Bias in Estimating Forest Disturbance Rates and Tree Lifetimes

Intervals between fires, tree blowdowns, and other large and small forest disturbances are often estimated by dating tree rings. Dates are taken from living trees, fallen logs or stumps, or other indicators such as peaks in tree age distributions and ages of understory plants damaged by treefalls. Such indicators disappear gradually over time; this will bias estimates of disturbance recurrence times or tree lifetimes. Models and simulations employing plausible rates of disturbance and indicator disappearance show that the bias can be substantial. Comparison of a realistic simulation with simpler models suggests that it may be difficult to correct the bias, given estimates for indicator survival rate. I recommend that investigators estimating forest—disturbance recurrence items also estimate survival rates for their indicators, and collaborate with statisticians (about estimating survival times with censored data) before designing field studies.