The red light signals a high-vigor zone, the white light signals a low-vigor zone, and the amber light in the middle alerts the crew to an approaching change in zones.

The red light signals a high-vigor zone, the white light signals a low-vigor zone, and the amber light in the middle alerts the crew to an approaching change in zones.

The grape harvester signal, a light signal attached to grape harvesters that is being used to differentiate fruit quality during harvest, combines new technology with old-fashioned farmer ingenuity.

When Jenn Smithyman, precision agriculture specialist at Ste. Michelle Wine Estates, couldn’t find a portable signal that could incorporate the technology she wanted to use in separating fruit quality during harvest, she created a device and helped develop a system that could be used on any harvester.

For several years, Washington State’s largest wine producer, Ste. Michelle Wine Estates, has searched for a way to improve the quality of mediocre vineyards—ones that overall are so-so in quality, but have pockets of high quality grapes within the vineyard.

"We wanted to pull out areas or ‘sweet spots’ within vineyards that were superior in terms of quality," said Smithyman, adding that Ste. Michelle viticulturists and winemakers look at grapes throughout the season and have a general idea of those vineyards that have good spots within them but are made into lower-priced wines because of the overall quality.

Aerial mapping

In 2003, Ste. Michelle began evaluating mapping technology called normalized difference vegetative index (NDVI) to identify variability in vineyards. They wanted to identify higher quality fruit that could produce high-tiered wines and give their winemakers blending options in the winery. Smithyman used the NDVI imagery to map high- and low-vigor zones within the block in order to influence management practices that would improve vineyard uniformity. Also, she mapped vineyard variability so that distinct areas within blocks could be harvested separately.

To evaluate if it was worth all the effort to separate grape quality in the field, Ste. Michelle produced wines from these different zones created by the NDVI imagery in 2005. "In blind tastings, the wines showed significant differences," she said.

Initially, Smithyman tried to separate blocks into ten different vigor zones, but found that such a high number was unmanageable and the differences too slight between zones. "However, there are greater differences between high- and low-vigor zones," she said.

They tried to harvest the vigor zones from the NDVI mapping by flagging vines and distributing maps to hand-harvest crews, but the efforts were time consuming, expensive, and created confusion among the workers. Efforts to separate grapes with the machine harvester also proved inefficient, she said, especially when harvesting was done at night. They even tried harvesting the vigor zones by entire rows. Although it reduced the accuracy of the zonal harvest, four out of five tasters could distinguish wines produced from the zones that were differentially harvested in 2005.

Knowing they were onto something, they continued their efforts to improve differential harvesting. In 2006, some 3,500 acres of Ste. Michelle and grower vineyards were aerially imaged, with 20 blocks consisting of 400 acres chosen for differential harvesting.

Smithyman, with assistance from Washington State University’s Center for Precision Agriculture Systems, found a way to link the aerial maps with the grape harvester signal.

The harvester signal technology is still evolving.

The signal bar consists of a series of lights (red, white, and amber) that are controlled by a personal digital assistant or PDA with a Global Positioning System (GPS) card. The NDVI maps were loaded into the PDA, and with the help of customized software, the PDA turns the signal lights on and off based on the harvester location within the field.

To prepare the NDVI maps, she draws polygons around the high- and low-vigor zones with a 15-foot buffer to allow gondola drivers working with the grape harvester time to make changes.

Gondola drivers can decide for themselves how they want to coordinate receiving different grades of fruit. In most situations, the gondola drivers follow each other down the row, with each designated to a specific light color (vigor zone), and collect fruit from the harvester to deliver grapes to different trucks. When it is the second driver’s turn to receive grapes, the first driver moves forward and out of the way.

Smithyman estimated that the signal arm was built for $200. The GPS software they used was $500, and it cost $3 per acre for the aerial imagery, which was collected in early August. She doesn’t have a price tag on the customized software program that had to be written, but noted that demand for the technology would likely stimulate commercial development of the software needed to link the maps to the signal arm.