Geoscience Reference
In-Depth Information
IV. DISCUSSION
The spatially explicit, continuous, time-dependent model of community
size composition shows that the passive transport and active movement of
individuals can influence local growth and size spectra. Prey-seeking and
predator-avoiding behaviour led to changes in the relative abundance of
different-sized individuals with changes in the abundance of primary produ-
cers. In areas of high phytoplankton abundance, community size-spectrum
slopes were shallower and larger individuals were present, whereas in low
production areas, slopes were steeper and size spectra truncated. This agrees
qualitatively with empirical patterns in the Celtic Sea (
Blanchard et al.,
2005a
) and theoretical results from an alternate size-spectrum model
(
Blanchard et al., 2010
). As size spectra are compiled at increasing spatial
scales, that span gradients in primary production, truncation becomes less
apparent. The results suggest that the apparent size structure of marine
communities will depend on the scale of sampling and emphasise the need
to sample at adequate space and time scales if seeking to compare emergent
properties, such as size spectra, among regions or ecosystems. At local scales,
(i.e. those scales that are small in relation to the scale of gradients in primary
production), the size composition of the community can respond to the
movements of predators and prey, and the structures and slopes of size
spectra are expected to vary in the absence of human impacts. This underly-
ing variation has to be recognised and considered if variations among size
spectra in space are used to describe the effects of human impacts such as
fishing, and suggests that responses to human impacts will be more consistent
at larger spatial scales. With limited resources to allow replication in space,
our results show that more reliable descriptions of regional size spectra will
be obtained by sampling in areas of relatively high primary production.
The model structure was based on relatively few assumptions. We assumed
predation was the main factor driving growth and mortality, and a driver of
behavioural-based movement. We also included senescence mortality and a
density-dependent movement term. Of the assumptions, the evidence for
senescence mortality is the least well supported by available data, owing to
the absence of marine ecosystems impacted by humans where the role of
senescence can be studied.
The set of cost functions used in the model was intended to reflect the
behaviour of individuals. For generality, we assumed cost functions were
spatially and temporally invariant, but these could be modified to incorpo-
rate and assess the effects of variations in space and time. Modelled move-
ment was based on 'local' rather than 'ideal' knowledge of the fitness
landscape, but adopting a 'local' or 'ideal' assumption necessarily simplifies
potentially complex behaviour. For instance, migrating fishes may suppress
