Development and testing of a new chilling model for low chill conditions

In order to develop an accurate Chilling Model for low chill conditions a data base of progression of dormancy patters relative to temperature was produced. This data was derived from 24 sites throughout the apple growing regions of South Africa over a 5 year period using the cultivars Granny Smith and Royal Gala. A total of 204 dormancy progression graphs were produced. An 85,7 % fitted model was obtained when SAS was used to fit a two step mathematical model to the data. Eleven criteria (e.g. start day of dormancy, date of maximum dormancy, depth of dormancy and end day of dormancy) associated with the model were identified, standardized and were used to cluster the 24 sites (farms). Three distinct clusters were obtained. Average hourly temperatures for all the sites were collected and verified to produce an accurate data set. Numerous statistical processes have been applied to the two data sets to try to emulate these clusters using the temperature data. This is an ongoing process and advanced statistical modelling is required to develop the 2 stage chilling model accurately. A number of options are being considered to continue this project in the field of modelling statistics. The two stage chilling model should accurately predict the entrance into dormancy, the date of “maximum dormancy” and be able to quantify the extent of the subsequent release from dormancy. The model should therefore ultimately predict the inadequate chilling for each season and farm cluster.