by Andreas Neuman

Published on February 6, 2016

Some of the ways in which growers of high-value agriculture such as wine grapes can turn remote sensing data into very substantial returns on investment are intuitive. For example, it is easy to understand that detecting a water leak right after bud break or identifying an infestation before it spreads can save significant money for growers. It is also fairly intuitive that identifying underperforming zones early in the season, analyzing the causes, and subsequently applying customized cultivation techniques to those areas throughout the growing cycle can dramatically boost yields. However, there is another powerful way in which savvy growers are unlocking remote sensing’s potential value.
*This article is commenting on Australian studies, all currency is indicated in Australian dollars
This method is called selective harvesting, defined as the segmented picking of fruit at harvest according to different yield/quality criteria in order to exploit the observed variation.


New techniques capitalize on proven science

The concepts behind differential cultivation and selective harvesting are not new. Robert Bramley of Australia’s Commonwealth Scientific and Industrial Research Organisation (CSIRO), one of the world’s leading researcher into applied precision viticulture, has been researching applied techniques for differential cultivation based upon remote sensing products since the 1990s. In a peer-reviewed journal article from 2004, he detailed how both Australian grape growers and winemakers (an important distinction when describing use-cases for tools and return on investment) have successfully used remote sensing technology to deliver to wineries parcels of fruit which are as uniform as possible and meet the specific quality criteria necessary for their intended use.


Optimal harvests may require selecting crops at different times

In one of the case studies they detail, the authors illustrate how a 3.3 ha (8.15 acre) vineyard was divided in two zones using remote sensing data collected at veraison and processed into crop health maps based upon the Normalized Difference Vegetation Index (NDVI). An experiment was conducted; in 2002 the two zones were harvested on different dates so that only one chaser bin had to be used and, perhaps more importantly, so that one zone could continue to mature on the vines to reach peak quality. In order to benefit from the selective harvest, the grapes from the two zones were vinified separately.
An economic analysis was performed to analyze the fruit value in multiple scenarios:
(a) the baseline – uniform harvest;
(b) selective harvest on the same date (which requires a second trailer bin);
(c) selective harvest on different dates;
(d) selective harvest on different dates with the added investment in a headwall to clearly delineate the zones (which causes some loss of planted area as well as a capital investment but improves harvest efficiency).
Adjusted for the varying cost of the harvest, the gross returns per hectare were all greater in the selective harvest events.

  • Scenario B resulted in a 12.73% increase ($1,412/ha or $572/acre);
  • Scenario C resulted in a 21.43% increase ($2,377/ha or $962/acre);
  • Scenario D in 23.36% increase ($2,591/ha or $1,049/acre).

Selective harvesting can command higher wine prices

In a separate study conducted in 2002, multispectral imagery was collected during veraison and was used to divide a Western Australian vineyard into two zones, one high vigor and one low vigor. Selective harvesting was implemented, and after vinification, a determination was made that the quality was sufficiently differentiated to bottle the wine from the higher quality zone at a different price point. If the grapes were combined, the monetary value of the higher quality would have been lost as the winemaker would have sold the blend at $19 (retail price) – which is what the lower quality zone’s wine was bottled at. The selective harvest allowed the winemaker to use the grapes from the higher quality zone to a produce a $30 (retail price) wine, which for this 3.3ha (8.15 acre) field, resulted in an increase of $139,480 on the retail value over what uniform harvesting would have produced. This is equivalent to an additional $3,653/ton of harvested fruit or an additional $42,267/ha ($17,112/acre).
Clearly, these studies show that separating higher quality grapes and harvesting them at their peak quality creates more value to growers who can produce reserve quality grapes and winemakers who produce more expensive (and better quality) wine as a result. s we mentioned earlier the returns on investment for growers and for winemakers using contracted grapes merit individual analysis. Another Australian study helps break down the numbers for both stakeholders. For this study, the subject vineyard was divided into high and low yield zones and selectively harvested. After harvest, the grape quality in the high yield zone was assessed as C grade, with a corresponding price per tonne of $1,200 to the grower and a per bottle price of $14 for the winemaker. The grape quality in the low yield zone was graded as B+, with a corresponding price per tonne of $2,100 and a per bottle price of $23.50. If the two qualities were blended, it would have been graded as a B-, worth $1,800/tonne or $18/bottle. A cost/benefit analysis was performed to determine if the additional cost of selective harvest (using two trailing bins for same-day selective harvesting) created a positive return on investment (ROI) for both the grower and winemaker. The result was that the increased value of the grapes resulted in a 3.25% increase ($4,657 for the 4.25 ha (10.5 acres) field) to the grower and a 20.46% increase ($272,971) to the winemaker!

Drones make it easy (and profitable) to implement selective harvest

While these examples used two zones to simplify the studies, experience with precision viticulture has shown that most vineyards and even varietal blocks can be divided into more than two quality zones. In modern precision viticulture, the choice of how many zones to divide a plot into stays in the hands of the same stakeholders – usually the vineyard manager and/or the winemaker – however, now there are more tools at their disposal to support their decision-making process. Remote sensing has never been more accessible or affordable thanks to the advent of the commercial drones and sensors designed for agriculture, and data management platforms have been custom built to ensure that the full value of the information can be unlocked by providing avenues for collaboration and expert advice from off-site consultants.
For most vineyards, crop health surveys flown by a commercial drone operator three times per season can represent less than a 1% increase over existing input costs. However, for vineyards interested in trying out remote sensing for the first time and wish to have a single “prove it to me” crop health survey before committing to the recommended series of flights, there is one particularly useful time to have these flights. Multiple studies (Lamb et al., 2004 for example) have shown that veraison (starting late July in California and much of the Northern hemisphere) is the best time for the acquisition of imagery of the vineyard. The season has sufficiently progressed and many of the most important characteristics of the grapes are observable – not just by the naked eyes and sampling tools of field workers but at a larger scale in crop health maps that are built by using one or more crop indexes such as NDVI.

Data collection can be synced to the growth cycle

At veraison, the information is highly actionable as it comes at a strategic time when both the vineyard manager and/or the winemaker can positively affect the volume and quality of the harvest. Vineyard managers can see the results of their field management decisions from the early part of the season and assess the spatial variation in the crops as well as the need to adjust management zone boundaries. At this point in the growth cycle there confidence levels in yield predictions begin to increase and there still is time to take effective action. Information gained at this time can be used to cultivate a more homogenous crop and adjust the composition of the grapes to meet the desired qualities.
At this time of the season, fruit sampling is quite important, as BRIX, pH, titratable acidity and a few other variables are analyzed. However, without actionable information such as the kind provided by drones equipped with multispectral sensors, sampling is usually performed randomly in an attempt to conduct a representative survey of the fields. After those surveys, most growers are aware that differences exist within their field but drawing precise boundaries can be extremely challenging (not to mention time-consuming). That’s when remote sensing and the resulting vigor (crop health) maps of vineyards come in handy, as they provide comprehensive health assessments that identify individual vines that need attention as well as wider trends that identify zones can be made in hours.
New technology is affordable and available to vineyards of all sizes, and early adopters are putting new tools to good use. For example, now that remote sensing data can be used to provide a precise map that displays variance, a winemaker, vineyard manager, consultant or any other stakeholder can collaborate with field workers to design and execute a targeted instead of random sampling plan by pinpointing specific vines that representative identified zones.
The same targeted sampling method using a vigor map can be used to improve harvest prediction because there is a high correlation between plant vigor and weight of the grapes. Vineyard managers performing fruit counting for yield estimation can now get more accurate results by creating sampling zones, sampling within those defined zones and creating zonal estimates which can be summed together instead of building one field-level average.

Join Our Newsletter

Discover Precision Agriculture and Biocontrol news and tips, learn about your fellow UAV-IQ users, and stay up to date with what’s happening at UAV-IQ.

Related Articles