Abstract
Three two-year studies evaluated the effects of several reflective mulches on yield and fruit composition in southern New England vineyards. A white reflective woven material (WRM) was tested under vines (covering the herbicide strip) of Chardonnay, Pinot noir, and Merlot in 2004 and 2005. Crushed quahog shells (QS) and a silver aluminized reflective mulch (SRM) were tested under vines of Merlot in 2005 and 2006, and quahog shells were compared with untreated controls on Cabernet franc and Chancellor in 2006 and 2007. WRM increased photosynthetically active radiation (PAR) reflectance into the canopy, but had no measurable impact on fruit composition. SRM improved PAR reflectance into the canopies when compared to the control, but did not impact canopy density, yield, or fruit composition. Rapid soiling and tearing were common problems with the SRM. The addition of a thin layer of quahog shells to the herbicide strip increased soil calcium levels, which resulted in higher pH, and substantially increased Ca:Mg ratio. QS increased canopy density, yield, cluster number, cluster weight, and Brix in some cultivars and some years.
High-quality grapes are required for making premium wine. Adequate fruit ripening, particularly in red cultivars, is difficult to achieve in coastal New England vineyards because of inadequate heat units. Reflecting radiation (specifically photosynthetically active radiation, PAR) from the vineyard floor into the canopy may enhance the maturity of grapes and add complexity to the wine (Robin et al. 1996). Because of their dark color, few soils have the natural tendency to reflect light upward from the vineyard floor (Martyn 1983, Post et al. 2000), so researchers have investigated the use of mulches with the intention of directing more sunlight into the canopy where it will be intercepted by leaves and clusters. Potential reflective mulches have included aluminized polypropylene (Coventry et al. 2005, Razungles et al. 1996, Reynolds et al. 2008, Robin et al. 1996), white geotextiles (Hostetler et al. 2007a, 2007b), and crushed mollusk shells (Creasy et al. 2006).
Previous reflective mulch studies have been limited due to either presentation of only a single year of data (Coventry et al. 2005, Razungles et al. 1996, Reynolds et al. 2008) or absence of statistical analysis (Coventry et al. 2005, Razungles et al. 1996, Robin et al. 1996). Among studies that have included statistical analyses (Hostetler et al. 2007a, 2007b, Reynolds et al. 2008), reflective mulches have demonstrated variable effects on yield components, fruit composition, and/or perceived wine quality. The objective of this research was to evaluate the effect of several reflective mulches (white woven geotextile, silver aluminized reflective product, and crushed quahog shells) placed in the herbicide strip of coastal New England vineyards on yield and fruit composition of winegrapes.
Materials and Methods
Three studies were conducted using one or more of the following materials: a reflective white woven material (WRM) (Extenday reflective fabric; Extenday USA LLC, Yakima, WA), a silver aluminized reflective mulch (SRM) (Brite ‘Nup; Adcock Manufacturing, Gardena, CA), and a thin layer of crushed quahog (Mercenaria spp.) shells (QS) that are a byproduct of the local seafood industry and are readily available in coastal New England.
Vineyard sites and experimental designs.
Study 1: WRM.
Test plots were established in 2004 at Newport Vineyards at 60 m elevation in Middletown, Rhode Island (lat: 41.55° N; long: 71.29°W). Six-year-old Merlot, five-year-old Chardonnay, and three-year-old Pinot noir, all grafted on 3309 rootstock, were used in study 1. The soil was characterized as a Newport coarse-loamy, mixed, active, mesic Typic Dystrudepts (Soil Survey Staff 2008) with an 8% slope facing east, and rows were oriented north-south. All vines were trained to a cane-pruned vertical shoot positioning (VSP) system and planted on 1.5 m × 2.7 m spacing. For each variety, the experimental design was a randomized complete block with four replicates. Each plot consisted of three panels (five vines each) and the middle panel was used for data collection. Vines were shoot-thinned, cluster-thinned, and leaf-pulled by the grower according to normal vineyard practices (Wolf 1995).
Two treatments were investigated in study 1: WRM and untreated controls (1-m wide herbicide strip). The WRM was applied on 11 May 2004 as lengths of 0.5-m strips on either side of the vine row. The strips met directly under the vines so that no soil was visible. The WRM was secured to the ground with large (20-cm long) 11-gauge metal staples at ~1-m intervals. The reflective fabric was installed at the beginning of the study and was maintained by the grower during the two-year period. The projected life expectancy of the WRM reported by the manufacturer is five years (Bertelsen 2005).
Study 2: SRM and QS.
Test plots were established at the Newport site in a Merlot vineyard (described for study 1) in 2005. Three treatments were evaluated in this study: SRM, QS, and untreated controls (1.0-m wide herbicide strip). The SRM was applied as lengths of 0.5-m strips on either side of the vine row and secured with large staples at ~1-m intervals. The strips met directly under the vines so that no soil was visible. The crushed QS was applied by hand as a thin layer (~2 cm thick), extending ~0.5 m on either side of the vine. Both applications were made on 14 June 2005. The experimental design was a randomized complete block with four replicates. Each plot consisted of one panel of five vines trained to a cane-pruned VSP system. The middle three vines were used for data collection. Vines were shoot-thinned, cluster-thinned, and leaf-pulled by the grower according to normal vineyard practices (Wolf 1995).
Study 3: QS.
Test plots were established in 2006 at COJ Vineyards, in Brewster, Massachusetts, situated at 5-m elevation (lat: 41.76° N; long: 70.08°W). Vines were four-year-old Cabernet franc and Chancellor, both grafted on 3309 rootstock. The soil at the site was characterized as a Carver mesic, uncoated Typic Quartzipsamments (Soil Conservation Service 1993) with a 1% to 3% slope, with rows running north-south, at 2.4 m × 1.8 m spacing. All vines were trained to a cane-pruned VSP system. Treatments, applied under the vine canopy on 6 June 2006, were a thin layer of QS in the herbicide strip and untreated controls. For each variety, the experimental design was a randomized complete block with five replicates. Each plot consisted of one panel of four vines with guard panels between treatments. Vines were shoot-thinned, cluster-thinned, and leaf-pulled by the grower according to normal vineyard practices (Wolf 1995).
Bud counts and canopy measurements.
Bud survival (percentage live buds) was determined in May 2005 for study 1 and study 2. Canopy characterization consisted of point quadrat analysis (PQA) and measurement of photosynthetically active radiation (PAR) reflected from the mulch treatments. PAR reflectance was quantified with a 1.0-m AccuPAR ceptometer (model PAR-80, Decagon Devices, Pullman, WA) and a LI-COR quantum sensor (LI-COR, Lincoln, NB). The ceptometer was placed facing downward within and parallel to the row direction at fruiting zone height, while the quantum sensor was placed above the canopy facing upward to determine the intensity of full sunlight. Reflected PAR was determined as a percentage of full sunlight above the canopy. All radiation measurements were taken on cloudless days between 1100 and 1400 EST. Light measurements were taken in both years for study 1 and at the end of the first year for study 2.
Characterization of the canopy by PQA was performed at 10-cm intervals (Smart and Robinson 1991). PQA was performed midway between fruit set and veraison once in each year of the two-year duration for the three studies: study 1, 21 July 2004 and 19 July 2005; study 2, 14 July 2005 and 12 July 2006; and study 3, 10 July 2006 and 24 July 2007.
Soil temperatures.
Soil temperatures were monitored using Hobo Pro Series 8 dataloggers with thermocouples (Onset Computers, Bourne, MA) in the second year of study 1 in Merlot. A soil probe was buried 2.5 cm beneath the soil surface within each replicate of each treatment. Temperatures were recorded hourly from 29 Mar through 18 May 2005.
Soil sampling.
Soil samples were taken from treatment plots for study 2 in spring 2006, 2007, and 2008 and for study 3 in 2007 and 2008. For each sample, a 30-mm diam by 31-cm long metal soil sampling tube was used to sample to a depth of 15 cm at four random points within the panel. All cores for each plot were mixed together and then air-dried for at least 24 hr. Samples were analyzed at the University of Massachusetts Soil and Tissue Testing Laboratory for pH, cation exchange capacity (CEC), and organic content (Sims and Wolf 1995) and for levels of ammonium (NH4+), nitrate (NO3−), phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), and micronutrients. Cation exchange capacity was determined on the basis of Ca+2, Mg+2, and K+ in Mehlich 3 or Morgan’s soil test solution extractions. Organic content was determined by loss of weight after heating to 450°C. Soil K, P, Ca, and Mg concentrations were analyzed by Morgan extraction and atomic absorption spectrophotometry (Sims and Wolf 1995).
Yield components and fruit composition.
Timing of harvest at all sites was determined by the grower. Yield components (yield/vine, cluster number/vine, cluster weight, berries/cluster, and berry weight) were measured in each year for each study. At harvest, yield per vine (measured with a hanging scale accurate to 0.1 kg; Pro Series, model 71S-018215, Cabela’s, Sydney, NE) and cluster number per vine were recorded. Cluster weight was determined by dividing vine yield by cluster number. A random 12-cluster sample was removed from the total volume of fruit from the harvested panel. Samples were placed into large plastic, resealable bags and frozen in a chest freezer (Revco Technologies, Asheville, NC) at −31°C until processed.
When processed, samples were thawed at room temperature for ~12 hr. Two 100-berry samples were counted and weighed on a XP-300 balance (Denver Instruments, Denver, CO). Berries were then crushed and pressed through a screen-mesh filter to collect at least 100 mL juice. Soluble solids were measured on a Brix scale using a pocket digital refractometer (Fisher Scientific, Agawam, MA). Juice pH was measured with an Accumet AB15 pH meter (Fisher Scientific). Titratable acidity was quantified according to previously published procedures (Amerine and Ough 1988).
Total phenolic, anthocyanin, and flavonol content were quantified spectrophotometrically and expressed as a function of the whole berry, according to a published procedure (Vanden Heuvel et al. 2005). Pruning weights were collected on all vines for study 3 on 23 Mar 2007 and 21 Jan 2008, after balanced pruning by the grower. Yield was divided by pruning weight to calculate cropload ratios.
Statistical analysis.
Data were analyzed with Statistical Analysis System software (v. 9.1; SAS Institute, Cary, NC) using Proc GLM, Proc Mixed, or Proc T-test. Model assumptions were tested through residual analysis and no transformations were needed. Means were separated using Student t tests or Tukey’s honestly significant difference (HSD) test as appropriate at α = 0.05.
Results
Study 1: Fabric (WRM).
PAR reflectance was higher with the WRM when compared with the untreated controls in all cultivars (all p values were lower than p ≤ 0.011) (Table 1⇓). WRM PAR reflectance decreased rapidly as the season progressed. PAR reflected during the second year remained unchanged during the growing season and was still significantly higher than PAR reflectance from the herbicide strip (Table 1⇓).
Canopy densities, as expressed through PQA, were similar for both treatments in both years for Pinot noir and Chardonnay, except for percent interior clusters (PIC) in 2005 (Table 2⇓). Vines with WRM had greater PIC than untreated vines in Chardonnay. For Merlot, leaf layer number (LLN) and percent interior leaves (PIL) were higher for vines treated with WRM compared with untreated vines in both years, and PIC was greater for vines in the WRM treatment in 2005. WRM did not affect budbreak in 2005.
Spring soil temperatures beneath the WRM were generally lower than in the control (Figure 1⇓); 11°C was the maximum differential measured for any hourly recording. For more than half of the days, average daily soil temperatures ranged from 1 to 5°C cooler under the WRM, averaging 1.8°C cooler (daily mean difference) than the untreated control.
Yield components (number clusters/vine, yield/shoot, yield/vine, berry weight, and weight/cluster) were not affected by treatment in any year or by variety (data not shown), except in 2005 when the number of Merlot clusters per vine was higher (p = 0.039) for vines with the WRM (35.1 ± 2.2 clusters; value ± SE) compared with untreated vines (29.4 ± 0.7 clusters), although yield per vine was not significantly affected and node number may not have been constant among treatments because of grower pruning practices. Must composition tests on 2004 and 2005 harvest fruit revealed no differences among treatments in Pinot noir, Merlot, or Chardonnay for Brix, pH, and titratable acidity (data not shown). Total phenolics, total flavonols, and total anthocyanins of Pinot noir and Merlot were also not affected by treatment (data not shown).
Study 2: SRM and QS.
Percentage of PAR reflected into the fruiting zone was higher (p < 0.001) for both SRM (13.4% of ambient flux ± 0.75) and QS (12.1% of ambient flux ± 0.44) treatments compared with the untreated control (4.7% of ambient flux ± 0.15) in the first year of the study. Treatments had no effect on bud survival in the year following application of the SRM and QS (data not shown). Use of SRM or QS did not affect PIC, PIL, or LLN in either year of the study (data not shown).
Calcium concentrations were substantially greater in soil taken from under the QS treatment compared with the SRM and control in all three years (Table 3⇓). Cation exchange capacity (CEC) was greater in the QS treatment compared with the SRM and control in 2007 and 2008. Base saturation levels of K and Mg were ~50% lower in the QS treatment compared with the SRM and control in 2007 and 2008. Other treatment effects included higher sulfur levels and higher percentage base saturation for Ca in soil of QS treatments compared with SRM in 2006.
Yield components (number of clusters per vine, yield per shoot, yield per vine, berry weight, and weight per cluster) were not affected by treatment in this study (data not shown). Total phenolics, total flavonols, and total anthocyanins of Merlot were not affected by treatment (data not shown). Must composition was not affected by treatment in either year, except in 2006 when Brix was higher (p = 0.025) for fruit on QS vines (20.6 ± 0.10; value ± SE) compared with SRM vines (20.0 ± 0.19). Both treatments were similar to the control (20.5 ± 0.14).
Study 3: QS.
Point quadrat analysis indicated that canopy density was not affected by treatment in Cabernet franc. Chancellor vines in the QS treatment had higher LLN than the control (1.57 ± 0.08 vs. 1.18 ± 0.05; p = 0.002) and higher PIL (19.7 ± 2.62 vs. 12.0 ± 2.29; p = 0.056) and PIC (33.6 ± 5.16 vs. 19.5 ± 1.40; p = 0.050), respectively.
Soil nutrient and characterization parameters varied by variety, treatment, and year. In both years, Ca levels were at least twice the concentration for soil under the Chancellor vines that received the QS treatment compared with the control (Table 4⇓); in 2008, Ca levels were twice the concentration for Cabernet franc vines on soils treated with QS. For soils under Chancellor vines, CEC, pH, sulfur, and percent Ca, K, and Mg base saturation were higher in both years for QS-treated soils compared with untreated soils. Soil analyses were more variable for treatments in Cabernet franc. Sulfur concentrations and pH (2008 only), and percent Ca base saturation (both years) were higher for QS-treated soils, while QS-treated soils had lower K in 2007 and lower Mg concentrations in both years.
Most yield components in both years for Chancellor and Cabernet franc were not affected by treatment (data not shown). However in 2006 cluster weights were higher for Cabernet franc vines in the QS treatment (0.132 kg ± 0.005 vs. 0.108 kg ± 0.007; p = 0.031). Chancellor vines receiving the QS treatment had greater yield per vine in 2007 (5.56 kg ± 0.43 vs. 4.21 kg ± 0.19; p =0.021) and higher numbers of clusters (24.7 ± 0.86 vs. 21.1 ± 0.71; p =0.012) compared with control vines. Pruning weights and crop-load ratios (yield/pruning weight) were affected by treatment only in 2007 when the cropload ratio was higher (p = 0.030) for Chancellor vines in the QS treatment (5.88 ± 0.52) compared with control vines (4.24 ± 0.34). Node numbers may have differed among treatments as vines were balance-pruned.
Must composition was affected by QS treatment only in 2006 when Brix was higher (p = 0.029) for Chancellor vines in the QS treatment (20.9 ± 0.47) compared with control vines (19.2 ± 0.23). Concentrations of total anthocyanins, flavonols, and phenolics were affected by QS treatment only in 2006 when total phenolics was higher ( p = 0.016) for Cabernet franc vines in the QS treatment (5.02 ± 0.20 mg/g dw) compared with control vines (4.26 ± 0.15 mg/g dw).
Discussion
These studies tested three different reflective mulches (WRM, SRM, and QS) compared to a control (herbicide strip) on a variety of cultivars. Impacts on measured parameters varied by mulch, and to address these differences, each mulch is discussed separately.
The WRM reflected more PAR into the canopy when compared to the control, particularly in the early portions of the growing season. The WRM we tested reflected considerably more PAR into the canopy than a similar material used by other researchers (Hostetler et al. 2007a), although this variation may be due to differences in physical properties, as the mulch was sold by different manufacturers. While supposedly usable for four full seasons according to the manufacturer, the WRM soiled quickly, resulting in reduced PAR reflectance. The decline in reflectance has been noted previously with similar materials (Hostetler et al. 2007a, 2007b).
The WRM was sold as a product that could be left in place for up to five years. Because of economic and logistical constraints articulated by the vineyard manager, we decided to leave the WRM in the vineyard throughout the entire year and evaluate its durability. If left in place throughout the year, its use over multiple years would result in an inability to hill up and protect graft unions from cold damage. Increased usability of the material would be anticipated if the vineyard manager is willing to apply, remove, and store the WRM each year.
The WRM tested in this study had no measurable impact on the vines with the exception of an increase in canopy density, which did not detrimentally impact fruit composition. While WRM has increased pruning weight over control vines when applied in alleys (Hostetler et al. 2007b), the response has been attributed to suppression of weed competition, which would not have occurred in this study as the controls were herbicide strips. No other studies have reported an impact of any reflective mulch on vine vigor.
Cooler soil temperatures recorded under the WRM treatment could have mixed effects on vine growth and production, although it is unclear how much of an impact the temperature differences noted here (average 1.8°C difference) would have on the vine. Reduced soil warming in the spring might delay budbreak, which could be positive for early budding cultivars in cool climates with respect to reduced risk of frost damage, but negative for later-budding cultivars where fruit ripening could be limited. Future research on reflective mulches should focus on the long-term impact of reduced soil warming on the vine.
SRM is probably the most frequently tested reflective mulch (Coventry et al. 2005, Razungles et al. 1996, Reynolds et al. 2008, Robin et al. 1996), although most authors do not provide manufacturers names, so it is likely these studies are not comparing the same product. The SRM tested in this study also improved PAR reflectance into the canopies when compared to the control (and to the QS mulch). Several logistical issues were associated with the SRM. The aluminized coating deteriorated when left on for the winter, a problem reported by previous researchers (Coventry et al. 2005). SRM became torn and often blew away, resulting in the need for removal and reapplication on an annual basis.
SRM has been shown to impact yield in a wide variety of cultivars (Reynolds et al. 2008); however, we saw no impact of SRM on yield components in this study. We also saw little effect of SRM on fruit composition, in contrast to previous work in this area (Robin et al. 1996, Coventry et al. 2005, Reynolds et al. 2008). It has been suggested that improved flavonols may have reduced insect numbers in the canopy (Coventry et al. 2005), as SRM treatments had fewer aphids and leafhoppers on two sampling dates through the season. Effects of mulches on insect pressure were not assessed in our studies.
Quahog shells are a byproduct of the local seafood industry in coastal New England and can provide a permanent cover under the vines. QS improved reflectance into the canopy when measured prior to harvest in the season of application. While we did not measure PAR reflectance in multiple years from the QS, we noted that the shells were always bright white and did not become soiled like the inorganic mulches. The maintenance of the brightness may be due to the fact that the QS allowed water to permeate the shells while water pooled on the WRM and SRM. Like the WRM, the QS treatments resulted in increased canopy density as indicated by PQA and also improved yield components. These observations are similar to those in other research (Creasy et al. 2006), which investigated the impacts of crushed mussel shells on Pinot noir in New Zealand and which noted increased pruning weights and cluster numbers in the shell treatments.
Although a few improvements in fruit composition were noted in our studies (QS increased Brix in Chancellor and Merlot and improved total phenolics in Cabernet franc), these improvements were not consistent across years or cultivars. In 2007, our grower-cooperator fermented the Cabernet franc fruit from the QS and control treatments separately. Fruit lots were treated identically, but fermentations were not replicated, so results must be interpreted with caution. When the unreplicated wines were scored by a trained Cabernet franc tasting panel (panel training described in Brock 2008), the QS wines were determined to have reduced floral and pomegranate aromas, increased earthy aromas, and reduced cherry and earthy taste compared to the control. However, the New Zealand study reported that mouthfeel of microvinifications from shell plots was rated lower relative to the control for acidity, unripe tannins, and drying tannins, but higher for surface smoothness, heat, texture, and complexity by a group of NZ Pinot noir winemakers (Creasy et al. 2006). In this NZ blind tasting, 29 of the 39 winemakers preferred the wines from the shell plots. A SRM product has also been demonstrated to impact sensory scores positively (Reynolds et al. 2008). Wines from mulched treatments could not be segregated from nonmulched (control) treatments using principal component analysis across red wine cultivars; however, Cabernet franc, Cabernet Sauvignon, Pinot Meunier, and Pinot noir wines from mulched plots all demonstrated reduced vegetable flavors or aromas compared to wines from non-mulched plots. Pinot noir wines from mulched treatments also displayed increased plum aroma, currant flavor, and color compared with control wines, although control wines of Cabernet Sauvignon had higher color intensity compared with wines from mulched plots.
While SRM and WRM will have little impact on soils, the addition of even a single layer of QS significantly increased soil Ca at both sites during the studies, resulting in increased soil pH and a substantially increased Ca:Mg ratio as high as ~19:1 on the Chancellor. The increased Ca did not result in a significant increase in soil pH of the Merlot vineyard, even three years after application. As pH increases, the nutrition program may need to be modified as nutrients such as iron, zinc, manganese, and phosphorus will become less available. The impact of crushed QS on soil properties and fertility warrants further investigation.
Conclusions
Cost of material application and maintenance must be factored into the discussion when considering the use of reflective mulches in producing winegrapes. All mulches examined in this study would likely result in decreased weed competition, at least early in the season. Based on the variable results relating to improved fruit composition, it is unlikely that product costs and labor expenditures could justify the investment in applying reflective mulch to New England vineyards. Additionally, mulches that remain on the vineyard floor for multiple years (e.g., WRM and QS) would result in an inability to hill up to protect from winter damage, so any long-term cost projections that use these materials should include either additional labor for yearly application and removal of the mulch or periodic replanting costs.
Footnotes
Acknowledgments: This material is based on work supported by the Cooperative State Research Extension Education Service, USDA, the Massachusetts Agricultural Experiment Station and the UMass Cranberry Station, under project no. MAS008533. Additional support was provided by Northeast SARE Project Grant LNE04-198 for study 1.
The authors thank vineyard managers P. Nunes and Br. Nate for use of their properties and K. Kumler for donating one of the reflective mulches. We also acknowledge the technical assistance of A. Awad, K. Demoranville, N. DePaulo, K. Ghantous, A. Liberty, M. Salvas, and M. Walsh
- Received August 2008.
- Revision received December 2008.
- Revision received February 2009.
- Accepted February 2009.
- Published online September 2009
- Copyright © 2009 by the American Society for Enology and Viticulture