Project

Digital Vineyards

Goal: To test new and emerging technologies for vineyard operations and management, ranging from sensor network technology and remote sensing from UAV and satellite. This project also considers the use of machine learning and artificial intelligence to analyse data obtained with these systems and big data available from vineyards to estimate phenological stages occurrence and prediction of yields.

Updates
0 new
0
Recommendations
0 new
0
Followers
0 new
34
Reads
0 new
380

Project log

Sigfredo Fuentes
added a research item
Artificial intelligence (AI) is being developed and used in four main areas of viticulture and winemaking in order to tackle a wide range of challenges, from climate change to counterfeit wine.
Claudia Gonzalez Viejo
added 4 research items
Berry cell death (BCD) has been shown to be an objective and important parameter to assess berry and wine quality traits such as aroma profiles. Most common methods to measure berry cell death involve the collection of berries from veraison to harvest, followed by analysis using dyes (fluorescent and non-fluorescent) and recording images for computer vision analysis, which is a cumbersome process. Therefore, these studies aimed to develop rapid and non-destructive methods to assess BCD and wine aroma profiles using near-infrared spectroscopy (NIR), canopy architecture parameters (CAP), and machine learning (ML) modelling. Two studies were conducted using berries and wines from (i) Pinot Noir (Study 1) and (ii) Shiraz and Chardonnay (Study 2) from Australian vineyards. For Study 1, two ML models were developed using (i) NIR absorbance values as inputs to predict living and dead tissue (Model 1), resulting in high accuracy (R = 0.94), and (ii) the outputs from Model 1 (living and dead tissue) along with total soluble solids (Brix) were used as inputs to predict the intensity of 19 different aromas in the wine (Model 2) with accuracy R = 0.80. Study 2 consisted of four different highly accurate ML models developed using (i) NIR absorbance values as inputs to predict living and dead tissue (Model 3; R = 0.87), (ii) CAP assessed using computer vision from cover photography of grapevine canopies as inputs to predict living and dead tissue (Model 4; R = 0.98) and (iii, iv) to predict the intensity of 10 different aromas in the wine for Chardonnay (Model 5; R = 0.99) and Shiraz (Model 6; R = 0.99). These methods may be fully automated and used as non-destructive, accurate, affordable, and real-time techniques to assess BCD and predict wine quality traits for decision making. Furthermore, Models 3 – 6 can be integrated into visible cameras as the payload of unmanned aerial vehicles (UAV).
Due to increased rates of fraud through counterfeiting and adulteration of quality wines, it is important to develop novel non-destructive techniques to assess wine quality and provenance. Therefore, our research group developed a novel method using near-infrared (NIR) spectroscopy (1596 – 2396 nm) coupled with machine learning (ML) modelling to assess wine vintages and quality traits such as Brix, pH, alcohol content, and intensity of sensory descriptors through the bottle. These were developed using samples from two Australian vineyards for Shiraz, Chardonnay and Pinot Noir wines. Models resulted in high accuracy > 90% for classification (vintages) and regression (quality traits). The proposed method will allow to assess authenticity and quality of any wines in the market without the need to open the bottles, which is effective and convenient. Furthermore, currently there are low-cost NIR devices which can be affordable for winemakers and retailers.
Due to climate change, the higher incidence and severity of bushfires is a major challenge for wine producers worldwide as an increase in smoke contamination negatively affects the physicochemical components that contribute to lower quality of both fresh produce and final products (smoke taint in wines). This results in reduced prices and consumer acceptability, impacting the producers and manufacturers. Current methods available to winemakers for the assessment of contamination in berries and wine consist of costly laboratory analyses which require skilled personnel, are time-consuming, cost prohibitive and destructive. Therefore, novel rapid, cost-effective, and reliable methods using digital technologies such as the use of near-infrared (NIR) spectroscopy, electronic nose (e-nose) and machine learning (ML) have been developed by our research group. Several ML models have been developed for smoke taint detection and quantification in berries and wine from different varieties using NIR absorbance values or e-nose outputs as inputs to predict glycoconjugates, volatile phenols, aromatic volatile compounds, smoke-taint amelioration techniques efficacy and sensory descriptors, all models with > 97% accuracy. These methods and models may be integrated and automated as digital twins to assess smoke contamination in berries and smoke taint in wine from the vineyard for early decision-making.
Claudia Gonzalez Viejo
added a research item
Climatic anomalies, such as heatwaves and bushfires, are increasing in number, intensity, and severity worldwide due to climate change. Bushfires are especially important in winemaking countries since smoke contamination can reach vineyards in critical periods of berry development, producing smoke contamination, which is passed to the wine as smoke taint in the winemaking process. The only alternative for winemakers to assess berry or wine contamination is sending samples to specialized laboratories, which can be time-consuming, cost-prohibitive and only sentinel plants or batches can be monitored. New and emerging technologies based on non-destructive remote sensing, such as near-infrared spectroscopy and the development of low-cost e-noses coupled with artificial intelligence (AI) tools, have been developed by the Digital Agriculture, Food and Wine Group from The University of Melbourne. The machine learning (ML) classification models developed showed high accuracy (97–98%) for berries, leaves and wine assessment to predict the level of smoke contamination. Furthermore, ML regression models to predict smoke-derived compounds in berries, must, and wine also presented very high accuracy (R = 0.98–0.99). On the other hand, ML models to predict consumers acceptability of smoke-tainted wines were also successfully developed (R = 0.97–0.98). These models can result in cost-effective and accurate technologies applicable to the vineyard and wineries to assess levels of smoke taint and associated compounds for decision-making purposes.
Claudia Gonzalez Viejo
added a research item
The winemaking industry can benefit greatly by implementing digital technologies to avoid guesswork and the development of off-flavors and aromas in the final wines. This research presents results on the implementation of near-infrared spectroscopy (NIR) and a low-cost electronic nose (e-nose) coupled with machine learning to detect and assess wine faults. For this purpose, red and white base wines were used, and treatments consisted of spiked samples with 12 faults that are traditionally formed in wines. Results showed high accuracy in the classification models using NIR and e-nose for red wines (94%–96%; 92%–97%, respectively) and white wines (96%– 97%; 90%–97%, respectively). Implementing new and emerging digital technologies could be a turning point for the winemaking industry to become more predictive in terms of decision-making and maintaining and increasing wine quality traits in a changing and challenging climate.
Sigfredo Fuentes
added a research item
The increase of vineyard's water consumption due to the Global Warming Phenomenon (GWP) has forced the winegrowers to strengthen their irrigation and water stewardship efforts, intended for maintaining this resource's long-term sustainable use. Due to water being a limited resource, implementing the Water Footprint (WF) concept in winegrapes production provides helpful information for sustainable water stewardship. Currently, an automated version of the satellite-based METRIC (Mapping Evapotranspiration with Internalized Calibration) model, the Google Earth Engine Evapotranspiration Flux (EEFlux) platform, has been suggested as an alternative to analyzing the spatial variability of an entire field's water consumption throughout the growing season. This work aimed to evaluate the potential application of the EEFlux satellite's actual evapotranspiration (ET a) products and ancillary field data to obtain the WF blue (WF b) and green (WF g) of six commercial vineyards placed in the Chilean central zone. Firstly, the reliability of the daily actual evapotranspiration data from EEFlux (ET a EEFlux) was assessed against measured ET a data, using an available database from previous studies. The results of ET a EEFlux estimations against measured ET a were impressive, presenting a root square error (RMSE) of 0.8 mm day −1. The satellite-derived crop coefficients (K c Sat) allowed to estimate the total WF of each vineyard, in a range of 200 to 900 m 3 t −1 , showing an average relative error (RE) of 101%, between the satellite-based WF b (WF b Sat) and those calculated from irrigation records (WF b). These results reflected the particular conditions of each vineyard and can be considered reasonable since they were estimated from ancillary data and EEFlux products. This study provides new insights that may represent opportunities to sustainably managing the irrigation of vineyards.
Claudia Gonzalez Viejo
added a research item
Wildfires are an increasing occurrence throughout the world. Grapevine smoke exposure can result in wines characterized by unpalatable smoke aromas. Current chromatographic techniques to assess levels of smoke-derived compounds in berries and wine are destructive, expensive to operate and time-consuming. This research assessed the use of near-infrared technology along with artificial neural network modelling as rapid, non-destructive, and cost-effective tools to assess smoke-derived compounds in berries, must and wine.
Sigfredo Fuentes
added a research item
New and emerging technologies could play a critical role in the viticulture and winemaking of the future. Climate change has threatened the status quo within the viticultural and wine industry due to increased ambient temperatures, the variability of precipitation, and the increase of climatic risks. These main threats are specifically related to the compression of phenological stages, earlier harvests, many of these within the hottest months producing a dual warming effect. Furthermore, the increase of climatic anomalies, such as floods, frosts, and bushfires, in number, intensity, and window of opportunity within the growing season directly impacts yield and grape and wine quality. The viticulture and winemaking of the future need to have a transformational process to be more predictive rather than only reactive by implementing disruptive technology supported by artificial intelligence (AI). Link to the book: https://www.elsevier.com/books/improving-sustainable-viticulture-and-winemaking-practices/costa/978-0-323-85150-3
Sigfredo Fuentes
added a research item
While starting a new 2021-22 grape-growing season in Australia, potential climatic anomalies such as bushfires are in the consciousness of many grape growers and winemakers. Increasing ambient temperatures has resulted in associated climatic anomalies, such as extreme wildfires in Australia, California, Siberia, Greece and Turkey as forecasted and reviewed by the latest Intergovernmental Panel on Climate Change report (IPPC-AR6). The latest IPCC report also claims a “virtually certain” increase in the frequency and intensity of heatwaves due to greenhouse emissions from burning fossil fuels. Many of these wildfires or bushfires occur in Mediterranean countries that cultivate grapes for winemaking. Recent review papers on smoke taint in wines related to mitigation techniques have assessed the state of the art research efforts related to smoke taint and developed methods for mitigation and incorporating some of the latest digital technologies for its assessment. Both reviews agreed that the best practices to remediate smoke taint in wine are activated carbon fining and reverse osmosis treatments. This article focuses on the latest advances to monitor the levels of smoke contamination in grapevines and grapes and smoke taint in wines using novel digital and non-invasive technologies. Much of these latest efforts are the product of research from the Digital Agriculture, Food and Wine group from The University of Melbourne. The latest research has produced tools, and artificial intelligence (AI) models with high accuracy (Table 1) to be deployed in near-real-time using affordable technologies either in the field between veraison to harvest and in the winery that can be accessible to winegrowers and winemakers.
Claudia Gonzalez Viejo
added a research item
Berry cell death assessment can become one of the most objective parameters to assess important berry quality traits, such as aroma profiles that can be passed to the wine in the winemaking process. At the moment, the only practical tool to assess berry cell death in the field is using portable near-infrared spectroscopy (NIR) and machine learning (ML) models. This research tested the NIR and ML approach and developed supervised regression ML models using Shiraz and Chardonnay berries and wines from a vineyard located in Yarra Valley, Victoria. Australia. A ML model was developed using NIR measurements from intact berries as inputs to estimate berry cell death (BCD), living tissue (LT) (Model 1). Furthermore, canopy architecture parameters obtained from cover photography of grapevine canopies and computer vision analysis were also tested as inputs to develop ML models to assess BCD and LT (Model 2) and the intensity of sensory descriptors based on visual and aroma profiles of wines for Chardonnay (Model 3) and Shiraz (Model 4). Results showed high accuracy and performance of models developed based on correlation coefficient (R) and slope (b) (M1: R=0.87; b = 0.82; M2: R=0.98; b=0.93; M3: R=0.99; b=0.99; M4: R=0.99; b=1.00). Models developed based on canopy architecture, and computer vision can be used to automatically estimate vigor and berry and wine quality traits using proximal remote sensing and with visible cameras as payload of unmanned aerial vehicles (UAV).
Claudia Gonzalez Viejo
added a research item
Climatic anomalies, such as heatwaves and bushfires, are increasing in number, intensity, and severity worldwide due to climate change. Bushfires are especially important in winemaking countries since smoke contamination can reach vineyards in critical periods of berry development, producing smoke contamination, which is passed to the wine as smoke taint in the winemaking process. The only alternative for winemakers to assess berry or wine contamination is by sending samples to specialised laboratories, which can be time consuming and cost-prohibitive and only sentinel plants or batches can be monitored. New and emerging technologies based on non-destructive remote sensing, such as near-infrared spectroscopy and the development of low-cost e-noses coupled with artificial intelligence (AI) tools, have been developed by our research group, which can result in cost-effective and accurate technologies applicable to the vineyard and wineries to assess levels of smoke taint and associated compounds for decision-making purposes.
Claudia Gonzalez Viejo
added a research item
Wine aroma is an important quality trait in wine influenced by its volatile compounds. Many factors can affect the composition and levels (concentration) of volatile aromatic compounds, in-cluding the water status of grapevines, canopy management and the effects of climate change, such as increases in ambient temperature and drought. In this study, a low-cost and portable electronic nose (e-nose) was used to assess wines produced from grapevines exposed to different levels of smoke contamination. Readings from the e-nose were then used as inputs to develop two machine learning models based on artificial neural networks. Results showed that regression Model 1 displayed high accuracy in predicting the levels of volatile aromatic compounds in wine (R = 0.99). On the other hand, Model 2 also had high accuracy in predicting smoke aroma intensity from sensory evaluation (R = 0.97). Descriptive sensory analysis showed high levels of smoke taint aromas in the high-density smoke-exposed wine sample (HS), followed by the high-density smoke exposure with in-canopy misting treatment (HSM). Principal component analysis further showed that the HS treatment was associated with smoke aroma intensity, while results from the matrix showed significant negative correlations (p<0.05) were observed between ammonia gas (sensor MQ137) and the volatile aromatic compounds octanoic acid, ethyl ester (r = -0.93), deca-noic acid, ethyl ester (r = -0.94) and octanoic acid, 3-methylbutyl ester (r = -0.89). The two models developed in this study may offer winemakers a rapid, cost-effective and non-destructive tool for assessing levels of volatile aromatic compounds and the aroma qualities of wine for decision making.
Sigfredo Fuentes
added a research item
The incidence and intensity of bushfires is increasing due to climate change, resulting in a greater risk of smoke taint development in wine. In this study, smoke tainted and non-smoke tainted wines were subjected to treatments using activated carbon with/without the addition of a cleaving enzyme treatment to hydrolyze glycoconjugates. Chemical measurements and volatile aroma compounds were assessed for each treatment, with the two smoke taint amelioration treatments exhibiting lower mean values for volatile aroma compounds exhibiting positive ‘fruit’ aromas. Furthermore, a low-cost electronic nose (e-nose) was used to assess the wines. A machine learning model based on artificial neural networks (ANN) was developed using the e-nose outputs from the unsmoked control wine, unsmoked wine with activated carbon treatment, unsmoked wine with a cleaving enzyme plus activated carbon treatment, and smoke tainted control wine samples as inputs to classify the wines according to the smoke taint amelioration treatment. The model displayed a high overall accuracy of 98% in classifying the e-nose readings, illustrating it may be a rapid, cost-effective tool for winemakers to assess the effectiveness of smoke-taint amelioration treatment by activated carbon with/without the use of a cleaving enzyme. Furthermore, the use of a cleaving enzyme coupled with activated carbon was found to be effective in ameliorating smoke taint in wine and may help delay the resurgence of smoke aromas in wine following aging and hydrolysis of glycoconjugates.
Sigfredo Fuentes
added a research item
Unmanned aerial vehicle (UAV) remote sensing has become a readily usable tool for agricultural water management with high temporal and spatial resolutions. UAV-borne thermography can monitor crop water status near real-time, which enables precise irrigation scheduling based on accurate decision-making strategy. The crop water stress index (CWSI) is a widely adopted indi-cator of plant water stress for irrigation management practices, however dependence of its efficacy on data acquisition time during the daytime is yet to be investigated rigorously. In this paper, plant water stress captured by a series of UAV remote sensing campaigns at different times of the day (9h, 12h and 15h) in a nectarine orchard were analyzed to examine the diurnal behavior of plant water stress represented by the CWSI against measured plant physiological parameters. CWSI values were derived using a probability modelling, named ‘Adaptive CWSI’, proposed by our earlier research. The plant physiological parameters, such as stem water potential (ψstem) and sto-matal conductance (gs), were measured on plants for validation concurrently with the flights un-der different irrigation regimes (0, 20, 40 and 100 % of ETc). Estimated diurnal CWSIs were com-pared with plant-based parameters at different data acquisition times of the day. Results showed a strong relationship between ψstem measurements and the CWSIs at midday (12 h) with a high co-efficient of determination (R2 = 0.83). Diurnal CWSIs showed a significant R2 to gs over different levels of irrigation at three different times of the day with R2 = 0.92 (9h), 0.77 (12h) and 0.86 (15h), respectively. The adaptive CWSI method used showed a robust capability to estimate plant water stress levels even with the small range of changes presented in the morning. Results of this work indicate that CWSI values collected by UAV-borne thermography between mid-morning and mid-afternoon can be used to map plant water stress with a consistent efficacy. This has important implications for extending the time-window of UAV-borne thermography (and subsequent areal coverage) for accurate plant water stress mapping beyond midday.
Sigfredo Fuentes
added a research item
Climate change has posed major risks for viticulture and winemaking around the world, related to increased ambient temperatures, the variability of rain events, higher occurrence and intensity of climatic anomalies, such as frosts and bushfires. These changes have directly impacted grapevine phenology by compressing stages and pushing forward in time harvest to hottest months, producing a dual warming effect. Bushfires also directly impact berry smoke contamination, which can be passed to the wine in the winemaking process producing smoke taint. Due to these events' complexities and their effects on viticulture and winemaking, a smarter approach is required to obtain relevant information and process it efficiently for more appropriate decision-making by different practitioners. In the last 10 years, artificial intelligence has offered various applications to be included in viticultural and winemaking operations, which has rendered important advances and information to deal with climate change adversities.
Claudia Gonzalez Viejo
added a research item
The number and intensity of wildfires are increasing worldwide, thereby also raising the risk of smoke contamination of grapevine berries and the development of smoke taint in wine. This study aimed to develop five artificial neural network (ANN) models from berry, must, and wine samples obtained from grapevines with different levels of smoke exposure (i) Control (C), i.e., neither misting nor smoke exposure; (ii) Control with misting (CM), i.e., in-canopy misting but no smoke exposure; (iii) low-density smoke treatment (LS); (iv) high-density smoke treatment (HS) and (v) a high-density smoke treatment with misting (HSM). Models 1, 2 and 3 were developed using the absorbance values of near-infrared (NIR) berry spectra taken one day after smoke exposure to predict levels of 10 volatile phenols (VP) and 18 glycoconjugates in grapes at either one day after smoke exposure (Model 1: R= 0.98; R2= 0.97; b= 1) or at harvest (Model 2: R= 0.98; R2= 0.97; b= 0.97), as well as six VP and 17 glycoconjugates in the final wine (Model 3: R= 0.98; R2= 0.95; b= 0.99). Models 4 and 5 were developed to predict the levels of six VP and 17 glycoconjugates in wine. Model 4 used must NIR absorbance spectra as inputs (R= 0.99; R2= 0.99; b= 1.00), while Model 5 used wine NIR absorbance spectra (R= 0.99; R2= 0.97; b= 0.97). All five models displayed high accuracies and could be used by grape growers and winemakers to non-destructively assess at near real-time the levels of smoke-related compounds in grapes and/or wine in order to make timely decisions around grape harvest and smoke taint mitigation techniques in the winemaking process.
Sigfredo Fuentes
added a research item
Aim: To analyse unmanned aerial vehicle (UAV)-based imagery to assess canopy structural changes after the application of different canopy management practices in the vineyard. Methods and results: Four different canopy management practices: i–ii) leaf removal within the bunch zone (eastern side/both eastern and western sides), iii) bunch thinning and iv) shoot trimming were applied to grapevines at veraison, in a commercial Cabernet-Sauvignon vineyard in McLaren Vale, South Australia. UAV-based imagery captures were taken: i) before the canopy treatments, ii) after the treatments and iii) at harvest to assess the treatment outcomes. Canopy volume, projected canopy area and normalized difference vegetation index (NDVI) were derived from the analysis of RGB and multispectral imagery collected using the UAV. Plant area index (PAI) was calculated using the smartphone app VitiCanopy as a ground-based measurement for comparison with UAV-derived measurements. Results showed that all three types of UAV-based measurements detected changes in the canopy structure after the application of canopy management practices, except for the bunch thinning treatment. As expected, ground-based PAI was the only technique to effectively detect internal canopy structure changes caused by bunch thinning. Canopy volume and PAI were found to better detect variations in canopy structure compared to NDVI and projected canopy area. The latter were negatively affected by the interference of the trimmed shoots left on the ground. Conclusions: UAV-based tools can provide accurate assessments to some canopy management outcomes at the vineyard scale. Among different UAV-based measurements, canopy volume was more sensitive to changes in canopy structure, compared to NDVI and projected canopy area, and demonstrated a greater potential to assess the outcomes of a range of canopy management practices. Significance and impact of the study: Canopy management practices are widely applied to regulate canopy growth, improve grape quality and reduce disease pressure in the bunch zone. Being able to detect major changes in canopy structure, with some limitations when the practice affects the internal structure (i.e., bunch thinning), UAV-based imagery analysis can be used to measure the outcome of common canopy management practices and it can improve the efficiency of vineyard management.
Claudia Gonzalez Viejo
added 2 research items
Wildfires are an increasing problem worldwide, with their number and intensity predicted to rise due to climate change. When fires occur close to vineyards, this can result in grapevine smoke contamination, and subsequently, the development of smoke taint in wine. Currently, there are no in-field detection systems that growers can use to assess whether their grapevines have been contaminated by smoke. This study evaluated the use of near-infrared (NIR) spectroscopy as a chemical fingerprinting tool, coupled with machine learning, to create a rapid, non-destructive in-field detection system for assessing grapevine smoke contamination. Two artificial neural network models were developed using grapevine leaf spectra (Model 1) and grape spectra (Model 2) as inputs, and smoke treatments as targets. Both models displayed high overall accuracies in classifying the spectral readings according to the smoking treatments (Model 1: 98.00%; Model 2: 97.40%). Ultraviolet to visible spectroscopy was also used to assess the physiological performance and senescence of leaves, and the degree of ripening and anthocyanin content of grapes. The results showed that chemical fingerprinting and machine learning might offer a rapid, in-field detection system for grapevine smoke contamination that will enable growers to make timely decisions following a bushfire event, e.g., avoiding harvest of heavily contaminated grapes for winemaking or assisting with a sample collection of grapes for chemical analysis of smoke taint markers.
Bushfires are increasing in number and intensity due to climate change. A newly developed low-cost electronic nose (e-nose) was tested on wines made from grapevines exposed to smoke in field trials. E-nose readings were obtained from wines from five experimental treatments: i) low-density smoke exposure (LS), ii) high-density smoke exposure (HS) and iii) high-density smoke exposure with in-canopy misting (HSM), and two controls: iv) control (C; no smoke treatment) and v) control with in-canopy misting (CM; no smoke treatment). These e-nose readings were used as inputs for machine learning algorithms to obtain a classification model, with treatments as targets and 7 neurons, with 97% accuracy in the classification of 300 samples into treatments as targets (Model 1). Models 2 to 4 used 10 neurons, with 20 glycoconjugates and 10 volatile phenols as targets, measured: in berries one hour after smoke (Model 2; R=0.98; R2=0.95; b=0.97); in berries at harvest (Model 3; R= 0.99; R2 = 0.97; b = 0.96); and in wines (Model 4; R=0.99; R2=0.98; b=0.98). Model 5 was based on the intensity of 12 wine descriptors determined via a consumer sensory test (Model 5; R=0.98; R2=0.96; b=0.97). These models could be used by winemakers to assess near real-time smoke contamination levels and to implement amelioration strategies to minimize smoke taint in wines following bushfires.
Sigfredo Fuentes
added a research item
I DIRADARE ne vale la pena? Il diradamento dei grappoli e dei germogli sono due diverse pratiche comunemente utilizzate nella ge-stione dei vigneti. L'obiettivo è quel-lo di migliorare la qualità dell'uva in linea con la diffusa convinzione che i vini migliori siano prodotti con uve provenienti da vigneti a bassa vigo-ria e produttività. DIRADAMENTO DEI GRAPPOLI Il diradamento dei grappoli consi-ste, per l'appunto, nella rimozione di una porzione dei grappoli presenti sulla vite e il suo effetto dipende da fattori quali: tempistica, percentuale di grappoli rimossi, cultivar, vigore e clima. Questa tecnica, basata sul concetto che la rimozione di parte dei grappoli migliorerà il processo di maturazione di quelli rimanenti, è aneddoticamente associata a pro-duzioni di qualità nella viticoltura a clima temperato. Nonostante alcuni studi abbiamo dimostrato che il diradamento dei grappoli migliori la qualità dell'uva, altri, invece, non hanno evidenziato differenze significative. Nei casi in cui si è avuto un incremento della qualità dell'uva le variazioni, da un punto di vista della composizione chimica dell'acino, erano relative: all'accumulo di zuccheri, alla concentrazione in acido malico, al pH, al potassio, all'acidità totale e alla concentrazione in antociani e composti fenolici. Alcune ricerche, inoltre, hanno mostrato influenze del diradamento dei grappoli sulle caratteristiche sensoriali dei vini. DIRADAMENTO DEI GERMOGLI Simile al diradamento dei grappoli, il diradamento dei germogli vie-ne usato per ridurre la densità della chioma, la produzione e migliorare la composizione chimica delle uve.
Sigfredo Fuentes
added a research item
Digital Agriculture deals with the implementation and integration of digital data, sensors and tools on agricultural applications from the paddock to consumers. These technologies can range from: big data, sensor technology, sensor networks, remote sensing, robotics, and unmanned aerial vehicles (UAV). Data processing is performed using new and emerging technologies, such as computer vision, machine learning and artificial intelligence, among others. The latest advances made by the DAFW group for crop monitoring / decision making, assessment of quality of produces, sensory analysis for consumer perception assessment and animal stress and welfare assessment are presented in this article.
Sigfredo Fuentes
added a research item
Research and innovation activities in the areas sensor technology can accelerate the adoption of new and emerging digital tools in the agricultural sector by the implementation of precision farming practices such as remote sensing, operations, and real-time monitoring. The agricultural industry has been greatly affected by climate change; therefore, to be successful in overcoming these effects and remain competitive and sustainable in the market, there is the need to support research and application development of new and emerging sensor technologies and their applications in agriculture. A total of 13 papers were published in this Special Issue entitled: Emerging Sensor Technology in Agriculture and the topics addressed include different emerging technologies with applications on ecosystems (grasslands) and several agriculture crops such as peppers, apples, grapevines, cocoa trees, citrus, legumes, wheat and rice. Two papers were also related to the use of remote sensing to detect forage quality, regions of interest of pigs, and pesticide droplet deposition using machine learning.
Claudia Gonzalez Viejo
added a research item
Important wine quality traits such as sensory profile and color are the product of complex interactions between the soil, grapevine, the environment, management, and winemaking practices. Artificial Intelligence (AI) and specifically Machine Learning (ML) could offer powerful tools to assess these complex interactions and their patterns through seasons to predict quality traits to winegrowers close to harvest and before winemaking. This study considered nine vintages (2008-16) using near-infrared spectroscopy (NIR) of wines and corresponding weather and management information as inputs for artificial neural network (ANN) modeling of sensory profiles (Model 1 and 2 respectively). Furthermore, weather and management data were used as inputs to predict the color of wines (Model 3). Results showed high accuracy in the prediction of sensory profiles of vertical wine vintages using NIR (Model 1; R=0.92; slope=0.85), while better models were obtained using weather/management data for the prediction of sensory profiles (Model 2; R=0.98; slope=0.93) and wine color (Model 3; R=0.99; slope=0.98). For all models, there was no indication of overfitting as per ANN specific tests. These models may be used as powerful tools to winegrowers and winemakers close to harvest and before the winemaking process to maintain a determined wine style with high quality and acceptability by consumers.
Sigfredo Fuentes
added 2 research items
Shoot and bunch thinning are common practices used in vineyard management with the overall aim of improving fruit quality based on the widespread findings that have related the production of good wines with lower canopy vigour and low yielding vineyards.
Sigfredo Fuentes
added a research item
The midday stem water potential (Ψs) and stomatal conductance (gs) have been traditionally used to monitor water status of cherry trees (Prunus avium L.). Due to complexity of direct measurement, the use of infrared thermography has been proposed as an alternative. This study compares Ψs and gs against crop water stress indexes (CWSI) calculated from thermal infrared (TIR) data from high (HR) and low (LR) resolution cameras for two cherry tree cultivars: 'Regina' and 'Sweetheart'. For this purpose, a water stress-recovery cycle experiment was carried out at the post-harvest period in a commercial drip-irrigated cherry tree orchard under three irrigation treatments based on Ψs levels. Water status of trees was weekly measured using Ψs, gs, and compared to CWSI's, computed from both thermal cameras. Results showed that the accuracy in the estimation of CWSIs was not statistically significant when comparing both cameras for the representation of Ψs and gs in both cultivars. The performance of all physiological indicators evaluated presented similar trends for both cultivars and the averaged differences between CWSI's from both cawere 11±0.27 %. However, these CWSI's were not able to detect differences among irrigation treatments as compared to Ψs and gs due to narrow differences between treatments.
Sigfredo Fuentes
added a research item
One of the most significant climatic anomalies, related to climate change that is impacting the wine-growing industry is bushfire events. Grapevine smoke contamination and smoke taint in wines are difficult to assess in the vineyards and wineries. Current assessment methods require berry or wine sample collections and specialised laboratory analysis, which can be time-consuming, cost-prohibitive, and non-representative of the real level of contamination within vineyards. Recently, the Digital Agriculture, Food and Wine group (DAFW) have implemented Artificial Intelligence (AI) based on short and proximal remote sensing and Machine Learning (ML) modelling to assess and monitor smoke contamination and smoke taint in wines. The technology developed has rendered rapid, accurate, and affordable systems to monitor smoke contamination in grapevines, berries, and potential contamination in wines for seven different cultivars. This technology applied to grapevines may be implemented using Unmanned Aerial Vehicles (UAV) and infrared thermal imagery (IRTI) to map regions of vineyards according to smoke contamination levels. Applications to berries and wines using near-infrared spectroscopy (NIR) could offer a quick assessment of the implementation of amelioration techniques to reduce smoke-related compounds in berries and taint in wines. Finally, an electronic nose (e-nose) has been recently developed to assess smoke-related gases in wines to predict smoke taint, and it can be applied to the vineyard to monitor ambient gases and levels of smoke contamination in bushfire events. Further research is required to make these AI applications available to more viticultural regions, grapevine cultivars, and bushfire scenarios.
Sigfredo Fuentes
added a research item
Aim – Shoot thinning is a common canopy management practice used to obtain a desired canopy shoot density and improve the canopy microclimate. The cost of such technique, since it is often carried out manually, can be high. In this study the effect of severe shoot thinning (50% of shoots removed) applied at EL 15 was investigated by comparing yield components, canopy size, berry and wine chemistry and sensory attributes to a non-thinned control for the variety Semillon. Methods and Results – The trial was carried out over four seasons (starting in 2014-15) in the Semillon block of the Coombe vineyard (Waite Campus, the University of Adelaide). Canopy architecture was monitored at three phenological stages in each season and yield components were assessed at harvest. The harvested fruit was used for chemical and sensory analysis of the berries. Wines were made and their chemistry and sensory assessed. Results showed that shoot thinning affected the total leaf area only in two of the four seasons however the single shoot leaf area was higher in the shoot thinned vine. Similar results were obtained with pruning weight and cane weight. Shoot thinning did not reduce yield, despite a large reduction in bunch number, due to increases in bunch weight. Shoot thinning did not improve berry and wine chemistry. Similarly, little differences were observed in the sensory profile of berries and wines and only in the last season the assessors preferred the wines obtained from shoot thinned vines. Conclusions - Shoot thinning, in this study, failed to alter canopy architecture, yield and grape and wine chemistry. The vine balance indices leaf area/yield and yield/pruning weight were also unaffected by the treatment despite its intensity (50% of shoots removed). Significance and impact of the study – The practice of shoot thinning when applied at EL stage 15 (8-9 leaves separated) was not effective as a technique to improve canopy microclimate and berry and wine chemistry for Semillon in a hot Australian climate. Significant savings can be made if the use of this technique is better understood when applied in Australia. Further research is needed to explore the effect and timing of shoot thinning on other varieties and in different environments.
Sigfredo Fuentes
added a research item
Heat-pulse techniques are routinely used to estimate transpiration from canopies of woody plants typically without any local calibration, mainly because of the difficulty of doing so in the field and, frequently, lack of detailed weather data. This is despite concerns that the techniques may produce erroneous values under certain conditions, such as when evaporative demand is high. In this study, we used a micrometeorological approach to validate transpiration from irrigated olives deduced from heat-pulse technique by ascertaining precise values for the parameters that are critical for converting heat-pulse velocity to sapflow. The micrometeorological approach involved limited data on stomatal conductance (gs), obtained hourly with a porometer on four contrasting days, and was used to calibrate a simple model for predicting conductance. Predicted stomatal conductance (gsm) agreed well with that measured, and when both were used to calculate hourly transpiration, they produced values that were within 10% of each other. This was despite brief underestimations of transpiration based on gsm (Tm) in the early hours of the day that arose from poor determination of incident radiation at this time. We then used Tm to iteratively set the values for the various parameters, including the time-out value that accounts for zero-flow conditions, needed to convert heat-pulse velocity to sapflow, for the four days. The best fit between Tm and transpiration from sapflow (Ts) was obtained with time-out value set to 120s. All heat-pulse velocity data were therefore analysed with this time-out value to obtain sapflow and, hence, transpiration (Ts). Comparison of Tm and Ts for the whole season showed that the former tended to produce higher values on certain days when vapour pressure deficit (D) was high in summer (December–February). While Ts occasionally produced larger values than Tm under the mild conditions of autumn (March–April). Totals of the daily transpiration during the 190-day period were within 10% of each other.
Sigfredo Fuentes
added 6 research items
Leaf area index (LAI) is a key physical variable which controls the exchange of water and CO2 between the earth and the atmosphere. Recent improvements in the quality of satellite-derived estimates of LAI, specifically the MODIS LAI product, have led to increased confidence in their operational use. In this study, we examined the relationship amongst MODIS LAI (Collection 5), pre-dawn leaf water potential (Ψpd) (a surrogate for plant water availability), vegetation water use (ET) and pan evaporation (E0) in forest, evergreen woodland, open shrubland and savanna in Australia. We present three models which demonstrate how the MODIS LAI product can be used to make spatially explicit predictions of the state of three key physical variables, namely Ψpd, the aridity index and vegetation water use. The LAI-Ψpd model explains how plant available soil moisture can be predicted at a continental scale for intact forests, woodlands and savannas. In a similar manner, the aridity index (mean annual rainfall/E0) shows a powerful relationship with the MODIS LAI values. Using vegetation water use data from 16 field campaigns and published studies, we developed a MODIS LAI-ET model that provides the ability to predict both site and catchment-scale annual evapotranspiration. We test the model against independent estimates of site and catchment-scale vegetation water use. Copyright © 2009 John Wiley & Sons, Ltd.
Precision irrigation in grapevines could be achieved using physiologically based irrigation scheduling methods. This paper describes an investigation on the effects of three midday stem water potential (midday ΨS) thresholds, imposed from post-setting, over water use, vegetative growth, grape quality and yield of grapevines cv. Cabernet Sauvignon. An experiment was carried out on a vineyard located at the Isla de Maipo, Metropolitana Region, Chile, throughout the 2002/03, 2003/04 and 2004/05 growing seasons. Irrigation treatments consisted in reaching the following midday ΨS thresholds: −0.8 to −0.95 MPa (T1); −1.0 to −1.2 MPa (T2) and −1.25 to −1.4 MPa (T3) from post-setting to harvest. Results showed significant differences in grape quality components among treatments and seasons studied. In average, T3 produced smallest berry diameter (6% reduction compared to T1), high skin to pulp ratio (13% increment compared to T1) and significant increments in soluble solids and anthocyanins. Improvements in grape quality attributes were attributed to mild grapevine water stress due to significant reductions in water application (46% for T2 and 89% for T3 less in average, both compared to T1). This study found significant correlations between midday ΨS and berry quality components, no detrimental effects on yield by treatments were found in this study. This research proposes a suitable physiological index and thresholds to manage RDI and irrigation scheduling on grapevines to achieve high quality grapes on mild water stress conditions.
Sigfredo Fuentes
added a research item
Significant water losses at night-time were registered in an Almond tree trial near Berri, South Australia in the season 2009-10. To quantify nocturnal tree water uptake (Sn), two sets of sap flow sensors (heat-pulse compensated) were installed per tree in the North-East (NE) and South-West (SW) sides of the trunk for three trees per treatment. The treatments were 100% ETc and 60% ETc with daily irrigations at the peak atmospheric demand period (Dec – Jan). Nocturnal water uptake by trees was in the order of 20% and 15%, compared to diurnal, for the 100% ETc and 60% ETc treatments respectively. Night-time water uptake was correlated with plant water stress from the previous day measured as midday stem water potential (s). Nocturnal water uptake was also correlated with nocturnal VPD indicating that nocturnal transpiration (En) was significant for both treatments. Differences in Sn between NE and SW sensors were registered in 60% ETc treatment only, being lower Sn for the NE side. This is consistent with the sun path in the area, with the NE side receiving maximum direct solar radiation on the canopy, resulting in lower leaf water potentials (l) for that side and less leaf stomata conductance. More research needs to be conducted to reduce Sn in the form of transpiration, since it is not coupled with photosynthesis, and hence is associated with reduced water use efficiency (WUE). According to our results, reductions of water application in the order of 40% (60% ETc treatment) contributed to increased WUE and reduced Sn and therefore En.
Sigfredo Fuentes
added a research item
Vineyards are inherently variable however, the majority of them are managed on the assumption of uniformity. Understanding vineyard variability to target management strategies, apply inputs efficiently and deliver consistent grape quality to the winery is now essential. Moreover, canopy management, together with irrigation and fertilisation, is one of the main strategies applied in the vineyard to modify vine growth and achieve the desired fruit quality outcome. Significant savings can be made if the use of these practices is well-understood and targeted. The wine industry uses significant resources to enable practitioners to measure vineyard vigour and its variability using techniques ranging from simple visual observations to sophisticated precision/digital viticulture (PV/DV). A survey on PV1 showed that, in Australia, the main constraint to the adoption of these technologies was its cost, together with the lack of technical advice. Also, the survey respondents thought that PV tools and software were too complicated (less user friendly) for growers to use.
Sigfredo Fuentes
added a research item
The appearance of bubbles and foam can influence the likeability of a wine even before its consumption. Since foams are essential to visual and taste attributes of sparkling wines, it is of great importance to understand which compounds affect bubbles and foam characteristics. The aim of this work was to investigate the effect of interactions among proteins, amino acids, and phenols on the characteristics of foam in sparkling wines by using synchronous fluorescence spectroscopy techniques. Results have shown that several compounds present in sparkling wines influence foam quality differently, and importantly, highlighted how the interaction of those compounds might result in different effects on foam parameters. Amongst the results, mannoproteins were found to be most likely to promote foam and collar stability, while phenols were likely to increase the small bubbles and collar height in the foam matrix. In summary, this work contributes to a better understanding of the effect of wine compounds on foam quality as well as the effect of the interactions between those compounds.
Claudia Gonzalez Viejo
added a research item
Bushfires are a common occurrence in Australia and their number and intensity is predicted to rise due to climate change. Consequently, grapevine smoke exposure and the subsequent development of smoke taint in wine will also most likely increase. This study investigated the use of near infrared (NIR) spectroscopy and machine learning (ML) for the rapid, in-field detection of smoke contamination in grapevines. Five different smoke and water misting treatments were applied to Cabernet Sauvignon grapevines for one hour at approximately seven days post-veraison under experimental conditions described previously by Kennison et al. (2008) and Ristic et al. (2011). Near-infrared measurements were then taken from berries and leaves a day after smoking. Two artificial neural networks (ANN) were developed for berry (model 1) and leaf (model 2) NIR readings used as inputs to classify different smoke treatments. Model 2 had an overall accuracy of 95.2%, with a high perfo97.7% accuracy during training with a mean square error (MSE) of 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. Model 1 had an overall accuracy of 91.7%, 95.2% 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 ANN modelling.
Sigfredo Fuentes
added a research item
Bushfires are becoming more frequent and intensive due to changing climate. Those that occur close to vineyards can cause smoke contamination of grapevines and grapes, which can affect wines, producing smoke-taint. At present, there are no available practical in-field tools available for detection of smoke contamination or taint in berries. This research proposed a non-invasive / in-field detection system for smoke contamination in grapevine canopies based on predictable changes in stomatal conductance patterns based on infrared thermal image analysis and machine learning modeling based on pattern recognition. A second model was also proposed to quantify levels of smoke-taint related compounds as targets in berries and wines using near-infrared spectroscopy (NIR) as inputs for machine learning fitting modeling. Results showed that the pattern recognition model to detect smoke contamination from canopies had 96% accuracy. The second model to predict smoke taint compounds in berries and wine fit the NIR data with a correlation coefficient (R) of 0.97 and with no indication of overfitting. These methods can offer grape growers quick, affordable, accurate, non-destructive in-field screening tools to assist in vineyard management practices to minimize smoke taint in wines with in-field applications using smartphones and unmanned aerial systems (UAS).
Sigfredo Fuentes
added 3 research items
Background and Aim: Berry water loss at late stages of ripening is a cultivar dependent-trait correlated with mesocarp cell death. We tested the hypothesis that elevated temperature anticipates the onset and increases the rate of mesocarp cell death. The implications of these putative effects on the time course of berry shrivel were also investigated. Methods and Results: We assessed the progression of mesocarp cell vitality and the degree of shrivelling in berries from a factorial field experiment combining two thermal regimes (elevated temperature and control) and two cultivars (Shiraz and Chardonnay). A bilinear model was used to objectively discriminate the onset of cell death and to quantify the rates of cell death before and after the inflection point in chronological and thermal scales. Elevated temperature advanced the onset of mesocarp cell death of berries and increased the rate of cell death in the period onset-harvest for both cultivars. There was a close correlation between the proportion of living tissue and shrivel for Shiraz, but no shrivel was observed in Chardonnay despite significant mesocarp cell death. Conclusion: Elevated temperature accelerated both mesocarp cell death and berry shrivelling in Shiraz and accelerated mesocarp cell death but had no impact on shrivel in Chardonnay. Mesocarp cell death seems necessary but not sufficient to explain berry shrivelling. Significance of the Study: Understanding the functional links between berry shrivel and mesocarp cell death and their responses to environmental drivers would likely contribute to management practices that could reduce the severity of shrivel in a context of warmer conditions.
Grapevine reproduction is an intricate process that extends over two growing seasons The conditions of the season not only influence reproductive growth of the current year, but also affect bud fruitfulness and hence potential yield for the following year. Canopy management practices are widely adopted in vineyards to maintain a balance between vegetative growth and reproductive performance with the objective to improve yield and/or berry quality (Smart 1985). Grapevine reproductive performance response varies with different management practices as source sink relationship and microclimate can be manipulated at different levels. This study aimed to investigate reproductive responses of Semillon and Shiraz vines where different canopy management practices were applied, including bunch thinning, shoot thinning, leaf removal and lighter pruning.
Sigfredo Fuentes
added 2 research items
Climate change predictions of increasing CO2 concentration, temperature and a decline in water availability for irrigation have prompted research to develop rapid and accurate plant-based monitoring tools to help assess plant performance under environmental stresses. In this study, an adaptation of Near Infrared (NIR) spectroscopy was investigated as a rapid and non-destructive method to estimate water status of almond trees. Leaves from 12 almond cross-lines were scanned using an ASD FieldSpec® 3 (350-1850 nm) spectrophotometer (Analytical Spectral Devices, Boulder, Co, USA) and then destructively measured using a pressure chamber to obtain stem water potential (Ψstem). All calibrations were developed using partial least squares (PLS) regression with full cross validation. Calibrations for Ψstem assessment showed good prediction ability (SECV <0.4 MPa) for the 12 almond crosses studied thus demonstrating that, with further research, NIR spectroscopy could become a tool to estimate water status of almond trees.
Bushfires are a common occurrence throughout the world including Australia, the Mediterranean as well as North and South America. Unfortunately, their incidence is predicted to rise due to increases in temperature, wind and drought brought on by climate change. Many of these bushfires occur in areas close to wine regions, resulting in grapevine smoke exposure. Wine produced from these smoke-affected grapes are characterised by unpalatable smoky aromas such as "burning rubber", "smoked meats" and "burnt wood". These wines are unprofitable and result in significant financial losses for winegrowers. Currently there is no in-field detection system that growers can use to assess whether their grapevines have been contaminated by smoke, instead they must harvest grapes and conduct mini-ferments which are then sent off to a commercial laboratory for analysis. This process is incredibly time consuming and destructive. This study aimed to assess 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. 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 treatments (high smoke coupled with water misting (HS M) (Fig. 2), high smoke with no water misting (HS NM), low smoke (LS) using half the amount of fuel used in the high smoke treatments to achieve half the smoke density, control with mist (Con) and control with no mist (Con NM) to Cabernet Sauvignon grapevines at approximately seven days post-veraison. NIR measurements were then taken a day after smoking using the microPHAZIR TM RX NIR Analyser, which has a spectral range of 1600-2396 nm. Spectral readings were then used as inputs to train different ML algorithms using a customised code written in Matlab® R2019a, which resulted in two artificial neural network (ANN) models with the best classification performance for either berry (Model 1) or leaf (Model 2) readings according to the different smoke treatments. The ANN models were trained to classify the leaf or berry NIR readings according to the smoke treatments (HS M, HS NM, LS, Con NM or Con M). The models were able to correctly classify the leaf and berries using the spectral readings as inputs with high accuracy. The leaf models had an overall accuracy of 92% (Model 1), 95% (Model 2). Therefore, NIR and ANN modelling may be a great promise for the detection of grapevine smoke contamination. Further research is required to relate the spectral readings to the level of volatile phenols in grapes and smoke taint development in wine.
Sigfredo Fuentes
added a research item
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.
Sigfredo Fuentes
added a research item
Cocoa is an important commodity crop not only to produce one of the most complex products such as chocolate from the sensory perspective but one that commonly grows in developing countries close to the tropics. This paper presents novel techniques applied using cover photography and a novel computer application (VitiCanopy) to assess the canopy architecture of cocoa trees in a commercial plantation in Queensland, Australia. From the cocoa trees monitored, pod samples were collected, fermented, dried, and ground to obtain the aroma profile per tree using gas chromatography. The canopy architecture data were used as inputs in an artificial neural network (ANN) algorithm and the aroma profile considering six main aromas as targets. The ANN model rendered high accuracy (R = 0.82; MSE = 0.09) with no overfitting. The model was then applied to an aerial image from the whole cocoa field studied to produce canopy vigor, and aroma profile maps up to the tree-by-tree scale. The tool developed could aid significantly the canopy management practices in cocoa trees that have a direct effect on cocoa quality.
Sigfredo Fuentes
added a research item
Leaf area index (LAI) is one of the most important parameters in physiological and functional plant models to estimate tree canopy vigor and photosynthesis. However, LAI requires either destructive or indirect methods for accurate assessment, which can be time consuming, costly, and requires specialized instrumentation. Cover photography to estimate canopy architectural parameters has shown to be effective and accurate for several forest species and horticultural tree crops such as apple trees, grapevines and cherry trees. The accuracy of the LAI estimation is highly dependent on the appropriate use of the variable light extinction coefficient (k) parameter per image. Canopy cover photography was tested on a commercial cherry plantation in Maule, Chile during seasons 2011-12 and 2013-14. Two cultivars were assessed, 'Bing' (n=80 images) and 'Sweet Heart' (n=80 images), with 10 trees per cultivar, and 4 photos representing each canopy quadrant per tree. Real LAI (LAIreal) was measured allometrically from every tree photographed for both cultivars. Real k was computed based on the inverted LAI formula and LAIreal. Artificial Neural Networks (ANN) modeling for fitting was implemented per cultivar using a customized code written in MATLAB with canopy cover (ff), crown cover (fc), canopy porosity (Φ) and clumping index (Ω) obtained from image analysis algorithms as inputs, and real k as target. The ANN fitting model to obtain a variable k showed determination coefficients (R 2) for training = 0.98 and 0.92, validation = 0.96 and 0.94, testing = 0.98 and 0.90, and final k model = 0.98 and 0.94, for 'Bing' and 'Sweetheart', respectively, in both seasons studied. This resulted in improvements in the LAI estimation for cherry trees when compared to LAIreal with R 2 of 0.80 for 'Bing' and 0.90 for 'Sweetheart'. This is a significant improvement in the assessment of canopy vigor and water requirement for tools such as VitiCanopy ® , a free LAI estimation App available for iOS and Android devices based on canopy cover photography, which can incorporate a variable k.
Sigfredo Fuentes
added a research item
Traditional grape growing areas world wide are experiencing earlier and compressed phenology due to higher temperature and irradiance which has led to excessive sugar accumulation in grapes. This is associated to negative effects on the grape and wine quality due to higher alcohol content in wines. Recent research has shown that this early, climate-related, ripening can be ameliorated through vineyard management techniques such as late leaf removal above the bunch zone. This practice was found to delay sugar accumulation in berries by up to two weeks without affecting berry chemistry. Therefore, in this study, late leaf removal was investigated as a canopy management technique to delay ripening in Semillon and Shiraz varieties in a hot climate in Australia. Yield and yield components, canopy size, berry and wine chemistry and berry and wine sensory were measured over four seasons for Semillon and two seasons for Shiraz. The results showed no effect of the proposed technique in the first year of the implementation for both varieties. However, from the second year a delay in ripening of up to ten days in Semillon and 20 days in Shiraz was observed. Furthermore, no differences were observed in berry chemistry and sensory attributes in Semillon. In the case of Shiraz, the wines obtained from defoliated vines in the second season were significantly different from the control wines and characterised by attributes consistent with overripe berries such as dry fruit and jammy aromas and flavours. Results from this study showed that this technique could be an effective tool for growers to delay ripening however the timing and severity of the treatment should be carefully considered in accordance with the seasonal climate.
Sigfredo Fuentes
added a research item
Condé, B.; Robinson, A.; Bodet, A.; Monteau, A.; Fuentes, S.; Scollary, G.; Smith, T.; Howell, K.S. Using Synchronous Fluorescence to Investigate Compounds and Interactions Influencing Foam Characteristics in Sparkling Wines. Preprints 2019, 2019030119
Sigfredo Fuentes
added a research item
The identification of Chinese medicinal plants has relied in the past on ampelographic manual assessment by experts. However, more recently machine learning algorithms for pattern recognition, have been successfully applied to leaf recognition in other plant species. These new tools make the classification of Chinese medicinal plants easier, more efficient and cost effective. This paper showed comparative results between machine learning models obtained from two methods: i) a morpho-colorimetric method and ii) a visible (VIS) / Near Infrared (NIR) spectral analysis from sampled leaves of 20 different Chinese medicinal plants. Specifically, the automated image analysis and VIS / NIR spectral based parameters obtained from leaves were used separately as inputs to construct customized artificial neural network (ANN) models. Results showed that the ANN model developed using the morpho-colorimetric parameters as inputs (Model A) had an accuracy of 98.3% in the classification of leaves for the 20 medicinal plants studied. In the case of the model based on spectral data from leaves (Model B), the ANN model obtained using the averaged VIS / NIR spectra per leaf as inputs showed 92.5% accuracy for the classification of all medicinal plants used. Model A has the advantage of being cost effective, requiring only a normal document scanner as measuring instrument. This method can be adapted for non-destructive assessment of leaves in-situ by using portable wireless scanners. Model B combines the fast, non-destructive advantages of VIS / NIR spectroscopy, which can be used for rapid and non-invasive identification of Chinese medicinal plants and other applications by analyzing specific light spectra overtones from leaves to assess concentration of pigments such as chlorophyll, anthocyanins and others that are related active compounds from the medicinal plants.
Sigfredo Fuentes
added a research item
Information about canopy vigor and growth are critical to assess the potential impacts of biotic or abiotic stresses on plant development. By implementing a Digital Surface Model (DSM) to imagery obtained using Unmanned Aerial Vehicles (UAV) it is possible to filter canopy information effectively based on height, which provides an efficient method to discriminate canopy from soil and lower vegetation such as weeds or cover crops. This paper describes a method based on the DSM to assess CG as well as missing plants from a kiwifruit orchard on a plant-by-plant scale. The DSM was initially extracted from the overlapping RGB aerial imagery acquired over the kiwifruit orchard using the Structure from Motion (SfM) algorithm. An adaptive threshold algorithm was implemented using the height difference between soil / lower plants and kiwifruit canopies to identify plants and extract canopy information on a non-regular surface. Furthermore, a customized algorithm was developed to discriminate single kiwifruit plants automatically, which allowed the estimation of individual canopy cover fractions (fc). By applying differential fc thresholding, four categories of the CG were determined automatically: i) missing plants, ii) low vigor, iii) moderate vigor and iv) vigorous. Results were validated by a detailed visual inspection on the ground, which rendered an overall accuracy of 89.5% for the method proposed to assess CG at the plant-by-plant level. Specifically, for CG category i) the accuracy was 94.1% and for ii) was 85.1%, iii) and iv) were 86.7% and 88.0% respectively. The proposed method showed also to be appropriate to filter out weeds and other smaller non-plant materials which are extremely difficult to do by common color thresholding or edge identification methods. This method can be applied on a number of agricultural operations.
Sigfredo Fuentes
added 13 research items
Projected climate change and its variability predict increased temperatures, higher evaporative demand and water scarcity over a large portion of Australia. These projections have increased the pressure to obtain high quality fruit production with more efficient water use. Several studies have demonstrated that carefully imposed water stress improves stone fruit quality parameters. However, most of the conventional methods to monitor plant water status are based on manual point measurements, which have low spatial coverage, and are resource expensive and time consuming. This seriously constraints the efficient assessment of spatial variability of water status from orchards caused by heterogeneous soil characteristics, poor irrigation distribution uniformity and/or variability in canopy structure/architecture. This work has demonstrated the capability of unmanned aerial vehicles (UAVs) in detection of water stress for a stone fruit orchard (peach and nectarine). Furthermore, it was thought until recently that night-time water uptake for C3 and C4 plants corresponded to stem and organ rehydration. Any water loss at night-time was considered minimal, corresponding to cuticle conductance of leaves with no significant conductance through the stomata. Recent studies have shown that this knowledge was incomplete. In a climate change scenario, where night-time temperatures have been forecasted to increase at a higher rate compared to diurnal temperatures, night-time water losses for non-stress and mild stress water conditions may be exacerbated (Fuentes et al. 2013; 2014). Furthermore, counter-intuitively, elevated CO2 may increase nocturnal water loss (Zeppel et al. 2012). Current methodologies and modelling algorithms to estimate ETreal, assume stomata conductance at night-time to be zero. Therefore, accurate crop coefficient calculations need to incorporate these important night-time water losses, which are not coupled to photosynthesis, hence reducing water use efficiency.
To quantify night-time (S n) and diurnal (S d) tree water uptake, two sets of sap flow sensors (heat-pulse compensated) were installed per tree in the north-east and south-west sides of the trunk in three trees per treatment. There were two treatments: (1) control, irrigated with 100 % ETc (T100), and (2) deficit, irrigated at 60 % ETc (T60) with daily irrigations at the peak atmospheric demand (December–January). Normalised S n by trees was in the range of 15–25 % throughout the season, compared to normalised S d, for T100 and T60, respectively. Furthermore, S n was parabolically correlated to plant water status from the previous day, measured as midday stem water potential. We also found strong correlations between S n and nocturnal vapour pressure deficit for T100 and T60, indicating that nocturnal transpiration was significant for both treatments. Differences in S n were observed for the NE and SW sensors for T60, being significantly less for the NE side (sunny side) compared to the SW side (more shaded). No differences were observed for T100 regarding probe positioning.
Sigfredo Fuentes
added 3 research items
Australian vineyards need to continue to produce high quality fruit to remain competitive in the global market. Vine development, grape ripening and wine sensory attributes are highly influenced by the physical environment in which the vines grow and their management. Hence, vineyard practitioners use a variety of management practices in order to control canopy size and to achieve vine balance. However, the link between canopy management, canopy architecture and wine quality is not clearly understood. In this study five Shiraz vineyards from five Australian wine regions, were subjected to different canopy management treatments with the aim to alter yield or canopy size compared to an untreated control. Measurements were conducted on canopy growth and architecture, yield and yield components and grape and wine chemistry. Canopy architecture correlated better than yield or pruning weight with berry and wine chemistry and sensory measures. This study produced valuable information on the links between canopy architecture, vine balance and berry quality for different varieties, training systems and wine regions.
Canopy management trials conducted in commercial vineyards and at the Waite campus, University of Adelaide, demonstrated that a new tool, VitiCanopy, was effective at capturing canopy architecture differences that resulted in fruit and wine compositional changes and berry and wine sensory differences. Vine balance is more easily measured with VitiCanopy than other traditional measures that are labour intensive. VitiCanopy measurements of leaf area index, canopy porosity and growth rate were also highly correlated with vine performance and quality and can be measured early in the growing season, allowing within season management decisions to be made. These outcomes provide new information and tools that will lead to more informed vineyard management in Australian vineyards.
In the Chilean viticultural industry, Carménère is considered an emblematic cultivar that is cultivated mainly in arid and semi-arid zones. For this reason, it is necessary to use precise irrigation scheduling for improving water use efficiency (WUE), water productivity (WP), yield and wine quality. This study evaluated the effects of three deficit irrigation strategies on gas exchange variables, WUE, WP and yield components in a drip-irrigated Carménère vineyard growing under semi-arid climatic conditions during two consecutive seasons (2011/12 and 2012/13). The irrigation strategies were applied in completely randomized design from fruit set (S) to harvest (H). The first irrigation strategy (T1) involved continuous irrigation at 100% of actual evapotranspiration (ETa) from S to the veraison (V) period and at 80% of ETa from V to H. The second irrigation strategy (T2) involved irrigation at 50% of ETa from S to H and the third one (T3) involved no-irrigation from S to V and at 30% of ETa from V to H. The results indicated that there was a significant non-linear correlation between net CO2 assimilation (AN) and stomatal conductance (gs), which resulted in three zones of water stress (zone I = gs > 0.30 mol H2O m-2s-1; zone II = between 0.06 and 0.30 mol H2O m-2s-1; and zone III = gs < 0.06 mol H2O m-2s-1). The use of less water by T2 and T3 had a significant effect on yield components, with a reduction in the weight and diameter of grapes. A significant increase in WP (7.3 kg m-3) occurred in T3, which resulted in values of WUE that were significantly higher than those from T1 and T2. Also, a significant non-linear relationship between the integral water stress (SIΨ) and WP (R2 = 0.74) was established. The results show that grafted Carménère vines were tolerant to water stress although differences between cultivars/genotypes still need to be evaluated.
Sigfredo Fuentes
added a research item
The application of computer vision algorithms and chemometric fingerprinting using near-infrared spectrometry (NIR) of plant leaves, offer enhanced capabilities for ampelography by providing more accurate methods to discriminate leaves based on morphological parameters, and chemometrics, respectively. This paper showed that machine learning algorithms based on morpho-colorimetric parameters and NIR analysis separately, were able to automatically classify leaves of 16 grapevine cultivars. The artificial neural network (ANN) model developed with morpho-colorimetric parameters as inputs (Model 1), and 16 cultivars as targets, rendered an accuracy of 94% to classify leaves for all cultivars studied. The ANN model obtained with the NIR spectra per leaf as inputs (Model 2), and the real classification as targets, rendered 92% accuracy. The automatic extraction of morpho-colorimetric data, NIR chemical fingerprinting and machine learning modeling rendered rapid, accurate and non-destructive methods for cultivar classification, which can aid management practices.
Sigfredo Fuentes
added a research item
Artificial intelligence (AI) and machine learning (ML) has been used almost interchangeably since these two disciplines were founded. Albeit the subtle differences between these two concepts, these nuances matter in terms of their specific potential applications, such as in viticulture and oenology. The development of ML algorithms is related to its definition of a machine/computer to produce an output or perform some actions that it wasn’t programmed to do. This process can be achieved using vast and complicated data in a more efficient way. Furthermore, as the name states it, the algorithms are capable of “learning” when more data is incorporated in the process. We can think about ML algorithms as small software codes that have been programmed through an automatic training of data to fit a determined target. On the contrary, AI is the whole discipline dedicated to make machines more smart, in which ML is a part of .
Sigfredo Fuentes
added 7 research items
The use of small-site intensive soil water monitoring with multiple depth sensor arrays in representative soil volumes can indicate plant water usage rates in the broader scale. Multiple numbers of Sentek soil moisture monitoring probes were installed into sites within a range of different crops (including: grapevine, almond, cotton and watermelon) and irrigated with a range of different methods (drip, sprinkler and sub-surface drip). The probes were arranged within the plant root zones in transect lines extending from the water emitters. Soil water data was collected on a near-continuous basis simultaneously across the probe array. This was analysed using a new wetting pattern analysis software called IrriMAX™ 2D Imager to visualise the water distribution in the soil in 2-dimensions. This software uses automated cubic spline interpolation techniques to generate contour maps and video animation of water dynamics according to different levels of soil water. The soil water scale is obtained from the whole dataset and painted with false colours to visualise the dynamics of water movement throughout the soil profile. This method allows the visualisation of the spatial distribution of water in the soil and the duration of time over which such a distribution persists. This gave insights into the plant water uptake dynamics and soil drainage characteristics that could be extrapolated to the whole field for irrigation management purposes. Careful consideration of the data showed that root architecture and the passage of mobile salts could also be assessed at the root-zone scale. This visualization technique was shown to be a valuable adjunct to understanding the movement of water and nutrients throughout the root zone.
Night-time sap flow is an important unseen factor that contributes significantly to total evapotranspiration (ET) of horticultural and fruit tree crops as It has been shown in literature. • There is a direct and strong correlation between night-time transpiration and vapor pressure deficit (VPD) for non-stress situations. • The nocturnal process will be likely increased in a climate change scenario, where it has been forecasted increased rates in night-time temperatures compared to diurnal temperatures. • Recent research has also shown that night-time water consumption accounts for 15-30% for field-grown horticultural crops compared to diurnal transpiration, which is not considered in ET models (Fuentes et. al. 2014; Fuentes et al. 2013; Escalona et al. 2013).
Irrigation scheduling is critical for olive orchards, since it affects both fruit yield and olives composition. Regulated deficit irrigation (RDI) strategies have been applied with positive results in the past. However, to successfully regulate stress levels, it is necessary to have accurate measurements of plant water status, which is usually done using a pressure chamber. Canopy temperature (Tc) is another potential accurate indicator of water stress. Therefore, the objective of this study was to evaluate three methods to obtain Tc values from infrared thermal images to calculate the crop water stress index (CWSI). Furthermore, the relation between CWSI and midday stem water potential (MSWP) was also studied. The methods used to obtain Tc were: i) Tc1 obtained from a region of interest within the image; ii) Tc2 obtained from whole image; iii) Tc3 obtained from a filtered image using an interactive filtering process to exclude non-leaf material (low and high temperature values). The infrared thermal images were obtained using an infrared camera (Model i40, FLIR Instruments) in parallel with MSWP measurements from trees under different RDI strategies in a drip irrigated olive orchard (Olea europaea L. ‘Arbequina’) located in Pencahue valley, Maule Region, Chile (35°23’L.S; 71°44’L.W; 96 m a.s.l.) during the 2011-2012 season. Results obtained in this study showed that CSWI3 calculated using Tc3 had a better correlation with MSWP compared to the two other methods studied. The interactive filter process to obtain Tc values could be used in olive orchards as a fast and cheap indicator of water stress. Further studies are required to automate the analysis process.
Sigfredo Fuentes
added 77 research items
Adequate and timely irrigation based on real-time monitoring of crop water status is critical for efficient and sustainable water use. However, detection of water status in large crop fields is not a trivial task as manual inspection can be time consuming and costly. Moreover, the symptoms of water stress are visually detectable only after the crops are already in a significantly water deficient stage. Consequently, capability of monitoring water status in crops on a regular basis could maximize productivity and water use efficiency. As an indicator, the crop water stress index (CWSI) has been widely used to estimate water status in the crop fields. CWSI can be derived from ground-based leaf temperature measurements, however, airborne or UAV-borne high-resolution thermal sensing provides a superior platform to cover large regions within a short time window. In this paper, UAV-borne thermal sensing was conducted to map plant water stress and spatial variability in water control and deficit plots over 1 ha of a nectarine orchard at an altitude of 100 m from ground level. Targets of ground control points (GCPs) were designed to suit the image spatial resolution as well as the visibility in the thermal infrared spectral range. The target was made of aluminium body marked with a black cross, which can be detected as a cool object in the thermal infrared image due to its low emissivity. Thermal infrared images were post-processed to generate single temperature-based orthomosaic image for the entire study field. CWSI map was computed using canopy temperatures at the centre of canopies from the mosaic image. Histogram analysis was used to estimate the lower boundary temperature (Twet), representing the temperature of fully transpiring leaves. The upper boundary temperature (T dry) was determined by air temperature + 6 ⁰C. Ground measurements of midday stem water potential (SWP) and stomatal conductance (g c) were collected concurrently with UAV operation and used to correlate the thermal measurement to crop biophysical parameters. Results showed that CWSI was in good agreement with both SWP and g c with determination coefficients (R 2) of 0.92 and 0.97, respectively. Thus, remotely estimated CWSI from a UAV platform can play an important role in effective mapping of spatial variability of nectarine water stress and subsequently in optimal management of irrigation. Keywords: Unmanned Aerial Vehicle (UAV), thermal infrared imagery, Crop Water Stress Index (CWSI), canopy temperature, Stem Water Potential (SWP)
The Wine & Viticulture Journal has previously kept readers informed of some of the latest outcomes from the Vineyard of the Future (VoF) - a multinational project whose aims include testing new technologies and investigating the potential effects of climate change in trial vineyards. This article explores the latest successes from the VoF project – the development of an app that estimates leaf area index and canopy porosity, biological sensors for vineyards and wineries, modelling strategies using big data, robotic pourers, the use of unmanned aerial systems and remote sensing.
Terpenoids were extracted from grape vine bunches during plant development and analysed by GC-MSD. The grapevines analysed were from a commercial harvest of Vitis vinifera cv. Shiraz. The terpenoids were analysed from 4 weeks post flowering (wpf) to harvest in one season. The data are presented with the structure of the compound and aroma profile and semi-quantified. The sub-class of sesquiterpenes was given special attention, and this data set describes the first analysis of these compounds during ripening of this important economic crop. Sesquiterpenes may have a hitherto described contribution to wine aroma. This data set may provide insight into biosynthetic pathways and aroma chemistry. Interpretation of our data and further discussion can be found in “Terpene evolution during the development of Vitis vinifera L. cv. Shiraz grapes” (Zhang et al., 2016) [1].
Sigfredo Fuentes
added a project goal
To test new and emerging technologies for vineyard operations and management, ranging from sensor network technology and remote sensing from UAV and satellite. This project also considers the use of machine learning and artificial intelligence to analyse data obtained with these systems and big data available from vineyards to estimate phenological stages occurrence and prediction of yields.