Prediction of Diameter, Weight and Quality of Apple Fruit (Malus domestica Borkh.) cv. 'Elstar' using Climatic Variables and their Interactions

The objective of the present work was to determine the relationship between fruit size, fruit weight, fruit quality and climate factors and the interaction of these factors at harvest in order to predict the optimal harvest date to obtain high quality apples after storage. The first evaluation of 'Elstar' apples, included weekly determination of apple diameter and the second evaluation included at harvest measurement of fruit weight and fruit quality (Streif index). All apples were cold-stored in normal atmosphere at 3 °C or in low oxygen atmosphere at 1 °C immediately after picking. Analytical data were combined with daily registered meteorological data including: the sums of degree days, rainfall in the morning including minimum precipitation and/or dew, maximum precipitation (rainfall), evaporation potential, wind velocity, average daily relative humidity, vapour pressure deficit, global radiation, soil temperature and surface wetness. Factor analyses identified three non-correlated factors; 1) energy and water, 2) heat and wind and 3) vapour deficit. The factor loadings were used to determine the importance of each climate variable and its interactions on characteristics of the apple fruit. Multiple regression analyses, with forward selection of climate variables or their interactions showed that apple fruit diameter was significantly affected by two climate factors: 1) degree days and 2) evaporation potential. Only degree day had a significant effect on apple fruit weight. Fruit quality (Streif index) was affected by two climate factors: 1) degree days and 2) relative humidity. Apple fruit diameter, fruit weight and fruit quality were modelled using climate. The differences between measured and predicted diameter, weight and fruit quality were 0.2 mm, 1 g and 0.09 kg cm-2 ww%-1 scores-1, respectively. In view of climate change with expected warmer temperatures, the value establishing and using statistical models for prediction of quality characteristics will increase considerably.