A time-to-event analysis for temperature dependence of seed germination in four conifers: Ecological niche and environmental gradients

Seed germination is a vital process in the plant life cycle. It directly influences seedling establishment and subsequently determines population persistence and species distribution. Germination is highly sensitive to climatic conditions, and germination traits may vary among species and populations within the same species. This study aims to improve our understanding of intraspecific and interspecific variation in temperature dependence of seed germination and how this variation relates to the environment where seeds were collected. We focused on four conifer species (Pinus contorta, Picea engelmannii, Pinus ponderosa, and Pseudotsuga menziesii) and seeds collected at 20 sites spanning the elevation range of each species in the Western United States. We conducted seed germination tests at temperatures from 5 to 40 degrees C in 5 degrees C intervals and tracked germination for 30 days. We used the log-logistic time-to-event model and Gaussian function to fit germination data and analyze their temperature dependence. The climate variables at the seed-collecting sites were summarized by long-term normals (1971-2000) covering most seed-collection years. Climate variables were correlated and thus were combined via Principal Component Analysis (PCA) into several uncorrelated principal components (PC scores) that describe climate at the seed collection sites. The estimated optimal temperature (To), shape coefficient (b), and maximum value at optimal temperature (GPmax) of germination percentages and median germination rates (GR50) were then fit to linear models to determine whether they varied with species identity and the climatederived PC scores. Results showed that seed germination traits from different species exhibit climate gradients that are correlated with their ecological niche. However, seeds from different elevations within the same species had variable germination traits irrespective of environmental gradients. We also found that variation in germination was more significant near upper and lower temperature limits. Including inter- and intra-specific temperature dependence of seed germination would improve predictions of population dynamics and range shifts under climate warming and could inform forest management, seed transfer, and assisted migration practices.