Viticultural Evaluation of Eleven Clones and Two Field Selections of Pinot noir Grown for Production of Sparkling Wine in Los Carneros, California

Materials and Methods

The experiment was conducted over a 3-year period at Gloria Ferrer Champagne Caves located in Sonoma County within the Los Carneros American Viticultural Area (AVA). The site was characterized as a region I with ~1,390 degree days of heat summation (above 10°C). The Pacific Ocean exerts strong influence on the Carneros area, resulting in a cool maritime climate. Early morning fog and afternoon winds are common during the summer months. The fog typically dissipates around midmorning and temperature increases thereafter until cooling winds develop in the early afternoon. The soil type was Clear Lake clay (Miller 1972), pH 6.0 to 7.0, with an effective root depth of 60 to 75 cm.

Eleven clones and two field selections of Pinot noir (Vitis vinifera L.) were evaluated (Table 1⇓). Eight clones were selected from FPS, UC Davis: FPS 1, 2A, 13, 23, 29, 31, 32, and 33. Three clones were obtained from Oregon State University: Espiguette (ESP) 236 and 374 and Colmar (COL) 538. In addition, two field selections of Pinot noir, locally known as Wente (WEN) and Gamay Beaujolais (GB), both provided by Gloria Ferrer, were included in the trial because they were extensively planted in Carneros and elsewhere in northern California for sparkling wine use. The Gamay Beaujolais clone of Pinot noir had been misidentified as a variety for many years in California. It is not related to true Gamay noir of Beaujolais, France. FPS 31 and ESP 236 have the same French clonal identification number, that is, they are, theoretically, the same material. However, because they were obtained from different sources, they were treated as separate clones throughout this study.

Fifty vines of each clone or field selection were grafted onto AXR#1 rootstock (V. vinifera L. Aramon x V. rupestris Ganzin #1), and planted in July 1988. Vines were spaced at 1.5 m x 3.0 m, equivalent to a density of 2,160 vines/ha. Row orientation was east-west. Vines were head-trained, cane-pruned, and trained to a vertical shoot-positioned trellis system. The trellis consisted of a fruiting wire at 90 cm, one wire at 115 cm for protection against wind damage, two pair of movable shoot-positioning wires at 120 and 150 cm, and a single fixed foliage wire at 180 cm. The experiment was designed as a randomized complete block, consisting of 10 vines of each clone per block, replicated five times. The experiment was established as a separate block in the vineyard, and guard vines were employed to separate the data vines from edge effects. Two rows at each side of the vineyard block and six vines at both ends of each row were planted to a single clone.

Vines were pruned by the Gloria Ferrer’s vineyard staff, and the number of buds retained was set in accordance with common practice for the observed vine growth. This corresponded to an average bud count of between 28 and 34 buds per vine over the three years that data were collected. A cover crop of vetch, peas, and oats was planted each fall, with the primary purpose of improving soil organic matter. Interrow spaces were cultivated during the growing season while weeds were controlled under vine rows by application of herbicides. The trial was drip-irrigated, and water was applied according to gypsum block readings. The average irrigation was about 90 L per vine on 3-week intervals, with approximately four irrigations per season. The plot did not experience any major pest or disease problems and very little incidence of Botrytis bunch rot was observed before harvest. No shoot thinning or leaf removal was performed in the plot. All other cultural practices were those of a typical Carneros Pinot noir vineyard for sparkling wine production.

The number of shoots and weight of one-year-old cane prunings was recorded on a per-vine basis in the dormant season. In the fall, each vine was harvested individually and yield and cluster number recorded. Individual vine data were averaged to give a single value per experimental unit for statistical analysis.

Fruit from each clone was harvested on a maturity basis with a target of 19 Brix. Frequent sampling was done to predict fruit maturity. One day before harvest of the first clone, a 100-berry sample was collected per 10-vine experimental unit. Each berry sample was weighed and pressed, and the juice analyzed for soluble solids (Brix), titratable acidity (TA), and pH by a previously described method (Wolpert et al. 19940. Berry weight for 1992 and 1993 are not presented here because in those two years, berry samples were not collected for each experimental unit on the harvest date of each clone. Berry weight would likely have continued to increase for later harvested clones. Comparing berry weight at a given date would not be indicative of final weight, nor could a one-time measure be used to accurately calculate berries per cluster in the normal manner (Wolpert et al. 1994). In 1994, in addition to the same-day berry sample just before harvest of the first clone, a replicated 100-berry sample was taken on the harvest date of each clone. Berry weight and cluster weight on the day of harvest were used to calculate berries per cluster for 1994 data.

Multiyear data were subjected to analysis of variance (ANOVA) using a split-plot model (repeated measures). All possible two-way interactions were included in the model.

Results

Yield components.

Yield averaged across all treatments varied from 9.5 kg/vine in 1992 to 7.7 kg/vine in 1994 (Table 2⇑), a pattern inversely related to changes in average pruning weight. Poor fruit set in 1993, presumably because of inclement weather at bloom, resulted in an experiment-wide average cluster weight of only 108 g, versus 148 g and 123 g in 1992 and 1994, respectively (Table 2⇑).

High-yielding clones averaged 9 to 10 kg/vine and included FPS 23, 31, and 33, ESP 236, and WEN, while the lowest yielding clones were FPS 1, 13, and 29, which averaged between 6.5 and 7.0 kg/vine. Yield differences were due to both clusters per vine and cluster weight. FPS 2A and 33 had greater than 70 clusters per vine while FPS 1, FPS 29, and ESP 374 had less than 60 clusters per vine. FPS 23 and 31 had the heaviest clusters at 147 g and 151 g, respectively, while FPS 13 and 29 had the lightest at 112 g and 115 g, respectively. Mean yield was positively correlated with mean cluster weight (r = 0.79, p < 0.05) (Figure 1⇓). However, vines with similar yields had significantly different cluster weight values, indicating other yield components had significant effect. For example, FPS 2A and 23, with similar 3-year average yields of 8.8 and 9.0 kg/vine, respectively, had the opposite pattern of cluster number and cluster weight, with FPS 2A having more (72) and lighter (122 g) clusters while FPS 23 had fewer (61) and heavier (147 g) clusters.

• Download figure
• Open in new tab

Figure 1

Correlation of cluster weight and yield of 11 clones and two field selections of Pinot noir, Los Carneros, CA. Data points are averages of three years, 1992–1994.

The combination of lower yield and greater vegetative growth resulted in a range of average yield:pruning (Y:P) ratios from a high of 6.2 in 1992 to just over 4 in 1993 and 1994. Individual selections varied widely in average Y:P ratios from a low of 2.7 for FPS 13 to a high of 7.9 for ESP 236. The Y:P ratio was negatively correlated with the average shoot weight (r = −0.83) (Figure 2⇓).

• Download figure
• Open in new tab

Figure 2

Correlation of yield:pruning weight ratio and shoot weight for 11 clones and two field selections of Pinot noir, Los Carneros, CA. Data points are averages of three years, 1992–1994.

Fruit maturity

Before the beginning of harvest, a berry sample established the relative sugar maturity levels of the 13 treatments. Because of its importance as a harvest criterion for sparkling winegrapes, Brix data are presented for individual years as well as for the 3-year mean. FPS 1 was the first clone to reach the target maturity of 19 Brix in all three years. The lowest Brix values were seen in FPS 31 and 33 and ESP 236. The difference between the earliest and latest ripening selections was 3.9 to 4.4 Brix over the three years (Table 3⇓). Mean Brix of the 13 treatments, on the harvest date of the first clone, showed a strong negative correlation to yield (r = −0.85) (Figure 3⇓).

View this table:

• View inline
• View popup
• Download powerpoint

Table 3

Juice composition for 11 Pinot noir clones and two field selections, over three years, 1992 to 1994, Los Carneros, CA.

• Download figure
• Open in new tab

Figure 3

Correlation of yield and Brix on the date of harvest of the first clone(s), Los Carneros, CA. FPS 1 was among the clones harvested first every year. Data points are averages of three years, 1992–1994.

Titratable acidity values were very high as is typical for grapes intended for sparkling wine use. The range was even greater given that the maturity comparisons were made at the very beginning of the harvest period, not at the final harvest sugar levels. There was no correlation between Brix and TA values (data not shown). As expected, early-ripening selections had lower TA levels, such as FPS 1 (18.9 Brix and 14.2 g/L TA), and later-ripening selections had higher TA levels, such as FPS 33 (15.6 Brix and 16.2 g/L TA). However the pattern did not hold for all treatments; for example, WEN had the lowest mean TA value (13.5 g/L) but the third highest mean Brix value (17.6). Values for pH ranged from 2.89 to 3.15 and were normal for maturity levels.

In 1994, berry samples at harvest permit a one-year examination of the contribution of berry weight and berries per cluster to yield differences (Table 4⇓). Berry weight varied by 20% from lowest to highest, with FPS 13 and WEN having the smallest berries (1.03 g and 1.06 g, respectively) and FPS 23 and 31 having the largest berries (1.29 g and 1.24 g, respectively). Numbers of berries per cluster also played a significant role in yield differences, varying by 55% from a low of 89 for FPS 1 to a high of 150 for FPS 31.

View this table:

• View inline
• View popup
• Download powerpoint

Table 4

Yield components of 11 Pinot noir clones and two field selections, data taken at harvest when clones reached 19 Brix, 1994, Los Carneros, CA.

Discussion

Differences among Pinot noir clones and selections were significant, with vegetative growth varying more than two-fold and yield by 1.5-fold (Table 1⇑). Differences were greater than those reported for clones of Chardonnay (Wolpert et al. 1994), Cabernet Sauvignon (Wolpert et al. 1995), and Zinfandel (Wolpert 1996). For some clones, vegetative growth was excessive. Pruning weight of 2.4 to 2.6 kg/vine (1.6 to 1.7 kg/m) for FPS 32 and 13, respectively, exceeded the recommended upper level of 1.0 kg/m (Smart and Robinson 1991). This propensity for strong vegetative growth was exacerbated by low yield, largely because of small clusters. Increasing shoot number to increase yield is not a reasonable strategy because shoot numbers were already high at ~30/vine (20/m). A more appropriate strategy for balancing small-clustered Pinot noir clones would be to reduce vine growth, such as by selecting a less vigorous rootstock. The goal would be to decrease shoot weight from excessive levels, 75 to 89 g for FPS 1, 13, and 32, to moderate levels 50 to 60 g (Figure 3⇑).

Conversely, ESP 236 had the least growth (1.2 kg/vine, 0.8 kg/m) and high yield (9.1 kg/vine), resulting in the highest Y:P ratio at 7.9. While the Y:P ratio is not high enough to cause concern, the resulting shoot weight of 45 g was the lowest of the clones and lower than recommended for adequate performance (Smart and Robinson 1991). One strategy to ensure shoot weight for high-yielding clones would be to reduce shoot number, but that would reduce yield per hectare, unless vine spacing was concomitantly reduced (Intrieri and Filippetti 2000).

As with other varieties in California (Wolpert 1996, Wolpert et al. 1994, 1995), and in Pinot noir in Oregon (Watson et al. 1988), differences in yield among Pinot noir selections were largely due to cluster weight (Table 2⇑). However, more so than for other varieties referenced, yield of Pinot noir was affected by cluster number (Table 2⇑), although it is not clear whether this difference was due to fruitfulness of nodes retained, as a function of clone, or differences in number of blind buds. These differences in crop load were significantly correlated with maturity level as evidenced by Brix values at harvest, for data averaged over three years (Figure 3⇑). This well-known relationship of yield and maturity suggests that clones are not “early ripening” or “late ripening,” as they are referred to in popular literature, but rather low yielding and high yielding.

However, performance of two clones in the Carneros study can be discussed relative to previous research. The German experience with Mariafeld as a clone with large crop and large berries (Becker et al. 1988) is similar to the performance of FPS 23, also known as Mariafeld (Tables 2⇑ and 4⇑). Characterization of ESP 236 as fruitful with a tendency to overcrop (Bernard et al. 1983) was supported in our study, as ESP 236 had the highest Y:P ratio.

Another confounding factor that often calls the comparison of clone performances into question is the inability to determine clone identity via DNA analysis (C. Meredith, personal communication, 1998). Thus, in the absence of a unique phenotypical characteristic that identifies a clone, “trueness to type” is entirely dependent on keeping accurate records during importation and propagation. The performance data of FPS 31 (reported to be Espiguette 236) and the Oregon State University selection of Espiguette 236 (Table 1⇑) would indicate that the two clones significantly differ in pruning weight, shoot weight, yield, Y:P ratio, cluster weight, and, in 1994, berry weight. One could reasonably conclude that they are not the same clone. This example highlights the importance of developing confidence of the clonal source and the fidelity of recordkeeping during propagation.

Conclusion

This study is the first report of performance of sparkling wine clones of French origin (FPS 31, 32, and 33, ESP 236 and 374, and COL 538) in California. Their performance indicates some have potential for high yield and others for excessive vegetative growth. Viticultural performance data should be helpful to viticulturists responsible for making rootstock, spacing, and trellis choices for vineyards intended for sparkling wine production.