Sampling

Krstic, Mark. (2004, July/August). Sampling for maturity and colour. Australian Viticulture, pp 75-77.

The harvest decision is the single most important decision made by the grape grower or winemaker with respect to a particular vineyard.

The crop is usually sampled on a regular basis to enable the harvest decision to be made. For a statistically sound sampling strategy to be employed the sampler requires a detailed understanding of all potential sources of variation.

The author's research suggests that a representative 20 bunch sample could deliver a 4% error margin for maturity sampling. A representative 40. berry sample would deliver a similar error.

The paper examined the variation in brix and total anthocyanin levels with respect to yield within a single vineyard and found that while yield may vary 5 - 6 fold the relative change in brix can be pretty small ie 22 - 22.5 Brix. The variation in anthocyanin levels can be 3 - 4 fold ie 0.8 - 2.4 mg/g berry weight. Therefore a different sample size would be required to estimate brix, total anthocyanins or yield to a similar error level.

The paper also examined the variation between brix and total anthocyanins across a range of vineyards. It is well-known that a number of differences exist between vineyards such as soil type, vineyard area, crop load etc and these all contribute to the variability seen between vineyards. It was found that the variability was again greater with total anthocyanins than with brix.

The effect of vineyards size on variation was investigated and it would appear that sampling intensity does not need to be adjusted according to the area of individual vineyards.

While the ideal would be to know exactly the variations in the parameters being tested so that error and waste could be minimised it is not currently practicable to do so.

The paper concludes that a 20 bunch sample can provide an estimate of the mean for most vineyards with a 3.5% level of error for brix and 7.6% level of error for total anthocyanins.

References