Hydrothermal sensitivities of seed populations underlie fluctuations of dormancy states in an annual plant community

Plant germination ecology involves continuous interactions between changing environmental conditions and the sensitivity of seed populations to respond to those conditions at a given time. Ecologically meaningful parameters characterizing germination capacity (or dormancy) are needed to advance our understanding of the evolution of germination strategies within plant communities. The germination traits commonly examined (e.g., maximum germination percentage under optimal conditions) may not adequately reflect the critical ecological differences in germination behavior across species, communities, and seasons. In particular, most seeds exhibit primary dormancy at dispersal that is alleviated by exposure to dry after-ripening or to hydrated chilling to enable germination in a subsequent favorable season. Population-based threshold (PBT) models of seed germination enable quantification of patterns of germination timing using parameters based on mechanistic assumptions about the underlying germination physiology. We applied the hydrothermal time (HTT) model, a type of PBT model that integrates environmental temperature and water availability, to study germination physiology in a guild of coexisting desert annual species whose seeds were after-ripened by dry storage under different conditions. We show that HTT assumptions are valid for describing germination physiology in these species, including loss of dormancy during after-ripening. Key HTT parameters, the hydrothermal time constant (θ_HT) and base water potential distribution among seeds (Ψ_b(g)), were effective in describing changes in dormancy states and in clustering species exhibiting similar germination syndromes. θ_HT is an inherent species-specific trait relating to timing of germination that correlates well with long-term field germination fraction, while Ψ_b(g) shifts with depth of dormancy in response to after-ripening and seasonal environmental variation. Predictions based on variation among coexisting species in θ_HT and Ψ_b(g) in laboratory germination tests matched well with 25-yr observations of germination dates and fractions for the same species in natural field conditions. Seed dormancy and germination strategies, which are significant contributors to long-term species demographics under natural conditions, can be represented by readily measurable functional traits underlying variation in germination phenologies.