Geoscience Reference
In-Depth Information
sense, the limited size of communities could disproportionately affect larger
individuals (
Burness et al., 2001; Marquet and Taper, 1998
). However, com-
munities from other temporary pond ecosystems presented a shallower scaling
(
Hayward et al.,2009
). Therefore, the large scaling might be more closely
related to a better statistical estimation of parameters as opposed to the more
biased and imprecise binned methods (
White et al.,2008
). In
Figure 4
, estima-
tion of the scaling exponent with binned and unbinned methods is contrasted.
This comparison suggests that shallow slopes derived from binning methods
could be related to the consideration of noisy information at either extreme of
body size (see
Figure 4
). Our analysis of methodological performance also
suggests that histogram methods are probably underestimating the exponent
of the DMR (see below). In addition, it should be considered that recent
theoretical predictions also report the occurrence of larger exponents than
formerly considered (e.g.
Loeuille and Loreau, 2006
).
C. Cross-Community at Different Levels
Self-thinning has been proposed to operate in groups of individuals that
share a limiting resource (
Westoby, 1984; White et al., 2007; Yoda et al.,
1963
). The pattern reported for the whole community in
Figure 6
mixes
individuals from different species, guilds, and trophic levels. The reduction
in individual density compensating for the increase in energetic demands with
increasing body size is expected to be more evident when individuals of the
same guild or species are considered. The self-thinning analysis for the guild
of annual fishes denotes a strong density-mass association (
Figure 8
). Final-
ly, at the population level the scaling in the self-thinning pattern was even
more pronounced. This last result highlights the dependence of the DMR on
the ecological level of analysis. In the estimation of abundance for each
community, several sampling units without fishes were observed. Abundance
was estimated as the number of individuals collected divided by the number
of sample units taken in the pond. As a consequence, these null observations
have the potential to affect scaling since areas without individuals could
represent regions without suitable conditions for fish presence. The exclusion
of these null observations produced a biologically meaningful change in
scaling from values close to
1 to exponents very close to
0.75 (
Figure 8
).
D. Amphibians as an Example of Discontinuous DMR
Amphibian species in the pond metacommunity are a good example of the
coexistence of different scalings within the same body size distribution
(
Figure 7
A). The frequency distribution of body sizes at the individual level
clearly indicates the existence of two modes separated by a discontinuity.
