Multi-scale analysis of fire effects in alpine treeline ecotones
Ph.D. Dissertation Abstract by Alina Cansler (2015)
Although the direct effects of climate change have been studied though observational
and experimental methods in alpine treeline ecotones (ATEs), indirect effects due to
shifts in disturbance regimes have received less attention despite evidence that the
frequency and extent of large disturbances are increasing in many other ecosystems.
At a regional scale, I analyzed wildfires occurring over a 29-year period (1984-2012)
in ATEs in eight mountainous ecoregions of the Pacific Northwest and Northern Rocky
Mountains. I focused on two components of the ATE: (1) subalpine parkland, which extends
from closed subalpine forest through a fine-scale mosaic of forests and non-forest, and
(2) alpine vegetation, which includes meadow, shrubland, and alpine tundra. I expected
that subalpine parkland and alpine vegetation would burn less, proportionally, than the
entire ecoregion. In four of eight ecoregions—three in Rocky Mountains and one in the
Cascades—the proportion of subalpine parkland burned was comparable or greater than the
proportion of the entire ecoregion that burned. In alpine ecosystems little of the area
(<7%) burned during the 29-year study period.
At a local scale, we examined variability in fire severity and changes in plant
structure, using data from >500 plots within four alpine treeline ecotones sites in the
Cascade Range and Northern Rocky Mountains, which had burned 18-27 years prior. We
assessed the likelihood of different pre-fire canopy-cover structural classes—closed
forest (>40% tree cover), open forest (10%-40%), parkland (<10%), and unforested
areas (alpine, meadow, and Krummholz)—to burn and to change to a different structural
class after fire. We also evaluated changes in forest structure—specifically the
abundance of live trees within five diameter at breast height (DBH) classes—using
non-metric multidimensional scaling (NMDS) to visualize differences and Permutational
Multivariate Analysis of Variance (PERMANOVA) to test statistically for differences from
pre-fire to post-fire, and between unburned and three higher-severity class. Non-forested
areas were less likely to burn and fire increased the proportion of non-forested area.
The effects of the fire on forest structure were mixed: previously forested stands had a
greater probability of retaining forest cover than they had of becoming non-forested.
Greater fire severity decreased the abundance of larger, relative to smaller, overstory
trees; the latter suffered greater mortality. Of the four common high-elevation tree
species observed in burned plots, Abies lasiocarpa had the highest rates of mortality
(60%), Larix lyallii had the lowest rate (11%), with intermediate levels in
Pinus albicaulis (52%) and Picea engelmannii (37%).
Strong significant correlations between the overall annual area burned across all vegetation types,
and the area burned in subalpine parkland and alpine vegetation (? = 0.89 and ? = 0.88, respectively),
indicate that fire may become more prevalent in both subalpine parkland and alpine vegetation if the
overall area burned increases due to climate change. Within burned ATEs fire effects are moderate, and
highly heterogeneous. The combined effect of climate change and fire may cause ATEs to expand upward and
trees to infill previously snow-dominated sites, while simultaneously increasing fine- and course-scale
heterogeneity within the ecotone, due to fire-cause mortality.