of human predation are consistent with expectations
arising from theoretical and experimental communi-
ty ecology, particularly regarding trophic cascades,
the influence of biodiversity on ecosystem stability
and the nature of regime shifts among alternate
semi-stable states. I explore what the impacts of
human predation on marine communities can tell
us about how and whether basic principles in com-
munity ecology extrapolate to complex ecosystems
at large spatial and temporal scales, and thus assess
their value, if any, for conservation and manage-
ment.
8.1.1 Fishing as a global experiment in community manipulation
Of the several human impacts on marine systems,
the strongest and most pervasive is the continuous
removal of large quantities of animal biomass
through fishing. In general, fishery management
aims to reduce a fish stock to50% of its unfished
biomass in order to maximize productivity. In prac-
tice, many stocks are fished well beyond this target
(Hilbornet al. 2003a; FAO 2007). Modern fisheries,
including both landings and by-catch, currently
consume 24–35% of global marine primary produc-
tion in the continental shelf and major upwelling
areas (Pauly and Christensen 1995). Thus, any at-
tempt to understand modern marine communities
must reckon with the fact that humans are now the
dominant predator throughout the world ocean.
The manydirectimpacts of human exploitation on
marine fish populations and communities are well
documented (Jennings and Kaiser 1998; Paulyet al.
1998; Jacksonet al. 2001; Myers and Worm 2003).
But this intense predation is expected to have ex-
tensiveindirecteffects on marine communities as
well. Indeed, effects of fishing provide a uniquely
useful case for testing how community models
scale up to real ecosystems, for several reasons.
First, fishing impacts have followed similar pat-
terns in many regions (Jennings and Kaiser 1998;
Paulyet al. 1998; Jacksonet al. 2001; Hilbornet al.
2003a; Myers and Worm 2003), providing a degree
of replication and potential generality. Second, for
pelagic fishes specifically, harvesting provides a
relatively ‘clean’ test of community manipulation
in that removal of individuals or species from the
system has relatively low impact on the habitat
(although ‘ghost’ nets can continue to ensnare fish-
es indiscriminately long after they are lost or aban-
doned); this is in contrast to most human impacts
on land (Wilcoveet al. 1998) and in benthic habitats
(Watling and Norse 1998), where habitat destruc-
tion or modification confounds species removals
with other impacts (Srivastava and Vellend 2005).
Third, the major commercial value of fisheries
means that there is a large body of detailed data
on which, when and how many fish have been
removed from the oceans (FAO 2007). Finally, the
growing database on ecological changes within ma-
rine protected areas (MPAs) offers important large-
scale experimental controls against which to evalu-
ate the effects of fishing (Halpern and Warner 2002;
Micheliet al. 2004).8.1.2 Physical forcing and the uniqueness of marine ecosystems
The ecology of marine communities and their re-
sponses to perturbations are strongly influenced by
the unique nature of the marine environment
(Steele 1985, 1991). The fundamental physical dif-
ference between terrestrial and (pelagic) aquatic
systems is the greater density of the liquid medium
of water (Strathmann 1990), which has three impor-
tant consequences for understanding the ecological
structure and dynamics of terrestrial compared
with pelagic ecosystems. First, in water, buoyancy
allows primary producers (and other organisms) to
float and obviates the need for large, expensive,
metabolically inert structural tissues required to
compete for light on land. Thus, the dominant
marine pelagic autotrophs are microscopic, fast-
growing and highly nutritious (floatingSargassum
accumulations being a conspicuous exception in the
Atlantic gyre). Consequently, compared with land,
marine ecosystems show higher grazing rates and
production: biomass ratios, a much larger faction of
primary production grazed and more efficient con-
version of production to herbivore biomass (Steele
1991; Cebrian 1999; Shurinet al. 2006). The higher
growth rates, nutritional content and vulnerability
of marine autotrophs to grazing in turn have
important consequences for the structure and
functioning of marine communities, including96 APPLICATIONS