Validation of winter chill models using historic records of walnut phenology

Many fruit and nut species require cold temperatures during the dormancy season to initiate flowering and bear fruit. Quantifying these chilling requirements is crucial for identifying appropriate cultivars for a given site, for timing applications of rest-breaking chemicals and for predicting consequences of climate change. We present a new method to test temperature models describing chilling and heat requirements of perennial plants, and use this method to compare the ability of four chilling models (Chilling Hours, Utah Model, Positive Utah Model and Dynamic Model) to explain walnut phenology in California. When plotting remaining heat before a phenological stage is reached against accumulated winter chill, observational curves for all years should intersect in one common point, assuming fixed chilling and heat requirements and a sequential fulfillment of these requirements. This point defines the chilling and forcing requirements of the plant, and the quality of the chilling/heat model combination is indicated by how well defined the intersection point is. We used this method on a total of 1297 phenological observations, including four walnut cultivars, seven phenological stages and eight locations in California. Using an hourly temperature record, winter chill was quantified by the four chilling models and remaining heat was estimated using the Growing Degree Hour concept. The theoretical intersection point was more clearly defined for the Dynamic and Positive Utah Models than for the Chilling Hours and Utah Models in almost all cases, indicating that these are superior in explaining walnut phenology. It was also apparent that chilling models were not equivalent and that chilling requirements determined under constant temperature conditions, when quantified in Chilling Hours, were not representative of chilling requirements in orchards.