Geoscience Reference
In-Depth Information
a slice in time and not the cumulative diet throughout the ontogeny of the
individuals. Further data, mainly from experiments in mesocosms, linking
size, ontogeny and behaviour, are required to test expectations from random
encounters and more realistic models to understand the variance in prey
consumption during the lifespan of individuals in natural populations
(
Moran, 1962; O'Dwyer et al., 2009; Woodward et al., 2010
).
In addition to the individual rank connectivity curves, we test the
eco-evolutionary model with the species abundance curves for a predator-
prey dataset. This has been challenging. We feel that the approach developed
here to scale from individual-level processes to food web dynamics in space
and time has several gaps and it still requires new generations of researchers
combining testable theory with experiments and large datasets. Some useful
information, however, can be obtained. For example, the model predicts
more rare species than observed in the fish metacommunity. Two different
mechanisms might cause this pattern (1) higher probability of extinction for
the rare species in a highly fluctuating and stressed ecosystems such as the
Guadalquivir estuary (
Rapport et al., 1985
) and (2) strong dispersal limita-
tion even if sampled species are the most common in the estuary. These two
mechanisms were not implemented by our neutral expectation because we
have assumed only demographic (and not environmental) stochasticity and
no dispersal limitation, respectively. Thus stress and dispersal limitation
might be strong factors shaping diversity in space and time in the Guadal-
quivir estuary. This is not surprising because dispersal limitation has been
recognized as one of the main factors driving diversity in several communities
under different neutral models with different speciation modes (
Etienne and
Haegeman, 2011; Rosindell et al., 2010
). A future challenge to improve the
link between individual-based datasets and eco-evolutionary models of indi-
vidual-based food webs will be to compare neutral expectations using a suite
of mechanisms with a range of dispersal values against individual-level data
sampled over a broad range of spatial and temporal scales.
There are a greater number of abundant fish at lower salinity and temper-
ature for all temperature and salinity ranges than predicted by the model
(
Figure 10
). This result suggests that the diversity of abundant species
increases not only with lower salinity (
Attrill and Rundle, 2002; Odum,
1988
) but also down temperature gradients. This is most likely driven by
specific combinations of spatial heterogeneity in the Guadalquivir estuary
increasing the probability of higher diversity caused by overlapping between
freshwater and marine fish species. The expectations from the random en-
counter model, however, more closely predict the species abundance curves
at higher values of salinity and temperature, when environmental conditions
are likely to be more homogeneous. These patterns of diversity in the differ-
ent environmental conditions are not observed in the mysid metacommunity,
possibly a consequence of the low number of species sampled.
