2.9 Statistical analyses
One-way repeated-measures analysis of variance (RM ANOVA) was used to
assess the effect of girdling on respiration of leaves and branches. We
used linear mixed effect modelling to assess differences in St and Ss in
leaves and bark before and after girdling (fixed effects), accounting
for tree number, sample position and sampling date (random effects).
Linear mixed effect modelling was also used to account for tree number,
sample position and sampling date (random effects) when assessing
differences in D x and D bprior and after girdling (fixed effects). Student’s t-tests were used to
assess differences in D G+ before and after
girdling. We report the marginal R2 (proportion of
total variance explained through fixed effects) and conditional
R2 (proportion of total variance explained through
fixed and random effects) of the mixed effect model.
Regression analyses were used to determine the effect of girdling on the
relationships between whole-tree fluxes of H2O and
CO2 and environmental drivers, namely PAR,
Tair and VPD prior to and after girdling. Single factor
ANOVA and Fisher’s LSD test were used to assess the effect of sample
location (base, top) on the daily amplitude of expansion and contraction
in D x, D b, andD Gempir. These tests also
included comparing pre-girdle (6 to 10 May) to post-girdle (20 to 24
May) measurements. Violation of the condition of homogeneity of variance
of these data was tested (Barlett’s chi-square) and rejected in all
cases as p >0.05.