Influence of quadrat sampling intensity on precision of grain sorghum yield data

Abstract

Variation in grain yield was studied within 3 contour bays of commercial grain sorghum. The objective was to determine the optimal quadrat sampling strategy for estimating mean grain yield of the whole bay, and of positions within the bay. A series of strategies were identified for attaining coefficients of variation (c.v.) of 5, 10 and 15%, using 9-m2 quadrat samples. For estimating the bay mean at a c.v. of 10%, the outcome preferred was to collect 1 quadrat sample at each of 6 positions (plots) within the bay. To estimate the plot mean at the same precision as the bay mean, 4 quadrat samples would be required at each position. In subsequent crops, a high correlation (r2=0.93) was obtained between estimates of mean grain yield from the chosen strategy and total bay yield from header harvesting. Alternative strategies to suit individual purposes may be identified from the data presented.