Author(s): Martins, João Duarte Victor Franco
Date: 2011
Persistent ID: http://hdl.handle.net/10400.5/4131
Origin: Repositório da UTL
Subject(s): vitis vinifera; yield prediction; number of bunches
Author(s): Martins, João Duarte Victor Franco
Date: 2011
Persistent ID: http://hdl.handle.net/10400.5/4131
Origin: Repositório da UTL
Subject(s): vitis vinifera; yield prediction; number of bunches
Mestrado em Viticultura e Enologia - Instituto Superior de Agronomia
With the objective of finding estimated yield models, it was collected data from yield components of a Sauvignon Blanc vineyard variety from Quinta do Pinto, Merceana, Denomination of Origin Alenquer, Lisboa wine region. The method was based on the sampling of 32 spaces that correspond to 192 vines in four different phenological stages (separated flowers, berry pea-size, veraison and ripe bunch). The data obtained was then submitted to an statistical analysis in order to find the correlation coefficients between the several variables and yield. The results indicate significant allometric relations between the inflorescences and bunches variables measured. The number of bunches explains most of the production variability (58%). Based on a stepwise regression, two models were obtained for yield estimation: One based on three variables measured during the berry pea-size stage which explains 57% of the yield variability and other based on two variables measured during the veraison that explain 76%. The model produced at veraison seems to be the more robust and includes fast and easy to determine on the field variables. These models, when adjusted, tend to be an useful tool, allowing early previsions in an easy and economical way.