Pair-wise analyses of the effects of demographic processes, vital rates, and life stages on the spatiotemporal variation in the population dynamics of the riparian tree Aesculus turbinata Blume

Population growth rates (λ) of the riparian tree Aesculus turbinata varied from 0.9988 to 1.0524 spatiotemporally. We conducted a series of pair-wise demographic and matrix analyses, including randomization tests, three types of life table response experiments (LTREs), analysis of variance and χ² tests, to test which life stages had the greatest effect on this variation in λ. Randomization tests detected significant variations in λ between plots affected or not by typhoons in three habitats and between periods with high and low recruitment in one habitat. Mixed-level LTREs identified that the demographic processes and life stages that had the strongest effect on the actual variation in λ were: (1) progressions of small and intermediate juveniles and (2) founding process from seeds to 1-year-old seedlings. These juvenile stages had medium sensitivities and variances that explained high upper-level LTRE contributions. Lower-level LTREs showed that the vital rates contributing the most were the growth rates of these juvenile stages. These findings demonstrate that progression from one stage to the next, growth rates of 1-year-old seedlings, and stunted aging juveniles are the most important stages in the population dynamics of this long-lived primary tree species. Transition matrix elements with high elasticities had little effect on the variation in λ, indicating that high-elasticity vital rates do not necessarily drive variation in population growth. As compared with the results of randomization tests, significant differences in vital rates examined using ANOVA or χ² tests showed that typhoon disturbance had the greatest effect on the demographic parameters of individual trees.