High-Tech Solutions Could Improve Grape Crop Estimates

Determining accurate estimates of grape crop yield is notoriously difficult, but new methods involving precision sensing technology could help growers and, eventually, producers who purchase grapes.


Terence Bates, director of the Cornell Lake Erie Research and Extension Lab in the College of Agriculture and Life Sciences, is heading up a $6 million USDA grant that will take much of the guesswork out of crop estimation.

“It’s really about bringing precision ag tools to viticulture,” Bates says.

Traditionally, growers might count the number of shoots on a vine and clusters on those shoots. Factoring in cluster weight and berry weight, they come up with estimates that Bates says aren’t scientific.

“Every time you measure something like that, you’re adding human error in that,” he says.

Miscounting can be an issue. And sampling error can come into play. If, for instance, a grower only counts and samples from vines that have strong growth potential, he or she could easily overestimate crop yield.

Bates is testing sensors that can scan entire vineyards and measure the characteristics of soil, canopy and the grapes themselves. The data is tracked spatially with GPS so that researchers can determine areas where vines might produce an average, strong or weak yield based on vine characteristics and the availability of soil nutrients.

Once the vineyard is analyzed, a grower can find representative areas to clean harvest and use as samples to estimate crop yield. Bates says growers will pick and weigh the fruit from a sample size that is about 1% of a vineyard acre. The number of samples in an acre or across an entire vineyard block would be determined by how much variation there is in the spatial soil and canopy data. A uniform vineyard will require fewer samples than a highly variable vineyard.

The fruit weight would be plugged into a formula to estimate overall harvest yield. Bates says the system is usually accurate within 5%.

“We’ve gotten much more accurate,” Bates says.

The method has worked well at the grower level, but it doesn’t yet solve the problem faced by companies that use those grapes in products such as juice or wine. They need accurate estimates to determine resources they’ll need to process grapes.

“You’re talking about vineyards that have a wide range of environmental conditions and a wide range of management conditions,” Bates says. “It’s a big challenge, and we haven’t improved our estimate enough to make the industry happy.”

But Bates says that could change if using sensing tools becomes a best practice and product producers require those accurate estimates from growers. It would be akin to the requirements growers have to keep pesticide records, for instance.

“If we develop a method that works and is easy for the growers to do, in the long run, it would help them to make it mandatory,” Bates says.

The data gathered during analysis could also help growers determine how to deploy resources to their vineyards to reduce variability in crop quality.