Accurate crop estimation is critical for planning in both vineyards and wineries.
Viticulturists require early season estimates to make decisions on canopy management, crop reduction and vine balance in order to optimize tonnages to maximize quality and productivity. Wineries rely on the crop estimate to ensure the grape intake requirements are sufficient and there is allowable tank space and to order adequate enological supplies for the upcoming vintage.
It is difficult to accurately assess the crop load as environmental factors, management practices and grapevine physiology all affect yield. Often, crop estimates vary from the actual harvest up to and exceeding 30%. In general, a good estimate will not vary by more than 10%.
In order to provide a reasonable crop estimate, the number of vines per block, the average number of bunches per vine and the bunch mass at harvest must be determined or predicted. Taking the time to properly estimate the potential crop load should result in an estimate that is accurate.
In order to estimate the potential crop in a block, the total number of vines per block must be determined. Most vineyards already have an accurate count of the number of vines per acre and block.
If there is no record of plants per block, an accurate vine count is required. For most vineyards the number of vines per block is calculated based on planting density. Density is determined by multiplying the distance between vines (vine spacing) by the distance between rows (row spacing). For example:
| Vine spacing (ft) |
Row spacing (ft) |
Vines per acre |
Vine spacing (m) |
Row spacing (m) |
Vines per hectare |
|---|---|---|---|---|---|
| 3 | 7 | 2,074 | 0.9 | 2.1 | 5,126 |
| 4 | 7 | 1,556 | 1.2 | 2.1 | 3,844 |
| 3 | 8 | 1,815 | 0.9 | 2.4 | 4,485 |
| 4 | 8 | 1,361 | 1.2 | 2.4 | 3,364 |
| 5 | 8 | 1,089 | 1.5 | 2.4 | 2,691 |
| 4 | 9 | 1,210 | 1.2 | 2.7 | 2,990 |
| 5 | 9 | 968 | 1.5 | 2.7 | 2,392 |
| 6 | 9 | 807 | 1.8 | 2.7 | 1,993 |
Most vineyard blocks will have a slightly lower number of vines per acre than the calculated number due to missing vines from disease, winter injury, mechanical damage, replants, etc. These missing vines may cause an error in crop estimation if not accounted for.
One method of accounting for the missing vines is to count the total number of missing vines in each block and subtract from the total plants per block. This requires spending the time to count all missing vines in every block in the vineyard.
Another method is to include missing vines when counting the number of bunches per vine (see below). In this situation, record the plant as present but with 0 bunches per vine. This eliminates the time required to count all missing vines in each block.
Bunch counting can be done during early shoot growth once the inflorescences are visible. It is best to count bunches as early as possible as they are much easier to see at this stage and it takes less time. Secondary bunches are not visible at this time.
It is more difficult to see bunches once the canopy is fully grown as leaves tend to obstruct one's view. If counting bunches later in the season, it is important to count only primary bunches found on the primary and secondary shoots. The secondary bunches on lateral shoots should be ignored as they are not harvested.
Determining the number of bunches to count to get an accurate estimation of a particular block is dependent upon several factors. These include:
The accuracy of the estimate increases with the number of vines counted per block. The sample size should be determined by considering what level of variance is acceptable and how much time is to be allocated towards completing the crop estimation.
In general, on very small blocks (less than 1.5 ha or 3 acres in size), 25-40 vines should be counted. On large blocks (2-4 ha or 5-10 acres), a minimum of 100 vines should be counted. For practical purposes, it works well to count 5-10 vines in a row every 6-10 rows.
Randomly but uniformly select sections of vines for counting over the entire block. If there are sections with missing vines that have not been accounted for, record the vines as 0 bunches per vine. If there is a section of a particular block that varies from the rest of the block, such as from frost damage or any other factor, it should be estimated separately.
The most difficult process in crop estimation is predicting the harvest bunch weight.
The harvest bunch weight changes each year as it is influenced by many factors, including vine physiology, vine nutrition status, vine water status, pest and disease pressure, environmental conditions and vineyard management practices. Three commonly used methods of predicting bunch weight are:
Each method has benefits and drawbacks. Historic bunch weight estimates provide the first estimate in the growing season. The berry number and lag phase estimates account for the annual variation in bunch weights. A combined use of these methods over the growing season allows for early crop estimates which can be adjusted later on in the season to improve the accuracy of the estimate.
Historic bunch weight averages are useful in providing an early season crop estimate. This estimate should be completed in May, well before bloom. An average of at least five years of historic bunch weights by block or variety is required in order to give a reasonable crop estimate.
It is important to collect bunch weights of all blocks in production on the day of or just prior to harvest each year. An assessment of the variation in the historic bunch weights will give a good indication of how accurate the crop estimation may be using the historic bunch weight.
In order to estimate the crop load simply multiply the average number of bunches per vine by the total vines per block and the historic bunch weight.
This estimate fails to account for the annual variation in bunch weights, which in some varieties and on younger vines can vary significantly. However, it does provide an early season estimate that serves as a good indicator of the crop potential in the vineyard.
Determining lag phase weights involves sampling bunches during lag phase and determining the average bunch weights. These lag phase weights are multiplied by a factor to predict the final harvest bunch weight.
The benefit of lag phase crop estimation is that it takes into account the annual variation in bunch size, which can be as much as 100% of the bunch weight in some circumstances.
The lag phase method of bunch weight prediction is more accurate than the historic bunch weight method as it takes into account the annual variation in bunch weights. The disadvantage of this method is that the estimate cannot be completed until mid- to late-July. It is also critical to time the sampling of the blocks to the same stage of lag phase each year in order to improve the accuracy of the estimate.
Lag phase occurs about halfway between bloom and harvest and is identified by the hardening of the seeds. An increase in sugar production signals the end of lag phase and the onset of veraison. Lag phase is a useful time to collect samples to predict the harvest bunch weight, which changes very slowly at this time. The lag phase weights are close to half of the actual harvest weight.
When sampling, it is important to accurately identify the exact stage of lag phase to ensure the samples are always taken at the same stage.
The lag phase factor is the difference between the lag phase weight and the harvest weight. Divide the harvest bunch weight by the lag phase weight to determine the factor. For example:
The lag phase factors vary according to variety and stage of lag phase at which bunches are sampled. In general, lag phase factors vary from 1.5-2.5. The factor for a particular variety will range from 1.9-2.5 at the very start of lag phase.
This phase is indicated by the onset of hardening of the seeds in the grape. Dissecting the berry with a razor blade can test this. The berry is at the start of lag phase when the razor blade meets resistance when cutting due to the hard seeds.
Later stages of lag phase can be determined by measuring the % Brix in the berries. The later stages of lag phase will have % Brix values of 6.0-9.0 % Brix. At this stage, the lag phase factor will range between 1.5-1.9.
It is critical that the lag phase at which the samples are taken are the same from year to year. This will help keep the lag phase factor consistent. Similar to historic bunch weights, an average of lag phase factors over five years will provide a good crop estimate using this method.
Another method of bunch weight prediction is to count the average number of berries per cluster. This number is multiplied by the historic average berry weight, which should be recorded at harvest for each block every year.
It is best to use individual block berry weight values as they tend to be more accurate than variety averages. In the absence of individual block averages, varietal averages can give reasonable estimates.
Bunches can be sampled for berry counting once the berries have set. Sampling bunches should be completed in random fashion as described in the lag phase method. A minimum of 40 bunches should be collected in order to obtain a sufficient sample size. Once collected, one can count all berries from all of the bunches to determine the average berry number.
Although very accurate, this method is extremely time consuming and in many cases may not be practical.
An alternative method is to separate the bunches into tiers of similar sized bunches. Usually the bunches can be split into 3-5 tiers. From these tiers, select three of the most representative bunches for berry counting. Using the average number of berries from each tier, one can calculate the average berry number for the sample as shown in the following table.
| Tier | # of bunches | Sample size | % of sample | Bunch 1 | Bunch 2 | Bunch 3 | Average berry number (ABN) | ABN* % of sample |
|---|---|---|---|---|---|---|---|---|
| 1 | 5 | 50 | 0.10 | 233 | 213 | 209 | 215.0 | 21.5 |
| 2 | 9 | 50 | 0.18 | 175 | 190 | 182 | 182.3 | 32.8 |
| 3 | 17 | 50 | 0.34 | 156 | 143 | 143 | 149.0 | 50.7 |
| 4 | 13 | 50 | 0.26 | 128 | 122 | 122 | 129.7 | 33.7 |
| 5 | 6 | 50 | 0.12 | 76 | 90 | 90 | 83.7 | 10.0 |
| Sample average berry # = 148.7 |
Definitions:
The average berry # is then multiplied by the historic average berry mass. An additional 5% should be added to the value in order to account for the rachis weight.
The berry number method of bunch weight prediction is valuable in estimating blocks where the set has been quite variable and there is a very large range of bunch sizes. The berry method can also produce a crop estimate at berry set earlier in the season than lag phase estimates.
It is necessary to adjust the crop estimate from the predicted value if the block is going to be machine harvested.
The amount to adjust depends on the harvesting equipment, the variety and the status of the canopy. There will be a reduction in tonnage due to some juice and fruit loss and the elimination of the rachis that is left on the vine. In general, the rachis makes up approximately 5% of the total bunch weight.
The estimated % juice and fruit loss is added to the rachis value and subtracted from the total estimated tonnage. Industry standards use values averaging around 10% depending upon the block being harvested.
(a) Berry set: post shatter
(b) Lag phase
