Predictability of demographic rates based on phylogeny and biological similarity

Lack of demographic data for most of the world's threatened species is a widespread problem that precludes viability‐based status assessments for species conservation. A commonly suggested solution is to use data from species that are closely related or biologically similar to the focal species. This approach assumes that similar species and populations of the same species would have similar demographic rates, an assumption that has yet to be thoroughly tested. In this study, we constructed a Bayesian hierarchical model to predict demographic rates (intrinsic rate of population growth, r; recruit survival; juvenile survival; adult survival; fecundity) based on biological traits and phylogenetic relatedness, using data for 425 plant species. We found generally small effects of species‐level traits (except woody polycarpic species tend to have high adult survival rates that increased with plant height), and a weak phylogenetic signal for four of the five demographic parameters examined. Patterns were stronger in adult survival and fecundity than other demographic rates, however the unexplained variances at both the species and population levels were high for all demographic rates. For species lacking demographic data, our model produced large, often inaccurate, prediction intervals that may not be useful in a management context. Our findings do not support the assumption that biologically similar or closely related species have similar demographic rates, and provide further evidence that direct monitoring of focal species and populations is necessary for informing conservation status assessments.