A general linear mixed model with repeated measurements was
used to test the effect of treatments on total vegetation cover
(woody + herbaceous native plants) and on seeded grass cover.
Treatments were the same as explained above for erosion.
Treatment, sampling date, and their interactions were included
as categorical
fixed effects in the model. Plot and quadrat were
included as random effects to account for positive correlations
between repeated measurements on the same plot and quadrat.
When significant mixed effects were indicated, post-hoc pairwise
comparisons (with Bonferroni adjustment for multiple compar-
isons) were conducted to assess differences between the main
effects of treatments and their interactions. Residuals were tested
for autocorrelation, normality, and homogeneity of variance.
A generalized linear model was used to explore the possible
influence of vegetation patch area and contiguity index, in each
vegetation measurement date, on accumulated in three months
interval soil erosion measured in untreated burned and seeded
plots in the
first post-fire year. Erosion data were log transformed
to achieve normality. Mulching treatment was not evaluated as we
considered that the change in soil cover conditions and associated
hydrological parameters caused by the mulch masked the effect of
vegetation.