|Chaves, Bernardo -|
|Sanchez, Luis -|
|Dokoozlian, Nick -|
Submitted to: Good Fruit Grower
Publication Type: Trade Journal
Publication Acceptance Date: December 30, 2013
Publication Date: January 29, 2014
Citation: Tarara, J.M., Chaves, B., Sanchez, L., Dokoozlian, N. 2014. Automatically determining lag phase in grapes to assist yield estimation practices. Good Fruit Grower. February 1st 2014 Issue. Interpretive Summary: Estimating crop yield in a vineyard is an important but often difficult task. Accurate projections ensure that enough processing equipment will be available because juice and wine grapes must be processed immediately and cannot be stored for later handling. Yield estimation is based on counting the clusters on a grapevine then weighing them. This is labor intensive and must be completed during a very short window of time because the stage of growth in the berries when this estimation is made may only be a few days. Growers must cover hundreds to thousands of acres rapidly for the highest likelihood of accurate yield estimates. A technology developed by USDA-ARS was used in two commercial vineyards to indicate the increasing weight of the crop from the tension in the wire that supports the grapevines on their trellis. The device is automated and provides daily data, something that is not presently available to growers. We mathematically process the device's output to produce this estimate of crop growth. To help growers with traditional sampling techniques, meaning manual sampling, we estimated the date at which this sampling should begin. We used a calculus-based mathematical analysis of the wire tension data to identify this date. We then compared growers' selected dates with those identified mathematically. On the whole, the industry representatives selected dates later than the mathematical solution, with an average of overestimation of seven days. Because the growth pattern of grape berries is well known, finding techniques like monitoring trellis wire tension could remove some of the guesswork involved in estimating the date for starting to sample for yield estimation. A technology that provides daily updates improves a grower's ability to make rapid management decisions, in this case better managing the critical timing for estimating yield.
Technical Abstract: Estimating grapevine yield is an important though often difficult task. Accurate yield projections ensure that enough physical infrastructure is available to process fruit that cannot be stored for later handling. Crop estimation is based on recording cluster numbers and cluster weights of representative samples in each vineyard. The task is labor intensive and typically must be completed during a short window, often only a few days. The 'lag phase', the period of slowest berry growth, often is used as the trigger to commence sampling for yield estimation so that established scalars are correctly applied. Lag phase is between the first period of rapid growth after flowering and the second period of rapid growth during ripening, the well-known double-sigmoid curve of growth in fleshy fruits. The trellis-tension technology, developed by USDA-ARS, involves an automated device that continuously monitors and records tension in the main load-bearing wire of the trellis. After post-processing the output to remove signal noise, the data produce an indirect indicator of vine and fruit growth and show the distinctive double-sigmoid curve. We defined the onset of lag phase as the point at which the slope of the first curve approaches zero, determined by taking its second derivative. We compared industry representatives' selected dates for the onset of lag phase with those identified analytically. On the whole, the industry representatives selected dates later than the analytical solution, with an average of overestimation of seven days. Finding techniques like monitoring wire tension in a load-bearing wire on a trellised crop could remove some of the guesswork involved in estimating the onset of lag phase in the field. Any technology with continuous data improves the industry's real-time management decisions--in this case, better managing the critical timing for estimating yield.