Analyzing treatment effects on Douglas-fir seedling survival with an extreme value regression model

Abstract

First year survival patterns in a cohort of natural Douglas-fir (Pseudotsugamenziesii (Mirb.) Franco) seedlings were assessed with an extreme value regression model. Extreme value distributions were fit to the data, assuming equal scale parameters among treatments but allowing location parameters to vary. The regression approach using indicator variables facilitated statistical comparison of survival curves among several habitats and treatments.