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…

Protected Designation of Origin (D.P.O.) Valdepeñas: classification and map of soils

The objective of the work described here is the elaboration of a map of the different types of vineyard soils that to guide the famers in the choice of the most productive vine rootstocks and varieties. 90 vineyard soils profiles were analysed in the entire territory of the Origen Denominations of Valdepeñas. The sampling was carried out in 2018 (June to October) by making a sampling grid, followed by photointerpretation and control in the field. The studied soils can be grouped into 9 different soil types (according to FAO 2006 classification): Leptosols, Regosols, Fluvisols, Gleysols, Cambisols, Calcisols, Luvisols and Anthrosols. A map showing the soil distribution with different type of soils has been made with the ArcGIS program. Regarding to the choice of rootstock, Calcisoles are soils with a high active limestone content, so the rootstocks used in these soils must be resistant to this parameter; Luvisols are deep soils with high clay content, so they will support vigorous rootstocks. Because the cartographic units are composed of two or more subgroups, with are associated in variable proportions, 9 different soil associations have been established; Unit 1: Leptosols, Cambisols and Luvisols (80%, 15% and 5% respectively); Unit 2: Cambisols with Regosols and Luvisols (40%, 30% and 30% respectively); Unit 3: Cambisols and Gleysols with Regosols (40%, 40% and 20% respectively); Unit 4: Regosols with Cambisols, Leptosols and Calcisols (40%, 30%, 15% and 15% respectively); Unit 5: Cambisols, Leptosols, Calcisols and Regosols (25% each of them); Unit 6: Luvisols with Cambisol and Calcisols (80%, 10% and 10% respectively); Unit 7: Luvisols and Calcisols with Cambisols (40%, 40% and 20% respectively); Unit 8: Calcisols with, Cambisols and Luvisols (80%, 10% and 10% respectively); Unit 9: Anthrosols. These study allow to elaborate the first map of vineyard soils of this Protected Designation of Origin in Castilla-La Mancha.

δ13C : A still underused indicator in precision viticulture  

The first demonstration of the interest of carbon isotope composition of sugars in grapevine, as an integrated indicator of vineyard water status, dates back to 2000 (Gaudillère et al., 1999; Van Leeuwen et al., 2001). Thanks to the isotopic discrimination of Carbon that takes place during plant photosynthesis, under hydric stress conditions, it is possible to accurately estimate the photosynthetic activity. Ever since, δ13C has been widely applied with success to zonation, terroir studies and vine physiology research, but is still not widely used by viticulturists. This is quite astonishing by considering the impact of global warming on viticulture and the need to improve water management, that would justify a widespread use of δ13C.
The lack of private laboratories proposing the analysis, the cost of the technology, as well as the long analytical delays, have been detrimental to its development. Some laboratories tried to overcome the analytical difficulties of isotopic analysis by using fourier transformed infrared spectroscopy, as a fast and cheap alternative to the official OIV method (IRMS). These claimed FTIR models have never been published or peer reviewed and cannot be considered robust. In this work, thanks to the recent acquisition of IRMS technology, new modern and robust applications of δ13C for viticulture are proposed. This includes the use of the analysis to make parcel separations at harvesting, the possibility to increase the precision of hydric stress cartography and the potential cost reduction when compared with Scholander pressure bomb analysis.

The impact of sustainable management regimes on amino acid profiles in grape juice, grape skin flavonoids, and hydroxycinnamic acids

One of the biggest challenges of agriculture today is maintaining food safety and food quality while providing ecosystem services such as biodiversity conservation, pest and disease control, ensuring water quality and supply, and climate regulation. Organic farming was shown to promote biodiversity and carbon sequestration, and is therefore seen as one possibility of environmentally friendly production. Consumers expect organically grown crops to be free from chemical pesticides and mineral fertilizers and often presume that the quality of organically grown crops is different or higher compared to conventionally grown crops. Integrated, organic, and biodynamic viticulture were compared in a replicated field trial in Geisenheim, Germany (Vitis vinifera L. cv. Riesling). Amino acid profiles in juice, grape skin flavonoids, and hydroxycinnamic acids were monitored over three consecutive seasons beginning 7 years after conversion to organic and biodynamic viticulture, respectively. In addition, parameters such as soil nutrient status, yield, vigor, canopy temperature, and water stress were monitored to draw conclusions on reasons for the observed changes. Results revealed that the different sustainable management regimes highly differed in their amino acid profiles in juice and also in their skin flavonol content, whereas differences in the flavanol and hydroxycinnamic acid content were less pronounced. It is very likely that differences in nutrient status and yield determined amino acid profiles in juice, although all three systems showed similar amounts of mineralized nitrogen in the soil. Canopy structure and temperature in the bunch zone did not differ among treatments and therefore cannot account for the observed differences in favonols. A different light exposure of the bunches in the respective systems due to differences in vigor together with differences in berry size and a different water status of the vines might rather be responsible for the increase in flavonol content under organic and biodynamic viticulture.

Green berries on Gewürztraminer (Vitis vinifera L.) in South Tyrol (Italy)

The grape variety Gewürztraminer is known to be affected by two physiological disorders namely berry shrivel and bunch stem necrosis. During the season 2014 we noticed a new symptomatology type of ripening disorder on the variety. The new symptom showed not all berries fallowing the normal maturation stages, but single berries remaining at a soft but green stage till harvest. The broad distribution of these so called “green berries” symptoms in different production sites of our region, caused huge damage due to the difficulty of eliminating single berries per bunch before harvesting. Therefore, the Research Centre Laimburg began to investigate the reasons and origins of this new symptom. This work shows the results of first attempts to find causes for the symptom as well as the resulting approach to mitigate symptoms. Applications of magnesium leaf fertilizer showed first promising results against this putative disorder. To study the causal effect of the green berries 30 symptomatic vineyards in 2014 have been selected for a monitoring during the season 2016. To evaluate the foliar nutrient treatment two vineyards have been selected for application of magnesium sulfate and magnesium chloride. Leaf and berry nutrient analysis, as well as the main quality parameters during ripening have been performed. As soon as “green berries” symptoms appeared, incidence and severity have been evaluated. Most of the symptomatic vineyards of the 2016 monitoring showed light to clear magnesium deficit symptoms on their foliage. Only during the seasons 2020 and 2021 “green berries” symptoms could be found in the leaf fertilizer treatment vineyards. Both seasons showed a significant effect of the magnesium treatments to reduce the incidence and severity of the symptom. It seems that the appearance of the “green berries” symptom on Gewürztraminer is correlated to a disturbed uptake of magnesium of the vines.

The modification of cultural practices in grapevine cv. Syrah, does it modify the characteristics of the musts?

The work shows the results of a year of experimentation (2020) in a Syrah variety vineyard in La Roda (Castilla-La Mancha, Spain). The trial approach was on a randomized block design with two factors: Irrigation (I) and Pruning (P).
Irrigation schedules were adjusted to apply amounts close to 1,500 m3/ha. With this provision, 2 different irrigation treatments were proposed: I1) Start of irrigation from pea-sized grape to post-harvest (providing at least 20 % of the total amount of irrigation water to be provided post-harvest); I2) Start of irrigation from pea-sized grape to harvest (usual irrigation practice in the study area). Pruning was proposed with two treatments, one at the end of January (P1), which is pruning on a conventional date; and P2) pruning carried out at the beginning of budding. In total, 4 repetitions were designed with 4 elementary plots, each one of them representing one of the proposed treatments (I1P1; I1P2; I2P1; I2P2). In total, 16 plots were worked on and each elementary plot consisted of 30 strains, distributed in 3 lines.
The productive response was evaluated with the yield results of the harvest harvested at 23 ºBrix. The qualitative response was measured in the musts through the indices of technological (acidity, pH and potassium) and phenolic maturity and aromatic compounds in free and glycosylated fractions. The treatments tested had, in general, an effect on the different variables analyzed.