Geoscience Reference
In-Depth Information
nodes or species, S, in a food web and the number of links, L, between them
(
Hall and Raffaelli, 1993
).
A food chain is an ordered sequence of at least two species that runs from a
basal species (i.e. a primary producer or detritus) to a top predator. Food
chain length is the number of links within this particular path (
Hall and
Raffaelli, 1993; Pimm, 1982
).
Linkage density (L/S) is the number of links per species (S), connectance (C),
the proportion of realised links within a web, is calculated as 2
L/(S
2
S)
(
Hall and Raffaelli, 1993
; Warren, 1989), linkage complexity is calculated by
S
C (
Briand, 1985
). Trophic vulnerability (V) and trophic generality (G)of
a species are the numbers of its predator and prey species, respectively
(
Schoener, 1989
).
With respect to the number of links from detritus to its consumers, we
followed a conservative approach here as we introduced a single virtual
'phytodetritus species'. As there are 59 phytoplankton species in our system,
one could argue that there exist 59 feeding links between phytodetritus and
any species that feeds upon it. This would change all parameters significantly
and make the Weddell Sea system even more unique in terms of linkage
density and generality.
There are various ways to calculate the trophic height of a species within a
food web. The prey averaged trophic height is the TL calculation many prior
studies have used, which is equal to 1 plus the mean trophic height of all the
consumer's trophic resources (Williams and Martinez, 2004). Here, we use
the short-weighted trophic height, where the prey averaged trophic height is
weighted by the shortest chain within the network, as it is a better estimate of
trophic height (
Williams and Martinez, 2008
).
D. Data Analysis: Statistics, Extinction Scenarios and
Robustness of Weddell Sea Food Web
To explore whether species body mass was correlated with trophic level as
well as whether a consumer's trophic generality and vulnerability (e.g.
Memmott et al., 2000
) are related to the functional consumer classifications,
we conducted simple pairwise correlations between body size and trophic
level for all parameters for the entire set of species. We also conducted this
analysis for all consumers combined (i.e. ignoring categories). The analysis
was performed using R.
To investigate robustness of the Weddell Sea food web, we carried out
computer simulations that quantified how susceptible the food web was to
collapse as a result of secondary extinction cascades. The method employed
is as follows: first, a species is removed from the network; following this, any
non-basal species that loses all of its prey items, or cannibalistic species that
