## dbMEM analysis of the mite data - broad scale
(mite.dbmem.broad anova(mite.dbmem.broad)
(axes.broad <- anova(mite.dbmem.broad, by = "axis"))
# Number of significant axes
(nb.ax.broad # Plot of the two significant canonical axes
mite.dbmembroad.axes choices = c(1,2),
display = "lc",
scaling = 1)
par(mfrow = c(1, 2))
sr.value(mite.xy, mite.dbmembroad.axes[ ,1])
sr.value(mite.xy, mite.dbmembroad.axes[ ,2])
# Interpreting the broad-scaled spatial variation: regression of
# the two significant spatial canonical axes on the environmental
# variables
mite.dbmembroad.ax1.env summary(mite.dbmembroad.ax1.env)
mite.dbmembroad.ax2.env summary(mite.dbmembroad.ax2.env)
The broad-scale relationships are clearly related to microtopography and the
absence of shrubs.
## dbMEM analysis of the mite data - medium scale
(mite.dbmem.med anova(mite.dbmem.med)
(axes.med <- anova(mite.dbmem.med, by = "axis"))
# Number of significant axes
(nb.ax.med <- length(which(axes.med[ ,ncol(axes.med)] <= 0.05)))
# Plot of the significant canonical axes
mite.dbmemmed.axes display = "lc",
scaling = 1)
par(mfrow = c(1, 2))
sr.value(mite.xy, mite.dbmemmed.axes[ ,1])
sr.value(mite.xy, mite.dbmemmed.axes[ ,2])
326
7 Spatial Analysis of Ecological Data