landscape consisting of habitat patches of varying
quality, and a matrix of varying but lower quality
between patches. Crucial for this view being accept-
ed was the accumulation of observations that pat-
terns in local diversity and local interactions could
not be fully explained by local processes and local
environmental conditions (e.g. Ricklefs and Schlu-
ter 1993; Poliset al. 1997; Holyoaket al. 2005). This
led a new generation of ecologists to design studies
in which the effects of local conditions and regional
dynamics or landscape composition could be sepa-
rated more rigorously. Many such studies tested
very applied questions. Some examples are the im-
portance of the amount of natural habitats in the
surrounding landscape for parasitoids on pest in-
sects in crops (Thies and Tscharntke 1999), and the
effects of different agricultural practices and land-
scape heterogeneity on insect diversity (Weibull
et al. 2000). In both these as well as many other
cases, the effects of landscape were significant and
sometimes larger than effects of local management.
9.2.2 Implications of metacommunity theory
Allowing theoretical predictions to guide manage-
ment is risky. Many applied questions in ecology
require both knowledge of ecological theory and
intimate understanding of the natural history of
the system under study. Furthermore, for many
applied questions the system that needs to be un-
derstood includes not only the ecosystem, but also
many aspects of humans as part of linked social–
ecological systems, including sociology, economy,
politics and the conflicting interests of those advo-
cating different ideological views of the world (see,
for example, Lawton 2007). Nonetheless, it is valid
to examine what the emerging metacommunity the-
ories might predict about the management of diver-
sity and ecosystem services in human-dominated
landscapes (see, for example, Bengtssonet al. 2003
in the context of reserves). An analogy from evolu-
tionary biology is useful: Predictions on evolution-
ary changes in pests or pathogens in response to
human actions are of interest, although we know
that decisions in society do not fully take these into
account.
Leiboldet al. (2004; see also Chapter 5) offered four
simplified perspectives of how metacommunity the-
ory might view the world, i.e.neutral,patch dynamics,
species sortinganddispersal-drivenmetacommunities
(also termedmass effects,but in this author’s view this
term is ambiguous and should be replaced to empha-
size the main role of dispersal for the dynamics of this
type of communities). Each of these offer interesting
insights into how diversity, species composition and
ecosystem services are affected by local (sorting) and
regional (dispersal) processes, but, as noted in Chase
and Bengtsson (Chapter 5), the four perspectives are
not exclusive. The basic processes and management
implications of the four metacommunity perspec-
tives are summarized in Table 9.1.
The metacommunity view of the world is mainly
one of discrete patches and a matrix between patches
that is non-habitable environment. The quality of the
matrix in this view mainly affects dispersal (migra-
tion rates). However, as pointed out by Vandermeer
and Perfecto (2007) the matrix is often habitat for
many species in agricultural and forest systems.
The effect of a matrix that is habitable or even similar
in quality to patches is most easily included in the
dispersal-driven(mass effects) perspective. This view of
the matrix highlights the points of Oksanen (1990)
and Holt (2002) that the relative productivity of
patches or of the matrix compared with patches is
of great importance when examining the effects of
spatial heterogeneity. Spill-over from more produc-
tive habitats or patches can greatly affect species
interactions. It may complicate management, but is
important for the delivery of ecosystem services such
as biological control and pollination (see below).
Some general insights for ecosystem management
can be deduced from these perspectives (e.g. Mouil-
lot 2007), although for particular systems and ques-
tions predictions have to be much more specific. In
the following, unless otherwise stated, communities
are assumed to be at one trophic level, structured by
competition for space or other resources, and there
is a trade-off between competitive ability and traits
associated with dispersal.
The species sorting perspective implies that
management of local conditions and local habitat
heterogeneity is most important to maintain diversi-
ty in a patch (Table 9.1). Becausespecies sortingallows
the most efficient competitors to dominate patches,118 APPLICATIONS