Geoscience Reference
In-Depth Information
VI. CONCLUSION
Although the PPMR has been widely used in size-based food-web modelling,
the empirical estimation of PPMR is not simple and remains controversial.
This is because PPMR may be defined in different ways, depending on the
choice of biological scale from individuals to communities. Therefore, there
is variation in PPMR with factors such as species identity and body mass,
which contrasts with the conventional assumption of these models. Using
recently compiled gut content data of marine food webs (
Barnes et al., 2008
),
we conducted a detailed study of scale dependence and determinants of
PPMR. We illustrated that the scale dependence of PPMR is determined in
a complex way and that the averaging and sampling effects may result in
different values of PPMR, depending on data elements such as body mass
and sample counts of predators and prey. The results of our study comple-
ment previous arguments that species averaging underestimates PPMR
(
Woodward and Warren, 2007; Yvon-Durocher et al., 2011b
). We also
used AIC to elucidate how PPMR is explained by predator species identity
and body mass for different PPMR definitions. We observed that the possible
determinants of PPMR become clearer with increasing resolution of data
analysis. For species-averaged and link-averaged PPMRs, different statisti-
cal models were selected among food webs, with the best model remaining
unclear. For individual-predator and individual-link PPMRs, the model that
combined species identity and body-mass effects gave the best explanation in
most of food webs. Based on these results, we discussed the application of
PPMR in food-web models. The species-based allometric food-web model
relies on the species-averaged or link-averaged PPMR, the determinants of
which are uncertain, and thus caution is necessary when applying the model.
Meanwhile, the size-based community-spectrum model, which relies on indi-
vidual-predator or individual-link PPMR, should consider taxonomic dis-
similarities, although it runs counter to the initial objective to simplify the
complexity of the food-web structure. We also suggest that PPMR may vary
with factors other than predator species identity and body mass. To date, no
theoretical models have been developed to predict the observed patterns of
PPMR, such as scale dependence and interspecific or intraspecific variations.
It is important to recognise that we are still at an early stage of understanding
size-dependent trophic interactions and their resulting food-web dynamics.
Further studies are required to accumulate high-resolution data on feeding
relationships in various ecosystems and to establish a more reliable form of
size-structured food-web models. Laboratory experiments are also useful for
identifying the determinants of predator-prey body-size relationships. Ulti-
mately, future research with respect to PPMR is expected to contribute to
our understanding of the structure and dynamics of complex food webs.