"Flavescence doree" ( FD) is a highly transmissible disease very closely monitored in Europe as it reduces vine productivity and causes vine death. Currently, this disease is controlled by a two-pronged approach: spray insecticide on a regular basis to kill the vector and by experts surveying each row in a vineyard. Unfortunately, these experts are not able to carry out such a task every year on every vineyard and need an aid for planning their surveys. In this study, we propose and evaluate an original automatic method for the detection of FD, based on computer vision and artificial intelligence applied to images acquired by proximal sensing. A two-step approach is used, mimicking expert's scouting in the vine rows: i) the three known isolated symptoms are detected, ii) isolated detections are combined to make a diagnosis at vine scale. To achieve this, a detection deep neural network is used to detect and classify non-healthy leaves into three classes - 'FD symptomatic leaf', 'Esca leaf' and 'Confounding leaf' - while a segmentation network retrieves FD symptomatic shoots and bunches. Finally, the association of the detected symptoms is performed by a RandomForest classifier allowing a diagnosis at the image scale. The experimental evaluation is conducted on images collected on 14 blocks planted with 5 grape cultivars, allowing the study of the impact of acquisition conditions and variability of symptom expressions among grape cultivars.

The grapevine is vulnerable to diseases, deficiencies, and pests, leading to significant yield losses. Current disease controls involve monitoring and spraying phytosanitary products at the vineyard block scale. However, automatic detection of disease symptoms could reduce the use of these products and treat diseases before they spread. Flavescence dorée (FD), a highly infectious disease that causes significant yield losses, is only diagnosed by identifying symptoms on three grapevine organs: leaf, shoot, and bunch. Its diagnosis is carried out by scouting experts, as many other diseases and stresses, either biotic or abiotic, imply similar symptoms (but not all at the same time). These experts need a decision support tool to improve their scouting efficiency.To address this, a dataset of 1483 RGB images of grapevines affected by various diseases and stresses, including FD, was acquired by proximal sensing. The images were taken in the field at a distance of 1-2 meters to capture entire grapevines and an industrial flash was ensuring a constant luminance on the images regardless of the environmental circumstances. Images of 5 grape varieties (Cabernet sauvignon, Cabernet franc, Merlot, Ugni blanc and Sauvignon blanc) were acquired during 2 years (2020 and 2021).Two types of annotations were made: expert diagnosis at the grapevine scale in the field and symptom annotations at the leaf, shoot, and bunch levels on computer. On 744 images, the leaves were annotated and divided into three classes: ‘FD symptomatic leaves’, ‘Esca symptomatic leaves’, and ‘Confounding leaves’. Symptomatic bunches and shoots were, in addition of leaves, annotated on 110 images using bounding boxes and broken lines, respectively. Additionally, 128 segmentation masks were created to allow the detection of the symptomatic shoots and bunches by segmentation algorithms and compare the results to those of the detection algorithms.

“Flavescence dorée” (FD) is a grape vine disease caused by the bacterial agent “Candidatus Phytoplasma vitis” and spread by the leafhopper Scaphoideus titanus Ball (Hemiptera: Cicadellidae). The disease is very closely monitored in Europe, as it reduces vine productivity and causes vine death and is also highly transmissible. Currently, the control method used against this disease is a two-pronged approach: i) the spraying of insecticide on a regular basis to kill the vector, and ii) a survey of each row in a vineyard by experts in this disease. Unfortunately, these experts are not able to carry out such a task every year on every vineyard and need an aid for planning their survey.In this study, we propose and evaluate an original automatic method for the detection of FD based on computer vision and artificial intelligence algorithms applied to images acquired by proximal sensing. A two-step approach was used, mimicking an expert’s scouting in the vine rows: (i) the three known isolated symptoms (red or yellow leaves depending on variety, together with a lack of shoot lignification and the presence of desiccated bunches) were detected, (ii) isolated detections were combined to make a diagnosis at image scale; i.e., vine scale. A detection network was used to detect and classify non-healthy leaves into three classes: ‘FD symptomatic leaf', 'Esca leaf' and 'Confounding leaf'; while a segmentation network was used for the retrieval of FD symptomatic shoots and bunches. Finally, the association of detected symptoms was performed by a RandomForest classifier for diagnosis at the image scale. The experimental evaluation was conducted on more than 1000 images collected from 14 blocks planted with five different grape varieties. The detection of the isolated symptoms achieved a precision of between 0.67 and 0.82 and a recall of between 0.39 and 0.59. The classification at the image scale obtained very good results when applied to images acquired under the same conditions, with the same grape varieties as the training images (precision and recall of more than 0.89). The results of the tests on the other grape varieties show the importance of having some of them in the training base in these AI-based approaches.

A new model for grapevines (Vitis vinifera) is the first perennial fruit crop model using the Agricultural Production System sIMulator (APSIM) Next Generation framework. Modules for phenology, light interception, carbohydrate allocation, yield formation and berry composition were adapted or added into APSIM Next Generation to represent the nature of fruit-bearing vines. The simulated grapevine phenological cycle starts with the dormancy phase triggered by a critical photoperiod in autumn, and then goes through the subsequent phenophases sequentially and finally returns to dormancy for a new cycle. The canopy microclimate module within APSIM Next Generation was extended to allow for row crop light interception. The carbohydrate arbitrator was enhanced to consider both sink strength and sink priority to reflect carbohydrate reserve as a concurrent competing sink. Weather conditions and source-sink ratio at critical developmental stages were used to determine potential grapevine yield components, e.g. bunch number, berry number and berry fresh weight. The model was calibrated and tested extensively using four detailed data sets. The model captured the variations in the timing of measured budburst, flowering and veraison over 15 seasons across New Zealand for five different varieties. The calculated seasonal dynamics of light interception by the row and alley were consistent with field observations. The model also reproduced the dynamics of dry matter and carbohydrate reserve of different organs, and the wide variation in yield components caused by seasonal weather conditions and pruning regimes. The modelling framework developed in this work can also be used for other perennial fruit crops.

Larger easily visible animals and plants are negatively affected by agrochemicals used for intensive food production, but we do not understand the general spatial and temporal effects of agrochemicals on the multitudes of bacteria, fungi, and small invertebrate animals that underpin ecosystem productivity. We sequenced the 16S, ITS2, and COI DNA barcode regions from 648 New Zealand vineyard soil samples managed under either conventional or low-agrochemical-input conservation approaches across two regions and three seasons in 1 year and discovered at least 170,000 phylotypes (taxa) with >97% genetic identity. Management approach correlated with a significant 2%-10% difference in the abundances of phylotypes that differed over regions and seasons. Although the data show that agrochemicals do not have a large effect on soil biodiversity on average, the important finding is that the magnitude of impact differs between taxa types and locations, and some taxa most affected also influence the quality of agricultural produce.

Seasonal differences in weather conditions cause marked variation in grapevine yield. However, quantitative relationships between various yield components and climatic factors at field scales are still lacking. By using a long-term field trial, we quantified the correlation between weather conditions during the key development stages and the yield components of Vitis vinifera L. Sauvignon blanc growing under cool-climate conditions. A long-term phenology and yield monitoring trial using both two-cane and four-cane trained vertically shoot positioned (VSP) Sauvignon blanc vines was established in four vineyards in Marlborough, New Zealand in 2004. Phenology, bunch number, berry mass, yield and meteorology records were collated. A multivariable mixed linear model was used to assess the relationship between various yield components and weather conditions. The critical periods for each yield component and weather factor were optimised based on the maximum likelihood returned from the mixed linear model. The optimised critical periods of temperature for all yield components occurred mainly before 50 % flowering either in the previous season (during inflorescence initiation) and the current season, indicating the importance of the pre-flowering period on yield formation. Out of all weather factors, maximum daily temperature had the largest effect on bunch number and overall yield and strongly influenced berry number and bunch mass. Rainfall near flowering time had a negative effect on berry mass and bunch mass, but post-flowering rainfall had a strong positive effect. The statistical model explained 60 to 85 percent of the seasonal variations in bunch number, berry number, berry and bunch mass and yield per vine.

Although Sauvignon Blanc (SB) grapes are cultivated widely throughout New Zealand, wines from the Marlborough region are most famous for their typical varietal combination of tropical and vegetal aromas. These wines differ in composition from season to season as well as among locations within the region, which makes the continual production of good quality wines challenging. Here, we developed a unique database of New Zealand SB grape juices and wines to develop tools to help winemakers to make blending decisions and assist in the development of new wine styles.

We quantified the importance of postharvest carbohydrate assimilation and nitrogen availability to replenish vine reserves, over and above maintaining optimal growth, productivity, and fruit quality of high-yielding, vigorous Sauvignon blanc grapevines. To create different carbohydrate (CHO) and nitrogen (N) reserve concentrations, our factorial-design trial consisted of a postharvest defoliation treatment overlaid with a pruning treatment in which 48 or 72 nodes were retained on four-or six-cane vertical shoot positioned vines, respectively. In defoliation (Defol) vines, all leaves were removed immediately after fruit harvest, while foliated vines (Fol) went through normal senescence. From just after ectodormancy in 2008, samples of root and trunk tissue were taken throughout the years for CHO and N analyses and results were compared with annual yield data. Both defoliation and node number treatments reduced vine growth and yield. Additionally, differences in CHO and N of the permanent structure were found. Depleted winter reserves in trunk and root were replenished during the next growth cycle, suggesting that grapevine N and CHO partitioning favor survival of the permanent structure over increasing vine size and yield. However, after two consecutive years of defoliation, the cumulative effects of smaller, less fruitful canes from year one and reduced carbohydrates from the subsequent year reduced both yield and vegetative growth in the third growing season. Therefore, even the short-lived postharvest canopy in cool climates contributes to the vine CHO economy. Defoliation or excessive crop loads affected carbohydrate reserves in vines, but only after several consecutive years of low recharge; this manifested iteself in lower yields and poorer vegetative growth.

This study presents a comprehensive lipidome analysis of Sauvignon blanc grape juice by combining GC-MS based fatty acid profiling with shotgun lipidomics strategy. We observed that despite grape juice being a water based matrix it contains a diverse range of lipid species, including common saturated and unsaturated free and intact fatty acids as well as odd-numbered and hydroxy fatty acids. Based on GC-MS quantitative data, we found that the total lipid content of grape juice could be as high as 2.80 g/L. The majority of lipids were present in the form of complex lipids with relatively small amount of free fatty acids (<15%). Therefore we concluded that the lipidome should be considered an important component of grape juice with the potential to impact on fermentation processes as well as on the sensorial properties of fermented products. This work serves as a hypothesis generating tool, the results of which justify follow-up studies to explore the influence of the grape juice lipidome and lipid metabolism in yeast on the aroma profile of wine.

In Marlborough, New Zealand, olives are becoming an important crop alongside grapes. However, despite olives being drought resistant, they are generally planted on the poorer free-draining soils. Also, with the strong increase in cropping area, the demand for irrigation water has increased dramatically. In this research, we investigate the impact of short-term water stress on plant physiological processes, crop yield and oil quality in Marlborough, New Zealand. For that purpose, during the dry summer of 2000–2001, two trees were kept without irrigation for 64 days while two neighbouring trees were irrigated following standard practice. The trees were measured for transpiration (E), leaf and stem water potential (ΨL and ΨS), every other day, from dawn to dusk for three weeks from just before irrigation was started up again. All four trees were wired up for measuring stem sap flow (T) which was recorded hourly and a basic meteorological station provided weather data. Fruit and shoot development was measured weekly. It was found that under the short period of dry conditions with soil moisture (() dropping to <5%, olive trees kept functioning at a very low level with ΨL and ΨS reduced from −1 to <−4.0MPa (T) reduced from 20 to 5mm/h and (E) reduced from 1.5 to 1.0mmolm−2s−1. Within 10 days of restarting irrigation all these parameters were back to pre-drought levels. Both fruit and shoot growth came to a standstill within a week after drought was induced. During the first few days after re-watering, a high variability in ΨL was found between leaves from the same trees. This variability disappeared after ∼six days. Shoot growth did not recover after re-watering but fruit growth rate, became the same as for continuously irrigated trees within days, but fruit size did not manage to recover before harvest. Yield from the dry trees was low because berry and pit weight were reduced by almost 50% at harvest, had a lower oil and percentage and were lower in phenolics. Stem sap flow was found to give a very good continuous measurement for the hydration status of the olive trees.

The flowering of Vitis vinifera spreads over two seasons. Tendrils and inflorescences have a common origin known as anlage or uncommitted primordia. The fate of the uncommitted primordia depends oil the cytokinin-gibberellin balance, with cytokinins promoting transition to flowering and gibberellins inhibiting it. High temperature and high light are induction stimuli for flowering. Neither photoperiod nor vernalization is very relevant for flowering induction. Inflorescence primordia development in latent buds stops after the formation of secondary and tertiary branches, approximately one month before shoot periderm formation. Buds resume growth after dormancy, with further branching of inflorescences before differentiation of individual flowers. Warm weather at budburst favors further inflorescence differentiation, resulting in more clusters per shoot, while cool weather favors differentiation of more flowers per clusters and fewer clusters per shoot. Environment and cultural practices influence flowering, either directly or indirectly via their impact on photosynthesis and nutrient availability. Cultural practices encouraging light penetration into the canopy favor flower initiation, while practices resulting in shading have a detrimental impact. Flower formation occurs through a series of sequential steps under hormone-mediated genetic control. The first genetic change involves the switch from the vegetative to the floral state, in response to different environmental and developmental signals, through the activity of floral-meristem identity genes. Second, the floral meristem is patterned into the whorls of organ primordia through the activity of floral-organ identity genes. Third, the floral-organ identity genes activate downstream effectors that specify the various tissues which constitute the different floral structures. The flowers are hermaphroditic and most are self-pollinated but cross-pollination also occurs. Fertilization is hindered by cool rainy weather and favored by warm dry weather.

Aims: The young alluvial soils of the Wairau Plains, Marlborough, New Zealand, are considered to play an important role in determining this unique wine style. The aim of this experiment was to investigate, within a single vineyard, the impact of soil texture on vine vigour, vine earliness and fruit composition. Methods and results: Trunk circumference and pruning weights, were greater as the depth to gravel increased. Soil conductivity measurements, using an electromagnetic sensor (EM38) in conjunction with global positioning related well to vine trunk circumference. Where gravels came to the surface, soil temperatures measured at 30 cm depth were consistently higher by 1 to 2 °C (air temperature was unaffected) and vine phenology was more advanced, when compared to the deep silt soils. At harvest, fruit soluble solids and pH were higher and titratable acidity lower when vines grown on shallow soils, but soil type had no significant effect on fruit yield. Conclusions and significance of the study: Vine phenology during the growing season and fruit composition at harvest but not yield, reflect changes in soil texture over quite short distances within vineyards on the Wairau Plains. Within a vineyard, the higher the proportion of gravelly soils, the more advanced the vine phenology and the riper the fruit and wine style will exhibit riper (more tropical) and lower unripe (herbaceous) characteristics on a particular date.

For six years, different field trials were used to examine the effect of irrigation on the yield and fruit quality of Sauvignon blanc grapevines in Marlborough, New Zealand. Control vines were given irrigation volumes equivalent to 100% of ETcrop (IR 100 ), minus an allowance for effective rainfall. Different levels of regulated deficit irrigation (RDI) were applied to other vines. Treatments varied from 80% (IR 80 ) to 0% (IR 0) of the control irrigation volume. During the first two years of the trial, soil moisture content (SMC) of the control vines was maintained above 0.20 L L -1 . In subsequent years, SMC was gradually lowered from 0.20 L L -1 at flowering to 0.10 L L -1 from veraison through to harvest. In only one year (2004/05), a significant difference in yield was found when IR 100 and IR 20 both yielded >8 kg vine -1 compared with IR 80 vines where yields were > 6.5 kg vine -1 . We attribute this lower yield to smaller bunch weights. Minor differences in berry quality attributes were also observed in just one season (2006/07) when o Brix at harvest was significantly higher for IR 100 compared with the other treatments. Also sensory analysis revealed only minor influences of irrigation treatments within single years, but large differences between years. The year to year variation in yield, juice quality and sensory aspects suggested that a change in climate might have a greater impact on wine character compared with irrigation effects. Unravelling the cause for these changes is the focus of our research programme.

Grapes in Marlborough are typically grown on a vertical shoot positioned trellis system (VSP). For this purpose Pinus radiata posts are treated with CCA, a mixture of copper (Cu), chromium (Cr) and arsenic (As), giving a wood concentration of 1,730, 3,020 and 2,410 mg/kg, respectively on a dry matter basis. The CCA levels around the posts in different soils were investigated and assessed for the potential leaching of CCA into ground water. An initial survey showed leaching of all three heavy metals from the treated posts into the soil surrounding the posts (0.2% of the total vineyard area) compared with the control, depending on vineyard age and soil type. The rate of movement out of the posts was calculated from posts placed in lysimeters. HortResearch's Soil Plant Atmosphere Model (SPASMO) was used to predict the leaching rate of CCA. For As, leaching was found to be 5 mg/post/month, with the Cr rate being about twice that. Further modelling revealed a steady plume of As moving downwards after about 200-300 years. However, longterm hydrogeological modelling showed that sufficient aquifer water flow prevented the accumulation of CCA in the ground water. The modelling approaches are discussed.

There have been conflicting reports as to the extent that copper–chromium–arsenic (CCA) treatments leach from timber. In New Zealand, vineyards utilise CCA-treated posts at a rate of 579 posts per hectare. This represents a potential CCA burden on the soil of 12, 21, and 17 kg/ha, respectively, for the three elements. Given a replacement rate of 4% per year, the use of CCA-treated posts may result in an accumulation of these elements in the soil, possibly leading to groundwater contamination. We undertook a general survey to determine the extent of CCA leaching from treated vineyard posts. Treated Pinus radiata posts were sampled at six sites around the Marlborough region of New Zealand to represent a range of post ages and soil types. For each post, above- and belowground wood samples were taken. As well, the soil adjacent to the post was sampled at a 50 mm horizontal and 100 mm vertical distances from the post. The belowground wood samples of the posts had significantly lower CCA concentrations than the aboveground portions, which were not significantly different from new posts. This indicates leaching. Soils surrounding the posts had significantly higher CCA concentrations than control soils. Higher CCA concentrations were measured under the posts than laterally. Some 25% of the samples exceeded 100 mg/kg As, the Australian National Environment Protection Council (ANEPC) guideline level for As in agricultural soil, and 10% exceeded 100 mg/kg Cr, the ANEPC limit for chromium. At one site, we found a significant positive correlation between post age and CCA-leaching. The CCA issue could be eliminated by using alternative posts, such as steel, concrete, or untreated woods such as Eucalyptus or beech. Alternatively, CCA-treated posts could, for example, be lacquered or otherwise protected, to reduce the rate of CCA leaching.

Abstract. Winegrape growing in many parts of the world, including Marlborough, New Zealand, uses treated-timber posts to act as supports for the grapevine's canopy. At a density of 580 posts per hectare, the H4-process treated supports result in an areal loading of CCA of: Copper (12 kg-Cu ha−1), Chromium (21 kg-Cr ha−1) and Arsenic (17 kg-As ha−1). Arsenic is the most mobile and toxic of the CCA-treatment cocktail. We describe experiments which indicate that about 4–6 mg-As month−1 post−1 is released from the subterranean part of the post. We have used SPASMO (Soil Plant Atmosphere System Model) to predict post-to-soil leakage, as well as the pattern dynamics of leaching and exchange around the post. Locally the pattern dynamics of transport and fate are controlled by the soil's chemical characteristics and the prevailing weather. Over its 20-year lifetime, the concentration of arsenic, both that adsorbed on the soil and in the soil solution, exceeds guideline values for soils (100 mg-As kg−1) and drinking water (10 μg-As L−1). Under a regime of 5% annual replacement of posts, the spatially averaged concentration of arsenic leaching through the soil is predicted to rise to 1.25 to 1.7 times the drinking water standard, depending only slightly on the soil type. The steady value is primarily controlled by the arsenic-release rate from the post. These steady values were used in a simple hydrogeological model of the major Marlborough aquifer systems to determine whether the subterranean flow of water could dilute the descending plumes of arsenic coming from above. Except for the sluggish aquifers of the southern valleys in Marlborough, most of the aquifer systems seem capable of diluting the leachate to between one tenth and one twentieth of the drinking water standard. The upscaling of our modelling of the local pattern dynamics spanned six orders of spatial magnitude, and four orders of time dimension.

We have tested the inverse modeling approach to derive macroscopic water stress parameters (MWSP) using different types of information, such as soil water content, pressure head, and transpiration rate. This testing was performed by numerical experiments considering a multilayered soil growing grapevines under three different irrigation regimes and two contrasted water stress scenarios. The results indicate that measurements of the soil water content alone do not contain enough information to estimate MWSP. Nonuniqueness is likely to occur, and the MWSP estimates may contain large uncertainties. However, the incorporation of only transpiration measurements into the objective function does allow the deriving of accurate MWSP. This contrast is mainly due to the difference of sensitivity to the MWSP, which is much higher for the transpiration than for the soil water content. Results obtained using only soil water pressure head measurements are similar to or poorer than those obtained using transpiration data. Moreover, visual inspection of response surfaces of the objective function suggests that the incorporation of further information in addition to transpiration into the objective function is not of great value for the identification of MWSP. Uncertainties for the MWSP estimated using the three types of information combined are in most cases only 1.3 times smaller than when transpiration measurements alone are incorporated into the objective function. Beyond the specific results obtained for the estimation of MWSP we find that the parameters estimates and their associated uncertainties are strongly dependent upon the type, quantity, and quality of the information included into the objective function. Hence inverse modeling may provide a means to design better experiments.

With a fast change of land use in Marlborough from extensive pastoral farming to intensive irrigated viticulture, a need has risen to investigate the sustainable use of the available water. In 2001 a 5 ha irrigation research project was installed in a Marlborough Sauvignon Blanc vineyard. Irrigation treatments installed were control (compensate 100% for crop evapotranspiration (ETO)), 80%, 70% and 60% of ETO. During the two years that the Regulated Deficit Irrigation (RDI) trial has run so far, very different climatic conditions created much greater differences in yield and vegetative growth, than up to 40% reduction in irrigation, none of which were significant. The use of sap flow in the vines has been fine-tuned and is now giving reliable results on which to base vine water need.

The Marlborough District Council (MDC) and the Marlborough Grape Growers Association are concerned about minimising the environmental risks associated with intensive pesticide use and/or the inefficient use of irrigation water. HortResearch has been commissioned to carry out a measurement and modelling exercise to determine the irrigation demand for Marlborough vineyards, and to assess the fate of surface-applied chemicals that might be used under the export wine grape spay schedule. This paper describes field experiments to measure the vineyard water balance, and presents modelling results from a risk assessment of pesticide fate. Marlborough is New Zealand's largest grape growing area. Vineyards are expanding rapidly across the region. This expansion is placing increased pressure on developers, who are seeking more water for irrigation of their high-value crops, as well as resource and policy analysts who are seeking a wiser stewardship of this precious resource. Effective management of water is critical because all of the main population centres are totally or partially dependent on groundwater for their domestic supply. A recent survey in Marlborough noted that '… the region is approaching a crossroads in water management, with a growing deficit between natural supply and the demand of water' (Davidson, 2001). The challenge for the irrigators and the regulators, as well as the wider community of domestic users, is to manage the use of water more efficiently. Vineyards also adopt an intensive pesticide spray programme, with a range of fungicides used to control mildews during the growing season and Botrytis cinerea at harvest, insecticides used for leaf roller, mealy bug and mite control, and herbicides used for general weed control under the vines. The intensive use of pesticides could pose an environmental risk of leaching to groundwater, especially if chemicals are inappropriately applied to the land. The risk of contamination will depend on the timing and rate of application, as well as the specific physio-chemical properties of each chemical. But soil and climatic conditions can also play an important role in determining the environmental fate. The Marlborough District Council (MDC) and the Marlborough Grape Growers Association are concerned about minimising the environmental risks associated with intensive pesticide use and/or the inefficient use of irrigation water. HortResearch has been commissioned to carry out a measurement and modelling exercise to determine the irrigation demand for Marlborough vineyards, and to assess the fate of surface-applied chemicals that might be used under the export wine grape spay schedule. The computer model that we used for this task (SPASMO - Soil Plant Atmosphere System Model), links the mechanisms of soil water flow through the root zone, with the complex pesticide transformations that result from both natural processes, and those consequent upon the application of a pesticide to the soil surface (Sharma et al., 2003). The calculations are based on local soil and climate data, they use published chemical-transport properties, and they assume 'normal' spray diaries for the export wine grapes. This paper describes field experiments to measure the vineyard water balance, and presents modelling results from a risk assessment of pesticide fate.