Geoscience Reference
In-Depth Information
b. Prey Mass-Variance of Predator Mass (Grouping). The paired t-test
found no significant difference between the slopes (estimated by linear
regression) of the different study systems under the taxonomic (B) and size-
class (E) groupings (t
0.314, Figure 11 C).
c. Prey Mass-Range of Predator Mass (Resolution). The paired t-test
found no significant difference between the slopes (estimated by linear
regression) of the different systems under the high (B) and the low (D)
resolutions (t
¼
1.098, df
¼
6, p
¼
0.337, Figure 11 D).
d. Prey Mass-Range of Predator Mass (Grouping). Again, the paired t-test
found no significant difference between the slopes (estimated by linear re-
gression) of the different systems under the taxonomic (B) and size-class (E)
groupings (t
¼
1.043, df
¼
6, p
¼
0.118, Figure 11 E).
e. Species Mass-Out-Degree (Grouping). This response also had a unim-
odal relationship with the size-class mass ( Figure 12 ) and, as with the preda-
tor mass-range of prey mass relationship, a quadratic term was introduced
into the linear regression. The paired t-test revealed that there was a signifi-
cant difference between the size of the quadratic term in the two groupings
with the size-class food web grouping having a much greater negative coeffi-
cient (mean of
¼
1.824, df
¼
6, p
¼
24.632) than did the taxonomic food web grouping (mean of
0.032, Figure 11 A). For the taxonomic food
web grouping, only two of the seven relationships had a significant quadratic
term. If we used a model without the quadratic term instead, five of the
systems had significantly negative slopes (Coilaco and Guampoe were the
exceptions: data not shown).
The differences in all resolution comparisons are summarized in Figure 13 A,
while the differences in all grouping comparisons are summarized in
Figure 13 B. These summaries show that in 7 of 15 comparisons done, the
scaling of the response variables with their respective focal mass differed
depending on the resolution or grouping of the data used. More specifically,
in Size Structure Dimension Set #1: Trophic Orderings (Section III.A.1 ), all
response variables (prey body mass, PPMR and TH) scaled differently depend-
ing on resolution or grouping. In Size Structure Dimension Set #2: Diet Varia-
tion (Section III.A.2) , prey range scaled differently depending on the grouping
used. Lastly, in Size Structure Dimension Set #3: Predator Variation (Sec-
tion III.A.3 ), out-degree scaled differently depending on the grouping used.
3.991, t
¼
2.771, df
¼
6, p
¼
IV. DISCUSSION
A. Individuals and Species Averages—Effects of Resolution
One aim of this study was to investigate if our perception of patterns in size
structure (e.g. predator-prey relationships) in ecological communities will
be changed as the resolution of empirical datasets becomes finer. We show
Search WWH ::




Custom Search