Geoscience Reference
In-Depth Information
10
0
10
-
1
10
-
2
10
-
3
0
10
-
4
5
0
100
200
10
300
Sampling effort
Figure 14
Sampling effort and species-level connectance for the whole food web:
Observed (red (dotted lines) ) and predicted (black represents the mean after 10
2
replicates) connectance, C
¼
L/S
2
, (where L is the number of tropic links and S is
the number of species) for the Guadalquivir estuary food web at a given sampling
effort (i.e. number of individuals) for all the spatio-temporal situations analysed. The
observed connectance decays much faster than the predicted connectance at any given
number of individuals sampled when a
ij
¼
1 if at least 1, 2, 3,
...
, N individual predator
of species j has 1 individual prey of species i. Observed connectance is one order of
magnitude smaller after the first 10
2
individuals sampled. Connectance decays from
3
10
2
(model expectation) to 3
10
3
(observed data). This confirms the high
sensitivity of species-level connectance to sampling and individual variance as in the
analysis for the fish-mysid food web. The pattern remains for all the spatio-temporal
situations. The decay of the observed connectance is even faster for the scenario with a
small increment in the number of items of each individual predator.
approximately 25,000 individuals sampled across prey and predators, a
result that is against the expected decrease of the variance with increasing
sampling effort.
C. Effect of Total Length on Intraspecific Variance in Prey
Consumption
The difference between model prediction and data is partially explained by
the length of individual predators. Specifically, 35% of this variance is
accounted by the length of the individual predators in all the environmental
situations (
Figure 7
) and less than 20% is accounted by the identity of each
species (
Figure 8
). The variance explained is even lower because we did the
