Does fire severity difference create divergent plant communities in black spruce-Kalmia forests of eastern Canada?
Abstract
After fire, some conifer-ericaceous plant communities in Western Europe and eastern
North America have demonstrated divergent succession from traditional conifer stand
replacement to Ericaceae-dominated community states. Divergence may be of function of
species‟ responses to changes in fire severity as determined by the amount of residual organic
matter (ROM) remaining after fire. ROM depth may act as an abiotic habitat filter to conifer
species by physically inhibiting species establishment or through chemical afterlife effects in the
unburned organic matter. If post-fire community assembly is predominantly determined by ROM
depth, then post-fire species composition and their functional traits should correspond with the
depth and distribution of ROM. But if species interactions are the primary constraint on
community assembly, then post-fire species composition and their functional traits should be
independent of the depth and distribution of ROM. I tested these predictions in black spruce
(Picea mariana (Mill.) B.S.P.)-Kalmia (Kalmia angustifolia L.) plant communities of eastern
Newfoundland among three post-fire plant community types (heath, wood savannah and forest)
with similar pre-fire stand types (forest).
To determine if the three communities were similar prior to fire, I reconstructed stand
types by sampling unburned portions of adjacent stands and by compiling pre-fire forest resource
inventory data. I quantified post-fire differences in species composition by determining percent
cover of all vegetation along three 20 x 1 m belt transects in replicated 20 x 20 m plots. From the
species list, I compiled functional trait information from the literature for each species, and
weighted traits by the total cover of species that possessed a given trait. Species and trait
response (total abundance, richness and diversity) was then compared using one-way ANOVA.
Indicator analysis was also performed to characterize species and functional traits among
communities. I estimated fire severity as ROM by taking soil cores sampled contiguously in one
20 x 1m belt transect from the replicated 20 x 20 m plots in each community. ROM was
estimated by the depth of organic matter not consumed by the most recent fire under the
uppermost charcoal layer. Multivariate ordination (Canonical correspondence analysis) and
exponential growth and decay models were then used to test the effect of ROM on species
composition and their functional traits.
I found that the relative importance of abiotic and biotic filtering in post-fire community
assembly was a function of ROM. When ROM was less than 2 cm, abiotic filtering favored the
establishment of black spruce and other seed regenerating species, leading to black spruce
recovery. When ROM was greater than 2 cm, substrate conditions and biotic interactions favored
Kalmia and other vegetatively regenerating species, resulting in successional divergence. Also,
when abiotic filtering was the predominant habitat constraint, it controlled species composition
and functional traits directly by filtering species mode of regeneration, and indirectly through
differences in the dominant species‟ performance between the heath and forest. The
independence of species composition and functional traits to shallow ROM in the wood
savannah suggests that priority effects may play an important role in determining the relative
importance of abiotic vs. biotic filtering when ROM is around a threshold value (~2 cm).