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…

Analysis of some environmental factors and cultural practices that affect the production and quality of the Manto Negro, Callet and Prensal Blanc varieties

45 non irrigated vineyards distributed in the DO (Denomination) Pla i Llevant de Mallorca and the DO Binissalem Mallorca were used to investigate the characteristics of production and quality and their relationships certain environmental factors and cultural practices. The grape varieties investigated are autochthonous to the island of Mallorca, Manto Negro and Callet as red and Prensal Blanc as white. All plants were measured for four consecutive years in the main production and quality parameters. Among the environmental factors, the type of soil has been studied, more specifically its water retention capacity, the planting density, the age of the vineyard and the level of viral infection. The presence or absence of virus seems to have no effect on any component studied in the varieties studied. For the white variety Prensal Blanc age is negatively correlated with production and the number of bunches, nevertheless it does not cause any effect on the required quality parameters. However, for the red varieties Callet and Manto Negro, the age of the plantation is the variable that best correlates with the quality parameters, therefore the old vines should be the object of preservation by the viticulturists and winemakers in order to guarantee its contribution to the quality of the wines made with these varieties.

‘Cabernet Sauvignon’ (Vitis vinifera L.) berry skin flavonol and anthocyanin composition is affected by trellis systems and applied water amounts

Trellis systems are selected in wine grape vineyards to mainly maximize vineyard yield and maintain berry quality. This study was conducted in 2020 and 2021 to evaluate six commonly utilized trellis systems including a vertical shoot positioning (VSP), two relaxed VSPs (VSP60 and VSP80), a single high wire (SH), a high quadrilateral (HQ), and a guyot (GY), combined with three levels of irrigation regimes based on different crop evapotranspiration (ETc) replacements, including a 25% ETc, 50% ETc, and 100% ETc. The results indicated SH yielded the most fruits and accumulated the most total soluble solids (TSS) at harvest in 2020, however, it showed the lowest TSS in the second season. In 2020, SH and HQ showed higher concentrations in most of the anthocyanin derivatives compared to the VSPs. Similar comparisons were noticed in 2021 as well. SH and HQ also accumulated more flavonols in both years compared to other trellis systems. Overall, this study provides information on the efficacy of trellis systems on grapevine yield and berry flavonoid accumulation in a currently warming climate.

A predictive model of spatial Eca variability in the vineyard to support the monitoring of plant status

[lwp_divi_breadcrumbs home_text="IVES" use_before_icon="on" before_icon="||divi||400" module_id="publication-ariane" _builder_version="4.19.4" _module_preset="default" module_text_align="center" module_font_size="16px" text_orientation="center"...

Grapevine sugar concentration model in the Douro Superior, Portugal

Increasingly warm and dry climate conditions are challenging the viticulture and winemaking sector. Digital technologies and crop modelling bear the promise to provide practical answers to those challenges. As viticultural activities strongly depend on harvest date, its early prediction is particularly important, since the success of winemaking practices largely depends upon this key event, which should be based on an accurate and advanced plan of the annual cycle. Herein, we demonstrate the creation of modelling tools to assess grape ripeness, through sugar concentration monitoring. The study area, the Portuguese Côa valley wine region, represents an important terroir in the “Douro Superior” subregion. Two varieties (cv. Touriga Nacional and Touriga Franca) grown in five locations across the Côa Region were considered. Sugar accumulation in grapes, with concentrations between 170 and 230 g l-1, was used from 2014 to 2020 as an indicator of technological maturity conditioned by meteorological factors. The climatic time series were retrieved from the EU Copernicus Service, while sugar data were collected by a non-profit organization, ADVID, and by Sogrape, a leading wine company. The software for calibrating and validating this model framework was the Phenology Modeling Platform (PMP), version 5.5, using Sigmoid and growing degree-day (GDD) models for predictions. The performance was assessed through two metrics: Roots Mean Square Error (RMSE) and efficiency coefficient (EFF), while validation was undertaken using leave-one-out cross-validation. Our findings demonstrate that sugar content is mainly dependent on temperature and air humidity. The models achieved a performance of 0.65

Is wine terroir a valid concept under a changing climate?

The OIV[i] defines terroir as a concept referring to an area in which collective knowledge of the interactions between the physical and biological environment (soil, topography, climate, landscape characteristics and biodiversity features) and vitivinicultural practices develops, providing distinctive wine characteristics. Those are perceptible in the taste of wine, which drives consumer preference and, therefore, wine’s value in the marketplace. Geographical indications (GI) are recognized regulatory constructs formalizing and protecting the nexus between wine taste and the terroir generating it. Despite considering updates, GIs do not consider the nexus as a dynamic one and do not anticipate change, namely of climate. Being climate a fundamental feature of terroir, it strongly impacts wine characteristics, such as taste. According to IPCC[ii], many widespread, rapid and unprecedented changes of climate occurred, some being irreversible over hundreds to thousands of years. Climatic shifts and atmospheric-driven extreme events have been widely reported worldwide. Recent climatic trends are projected to strengthen in upcoming decades, whereas extremes are expected to increase in frequency and intensity, forcing wines away from GI definitions. Geographical shifts of viticultural suitability are projected, often moving into regions and countries different from current ones. Some authors propose adaptation in viticulture, winemaking and product innovation. We show evidence of climate changing wine characteristics in the Douro valley, home of 270-year-old Port GI. We discuss herein resist or adapt stances for when climate changes the nexus between terroir and wine characteristics. Using the MED-GOLD[iii] dashboard, a tool allowing for easy visual navigation of past and future climates, we demonstrate how policymakers can identify future moments, throughout the 21st century under different emission scenarios, when GI specifications will likely need updates (e.g., boundaries, varieties) to reduce climate-change impacts.