Geoscience Reference
In-Depth Information
(
Gower et al., 1980
), localised production in the vicinity of fronts (
Holligan,
1981
) and high levels of production associated with coastal (
Barber and
Smith, 1981
) or equatorial upwellings (
Vinogradov, 1981
). These areas
of high primary production support production at higher trophic levels
(e.g.
Herman et al., 1981; Vinogradov, 1981
).
Phytoplankton are the main contributors to total primary production in
marine ecosystems (
Duarte and Cebri
´
n, 1996
) and support strongly size-
structured food webs where most predators are larger than their prey
(
Boudreau et al., 1991
). Since the transfer of energy from prey to predators
is inefficient, the production of the community falls as body size and trophic
level increase (
Jennings et al., 2002
). Further, since the relationship between
production and biomass depends on body mass, biomass also changes
systematically with body mass. Consequently, relationships between (log)
abundance and (log) body mass, commonly dubbed size spectra, are widely
used to describe the structure of size-based communities.
Dynamic models of community size spectra can be used to predict how the
abundance (N) of individuals changes as a function of body mass (M) and time
(t) through the processes of size-based feeding interactions driving growth and
mortality (
Benˆ it and Rochet, 2004; Blanchard et al., 2009; Camacho and Sol´,
2001; Gilljam et al., 2011, Law et al., 2009; Maury et al., 2007a,b; Meli´n et al.,
2011; Silvert andPlatt, 1980
) although its effectiveness can depend on the nature
of the systemunder investigation (
Henri and vanVeen, 2011
) and the diversityof
the predator species (
Jacob et al., 2011
). The scaling of N with M from these
models is consistent with empirical size spectra for communities (
Arim et al.,
2011; Boudreau and Dickie, 1992
). These models typically assume that indivi-
duals are distributed homogeneously in space. Similarly, the data used to
describe the structure of size spectra in empirical studies are often pooled across
large spatial scales in an attempt to represent the community. While a few
empirical studies have investigated how size-spectra slopes vary spatially
(
Blanchard et al., 2005a; Piet and Jennings, 2005
), there has been no attempt
to assess systematically the consequences of describing size spectra on different
spatial scales. However, there is a history of species-specific size-, age- and stage-
based models incorporating basic spatial processes (
Bryant et al., 1997
,etc.),
and a spatial advection and diffusion model has been developed to model the
role of top predators at the scale of large marine ecosystems using a size-
spectrum approach (APECOSM;
Maury, 2010
). Analyses of the effects of
spatial processes on the structure and dynamics of community size spectra are
required to understand how localised zones of high primary production might
influence community structure more widely and to provide insight into the
effects of sampling at different scales on the apparent structure of the commu-
nity size spectrum.
Biotic processes that may influence predator-prey interactions (
Arim et al.,
2011; Nakazawa et al., 2011
) and hence the local structure of the size spectrum
