Document: BRI-3-91-6

Estimating self-thinning relations in plants with regression quantiles.

CADE, B.S.*

U. S. Geological Survey, Fort Collins, CO 80525 USA ¹

Abstract:
Density dependent mortality due to intra- and interspecific competition for resources can be viewed as a constraint on plant performance. Viewing the self-thinning process as a constraint on plant performance suggests that changes in the upper quantiles of the density distributions will convey more reliable information on this process than changes in the center or lower quantiles of the density distributions, where other processes may be the active constraint. I demonstrate recent advances in a class of statistical estimators, termed regression quantiles, that provide estimated rates of change for all percentiles of probability distributions. Example analyses of density dependent mortality due to intraspecific competition for two Chihuahuan Desert annuals, Eriogonum abertianum and Haplopappus gracilis, indicated an exponential functional form was appropriate for both species but parameter estimates differed between them. Constraints on densities at the end of the growing season initially increased and then decreased with increasing germination densities for both species, but the decrease occurred at higher germination densities for Haplopappus than for Eriogonum. A modification of the basic regression quantile analysis using geometric means of two estimates was made to estimate linear density dependent changes due to interspecific competition between Eriogonum and Haplopappus, where neither species can be regarded as the dependent variable in a regression model. Advantages of the regression quantile estimates are discussed relative to those from other regression procedures and nonstatistical methods.

Keywords: density dependence, limiting factors, regression quantiles, self-thinning

This abstract is being presented at: 3:30 PM in session:
Oral Session #46: Modeling Populations and Statistical Ecology.