Flowering prediction of pear tree (Pyrus communis L.), apple tree (Malus domestica Borkh) and plum tree (Prunus domestica L.) - similarities and differences

Abstract

Paper discusses the similarities and differences in predicting the full flowering of apple tree (cv. 'Bobovec'), pear tree (cv. 'Pastorjevka') and domestic plum tree. The study was conducted at University of Ljubljana. Input data were standard weather and phenological observations of Hydrometeorological Institute of Slovenia. First average monthly air temperature and average monthly amount of precipitation from January to April were used as predictors for flowering of fruit trees during the time period 1967-1996 and in second case input data for models were phenological data of autochthonous plants: first leaves unfolded of Betula pendula ROTH, Fagus sylvatica L. and Tilia platyphyllos Scop. Two different approaches were employed for predicting of full flowering of apple tree, pear tree and plum tree: the linear multiple regression and correlation analysis. The date of full flowering for all tree species is strongly correlated with the first leaves unfolding of birch (R2 is from 0.61 to 0.91) and in case of pear tree and apple tree also with first leaves unfolding of beech (R2 is from 0.62 to 0.66). Also strong correlation with combination of average March and April air temperature was found for flowering of apple tree and plum tree, for pear tree the full flowering was correlated with combination of average February and March air temperatures. In case of apple tree 76% of total variance was explained by linear multiple regression model between full flowering and air temperatures, for plum tree 55% and for pear tree even 91%. The multiple regression model for predicting the full flowering on the base of March and April temperatures have great disadvantage, because the predictions can be made in case of apple tree only 2 days ahead and in case of plum tree not even that. On the other hand, the prediction of flowering for pear tree can be made 22 days ahead. The full flowering for apple tree and plum tree in our case were not correlated to average monthly amount of precipitation at all. This could be expected since in Slovenia, water availability in the spring is sufficient for plant needs.

Keywords

• phenology
• statistical models
• slovenia