zooplankton (Verity and Smetacek 1996). The
striking example of the Black Sea illustrates both
the complexity of processes driving cascading
ecosystem shifts and the potentially high stakes
for human society. Analysis of four decades of
time-series data revealed patterns consistent
with a cascade through five levels, from predatory
fishes, to planktivorous fishes, to zooplankton, to
phytoplankton, to water-column nutrient stocks
(Daskalov 2002; Daskalovet al. 2007). A major
shift from a clear-water phase supporting abundant
large fish to a turbid phase began in the early 1970s
after industrial fishing depleted apex predators
(Daskalov 2002), a situation strongly reminiscent
of phase shifts in north temperate lakes (Scheffer
and Jeppesen 2007). Because these changes also
coincided with eutrophication and the invasion of
an exotic predatory ctenophore, mass balance mod-
els were developed to evaluate the relative impor-
tance of bottom-up and top-down mediation of
these changes. The model simulations produced
cascading changes in biomass of lower trophic
levels quite similar to the observed pattern when
predators were removed, whereas simulated eutro-
phication produced biomass increases across all
levels, in contrast to observed patterns (Daskalov
2002). In this system, then, it appears that restora-
tion of predatory fishes should be at least as effec-
tive in restoring water quality as reducing nutrient
loading. Negative correlations across trophic levels,
from predatory fishes through zooplankton to phy-
toplankton and even down to water-column nutri-
ent stocks, in time series from the North Pacific
(Shiomotoet al. 1997) and coastal north Atlantic
Oceans (Franket al. 2005), suggest that cascading
trophic interactions can occur even in open pelagic
ecosystems.
The question remains whether these patterns are
general. Micheli (1999) conducted a meta-analysis
of marine pelagic systems, using data from both
mesocosm experiments and time series from
unmanipulated systems to ask whether nutrient
loading and predation penetrated through the
food chain. Although both zooplanktivorous
fishes and nutrient loading significantly affect the
adjacent trophic level (zooplankton and phyto-
plankton respectively), these effects attenuated
rapidly through the food chain. Trophic cascades
thus seemed to be the exception rather than the
rule in marine pelagic ecosystems. But the story
maybemorecomplex.Detailedanalysisofexperi-
ments found that removal of predators in the ma-
rine pelagic frequently did cascade to affect
phytoplankton biomass, but that the sign of pred-
ator influence on phytoplankton depends on
foodchainlength,whichinturndependsoncell
size and thus taxonomic composition of the domi-
nant algae (Stiboret al. 2004). When data from
three- and four-link experimental food chains
were averaged, the strong influence of predators
on phytoplankton was masked. It remains uncer-
tain, however, whether these pelagic cascades are
also common in unmanipulated, open marine
systems where food chains with three (classical)
and four (microbial loop) links operate in parallel.8.4 Biodiversity and stability of marine ecosystems
8.4.1 Conceptual background
Early ecologists, from Darwin (Hector and Hooper
2002) to MacArthur (1955) and Elton (1958), be-
lieved that diverse communities were more stable
and better able to resist disturbances than depau-
perate ones. These ideas have received renewed
attention as concern about declining biodiversity
has grown (McCann 2000). There are several me-
chanisms by which biodiversity might increase
stability of community- or ecosystem-level proper-
ties. The most general is simple statistical averaging
(also called the ‘portfolio effect’): as long as tempo-
ral fluctuations of co-occurring species are not per-
fectly correlated, variance of their aggregate
abundance in response to stochastic environmental
variance will be lower than the average variability
of component species (Tilmanet al. 1997; Doaket al.
1998). Biodiversity may also stabilize ecosystem
properties against perturbations by enhancing
the system’s ability to absorb a stress without
changing (resistance) or the rapidity with which
it returns to its original state after perturbation
(resilience), through at least two biological mechan-
isms. First, niche differentiation (functional diver-
sity) among species increases the probability that at
least some species will thrive as environmental104 APPLICATIONS