2. Discussion

In New Zealand, field conditions, natural product (NP) fungicides, NP1 (based on anhydrous milk fat) and NP2 (based on soybean oil), provided the effective control of Botrytis bunch rot on three green winegrape varieties (Chardonnay, Sauvignon Blanc, Riesling), and powdery mildew control on Chardonnay and Riesling, in two different geographical regions (Hawke’s Bay, Gisborne) over multiple seasons and under a full range of disease pressures. Disease control was usually as efficacious as the synthetic fungicide treatment(s) and total yields in NP-treated crops were normally equivalent to those obtained from the synthetic fungicide treatments. Microvinification of grape berries from the NP- and full fungicide treatments produced wines were evaluated by a trained sensory panel and subjected to chemical residue analysis. Results showed that the NP wines were residue-free and that there were no adverse effects on sensory characteristics. Taken together, the data indicate that disease control of two key pathogens in vineyards is possible using fungicides based on these natural products.

The use of a plethora of naturally sourced products for plant disease control has been practised since ancient times and is often utilised by native peoples ^[17]. There is an abundance of global research on the use NPs under controlled laboratory conditions ^[17], but relatively fewer products have been developed for commercial use in variable field situations ^[15], often because of inadequate field research, problems associated with inconsistent efficacy, undesirable side effects such as growth of unwanted organisms, or difficulties with formulation ^[25]. Some of the most successful and widely used NPs that are commercially available include: sulphur, which remains the mainstay of powdery mildew control ^[26]; harpin proteins, which elicit plant defences ^[14]; chitosan, which is both an elicitor and directly antifungal ^[16]; giant knotweed extract (Regalia^®) which is effective against PM ^[18]; seaweeds, although seaweed extracts are often used for growth enhancement rather than for disease control, per se ^[27]. Although plant oils are used frequently in horticulture as adjuvants ^[28], they are rarely the active ingredient in biofungicides, the exception being a tea tree oil extract (Timorex Gold^®), which is effective against PM ^[29]. This study reports the novel use of an animal fat (NP1) and a plant oil (NP2) for the control of both PM and Botrytis bunch rot in a range of field trials, with data obtained on product efficacy over numerous seasons on different winegrape varieties, vineyards and geographic locations, and under different disease pressures.

Both winegrape variety and geographical region appear to influence treatment efficacy against Botrytis. The NPs gave the most effective control of Botrytis rots on Chardonnay (ranging from 63–97% efficacy, as measured at harvest over four seasons), and much more variable control on Sauvignon Blanc (0–96% efficacy, measured over three seasons) and Riesling berries (46–58% efficacy, one season of data). Inter-variety variation existed even when trials were carried out at the same time, in the same vineyard, and using the same spray programme (e.g., Figure 2c versus Figure 3a), hence the grapevines would have been subjected to similar climatic conditions and vineyard management factors. Possible reasons for variations within the same vineyard, apart from inherent differences in resistance between the different varieties associated with microbiome, bunch architecture and berry biochemistry, might include the different ages of the vines and rootstocks (Table 3), along with possible variation in microclimate and soil type and microbiome within a particular vineyard. Whilst only results for green varieties are presented here, Calvo-Garrido et al. ^[30] independently tested an integrated disease management (IPM) programme containing the same products as our BZ/NP2/AZ treatment on red grape varieties in France and found it to be the most effective for Botrytis control out of all the biological programmes tested. In general, Botrytis disease control was much more effective in Hawke’s Bay and Gisborne than in the dryer climate of Marlborough, which is less conducive to Botrytis infection (Figure 2 and Figure 3). Regional differences may also be attributable to subtle differences in B. cinerea ecology and epidemiology ^[5], climate, soils, rootstock, and vine age, as well the effect of vineyard management practices. For example, Marlborough growers normally maintain denser Sauvignon Blanc canopies (≤40% bunch exposure) than Hawke’s Bay growers (typically ≥70% bunch exposure) to create the grassy flavours that are preferred in wines from this region. The effect of canopy density on NP spray penetration and Botrytis disease control efficacy was investigated in the current study. Spray penetration of the non-systemic NPs was much more effective in the low-density canopy (LCD, approximately 80% bunch exposure) versus the standard density canopy (SCD, approximately 26% bunch exposure) at both sites. Disease control efficacy results in Marlborough were inconclusive, due to very low levels of Botrytis, but the Hawke’s Bay results indicated that a more open canopy helps to reduce Botrytis in its own right, and that canopy density also has a significant effect on the efficacy of the NPs. An LCD allows for greater air circulation around the bunches, thus lowering humidity, and creating a micro-environment that is less favourable to Botrytis ^[31]. However, the most likely reason for NP2 working better in an LCD is that it enables greater spray penetration to internal bunches (Figure 3 and Figure 4), which is important to NP2 efficacy, because it has contact-only activity ^[23][32]. However, yield tended to be significantly lower in the LCD, possibly because insufficient leaf material remained to create adequate photosynthates for berry development. Further work is, therefore, required to determine the optimal canopy density for excellent disease control without adverse effects on yield. Another possibility might be plucking just one side of the canopy, as was successfully used in the study by Calvo-Garrido et al. ^[30]

Figure 1. Percent incidence of Botrytis latent infections (latents) on immature berries from Chardonnay and Sauvignon Blanc grape varieties, following freezing and incubation in high humidity chambers. The efficacy (E) of each spray programme for Botrytis control (shown in red text) was calculated using the formula, E = ((U − T)/U) × 100, where T is the percentage of latents in the test treatment and U is the percentage disease in the Nil botryticide control. A negative E value indicates that disease is higher in the treatment than in Nil botryticide control. Treatments are described in full in Tables S1 and S2, but briefly comprised: Nil botryticides (negative control), where 1–3 sprays of a powdery mildew-specific fungicide were applied and no botryticides; full fungicide (positive control), where a mixture of fungicide products were applied at 7- to 21-day intervals from late-November through to mid-March to provide season long disease control; NP1 and NP2, which were applied 1–4× during the early- and mid-season of phenological development, i.e., between post-bloom and post-bunch closure. For the NP treatments, BOTRY-Zen® (BZ, 4 kg/ha formulation of the biocontrol agent, Ulocladium oudemansii) was also applied during the remainder of the early-season, and BCA-L1, an experimental formulation of Aureobasidium pullulans (2 × 10⁷ spores/mL) was applied during the late season (mid-February to mid-March) to provide season-long disease control using only biopesticide products. Switch® fungicide (800 g/ha) and ARMOUR-Zen (AZ, 5 L/ha formulation of chitosan) were sometimes used as alternative mid- and late-season treatment comparisons. Each lettered graph represents a separate experiment, set up in a randomised block design. Different ascending numbering systems are used for Chardonnay vs. Sauvignon Blanc experimental seasons to indicate that trials for both varieties were performed over consecutive seasons, but that season 1 in Chardonnay does not necessarily correspond to the same calendar year as Season i in Sauvignon Blanc. On each graph, error bars indicate standard errors, and the boxed values provide the probabilities obtained from analysis of variance (ANOVA). Different letters indicate statistically significant differences from pairwise likelihood ratio tests but are only presented when the ANOVA p ≤ 0.05. The concentration of the active ingredient in the NP formulation is indicated in parentheses for each specific experiment. (a) Chardonnay, Season 2, Hawke’s Bay site, with assessments at two times—pre-bunch closure and post-bunch closure, NP1 (7 g/L) and NP (15 g/L); (b) Chardonnay, Season 4, with two assessment sites—Hawke’s Bay and Gisborne, NP2 (5 g/L); (c) Sauvignon Blanc, Season i, Hawke’s Bay site, NP2 (5 g/L).

Figure 2. Percentage crop loss resulting from Botrytis bunch rot infections of Chardonnay grapes, as assessed at harvest in April, over four successive seasons. Crop loss was calculated as (%bunch severity × %bunch incidence)/100 for 50 randomly harvested bunches per plot. The efficacy calculation is outlined in Figure 1. Treatments are described in full in Tables S1 and S2, and briefly in Figure 1. In Season 1 only (Figure 2a only), there was an unsprayed treatment, and the Kumulus^® DF (3 g/L), Kocide^® 2000 DF and NP1 (7 g/L) treatments were applied right throughout the growing season (from mid-November until mid-March) instead of just between post-bloom and post-bunch closure (Figure 2b–d). Each graph represents data from a separate season, with error bars indicating standard errors, boxed values showing probabilities obtained from ANOVA, and different letters indicating statistically significant differences from pairwise likelihood ratio tests, which are only presented when the ANOVA p ≤ 0.05. The concentration of the active ingredient in the NP formulation is indicated in parentheses for each specific experiment. (a) Season 1, Hawke’s Bay site, NP1 (7 g/L); (b) Season 2, Hawke’s Bay site, NP1 (7 g/L) and NP (15 g/L); (c) Season 3, Hawke’s Bay site, NP2 (5 g/L); (d) Season 4, Hawke’s Bay and Gisborne sites, NP2 (5 g/L).

Figure 3. Percentage crop loss resulting from Botrytis bunch rot infections of Sauvignon Blanc (SB) and Riesling grape varieties, as assessed at harvest in April. Data were collected over 3 successive seasons for SB and one season for Riesling. Crop loss was calculated as described in Figure 2, and the efficacy calculation is outlined in Figure 1. Treatments are described in full in Tables S1 and S2, and briefly in Figure 1, except for the addition of treatments applied under different canopy densities, where leaf plucking and shoot thinning were used to create standard canopy densities (SCD), with bunch exposures of ca. 26%, and low canopy densities (LCD), with bunch exposures of ca. 80%. Each graph represents data from a separate trial, set up in a randomised block resign, with error bars indicating standard errors, boxed values showing probabilities obtained from ANOVA, and different letters indicating statistically significant differences from pairwise likelihood ratio tests, which are only presented when the ANOVA p ≤ 0.05. The concentration of the active ingredient in the NP formulation is indicated in parentheses for each specific experiment. (a) SB, Season i, Hawke’s Bay site, NP2 (5 g/L); (b) SB, Season ii, Hawke’s Bay and Marlborough sites, NP1 (5 g/L) and NP2 (5 g/L); (c) Season iii, Hawke’s Bay and Marlborough sites, new formulation of NP2 (30 mL/L); (d) Riesling, Hawke’s Bay, NP1 (7 g/L) and NP2 (5, 10 and 20 g/L).

An important aspect of the efficacy of the NPs against Botrytis is that they can inhibit latent infections, which generally occur around the mid-season (i.e., during the summer months) of the phenological stages of grape development. Switch® is the main (systemic) fungicide used to provide control against Botrytis latent infections, but its use is restricted to two applications/season and the final application date should be no less than 60 days before harvest to prevent maximum residue limits being exceeded in export wines and to the minimise the risk of developing fungicide resistance ^[33]. Use of the NPs offers a viable residue-free alternative for latent infection control. However, it must be noted that the NP1 and NP2 residue studies were carried out on wines produced from Chardonnay grapes in the Hawke’s Bay region. We recommend that further residue studies are performed on wines produced from different varieties and regions to confirm this result. Another advantage of the NPs is that pathogens do not tend to develop resistance to agricultural sprays containing lipids ^[19][20]. Given that NPs, like other lipids, appear to have contact-only action via a direct physical effect on the pathogen ^[23][32], control of latent infections is surprising and may suggest an, as yet undiscovered, mode of action, such as the direct induction of plant defences, or possibly indirect induction of defences via generation of fungal elicitors (compounds that activate plant defences) following the disruption of pathogen cells. This remains to be investigated.

The efficacy of the control of PM on berries was only recorded on Chardonnay (88–98% control, two seasons of data) and Riesling (65–77% control, one season) in the Hawke’s Bay. PM was either not measured or not observed in Sauvignon Blanc in the current study. However, approximately 40% control of PM on Sauvignon Blanc grapes was achieved with use of NP2 in an LCD in Marlborough in another study (Wurms, PFR, pers. comm.). The reasons for varietal and geographical variations are likely to be the same as discussed for Botrytis, but there is an additional complication. The teleomorphic stage of Eryspihe necator, a key source of inoculum for the next season, is becoming much more prevalent in commercial New Zealand vineyards ^[11] and is directly associated with increased disease severity ^[12]. At this stage, the relative efficacy of the NPs against the teleomorphic stage of PM is unknown, and this needs further research. More information is also needed on population structure and the distribution of mating types in New Zealand.

The BZ/NP2/AZ programme represents a successful season-long biopesticide programme using products that are commercially available in New Zealand and comprising one biocontrol agent (BCA) and two different natural products, all of which have different modes of action. The combined use of the three products is likely to increase the durability of the individual components compared with the durability if they were used alone. The efficacy of this treatment programme has recently been confirmed in France ^[30]. The fungus (Ulocladium oudemansii) in BOTRY-Zen works by outcompeting B. cinerea for the colonisation of floral tissue and bunch trash before bunch closure, thus preventing B. cinerea infections from becoming established within the bunches ^[34]. NP2, which has been registered in New Zealand under the trade name MIDI-Zen^®, directly affects PM and Botrytis pathogens by causing conidiophores to collapse and conidia and hyphae to wither and extrude cellular contents. ARMOUR-Zen^® contains chitosan, which is known to be both directly fungicidal and to induce plant defence mechanisms ^[35][36]. The BZ/NP2/AZ treatment is also known to be residue-free, which represents a major advantage of using NP2 as an alternative to Switch, and does not adversely affect wine sensory quality, or yield. However, it can cause a delay in °Brix when compared to BZ/Switch/AZ (Figure 4b).

Figure 4. Fruit maturity (°Brix) in Sauvignon Blanc berries, assessed at harvest in April. Treatments are described in full in Tables S1 and S2, and briefly in Figure 1 and Figure 3. Error bars indicate standard errors, boxed values show ANOVA probabilities, and different letters indicate statistically significant differences from LSD (p ≤ 0.05), only presented when the ANOVA p ≤ 0.05. The concentration of the active ingredient in the NP formulation is indicated in parentheses for each specific experiment. (a) Season i, Hawke’s Bay site, NP2 (5 g/L); (b) Season ii, Hawke’s Bay and Marlborough sites, NP1 (5 g/L) and NP2 (5 g/L); (c) Season iii, Hawke’s Bay site, new formulation of NP2 (30 mL/L), °Brix were measured on two dates—the normal commercial harvest date and the actual harvest date, as berries were left on the vines for another week to allow for additional Botrytis development; (d) Season iii, Marlborough site, new formulation of NP2 (30 mL/L), Brix° measured on two dates. Data for the Nil botryticide treatments on the actual harvest date are missing in Figure 4c,d because the fruit were accidentally harvested by the commercial pickers.

There is often a delicate balance between disease control efficacy and higher doses of oils/fats becoming phytotoxic to the plant, as fats and oils are often associated with chlorosis and necrosis of plant tissue ^{[37][38][39][40]}, so this may explain why higher doses of NPs are not always beneficial (Figure 3d). Combining the two NPs together has been shown in other studies to have a complementary effect on disease control, making it possible to reduce the concentrations of the active ingredients (a.i.) of each ^[25], possibly because the NPs appear to have different physical modes of action, with NP1 causing withering and distortion/ridging of fungal structures, whilst NP2 application leads to the extrusion of cellular contents ^[23]. However, in this study, using alternating NP1 and NP2 at different times in the spray programme, instead of the use of just one NP product, seemed to exacerbate Botrytis latent infections (Figure 1a). Significant phytotoxic effects associated with use of the NPs, such as leaf burning, were not observed at any time in the current study.

As with all agricultural products, the successful deployment of NPs depends on understanding their potential drawbacks. Accurate spray targeting, particularly under dense canopies, remains the most critical factor in the success of these products, because the NPs appear to have contact-only action. The effect of NPs on fruit maturity also needs to be considered, as °Brix (which is used an indicator of fruit maturity in grape berries) was often 1–1.5° lower at harvest time. Application of lipids over an extended period of time can be associated with a delay in the onset of véraison ^[41][42], hence this problem was minimised by instead restricting their use to the mid-season as part of an IPM programme. Figure 4 also shows that °Brix continues to rise normally over time in the NP treatments, so, if necessary, harvest can be delayed by 1–2 weeks to achieve optimal °Brix. However, further research is also needed to determine whether the reduction in acids in the berries is also delayed. If acid reduction is not delayed, then the fruit would need to be picked at a lower °Brix, in which case NP-use might be another tool to produce lower alcohol wines harvested at lower °Brix.

Overall, this research has shown that the effective control of PM and Botrytis can be achieved in winegrape vineyards using natural lipid-based products, without any adverse effects on yields, and offering the additional advantage of being residue free, not easily overcome by pathogen resistance, and consisting of ingredients that are generally regarded as safe.