Submitted to: Transactions of the ASAE
Publication Type: Peer reviewed journal
Publication Acceptance Date: 12/5/2003
Publication Date: 5/24/2004
Citation: Tarara, J.M., Ferguson, J.C., Blom, P.E., Pitts, M.J., Pierce, F.J. 2004. Estimation of grapevine crop mass and yield via automated measurements of trellis tension. Transactions of the American Society of Agricultural Engineers. 47(2) p.647-657. Interpretive Summary: Grape growers, juice processors, and wineries require accurate estimates of yield as early as possible in the growing season because grapes are a high-value crop, are perishable, and require expensive equipment for processing. The standard method of estimating yield requires a grower or processor to collect grapes by hand, count the number of berries per cluster, and hand weigh the berries. The procedure is slow, expensive, and provides limited information. Using an instrument to automate yield estimation could increase production efficiency and provide continuous information about the size of the crop, allowing growers to make better predictions of yield and better judgments about whether to thin a crop that is too large. We devised a new approach that automates yield estimation in a vineyard by measuring the tension in support wire of the trellis, on which most commercial grapevines grow. Because the trellis wire supports the vine, as the vine grows, tension on wire increases. Likewise, as the fruit clusters grow larger, their weight can be related to the corresponding increase in wire tension. Environmental effects on the wire, like temperature and wind, must be accounted for. The method had acceptable accuracy in a single-wire trellis and needs to be tested on complex trellises where the tension can be distributed among multiple wires.
Technical Abstract: A novel approach was devised to estimate vegetative growth and fruit mass (i.e., yield) in grapevines by measuring continuously the tension in the horizontal (cordon) support wire of the trellis. Load cells installed in-line with the cordon wire were connected to an automated data acquisition system, a major departure from the viticulture industry's standard method of collecting fruit samples by hand two or three times per growing season. Each experimental row in the vineyard was calibrated to determine the change in tension in that row in response to an increase in known mass on the cordon wire. The effects of temperature on wire tension were removed by post-processing with a regression-based empirical protocol, which corrected the raw data to a standard temperature. Because data were averaged over 15 min., wind gusts appeared to have little measurable effect on the tension measurement. A smoothing algorithm removed remaining transient disturbances in the data without masking significant events like crop thinning or harvest. Results to date suggest a linear relationship between wire tension and fruit mass that varies among rows, but not within a row during a single season. Yields were between 4.5 and 20.2 t ha-1 (2 to 9 t acre-1).