| dc.description.abstract | Stochastic demography is essential to accurately predict how populations persist in variable environments. In age- or stage-structured populations, the long-run growth rate is negatively affected by temporal variance in vital rates (e.g., survival and fecundity). Consequently, natural selection is predicted to minimize the variation in the most influential vital rates, resulting in demographic buffering. This thesis aimed to investigate the demographic buffering hypothesis in detail. In the first paper (Paper I), we did a thorough review of the existing literature on demographic buffering and found an overall support of the demographic buffering hypothesis. However, several issues related to testing for demographic buffering were identified. We found that solving scaling problems when decomposing, measuring, and comparing stochastic variation in vital rates is crucial when studying demographic buffering. From the very first studies on demographic buffering, a major problem when comparing the temporal variance of vital rates bound between 0 and 1 has been how to scale the variance to make it comparable between age- or stage-classes with differing mean values. In the second paper (Paper II), we discuss how the coefficient of variation (CV; standard deviation divided by the mean) can be used (and misused) when comparing variation of quantitative traits. We also discuss for which scales the variation can be meaningfully compared and show that for traits described by statistical distributions where the mean and variance are nonproportional (e.g., survival) the CV must be standardized to make meaningful comparisons. In age-structured populations, the influence of survival on the population growth rate decreases with ageing after age at first (AFR) reproduction. Consequently, demographic buffering should be strongest at AFR and decrease thereafter (Paper I). In the third paper (Paper III), using unique individual data from 16 species with differing life-histories, we found that the temporal CV of survival increases with age after AFR. In addition, we tested if the strength of buffering, defined as the slope between temporal variance and elasticity, depends on a species position along the slow- fast continuum (measured as generation time). Taking phylogeny into account, we found that the strength of buffering decreases with generation time when scaled by absolute time. However, using biological time (absolute time divided by generation time), the strength of buffering increases with generation time. | en_US |
