Grapes 101 is a series of brief articles hilighting the fundamentals of cool climate grape and wine production.
How yield components vary
By Tim Martinson
The 2021 growing season in New York has been a moderately warm one (so far), but one with ample moisture, and frequent rain events. Bloom was about one week early in many parts of NY. And last winter was mild, with negligible amounts of winter injury. These ingredients and some early crop estimates in Western NY, reported by Jennifer Russo, Lake Erie Regional Grape Extension Program, suggest that many growers will experience a heavier-than average crop in 2021.
Concord Crop Estimates. Jennifer’s review of 182 crop estimation blocks (Table 1), showed that half of the 182 blocks surveyed had projected yields of over 9 T/acre, with 19% estimated at 16 Tons/acre and higher.
Table 1. The Lake Erie Regional Grape Program Beltwide Concord Grape Crop Estimation percentage of total samples. Source: NY & PA Lake Erie Regional Grape Program weekly Crop Update, July 22, 2021
For Concords, mid-season crop estimation with mechanical harvesters is an accepted and common practice. It ties in with mechanical pruning to a standard ‘node number’ (often 120 nodes), along with later crop adjustment through mid-summer mechanical crop thinning to prevent overcropping if needed – and adjust to weather conditions while meeting processors’ maturity standards.
For growers of other hybrid and vinifera wine grapes, we don’t have a huge database of crop estimates – but observations tend to point to a large crop this year. For some, it seems like a possible repeat of 2017, when fruitful buds, exceptional fruit set, and ample moisture led to a huge crop – in some cases 30% above what growers had estimated.
So what distinguishes a ‘heavy crop year’ from an ‘average’ or ‘small crop’ year? Part of the answer is how different yield components combine to produce the final crop.
Yield and its components
Total yield is made up of several components:
- The number of vines per acre
- The number of clusters per vine
- How much the clusters weigh
Cluster weight is further composed of two elements:
- The number of berries per cluster
- Berry weight
With the possible exception of vines per acre, which is a fixed number (but there are skips and vine mortality that need to be taken into account), the yield components – clusters/vine, berries/cluster, and berry weight, vary from year to year. So how much do they vary and how much does each component contribute to the final yield?
To examine this, I’ll draw upon data collected over nine years (2008-2016) for Veraison to Harvest, and five years of complete yield data from the “NE1020 coordinated variety trial” at Cornell AgriTech.
1. Range of final berry weights for four varieties from 2008 to 2016. The box plots encompass the range of berry weights observed in four varieties from multiple vineyard blocks across New York. Concord (unsurprisingly) had the heaviest berries (median 3.6 g/berry) and berry weight varied by 1.4 g. The three V. vinifera varieties (also unsurprisingly) had smaller berries (median 1.6 to 1.8 g/berry) and berry weight varied by 0.9 g/berry.
2. How each 0.1 gram/berry translates to yield per vine and per acre. Simple math will show how each 1/10 gram per berry affects yield.
- 6x9 ft planting density = 807 vines per acre
- Vines are managed to 5 shoots/ft of canopy = 30 shoots per vine
- The average number of clusters per shoot = 2, so there are 30 x 2 = 60 clusters per vine
- Berries per cluster = 50 (to use a nice round number)
- Berries per vine = 60 clusters x 50 berries/cluster = 3000 berries/vine
- Change in cluster weight = 0.1 g/berry x 50 berries/cluster = 5 g/cluster
- Change in weight/vine = 5 g/cluster x 60 clusters = 300 g/vine
- Change in crop weight = 300 g/vine x 807 vines/acre = 242,100 g/acre or 242 kg/acre
- Change in tons/acre = 242 kg/acre x 2.24 lb/kg = 542 lb/acre = 0.27 T/acre
So, given these assumptions, each 0.1 gram change in berry weight, is equivalent to a quarter of a ton of yield per acre.
Cluster number, berry number, and berry weight in six years of a variety trial
3. Overall yield components across nine varieties.
From 2010 to 2015, we collected detailed yield data from a variety trial of nine new and standard interspecific hybrid grape varieties. Each year, a crew harvested each vine by hand, counted the number of clusters harvested, and collected a 100-berry sample to obtain berry weights and fruit composition.
Table 1 shows the yearly variation in overall yield (kg/vine) and each of the yield components. The overall yield across all varieties varied by 40-78%. Cluster number (34-84%) was the component that varied the most, followed by cluster weight (22-37%). Breaking down cluster weight into its two components of berry number and berry weight, berries per cluster (15-53%) had a wider range of variability than berry weight (11 to 24%).
This quick and dirty summary appears to indicate that in relative importance, cluster number > berries per cluster > berry weight in contributing to year-to-year variability in yield. Anecdotally, a common rule of thumb is that cluster number-berry weight contribute 60-30-10 % respectively to yield
Table 1. Yield variation 2010-2015 in Coordinated Variety Planting at Cornell AgriTech
4. Examples from 3 different varieties. So if we take the average numbers for berry weight, berries per cluster, and clusters per vine, and just change one of the factors using the observed minimum and maximum measured values, how does that affect overall yield?
In all three cases, changing berry weight (varying by 0.4 g/berry) had the least impact (11-18%) on yield. Berries per cluster (varying by 13/cluster in Chancellor and 37/cluster in Noiret) had the next higher impact (19-40%). Variability in clusters per vine had the greatest impact on predicted yield (37-53%).
Table 2. Effect of varying one yield component at a time on predicted yield per acre in Aromella, Chancellor, and Noiret, from 2010-2015 yield data.
Implications for Crop Estimation
Reliable crop estimates are important to growers and processors, but even rigorous, consistent crop estimation can miss the mark in some years. While one publication suggests that getting within 15% of the true number is an appropriate goal, there are years when estimates are off by 30% or more, despite growers’ or processors’ best efforts. But the ‘snapshots’ from multiyear yield metrics suggest the following:
- Cluster counts are the most important, but not enough for accurate crop estimates. Getting accurate cluster counts will get you most of the way only (60-80 %?) to an accurate crop estimate, but doesn’t capture variability in cluster weight. And it’s difficult to be able to either sample enough vines or to choose the most ‘representative’ panels to accurately represent the variability in a vineyard block. Automated imaging (see RIPE summary, Using cell phones to obtain accurate prebloom cluster counts) offers the prospect of making prediction much more accurate by sampling more vines, but it still can’t track variability in cluster weights.
- Estimates of cluster weight are also needed. Predicting in mid-season how much clusters will weigh at harvest is the other important crop estimation goal. One can collect and weigh clusters at a specific time (for example lag phase, when berries transition from cell division to cell enlargement) and multiply by a factor (commonly x2) to predict the final weight. An alternative to this mid-season estimate is to weigh a sample of clusters at harvest each year and use that information along with cluster counts to arrive at a crop estimate.
- Environmental conditions often intervene, requiring informal modification of estimates (i.e fudge factors). Even when growers have reliable cluster counts and historical cluster weights (or lag-phase estimates) as the basis for crop estimation, events during the growing season can intervene. Drought or disease can reduce final berry weight or lead to ‘shelling’. Significant rainfall after dry periods can unexpectedly swell berry size. Patchy spring bud injury that lowers shoot and cluster count in part of the vineyard can be underestimated. For these reasons, experienced growers often mentally adjust crop estimates based on prior experience and informal ‘fudge factors’.
Grapevine Biology and environmental conditions drive variability
Yield components are determined in part by the grower (how many buds retained at dormant pruning, shoot thinning, cluster thinning), but grapevine physiology both in the previous year and the current season are huge determinants of cluster number, berry number and berry size. Here are some links to previous articles that address the physiology behind yield component variation:
- Bud fruitfulness (the number of clusters per shoot) is determined by weather and sunlight exposure during and after the previous year’s bloom, when buds for next year are initiated (see previous Grapes 101 article Bud Fruitfulness and Yield).
- Floral branching (and flower number) is largely determined by leaf fall the previous year (See Grapes 101 article How Grapevine Flowers Form).
- Fruit set depends on weather during bloom and photosynthetic activity on leaves closest to clusters (cool, rainy = less fruit set). Early leaf removal (around trace bloom) of 5-7 leaves around the fruiting zone reduces fruit set. Removing shoot tips to interrupt shoot growth at bloom increases fruit set (see How Radical Manipulation of Sources and Sinks Affected Riesling Yield, Bud Hardiness, and Return Crop).
For more information
- Bates, T. 2018. Concord Crop Adjustment: Theory, Research, and Practice, posted at Lake Erie Regional Grape Program https://LERGP.com.
- Dami, I. and P. Sabbatini. 2011 Crop Estimation of Grapes. Ohio State University Factsheet HYG-1434-11. Good summary of common methods of crop estimation.
Tim Martinson is senior extension associate with the statewide viticulture extension program, based at Cornell AgriTech in Geneva, NY.