8. Biodiversity and Ecosystem Processes
125
In sum, it is now possible, within a relatively simple conceptual framework, to
embrace all the likely outcomes linking changes in ecosystem processes to changes
in plant species richness within a community (Fig. 1), and to suggest that the most
likely outcome is the relationship shown in Fig. 1b and Fig. 2: that is, a plot of
biomass, or primary production etc. against increasing plant species richness will
typically be a negatively accelerating, rising curve.
2.1.4 The Insurance Hypothesis
I have developed these arguments implicitly assuming a relatively benign world,
with communities exposed to average environmental conditions. How might species richness affect ecosystem processes under extreme events? The theoretical
answer again rests on niche differences between species. By exactly the same arguments developed in Section 2.1.2, niche differences between species 'spread risk',
so that more species rich assemblages ought to be more resilient to extreme events
(drought, late frost etc.) than species poor assemblages (Walker 1992; Lawton and
Brown 1993). The effect has been variously called the insurance hypothesis, or the
portfolio effect (from the analogous process of risk spreading by investing in a
range of stocks and shares) and has strong theoretical support (Naeem 1998; Tilman
et al. 1998).
A key point of the insurance hypothesis is that species that appear to be 'redundant' under more benign conditions (Fig. la, and the 'plateau' at higher species
richness in Fig. 1b) may play an essential role in maintaining ecosystem processes
under extreme events. In other words, more species are required to maintain ecosystem processes at a given level in a variable environment than in a constant one
(Fig. 2c). Determining how different kinds of environmental disturbances interact
with one-another, and with species richness, to influence ecosystem processes is an
unresolved theoretical and experimental problem.
2.1.5 Within System and Between System Comparisons
There is a crucial difference between the impacts of changes in biodiversity on
ecosystem processes within an ecosystem (within-system effects), and comparisons
between different ecosystems (between-system effects) (Fig. 3). So far, I have been
talking only about within-system effects, where the theoretical problem being addressed is about loss of species within one community, at one locality. Several
experiments have now been carried out to test this aspect of theory (Section 2.2)
(e.g. Naeem et al. 1994, 1995; Tilman and Downing 1994; Hooper and Vitousek
1997, 1998; McGrady-Steed et al. 1997; Naaem and Li 1997; Tilman et al. 1996,
1997b).
An alternative approach is to resort to correlative field studies on natural vegetation (e.g. Wardle et al. 1997a,b), using different sites to examine the relationship between biodiversity and ecosystem processes - that is to make betweensystem comparisons. This approach tests a fundamentally different problem, because it simultaneously examines changes in species richness and differences be-
125
In sum, it is now possible, within a relatively simple conceptual framework, to
embrace all the likely outcomes linking changes in ecosystem processes to changes
in plant species richness within a community (Fig. 1), and to suggest that the most
likely outcome is the relationship shown in Fig. 1b and Fig. 2: that is, a plot of
biomass, or primary production etc. against increasing plant species richness will
typically be a negatively accelerating, rising curve.
2.1.4 The Insurance Hypothesis
I have developed these arguments implicitly assuming a relatively benign world,
with communities exposed to average environmental conditions. How might species richness affect ecosystem processes under extreme events? The theoretical
answer again rests on niche differences between species. By exactly the same arguments developed in Section 2.1.2, niche differences between species 'spread risk',
so that more species rich assemblages ought to be more resilient to extreme events
(drought, late frost etc.) than species poor assemblages (Walker 1992; Lawton and
Brown 1993). The effect has been variously called the insurance hypothesis, or the
portfolio effect (from the analogous process of risk spreading by investing in a
range of stocks and shares) and has strong theoretical support (Naeem 1998; Tilman
et al. 1998).
A key point of the insurance hypothesis is that species that appear to be 'redundant' under more benign conditions (Fig. la, and the 'plateau' at higher species
richness in Fig. 1b) may play an essential role in maintaining ecosystem processes
under extreme events. In other words, more species are required to maintain ecosystem processes at a given level in a variable environment than in a constant one
(Fig. 2c). Determining how different kinds of environmental disturbances interact
with one-another, and with species richness, to influence ecosystem processes is an
unresolved theoretical and experimental problem.
2.1.5 Within System and Between System Comparisons
There is a crucial difference between the impacts of changes in biodiversity on
ecosystem processes within an ecosystem (within-system effects), and comparisons
between different ecosystems (between-system effects) (Fig. 3). So far, I have been
talking only about within-system effects, where the theoretical problem being addressed is about loss of species within one community, at one locality. Several
experiments have now been carried out to test this aspect of theory (Section 2.2)
(e.g. Naeem et al. 1994, 1995; Tilman and Downing 1994; Hooper and Vitousek
1997, 1998; McGrady-Steed et al. 1997; Naaem and Li 1997; Tilman et al. 1996,
1997b).
An alternative approach is to resort to correlative field studies on natural vegetation (e.g. Wardle et al. 1997a,b), using different sites to examine the relationship between biodiversity and ecosystem processes - that is to make betweensystem comparisons. This approach tests a fundamentally different problem, because it simultaneously examines changes in species richness and differences be-