidered when dealing with pollution problems in by the overall disturbance regime.
particular ecosystems. For instance, deposition of Landscape change may also be biotically forced.
nitrogen or sulphur arising from regional atmos- Species migrations, as discussed above, are important
pheric pollution can have important local impacts here, but the effects of animals and pathogens should
(Appleton, 1995), and applications of phosphorus also be noted. Changes in abundances of herbivores
fertiliser to agricultural land can run off and cause can lead to changes in patch types and configura-
pollution problems in adjacent lakes or estuaries tions, and pathogen spread can significantly alter
(e.g. Boggess et al., 1995; Flaig & Reddy, 1995). For landscape composition and structure (Knight, 1987).
large river systems, pollution problems may have to There has been considerable debate over the idea
be tackled not just at the landscape scale, but also at that, while individual landscape patches are in a
the national or continental level (Malle, 1996). constant state of flux, the landscape as a whole is in
a state of equilibrium. In other words, over the en-
tire landscape, the distribution of patches of differ-
LANDSCAPE CHANGE ent types or ages should remain constant through
time. For this to be the case a certain minimum area
Natural change is required, known as the ‘minimum dynamic area’
Natural landscapes are in a constant state of flux. (Pickett & Thompson, 1978; Baker, 1992). The validity
Patch composition and configuration changes in re- of the assumption of equilibrium landscapes de-
sponse to a variety of processes, particularly climatic pends greatly on the scale considered (Turner et al.,
changes and disturbance regimes. Climate varies 1993), and recent writers have questioned its gener-
greatly over a number of different time-scales, rang- ality. Examination of areas where large tracts of nat-
ing from the long-term changes which take place over ural ecosystems persist indicates a distribution of
thousands of years in response to glacial/interglacial patch types and ages that is far from equilibrium
cycles to relatively short-term phenomena such as (Sprugel, 1991).
drought cycles over periods of decades. A consider-
able body of evidence is available which shows that
natural vegetation responds to long-term climatic
Human-induced change: modification
changes by migrating across the landscape (Delcourt
and fragmentation
& Delcourt, 1991). However, individual species mi- Fragmentation of natural ecosystems has been
grate at different rates and, at any given time, the called one of the most pervasive changes in terres-
composition of the landscape depends on which trial ecosystems across the earth. It occurs wherever
species are already present, which species are migrat- land transformation results in the removal of the
ing in, and how these two sets of species interact. Past pre-existing land cover and its replacement with
landscapes in any particular area are liable to have other cover types, be it urban, agriculture, produc-
been quite different to those present today. tion forestry or other anthropogenic land uses. Such
Infrequent episodic climatic events are particu- activities may remove most of the pre-existing cover
larly important agents of landscape change. Events types, transforming landscapes and regions in rela-
such as droughts, exceptionally high rainfalls or tively short periods. Examples include deforestation
windstorms are capable of switching patches from in tropical areas such as Amazonia, and clearance
one type to another very quickly (Hobbs, 1994). For for agriculture in the Midwest of the United States
instance, a grassland patch can become a shrubland and in Australia (Hobbs & Saunders, 1993; Laurance
patch following a year of exceptionally high rainfall. & Bierregaard, 1997; Schwartz, 1997). In these and

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
30 RICHARD HOBBS

many other cases, the result of such activities is that year. If the fragmentation process removes a particu-
the pre-existing cover types are left as patches in a lar plant community or species which provided nec-
modified matrix of production land. These patches tar resources at a critical time of year, nectarivorous
may be of varying sizes and may be isolated from birds could no longer survive in the area.
other such patches to varying degrees.
Connectivity and corridors
Impact of fragmentation on biota As noted above, a further factor influencing popula-
Considerations of fragmented landscapes have fo- tion viability may be the ability of species to move
cused on factors that influence the likelihood of around the landscape. Species need to move for nu-
persistence of the biota retained in the remaining merous reasons, such as to fulfil resource require-
fragments of habitat. Obviously, the major impact of ments, to disperse to new territories, or to migrate.
transforming one ecosystem type to another is the Some species may move only short distances and
dramatic reduction in the amount of habitat left for hence be able to exist entirely within an individual
those species dependent on the pre-existing ecosys- habitat patch, but others may need to move either
tem type. The actual area of many ecosystem types is between different habitat types or across the land-
reduced to a few percent of their original extents, scape between different fragments. Others may need
and this not only affects their representation in the to move across entire continents as they follow mi-
landscape and in any reserve system that is devel- gration routes.
oped, but also affects the ability of individual In an unfragmented landscape, the ability of
species to survive. Species vary greatly in their habi- species to move depended entirely on the ‘perme-
tat requirements, and it is likely that in fragmented ability’ of the different ecosystem types present, i.e.
systems, populations of some species are limited the extent to which different patch types either
simply by the amount of habitat available. aided or hindered movement. The permeability of
In addition to the simple amount of habitat re- different landscape elements is likely also to vary
maining, the spatial arrangement of that habitat from species to species. Some species will only move
may also be important. Considerable debate centred in thickly vegetated cover types, whereas other
for many years on whether fragmented systems species favour more open country. In a fragmented
would perform better in terms of retaining species landscape, dense wooded cover is often replaced with
if they consisted of a few large fragments versus lots low, open crop or pasture. In this situation we might
of small patches (Simberloff, 1982; Willis, 1984; expect that the permeability of the landscape would
Lomolino, 1994; Boecklen, 1997). The so-called SLOSS decrease for species which require wooded cover. The
(single large or several small) debate stemmed from exact impact of the fragmentation depends both on
ideas generated from the theory of island biogeog- its extent and pattern and also on how particular
raphy, which related species numbers on islands to species view the landscape. Using simple geometric
the island size and isolation. There is no real resolu- models, it can be shown that the extent to which the
tion to this debate, since different types of organism various habitat patches in the landscape remain
will probably require different amounts and distri- ‘connected’ declines as the extent of habitat removal
butions of habitat. For instance, large mammalian increases. However, there is also evidence that the re-
predators probably require large uninterrupted lationship is not linear, and that there is a threshold
tracts of habitat, whereas insects and plants can in ‘connectivity’ at about 60% habitat removal, at
probably survive well in much smaller areas. which the level of connectivity declines sharply.
Other factors may also limit populations, how-
ever. For instance, some species may be limited, not Edge effects and processes
by the areal extent of habitat, but by the availability In addition to the actual removal of the pre-existing
of a particular resource on which they depend. An ecosystem type and the creation of the patch/matrix-
example may be nectarivorous birds which depend type landscape discussed above, ecosystem modi-
on a continuous supply of nectar throughout the fication can also occur, whereby the pre-existing

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
A landscape perspective 31

ecosystem type is not removed entirely, but is altered may occur only over tens of metres, while changes in
by human use, such as by grazing, timber harvesting predation rates or weed invasion may extend much
and so on. The two processes of fragmentation and further. For instance, edge-related impacts can result
modification can occur simultaneously, and often in considerable mortality of the species up to 100 m
fragments remaining in a modified matrix are also from the fragment edge (Laurance et al., 1997), and
severely impacted by modifying influences (Hobbs, this impact can be particularly severe in small frag-
1993b; McIntyre & Hobbs, 1999). Thus, often it is im- ments where virtually the entire fragment becomes
portant to consider not only the amount of a partic- ‘edge’ rather than ‘interior’ habitat.
ular ecosystem which is left, but also the condition In addition to the localised edge effects, the sur-
of the remaining patches. rounding landscape can influence fragments more
Factors which can influence the condition of generally. For instance, significant changes in hy-
patches include both factors operating within drology can result from the process of replacing one
individual patches and factors arising from the cover type with another. These large-scale changes
surrounding landscape. An important feature of influence not only the production matrix but also
fragmented ecosystems is that the pre-existing dis- the ecosystems remaining in fragments (George
turbance regime can no longer operate (Hobbs, et al., 1995). Also, animals and plants prevalent in
1987). For instance, fire frequency and intensity may the management matrix may exert profound influ-
increase or decrease in isolated fragments, and this ences on the biota in fragments. In particular, do-
in turn affects the survival and regeneration of the mestic stock can dramatically alter the structure
plant communities in the fragments. Also, if move- and composition of fragmented systems by grazing
ment of faunal species around the landscape be- and trampling. Similarly, feral predators can be re-
comes restricted, seed dispersal, and hence gene sponsible for the decline of faunal species which
flow, of plant species may be reduced or even cease. would otherwise survive in the fragmented system.
As species are lost from fragmented systems, so too Finally, invasive plant species may disperse into frag-
are any functions that these species performed. For mented systems from the surrounding matrix, and
instance, if burrowing mammals disappear from again these have the potential to significantly alter
fragmented systems, soil disturbance and turnover the structure and dynamics of the fragmented sys-
is reduced, and hence nutrient cycling and water tem. Hence fragmentation results not only in the
infiltration to the soil may be affected. formation of small, isolated fragments, but also in
The isolation of fragments in a modified matrix landscape-scale system changes.
opens the fragments up to a range of microclimatic,
hydrological and biotic changes (Saunders et al.,
1991). A primary result of this isolation is the creation
THE BASIS OF LANDSCAPE-SCALE
of an edge between the pre-existing ecosystem and
RESTORATION: UNDERSTANDING
the modified matrix. At this edge, the conditions
THRESHOLDS AND SETTING GOALS
prevalent in the matrix impinge into the fragment, Most of the information and methodologies on
and a series of ‘edge effects’ can occur. The edges of ecological restoration centre on individual sites, and
isolated fragments experience different microcli- ultimately restoration activities have to be con-
matic conditions, receive more nutrients transferred ducted in particular sites. However, site-based
from adjacent patches and may have a higher inci- restoration has to be placed in a broader context,
dence of weed invasion or predation than the interior and is often insufficient on its own to deal with
of the fragment (Kapos, 1989; Matlack, 1993; Murcia, large-scale restoration problems. Landscape- or re-
1995). Such changes can result in changes in vegeta- gional-scale processes are often either responsible
tion structure and floristic and faunal species compo- for ecosystem degradation at particular sites, or al-
sitions. The distance to which edge effects permeate a ternatively have to be restored to achieve restoration
patch varies with the type of patch and the feature be- goals. Hence restoration is often needed both within
ing considered: for instance microclimatic changes particular sites and at a broader landscape scale.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
32 RICHARD HOBBS

How are we then to go about restoration at a This ties in with recent attempts to use the concept
landscape scale? What are the relevant aims? What of ecosystem health as an effective means of dis-
landscape characteristics can we modify to reach cussing the state of ecosystems (Costanza et al.,
these aims, and do we know enough to be able to 1992; Cairns et al., 1993; Shrader-Frechette, 1994).
confidently make recommendations on priorities Central elements of ecosystem health are the sys-
and techniques? tem’s vigour (e.g. production), organisation (or the
There are a number of steps in the development diversity and number of interactions between sys-
of a programme of landscape-scale restoration, tem components) and resilience (the system’s ca-
which can be outlined as follows: pacity to maintain structure and function in the
presence of stress (Rapport et al., 1998). Attempts
1. Assess whether there is a problem which requires
have also been made to produce readily measurable
attention, for instance:
indices of ecosystem health for a number of differ-
(a) changes in biotic assemblages (e.g. species loss
ent ecosystems, although there is still debate over
or decline, invasion)
whether these are useful or not. In the same way,
(b) changes in landscape flows (e.g. species move-
there have been recent attempts to develop a set of
ment, water and/or nutrient fluxes)
measures of landscape condition.
(c) changes in aesthetic or amenity value (e.g. de-
Aronson and Le Floc’h (1996b) present three
cline in favoured landscape types).
groups of what they term ‘vital landscape attributes’
2. Determine the causes of the problem, for instance:
which aim to encapsulate (1) landscape structure and
(a) removal and fragmentation of native vegetation
biotic composition, (2) functional interactions among
(b) changes in pattern and abundance of vegeta-
ecosystems and (3) degree, type and causes of land-
tion/landscape types
scape fragmentation and degradation. Their list of 16
(c) cessation of historic management regimes.
attributes provides a useful start for thinking about
3. Determine realistic goals for restoration, for in-
these issues, but many of the attributes are either dif-
stance:
ficult to measure or hard to interpret, or both. It thus
(a) retention of existing biota and prevention of
remains difficult to conduct a practical assessment of
further loss
whether a particular landscape is in need of restora-
(b) slowing or reversal of land or water degradation
tion, and if so, what actions need to be taken. Steps to-
processes
wards this are being developed, at least for landscape
(c) maintenance or improvement of potential for bi-
flows, in the Landscape Function Analysis approach
ological production
developed for Australian rangelands (see Ludwig et al.,
(d) integration of approaches to tackle multiple
1997; Tongway & Ludwig, volume 2), which aims to
goals.
provide easily measurable and interpretable indices
4. Develop cost-effective planning and management
of landscape function. These authors indicate that
tools for achieving agreed goals:
the interpretation of measures of landscape condi-
(a) determining priorities for action in different
tion must be based on the goals set for that particular
landscape types and conditions
landscape. A particular condition may indicate that
(b) spatially explicit solutions
the landscape is quite acceptable for one type of use,
(c) acceptance and ‘ownership’ by managers and
but in need of restoration to allow the continuation
landholders
or adoption of another land use.
(d) an adaptive approach which allows course cor-
Once a problem has been perceived, the correct
rections when necessary.
diagnosis of its cause and prescription of an effec-
This short list hides a wealth of detail, uncertainty tive treatment is by no means simple. The assump-
and science yet to be done. For instance, the initial tion underlying landscape ecology is that landscape
assessment of whether there is a problem or not processes are in some way related to landscape
requires the availability of a set of readily measura- patterns. Hence, by determining the relationship be-
ble indicators of landscape ‘condition’ or ‘health’. tween pattern and process, better prediction of what

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
A landscape perspective 33

will happen to particular processes (biotic move- Considering system function provides a useful
ment, metapopulation dynamics, system flows, etc.) framework for the initial assessment of the state of
if the pattern of the landscape is altered in particu- the system and the subsequent selection of repair
lar ways, can occur. Thus, we are becoming increas- measures (Tongway & Ludwig, 1996; Ludwig et al.,
ingly confident that we can, for instance, predict the 1997). Where function is not impaired, goals for
degree of connectivity in a landscape from the pro- restoration can legitimately focus on composition
portion of the landscape in different cover types. and structure.
Similarly, as landscapes become more fragmented, a The same scheme can be considered at a land-
greater proportion of the biota drops out, and again scape scale. At broad scales it becomes even more
there may be thresholds or breakpoints where difficult to decide what should be restored, where
relatively large numbers of species are lost. and how. It can be hypothesised that restoration
As part of this process, we need to develop ecolog- thresholds might exist at the landscape scale as are
ical response models which capture the essence of apparent in particular ecosystems or sites (Fig. 3.2b).
the landscape and its dynamics. These models can One type of threshold relates to the loss of biotic
be simple or complex, quantitative or conceptual, connectivity as habitat becomes increasingly frag-
and there needs to be consideration of both general mented and modified, while another relates to
characteristics of landscapes and more specific ele- whether landscape modification has resulted in
ments relating to specific cases. General features of broad-scale changes in landscape physical processes,
many systems seem to be the potential for the sys- such as hydrology. If the landscape has crossed a
tem to exist in a number of different states, and the biotic threshold, restoration needs to aim at restor-
likelihood that restoration thresholds exist, which ing connectivity. If, on the other hand, a physical
prevent the system from returning to a less- threshold has been crossed, this needs to be treated
degraded state without the input of management ef- as a priority. Hence, for instance, in a fragmented
fort (Aronson et al., 1993; Scheffer et al., 1993; Milton forested landscape, the primary goal may be the pro-
et al., 1994; Hobbs & Norton, 1996). Whisenant (1999) vision of additional habitat or re-establishing con-
has recently suggested that two main types of such nectivity for particular target species, whereas in a
threshold are likely, one which is caused by biotic in- modified river or wetland system, the primary need
teractions, and the other caused by abiotic limita- may be to re-establish water flows (Middleton, 1999).
tions. The type of restoration response needed will Restoration activities required to overcome particu-
depend on which, if any, thresholds have been lar physical changes may also act to overcome biotic
crossed (Fig. 3.2a). If the system has degraded mainly thresholds. An example of this would be where ex-
due to biotic changes (such as grazing-induced tensive revegetation is required to counteract hydro-
changes in vegetation composition), restoration ef- logical imbalances, and at the same time can have a
forts need to focus on biotic manipulations which positive impact on biotic connectivity (Hobbs, 1993a;
remove the degrading factor (e.g. the grazing ani- Hobbs et al., 1993).
mal) and adjust the biotic composition (e.g. replant If we accept that different types of thresholds in
desired species). If, on the other hand, the system landscape function are possible, a number of im-
has degraded due to changes in abiotic features portant questions have to be asked in terms of
(such as through soil erosion or contamination), restoration. First, does the threshold work the same
restoration efforts need to focus first on removing way on the way up as it did on the way down, or is
the degrading factor and repairing the physical there a hysteresis effect? In other words, in a
and/or chemical environment. landscape where habitat area is being increased,
In the latter case, there is little point in focusing will species return to the system at the same rate as
on biotic manipulation without first tackling the they dropped out when habitat was being lost?
abiotic problems. In other words, it is important that Second, what happens when pattern and process
system functioning is corrected or maintained before are not tightly linked? For instance, studies in cen-
biotic composition and structure are considered. tral Europe have illustrated the important role of

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
34 RICHARD HOBBS

(a) Transition threshold Transition threshold


controlled by biotic interactions controlled by abiotic limitations

Level of
system
function/
health
1
Fully 2
functional

3
Recovery 4 Recovery requires
requires modification of the
improved physical environment
management
Recovery
requires vegetation 5
manipulation 6

Non-functional

Intact Degraded

Ecosystem state

Transition threshold Transition threshold


(b) controlled by loss of controlled by loss of physical
biotic connectivity landscape function

Level of
landscape
function/
health
1
Fully 2
functional

3
Recovery/ 4 Recovery requires
landscape-scale
maintenance
modification of the
requires
physical environment
improved
management Recovery
of retained requires habitat 5
habitat manipulation and 6
replacement

Non-functional

0% Proportion of habitat destroyed 100%


Low Degree of habitat modification High

Landscape state

Fig. 3.2. (a) Conceptual model of system transitions between states of varying levels of
function, illustrating the presence of two types of restoration threshold, one controlled by
biotic interactions and one controlled by abiotic limitations. Adapted from Whisenant (1999).
(b) A similar model applied to landscapes, indicating transition thresholds controlled by loss
of biotic connectivity and loss of physical landscape function. From Hobbs & Harris (2001).

traditional management involving seasonal move- tion of this process, and restoration in this case
ment of sheep between pastures in dispersing seeds will not involve any modification of the existing
around the landscape (Bakker et al., 1996; Fischer et al., landscape pattern – rather it will entail the rein-
1996; Poschlod et al., 1996). The long-term viability of statement of a management-mediated process of
some plant species may be threatened by the cessa- sheep movement. Hence, correct assessment of the

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
A landscape perspective 35

problem and its cause and remedy require careful ex- 3. Design repair activities at the proper scale.
amination of the system and its components rather 4. Design landscapes to increase retention of limiting
than generalised statements of prevailing dogma. resources.
5. Design spatial variation into landscapes.
6. Design landscapes to maintain the integrity of
DEVELOPING GUIDELINES FOR primary processes.
LANDSCAPE-SCALE RESTORATION 7. Design linkages into landscapes.
Several approaches are possible to developing guide- 8. Design propagule donor patches into landscapes.
lines and options for landscape restoration. A 9. Design landscapes to encourage animal dispersal of
hierarchy of guidelines can be considered, depend- desired seed.
ing on the scope and goals of the restoration and 10. Design landscapes to encourage wind dispersal of
management, and the level and detail of informa- desired seed.
tion available for making decisions. This can be set 11. Design landscapes to encourage positive animal
out as follows: interactions.

1. General guidelines which are applicable in most While these guidelines were primarily directed at
types of landscape (Whisenant, 1999). semi-arid areas, they may have broader application.
2. Guidelines which are related to particular broad Whisenant (1999) suggests that the guidelines
categories of landscape (McIntyre & Hobbs, 1999, in provide a framework for considering landscape in-
press). teractions, but recognises that they do not produce
3. Guidelines which are derived from consideration of specific quantitative designs. It is important that we
specific species or suites of species in a given land- attempt to move from these generalised guidelines
scape (Lambeck, 1997). to more specific recommendations for particular
4. Spatially specific options which are developed for types of landscape, and finally to spatially explicit
particular landscapes in relation to existing land- options for individual landscapes. Hence, I now
scape structure, target species and specified local consider attempts to categorise broad landscape
management goals. types and identify the priority activities in each, and
then examine options for developing spatially ex-
plicit landscape restoration plans.
General guidelines
How do you go about determining how to conduct Guidelines for broad landscape types
restoration at a landscape scale? What sort of
If we consider fragmented landscapes as an exam-
landscape-level management and restoration is appro-
ple, priority actions are likely to involve: (1) the pro-
priate for different landscapes? A set of general princi-
tection of existing habitat patches; (2) their effective
ples, derived from island biogeography theory, sug-
management; and (3) restoration both within
gest that bigger patches are better than small patches,
patches and at a broader landscape scale. But where
connected patches are better than unconnected, and
do we go from there? Which are the priority areas to
so on. For fragmented landscapes, such principles can
retain? Should we concentrate on retaining the ex-
be translated into the need to retain existing habitat
isting fragments or on restoration, and relatively
patches, especially large ones, and existing connec-
how many resources (financial, manpower, etc.)
tions, and to revegetate in such a way as to provide
should go into each? How much restored habitat is
larger patches and more connections (Hobbs, 1993a).
required, and in what configuration? When should
Whisenant (1999) provided a set of guidelines for land-
we concentrate on providing corridors versus addi-
scape repair, which included the following:
tional habitat? If we are to make a significant impact
1. Treat causes rather than symptoms. in terms of conserving remaining fragments and as-
2. Emphasise process repair over structural replace- sociated fauna, these questions need to be addressed
ment. in a strategic way.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
36 RICHARD HOBBS

McIntyre & Hobbs (1999) have examined these Habitat modification alters the condition of the
questions in terms of the range of human impacts remaining habitat, and can occur in any of the situ-
on landscapes. They identified four broad types of ations illustrated in Table 3.1. Modification acts to
landscapes (Table 3.1, Fig. 3.3), with intact and relict- create a layer of variation in the landscape over and
ual landscapes at the extremes, and two intermedi- above the straightforward spatial patterning caused
ate states, variegated and fragmented. In variegated by vegetation destruction. There is a tendency for
landscapes, the habitat still forms the matrix, habitats to become progressively more modified
whereas in fragmented landscapes, the matrix com- with increasing levels of destruction, owing to the
prises ‘destroyed habitat’. Each of the four levels progressively greater proportion of edge in remain-
described in Table 3.1 is associated with a particular ing habitats.
degree of habitat destruction, and the categories are The framework in Table 3.1 can assist in deciding
not entirely arbitrary. For instance, the distinction where on the landscape to allocate greater and
between variegated and fragmented landscapes re- lesser efforts towards different management actions
flects suggestions discussed earlier that landscapes (McIntyre & Hobbs, in press). Three types of action
in which habitats persist over more than 60% of the could be applied to habitats for their conservation
area are operationally not fragmented, since they management:
consist of a continuous cluster of habitat. This broad
division can be regarded as a ‘first cut’, and the 1. Maintain the existing condition of habitats by re-
provision of names for each category is for conven- moving and controlling threatening processes. It is
ience rather than to set up a rigid classification. generally much easier to avoid the effects of degra-
Further investigation is required to test these cate- dation than it is to reverse them.
gories and to examine the need for further subcate- 2. Improve the condition of habitats by reducing or
gories. For instance, functionally different types of removing threatening processes. More active man-
‘fragmented’ landscapes could be recognised. agement may be needed to initiate a reversal of

Table 3.1. Four landscape states defined by the degree of habitat destruction

Degree of destruction Degree of Pattern of


of habitat Connectivity a of modificationb of modificationb of
Landscape type (% remaining) remaining habitat remaining habitat remaining habitat

Intact Little or none High Generally low Mosaic with


(⬎90%) gradients
Variegated Moderate Generally high but Low to high Mosaic which may
(60–90%) lower for species have both gradients
sensitive to habitat and abrupt
modification boundaries
Fragmented High Generally low but Low to high Gradients within
(10–60%) varies with mobility fragments less
of species and evident
arrangement on
landscape
Relictual Extreme None Generally highly Generally uniform
(⬍10%) modified

a
From Pearson et al. (1996).
b
Modified from McIntyre & Hobbs (in press).

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
A landscape perspective 37

(a) Intact Variegated Fragmented Relictual

% destroyed <10% 10-40% 40-90% >90%

Habitat Destroyed

(b)

Unmodified Modified

Highly modified Destroyed

Fig. 3.3. (a) Four levels of habitat destruction, each characterised by a range of proportions of
habitat destroyed. (b) Pattern of habitat modification overlying landscape patterns of habitat
destruction depicted in (a). Although any combination of destruction and modification levels
is theoretically possible, those considered to be typical of different destruction levels are
illustrated. Modified from McIntyre and Hobbs (1999, in press).

condition (e.g. removal of exotic species, reintroduc- by rehabilitating degraded areas, and expanding habi-
tion of native species) in highly modified habitats. tat by revegetating to create larger blocks and restore
3. Reconstruct habitats where their total extent has been poorly represented habitats. The first priority is the
reduced below viable size using replanting and re- maintenance of elements which are currently in good
introduction techniques. Here reconstruction im- condition. This will be predominantly the vegetated
plies the return of key habitat elements to areas matrix in intact and variegated landscapes and the
where habitat has previously been completely re- remnants which remain in good condition in
moved. As this is so difficult and expensive, it is a last fragmented landscapes. There may well be no rem-
resort action that is most relevant to fragmented and nants left in good condition in relict landscapes.
relictual landscapes. We have to recognise that Maintenance will involve ensuring the continuation
restoration often does not come close to restoring of population, community and ecosystem processes
habitats to their unmodified state, and this rein- which result in the persistence of the species and com-
forces the wisdom of maintaining existing ecosys- munities present in the landscape. Maintaining frag-
tems as a priority. ments in good condition in a fragmented system may
also require activities in the matrix to control land-
The next stage is to link these activities to specific scape processes, such as hydrology. For instance,
landscape components (matrix, connecting areas, numerous examples have been documented where
buffer areas, fragments) in which they would be most maintenance of habitat patches depends on the effec-
effective, and to determine priorities for management tive management of surface and subsurface water
action in different landscape types. A general ap- movement in the surrounding landscape (Rowell,
proach might be to build on strengths of the remain- 1986; Barendregt et al., 1995; George et al., 1995).
ing habitat by filling in gaps and increasing landscape The second priority is the improvement of elements
connectivity, increasing the availability of resources that have been modified in some way. In variegated

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
38 RICHARD HOBBS

landscapes, buffer areas and corridors may be a prior- tion goals are for the area. Lambeck (1997) has re-
ity, while in fragmented systems, improving the sur- cently contended that more efficient solutions to
rounding matrix to reduce threatening processes will conservation problems can be developed if we take a
be a priority, as indicated above. In relict landscapes, strategic approach rather than a generalised one.
improving the condition of fragments will be essential This involves developing a clear set of conservation
for their continued persistence. Improvement may in- objectives rather than relying on vague statements
volve simply dealing with threatening processes such of intent. One set of objectives relates to the achieve-
as stock grazing or feral predators, or may involve ac- ment of a comprehensive, adequate and representa-
tive management to restore ecosystem processes, im- tive set of reserves or protected area networks. An-
prove soil structure, encourage regeneration of plant other, complementary set of objectives relate to the
species, or reintroduce flora or fauna species formerly adequacy of the existing remnant vegetation (not
present there (Hobbs & Yates, 1997). only reserves). The process of setting conservation
Reconstruction is likely to be necessary only in frag- objectives in any given area can be simplified by
mented and relict areas. Primary goals of reconstruc- identifying a set of key or ‘focal’ species which are
tion will be to provide buffer areas around fragments, most at risk from the main threats identified in the
to increase connectivity with corridors, and to provide area, in essence a multi-species indicator/umbrella
additional habitat (Hobbs, 1993a). While some basic species approach.
principles of habitat reconstruction have been put for- To identify focal species, Lambeck (1997) recog-
ward, the benefits of such activities have rarely been nised three distinct sets of species, each of which
quantified. Questions remain about which character- were likely to be limited or threatened by particular
istics of ‘natural’ habitat are the most important to try characteristics of the landscape (Fig. 3.4). These were:
to incorporate into reconstruction, and what land-
scape configurations are likely to be most effective. 1. Area or habitat limited species, i.e. species whose
In order to answer such questions, it becomes numbers are limited by the availability of large
very important to specify clearly what the conserva- enough patches of suitable habitat.

Survey
Yes No further action
Population secure?
No

Area- Dispersal- Resource- Management-limited


limited limited limited (Fire, Predators, Weeds, Grazing)

Define patch
attributes Define
connectivity
attributes
Define
compositional
requirements

Design Guidelines Management Guidelines

Implementation
Monitoring

Fig. 3.4. A process for selecting focal species in any given landscape, based on whether species
are limited by area, movement or management, and resulting in landscape design guidelines
and management guidelines. Modified from Lambeck (1997).

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
A landscape perspective 39

2. Movement limited species, i.e. species whose num- sure that their landscape has patches of habitat
bers are limited by the degree to which they can equal to or exceeding the specified size, and that
move between habitat patches. these patches are separated by distances less than
3. Management limited species, i.e. species whose those specified. These patches need to be distributed
numbers are limited by processes such as predation, from one boundary of the management area to the
disturbance, fire and the like, which can be manip- other and connected by high-quality strips of linear
ulated within particular sites. habitat. Such a design aims to ensure that popula-
tions of all species are linked across the area being
Design of landscape reconstructions is based on the managed.
requirements of the most sensitive species in each of This method clearly requires a minimum of spa-
these categories. For instance, if you can identify tial landscape data and observations of the biota, and
which species have the requirement for the largest further refinement is possible with more detailed
areas of habitat, you can start assessing the ade- information. Even with the minimal data require-
quacy of the current landscape for that species, and ments, the approach is still very time-consuming,
hence all other species with less-demanding habitat and requires that each case be analysed independ-
requirements, and can also start making recommen- ently. For the more general application of the ap-
dations on where and how much habitat reconstruc- proach, it may be possible to work back from the
tion needs to be undertaken. particular to a more generalisable approach, if
landscapes with similar characteristics (biophysi-
cal and anthropogenic) can be identified and
Spatially explicit options for specific landscapes grouped. In that case, the approach adopted in one
The ability to make spatially explicit recommenda- landscape could be applied to other landscapes
tions as to where restoration activities should oc- with similar characteristics.
cur is essential if real solutions are to be developed
and implemented. Moving from generalised guide-
lines to specific options for a particular area re-
CONCLUDING REMARKS
quires that we are able to translate the generalised In this chapter I have summarised what we know
guidelines into concrete recommendations of what about how landscapes work, and how this might be
to do where. There are relatively few examples utilised for the planning and implementation of
where this translation has been accomplished suc- landscape-scale restoration. Our level of understand-
cessfully. Part of the problem lies in the fact that ing of landscape patterns and processes is increas-
every landscape is unique in terms of its biophysi- ing, and our ability to work effectively at the land-
cal characteristics and its pattern of human alter- scape scale is being facilitated by improved tools
ation. Translating a general guideline thus has to such as remote sensing and GIS.
take account of the unique characteristics of the Landscape-scale restoration is still largely in its
area being managed/restored and refer directly to infancy. The recognition of the importance of em-
the spatial realities of each situation. Outlined in barking on restoration at landscape and regional
Box 3.1 is one attempt to do so in the Western scales is increasing, and examples of landscape-
Australian wheat belt. scale projects are beginning to accumulate. How-
While this approach enables us to identify the ever, these are often still in their early stages, and it
minimum patch sizes required for species to have a will take time to assess their efficacy. Because of
reasonable probability of occurrence, and to identify the temporal and spatial scales involved, novel and
patches that were too isolated to be occupied, the re- integrated approaches need to be taken to the
sults do not ensure that populations will persist in problem. Often the only way to assess different sce-
the long term. To assess this, further more detailed narios is by computer modelling, and it is usually
research and monitoring will be required. In the impossible to design replicated landscape-scale
meantime, land managers are being advised to en- experiments. Hence, there is a need to learn as we

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
40 RICHARD HOBBS

Box 3.1 The Western Australian wheat belt characteristics of habitat patches where species do
and don’t occur. This enables the specification of the
The wheat belt of Western Australia has been minimum fragment size and the maximum
extensively cleared for agriculture, leaving native interpatch distance that is required for these species
vegetation as mostly small isolated fragments (Hobbs & to have a specified probability of occurring. It is then
Saunders, 1993; Saunders et al., 1993). While the lower possible to identify all remnants that do not meet
parts of the landscape are under threat from these criteria and specify the amount of habitat
salinisation (George et al., 1995), in the remainder of the reconstruction required to produce fragments of
landscape the primary threats are habitat loss and adequate size. Similarly, it is possible to identify all
isolation. Restoration options are currently being remnants that are too isolated for the most dispersal-
sought to deal with threats to the conservation of the limited species and identify the need for the
region’s highly diverse biota. construction of intermediate habitat patches or
Using the scheme devised by Lambeck (1997), corridors. Maps are then produced which indicate
species that are considered to be threatened by each the extent to which each patch needs to be expanded
of the threatening processes are grouped and ranked or connected in order to have an equivalent
in terms of their sensitivity. Presence/absence probability of being occupied by the species which
surveys of vegetation remnants indicate the had the greatest demand for that patch type.
distribution of the various species whose This approach has been used to determine the
populations are limited by the amount of habitat requirements for bird habitat in four watersheds in the
available or by the degree of isolation of habitat wheat belt of Western Australia, each covering an area
patches. Analysis of the spatial attributes of the of approximately 20–30 000 ha (see Wallace, 1998;
vegetation remnants in the landscape undertaken Lambeck, 1999). An example is given in Fig. B3.1.
using GIS routines then determines the

A. GRE A
T T ERN
EAS Y
HI GHWA
TAMMIN
RO AD

NE
LS O
AD
RO

S
ED
I L
RO A

LDF DI XON RO AD
GO
D

IKTT O ROGERS RO AD
T ON
RALS

N
5km

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
A landscape perspective 41

B.
Tammin

N
5km
Fig. B3.1. An example of the development of management options for a highly fragmented
landscape, based on the estimated habitat and connectivity requirements of focal bird
species. (a) The landscape as it currently exists in the South Tammin subcatchment
(watershed) in southwest Western Australia. Shaded areas are native vegetation of different
types. (b) Recommendations for revegetation to enlarge existing vegetation patches and
connect isolated patches. From Hobbs & Lambeck (in press).

go, and to build a degree of adaptability into the mation, but also to be economically possible and prac-
restoration designs. tically achievable. Almost anything is possible if there
Once the options for restoration have been derived are enough human and financial resources available,
from guidelines or more detailed ecological response but usually the cost and social desirability of forego-
models, these then have to be considered in the ing other options limits the potential extent of
broader context of individual and societal goals restoration activities. Selling the benefits of restora-
(Hobbs & Lambeck, in press). Restoration is not an al- tion is often difficult due to the relatively long time-
ternative to conservation or sound management, but frames involved in achieving anything. Discount rates
is often a necessary part of these activities. Restora- often mean that short-term costs greatly outweigh
tion at the landscape scale often also involves tack- any long-term benefits. Often another primary driver
ling multiple issues at once, and balancing conserva- in deciding which options will be pursued is the pre-
tion and production. The idea of ‘reintegrating’ vailing political climate, which drives government
landscapes embodies the idea that successful restora- support and funding for restoration activities.
tion must encompass the biophysical, social and eco- Broad-scale restoration is going to become in-
nomic realities of the situation (Aronson & Le Floc’h, creasingly necessary as humans continue to modify
1996a, b; Hobbs & Saunders, 1993). and use the earth and its resources, and we need to
To succeed, restoration activities need not only ensure that we continue to improve our ability to
to be based on sound ecological principles and infor- provide restoration options at these broad scales.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
42 RICHARD HOBBS

REFERENCES Dawson, D. (1994). Are Habitat Corridors Conduits for Animals


and Plants in a Fragmented Landscape? A Review of the
Appleton, E.L. (1995). A cross-media approach to saving the Scientific Evidence. London: English Nature.
Chesapeake Bay. Environmental Science and Technology, 29, Delcourt, H.R. & Delcourt, P.A. (1991). Quaternary Ecology:
550A–555A. A Paleoecological Perspective. New York: Chapman & Hall.
Aronson, J. & Le Floc’h, E. (1996a). Hierarchies and Fischer, S.F., Poschlod, P. & Beinlich, B. (1996). Experimental
landscape history: dialoging with Hobbs and Norton. studies on the dispersal of plants and animals on sheep
Restoration Ecology, 4, 327–333. in calcareous grasslands. Journal of Applied Ecology, 33,
Aronson, J. & Le Floc’h, E. (1996b). Vital landscape 1206–1222.
attributes: missing tools for restoration ecology. Flaig, E.G. & Reddy, K.R. (1995). Fate of phosphorus in the
Restoration Ecology, 4, 377–387. Lake Okeechobee watershed, Florida, USA: overview and
Aronson, J., Floret, C., Le Floc’h, E., Ovalle, C. & Pontanier, recommendations. Ecological Engineering, 5, 127–142.
R. (1993). Restoration and rehabilitation of degraded Flannery, T. (1994). The Future Eaters: An Ecological History of
ecosystems in arid and semiarid regions. 1: A view from the Australasian Lands and People. Port Melbourne, Vic:
the South. Restoration Ecology, 1, 8–17. Reed Books.
Baker, W.L. (1992). The landscape ecology of large Forman, R.T.T. (1995). Land Mosaics: The Ecology of Landscapes
disturbances in the design and management of nature and Regions. Cambridge: Cambridge University Press.
reserves. Landscape Ecology, 7, 181–194. Forman, R.T.T. & Godron, M. (1986). Landscape Ecology. New
Bakker, J.P., Poschlod, P., Strykstra, R.J., Bekker, R.M. & York: John Wiley.
Thompson, K. (1996). Seed banks and seed dispersal: George, R.J., McFarlane, D.J. & Speed, R.J. (1995). The
important topics in restoration ecology. Acta Botanica consequences of a changing hydrologic environment
Neerlandica, 45, 461–490. for native vegetation in south Western Australia. In
Barendregt, A., Wassen, M.J. & Schot, P.P. (1995). Hydrological Nature Conservation, vol. 4, The Role of Networks, eds.
systems beyond a nature reserve, the major problem in D.A. Saunders, J.L. Craig & E.M. Mattiske, pp. 9–22.
wetland conservation of Naardermeer (The Netherlands). Chipping Norton, NSW: Surrey Beatty.
Biological Conservation, 72, 393–405. Gosz, J.R. (1993). Ecotone hierarchies. Ecological
Bell, S.S., Fonseca, M.S. & Motten, L.B. (1997). Linking Applications, 3, 369–376.
restoration and landscape ecology. Restoration Ecology, 5, Haines-Young, R., Green, D.R. & Cousins, S. (eds.) (1993).
318–323. Landscape Ecology and GIS. London: Taylor & Francis.
Bennett, A.F. (1999). Linkages in the Landscape: The Role of Hansen, A.J. & di Castri, F.D. (eds.) (1992). Landscape
Corridors and Connectivity in Wildlife Conservation. Gland, Boundaries: Consequences for Biotic Diversity and Ecological
Switzerland: IUCN. Flows. New York: Springer-Verlag.
Boecklen, W.J. (1997). Nestedness, biogeographic theory, Hanski, I.A. & Gilpin, M.E. (eds.) (1997). Metapopulation
and the design of nature reserves. Oecologia, 112, 123–142. Biology: Ecology, Genetics, and Evolution. New York:
Boggess, C.F., Flaig, E.G. & Fluck, R.C. (1995). Phosphorus Academic Press.
budget–basin relationships for Lake Okeechobee Hanski, I. & Simberloff, D. (1997). The metapopulation
tributary basins. Ecological Engineering, 5, 143–162. approach, its history, conceptual domain, and application
Cairns, J.J., McCormick, P.V. & Niederlehner, B.R. (1993). to conservation. In Metapopulation Biology: Ecology, Genetics,
A proposed framework for developing indicators of and Evolution, eds. I.A. Hanski & M.E. Gilpin, pp. 5–26.
ecosystem health. Hydrobiologia, 263, 1–44. New York: Academic Press.
Cale, P. & Hobbs, R.J. (1994). Landscape heterogeneity Hanski, I., Pöyry, J. Pakkala, T. & Kuussaari, M. (1995).
indices: problems of scale and applicability, with Multiple equilibria in metapopulation dynamics. Nature,
particular reference to animal habitat description. 377, 618–621.
Pacific Conservation Biology, 1, 183–193. Hargis, C.D., Bissonette, J.A. & David, J.L. (1998). The
Costanza, R., Norton, B.G. & Haskell, B.D. (eds.) (1992). behaviour of landscape metrics commonly used in the
Ecosystem Health: New Goals for Environmental Management. study of habitat fragmentation. Landscape Ecology, 13,
Washington, DC: Island Press. 167–186.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
A landscape perspective 43

Harris, L.D. (1984). The Fragmented Forest: Island Biogeographic Hobbs, R.J. & Yates, C.J. (1997). Moving from the general
Theory and the Preservation of Biotic Diversity. Chicago, IL: to the specific: remnant management in rural Australia.
University of Chigago Press. In Frontiers in Ecology: Building the Links, eds. N. Klomp &
Hill, C.J. (1995). Linear strips of rain forest vegetation as I. Lunt, pp. 131–142. London: Elsevier Science.
potential dispersal corridors for rain forest insects. Hobbs, R.J., Saunders, D.A. & Arnold, G.W. (1993).
Conservation Biology, 9, 1559–1566. Integrated landscape ecology: a Western Australian
Hobbs, R.J. (1987). Disturbance regimes in remnants of perspective. Biological Conservation, 64, 231–238.
natural vegetation. In Nature Conservation: The Role of Holland, M.M., Risser, P.G. & Naiman, R.J. (eds.) (1991).
Remnants of Native Vegetation, eds. D.A. Saunders, G.W. Ecotones: The Role of Landscape Boundaries in the Management
Arnold, A.A. Burbridge & A.J.M. Hopkin, pp. 233–240. and Restoration of Changing Environments. New York:
Chipping Norton, NSW: Surrey Beatty. Chapman & Hall.
Hobbs, R.J. (1992). Corridors for conservation: solution or Hornung, M. & Reynolds, B. (1995). The effects of natural
bandwagon? Trends in Ecology and Evolution, 7, 389–392. and anthropogenic environmental changes on ecosystem
Hobbs, R.J. (1993a). Effects of landscape fragmentation on processes at the catchment scale. Trends in Ecology and
ecosystem processes in the Western Australian wheatbelt. Evolution, 10, 443–449.
Biological Conservation, 64, 193–201. Hudson, W.E. (ed.) (1991). Landscape Linkages and Biodiversity.
Hobbs, R.J. (1993b). Can revegetation assist in the Washington, DC: Island Press.
conservation of biodiversity in agricultural areas? Kapos, V. (1989). Effects of isolation on the water status of
Pacific Conservation Biology, 1, 29–38. forest patches in the Brazilian Amazon. Journal of Tropical
Hobbs, R.J. (1994). Dynamics of vegetation mosaics: can Ecology, 5, 173–85.
we predict responses to global change? Ecoscience, 1, Knight, D.H. (1987). Parasites, lightning, and the
346–356. vegetation mosaic in wilderness landscapes. In Landscape
Hobbs, R.J. (1995). Landscape ecology. In Encyclopedia of Heterogeneity and Disturbance, ed. M.G. Turner, pp. 59–83.
Environmental Science, vol. 2, ed. W.A. Nierenberg, New York: Springer-Verlag.
pp. 417–428. San Diego, CA: Academic Press. Kolasa, J. & Waltho, N. (1998). A hierarchical view of
Hobbs, R.J. (1999). Restoration ecology and landscape habitat and its relationship to species abundance. In
ecology. In Issues in Landscape Ecology, eds. J.A. Wiens & Ecological Scale: Theory and Applications, eds. D. Peterson &
M.R. Moss, pp. 70–77. Guelph, Ontario: International V.T. Parker, pp. 55–76. New York: Columbia University
Association of Landscape Ecology. Press.
Hobbs, R.J. & Harris, J.A. (2001). Restoration Ecology: Lambeck, R.J. (1997). Focal species: a multi-species
repairing the Earth’s ecosystems in the new millenium. umbrella for nature conservation. Conservation Biology,
Restoration Ecology, 9, 239–246. 11, 849–856.
Hobbs, R.J. & Lambeck, R.L. (in press). Landscape science and Lambeck, R.J. (1999). Landscape Planning for Biodiversity
management in Western Australia. In Integrating Landscape Conservation in Agricultural Regions, Biodiversity Technical
Ecology into Natural Resource Management, eds. J. Liu & Paper no. 2. Canberra: Department of the Environment
W.W. Taylor. Cambridge: Cambridge University Press. and Heritage.
Hobbs, R.J. & Norton, D.A. (1996). Towards a conceptual Laurance, W.F. & Bierregaard, R.O. (eds.) (1997). Tropical Forest
framework for restoration ecology. Restoration Ecology, 4, Remnants: Ecology, Conservation and Management of Fragmented
93–110. Communities. Chicago, IL: University of Chicago Press.
Hobbs, R.J. & Saunders, D.A. (eds.) (1993). Reintegrating Laurance, W.F., Laurance, S.G., Ferreira, L.V., Rankin-de
Fragmented Landscapes: Towards Sustainable Production and Merona, J.M., Gascon, C. & Lovejoy, T.E. (1997). Biomass
Conservation. New York: Springer-Verlag. collapse in Amazonian forest fragments. Science, 278,
Hobbs, R.J. & Wilson, A.-M. (1998). Corridors: theory, 1117–1118.
practice and the achievement of conservation le Houérou, H.N. (1981). Impact of man and his animals
objectives. In Key Concepts in Landscape Ecology, eds. on Mediterranean vegetation. In Mediterranean-Type
J.W. Dover & R.G.H. Bunce, pp. 265–279. Preston, UK: Shrublands, eds. F. di Castri, D.W. Goodall & R.L. Specht,
International Association for Landscape Ecology. pp. 479–521. Amsterdam: Elsevier.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
44 RICHARD HOBBS

Lomolino, M.V. (1994). An evaluation of alternative Pickett, S.T.A. & Cadenasso, M.L. (1995). Landscape ecology:
strategies for building networks of nature reserves. spatial heterogeneity in ecological systems. Science, 269,
Biological Conservation, 69, 243–249. 331–334.
Ludwig, J., Tongway, D., Freudenberger, D., Noble, Pickett, S.T.A. & Thompson, J.N. (1978). Patch dynamics and
J. & Hodgkinson, K. (eds.) (1997). Landscape Ecology, the design of nature reserves. Biological Conservation, 13,
Function and Management: Principles from Australia’s 27–37.
Rangelands. Melbourne, Vic: CSIRO Publishing. Poschlod, P., Bakker, J., Bonn, S. & Fischer, S. (1996). Dispersal
MacMahon, J.A. (1998). Empirical and theoretical ecology of plants in fragmented landscapes. In Species Survival in
as a basis for restoration: an ecological success story. In Fragmented Landscapes, vol. 35, eds. J. Settele, C. Margules,
Successes, Limitations, and Frontiers in Ecosystem Science, eds. P. Poschlod & K. Henle, pp. 123–127. Dordrecht: Kluwer.
M.L. Pace & P.M. Groffman. New York: Springer-Verlag. Rapport, D.J., Costanza, R. & McMichael, A.J. (1998).
Malle, K.-G. (1996). Cleaning up the River Rhine. Scientific Assessing ecosystem health. Trends in Ecology and
American, 274, 70–73. Evolution, 13, 397–402.
Matlack, G.R. (1993). Microenvironment variation within Rowell, T.A. (1986). The history of drainage at Wicken
and among forest edge sites in the eastern United States. Fen, Cambridgeshire, England, and its relevance to
Biological Conservation, 66, 185–194. conservation. Biological Conservation, 35, 111–142.
McCulloch, D.R. (ed.) (1996). Metapopulations and Wildlife Saunders, D.A., Hobbs, R.J. & Margules, C.R. (1991). Biological
Conservation. Washington, DC: Island Press. consequences of ecosystem fragmentation: a review.
McIntyre, S. & Hobbs, R.J. (1999). A framework for Conservation Biology, 5, 18–32.
conceptualizing human impacts on landscapes and its Saunders, D.A., Hobbs, R.J. & Arnold, G.W. (1993). The
relevance to management and research. Conservation Kellerberrin project on fragmented landscapes: a review
Biology, 13, 1282–1292. of current information. Biological Conservation, 64,
McIntyre, S. & Hobbs, R.J. (in press). Human impacts 185–192.
on landscapes: matrix condition and management Scheffer, M., Hosper, S.H., Meijer, M.-L., Moss, B. &
priorities. In Nature Conservation, vol. 5, Nature Jeppesen, E. (1993). Alternative equilibria in shallow
Conservation in Production Environments, eds. J. Craig, lakes. Trends in Ecology and Evolution, 8, 275–279.
D.A. Saunders & N. Mitchell. Chipping Norton, NSW: Scheiner, S.M. (1992). Measuring pattern diversity. Ecology,
Surrey Beatty. 73, 1860–1867.
Middleton, B. (1999). Wetland Restoration: Flood Pulsing and Schwartz, M.W. (ed.) (1997). Conservation in Highly
Disturbance Dynamics. New York: John Wiley. Fragmented Landscapes. New York: Chapman & Hall.
Milton, S.J., Dean, W.R.J., du Plessis, M.A. & Siegfried, Shrader-Frechette, K.S. (1994). Ecosystem health: a new
W.R. (1994). A conceptual model of arid rangeland paradigm for ecological assessment. Trends in Ecology
degradation: the escalating cost of declining and Evolution, 9, 456–457.
productivity. BioScience, 44, 70–76. Simberloff, D.S. (1982). Island biogoegraphic theory and
Murcia, C. (1995). Edge effects in fragmented forests: the design of wildlife refuges. Ékologiya, 4, 3–13.
implications for conservation. Trends in Ecology and Simberloff, D.S. & Cox, J. (1987). Consequences and costs of
Evolution, 10, 58–62. conservation corridors. Conservation Biology, 1, 63–71.
Ostfeld, R.S., Pickett, S.T.A., Shachak, M. & Likens, G.E. Simberloff, D.S., Farr, J.A., Cox, J. & Mehlman, D.W. (1992).
(1997). Defining the scientific issues. In The Ecological Basis Movement corridors: conservation bargains or poor
of Conservation: Heterogeneity, Ecosystems and Biodiversity, eds. investments. Conservation Biology, 6, 493–504.
S.T.A. Pickett, R.S. Ostfeld, M. Shachak & G.E. Likens, Sprugel, D.G. (1991). Disturbance, equilibrium, and
pp. 3–10. New York: Chapman & Hall. environmental variability: what is ‘natural’ vegetation
Pearson, S.M., Turner, M.G., Gardner, R.H. & O’Neill, R.V. in a changing environment? Biological Conservation,
(1996). An organism-based perspective of habitat 58, 1–18.
fragmentation. In Biodiversity in Managed Landscapes: Suffling, R. (1993). Induction of vertical zones in subalpine
Theory and Practice, eds. R.C. Szaro & D.W. Johnston, valley forests by avalanche -formed fuel breaks. Landscape
pp. 77–95. New York: Oxford University Press. Ecology, 8, 127–138.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
A landscape perspective 45

Taylor, P.D., Fahrig, L., Henein, K. & Merriam, G. (1993). Humphries, H.C. (1991). Pattern and scale: statistics
Connectivity as a vital element of landscape structure. for landscape ecology. In Quantitative Methods in Landscape
Oikos, 68, 571–573. Ecology: The Analysis and Interpretation of Landscape
Thirgood, J.V. (1981). Man and the Mediterranean Forest: Heterogeneity, vol. 82, eds. M.G. Turner & R.H. Gardner,
A History of Resource Depletion. London: Academic Press. pp. 17–49. New York: Springer-Verlag.
Tongway, D.J. & Ludwig, J.A. (1996). Rehabilitation of Wallace, K.J. (ed.) (1998). Dongolocking Pilot Planning Project for
semi-arid landscapes in Australia. 1: Restoring Remnant Vegetation, Final Report (Phase 1). Perth, WA:
productive soil patches. Restoration Ecology, 4, 388–397. Department of Conservation and Land Management.
Turner, M.G. (1987). Landscape Heterogeneity and Disturbance. Whisenant, S.G. (1999). Repairing Damaged Wildlands:
New York: Springer-Verlag. A Process-Orientated, Landscape-Scale Approach. Cambridge:
Turner, M.G. (1989). Landscape ecology: the effect of Cambridge University Press.
pattern on process. Annual Review of Ecology and Wiens, J.A. (1997). Metapopulation dynamics and landscape
Systematics, 20, 171–197. ecology. In Metapopulation Biology: Ecology, Genetics, and
Turner, M.G. & Gardner, R.H. (eds.) (1991). Quantitative Evolution, eds. I. A. Hanski & M. E. Gilpin, pp. 43–62. New
Methods in Landscape Ecology: The Analysis and Interpretation York: Academic Press.
of Landscape Heterogeneity, vol. 82. New York: Willis, E.O. (1984). Conservation, subdivision of reserves, and
Springer-Verlag. the anti-dismemberment hypothesis. Oikos, 42, 396–401.
Turner, M.G., Romme, W.H., Gardner, R.H., O’Neill, R.V. & Wilson, A.-M. & Lindenmayer, D. (1996). Wildlife Corridors
Kratz, T.K. (1993). A revised concept of landscape and the Conservation of Biodiversity: A Review. Canberra:
equilibrium: disturbance and stability on scaled Greening Australia Ltd.
landscapes. Landscape Ecology, 8, 213–227. With, K.A. (1997). The application of neutral landscape
Turner, M.G., Gardner, R.H. & O’Neill, R.V. (1995). Ecological models in conservation biology. Conservation Biology, 11,
dynamics at broad scales. Ecosystems and landscapes. 1069–1080.
BioScience, Supplement 1995, S29–S35. With, K.A. & Crist, T.O. (1995). Critical thresholds in species’
Turner, S.J., O’Neill, R.V., Conley, W., Conley, M.R. & responses to landscape structure. Ecology, 76, 2446–2459.

Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005
Downloaded from https://www.cambridge.org/core. University of Toledo, on 30 Nov 2019 at 09:48:22, subject to the Cambridge Core terms of use, available at
Cambridge Books Online © Cambridge University Press, 2009
https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9780511549984.005

You might also like