Grapevine yield variability: contribution of grape load and effects of the previous season’s growing conditions

Abstract

The grapevine yield may be highly variable over years, even in the same plot with the same cultural practices. The determinism of yield formation of grapevine is complex as it takes place over two consecutive years, starting around flowering at year n-1 with the initiation and differentiation of the inflorescences of year n in the latent buds. Therefore forecasting grapevine yield is still a challenge for grape growers. In this study the following hypotheses, based on a review of literature, were tested: (i) the number of bunches per vine (NBv) is the yield component that accounts for the main part of yield variability, and (ii) the potential NBv of current year (year n) is determined by water and nitrogen stresses experienced by grapevine around the previous year (year n-1) flowering period. These hypotheses were tested on two grape varieties, shiraz and aranel, over three and six years, respectively, in Languedoc-Roussillon, south of France. For each variety different treatments were applied each year, i.e. chemical or mechanical weeding, grass cover, irrigation and/or fertilization. The considered yield components were NBv, number of berries per bunch and berry weight. They were measured once at harvest. Hypothesis (i) was tested by performing a multiple linear regression analysis for each variety. NBv accounted for around 58 % of yield variability for cv. shiraz and 20% for cv. aranel. Hypothesis (ii) was tested by looking for correlations between the number of bunches per shoot (NBs), the predawn leaf water potential (ψ_predawn), and the leaves nitrogen content (LNC) around flowering. Our results show that NBs of year n is well and better correlated with ψ_predawn and LNC of year n-1 than those of year n. This study shows that the potential yield of year n is settled around flowering of year n-1 through bud fertility (NBs). This potential could be assessed by measuring or estimating water and nitrogen stresses. This early information could be used to adapt pruning practices depending on the target yield of the next year.