Spatial aspects of tree mortality strongly differ between young and old-growth forests

Rates and spatial patterns of tree mortality are predicted to change during
forest structural development. In young forests, mortality should be primarily density
dependent due to competition for light, leading to an increasingly spatially uniform pattern of
surviving trees. In contrast, mortality in old-growth forests should be primarily caused by
contagious and spatially autocorrelated agents (e.g., insects, wind), causing spatial aggregation
of surviving trees to increase through time. We tested these predictions by contrasting a threedecade
record of tree mortality from replicated mapped permanent plots located in young
(,60-year-old) and old-growth (.300-year-old) Abies amabilis forests. Trees in young forests
died at a rate of 4.42% per year, whereas trees in old-growth forests died at 0.60% per year.
Tree mortality in young forests was significantly aggregated, strongly density dependent, and
caused live tree patterns to become more uniform through time. Mortality in old-growth
forests was spatially aggregated, but was density independent and did not change the spatial
pattern of surviving trees. These results extend current theory by demonstrating that densitydependent
competitive mortality leading to increasingly uniform tree spacing in young forests
ultimately transitions late in succession to a more diverse tree mortality regime that maintains
spatial heterogeneity through time.
Key words: Abies amabilis; density dependence; forest structural development; long-term studies; oldgrowth
forest; Pacific silver fir; self-thinning; succession; tree mortality; western Cascade Range,
Washington, USA.