Remote Estimation of Lag Phase in Berries by Monitoring Trellis Tension in Vineyards.

Estimating yield in vineyards before the onset of ripening requires an accurate estimate of the onset of the 'lag' phase (L) in fruit growth: the relatively short period of slowest increase in mass between the first and second sigmoid curves that are used to describe growth in fleshy fruits. Field-scale estimates of yield are computed from scalars under an assumption that fruit mass was measured during L. The standard heuristic approaches to determining that date often lead to unacceptably large errors in the yield estimates. We used an automated, remote system that detected increases in vegetative and fruit mass in grapevines by monitoring the temporal increase in tension in the main load-bearing wire of the trellis. We fitted logistic curves to the season-long change in tension (n = 36) such that the parameters could be interpreted biologically. The curves fit the data well (RMSE 4.23 to 11.84) in three disparate years and two vineyards (two cultivars). Solution of the second derivative of the first logistic curve for its minimum identified the onset of L with a range of 3 to 6 d across years and cultivars. The asymptote of the first logistic curve indicated the onset of ripening, or end of L. Analytical solutions consistently identified the onset of L earlier (2 to 15 d) than dates selected by trained observers who examined individual and mean tension curves. With nearly continuous data, analytical solutions estimated the onset of L at finer temporal resolution (daily) than is possible with field scouting (ca. weekly). Remote sensing of trellis tension can be used to better time field scouting for yield estimation and to decrease the industry's reliance on visual and tactile assessment of berries to identify L.