PARENT SESSION
Oral Session 81: Forest Ecology IV: Seeds, Growth, and Recruitment.
Presiding: J Kush
Wednesday, August 4, 1:30 PM to 5:00 PM, Meeting Room C 120.

Dissecting the inverse relationship between tree growth rate and probability of death.

Stephenson, Nathan^*,1, van Mantgem, Phillip¹, ¹ U.S. Geological Survey, Three Rivers, CA

ABSTRACT- Models suggest that forest composition and dynamics can be profoundly affected by the nature of the relationship between tree growth rate and probability of death. Yet little is known about the underlying causes of the relationship, or how environmental changes might affect it. In particular, the widely-used "gap" models of forest dynamics rely on two untested assumptions: (1) causes of tree death fall into two groups: those independent of and those dependent on growth rate, and (2) the only way environmental changes affect probability of death is indirectly, by altering growth rate. To test these assumptions, we tracked the growth and survival of 10,691 trees in the Sierra Nevada, California. Over several years we recorded 775 deaths by cause: mechanical (breaking, uprooting, and crushing), biotic (mostly insects and fungi), and "stress" (probably including all deaths by direct physiological stress, plus some biotic deaths). Logistic regression revealed that, contrary to gap model assumptions, no cause of death was independent of growth rate. However, the strength of the relationship between growth and death differed significantly among causes: stress > biotic > mechanical. Additionally, an introduced pathogen increased the probability of death in Pinus lambertiana growing at all rates, demonstrating that changes in probability of death can be either growth-mediated, as assumed in gap models, or direct, resulting from a change in the nature of the relationship between growth rate and probability of death. We propose that in response to environmental changes, the relative importance of growth-mediated versus direct changes in tree mortality will vary among causes of death, in proportion to the relative strength of the relationship between growth rate and probability of death. Specifically, any environmentally-induced change in probability of stress mortality will mostly be growth-mediated, whereas any change in probability of mechanical mortality will mostly be direct. Change in biotic mortality may be either growth-mediated or direct (or both).

Key words: forest dynamics, forest demography, gap models, tree mortality