Research Note Flower Debris Removal Delays Grape Bunch Rot Epidemic

Research Note Flower Debris Removal Delays Grape Bunch Rot Epidemic Daniel Molitor, 1,2 * Lucien Hoffmann, 1 and Marco Beyer 1 Abstract: Field trials investigating the impact of removing flower debris (necrotic flower caps, filaments, anthers, aborted unfertilized ovaries, and aborted berries), which usually remain partly attached to grape clusters after bloom, on the epidemic of grape bunch rot caused by Botrytis cinerea were conducted in the white grapevine (Vitis vinifera L.) cv. (four experiments) and cv. Riesling (two experiments) in Luxembourg in the period of 2011 to 2014. Grape clusters remained untreated (control), or (i) flower debris was removed from the clusters (brush), or (ii) the clusters were treated with a botryticide (active ingredient fenhexamid; botryticide), or (iii) the clusters were brushed with a brush soaked in a botryticide suspension (botryticide-soaked brush) (all three treatments conducted at growth stage BBCH 73). On average of all six trials, bunch rot epidemics (day of the year reaching 5% disease severity) were significantly (p < 0.045) delayed compared to the untreated control by 3.7 (brush), 4.3 (botryticide), or 5.7 days (botryticide-soaked bruch), respectively. No significant differences in the delay of the epidemic were observed among the three treatments. Consequently, removal of flower debris might contribute to a reduction in or partial replacement of pesticide use in viticulture. Efficient technical solutions to automatically remove flower debris need to be developed. Key words: Botrytis cinerea, crop cultural practice, inoculum potential, pesticide reduction, pesticide replacement, Vitis vinifera Grape bunch rot caused by Botrytis cinerea Pers.:Fr. (teleomorph: Botryotinia fuckeliana [de Bary] Whetzel) is one of the major fungal diseases on grapevine (Vitis vinifera L.) worldwide. Bunch rot causes yield losses and reduces wine quality by promoting off-flavors, unstable color, wine damage through oxidation processes, premature aging, and difficulties in clarification (Ribéreau-Gayon 1983). In the past, bunch rot control strategies were mainly based on routine applications of fungicides (Shtienberg 2007) with specific activity against B. cinerea (botryticides). Since pesticide use is to be reduced in integrated pest management and chemical measures to control bunch rot are under more critical scrutiny (Shtienberg 2007), nonchemical alternatives to chemical treatments are gaining increasing importance. Besides an adequate selection of grape varieties and proper canopy management, crop cultural measures resulting in less compact clusters and/or in better sun and wind exposure represent 1 Environmental Research and Innovation (ERIN) Department, LIST Luxembourg Institute of Science and Technology, 41, rue du Brill, L-4422 Belvaux, Luxembourg; and 2 Department of Crop Sciences, Division of Viticulture and Pomology, University of Natural Resources and Life Sciences, Konrad Lorenz Str. 24, A-3430 Tulln, Austria. *Corresponding author (daniel.molitor@list.lu; tel: 00352 47 02 61 436; fax: 00352 47 02 64) Acknowledgments: The authors would like to especially acknowledge N. Baron, M. Schultz, R. Rausch, M. Pallez, M. Behr, R. Mannes, and S. Fischer for their support; Dr. Ruth Walter (DLR Rheinpfalz) for fruitful discussion; V. Peardon for language editing; and the Institut Viti-Vinicole (Remich/Luxembourg) for financial support in the framework of the research project ProVino Pesticide reduction in viticulture. Manuscript submitted March 2015, revised May 2015, accepted Jun 2015 Copyright 2015 by the American Society for Enology and Viticulture. All rights reserved. doi: 10.5344/ajev.2015.15019 efficient nonchemical tools in bunch rot control strategy (e.g., Molitor et al. 2012, 2015). Although applying a single cultural measure may delay the epidemic, this does not guarantee mature and healthy grapes every season. Therefore, especially under the given climatic conditions in Central Europe, efficient bunch rot control programs need to be built on several modules (Evers et al. 2010, Molitor et al. 2011). Young, immature grape berries are highly resistant to B. cinerea between fruit set and veraison (Hill et al. 1981), and most destructive B. cinerea infections commonly affect ripe berries (Shtienberg 2007). In contrast to young grape berries, flower debris consisting of necrotic flower caps, filaments, anthers, and aborted unfertilized ovaries, along with aborted fertilized berries, are easily colonized by B. cinerea because the defense mechanisms that operate in living grape organs are absent in necrotic tissues. Especially in compact clusters, the flower debris often remains attached to the expanding berries for several weeks and then becomes trapped in the interior parts of the clusters after bunch closure. Inside the grape clusters, the trapped flower debris may serve as an inoculum reservoir for infections of healthy berries (Wolf et al. 1997) once these become susceptible after veraison. Infection of ripe grape berries can occur during sufficiently humid periods, and its severity is influenced by the microclimatic conditions within the grape cluster (Savage and Sall 1984). Especially in compact grape clusters, microclimatic conditions are highly conducive for fungal infections because of low air circulation and limited exposure to sunlight (Zoecklein et al. 1992). Wolf et al. (1997) indeed observed that removing floral debris via compressed air or a backpack leaf blower reduced the severity of bunch rot disease. However, neither the effects of flower debris removal on the temporal position of the disease nor the potential of this crop cultural measure to reduce or partly replace the 548

Flower Debris Removal Effect on Grape Bunch Rot 549 use of pesticides in viticulture were investigated by these authors. The grape cultivars and Riesling are two of the most widely grown cultivars in the Luxembourgish grapegrowing region. Because of their frequently compact cluster structure, both cultivars are highly susceptible to bunch rot. In humid and cool grapegrowing regions, where the grape harvest date is more often determined by the progressing development of fungal pathogens on the grapes than by full grape maturity, growers focus on improving grape health to reach full grape maturity as a prerequisite for high-quality wines (Molitor et al. 2012). Hence, the hypotheses for the present investigations were the following: (i) removing flower debris after flowering reduces the B. cinerea inoculum, delays the onset of bunch rot, and thereby prolongs the ripening period, and (ii) flower debris removal may represent a crop cultural measure allowing for a partial replacement of pesticides in viticulture. Materials and Methods Vineyard site and experimental design. Field trials were carried out between 2011 and 2014 in the experimental vineyards of the Institut Viti-Vinicole in Remich, Luxembourg (lat. 49 32 40 N; long. 6 21 14 E) with the white V. vinifera L. cv. and cv. Riesling. Both vineyards were planted in the year 2000, and the vines, grafted onto SO4 rootstocks, were trained to a vertical shoot-positioning system. The space per plant was 2.4 m 2 (2 m between rows and 1.2 m between vines). The experimental vineyards were previously described (Molitor et al. 2015). Regular fungicide applications (at 10- to 12-day intervals) against Plasmopara viticola (Berk. and M.A. Curtis) Berl. and De Toni and Erysiphe necator Schwein. were carried out in all seasons (active ingredients: azoxystrobin, benthiavalicarb-isopropyl, copper hydroxide, cyazofamid, dithianone, fluopicolide, folpet, fosetyl-al, mancozeb, penconazole, phosphoric acid, potassium hydrogen carbonate, spiroxamine, sulfur, trifloxystrobin, and zoxamide). No fungicides with known specific activity against B. cinerea were applied. Manual cluster zone leaf-removal was applied on the southeast-exposed sides of the rows between the end of flowering (BBCH stage 69; Lorenz et al. 1995) and fruit set (BBCH 71). Each of the six experiments (four experiments on Pinot gris in 2011, 2012, 2013, and 2014; two experiments on Riesling in 2011 and 2012) was performed in a randomized complete block design with four replicates of eight vines per plot. Treatments were the same in all experiments and defined as follows: 1, untreated control; 2, brush (manual flower debris removal with a brush); 3, botryticide (botryticide application); and 4, botryticide-soaked brush (clusters were brushed with a brush soaked in a botryticide suspension). Treatments 2 to 4 were applied at growth stage BBCH 73 according to Lorenz et al. (1995). All single clusters in each plot were treated on application dates given in Table 1. In treatment 2, clusters were brushed individually using a flat-hair brush until flower debris was removed (Figure 1B). The workload amounted to ~65 hr per ha. Botryticide applications (Teldor; active ingredient 500 g/kg fenhexamid with a treatment concentration of 0.1%) in treatment 3 were carried out manually with a backpack sprayer (Solo Akku 416; Solo Kleinmotoren GmbH, Sindelfingen, Germany) equipped with injector nozzles (Albuz AVI 80; Agrotop Spray Technology, Obertraubling, Germany). No wetting agents or surfactants were used. Clusters were treated from both sides until runoff occurred. 780 L/ha of pesticide-water mix was applied, resulting in a fenhexamid dose of 0.39 kg/ha. Figure 1 Grape cluster in treatment 2 before (A) and after (B) cluster brushing. Pictures were taken in the experimental vineyard on 18 June 2014. Treatment Table 1 Treatments and application dates in the six cultivar year combinations. 2011 Riesling 2011 2012 Riesling 2012 2013 2014 1 (Untreated control) na a na na na na na 2 (Brush) 15 June 15 June 03 July 03 July 12 July 18 June 3 (Botryticide) 21 June 21 June 04 July 04 July 12 July 18 June 4 (Botryticide-soaked brush) 22 June 22 June 04 July 04 July 12 July 18 June a na: not applicable.

550 Molitor et al. In treatment 4, clusters were brushed individually with the flat brush as described for treatment 2. Between single clusters, the brush was dipped into the botryticide suspension (treatment concentration as for treatment 3 was 0.1%, with no wetting agents or surfactants). Sixty-five L/ha pesticide-water mix was used for cluster brushing in treatment 4, resulting in an application of 0.03 kg/ha fenhexamid. The workload was comparable to that in treatment 2. Assessment of B. cinerea disease progress. The progress of the B. cinerea disease severity was followed at weekly to biweekly intervals by visual assessments of 100 randomly selected clusters per plot (50 on each side of the row). Disease severity was assessed according to the European and Mediterranean Plant Protection Organization guideline PP1/17. Grape clusters were classified according to visual symptoms into seven classes: 0%, 1 to 5%, 6 to 10%, 11 to 25%, 26 to 50%, 51 to 75%, and 76 to 100% disease severity. Average disease severities were calculated as described by Molitor et al. (2015). Efficacy levels of the different treatments at the final assessment were calculated as 100% minus the disease severity in the respective treatment relative to that in the untreated control. To describe the temporal progress of the disease severity, the average values were plotted against the assessment date (expressed as day of the year [DOY]). Disease progress curves were fitted to this data using the sigmoidal equation (1), Eq. 1 where y is the disease severity, x corresponds to the assessment date (DOY), x 0 is the inflection point of the curve (when a disease severity of 50% is reached), and b is the slope factor of the curve. Solving this equation for x allowed to estimate the time point at which a specific disease severity level was reached. In the present study, the x 5% values (the DOY when a disease severity level of 5% was reached) were selected following Beresford et al. (2006) and Evers et al. (2010) to compare the effects of the different treatments on the timing of the disease. The delay of the time point at which disease severity reached 5% compared with the untreated control was calculated by subtracting the x 5% values of the untreated control from the x 5% values of the treatments 2 to 4. Statistical analysis. All statistical analyses were performed with the software package SPSS version 19 (Chicago, IL). Treatment effects on disease severities at every assessment date were analyzed by one-way ANOVA. In case null hypothesis were rejected (at a p 0.05), pairwise multiple comparisons according to Duncan (1955) were performed to determine the effects of the individual treatments. To check for statistically significant differences in the time point when a 5% disease severity was reached, x 5% values (temporally relative to x 5% values of the control treatment 1 in the same cultivar year combination) were compared pairwise (α 0.05) after independent-samples Kruskal- Wallis tests. Results and Discussion The final (immediately prior to harvest) disease severities in the untreated control ranged from 10.4 to 19.3%, except for 2012 in which the final disease severity in the Figure 2 Disease severity progression of bunch rot caused by B. cinerea in cv. (2011 to 2014) and cv. Riesling (2011 and 2012) as a function of the day of the year (DOY). Plot symbols represent the observed disease severities, the dashed lines represent the calculated progress according to Equation 1. Disease severities of treatments in the same cultivar year combination at the same assessment date marked with the same letter did not differ significantly (according to Duncan s multiple range test; α = 5%). If no significant differences were observed, no notations are indicated. Error bars represent the SE.

Flower Debris Removal Effect on Grape Bunch Rot 551 untreated control was 1.7%. In each experiment, the highest final disease severities were observed in the untreated control. Average efficacy levels (average values of the efficacy levels of all six experiments) at the final bunch rot assessment were 24.4% (treatment 2, brush), 28.8% (treatment 3, botryticide), and 38.9% (treatment 4, botryticidesoaked brush). In comparable trials with cv. Riesling in 2010, Walter (2012) observed that, compared with no treatment, removing flower residuals shortly after flowering reduced final disease severity at harvest by 49%. The higher level of efficiency due to flower debris removal in the trials of Walter (2012) compared to present investigations might be caused by the specific environmental conditions in the growing season 2010. In three out of the six present experiments, significant differences between the untreated control and treatment 4 (botryticide-soaked brush) were observed at the final assessment date (Figure 2). Generally lower bunch rot disease severities as a result of senescent floral debris removal in cv. Chardonnay grapes were also observed by Wolf et al. (1997), but these reductions were not always significant. Sigmoidal equations of the type enabled very precise modeling of the disease progress, confirming previous observations (Molitor et al. 2015). The coefficients of determination (R 2 ) ranged from 0.97 to 1 with p values ranging from < 0.0001 to 0.0605 (Figure 2, Table 2). Compared with the untreated control, in all six cultivar year combinations, treatments 2 to 4 delayed the time point at which a disease severity of 5% was reached (Figure 3, Table 2). The average times of delay were 3.7 days for treatment 2 (brush), 4.3 days for treatment 3 (botryticide), and 5.7 days for treatment 4 (botryticide-soaked brush) (Table 2, Figure 3). Under the climatic conditions in the Luxembourgish Moselle valley, the time of harvest is more often determined by declining crop health than by full grape maturity (Molitor et al. 2012). The observation that the removal of flower debris potentially delays a bunch rot epidemic suggests that the removal allows for a prolongation of the ripening period. No consistent difference in the temporal delay of the epidemic due to the treatments between the two cultivars used in this Table 2 Parameters describing the disease progress curves and the temporal delay of the day of the year on which a disease severity of x 5% was reached compared with the untreated control in the six cultivar year combinations. a Cultivar, year, treatment R 2 p value b x 0 x 5% Delay x 5% (d), 2011 1 (Untreated control) 0.9997 0.0106 11.27 279.02 245.8 na b 2 (Brush) 0.9958 0.0414 11.14 282.67 249.9 4.0 3 (Botryticide) 0.9910 0.0605 11.34 286.72 253.3 7.5 4 (Botryticide-soaked brush) 0.9958 0.0411 11.51 292.56 258.7 12.8 Riesling, 2011 1 (Untreated control) 0.9655 0.0027 14.16 291.56 249.9 na 2 (Brush) 0.9829 0.0010 11.92 289.58 254.5 4.6 3 (Botryticide) 0.9731 0.0019 12.95 291.37 253.2 3.4 4 (Botryticide-soaked brush) 0.9566 0.0039 14.14 297.25 255.6 5.8, 2012 1 (Untreated control) 0.9999 <0.0001 6.21 306.12 287.8 na 2 (Brush) 0.9941 0.0030 6.43 308.42 289.5 1.7 3 (Botryticide) 0.9990 0.0005 7.24 313.53 292.2 4.4 4 (Botryticide-soaked brush) 0.9973 0.0013 5.58 306.83 290.4 2.6 Riesling, 2012 1 (Untreated control) 0.9874 0.0063 6.81 303.59 283.5 na 2 (Brush) 0.9847 0.0077 6.95 307.45 287.0 3.5 3 (Botryticide) 0.9972 0.0014 6.56 307.10 287.8 4.3 4 (Botryticide-soaked brush) 0.9807 0.0097 7.39 309.75 288.0 4.4, 2013 1 (Untreated control) 1.0000 <0.0001 7.24 291.38 270.1 na 2 (Brush) 1.0000 <0.0001 6.31 293.08 274.5 4.4 3 (Botryticide) 1.0000 <0.0001 7.12 293.47 272.5 2.4 4 (Botryticide-soaked brush) 1.0000 <0.0001 7.30 297.80 276.3 6.2, 2014 1 (Untreated control) 0.9934 0.0002 11.88 295.86 260.9 na 2 (Brush) 0.9980 <0.0001 11.21 297.85 264.8 4.0 3 (Botryticide) 0.9963 <0.0001 10.53 295.58 264.6 3.7 4 (Botryticide-soaked brush) 0.9945 0.0002 11.04 295.99 263.5 2.6 a The curves were determined according to Equation 1. b: slope factor of the curve; R 2 : coefficient of determination; x 0 : inflection point of the curve; x 5% : day of the year (DOY) when a disease severity of 5% was reached. b na: not applicable.

552 Molitor et al. study were observed (Figure 3). Pairwise comparisons after an independent-samples Kruskal-Wallis test showed that all three treatments statistically significantly delayed the day of the year when a disease severity of 5% was reached, compared with the untreated control. No significant differences in the time point at which disease severity reached 5% were observed among treatments 2, 3, and 4 (Table 3). This shows that flower debris removal at BBCH 73 delays the disease with an efficacy similar to that of botryticide application. Hence, flower debris removal could represent an additional tool for a partial pesticide replacement in terms of crop cultural measures, as is the goal in integrated pest management. Especially in the two tested cultivars and Riesling (which frequently exhibit a compact cluster structure), flower debris often remains attached to expanding berries and is trapped in the interior parts of the clusters after bunch closure. Brushing off flower debris limits this inoculum potential in the interior parts of the clusters. In the present study as well as in the trials of Walter (2012), flower debris was removed manually with brushes. Due to the high costs of manual processing, manual methods are not economically profitable in large-scale grape production. Our experience (data not shown) indicates the airstream speed might be the limiting factor for removing flower debris, which is often strongly attached to the berry clusters. Therefore, efficient technical solutions for flower debris removal need to be developed. Although not statistically significant, treatment 4 (with a botryticide-soaked brush) had a higher average efficacy than the brushing and botryticide treatments alone. The botryticide-soaked brush treatment combines debris removal via brushing with the antifungal effect of the botryticide applied to cluster stems and berries. Interestingly, the amount of botryticide applied with the brushing technique was more than 10 times lower than typical spray application rates (0.03 kg/ha versus 0.39 kg/ha). With spraying, a higher percentage of the botryticide suspension misses the target organs or is lost because of runoff, whereas with cluster brushing, these losses are minimal. By means of cluster brushing, the active ingredients are directly applied on the target organs, i.e., the cluster stems and berries. Hence, this application Table 3 Pairwise comparisons of the time point at which disease severity reached 5% (independent-samples Kruskal-Wallis test). Treatment comparison a Adjusted p value 1 with 2 0.045 1 with 3 0.035 1 with 4 0.003 2 with 3 1.000 2 with 4 1.000 3 with 4 1.000 a 1 = untreated control; 2 = brush; 3 = botrycide; 4 = botrycide-soaked brush. Figure 3 Temporal delay of the day of the year (DOY) when disease severity reached 5% for each of the treatments compared with the untreated control in the six cultivar year combinations. The box plots represent data from all six experiments and indicate the median and the 25 and 75 percentiles; whiskers indicate minima and maxima.

Flower Debris Removal Effect on Grape Bunch Rot 553 technique might enable pesticide reduction without reducing the biological efficacy. To apply methods used here to practical vineyard situations, technological challenges need to be solved and further tests at different phenological stages are necessary. Conclusions Manual removal of flower debris from grape clusters at growth stage BBCH 73 delayed the onset of grape bunch rot, thereby prolonging the potential ripening period. This crop cultural measure delayed the fungal epidemic to a similar degree as the application of a botryticide at the same stage of berry development. Therefore, flower debris removal might contribute to a reduction in, or to a partial replacement of, pesticide use in viticulture. Technical solutions need to be developed to practically and efficiently implement this crop cultural measure into an integrated bunch rot control strategy. Literature Cited Beresford, R.M., K.J. Evans, P.N. Wood, and D.C. Mundy. 2006. Disease assessment and epidemic monitoring methodology for bunch rot (Botrytis cinerea) in grapevines. N. Z. Plant Protect. 59:355-360. Duncan, D.B. 1955. Multiple range and multiple F tests. Biometrics 11:1-42. Evers, D., D. Molitor, M. Rothmeier, M. Behr, S. Fischer, and L. Hoffmann. 2010. Efficiency of different strategies for the control of grey mold on grapes including gibberellic acid (Gibb3), leaf removal and/ or botrycide treatments. J. Int. Sci. Vigne Vin 44:151-160. Hill, G., F. Stellwaag-Kittler, G. Huth, and E. Schloesser. 1981. Resistance of grapes in different developmental stages to Botrytis cinerea. J. Phytopathol. 102:328-338. Lorenz, D.H., K.W. Eichhorn, H. Bleiholder, R. Klose, U. Meier, and E. Weber. 1995. Phenological growth stages of the grapevine (Vitis vinifera L. ssp. vinifera) Codes and descriptions according to the extended BBCH scale. Aust. J. Grape Wine Res. 1:100-103. Molitor, D., M. Rothmeier, M. Behr, S. Fischer, L. Hoffmann, and D. Evers. 2011. Crop cultural and chemical methods to control grey mould on grapes. Vitis 50:81-87. Molitor, D., M. Behr, L. Hoffmann, and D. Evers. 2012. Impact of grape cluster division on cluster morphology and bunch rot epidemic. Am. J. Enol. Vitic. 63:508-514. Molitor, D., et al. 2015. Postponing first shoot topping reduces grape cluster compactness and delays bunch rot epidemic. Am. J. Enol. Vitic. 66:164-176. Ribéreau-Gayon, P. 1983. Alterations of wine quality caused by Botrytis damages. Vigne Vini. 10:48-52. Savage, S.D., and M.A. Sall. 1984. Botrytis bunch rot of grapes: Influence of trellis type and canopy microclimate. Phytopathology 74:65-70. Shtienberg, D. 2007. Rational management of Botrytis-incited diseases: Integration of control measures and use of warning systems. In Botrytis: Biology, Pathology and Control. Y. Elad et al. (eds.), pp. 335-347. Springer, Dordrecht, The Netherlands. Walter, R. 2012. Fäulnispilze an weintrauben erregerkomplex, mykotoxine und bekämpfungsstrategien. J. Kulturpfl. 64:378-383. Wolf, T.K., A.B.A.M. Baudoin, and N. Martinez-Ochoa. 1997. Effect of floral debris removal from fruit clusters on botrytis bunch rot of Chardonnay grapes. Vitis 36:27-33. Zoecklein, B.W., T.K. Wolf, N.W. Duncan, J.M. Judge, and M.K. Cook. 1992. Effects of fruit zone leaf removal on yield, fruit composition, and fruit rot incidence of Chardonnay and White Riesling (Vitis vinifera L) grapes. Am. J. Enol. Vitic. 43:139-148.