A statistical study of rain damage of grapes grown for drying in Victoria

Abstract

Effect of harvest rainfall on loss of dried vine fruits was examined by step-wise regression with dried fruit packed per year as the dependent variable and rainfall, summed at half-monthly intervals, as an independent variable. Secular trend variables were fitted to account for trends in annual production. Fitted equations accounted for 81, 77, and 42 per cent of the variation in total production for the years 1925 to 1970. of sultanas, currants, and raisins, respectively. Half the non-trend variance of sultana production and one-third of that of currants and raisins was accounted for by rainfall during statistically selected periods. These were, first and second half of March for sultanas, first half of January and first half of February for currants, and second half of March for raisins. An attempt to enhance accuracy in an analysis of sultana production (Sunraysia district, 1946 to 1971) by including per cent fruitful buds and maximum daily temperatures (summed at six day intervals) in November as additional covariates was only partially successful. Per cent fruitful buds was not significantly related to production but daily maximum temperatures summed over the period November 25-30 inclusive accounted for a small (13 per cent) but significant portion of non-trend variance. A shift in time of maximum damage to sultanas was indicated; from the second half of March in the period 1925 to 1945 to the first half of March since then. These results are discussed in terms of economic importance and type of damage sustained. Ways of reducing the magnitude of the problem are suggested.