Previous  | Next Completed project

The yield: quality nexus. Substantiating similarity in the patterns of variation in grape yield/vine vigour and indices of fruit quality


This project sought to understand the ‘yield:quality nexus’ – the presumed link between fruit yield and quality such that higher yield means lower quality. It worked at the within-vineyard scale using the tools of precision viticulture (yield monitors, canopy sensors, GPS, etc). Albeit constrained by just two seasons of data collection, the evidence that yield and quality are inextricably linked was weak.

Developing a means of sensing fruit composition at harvest was an important part of the work. With much still to be done to deliver a commercial solution, the project established the potential of near-infrared (NIR) sensing as an on-the-go tool.


Much decision making in wine production systems is governed by an assumed interaction between grape yield and quality – the yield:quality nexus - which can be summarised as an inference that higher yields mean lower quality. Interestingly, it is almost impossible to find definitions for ‘higher’ or ‘lower’ in reference to either yield or quality, yet this assumed interaction may lead to consequences for growers who might be paid a lower price for a parcel of fruit simply because its yield is deemed too high, or who might decide to crop thin in the expectation of an improvement in quality which may or may not actually be realised. In collaboration with Taylors Wines and Kingston Estate Wines, this project sought to understand the yield:quality nexus using detailed measurements made in contrasting vineyards in the Clare Valley and at Kingston-on-Murray in the Riverland. With a long history of work in vineyard variability and the knowledge that vineyards may be highly variable production systems, our approach was to use the tools of Precision Viticulture (yield monitoring, remote and proximal canopy sensing, the global positioning system (GPS), etc.) and methods of spatial analysis, to see whether links between fruit yield and quality could be established at the within-vineyard scale. Circumstances beyond the control of either the project team or vineyard management meant that, at both of the sites studied, our analysis was confined to just two seasons of data collection, albeit with underpinning access to several years of yield monitor and or remotely sensed imagery data obtained pre-project. In doing this work, we focussed on the ‘standard’ industry measures of quality – juice pH, titratable acidity, total soluble solids (TSS) and the concentrations of anthocyanins (colour) and total phenolics. The results suggest that the link between yield and quality at the within vineyard scale is, at best, a weak one. Certainly, they suggest that making targeted vineyard management decisions aimed at fruit quality improvement based on measures of yield or vine vigour is something that growers should treat with great caution. Our results are therefore consistent with other recent Wine Australia-funded work which draws a similar conclusion about the merits of manipulating vine balance in pursuit of improved quality (CSP 1202). On the other hand, they are in contrast to previous work conducted in vineyards where contrasting zones of characteristic performance were reflective of a ‘terroir effect’.

Two complications in this work arise in relation to the mode and timing of fruit quality assessment. First, while the concentrations of anthocyanins (colour) and total phenolics have been shown to relate to grape and wine quality, they are not widely adopted as quality measures in commercial winemaking. On the other hand, juice pH, titratable acidity and total soluble solids, which are used for this purpose, are much more indicators of fruit maturity, and whilst maturity is certainly a factor that interacts with fruit composition, whether these measures can be regarded as robust indicators of ‘quality’ is a moot point. This is important because we observed that measures of both components of yield and of fruit composition tended to reach a peak some time before the date of harvest, and to then either stabilise or decline. Thus, the time of harvest, which is generally the time at which both yield and quality are assessed for commercial purposes (i.e. payment to growers, allocation of fruit to target products streams, etc.) is not necessarily the most appropriate time to make measurements aimed at understanding the yield:quality nexus. As such, the date of harvest is somewhat arbitrary in the context of grape physiology or phenology. Resolution of this issue by, for example, making measurements at a standard maturity (i.e. fixed value of TSS) is not feasible at the within-vineyard scale given that TSS, and therefore the maturity status of the crop, can be markedly spatially variable. Overall, our work suggests that additional and on-going effort is needed, aimed at identification of robust and objective measures of fruit quality and a means of normalising these to the maturity of the crop at the time of measurement.

The aforementioned work relied on the collection of samples from a small number of ‘target vines’ and attempts to relate analysis of these samples to measures of vine vigour and yield collected at a much higher spatial intensity using yield mapping and remote and proximal canopy sensing. A key secondary component of the project therefore was to explore how information on fruit quality or composition might be obtained at a similar spatial resolution and intensity to the data on vine vigour or yield. Initial attempts to do this (in collaboration with Force-A, Paris, France) using a fluorescence sensor were not as successful as previous work suggested it might be. Instead, we sought to prove the concept of on-the-go sensing of fruit composition during harvest using an NIR sensor mounted above the yield monitor on the discharge chute of a harvester. The results obtained suggest that this technology has much promise, although it needs considerable further development, a key part of which would be the generation of calibrations between NIR spectra and objective measures of fruit quality (see above). Aside from promoting more robust understanding of yield:quality interactions (including whether these exist), being able to map fruit composition at high spatial resolution could prove highly commercially significant. Within-vineyard variation in yield and vine vigour has been shown to have patterns of variation that are temporally stable from year to year; that is, the lower and higher yielding parts of vineyards are generally always lower or higher yielding in the absence of targeted manipulation. If high resolution sensing could be used to demonstrate that certain parts of vineyards inherently produce fruit of either lower or higher quality, or are inherently suited to particular styles of wines, the commercial implications could be profound. Selective harvesting has already been shown to be potentially highly profitable, but an ability to robustly plan such a strategy well in advance of harvest would deliver significant benefit to both grapegrowers and winemakers. Of course, the data provided by such a sensor would, in concert with yield monitoring and remote sensing, promote better understanding of the yield:quality nexus on a site-specific basis.