Fertile Virginia soils can sometimes lead to vigorous vine growth and high crop loads. Viticulturists commonly reduce crop load by pruning and cluster thinning to decrease shading and disease pressure, increase sunlight exposure and encourage ripening. Crop reduction is also thought to improve overall vine balance and prevent stress that can lead to lower fruitfulness in subsequent years (1). However, crop reduction comes at measurable cost. In her landmark crop load study in Oregon, where participants were routinely cropping at 2-2.75 tons/acre, Patty Skinis estimated that a 25-50% crop reduction came a cost of $700-800 per acre. In Virginia, the commercial cost of reducing premium Merlot crop from 4.5 to 3 tons/acre (as done in the Trump Crop Load trial) averages over $3750 (2). The question then arises: how much crop should be dropped to achieve ripeness and quality? Are we thinning too much or not enough? What is the right balance, for the vine as well as the bank account?

Previous studies examining the effect of altered crop load in different regions with different grape varieties have shown markedly mixed results. Some studies (3–5) find important differences with crop reduction including faster maturity, better color, higher anthocyanins, and higher wine quality scores while others find little to no differences at all (1,6,7). Some even report decreases in wine quality with lower crop level as negative flavor and aromas such as methoxypyrazine are concentrated (8) The attached report summarizes outcomes from many crop load studies. Interpretation of these studies is complicated by the differences in growing region, grape variety, and whether grapes were harvested on the same day or at the same level of maturity. Moreno Luna et al (2017)(1) studied both crop level and harvest date and conclude that differences between crop load treatments could be overcome with longer hang time, however in some regions (like Virginia), longer time includes greater risk of precipitation and disease. The method of crop management may also matter, with pruning more effective than crop thinning (8). Some studies find few differences until very high crop loads are reached (7,8). Many studies report vine phenology and grape chemistry but do not include examination of flavor and aroma of wine, either chemically or by sensory analysis. 

Figure 2: Lower crop load leads to greater accumulation of total glycosides over time. From Zoecklein (2000)

In a presentation to the Virginia Vineyards Association in 2000, Dr. Bruce Zoecklein of Virginia Tech presented data examining the relationship between yield and wine quality in Chardonnay and Cabernet Sauvignon in Winchester, Virginia (9). In addition to Brix, pH and TA, these studies included a measure of phenol-free glycosides, the pool of secondary metabolites that will be transformed into flavor and aroma in the wine. In Chardonnay cropped to 3.7, 7.3, and 10.2 tons/acre on an open lyre system, higher crop load led to decreased Brix with little change in pH or TA. The lower crop load treatment also had significantly higher total glycosides, which were correlated with wine quality through sensory analysis. When total glycosides were measured at different Brix levels during ripening, they were relatively the same for each crop load relative to Brix until grapes accumulated 20.5°Brix, at which point the lower crop load (4.4 tons/acre in this case) had notably higher accumulation of these metabolites. This relationship was shown to continue to diverge as Brix increased to 24.5°Brix. A similar study of in Cabernet Sauvignon harvested at 20.5°Brix showed differences in phenol-free glycosides between crop loads of 7.4 tons/acre and 10.4 tons/acre with the lower crop load again accumulating a larger pool of potential flavor/aroma compounds. In both Chardonnay and Cabernet Sauvignon, wines with higher phenol-free glycosides were perceived to be different in sensory analysis. These results underscore the fact that general grape chemistry (Brix, pH, TA) does not always fully describe the quality potential of the grapes.

When taken as a whole, the literature suggests that for any given situation, there is a range of crop load that will produce comparable quality wine, with lower quality wine produced when crop load is too high or too low. The most cost effective management strategy, then, would be to maintain the vineyard on the higher end of the range of comparable quality. Unfortunately, given the large ranges examined in the literature with differing results, there is no clear formula to determine the right amount of crop to grow on a given site. These values depend on factors such as vine vigor, leaf area, vintage, variety, water availability, length of the growing season and many others (3,4,7,8) In concluding his presentation on crop load, Zoecklein said

The relationships between yield and quality are complex...There may be no exact point when overcropping begins (9).

 The best approach to determine the right crop load for a given vineyard then is by practical experimentation.

...despite over 50 years of research investigating cluster thinning, crop size, and yield effects on vine physiology, vine size, yield components, berry composition, and wine composition and quality, there are many contradictions in the literature due to the cultivars chosen for experimentation and the circumstances under which each trial was carried out. (Luna et al, 2017)

The Oregon Statewide Crop Load Study

An impressive example of how to set up a practical experiment examining the effect of crop load on fruit and wine quality comes from the Statewide Crop Load Study conducted by Dr. Patty Skinkis of Oregon State University (7). This ongoing 10-year study (currently in year 8) has included 24 different companies across 6 AVA’s examining the effect of crop load on Oregon Pinot Noir. At each site, a randomized complete block design was used to test the effect of two or more crop levels determined by the practitioner, with three replicates per site. Crop levels were imposed as cluster thinning to 0.5, 1, 1.5, and 2 clusters per shoot as well as no reduction in clusters per shoot. Standardized sampling protocols from 10-vine sections were followed by all participants with data collected for vine phenology, yield and fruit composition (7).  Data collected so far indicate that there is considerable difference in yield per treatment each year, but differences in Brix have been seen only in high crop years. Few differences have been observed for pH, TA, or total phenolics based on yield (7). 

 

Figure 3 Anthocyanins across 15 sites in the Oregon Statewide Crop Load Study, from Skinkis 2017

The most consistent differences have been seen with 10-40% higher levels of anthocyanins in lower crop treatments in high crop load years, but even these are not seen at most sites. This change in anthocyanins might be due to physiological causes or those due to microclimate (fewer overlapping clusters). Quercetin glycoside, a phenolic molecule used by the grape as sunscreen, also differs with anthocyanins, indicating the differences in anthocyanin concentration may be due to shading more than vine physiology. Preliminary conclusions to this study indicate that the site and vintage have a higher impact on quality than yield, and that Oregon growers may be able to sustain higher crop loads than the current standard of 2-2.7 tons/acre with little loss in quality (7,10–12). 

The strengths of the Oregon study are many:

• The randomized complete block design accounts for vineyard variation and allows for more efficient statistical analysis.
• A standardized experimental design and sampling protocol allows for a broader application of the experimental results beyond a single vineyard.
• The practicality of the sampling regime (a 10-vine area) is not too complicated for practical application.

Virginia has higher vigor than Oregon, indicating that the trends seen during “high yield” years in Oregon may be seen more consistently in Virginia. The Skinkis study provides a roadmap for experimental design that can be applied to Virginia vineyards with potential impacts on grape and wine quality as well as cost of wine production. The crop load study done at Trump Winery in 2019 followed an adapted version of this protocol.

---

 References

(1) Moreno Luna, L. H.; Reynolds, A. G.; Di Profio, F. Crop Level and Harvest Date Impact Composition of Four Ontario Winegrape Cultivars. I. Yield, Fruit, and Wine Composition. American Journal of Enology and Viticulture 2017, 68 (4), 431–446. 

(2) SMS Research Advisors. 2019 Virginia Commercial Grape Report. 2020.

(3) Bravdo, B.; Hepner, Y.; Loinger, C.; Cohen, S.; Tabacman, H. Effect of Crop Level and Crop Load on Growth, Yield, Must and Wine Composition, and Quality of Cabernet Sauvignon. Am J Enol Vitic. 1985, 36 (2), 125–131.

(4) Wolf, T. K. Wine Grape Production Guide for Eastern North America; Plant and Life Sciences Publishing: Ithaca, New York, 2008.

(5) Sinton, T. H.; Ough, C. S.; Kissler, J. J.; Kasimatis, A. N. Grape Juice Indicators for Prediction of Potential Wine Quality. I. Relationship Between Crop Level, Juice and Wine Composition, and Wine Sensory Ratings and Scores. Am J Enol Vitic. 1978, 29 (4), 267–271.

(6) Matthews, M. A. Terroir and Other Myths of Winegrowing; University of California Press: Oakland, California, 2015.

(7) Skinkis, P. The Low Down on High Yields: Challenging Yield-Quality Standards for Oregon Pinot Noir, 2017.

(8) Chapman, D. M.; Matthews, M. A.; Guinard, J.-X. Sensory Attributes of Cabernet Sauvignon Wines Made from Vines with Different Crop Yields. Am J Enol Vitic. 2004, 55 (4), 325–334.

(9) Zoecklein, B. W. Update of Ongoing Projects, 2000.

(10) Skinkis, P. Pinot Noir Crop Load Wine Technical Tasting Challenges the Yield-Quality Relationship. Oregon State University Vine to Wine 2020.

(11) Skinkis, P.; Schreiner, R. P. Defining Crop Load Metrics for Quality Pinot Noir Production in Oregon; Unified Grant Management for Viticulture and Enology Interim Report Summary; 2016.

(12)  Uzes, D. M.; Skinkis, P. A. Factors Influencing Yield Management of Pinot Noir Vineyards in Oregon. Journal of Extension 2016, 54 (3).

Vineyards are variable places. We think about this when it is time to select a site, decide which varieties get planted where, how to fertilize or what (if anything) to grow between the rows. It is also crucial to consider the variation in the vineyard when setting up a vineyard trial or doing fruit sampling. Data analysis and interpretation are only as accurate as good experimental design allows, so it is crucial to consider the scale of each treatment and the physical and temporal plan for sampling before starting. 

The main issue when setting up a vineyard trial is minimizing the differences  between treatemens that are due to environmental effects in order to be able to recognize any differences that are due to the treatment itself (1). The solution to this challenge is replication, both in the treatment regime and in the sampling regime. If the treatment is applied in multiple places in the vineyard and shows the same effect each time, it is more likely that effect is due to treatment than simply because it occurred in a particular location in the vineyard. 

Figure 1 from The Low Down on High Yields, Skinkis, 2017

The randomized complete block design is a common method for setting up trials in the vineyard. Three main components contribute to the strength of this approach: blocking, randomization and replication (1).

A “block” is a defined, relatively homogenous region of the vineyard that is subdivided to receive each level of treatment. Several blocks are identified within the vineyard as replicates, with each block receiving each treatment level. Within each block which region receives which treatment is assigned randomly. Randomization reduces the bias that would be introduced by having a pattern of assignment. 

When setting up the blocks, it is important to take the overall layout of the vineyard into consideration. The size and location of each block is determined by the level of environmental heterogeneity. Heterogeneity within the block should be very low while between block heterogeneity can be large. For example, a vineyard located on a slope should have a block that contains each treatment on the top and another block at the bottom of that slope rather than treatment at the top and control at the bottom. Figure 3.4 from Design and analysis of Ecological Experiments(1) provides a good visualization of these issues. 

Figure 2 from Design and Analysis of Ecological Experiments, Scheiner, 1993

Another level of variation occurs during sampling. Any time a vineyard is sampled, there are a range of values that could found based on the variation that is present in that vineyard. Within a single grape cluster there can be berries with a wide range of sugar values, so it is unlikely a single berry can be chosen to represent the whole cluster. To narrow the value as close to the actual value as possible, good sampling must be practiced. This includes taking care to collect a representative sample, avoiding the biases of picking the most obvious or most attractive fruit. Many guides are available to ensure samples come from different locations on the vine and different locations on the cluster (2–4). However, sampling can also be in error due to environmental variation. Replicated sampling within a block (at the near, middle, and far portion of a row, for example) can help tease out variation due to location vs. variation due to treatment level. A key element to keep in mind while sampling for a vineyard trial is that the sample is not meant to determine the overall level of any metric for the whole vineyard, but rather it is meant to minimize environmental variation to determine differences due the treatment level. Individual sampling at multiple places sum together to determine if these changes are consistent throughout the vineyard. 

By carefully selecting treatment blocks placed at multiple sites within the vineyard, and collecting replicate samples within a block, statistical analysis can then be used to separate  which portion of the variation is due to the block location, the treatment level, and just random chance (1). Setting up the vineyard experiment with these elements allows much more clear interpretation of the data and a much better chance of accurately identifying treatment effects.

---

References:

(1) Design and Analysis of Ecological Experiments; Scheiner, S. M., Ed.; Chapman and Hall, Inc: New York, 1993.

(2) Zoecklein, B. W. Grape Sampling and Maturity Evaluation for Growers https://www.apps.fst.vt.edu/extension/enology/VC/Jan-Feb01.html (accessed Oct 30, 2019).

(3) Wolf, T. K. Wine Grape Produciton Guide for Eastern North America; Plant and Life Sciences Publishing: Ithaca, New York, 2008.

(4) Amerine, M.; Rossler, E. B. Field Testing of Grape Maturity. Hilgardia 28 (4), 93–114.

In 2019, the team at Trump Winery conducted a study of the Effect of Merlot Crop Load on Grape and Wine Quality. With the assistance of Dr. Tim Jordan, they set up replicated blocks within the vineyard and took an impressive amount of data to try to understand if wine quality in this Merlot vineyard is affected by altering crop load from 4.5 to 3 tons/acre. On April 30, 2020 we met virtually to discuss this study and to collect sensory data from our remote participants. 

Watch Video Here

Summary

Common inherited wisdom in viticulture says that smaller crop loads lead to better wine. However, most of the published reports on crop load were done in widely varying ranges (from 4-12 ton/acre) and in regions other than Virginia. Crop thinning as little as 1.5 tons/acre comes at considerable cost. In this study, crop thinning to a target of 3 vs. 4.5 tons/acre was done on the Merlot block of the Chateau vineyard of Trump Winery in a randomized complete block design. This resulted in significantly different crop load, with no significant difference in ripening parameters (Brix, pH, TA) or berry weight. Grape and wine color metrics were higher for the low crop treatment, however these were not significantly different in sensory analysis. Virginia has a wide range of variation between vintages, and vines may be able to carry more crop in warm, dry vintages such as 2019 while the results maybe different in cool, wet vintages such as 2018. Plans are to repeat this study in subsequent vintages.

Introduction

Common inherited wisdom in viticulture says that smaller crop loads lead to better wine^1,2. Smaller crop loads are often achieved by pruning, cluster thinning, or deficit irrigation¹ and are credited with hastening ripening, increased fruit quality, decreased disease (especially Botrytis), and better long term vine balance^1–3. However, crop reduction comes at measurable cost. In her landmark crop load study in Oregon, Patty Skinkis estimated that a 25-50% crop reduction came a cost of $700-800 per acre. Participants in the Oregon study were cropping at 2-2.75 tons/acre on average. In this study, reducing premium Merlot crop from 4.5 to 3 tons/acre can be estimated to forego $3600⁴. 

Previous studies examining the effect of altered crop load in different regions with different grape varieties have shown markedly mixed results. Some have found increases in sugar accumulation, color, and anthocyanins with lower crop load while others have found no difference, or even higher levels of unwanted characteristics such as vegetal/green pepper character^1,3,5–8. Positive effects of crop load reduction seem to be more pronounced in situations of over-cropping, such as hybrids with large cluster sizes, in low vigor sites or varieties, or in overall high crop years^2,3. 

In her landmark Statewide Crop Load Study in Oregon, Dr. Patty Skinkis of Oregon State University has coordinated study of crop load in Pinot Noir over 8 year (with two years remaining in the study) including 24 different sites across 6 AVA’s. Using standardized methodology in production scale vineyards, the study has found few consistent differences in Brix, pH, or TA. The most consistent differences have been seen with 10-40% higher levels of anthocyanins in lower crop treatments in high crop load years, but even these are not seen at most sites. Preliminary conclusions to this study indicate that site and vintage have a higher impact on quality than yield, and that Oregon growers may be able to sustain higher crop loads than the current standard of 2-2.7 t/acre with little loss in quality^2,9–11. 

Virginia has higher vigor and crop load than Oregon, indicating trends seen during “high yield” years in Oregon may be more seen more consistently in Virginia. The experimental design of Dr. Skinkis study in Oregon was adapted here to explore the effect of crop load on a premium Merlot vineyard at Trump Winery that is prone to high yields (4-5 tons/acre without adjustments) and non-uniform ripening. After cluster thinning, fruit maturity, yield, phenolics and wine chemistry were measured. 

Methods

Vineyard

The experiment was conducted at Trump Winery, near Charlottesville, Virginia, in a 10.77 acre block of Merlot. Planted in 2004, the Merlot is clone 181 on 3309C rootstock. The vineyard design is 7 ft by 3.33 ft spacing on a VSP trellis, and the vines are trained to a unilateral cordon. The site is situated at 1000-ft above sea level, on a 10-18° slope oriented 30°, east-southeast. 

A randomized complete block design with longitudinal measures was used to determine how cropping level affected disease incidence and ripening kinetics over the three to four weeks preceding harvest. Crop level treatments were applied at berry touch using lag phase cluster sampling to determine the balance needed to achieve desired yield. On June 29, cluster thinning was done at two treatment levels with large wings and stacked bunches also removed:

• 2 clusters/shoot (approximately 4 clusters per linear foot) to a target of 4.5 tons/acre 
• 1.5 clusters/shoot (alternating 1 cluster/shoot and 2 clusters/shoot)(approximately 3 clusters per linear foot) to a target of 3 tons/acre 

Block and treatment assignment:

There were four blocks identified within the vineyard, each containing 9 rows. Treatment position within each block was assigned randomly with four consecutive rows cluster thinned to two shoots per cluster while five consecutive rows were thinned to 1.5 shoots per cluster. The difference in number of rows was to ensure proper yields of each treatment for winemaking. Buffer rows bounded the experimental block on the top and bottom of the vineyard so that none of the treatment rows were exposed to edge effects except at the end of the rows. For each treatment, sampling locations were assigned for the inner rows only, again to act as a buffer against potential edge effects of the treatment. Figure 1 shows a map of blocks and treatments. 

Sampling

Randomly selected, paired samples for each block were determined for each sampling event. A sample location consisted of three consecutive panels of six mature vines per panel. If a panel did not have six mature vines, the next full panel was used. An effort was made to prevent sampling in a previously sampled location by flagging the beginning and end of a three-panel sample. 

Disease incidence: 

Disease incidence was measured using a visual count of clusters per vine. Only the first three vines of every panel were sampled. A tally was recorded of total clusters per vine and number of diseased clusters per vine. A diseased cluster included any indication of rot (Botrytis, sour, ripe, black rot, mildew) caused by any damage factor (insect, animal, disease, environmental, mechanical).  Dividing diseased clusters by total clusters per vine provided a proportion of diseased clusters per vine.

Ripening kinetics

Ripening was measured using weekly 100-count berry samples with three paired samples per treatment per block from the near, middle, and far ends of the rows. Sampling commenced at 18 °Brix and was done weekly up until harvest. The final berry sample was taken on the day of harvest. Each berry sample included selecting six berries per vine across all three panels (18 vines) for a total of 108 berries. The six berries were selected at random from six different clusters per vine by systematically sampling the top, middle, and bottom of clusters on the east and west faces of clusters. Weekly vineyard samples were processed and tested for Brix, pH, and TA.  At harvest, pooled samples from three treatment blocks were also sent to ETS labs (St. Helena, California) for phenolic testing. 

Yield

Yield was measured at harvest by the removal of six clusters from three locations per treatment (at the near, middle, and far ends of the rows). Cluster selection was random by removing no more than one cluster per vine and a total of two clusters per panel. An effort was made to remove two clusters on each side of the trellis panel (east, west). Clusters were weighed individually and as a group to record individual and average cluster weights. Each cluster was also broken down into individual berries to count the number of berries per cluster, and to weigh individual berries per cluster to record number of berries per cluster and average berry weights per cluster.  

Winery methods

Vineyard replicates were combined for winemaking. All winemaking operations were the same between lots. All blocks were harvested on the same day and refrigerated overnight prior to processing. Grapes were crushed and destemmed with addition of 50 ppm SO₂, 3 g/100kg Laffort Grand Cru Lafasse enzyme, and Laffort Tannin VR Supra into close-topped, temperature controlled stainless steel fermenters. Must was allowed to cold soak for 4 days at 40°F. Tartaric acid (3 g/L) was added during cold soak. At the end of cold soak, tanks were inoculated with 15 g/hL Laffort F15 rehydrated in 15 g/hL GoFerm. Fermentations were monitored daily for Brix and temperature with the addition of 25 g/hL Fermaid K at 1/3 sugar depletion. Fermentations were punched down twice per day and pumped over once per day. After the completion of fermentation, wine was drained and pressed, then inoculated with CHR Hansen Viniflora Oenos prior to racking to barrels for malolactic fermentation. Malic acid conversion was confirmed with enzymatic analysis prior to racking and addition of 50 ppm SO₂. Wine was monitored monthly for SO₂ and VA.

Statistical analysis of vineyard results

Analysis of variance (ANOVA) was used to test if between-subject (cropping level) sample variances were homogeneous. Within subject factors were blocks and time (week). A paired t-test was used to determine if sample means differed. An alpha = 0.05 was used to assess statistical difference.

Sensory analysis was completed by a panel of ­­­26 wine producers. Due to social distancing restrictions at the time of COVID-19, wines were shipped to panelists in randomly numbered sample bottles. Tasters were presented with three wines, two of one type and one of another, and asked to identify which wine was different (a triangle test). There were three tasting groups with the unique wine in the triangle test balanced between groups. Tasters were then asked to score each wine on a scale of 0 to 10 for aromatic intensity, fruit intensity, perception of ripeness, color and structure. They were also given open ended questions to describe the wines. Results for the triangle test were analyzed using a one-tailed Z test. Descriptive scores were analyzed using repeated measures ANOVA.

Results

Cluster removal successfully differentiated crop levels. At the time of harvest, the high crop averaged 4.09 clusters/foot with 4.12 lbs/foot while the low crop averaged 2.68 clusters/foot with 1.33 lbs/foot. These values were significantly different (t=13.6, p<0.0001)(Figure 2). There was no significant different between treatments in sugar accumulation (Brix), pH increase, or TA decrease (Figure 3). Differences in crop level also did not affect mean berry weight (Figure 4). Phenolic measurements in grapes at harvest were lower in the high crop treatment than the low crop treatment (Table 1) with notably lower anthocyanins and tannins in the high crop treatment. 

Vineyard replicates were pooled for winemaking. There was no difference in juice chemistry or basic wine chemistry between treatments (Tables 2&3). Fermentation was robust and progressed at the same rate between the two treatments (Figure 5). Wine from the low crop treatment had notably higher color and anthocyanin content than wine from the high crop treatment (Figure 6, Table 4). Wine from the low crop treatment also had higher tannin, though the other phenolics were largely the same between treatments (Table 5).                 

In a triangle test, 9 out of 26 respondents were able to distinguish the wines, indicating the wines were not significantly different. The wine from low crop load had significantly higher scores for color than the wine from high crop load (Table 6), indicating differences found in lab metrics were also visually perceptible. When asked what distinguished the wines in the triangle test, several respondents cited color. Open ended questions indicate that the wines were not very different. During discussion, many participants indicated the difference between the wine was small enough they did not feel it was worth the cost. Others felt that, for some wines, the cost was worth the increase in quality.

Figure 1: (A) The Chateau Vineyard at Trump Winery was divided into four blocks. (B) Each block was further subdivided into two treatments; high crop (pink) and low crop (blue) with high crop treatment receiving four consecutive rows and low crop treatment receiving five consecutive rows. Treatments were assigned randomly within the block. At each sampling event, each treatment was sampled three times, at the near, middle and far ends of the rows. (C) Image of high crop treatment taken on July 31,2020. (D) Image of low crop treatment taken on July 31, 2019)

                                                                                           

Figure 2: Yield from high and low crop treatments was significantly different at harvest. 

Figure 3: There were no significant differences in ripening parameters between treatments

Figure 4: There were no significant differences in berry weight between treatments at harvest.

Table 1: Phenolic compounds found in grape samples (mg/L) (ETS Labs)

Figure 5: Fermentation kinetics for high and low crop Merlot (In-house data)

Table 2: Juice Chemistry for high and low crop Merlot (in-house data)

Table 3: Chemistry for wine from high and low crop Merlot (ICV Labs)

Figure 6: Color Metrics for wine from high and low crop Merlot (ICV labs)

Table 4: Anthocyanins in wine from high and low crop Merlot (mg/L) (ETS Labs)

Table 5: Phenolics in wine made from two treatments of Merlot (mg/L)(ETS Labs)

Table 6: Sensory impressions of two treatments of Merlot using repeated measures ANOVA

---

References

(1) Matthews, M. A. Terroir and Other Myths of Winegrowing; University of California Press: Oakland, California, 2015.

(2) Skinkis, P. The Low Down on High Yields: Challenging Yield-Quality Standards for Oregon Pinot Noir, 2017.

(3) Moreno Luna, L. H.; Reynolds, A. G.; Di Profio, F. Crop Level and Harvest Date Impact Composition of Four Ontario Winegrape Cultivars. I. Yield, Fruit, and Wine Composition. American Journal of Enology and Viticulture 2017, 68 (4), 431–446. 

(4) Wood, V.; Custer, S.; Watson, K.; Alper, D. Virginia 2018 Commercial Grape Report. 11.

(5) Wolf, T. K.; Dry, P. R.; Iland, P. G.; Botting, D.; Dick, J.; Kennedy, U.; Ristic, R. Response of Shiraz Grapevines to Five Different Training Systems in the Barossa Valley, Australia. Australian Journal of Grape and Wine Research2003, 9 (2), 82–95. 

(6) Sinton, T. H.; Ough, C. S.; Kissler, J. J.; Kasimatis, A. N. Grape Juice Indicators for Prediction of Potential Wine Quality. I. Relationship Between Crop Level, Juice and Wine Composition, and Wine Sensory Ratings and Scores. Am J Enol Vitic. 1978, 29 (4), 267–271.

(7) Chapman, D. M.; Matthews, M. A.; Guinard, J.-X. Sensory Attributes of Cabernet Sauvignon Wines Made from Vines with Different Crop Yields. Am J Enol Vitic. 2004, 55 (4), 325–334.

(8) Bravdo, B.; Hepner, Y.; Loinger, C.; Cohen, S.; Tabacman, H. Effect of Crop Level and Crop Load on Growth, Yield, Must and Wine Composition, and Quality of Cabernet Sauvignon. Am J Enol Vitic. 1985, 36 (2), 125–131.

(9) Skinkis, P. Pinot Noir Crop Load Wine Technical Tasting Challenges the Yield-Quality Relationship. Oregon State University Vine to Wine 2020.

(10) Skinkis, P.; Schreiner, R. P. Defining Crop Load Metrics for Quality Pinot Noir Production in Oregon; Unified Grant Management for Viticulture and Enology Interim Report Summary; 2016.

(11) Uzes, D. M.; Skinkis, P. A. Factors Influencing Yield Management of Pinot Noir Vineyards in Oregon. Journal of Extension 2016, 54 (3).

---

Appendix A: Example sample grid for one block, one week. Four blocks were sampled per week.