Annual Under-Vine Cover Crops Mitigate Vine Vigor in a Mature and Vigorous Cabernet franc Vineyard

Transcription

1 Annual Under-Vine Cover Crops Mitigate Vine Vigor in a Mature and Vigorous Cabernet franc Vineyard Ming-Yi Chou 1 and Justine E. Vanden Heuvel 2 * Abstract: Excessive vine vegetative growth in wet, cool climates increases management costs and compromises grape quality. The standard practice of bare soil under vines exacerbates the vigor problem. This study examined the ability of under-vine cover crops to mitigate excessive vigor in a mature vineyard. A three-year study was conducted in a mature Cabernet franc vineyard in Ovid, NY. Under-vine cover crops of chicory (CHI), tall fescue (FES), tillage radish (TR), alfalfa (ALF), or natural vegetation (NV) were established annually since 2014 to compare with a control of glyphosate-maintained bare soil (GLY). Pruning weight was reduced 64% by CHI and 54% by FES or TR compared to GLY in Primary and lateral shoot growth was suppressed by CHI in Fruiting zone leaf layer number was reduced by all of the under-vine cover crops in 2015, and by CHI and FES in Yield per vine was not reduced by any of the under-vine cover crops in any year of the study. Under-vine cover crops had little impact on harvest parameters including soluble solids, titratable acidity, ph, or yeast assimilable nitrogen. Soil properties such as aggregate stability, microbial respiration rate, and carbon mineralizability were improved by some cover crops. Although NV, FES, and TR showed little, and ALF showed neutral, effect on vine growth or yield components, CHI effectively alleviated mature vine vigor without reducing yield in a cool climate. Since water or nutrient competition alone did not fully explain the mitigating effect of under-vine cover crops on vine vigor, other mechanisms such as microbial or allelopathic effects of under-vine cover crops should be examined. Key words: cover crop, soil properties, vine vigor, yield 1 Former Doctor of Philosophy Student, Horticulture Section, 134A Plant Science, Cornell University, Ithaca, NY 14853; and 2 Associate Professor, Horticulture Section, Cornell University, Ithaca, NY and Cornell AgriTech, Geneva, NY *Corresponding author (justine@cornell.edu) Acknowledgments: This study was financially supported by the Towards Sustainability Foundation. The authors thank Steve Lerch, Justin France, Raquel Kallas, Kirsten Kurtz, Taylor Mattus, and Connor Roberson for technical assistance, Dave Weimann of Sheldrake Point Winery for hosting the study, and Françoise Vermeylen of the Cornell University Statistical Consulting Unit for statistical guidance. Manuscript submitted April 2018, revised Sept 2018, Oct 2018, accepted Oct 2018 Copyright 2019 by the American Society for Enology and Viticulture. All rights reserved. doi: /ajev Maintaining bare soil under the vine row by applying herbicide is the most common vineyard floor management practice worldwide. Because vineyard groundcover competes for water and nutrients (Wheeler et al. 2005, Lopes et al. 2008, Celette et al. 2009), maintaining bare soil under-vine increases water and nutrient availability and promotes vine vegetative growth. In the northeastern United States, frequent growing season precipitation and high soil organic matter combine to result in excessive vine vigor, which is one of the primary viticultural challenges (Wolf 2008). Excessive vine vegetative growth can lead to high canopy management costs (Smart and Robinson 1991), reduced fruit sunlight exposure, and increased disease incidence (Valdés- Gómez et al. 2008, Austin et al. 2011), which compromise fruit quality (Smart 1985). Maintaining bare soil with either herbicide or soil tillage risks degrading soil health through soil erosion, breakdown of soil aggregates, depletion of organic matter, and deterioration of the microbial environment (Peregrina et al. 2010, Blanco-Canqui et al. 2011, Napoli et al. 2017). Moreover, vineyard soil without groundcover results in pesticide and nutrient leaching which contaminates groundwater (Karl et al. 2016a). Thus, alternative under-vine floor management practices are desirable. Many studies have investigated the effect of under-vine cover crops in young vineyards. Some cover crop species reduced vine vegetative growth measures like pruning weight, leaf layers, and shoot length (Hatch et al. 2011, Giese et al. 2014, Hickey et al. 2016, Karl et al. 2016a), but few studies have examined mature vineyards. One study of a mature vineyard in the cool climate region of the Finger Lakes in New York found that annual ryegrass inconsistently reduced vine pruning weight; yield was affected, but grape composition was not examined (Centinari et al. 2016). In the same region, vine growth, yield, and grape composition were not affected by annual ryegrass, buckwheat, or resident vegetation growing under-vine in three years on one site (Jordan et al. 2016), but chicory reduced shoot growth, pruning weight, and yield in the second year of establishment at another site (Jordan 2014). These inconsistent results make predicting the impact of under-vine cover crops in mature vineyards difficult. Thus, a study of aggressive under-vine cover crops and their effects on vine growth, yield, and berry composition was needed. Cover crop species were chosen based on extensive belowground growth and high biomass production (Gentile et al. 2003, Koscielny and Gulden 2012, Gruver et al. 2014). Alfalfa as a legume crop may increase soil N content in the long term but a previous study in the same region demonstrated that white clover reduced vine vegetative growth in the short term (Karl et al. 2016a). 98

2 Under-Vine Cover Crops Mitigate Vine Vigor 99 The objective of this study was to determine if high resource competition from under-vine cover crops would affect vine growth, yield, and fruit composition in a mature vineyard in the cool climate region of the Finger Lakes consistently. Soil properties were also analyzed to evaluate how under-vine cover crops impact soil health parameters over three years. It was hypothesized that aggressive under-vine cover crops would reduce vine vegetative growth through reduced vine water status and nutritional status, and improve soil properties. Materials and Methods Experimental setup. This study was conducted from 2014 to 2016 at a commercial vineyard in Ovid, NY (42.66 N; W). The soil type was Howard gravely loam with less than 5% slope (Soil Survey Staff 1975). The climate in 2016 was warmest during the three years of the experiment, with total 1648 growing degree days (GDD), followed by 2015 (1586 GDD), and 2014 (1431 GDD), based on 10 C (Table 1). Although the sum of the precipitation was higher in 2016, the early growing season from late May to early August was the driest compared to 2014 and The early season (June and July) precipitation in 2014 and 2015 was 127% and 93% more than in In 2016, many sites in the Finger Lakes had a level three drought according to U.S. Drought Monitor ( but the experimental site had ample precipitation through much of the season. The Cabernet franc (Vitis vinifera) clone 1 grafted onto Couderc 3309 (V. riparia V. rupestris) rootstock vines were planted in 1999 in a north-south row orientation and trained on a Scott Henry trellis. The in-row vine spacing was 2.13 m and interrow spacing was 2.74 m. According to the standard practices of the region (Wolf 2008), the vines were late winter cane-pruned around February each year to a consistent bud number, ~16.4 buds per linear meter and 40 buds per vine, not including one extra cane that served the dual function of kicker and winter damage back-up cane, which was removed by bloom. An experimental plot was set up to compare five cover crop treatments with glyphosate-maintained bare soil as a control. The randomized complete block design (RCBD) was applied across four adjacent vineyard rows of the experimental site for four replications, with treatments and control randomly assigned within each replicate. Each experimental unit was comprised of four consecutive panels with three vines per panel (12 vines per experimental unit). The middle two panels (six vines) were used for data collection for a total of 48 panels (144 vines) in the experiment. Cover crops were seeded and the herbicide was applied to an ~1.2 m wide under-vine strip along the row. A permanent between-row cover crop, a mix of fescue, white clover, and weeds, was maintained separately and mowed periodically. Under-vine cover crop establishment. The five undervine cover crop treatments were natural vegetation (NV), alfalfa (ALF), tall fescue (FES), tillage radish (TR), and chicory (CHI). Seeding rates varied by treatment (Table 2). The control was maintained by applying Roundup (Roundup PRO concentrate, Monsanto). The under-vine cover crops and herbicide strips were established on an annual basis due to the need to hill soil around the graft unions each fall to protect latent scion buds for the winter, then de-hill in spring to reduce opportunities for scion rooting. For the NV treatment, weeds were allowed to grow freely; the other cover crop treatments were seeded. Seeds of ALF, FES, CHI, and TR treatments were hand-broadcast between 26 May and 2 June 2014, 13 and 15 May 2015, and 25 to 26 May 2016, when the vines were roughly at Eichhorn- Lorenz scale (Dry and Coombe 2004). The seeding rates were the same for FES across all three years but rates Table 2 Scientific name, common name, and seeding rate of under-vine cover crop treatments used in the experiment. The seeds were purchased from Ernst Seeds, PA. Abbrev. Scientific name Variety/ common name Seeding rates (kg/ha) , 2016 CHI Cichorium intybus Blue chicory ALF Medicago sativa Alfalfa, vernal TR Raphanus sativus Tillage radish, Ground hog FES Festuca arundinacea Tall fescue, Kentucky Table 1 Growing degree days (GDD base 10 C) and precipitation for the experimental site during the growing season from 2014 to GDD ( C) a Precipitation (cm) a Month Historic b Historic b April May June July August September October Sum a In 2015, GDD data were obtained from the Romulus, NY station, ~5km northwest of the experimental site, and precipitation data were obtained from the Varick, NY station, 17.8 km northwest of the experimental site. In 2015 and 2016, both GDD and precipitation data were obtained from the Ovid, NY station, 4.1 km northwest of the experimental site. b Historic GDD and precipitation data were the average from 1973 to 2015 obtained from the Geneva, NY station.

3 100 Chou and Vanden Heuvel were increased in 2015 and 2016 for the other cover crops due to poor establishment in 2014 (Table 2). The control was established with Roundup application, in which glyphosate was the active ingredient, at a rate of 2.9 kg a.i./ha of a 2% solution on 24 June and 16 July 2014, 16 June 2015, and 15 June The cover crop treatments were trimmed in only one year (as they approached the height of the fruiting zone) using a string trimmer on 8 to 9 Aug Trimming/mowing of cover crops was not required (based on cover crop height) in the other years of the study. Ground coverage assessment and weed identification. In each experimental unit, two 400 cm 2 square-shaped grids were chosen randomly at veraison in 2015 and 2016 using a wooden frame with 20 cm inner length. A digital photo was taken at 1.5 m vertically above each chosen grid, with a measuring tape placed horizontally on the ground as a photo scaling reference. The aboveground tissue was harvested from each of the chosen grids, cover crops were separated from weeds, and both were placed into separate paper bags, dried in an oven at 60 C overnight, and weighed. The chosen grid within each digital photo was analyzed with ImageJ Version 1.50b (open resource via to define the proportion of ground covered with biomass with image processing steps similar to Ricotta et al. (2014). Cover crop coverage was determined by dividing cover crop biomass by total biomass for each experimental unit. The weeds in the NV treatment were identified visually using the same digital photos. Shoot growth measurement. Four shoots per data vine were selected randomly and marked at the beginning of each growing season for growth measurements. Shoot diameters were measured (mm) with calipers at the middle of internode one above the first fully developed bud, where two measurements were taken per shoot and then averaged. Shoot length was measured (cm) with measuring tape from the base to shoot tip. Lateral shoots at the second node from the marked primary shoots were measured using the same methods as for the primary shoot, beginning after the first hedging and continuing until the second hedging or shoot thinning, which resulted in missing tagged shoots. The presence/absence of the lateral shoot at the second node of the marked primary shoot was recorded to calculate the percentage of primary shoots with lateral shoot growth. Canopy architecture EPQA. Point quadrat analysis (PQA) (Smart and Robinson 1991) and enhanced point quadrat analysis (EPQA) (Meyers and Vanden Heuvel 2008) were conducted 25 Aug 2015 and 19 Aug 2016, to characterize the canopy light environment at veraison. A thin wooden stick was inserted horizontally through the fruiting zone, perpendicular to the row, at 20 cm intervals on a per panel basis while recording leaf and cluster contacts. Light environment in the fruiting zone was measured on the same day using a ceptometer (Decagon, model AccuPAR LP-80), which recorded the photon flux, with an ambient flux sensor attached. Two measurements were taken on the fruiting zone per data vine, within an hour of solar noon. The ambient flux sensor was pointed vertically toward the sky above the canopy, without shading the ceptometer. The proportion of sunlight interception was calculated by dividing the fruiting zone photon flux by ambient photon flux. Light interception and PQA data were uploaded into Canopy Exposure Mapping Tools, version 1.7 (available via Jim Meyers, jmm533@cornell.edu), to calculate leaf layer number, occlusion layer number, interior leaf percentage, interior cluster percentage, cluster exposure layer, and cluster exposure flux availability. Vine water and nutrient status measurements. Vine midday stem (Ψ Midday ) and predawn leaf (Ψ Predawn ) water potential were measured with a pressure chamber (Soil Moisture Equipment Corporation, model 3005F01) as described (Fulton et al. 2001). Measurements of Ψ Midday were taken within an hour of solar noon on a biweekly basis, and Ψ Predawn was measured during fruit ripening on 11 Sept 2015 and 13 Sept 2016 within an hour of 0400 hr Eastern Standard Time. For Ψ Midday measurements, fully expanded healthy young leaves were bagged with a 500 ml aluminum foil-covered Ziploc bag for 15 min before measurement. Leaves were cut from the shoot with a sharp blade and transferred immediately into the pressure chamber. The chamber was pressurized at ~1 bar/ sec until xylem sap moistened the cut surface of the petiole. Three measurements (one leaf per measurement) were taken for each experimental unit on each measurement date. One hundred petiole samples per experimental unit, from young, fully expanded leaves, were collected at roughly full bloom on 20 June 2015 and 24 June 2016, and at veraison on 4 Sept 2014, 24 Aug 2015, and 26 Aug The petioles were washed with mild soap, rinsed with deionized water, and sent to the Cornell Nutrient Analysis Laboratory (CNAL) to determine total carbon and nitrogen using combustion and macro- and micronutrients (Al, B, Ca, Cu, Fe, K, Mg, Mo, Mn, Na, P, Zn) using a dry ash extraction method as described (Campbell and Plank 1998). Yield components, pruning weight, and juice composition measurements. Yield data were collected at commercial harvest as determined by the grower on 25 Oct 2014, 17 Oct 2015, and 22 Oct The clusters from each vine were clipped manually, counted, and pooled in a plastic lug to determine the yield per vine using a hanging scale (Salter Brecknell, model SA3N340, accuracy ±0.1 kg). The total yield per vine was then divided by the number of clusters to determine the average weight per cluster. An extra 100 berries per experimental unit were collected at harvest, stored in a Ziploc bag at -20 C, then weighed (Santorius ELT103, accuracy ±0.001 g) to determine the average berry weight. The cooperating grower s standard practice was to prune the lower tier of the Scott Henry vines in late fall and then prune the upper tier in early spring. Pruning weight was collected on a per vine basis from upward shoots in 2014, and from all shoots in 2015 and 2016 over the two grower-implemented pruning dates. Only upward shoot pruning weight was collected in 2014 because in this year, the downward shoots were pruned by the grower in an untraceable manner. In 2016, the downward shoots were pruned by the grower prior to data collection. However, the shoots remained directly under the vine, so the data were collected by reconstructing the vine (by

4 Under-Vine Cover Crops Mitigate Vine Vigor 101 matching shoots to the vine based on size, shape, and color of the cut surfaces). Pruning weight was used as an indicator for vine vegetative growth and to determine the Ravaz index by dividing yield by pruning weight. Twenty clusters were collected randomly from each experimental unit at harvest and stored in a -20 C freezer before juice composition analysis. The clusters were then thawed at room temperature, whole cluster-pressed, and the juice was filtered with cheesecloth. Juice soluble solids was measured using a temperature compensating digital refractometer (SPER Scientific, ). Titratable acidity (TA) was measured by titrating 50 ml juice against 0.10 M NaOH to ph 8.2, and ph, with a benchtop ph meter (VWR Symphony ph Meter, model SB80RI). Yeast assimilable nitrogen (YAN) was calculated by combining the content of ammonia determined with a UniTAB Kit (Unitech Scientific) and primary amino nitrogen analyzed by derivatization using a Chemwell 2910 autoanalyzer (Unitech Scientific) and spectrophotometrically at 340 nm (Nisbet et al. 2013). YAN was quantified in 2015 and 2016 only. Cluster compactness was measured in 2016 only. Ten clusters from each experimental unit were collected at harvest, then berries from each cluster were counted and removed to measure the length of the rachis. The compactness was calculated as the number of berries per cm rachis. Analysis of soil properties. Under-vine soil samples were collected at the end of the growing season in November 2015 and Six soil cores to 20 cm deep were taken from each experimental unit, combined, and analyzed for wet aggregate stability, organic matter content, and microbial respiration rate. Soil properties were measured as described (Gugino et al. 2009, Karl et al. 2016a). Briefly, aggregate stability was measured with dried soil that was sieved to select particle sizes between 0.25 and 2 mm. The sample was placed on a 0.25 mm sieve and water droplets generated from a rain simulator were applied at mm/sec for 5 min. Soil particles retained on, and passed through, the sieve were collected, dried, and weighed to determine the proportion of stable soil aggregates. Soil organic matter content was measured by dry combustion at 550 C for two hours. To measure cumulative microbial respiration, 50 g soil from each experimental unit with particle size smaller than 2 mm diameter was placed in a 250 ml airtight and sterilized glass jar with 20 ml 0.5 M NaOH contained in a plastic tube. The jars were placed in darkness at 30 C for two weeks. Electrical conductivity of the NaOH in each plastic tube was measured and compared with a control solution to calculate the CO 2 generated during incubation. Carbon mineralizability was calculated by dividing the CO 2 generation rate by organic carbon content. Additional tests of Morgan-extractable P, K, and micronutrients were conducted in Briefly, soil nutrients were extracted using Morgan s solution and quantified with inductively coupled argon plasma spectrophotometry. In November 2016, four intact soil cores per experimental unit were collected for soil bulk density measurement. The soil samples were stratified into 0 to 5, 5 to 10, and 10 to 15 cm by hand, dried in oven at 60 C for 24 hrs, weighed, and weight was divided by volume. Statistical methods. The data were checked for normality and analyzed with mixed-model analysis of variance (ANOVA), where under-vine floor treatments were classified as a fixed effect and blocks as random effects, using JMP Pro version The Dunnett s test was adopted for post-hoc comparison of treatment means compared to the mean of the GLY control at α = Results Cover crop establishment. Cover crops did not establish well in the first year of the experiment, likely due to residual grower-applied herbicide remaining in the treatment plots from previous seasons. In the second and third years of the experiment, the area under the vines was well covered with cover crops and weeds (Figure 1). The glyphosate control remained relatively bare, with ground coverage at veraison of 30% in 2015 and <10% in 2016 (Figure 1), while >70% ground coverage was achieved with cover crops. Natural vegetation had >70% ground coverage in 2015 and The weed species identified have been listed (Chou 2018). Among the cover crops, TR was the most difficult to Figure 1 Proportion of under-vine soil covered with weeds and cover crops in a Cabernet franc vineyard at veraison in (A) 2015 and (B) The bars indicate standard errors. The significant differences between each treatment and the control (GLY) were found using mixed-model analysis of variance following with Dunnett s test at α = NV: natural vegetation, GLY: glyphosate, CHI: chicory, ALF: alfalfa, TR: tillage radish, and FES: fescue.

5 102 Chou and Vanden Heuvel establish and resulted in the lowest coverage at ~27 and 38% in 2015 and 2016, respectively. Unlike TR, ALF established poorly in 2015 at 24% coverage, but grew well in 2016 and reached 67% ground coverage. FES and CHI established relatively well in the second and third years: CHI had 50% coverage in both years and FES had 53% in 2015 and 62% in Shoot growth. Primary shoot length was not different whether cover crops were or were not utilized throughout the first half of the 2015 growing season (Figure 2A). In early season 2016, the primary shoot length was longer in FES and TR vines than in GLY vines (Figure 2B). Primary shoot length of FES was ~16 and 30% longer than that of GLY on 6 and 20 June, respectively. Primary shoot length of TR was ~25% longer on both dates. Primary shoot length in the middle of the growing season was reduced by CHI by 29 and 39% compared to GLY on 12 and 25 July, respectively. Primary shoot diameter and lateral shoot length did not differ among vines growing with under-vine cover crop and vines in GLY in 2015 and 2016 (Chou 2018). The proportion of primary shoots with laterals was not different between any of the under-vine cover crops and GLY in In 2016, there were fewer primary shoots with laterals in CHI than in the GLY control by 12 Sept (Chou 2018). Figure 2 Primary shoot length of Cabernet franc vines growing with different under-vine cover crop treatments throughout the early to midgrowing season in (A) 2015 and (B) The significant differences between each of the treatment and the control (GLY) were tested using mixed-model analysis of variance following with Dunnett s test at α = NV: natural vegetation, GLY: glyphosate, CHI: chicory, ALF: alfalfa, TR: tillage radish, and FES: fescue. EPQA. Grapevine canopy structure was impacted in the second and third years of the experiment; EPQA data was not collected in 2014 (Table 3). In 2015, planting cover crops reduced the leaf layer number in the fruiting zone, and NV, TR, and FES resulted in reduced occlusion layer numbers compared to the GLY. Although canopy structure was affected by the treatments in 2015, the light environment parameters, including cluster exposure layer and cluster exposure flux availability, were not affected. In 2016, the fruiting zone leaf layer number was affected by cover crop: CHI reduced leaf layer by 35% and FES, by 28%. Cluster exposure layer and percent interior clusters were affected by under-vine cover according to mixed model ANOVA, where CHI had only 9.48% interior clusters and 0.11 cluster exposure layer, compared to 22.41% and 0.24 in GLY. Yield components and berry composition. Yield per vine was not affected by under-vine cover crops in 2014 and 2015, but in 2016, NV and TR increased yield by 100 and 77%, respectively, compared to GLY (Table 4). Since cluster weight was not affected by any of the under-vine cover crops, the yield increase in NV and TR was due mainly to the increased number of clusters per vine. Only upward shoot pruning weight was collected in 2014, which indicated no differences (p = , data not shown). In 2015, CHI reduced pruning weight by 65% compared to GLY. Although the mean pruning weight of CHI (0.18 kg/vine) was less than half that of GLY (0.44 kg/vine) in 2016, there was no difference according to mixed model ANOVA (p = ). Pruning weight was also reduced ~54% by TR and FES compared to that of GLY in However, the Ravaz index (yield/pruning weight) was only affected by CHI (increase of 129%) in In 2016, cluster number per vine was increased 95, 66 and 73% by NV, TR, and FES, respectively. Berry size was increased by under-vine cover crops in Number of berries per cluster was 31 and 25% less in CHI and FES in 2016, respectively. Berry weight increased 22% in TR in 2014 and 11% in NV in 2015 compared to berries of GLY. In 2016, the third year of the experiment, berries from all cover crop treatments increased berry size by 10 to 17% compared to the mean berry weight of 1.26 g in GLY. Cluster compactness was only measured in Cluster compactness was affected by the under-vine floor treatments as revealed by ANOVA, but no pairwise difference was detected by Dunnett s test. The modification of the cluster compactness was likely because rachis length was reduced 18.6% by CHI compared to the control (Chou 2018). Berry soluble solids and under-vine cover crops had little impact on the berry harvest parameters. TA was not impacted by any of the under-vine cover crops in any year (Chou 2018). Juice ph was reduced 5% by FES in 2014 and YAN was reduced 40% by CHI in 2015, compared to ph 3.42 and 112 mg/l of YAN in GLY in 2014 and 2015, respectively. Nutrient and water status. Petiole nutrient differences were found at bloom in 2015 and at veraison in At bloom 2015, FES reduced Na by 0.05 g/kg (28%), all treatments but NV increased P by up to 1.21 g/kg (48%), and NV, CHI, and FES reduced Zn by up to 10 mg/kg (20%) compared to GLY (Table 5). Mg was also affected by cover

6 Under-Vine Cover Crops Mitigate Vine Vigor 103 crops tested by ANOVA, without post-hoc differences. At veraison in 2016, the petiole C content was reduced 1.4% by NV and increased 0.9 and 1% by CHI and FES, respectively, compared to GLY (Table 5). At the same time, NV increased petiole Fe and Na content by 16.3 mg/kg (75%) and 0.94 g/ kg (300%), respectively, compared to GLY. N did not differ among treatments, ranging from 0.80 to 0.91% at bloom 2015, 0.67 to 0.81% at veraison 2015, and 0.51 to 0.61% at veraison Other nutrients not impacted by the cover crops, including data of veraison 2015, can be found in Chou There were no Ψ Midday differences found between any of the under-vine cover crop regimes and GLY in 2015 throughout the growing season (Figure 3A). In 2016, FES vines had reduced mean Ψ Midday (-8.1 bar) on 20 June; increased mean Ψ Midday (-6.6 bar) on 3 Aug; NV vines had reduced mean Ψ Midday (-9.9 bar) on 27 Aug; and NV, ALF, and FES vines had reduced mean Ψ Midday, ranging from -12 to -13 bars on 13 Sept, when the lowest Ψ Midday in the growing season was observed (Figure 3B). Differences between under-vine cover crop and GLY in late-season Ψ Predawn were found in 2015 and TR vines consistently had lower Ψ Predawn, by 42% in 2015 and by 55% in 2016 (Chou 2018). In 2016, FES reduced Ψ Predawn by 60% of GLY. Soil properties. Soil properties were generally improved by NV, FES, and ALF compared to GLY (Table 6). Stable wet soil aggregates improved by 20.3% in NV and by 20.5% in ALF compared to 11.1% in GLY in Organic matter content did not differ between any of the under-vine cover crops and GLY, even though the mean organic matter content in GLY was constantly the lowest, with 2.97% in 2015 and 3.02% in Microbial respiration rate increased 75% by FES in Soil carbon mineralizability of NV in 2015 and FES in 2016 were 54 and 68% greater, respectively, than GLY in each year. In 2015, soil morgan-extractable nutrients were analyzed (Chou 2018). There were no differences found among treatments and control soil in P, K, Fe, or Zn content. Soil Mg content increased 21, 18, and 17% in NV, CHI, and FES, respectively, while Mn content increased 35 and 38% in CHI and FES, respectively. Soil nutrient data was not collected in There were no soil bulk density differences found between any of the under-vine cover crops and GLY in any soil depth from 0 to 15 cm (Chou 2018). Discussion Our study showed that under-vine cover crops can reduce vine vegetative growth in a vigorous, mature, cool-climate vineyard without reducing yield, which is a major economic concern to growers. Alleyway cover crops altered vine balance by maintaining the same level of yield with reduced pruning weight in warm- and hot-climate mature vineyards in Spain (Pérez-Álvarez et al. 2015) and Portugal (Lopes et al. 2008). However, under-vine cover crop studies in cool climates reported little to no impact on vegetative growth and Ravaz index (Jordan 2014, Centinari et al. 2016). Under-vine chicory was found to reduce vine canopy density and pruning weight, but did not impact yield in New Zealand (Wheeler et al. 2005). However, this study was conducted in a young vineyard and presented only a single year of results. In this study, under-vine chicory most effectively reduced vine vegetative growth, including pruning weight and fruiting zone canopy density, while other cover crops showed relatively inconsistent effects on the same parameters. The Table 3 Approximation of canopy architecture at veraison using enhanced point quadrat analysis of Cabernet franc vines with different under-vine cover crops in 2015 and Values are means +/- standard errors. Treatment a 2015 veraison Leaf layer number Occlusion layer number % Interior leaves % Interior clusters Cluster exposure layer Cluster exposure flux availability GLY 1.75 ± 0.24 b 2.33 ± ± ± ± ± 0.07 NV 1.18 ± 0.23* c 1.84 ± 0.15* 15.0 ± ± 7.20* 0.19 ± ± 0.06 CHI 1.22 ± 0.23* 2.11 ± ± ± ± ± 0.06 TR 1.08 ± 0.23** 1.82 ± 0.15** 14.7 ± ± ± ± 0.06 ALF 1.21 ± 0.23* 1.95 ± ± ± 7.20* 0.20 ± ± 0.06 FES 1.20 ± 0.23* 1.80 ± 0.15** 8.93 ± ± ± ± 0.06 p-value d veraison GLY 1.34 ± ± ± ± ± ± 0.06 NV 1.27 ± ± ± ± ± ± 0.05 CHI 0.87 ± 0.06*** 1.93 ± ± ± ± ± 0.05 TR 1.16 ± ± ± ± ± ± 0.05 ALF 1.25 ± ± ± ± ± ± 0.06 FES 0.96 ± 0.07* 2.01 ± ± ± ± ± 0.06 p-value a Treatments were GLY: glyphosate, NV: natural vegetation, CHI: chicory, TR: tillage radish, ALF: alfalfa, and FES: fescue. b Pooled standard error. c Significance designation of Dunnett s test, compared to GLY: *p < 0.05, **p < 0.01, ***p < d The p-value was derived from mixed model analysis of variance following at α = 0.05.

7 104 Chou and Vanden Heuvel inconsistency of the cover crop effect on vine growth may result from different growing habits of the cover crops, such as rooting pattern, depth, and density, that may trigger different water and nutrient dynamics in the soil (Sainju et al. 1998, Perkons et al. 2014, Karl et al. 2016a). The timing of the competition was likely different between weeds and cover crop due to different timing of establishment and growth. Also, the effect of year-to-year weather variation, especially uneven precipitation, on cover crop coverage may be significant. At our experimental site, drastically different weather patterns were found during the experiment which may have confounded the effects. Reduced vine vegetative growth with under-vine CHI produced the high crop load, as shown by Ravaz index, that exceeded the recommended range of 5 to 10 for quality table- and winegrape production under various trellis systems (Kliewer and Dokoozlian 2005). However, the grape ripeness level was not compromised, as the berry soluble solids and TA of CHI treatment did not differ from that of the GLY control. Although wines were not made in this study, the grape harvest parameters indicated that the reduced vegetative tissue, including leaf layers and shoots, were unnecessary to ripen the fruit to commercially acceptable standards for the region. Further study on wine sensory properties and other chemical compounds such as secondary metabolites is required. Resource competition is the reason cited for reduced vine growth in previous cover crop studies (Tan and Crabtree 1990, Wheeler et al. 2005, Monteiro and Lopes 2007). In our study, however, there were no clear associations of reduction of water potential and nutrient content with the vine growth in the cover crop treatments, although reduced vine water Treatment a Table 4 Yield components of Cabernet franc vines growing with different under-vine cover crops. Values are means +/- standard errors Yield (kg/vine) Pruning weight (kg/vine) GLY 4.61 ± 0.85 b 3.90 ± ± ± ± 0.09 NV 4.25 ± ± ± 0.63*** c 0.81 ± ± 0.08 TR 6.51 ± ± ± 0.63** 0.53 ± 0.18* 0.34 ± 0.08 CHI 6.11 ± ± ± ± 0.18** 0.18 ± 0.08 FES 5.57 ± ± ± ± 0.18* 0.36 ± 0.08 ALF 5.68 ± ± ± ± ± 0.09 p-value d Ravaz index (yield/pruning wt) Cluster per vine (no.) Treatment GLY 7.5 ± ± ± ± ± 7.1 NV 6.1 ± ± ± ± ± 6.0*** TR 10.5 ± ± ± ± ± 6.1* CHI 17.3 ± 3.6* 18.2 ± ± ± ± 6.2 FES 8.2 ± ± ± ± ± 6.3** ALF 13.6 ± ± ± ± ± 7.2 p-value < Berry per cluster (no.) Cluster weight (g/cluster) Treatment GLY 70.7 ± ± ± ± ± ± 10.0 NV 73.6 ± ± ± ± ± ± 8.5 TR 75.6 ± ± ± ± ± ± 8.5 CHI 78.0 ± ± ± 6.3** ± ± ± 8.7 FES 68.0 ± ± ± 6.4* ± ± ± 8.8 ALF 64.0 ± ± ± ± ± ± 10.1 p-value < Berry weight (g/berry) Treatment GLY 1.60 ± ± ± 0.03 NV 1.75 ± ± 0.03* 1.47 ± 0.02**** TR 1.95 ± 0.05* 1.63 ± ± 0.02*** CHI 1.63 ± ± ± 0.02**** FES 1.75 ± ± ± 0.02**** ALF 1.83 ± ± ± 0.02*** p-value < < a Treatments were GLY: glyphosate, NV: natural vegetation, TR: tillage radish, CHI: chicory, FES: fescue, and ALF: alfalfa. b Pooled standard error. c Significance designation of Dunnett s test, compared to GLY: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < d p-value: The p-value was derived from mixed model analysis of variance following at α = 0.05.

8 Under-Vine Cover Crops Mitigate Vine Vigor 105 Table 5 Vine petiole nutrient analysis of Cabernet franc vines growing with different under-vine cover crops in 2015 at bloom and in 2016 at veraison bloom Treatment a Mg (g/kg) Na (g/kg) P (g/kg) Zn (mg/kg) GLY 3.99 ± 0.16 b 0.18 ± ± ± 2.4 NV 3.62 ± ± ± ± 2.4* c TR 3.88 ± ± ± 0.20* 50.4 ± 2.4 CHI 4.25 ± ± ± 0.20* 47.8 ± 2.4* FES 3.50 ± ± 0.01* 3.32 ± 0.20* 46.1 ± 2.4** ALF 4.52 ± ± ± 0.20** 54.2 ± 2.4 p-value d veraison Treatment C (%) Al (mg/kg) Na (g/kg) Fe (mg/kg) GLY 38.4 ± ± ± ± 4.2 NV 37.0 ± 0.3** 32.4 ± ± 0.13** 39.3 ± 4.2* TR 39.0 ± ± ± ± 4.2 CHI 39.3 ± 0.3* 17.6 ± ± ± 4.2 FES 39.4 ± 0.3* 15.7 ± ± ± 4.2 ALF 39.3 ± ± ± ± 4.2 p-value < a Treatments were GLY: glyphosate, NV: natural vegetation, TR: tillage radish, CHI: chicory, FES: fescue, and ALF: alfalfa. b Pooled standard error. c Significance designation of Dunnett s test, compared to GLY: *p < 0.05, **p < d The p-value was derived from mixed model analysis of variance following at α = status in TR and FES did couple with reduced vine vegetative growth. Predawn leaf water potential, often used as a soil water indicator (Winkel and Rambal 1993), was reduced by TR in 2015, as were pruning weight and leaf layer number. In 2016, the reduction in early and late growing season Ψ Midday and late season Ψ Predawn in FES may also explain the reduced leaf layer numbers. Since the tall fescue seeds were broadcast in late May, they likely started providing water competition in June. Water competition from FES may have affected vine vegetative growth, especially in the early growing season, as vines are sensitive to water deficit before veraison, and can lead to inhibition of vegetative and reproductive growth (Hardie and Considine 1976, Matthews and Anderson 1989); however, shoot length in FES was greater than GLY throughout June The mechanism behind the reduced vegetative growth with under-vine CHI remained unclear, as vine water and nutrient status were not reduced except for petiole Zn content in This fell within the recommended range for healthy vine growth (Wolf 2008), although analyses of leaf blades or whole leaves for nutrients may have better illuminated differences among treatments, particularly in N (Schreiner and Scagel 2017). Nitrogen competition was suggested to be the primary reason for reduced vine vegetative growth in many cover crop studies in various climates (Wheeler et al. 2005, Celette et al. 2009, Pérez-Álvarez et al. 2015). However, Figure 3 Midday stem water potential of Cabernet franc vines growing with different under-vine cover crops throughout the first half of the growing season in (A) 2015 and (B) The significant differences between each of the treatment and the control (GLY) were tested using mixed-model analysis of variance following with Dunnett s test at α = NV: natural vegetation, GLY: glyphosate, CHI: chicory, ALF: alfalfa, TR: tillage radish, and FES: fescue.

9 106 Chou and Vanden Heuvel we found no differences among treatments and control vines in the petiole N content. Although the nitrogen content of vines from both treatments and control were at the borderline or lower than the recommended value (1.2% at bloom and 0.8% at latter stage; Wolf 2008), no visual nutrient deficiency symptoms were observed. This phenomenon is consistent with previous studies done in the region (Centinari et al. 2016, Jordan et al. 2016, Karl et al. 2016b). Previous studies using barley as a cover crop reported reduced nitrate content in the surface soil since the first year of the experiment, yet there was reduced nitrogen in plant tissue only in the third year and in grape must only in the fourth (Pérez-Álvarez et al. 2015). This suggests that the nutrient competition begins in the soil, spreads to vegetative tissues later, and then finally to grapes. With only petiole N and fruit YAN measured here, it is unknown whether the N competition occurred in the soil. Although the same petiole nutrient concentrations were found in vigorous GLY vines and smaller CHI vines, the equality of the nutrient concentration may be due to a dilution effect (Jarrell and Beverly 1981), where a high amount of absorbed nutrient distributed in large vines had the same nutrient concentration as smaller vines that had absorbed less nutrient. Although the range of Ψ Predawn found in our study falls between -0.2 to -0.4 MPa, which is classified as mild water stress (Ojeda et al. 2002), no visual symptoms of water stress were observed during the growing season. Although CHI reduced vine vigor, it did not affect vine water status in any of the years. This was not the first report that reduced vine vigor as a result of cover cropping had no association with vine water status. Interrow tall fescue growth did not affect vine midday stomatal conductance and Ψ Predawn, but reduced pruning weight by ~5 t/ha (Celette et al. 2005) in a young Sauvignon blanc vineyard in a Mediterranean climate. In the cool-climate Finger Lakes region of New York, under-vine white clover reduced the pruning weight of young Cabernet franc vines without affecting Ψ Midday or Ψ Predawn (Karl et al. 2016b). Aside from water and nutrient competition, cover crops may also have allelopathic effects on vine growth (Wolpert et al. 1993, Celette et al. 2005). In fact, tall fescue can allelopathically suppress the below- and aboveground growth of young pecan trees (Smith et al. 2001). Berry composition was not impacted by most of the cover crops, as in previous studies (Monteiro and Lopes 2007, Tesic et al. 2007). The exception was CHI, which reduced berry YAN compared to GLY in Reduced grape must YAN as a result of cover crops has often been coupled with reduced petiole nitrogen content (Sweet and Schreiner 2010, Pérez- Álvarez et al. 2015, Karl et al. 2016b). Winegrapes from the Finger Lakes generally have low YAN (Nisbet et al. 2014, Karl et al. 2016a). The YAN of grapes from all treatments and control in this study also had lower than the recommended content for healthy fermentation (Bell and Henschke 2005, Boulton et al. 2013), so nitrogen adjustment may be required regardless of under-vine cover crop. Under-vine soil physical, chemical, and microbial properties were impacted by most under-vine cover crops. The lack of impact with TR may be due to its growth habit. TR grows actively in the fall, producing a high amount of biomass with a low C/N ratio (Weil et al. 2009), but does not favor growth of plant-beneficial fungi and other soil microorganisms due to its Brassicaceae biofumigation effect (Sarwar et al. 1998, White and Weil 2010). Thus, large amounts of biomass were still produced but did not build soil organic matter. Although barley and clover as cover crops did not impact vineyard soil P, K, and Mg content in Spain (Pérez-Álvarez et al. 2015), Mg increased in the soil of our NV, CHI, and FES treatments in Cover crops enriched vineyard soil organic matter, increased soil aggregation, improved microbial respiration rate, and resulted in higher nutrient mineralization in previous studies (Steenwerth and Belina 2008, Peregrina et al. 2010, Ruiz-Colmenero et al. 2013), some of which may be attributable to significant shifts in the soil microbiome as a function of under-vine cover crops (Chou et al. 2018). In this study, soil organic matter was not affected, but microbial respiration rate and mineralizability were improved by FES. The decoupling of soil organic matter content and microbial activity may indicate that FES effectively built the labile carbon in the soil and created a microbe-friendly soil environment, but the contribution to stable soil organic carbon was negligible in the short term. Microbial activity is sensitive to the short-term enrichment of labile carbon, such as the cover crop residues in this study, that are used readily as metabolic Table 6 Property parameters of the under-vine soil treated with different cover crops. Values are means +/- standard errors. Aggregate stability (%) Organic matter (%) Microbial respiration (mg CO 2 g/14 days) C mineralizability (mg CO 2 /g organic carbon in soil) Treatment a GLY 11.1 ± 2.1 b 17.1 ± ± ± ± ± ± ± 3.08 NV 20.3 ± 2.1* c 21.5 ± ± ± ± ± ± 4.78* ± 3.08 TR 12.1 ± ± ± ± ± ± ± ± 3.08 CHI 15.5 ± ± ± ± ± ± ± ± 3.08 FES 12.1 ± ± ± ± ± ± 0.11** 41.4 ± ± 3.08* ALF 20.5 ± 2.1* 23.6 ± ± ± ± ± ± ± 3.08 p-value d a Treatments were GLY: glyphosate, NV: natural vegetation, TR: tillage radish, CHI: chicory, FES: fescue, and ALF: alfalfa. b Pooled standard error. c Significance designation of Dunnett s test, compared to GLY, at *: p < 0.05, **: p < d The p-value was derived from mixed model analysis of variance following at α = 0.05.

10 Under-Vine Cover Crops Mitigate Vine Vigor 107 substrates (Sparling 1997). However, the stabilized soil organic carbon pool takes a long time to accumulate (Wander et al. 1994, Smith 2004). Economic analysis conducted in the same region found growing under-vine cover crops beneath young vines had a lower cost than maintaining bare soil by applying herbicide (Karl et al. 2016b). However, that study did not recommend using cover crops because yield reduction in cover crop treatment could lead to loss of total revenue by up to $4000/ha in a young vineyard. In the northeastern United States, yield is key to economic viability for grapegrowers due to the low profit margin (Yeh et al. 2014). Mature vines responded to under-vine cover crops differently than young vines, which have a less extensive root system and carbohydrate reserves (Holzapfel et al. 2010), potentially making mature vines more resilient to resource competition. In this study, yield was not reduced in any of the under-vine cover crop regimes, though yield was increased in NV and TR in 2016, suggesting that using under-vine cover crops in mature, cool-climate vineyards could be financially justifiable or even beneficial. Conclusion Growing under-vine cover crops can reduce the need for herbicides, mitigate vine vigor, and improve soil properties in a mature cool-climate vineyard. However, the reduction in vine vegetative growth was not explained solely by water or nutrient competition. Future examination of the mechanism through which under-vine cover crops reduce vine vegetative growth in mature cool-climate vineyards would be useful. The uncompromised yield of vines growing with undervine cover crops indicated the financial feasibility of using under-vine cover crops in mature vineyards. However, practical adoption of the cover crops requires investigation of their impact on wine sensory properties and careful assessment on their adaptation to the specific sites and grape varieties and their resulting financial outcomes. Literature Cited Austin CN, Grove GG, Meyers JM and Wilcox WF Powdery mildew severity as a function of canopy density: Associated impacts on sunlight penetration and spray coverage. Am J Enol Vitic 62: Bell SJ and Henschke PA Implications of nitrogen nutrition for grapes, fermentation and wine. Aust J Grape Wine Res 11: Blanco-Canqui H, Mikha MM, Presley DR and Claassen MM Addition of cover crops enhances no-till potential for improving soil physical properties. Soil Sci Soc Am J 75: Boulton RB, Singleton VL, Bisson LF and Kunkee RE Principles and Practices of Winemaking. Springer Science & Business Media, Berlin, Germany. Campbell CR and Plank CO Preparation of plant tissue for laboratory analysis. In Reference Methods for Plant Analysis. Kalra YP (ed.), pp Taylor & Francis Group, Boca Raton, FL. Celette F, Wery J, Chantelot E, Celette J and Gary C Belowground interactions in a vine (Vitis vinifera L.)-tall fescue (Festuca arundinacea Shreb.) intercropping system: Water relations and growth. Plant Soil 276: Celette F, Findeling A and Gary C Competition for nitrogen in an unfertilized intercropping system: The case of an association of grapevine and grass cover in a Mediterranean climate. Eur J Agron 30: Centinari M, Vanden Heuvel JE, Goebel M, Smith MS and Bauerle TL Root-zone management practices impact above and belowground growth in Cabernet franc grapevines. Aust J Grape Wine Res 22: Chou MY Vineyard floor management in the Finger Lakes region: Physiological and microbial perspectives. Thesis, Cornell University, Ithaca, NY. Chou MY, Vanden Heuvel J, Bell TH, Panke-Buisse K and Kao-Kniffin J Vineyard under-vine floor management alters soil microbial composition, while the fruit microbiome shows no corresponding shifts. Sci Rep 8: Dry P and Coombe B Grapevine growth stages - The Modified E-L System. Viticulture 1 Resources, 2nd ed. Winetitles Media, Broadview, Australia. Fulton A, Buchner R, Gilles C, Olson B, Bertagna N, Walton J, Schwankl L and Shackel K Rapid equilibration of leaf and stem water potential under field conditions in almonds, walnuts, and prunes. HortTechnology 11: Gentile RM, Martino DL and Entz MH Root characterization of three forage species grown in southwestern Uruguay. Can J Plant Sci 83: Giese G, Velasco-Cruz C, Roberts L, Heitman J and Wolf TK Complete vineyard floor cover crops favorably limit grapevine vegetative growth. Sci Hortic 170: Gruver J, Weil R, White C and Lawley Y Radishes: A new cover crop for organic farming systems. Michigan State University, MI. Gugino BK, Idowu OJ, Schindelbeck RR, van Es HM, Wolfe DW, Moebius-Clune BN, Thies JE and Abawi GS Cornell Soil Health Assessment Training Manual, ed Cornell University, Geneva, NY. Hardie WJ and Considine JA Response of grapes to water-deficit stress in particular stages of development. Am J Enol Vitic 27: Hatch TA, Hickey CC and Wolf TK Cover crop, rootstock, and root restriction regulate vegetative growth of Cabernet Sauvignon in a humid environment. Am J Enol Vitic 62: Hickey CC, Hatch TA, Stallings J and Wolf TK Under-trellis cover crop and rootstock affect growth, yield components, and fruit composition of Cabernet Sauvignon. Am J Enol Vitic 67: Holzapfel BP, Smith JP, Field SK and Hardie WJ Dynamics of carbohydrate reserves in cultivated grapevines. In Horticultural Reviews, vol. 37. Janick J (ed.), pp John Wiley & Sons, Inc., Hoboken, NJ. Jarrell WM and Beverly RB The dilution effect in plant nutrition studies. Adv Agron 34: Jordan L Evaluating the effects of using annually established under-vine cover crops in northeastern Riesling vineyards. Thesis, Cornell University, Ithaca, NY. Jordan LM, Björkman T and Vanden Heuvel JE Annual undervine cover crops did not impact vine growth or fruit composition of mature cool-climate Riesling grapevines. HortTechology 26: Karl A, Merwin IA, Brown MG, Hervieux RA and Vanden Heuvel JE. 2016a. Impact of undervine management on vine growth, yield, fruit composition, and wine sensory analyses in Cabernet franc. Am J Enol Vitic 67: Karl AD, Merwin IA, Brown MG, Hervieux RA and Vanden Heuvel JE. 2016b. Under-vine management impacts soil properties and leachate composition in a New York State vineyard. HortScience 51: Kliewer WM and Dokoozlian NK Leaf area/crop weight ratios of grapevines: Influence on fruit composition and wine quality. Am J Enol Vitic 56: