Seed germination model for Eucalyptus delegatensis provenances germinating under conditions of variable temperature and water potential

This paper develops a population-based threshold model to describe the combined action of sub- and supra-optimal temperatures, water stress and the release of dormancy by cool-moist stratification on the germination of seeds of Eucalyptus delegatensis R.T. Baker. Separate models were fitted for seed samples collected from five climatically differing regions. The model presumes that the time to germination of a given seed fraction is inversely proportional to the difference between the actual level of a given germination factor and the factor threshold. The model then assumes that variation in this factor threshold within a seed population as a whole can be characterised by a normal, or log-normal, distribution. By using physiological time rather than clock time as a metric, the model was extended to describe germination under varying conditions in the field. A number of applications of the model were demonstrated. The correlation of model parameters with regional climate was tested and it was concluded that site temperature affected both the mean and variation in base population sensitivities to stratification-dose but that site rainfall affected only the mean base population sensitivity to water stress with all populations having a common variance. Examination of the model parameters relating to the release of dormancy indicated that the increased germination rate associated with stratification could be accounted for by progress towards germination made at stratifying temperatures. Finally, the model was used to examine the fundamental regeneration niche of E. delegatensis and it was concluded that abundant germination in the field could only be expected when the soil water potential is above -0.4 MPa and temperatures exceed 7.5°C. The model is presented as a flexible framework that allows for the prediction of field germination and as a useful tool for exploring seed germination processes, and the fundamental regeneration niche of the species. The modelling framework is easily modified to include additional factors and factor interactions applicable to other situations and species.