Using several flowering phenology models to statistically determine 'Cripps Pink' apple chilling requirement

R. Darbyshire, I. Goodwin, K.S. Pope
Accumulation of sufficient winter chill under anthropogenic climate change may be a constraint for the future production of fruit and nut crops. To evaluate potential impacts and to implement adaptation strategies, reliable estimates of varietal chilling requirements (CR) are needed. Often CRs are estimated statistically as parameters in flowering phenology models. As such, estimated CR may differ depending on model or analysis choice. To test this, historical records of flowering phenology for 'Cripps Pink' apple were obtained from growing districts across southern Australia. These data were used to estimate CR via the sequential model, partial least squares (PLS) regression and the chill overlap model. Different CRs were found depending on the method used. Statistically, the sequential model did not perform well. PLS regression was difficult to interpret at some sites and provided different CR for different sites. The chill overlap model provided a good statistical fit, incorporates biological features of the relationship between chill and heat and aligned with industry experience. This study highlights that CR determined statistically can differ with method choice and hence projected impacts would also differ. Greater understanding of physiological mechanisms is needed to determine CRs that reflect species traits to create reliable estimates of future impacts of climate change.
Darbyshire, R., Goodwin, I. and Pope, K.S. (2017). Using several flowering phenology models to statistically determine 'Cripps Pink' apple chilling requirement. Acta Hortic. 1160, 185-192
DOI: 10.17660/ActaHortic.2017.1160.27
https://doi.org/10.17660/ActaHortic.2017.1160.27
chill overlap model, winter chill, Australia
English

Acta Horticulturae