GiESCO 2019 banner
IVES 9 IVES Conference Series 9 GiESCO 9 GiESCO 2019 9 Aroma and quality assessment for vertical vintages using machine learning modelling based on weather and management information

Aroma and quality assessment for vertical vintages using machine learning modelling based on weather and management information

Abstract

Context and purpose of the study ‐ Wine quality traits are usually given by parameters such as aroma profile, total acidity, alcohol content, colour and phenolic content, among others. These are highly dependent on the weather conditions during the growing season and management strategies. Therefore, it is important to develop predictive models using machine learning (ML) algorithms to assess and predict wine quality traits before the winemaking process.

Material and methods ‐ Samples in duplicates of Pinot Noir wines from vertical vintages (2008 to 2013) of the same winery located in Macedon Ranges, Victoria, Australia were used to assess different chemical analytics such as i) aromas using gas chromatography – mass spectrometry, ii) color density, iii) color hue, iv) degree of red pigmentation, v) total red pigments, vi) total phenolics, vii) pH, viii) total acidity (TA), and ix) alcohol content. Data from weather conditions from the specific vintages were obtained both from the bureau of meteorology (BoM) and the Australian Wine Availability Project (AWAP) climate databases. Such data consisted of: i) solar exposure from veraison to harvest (V‐H), ii) solar exposure from September to harvest (S‐H), iii) maximum January solar exposure, iv) degree days from S‐H, v) maximum January evaporation, vi) mean maximum temperature from veraison to harvest, vii) mean minimum temperature from V‐H, viii) water balance from S‐H, ix) solar exposure from V‐H, x) degree hour accumulation with base 25 – 30 °C, xi) degree hour accumulation with base 25 °C, xii) degree hour accumulation with base 30 °C, xiii) degree hour accumulation with base 35 °C, and xiv) total cumulative degree days accumulation with base 10 °C. All data were used to develop two machine learning (ML) regression models using Matlab® R2018b. The best models obtained were using artificial neural networks (ANN) with the Levenberg‐Marquardt algorithm with 5 neurons for Model 1 and 9 neurons for Model 2. Model 1 was developed using the 14 parameters from the weather data as inputs to predict 21 aromas found in the wines from the six different vinatges. Model 2 was developed using the same 14 parameters from weather data and the eight chemical parameters as targets and outputs.

Results ‐ Both models obtained presented very high accuracy to predict wine quality trait parameters. Model 1 had an overall correlation coefficient R = 0.99 with a high performance based on the mean squared error (MSE = 0.01), while Model 2 had an overall correlation coefficient R = 0.98 with a high performance (MSE = 0.03). These models would aid in the prediction of wine quality traits before its production, which would give anticipated information to winemakers about the product they would obtain to make early decisions on wine style variations.

DOI:

Publication date: June 22, 2020

Issue: GiESCO 2019

Type: Article

Authors

Sigfredo FUENTES, Claudia GONZALEZ VIEJO, Xiaoyi WANG, Damir D. TORRICO

School of Agriculture and Food, Faculty of Veterinary and Agricultural Sciences, University of Melbourne, VIC 3010, Australia

Contact the author

Keywords

wine quality, machine learning, weather, aromas

Tags

GiESCO 2019 | IVES Conference Series

Citation

Related articles…

A blueprint for managing vine physiological balance at different spatial and temporal scales in Champagne

In Champagne, the vine adaptation to different climatic and technical changes during these last 20 years can be seen through physiological balance disruptions. These disruptions emphasize the general grapevine decline. Since the 2000s, among other nitrogen stress indicators, the must nitrogen has been decreasing. The combination of restricted mineral fertilizers and herbicide use, the growing variability of spring rainfall, the increasing thermal stress as well as the soil type heterogeneity are only a few underlying factors that trigger loss of physiological balance in the vineyards. It is important to weigh and quantify the impact of these factors on the vine. In order to do so, the Comité Champagne uses two key-tools: networking and modelization. The use of quantitative and harmonized ecophysiological indicators is necessary, especially in large spatial scales such as the Champagne appellation. A working group with different professional structures of Champagne has been launched by the Comité Champagne in order to create a common ecophysiology protocol and thus monitor the vine physiology, yearly, around 100 plots, with various cultural practices and types of soil. The use of crop modelling to follow the vine physiological balance within different pedoclimatic conditions enables to understand the present balance but also predict the possible disruptions to come in future climatic scenarios. The physiological references created each year through the working group, benefit the calibration of the STICS model used in Champagne. In return, the model delivers ecophysiology indicators, on a daily scale and can be used on very different types of soils. This study will present the bottom-up method used to give accurate information on the impacts of soil, climate and cultural practices on vine physiology.

Effects of graft quality on growth and grapevine-water relations

Climate change is challenging viticulture worldwide compromising its sustainability due to warmer temperatures and the increased frequency of extreme events. Grafting Vitis vinifera L.

Grape must quality and mesoclimatic variability in Fruška Gora wine-growing region, Serbia

The Fruška Gora mountain is a traditional wine-growing region in Serbia situated in the Pannonian Basin. Due to such a position, the vicinity of the Danube River and the presence of concave configuration, it is suitable for grape production. This paper provides analyses of spatial variations in meteorological parameters and grape juice quality within Fruška Gora wine region over three consecutive vintages (2018-2020). The examined period can be defined as warm with cool nights during September (AVG 18,9°C; GDD 1918°C; CI 12°CF) and with the presence of mesoclimatic variability. The East part of the study area was somewhat drier and hotter compared to other parts of the region. The analyses of grape must samples (190 in total) of five cultivars (Cabernet-Sauvignon, Merlot, Chardonnay, Sauvignon blanc and Grašac (Welschriesling)) commonly grown across the region (19 sites), were performed using Fourier Transform Infrared Technology (FTIR). Among all cultivars, Sauvignon blanc was harvested first in the East area (DOY=246±5, GDD at harvest=1552±74, 22.2±0.7 °Brix), while the latest harvest was recorded for Cabernet-Sauvignon in the West (DOY=283±5, GDD at harvest=1936±187, 23.4±1.0 °Brix ). Both the red and white cultivars had higher acidity and YAN in the grape must if the vines were grown in the North and East compared to South and West areas. According to PCA analysis, Grašac showed the lowest variation in grape must chemical composition. Thus, the results confirm that Grašac is the most stable cultivar in Fruška Gora. All monitored cultivars reached technological fruit ripeness by the end of the growing season. However, it was difficult to reach full ripeness of red cultivars, mostly beacuse of uncoupling of technolocical and phenolic ripeness. Thus, Cabernet-Sauvignon had higher variations in GDD sums at harvest compared to other cultivars, which probably increased variations in grape must quality.

Updating the Winkler index: An analysis of Cabernet sauvignon in Napa Valley’s varied and changing climate

This study aims to create an updated, agile viticultural climate index (similar to the Winkler Index) by performing in-depth analyses of current and historical data from industry partners in several major winegrowing regions. The Winkler Index was developed in the early twentieth century based on analysis of various grape-growing regions in California. The index uses heat accumulation (i.e. Growing Degree Days) throughout the growing season to determine which grape varieties are best suited to each region. As viticultural regions are increasingly subject to the complexity and uncertainty of a changing climate, a more rigorous, agile model is needed to aid grape growers in determining which cultivars to plant where. For the first phase of this study, 21 industry partners throughout Napa Valley shared historical phenology, harvest, viticultural practice, and weather data related to their Cabernet sauvignon vineyard blocks. To complement this data, berry samples were collected throughout the 2021 growing season from 50 vineyard blocks located throughout 16 American Viticultural Areas that were then analyzed for basic berry chemistry and phenolics. These blocks have been mapped using a Geographic Information System (GIS), enabling analysis of altitude, vineyard row orientation, slope, and remotely sensed climate data. Sampling sites were also chosen based on their proximity to a weather station. By analyzing historical data from industry partners and data specifically collected for this study, it is possible to identify key parameters for further analysis. Initial results indicate extreme variability at a high spatial resolution not currently accounted for in modern viticultural climate indices and suggest that viticultural practices play a major role. Using the structure of data collection and analyses developed for the first phase, this project will soon be expanded to other wine regions globally, while continuing data collection in Napa Valley.

Anthocyanin profile is differentially affected by high temperature, elevated CO2 and water deficit in Tempranillo (Vitis vinifera L.) clones

Anthocyanin potential of grape berries is an important quality factor in wine production. Anthocyanin concentration and profile differ among varieties but it also depends on the environmental conditions, which are expected to be greatly modified by climate change in the future. These modifications may significantly modify the biochemical composition of berries at harvest, and thus wine typicity. Among the diverse approaches proposed to reduce the potential negative effects that climate change may have on grape quality, genetic diversity among clones can represent a source of potential candidates to select better adapted plant material for future climatic conditions. The effects of individual and combined factors associated to climate change (increase of temperature, rise of air CO2 concentration and water deficit) on the anthocyanin profile of different clones of Tempranillo that differ in the length of their reproductive cycle were studied. The aim was to highlight those clones more adapted to maintain specific Tempranillo typicity in the future. Fruit-bearing cuttings were grown in controlled conditions under two temperatures (ambient temperature versus ambient temperature + 4ºC), two CO2 levels (400 ppm versus 700 ppm) and two water regimes (well-watered versus water deficit), both in combination or independently, in order to simulate future climate change scenarios. Elevated temperature increased anthocyanin acylation, whereas elevated CO2 and water deficit favoured the accumulation of malvidin derivatives, as well as the acylation and tri-hydroxylation level of anthocyanins. Although the changes in anthocyanin profile observed followed a common pattern among clones, such impact of environmental conditions was especially noticeable in one of the most widely distributed Tempranillo clones, the accession RJ43.