Development of a hydrothermal time seed germination model which uses the Weibull distribution to describe base water potential

Abstract

Seed germination has been modelled extensively using hydrothermal time (HTT) models, that predict time to germination as a function of the extent to which seedbed temperature, T, and water potential, Ψ, exceed the base temperature, T b, and base water potential, Ψ b, of each seed percentile, g. Within a seed population the variation in time to germination arises from variation in Ψ b( g) modelled by a normal distribution. We tested the assumption of normality in the distribution of Ψ b( g) by germinating seed of two unrelated species with non-dormant seed ( Buddleja davidii (Franch.) and Pinus radiata D. Don) across a range of constant Ψ at sub-optimal T. When incorporated into a HTT model the Weibull distribution more accurately described both the right skewed distribution of Ψ b( g) and germination time course over sub-optimal T than the HTT based on the normal distribution, for both species. Given the flexibility of the Weibull distribution this model not only provides a useful method for predicting germination but also a means of determining the distribution of Ψ b( g).