GiESCO 2019 banner
IVES 9 IVES Conference Series 9 GiESCO 9 Machines and fire: developing a rapid detection system for grapevine smoke contamination using NIR spectroscopy and machine learning modelling

Machines and fire: developing a rapid detection system for grapevine smoke contamination using NIR spectroscopy and machine learning modelling

Abstract

Context and purpose of the study – Bushfires are a common occurrence throughout Australia and their incidence is predicted to both rise and increase in severity due to climate change. Many of these bushfires occur in areas close to wine regions, which receive different levels of exposure to smoke. Wine produced from smoke-affected grapes are characterised by unpalatable smoky aromas such as “burning rubber”, “smoked meats” and “burnt wood”. These smoke tainted wines are unprofitable and result in significant financial losses for winegrowers. This study investigated the use of near-infrared (NIR) spectroscopy and machine learning (ML) modelling for the rapid and non-destructive detection of grapevine smoke exposure by analysing grapevine leaves and/or grape berries.

Materials and methods – The trial was conducted during the 2018/2019 season at the University of Adelaide’s Waite campus in Adelaide, South Australia (34° 58’ S, 138° 38’ E) and involved the application of five different smoke and water misting treatments to Cabernet Sauvignon grapevines at approximately seven days post-veraison. Treatment vines were exposed to straw-based smoke for one hour under experimental conditions described previously by Kennison et al. (2008) and Ristic et al. (2011). Near-infrared (NIR) measurements were then taken from berries and leaves a day after smoking using the microPHAZIR TM RX NIR Analyser (Thermo Fisher Scientific, Waltham, USA) which has a spectral range of 1600-2396 nm. The NIR spectra were then used as inputs to train different ML algorithms, which resulted in two artificial neural networks (ANNs) with the best classification performance for either berry or leaf readings according to the different smoke treatments.

Results – Both ANN models found were able to correctly classify the leaf and berry spectral readings with high accuracy. The leaf model had an overall accuracy of 95.2%, 97.7% accuracy during training with a mean square error (MSE) 0.0082, 90.9% during validation with a MSE of 0.0353 and 88.1% during the testing stage with a MSE of 0.0386, while the berry model had an overall accuracy of 91.7%, 95.2% accuracy during training with a MSE of 0.0173, 86.4% during validation with a MSE of 0.0560 and 80.2% during the testing stage with a MSE of 0.0560. These results showed the potential of developing a rapid, non-destructive, in-field detection system for assessing grapevine smoke contamination following a bushfire using NIR spectroscopy and artificial neural network modelling.

DOI:

Publication date: September 28, 2023

Issue: GiESCO 2019

Type: Poster

Authors

Vasiliki SUMMERSON, Claudia GONZALEZ VIEJO, Damir TORRICO, Sigfredo FUENTES*

The University of Melbourne, School of Agriculture and Food, Faculty of Veterinary and Agricultural Sciences, Parkville 3010, Victoria, Australia

Contact the author

Keywords

bushfires, machine learning, smoke taint, climate change, non-destructive

Tags

GiESCO | GiESCO 2019 | IVES Conference Series

Citation

Related articles…

Projected changes in vine phenology of two varieties with different thermal requirements cultivated in La Mancha DO (Spain) under climate change scenarios

The aim of this work was to analyze the phenology variability of Tempranillo and Chardonnay cultivars, related to the climatic characteristics in La Mancha Designation of Origin, and their potential changes under climate change scenarios. Phenological dates referred to budbreak, flowering, veraison and harvest were analyzed for the period 2000-2019. The weather conditions at daily time scale, recorded during the same period, were also evaluated. The thermal requirements to reach each of these phenological stages were calculated and expressed as the GDD accumulated from DOY=60. Changes in phenology were projected by 2050 and 2070 taking into account those values and the projected temperatures and precipitation, simulated under two Representative Concentration Pathway (RCP) scenarios –RCP4.5 and RCP8.5– using an ensemble of models. The average phenological dates during the period under study were, April 16th ± 6.6 days and April 5th ± 6.0 days for budbreak, May 31st ± 6.0 days and May 27th ± 5.3 days for flowering, July 26th ± 5.6 days and July 25th ± 5.8 days for veraison, and Ago 23rd ± 10.8 days and Ago 17th ± 9.0 days for harvest, respectively, for Tempranillo and Chardonnay. The projected changes in temperature imply an average change in the maximum growing season (April-August) temperatures of 1.2 and 1.9°C by 2050, and 1.6 and 2.6°C by 2070, under the RCP4.5 and RCP8.5 scenarios, respectively. A reduction in precipitation is predicted, which vary between 15% for 2050 under RCP4.5 scenario and up to 30% by 2070 under RCP8.5. The advance of the phenological dates for 2050, could be of 6, 7, 7, and 8 days for Tempranillo and 4, 6, 6 and 9 days for Chardonnay, respectively for budbreak, flowering, veraison and harvest under the RCP4.5 scenario. Under the RCP8.5 emission scenario, the advance could be up to 30% higher.

Sustainable fertilisation of the vineyard in Galicia (Spain)

Excessive fertilization of the vineyard leads to low quality grapes, increased costs and a negative impact on the environment. In order to establish an integrated management system aimed at a sustainable fertilization of the vineyards, nutritional reference levels were established. For this purpose, 30 representative vineyards of the Albariño variety were studied, in which soil and petiole analyses were carried out for two years and grape yield and quality at harvest were measured. In both years of study, soil pH, calcium, sodium and cation exchange capacity were positively correlated with calcium content and negatively correlated with manganese in grapes. Irrigated vineyards had higher levels of aluminium in soil and lower levels of calcium in petiole. Climatic conditions were very different in the years of the study. The year 2019 was colder than usual, in 2020 there was a marked water stress with high summer temperatures. This resulted in medium-high acidity in grapes in 2019 and low acidity in 2020, with sugar levels being similar both years. A very marked decrease in must amino nitrogen was observed in 2020, with ammonia nitrogen remaining stable. The correlation of acidity and sugar values in grapes with soil and petiole analysis data made it possible to establish reference levels for the nutritional diagnosis of the Albariño variety in this region. Based on these results, an easy-to-use TIC application is currently being created for grapegrowers, aimed at improving the sustainability of the vineyard through reasoned fertilization. This study has now been extended to other Galician vine varieties.

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.

VineyardFACE: Investigation of a moderate (+20%) increase of ambient CO2 level on berry ripening dynamics and fruit composition

Climate change and rising atmospheric carbon dioxide concentration is a concern for agriculture, including viticulture. Studies on elevated carbon dioxide have already been on grapevines, mainly taking place in greenhouses using potted plants or using field grown vines under higher CO2 enrichment, i.e. >650 ppm. The VineyardFACE, located at Hochschule Geisenheim University, is an open field Free Air CO2 Enrichment (FACE) experimental set-up designed to study the effects of elevated carbon dioxide using field grown vines (Vitis vinifera L. cvs. Riesling and Cabernet Sauvignon). As the carbon dioxide fumigation started in 2014, the long term effects of elevated carbon dioxide treatment can be investigated on berry ripening parameters and fruit metabolic composition.
The present study aims to investigate the effect on fruit composition under a moderate increase (+20%; eCO2) of carbon dioxide concentration, as predicted for 2050 on both Riesling and Cabernet Sauvignon. Berry composition was determined for primary (sugars, organic acids, amino acids) and secondary metabolites (anthocyanins). Special focus was given on monitoring of berry diameter and ripening rates throughout three growing seasons. Compared to previous results of the early adaptative phase of the vines [1], our results show little effect of eCO2 treatment on primary metabolites composition in berries. However, total anthocyanins concentration in berry skin was lower for eCO2 treatment in 2020, although the ratio between anthocyanins derivatives did not differ.
[1] Wohlfahrt Y., Tittmann S., Schmidt D., Rauhut D., Honermeier B., Stoll M. (2020) The effect of elevated CO2 on berry development and bunch structure of Vitis vinifera L. cvs. Riesling and Cabernet Sauvignon. Applied Science Basel 10: 2486

1H-NMR-based Metabolomics to assess the impact of soil type on the chemical composition of Mediterranean red wines

The aim of this study was to evaluate the effects of different soil types on the chemical composition of Mediterranean red wines, through untargeted and targeted 1H-NMR metabolomics. One milliliter of raw wine was analyzed by means of a Bruker Avance II 400 spectrometer operating at 400.15 MHz. The spectra were recorded by applying the NOESYGPPS1D pulse sequency, to achieve water and ethanol signals suppression. No modification of the pH was performed to avoid any chemical alteration of the matrix. The generation of input variables for untargeted analysis was done via bucketing the spectra. The resulting dataset was preprocessed prior to perform unsupervised PCA, by means of MetaboAnalyst web-based tool suite. The identification of compounds for the targeted analysis was performed by comparison to pure compounds spectra by means of SMA plug-in of MNova 14.2.3 software. The dataset containing the concentrations (%) of identified compounds was subjected to one-way analysis of variance (ANOVA) to highlight significant differences among the wines. The untargeted analysis, carried out through the PCA, revealed a clear differentiation among the wines. The fragments of the spectra contributing mostly to the separation were attributed to flavonoids, aroma compounds and amino acids. The targeted analysis leaded to the identification of 68 compounds, whose concentrations were significant different among the wines. The results were related to soils physical-chemical analysis and showed that: 1) high concentrations of flavan-3-ols and flavonols are correlated with high clay content in soils; 2) high concentrations of anthocyanins, amino acids, and aroma compounds are correlated with neutral and moderately alkaline soil pH; 3) low concentrations of flavonoids and aroma compounds are correlated with high soil organic matter content and acidic pH. The 1H-NMR metabolomic analysis proved to be an excellent tool to discriminate between wines originating from grapes grown on different soil types and revealed that soils in the Mediterranean area exert a strong impact on the chemical composition of the wines.