Vineyard Ecosystems: Optimising the economics of grapevine removal response to leafroll virus

Vaughn Bell¹ and Alistair Hall²

The New Zealand Institute for Plant & Food Research Limited, Havelock North¹ and Palmerston North²


Grapevine leafroll virus is an economically important viral disease affecting many New Zealand vineyards. Among infected vines, the negative effects include reduced yield, reduced berry soluble solids and increased acidity, and ultimately, lower wine quality.

In white berry cultivars, virus-infected vines have no obvious foliar symptoms so visually distinguishing between infected and healthy vines is impractical. On the other hand, for red berry cultivars, virus-infected vines undergo distinctive foliar changes late in the growing season: dark red downward curling leaves with green veins (Figure 1). When vineyard personnel are trained to recognise these visual changes, the recommendation is to tag these vines for later removal (roguing).

Figure 1. Foliar symptoms of grapevine leafroll virus on a Merlot vine in Hawke’s Bay, April 2017.

Assuming all vines are sourced from nurseries accredited to the Grafted Grapevine Standard, spread of leafroll virus is most likely to be the result of mealybugs feeding on pre-existing infected vines within the vineyard before moving to healthy vines. Feeding on the new vines leads to virus transmission: the more mealybugs in the vines, the faster the virus will spread. Thus, if effective virus management is to be achieved and maintained, mealybug numbers must remain low (e.g. no more than 10 per 100 vine leaves inspected).

The New Zealand Winegrowers Virus Elimination study (2009 to 2015) found that in vineyards where mealybug numbers were consistently low, the incidence of leafroll virus quickly reduced to less than 1% per annum, where it was sustained until monitoring concluded. In these vineyards, effective mealybug management was attributed to adopting insecticide best practice, and/or the influence of biological control, whereby beneficial insects naturally maintained mealybugs at low population densities.

In several other study vineyards, however, it was common in most years to find 30 or more mealybugs per 100 vine leaves inspected, a result often attributed to poor adherence to insecticide labels. Under these circumstances, 25 to 75 (or 1 to 3% incidence) new virus-infected vines were recorded per hectare per year.

While all vines showing foliar symptoms of leafroll were rogued every year in every Virus Elimination study vineyard, there was some concern the response was suboptimal when mealybug numbers were high. One of the objectives of the Vineyard Ecosystems programme (jointly funded by New Zealand Winegrowers and the Ministry of Business, Innovation, and Employment) was to further evaluate data collected during the Virus Elimination project to see if roguing symptomatic vines was the optimal response under different virus incidence and mealybug abundance scenarios.

This article reviews the results of the Vineyard Ecosystems programme objective that compared the economic costs and effectiveness of the roguing response with other leafroll management responses. The model developed reveals that the economic annual costs estimated for various leafroll virus and mealybug scenarios support the Virus Elimination project recommendations: 1) starting to rogue early when virus incidence is low provides optimal conditions to achieve quick and effective virus management, and 2) low mealybug numbers results in the least loss of vines to the virus.

What did we do?

The results of this research were briefly conveyed by Vaughn Bell at Bragato 2018, with greater detail contained in the Vineyard Ecosystems Annual Report of the same year (Research Aim 1.2). The report is accessible on the members’ only website. Hence, the full details of the analyses undertaken and all the various assumptions used in the computational modelling are not repeated here. Instead, only contextual information is described.

The model used a theoretical vineyard block of 1 ha planted with 2,500 mature Merlot vines. Within this block, we simulated the spread of leafroll virus based on the virus incidence and patterns of virus spread observed annually in the 13 study vineyards of the Virus Elimination project.

Random simulations of 200 such standard blocks were run over 20 years, using three mealybug population densities as defined by data also from the Virus Elimination project: low (6 mealybugs per 100 vine leaves inspected), median (26 per 100), and high (75 per 100).

The model assumed initial virus incidence was low (0.4 and 5%) and moderate to high (10, 15 and 20%).

Five different management responses to leafroll virus were tested:

  1. ‘Rogue’ all infected vines with foliar symptoms of leafroll virus every year. Replacement virus-free vines were planted annually. This scenario assumed that the assessor was trained and accurate.
  2. ‘Rogue 50%’ resulted in just half the infected vines with foliar symptoms of leafroll virus being removed every year. Replacement virus-free vines were planted annually. This inefficient roguing response was assumed to be either the result of visual symptom identification undertaken too early in the growing season or was due to inadequate assessor training, with just 50% of symptomatic vines being rogued.
  3. ‘Rogue 1+2’ entailed removing all infected vines with foliar symptoms of leafroll virus every year plus removing the within-row immediate neighbours (the ‘first’ vines) either side of it, even if they had no foliar symptoms of leafroll virus. The logic was that one or both first vines might be infected but had not yet shown any foliar symptoms of leafroll. Replacement virus-free vines were planted annually.
  4. ‘Rogue 1+2 50%’ resulted in just half the symptomatic plus first vines being rogued. Replacement virus-free vines were planted annually. This second inefficient roguing option was also assumed to be the result of either poor timing for visual symptom identification or poor assessor training.
  5. ‘No action’ was a deliberate decision not to remove any leafroll virus-infected vines during the 20 years of model simulations.

When removing a virus-infected vine, the costs associated with roguing it and then planting and training a single replacement virus-free vine was estimated at $12.50. It was not until 5 years after planting that this vine was assumed to be 100% productive (0% productivity in years 1 and 2, 60% in year 3, 90% in year 4).

Although roguing costs were not incurred in the no-action decision, costs were nonetheless attributed to it. For example, average yield loss from a vine infected for one year was assumed to be 10% (in years 2, 3, and 4, these were assumed to be 20, 40, and 50%, respectively). Moreover, while there was no loss of crop value at harvest up to 12.5% virus incidence, a crop quality penalty of 5, 25, and 50% applied at 12.5-25%, 26-50%, and 51-100% incidence, respectively.

For each of the 15,000 simulations (3 levels of mealybug density x 5 initial percent virus incidence x 5 management responses x 200 simulations each), we recorded the numbers of virus-infected vines per year, and the numbers of vines removed and replaced per year. From this and the cost information, the average annual costs plus lost income per hectare over 20 years was calculated for each management response.

What did we find?

The 2018 Vineyard Ecosystems report to New Zealand Winegrowers outlined comprehensive results using all combinations of initial virus incidence and the three mealybug infestations. Rather than repeat these results, we instead highlight selected costs and benefits (or disadvantages) of the different management responses. In doing so, we focus on initial virus incidence of 0.4 and 20% (the two extreme scenarios tested), and review how variable mealybug abundance influenced average annual costs plus lost income per hectare over the 20 years tested.

The results of the model were unequivocal: roguing symptomatic vines was the optimal response to leafroll virus. Specifically, owners achieved the best results when roguing started at an initial low virus incidence (0.4%) and when mealybug abundance was low (6 per 100 vine leaves). While the virus was not entirely eliminated from the block, roguing and vine replacement averaged just five vines per year (0.2%) over 20 years. Average annual costs plus lost income was an estimated $131 per hectare (Table 1). Even at an initial virus incidence of 20%, roguing was viable when mealybug populations were low, but average annual costs plus lost income increased to $917 per hectare. Nonetheless, after removing the initial 500 virus-infected vines (20%) in year 1, and replacing them with high-health vines, average annual virus incidence never exceeded 1% from years 2 to 20.

By contrast, at an initial 0.4% virus incidence and at 20%, the feasibility of the roguing response reduced substantially when mealybug populations were median and high. For example, high numbers of mealybugs resulted in average annual costs plus lost income increasing precipitously to $4,012 and $5,457 per hectare at 0.4 and 20% initial virus incidence, respectively (Table 1).

For the other management responses tested – rogue 1+2, both inefficient roguing responses, and no-action – the influence of leafroll virus was relatively benign when both the initial virus incidence and mealybug numbers were low. Under these conditions, annual virus incidence did not exceed 1% over the 20 years modelled, except for the no-action decision, where an estimated 150 vines (or 6%) were virus-infected by year 20. For these responses, average annual costs plus lost income ranged from $145 to $332 per hectare (Table 1).

The implications of leafroll virus were more severe at 26 and 75 mealybugs per 100 vine leaves. Depending on mealybug infestations, annual virus incidence ranged from 4-20% for rogue 1+2 and the two inefficient responses. For the no-action decision at an initial virus incidence of 0.4%, 26 mealybugs per 100 vine leaves saw an estimated 50% of vines infected at year 20; at 20% initial virus incidence and 75 mealybugs, it was just 8 years before 90% of vines were infected. Not only was virus management negatively affected, the economic implications were also severe, with average annual costs plus lost income ranging from $7,400 to $10,500 per hectare (Table 1).

Table 1. The estimated average annual costs plus lost income (NZ dollar values) in a 1 ha Merlot block as measured over 20 years using the five different management responses to grapevine leafroll virus at three estimates of mealybug presence in the vine canopy (6, 26, & 75 per 100 vine leaves inspected), and when 0.4 and 20% of random vines were initially virus-infected.

What does this mean for the sector?

The results of this study fully supported the roguing recommendation from the Virus Elimination project (see Leafroll 3 virus and how to manage it, published by New Zealand Winegrowers, 2015). Specifically, the recommendation was to visually identify and tag infected vines in autumn so they could be rogued in winter. In doing so, owners reduced sources of leafroll infection so mealybugs had less opportunity to encounter virus-infected vines and less chance of transmitting the virus to neighbouring vines. Therefore, the emphasis was on roguing only those vines with foliar symptoms of virus, and none of the neighbouring vines.

For rogue 1+2, none of the virus incidence and mealybug abundance scenarios tested in the model was more effective at managing leafroll virus than roguing symptomatic vines only. Furthermore, rogue 1+2 was less cost-effective than roguing, particularly where mealybug numbers were either median or high. Despite a lack of support for the rogue 1+2, it is a response some owners may adopt in future as they consider tensions between the particular circumstances posed by leafroll virus in their vineyards and their own unique business imperatives.

For both inefficient roguing responses, 1+2 and 1+2 50%, neither was a better option for managing leafroll virus than roguing symptomatic vines. The assumption leading to these inefficiencies was that visual symptom identification was undertaken too early in the growing season and/or inadequate assessor training, either one of which meant just 50% of virus-infected (or symptomless first) vines were rogued in any given year. Therefore, as more vines succumbed to the negative influences of this disease over time, annual virus incidence was consistently higher for both inefficient responses relative to roguing. Ultimately, with more and more vines negatively affected by leafroll virus where roguing was inefficient, estimated average annual costs plus lost income also increased relative to one where roguing was efficient and mealybug management was effective.

In accepting there will be valid reasons why an owner might defer roguing indefinitely, the implications of a no-action decision was especially evident when there were low and median mealybug numbers. When mealybug numbers were high, the financial implications were relatively constant across all responses, except for the roguing response. Hence, if a roguing response is deferred because a block of virus-infected vines are to be removed in future and virus-free vines planted, then effective mealybug management leading up to redevelopment, is critical. Why? Remnant vine roots are a known reservoir of the virus. If they also sustain subterranean populations of mealybugs, there is a risk they will relocate to newly planted vines, with the resumption of mealybug feeding being a predicted virus transmission pathway. Mitigating this potential risk should include adopting mealybug insecticide best practice in the months (and possibly years) before the block is redeveloped.

Final comments

The model described relied on real data collected from New Zealand vineyards. It reflected multi-year measures of virus incidence, patterns of virus spread, and pre-harvest counts of mealybugs on vine leaves. The model supported the recommendations of the Virus Elimination project: starting to rogue early when virus incidence was low provided the optimal conditions to achieve quick and effective virus management. Importantly, what this research offers the sector is a financial context to the virus/mealybug relationship that quantifies estimated average annual costs plus lost income for different management responses at variable mealybug densities. And, as has been the refrain of New Zealand Winegrowers for some years, the fewer mealybugs in the vines, the better. In meeting this condition, owners have an opportunity to manage leafroll virus effectively whilst moderating the financial implications.


For red berry cultivars affected by leafroll virus, the recommendation is to remove (rogue) vines with foliar symptoms, and to maintain low numbers of mealybugs to reduce the risk of virus spread. New data analysis described in this article assessed if particular virus incidence and mealybug abundance conditions might identify economically optimal management practices. In a theoretical 1 ha block planted in 2,500 mature Merlot vines, variable combinations of initial virus incidence (range: 0.4 to 20%) and mealybug abundance (6 to 75 per 100 vine leaves) data were analysed to see if roguing optimised virus control and minimised costs. Roguing was compared with removing the symptomatic vine plus its immediate within-row neighbours, inefficient roguing, and a no action decision. The results reaffirmed that roguing symptomatic vines early when incidence and mealybug numbers were low was the best response. Thus, when compared with the other responses, roguing meant the lowest loss of vines, the least need to plant replacement vines, and the lowest estimated average annual costs plus lost income. The results highlighted that good mealybug control was critical to improved vineyard longevity. Without it, the virus imposed dire financial implications, even for roguing, the preferred response.

This article first appeared in the October/November 2019 issue of New Zealand Winegrower magazine