Investigating Commercialization of Sensor Technology and Mechanization
To determine the commercial potential for sensors and variable rate mechanization in bulk vineyards, the largest barrier will be return on investment (ROI). The majority of bulk juice acreage responds to a ROI even if there are technical or upfront cost barriers. Smaller growers, which are not an insignificant portion of the industry, will often have barriers to adoption beyond ROI. The goal of the project is to create tools that are useful and accessible to both large and small growers. It is important to forecast and predict adoption behavior beyond reasons of ROI, particularly with smaller growers.
To assess ROI, the project first investigated baseline behavior. The promise of sensor technology and variable rate management is to decrease vine size variability within a block and to improve vine size, thereby improving yields while maintaining quality, in a cost effective manner. The first step was to assess vine size in commercial vineyards.
NDVI sensors were deployed by the Extension team in grower vineyards. First, extension field staff used sensors to map grower vineyards. Later, a loaner program was set up to familiarize growers with the technology. NDVI provides a relative measurement that approximates vine size. Sensors were deployed in over 3,000 acres. About 20% of participating growers took dry matter measurements to calibrate NDVI readings that allowed sensors to accurately predict vine size. The primary motivation of this activity was to engage growers in the technology and observe their use of the technology with the hopes of building the future knowledge required for commercial implementation. In the meantime, we were also able to gather a significant amount of baseline data to measure vineyard performance without the assistance of technology. Average vine size was below 2.5 lbs. on standard spacing, many vineyard operators were averaging below 2.0lbs. The data indicated average vine size was between 35% and 60% smaller than research based recommendations. Variability within blocks was also extremely high. The differences between small and large vines within a block was often 50% or more. With this information we could start to create some assumptions about the potential of this technology.
With the effective application of technology, we hope to make bad vineyards good and good vineyards great. Compared to optimal levels of production preliminary results indicate that the region has a theoretical potential to increase yields by 40% while maintaining quality. Assuming just 50% of acreage could achieve 80% of full potential farm gate value of regional bulk juice production would increase from $32 million to $36 million annually. Current production methods tend to create cyclical crop production with years of high yield followed by years of low yield. This constant cycle undermines the value of the crop as new customers are expensive to reach and difficult to keep when the product is less available. The potential increases in yield also assume production is balanced, thereby severely limiting the economic impact of cyclical production.
We also took a look at baseline production costs and practices in the region. This allowed us to compare the performance we discovered with the current economic realities of the bulk juice industry. This cost survey shed light on the ability of growers to adopt technology that required an upfront capital cost. It also allowed us to measure reported costs of production practices to focus variable rate technology and extension guidance on areas that make the most economic sense.
Bulk juice, like commodities are in surplus. Our data reveals growers are responding by lowering costs below Cornell recommended levels. Despite those cost reductions, growers were still making capital investments. Growers also reported establishing new vineyards and expanding their operations. At current production levels and prices, participating growers were not operating at sustainable levels. While gross revenue did exceed operating costs, it did not exceed the total cost of production.
At current prices growers need to average 9 tons per acre to exceed the total cost of production. Such levels of production are only theoretically possible in an average season. This technology could provide growers with good sites the tools to be very close to sustainable. Being very close to sustainable when market prices are at a bottom would be a significant improvement for the average grower.
The idea behind this project is to lean on grower expertise, rather than reinvent that expertise. These tools provide growers with data and the practical ability to manage that data. Extension associates and researchers continue only in an advisory role. This should mirror a realistic commercial use of the technology. Growers that participated in NDVI studies were given maps and provided with the information to calibrate those maps. Interested growers also discussed results with extension viticulturists. Preliminary results ranged from extremely successful to GIS art. In the early stages of technology adoption, it is critical to demonstrate how sensors like NDVI create actionable data. Growers that decide not to act upon the data and just view it as art will see no economic impact of the technology. This pattern would undermine the adoption of sensor technology in field crops for years. There were, however, real success stories. Growers used the maps in different ways. Growers targeted renewal work to specific locations, selected areas (full rows) to thin, fertilize or prune differently. Growers also used to the maps to manage whole blocks differently. When the maps made it clear that problem areas were becoming larger than the high-performing areas, growers changed their strategy in the block.
While the adoption of variable rate mechanization requires a larger investment, the preliminary work made it clear that it was practical for larger farms. Current variable rate mechanization trials include variable rate shoot thinning, fruit thinning, and fertilizer management. All mechanized production practices can be updated with variable rate technology. Viticulture research and economic data indicates that crop load management is likely to provide the most impact. One focus of this economic evaluation has dealt specifically with variable rate crop thinning on a commercial vineyard. This began after 2014 when significant over-cropping led to widespread commercial thinning with mixed results.
In a research and commercial setting fruit thinning works very well to balance crop load, improve fruit quality and improve return crop. Overall the practice increases or maintains profitability at significantly lower levels of risk. On a commercial scale, the application of fruit thinning presents challenges. Many growers infrequently fruit thin and mistakes and unexpected results are to be expected when engaging in an unfamiliar or new production practice. 2013 was an excellent example of both very successful fruit thinning and mixed results. Growers indicated that acreage was under-thinned, over-thinned and unnecessarily thinned. We took a look at a dozen growers and their tonnage in 2013 and 2014. On average thinned vineyards had a yield of 7 TPA in 2013 and returned with a yield of 8.1 TPA in 2014. Un-thinned vineyards averaged 8.5 in 2013 and 5 in 2014. On average, growers were very successful.
However, looking at individual blocks showed room for significant improvement. One small block was thinned down to 2 TPA. After that experience the grower did not thin the block in 2014 and vine size and fruit quality suffered as it was over-cropped. While growers are not always calculating the benefits of fruit thinning, they’re certainly calculating the costs of unnecessary thinning.
In part, fruit thinning performance could be dramatically increased by additional samples and experience. Every harvester performs differently when harvesting green fruit. Vine size can also play havoc with samples. Technology will never be a replacement for sampling and experience. Sensor data will highlight areas where conditions have changed and growers need to take additional samples to prevent over-thinning. None of the growers that submitted yield data for 2013 and 2014 had adequately sampled yield prior to thinning. The most common reported practice was 1 – 2 samples per farm. Some reported no sampling.
Growers report that a successful thinning experience results in an increase in return crop equal to the tonnage removed. The data we gathered from 2013 seems to support this observation. With technology, there may be the potential to further improve those results.
Updating a harvester to allow a prescription map and computer controller to change bow rod speed on the fly is not an inexpensive investment. Growers should expect to invest at least $5,000 into the harvester itself. GPS technology, cabling, tractor display and controllers is also required hardware that can be moved from harvester to tractor. In total a grower might invest $20,000 - $30,000.
To put another way, we need to find a way a grower can leave an extra 150 tons by properly thinning spatial areas, while maintaining vine size and fruit quality. Payback in a single year might require 10% more total crop. It is probably more realistic to anticipate a 5-year payback period as thinning is not necessary every year. We continue to look at other ways to ensure a rapid and realistic payback period for variable rate technology. The core technology is precision agriculture software, a computer display, and GPS.
There is the opportunity to apply the technology across other production practices. Growers seem most excited about management and waypoint features. This would allow growers to reference points to return to at a later time. This technology also offers data tracking and mapping of farm activities. For our larger farms, monitoring activities accurately could add value and ease vineyard management challenges. Variable rate potassium applications may provide a real savings as well. Taken apart, these three practices would not likely justify the application of the technology. As tools to add value to the system, these applications could accelerate pay back.
CLEREL research staff and cooperating grower vineyards, like Betts farm, are a core test of commercial application of vineyard technology. Variable rate thinning on the Betts farm began in 2016. In 2016 an average of 1.5 tons were removed. The 2016 trial was really proof of concept, to learn more about how beater speed reacts to vine size and crop size. Adjustments in the harvester and variable rate settings did limit the amount of data available for review. That being said, the technology did show a lot of promise. For data that is available, it appears that return crop is larger than expected. Un-thinned areas were likely hit by drought stress and crop size appears to have significantly declined. Thinned areas took advantage of dry and sunny weather. By avoiding drought stress, it appears that return crop was above average.
These preliminary results reveal a return on investment increase of $194 per acre. Across two years the performance of the un-thinned area was less than total cost of production. In the thinned area it exceeded the cost of production, an unusual victory in the current market. While this result was surprising, we expect better results in 2017. Crop load is variable and significantly out of balance in some areas. This allowed us to test variable rate thinning over more acres that had a greater need for thinning. In 2017 variable rate thinning on the Betts farm was expanded, based on yield estimates, to 60 acres. This should provide enough data to adequately show how the technology performs at scale. Changes in yield, vine size, and return crop will all be monitored to validate utility and performance of their investments in technology.
Sensor technology and variable rate vineyard management offers a host of tools that would have an economic impact on many different production practices, all which have the potential to add value to the investment. While individual growers sometimes focus on particular production practices, in this case the low hanging economic fruit appears to be crop load management through delayed yield manipulation. It is an area that many growers seem to find most challenging to respond to variable conditions. It is also a practice that has one of the largest short term impacts on gross revenue. While we continue to look at other opportunities that add value, delayed and variable rate crop load management will likely remain the basic building block for commercialization in bulk juice. Thanks to Terry Bates, Rhiann Jakubowski, and Jackie Dresser for the integrated viticulture data, analysis and images used to make these observations and analyses.