By Amanda Garris Ph.D. ’04
periodiCALS, Vol. 6, Issue 2, 2016
Each morning, the same question greets Cornell Lake Erie Research and Extension Laboratory Director Terry Bates from his office white board: What are you doing for the grape growers? This summer, the answer has come easily. He’s systematically taking the guesswork out of managing vineyards, with help from a fleet of sensors that see the vineyard more clearly than the human eye.
We can be tricked by the orderly appearance of the trellised rows into thinking of the vineyard as pinstripes on a plain sheet. It’s more like a patchwork quilt. Areas of waterlogged soil adjoin the well-drained, and weaker vines are interspersed with the vigorous. A one-size-fits-all management approach—expecting each vine to produce the same fruit and have the same needs—has consequences. On a weak vine, producing too much fruit one year can hurt yields the next, and optimistic, robust vines produce many more grape clusters than they can ripen, requiring thinning of grape bunches midsummer. This variation is a headache for growers who need to decide on a management strategy before turning on the tractor, but Bates’ new approach aims to turn this variation into value.
“I don’t want to tell a grower what to do,” Bates said. “I want to give them the information that allows them to grow better grapes, more easily.”
That passion is why, in a Lake Erie Concord vineyard mid-July, with grapes the size of blueberries and cloudless skies overhead, Bates experienced a thrill, watching a family’s tractor-mounted computer automatically adjust the rate of crop thinning to match the vigor of the vines. He deemed it a “historic day for New York Concord production.”
The milestone was made possible by better data, and more of it. Over the past year, sensors dragged over the soil had measured its conductivity, reflectance sensors had mapped the leafy grapevine canopy and its gaps, and cameras took pictures of the developing grape clusters. Fully combined, the data created a geo-referenced map with an accurate picture of zones of productivity. The growers decided on a treatment for each zone and then uploaded it into the tractor’s computer.
For fully automated, information-age viticulture, Bates envisions that growers on the one million acres of vineyard in the United States will run the sensors through their plantings, upload the data to a website for processing, and download maps that classify the different zones of the vineyard. Then, precise and optimum management plans can be implemented on the fly by computers on the tractor. This summer’s Concord vineyard success represents the first commercial step toward this goal.
Bates’ larger vision is closer to reality because of a $6 million, four-year grant from the USDA National Institute for Food and Agriculture Specialty Crops Research Institute. It brings Bates together with collaborators from Carnegie Mellon—the “engineering brilliance” behind the work, according to Bates—as well as Penn State, Newcastle University in the U.K., and U.C. Davis on a nationwide project.
So what is Bates doing for the growers? The answer depends on the type of grape.
“For the Concord growers it’s all about maximizing yield and sweetness. In the table grape industry, there’s a lot of concern around crop size and color,” Bates noted. “For wine, the interest is in precision harvest to separate lots of wine based on fruit quality. And the big players in wine want to be able to predict their crop size accurately in the middle of the summer, so they can plan harvest logistics—like how many forklifts they will need. So everybody is interested in a slightly different application of the project.”