Enoforum 2021
IVES 9 IVES Conference Series 9 Enoforum Web 9 Enoforum Web Conference 2021 9 A new AI-based system for early and accurate vineyard yield forecasting

A new AI-based system for early and accurate vineyard yield forecasting

Abstract

Vineyard yield forecasting is a key issue for vintage scheduling and optimization of winemaking operations. High errors in yield forecasting can be found in the wine industry, mainly due to the high spatial variability in vineyards, strong dependency on historical yield data, insufficient use of agroclimatic data and inadequate sampling methods. Today, errors can reach values within the range of 20%-30% per block. Thus, improved methodologies for early and accurate vineyard yield forecasting are needed. We proposed a new system for vineyard yield forecasting that integrates: systematic cluster counting, sampling and weight measurement; key agroclimatic parameters; vineyards spatial variability and the use of forecasting models based on artificial intelligence (AI). We carried out trials in high yield Cabernet Sauvignon (CS) vineyards located in Maule Valley (Chile), during seasons 2019 and 2020. We covered 13 blocks (66 ha) and two trellis systems (pergola and free-cordon). We characterized the spatial variability of blocks using Sentinel 2 images and NDVI analysis. We defined sampling units based on NDVI levels and we counted and sampled grape clusters and measured their weights during fruit-set and veraison. Key agroclimatic data were taken from public databases and we collected yield historical data from 2017 onwards. We trained and applied machine-learning models based on MARS, Random Forest and SVR algorithms. For the 2020 trial, in veraison, we obtained an average error of 7.6% per block against a 10.1% given by the traditional method (error is 23.5% for all the CS grapes of the company). Time dedicated to counting and sampling was significantly lower. As a result, we obtained a cost-efficient, early and accurate new system for vineyard yield forecasting.

DOI:

Publication date: April 23, 2021

Issue: Enoforum 2021

Type: Article

Authors

Cuevas-Valenzuela, José1*; Caris-Maldonado, Carlos1; Reyes-Suárez, José Antonio2; González-Rojas, Álvaro1

1 Center for Research and Innovation (CRI) Viña Concha y Toro, Ruta k-650 km 10, Pencahue, Maule, Chile
2 Bioinformatics Department, Faculty of Engineering, Universidad de Talca, Campus Lircay, Talca, Maule, Chile

Contact the author

Tags

Enoforum 2021 | IVES Conference Series

Citation

Related articles…

Utilisation de données historiques pour caractériser le millésime en cours

Cet article propose la formalisation d’un modèle paramétrique pour représenter l’accumulation des sucres dans les baies de raisin durant la maturation. Le test de ce modèle sur des jeux de données réels a permis de valider l’approche proposée. Une seconde partie est axée sur l’adaptation de la méthode pour permettre la simulation du comportement du millésime en cours dès les premiers relevés de maturité. Ce travail possède de multiples applications dans le domaine de l’aide à la décision.
Read More

Impact of yeast strains on wine profiles of nine PIWIs: focus on volatile thiols

Disease resistant grapevine varieties (PIWI) are increasingly important for sustainable wine production, yet the impact of different yeasts on their wine profiles remains poorly studied. In this study, nine white interspecies varieties (i.e., caladris blanc, fleurtai, hibernal, johanniter, muscaris, sauvignon kretos, soreli, souvignier gris, and voltis) grown at the faculty of agriculture, university of Zagreb (Croatia) were vinified with three different saccharomyces cerevisiae yeasts (control strain, zymaflore x5, and zymaflore xarom).

Read More

Influence of canopy management on yield, grape and wine quality. Relationship between the potassium content and pH in must and wine of the cultivar “Tempranillo”

In recent years red wines are being produced in Andalusia from indigenous and foreign grape varieties, one of which is the Spanish variety Tempranillo.

Read More

Viticultural zoning in the province of San Juan, Argentina. Preliminary results, year 2000

La région viticole de San Juan (Argentine) est marquée par des températures très élevées et des variations diurnes faibles. La valorisation de la connaissance de cet environnement et de ses interactions avec le fonctionnement de la vigne et le lien au vin passent par l’étude de ses terroirs et de leur caractérisation. Le point de départ de ce travail est l’étude des zones mésoclimatiques aptes à la culture de la vigne de la Province de San Juan et à la caractérisation des sols de cette même région. L’objectif est de définir le potentiel vitivinicole des zones considérées.

Read More

SENSORY IMPROVEMENT OF DEALCOHOLISED WINES

Interest and willing-ness to buy alcohol-free wines by customers is increasing for several years [1]. Due to the rising relevance of dealcoholised wines it is the objective of this study to contribute to a better understanding of the flavor variation among dealcoholised wines and to explore enological measures, how to improve final quality.
First a range of commercial, alcoholfree white wines were analysed by the holistic sensory method projective mapping, including a question for hedonic acceptance. Based on the combination of a non-target-HS-SPME-GC/MS analysis with sensory analysis we obtained a clustering of the wines into three groups.

Read More