The answer to the winery's problem came from a seemingly unlikely source. During the early 1990s, most of the wineries in Napa Valley were struck by the worst insect epidemic in their 40-year wine growing history. Tiny, aphid-like insects, known as phylloxera, swept across Northern California infecting grapevine after grapevine. The insects would feed on the vine’s roots and in doing so leave them open to fungal infections that would kill them in two or three growing seasons.
 

Under the direction of remote sensing researcher Lee Johnson, scientists at NASA Ames Research Center became involved with the Mondavi winery to find ways to predict the devastation of the crops. They set up a series of trials wherein they used multi-spectral digital cameras mounted on aircraft to detect the insects by measuring the density of foliage across the vineyard. The studies helped the vineyards monitor the spread of the insects so that they would know precisely when and where they had to pull the existing grapevines and replant with phylloxera resistant vines. In the midst of this project, those researchers involved realized that this same technology could be used to separate vines of varying vigor.

“In these phylloxera experiments, we saw with the multi-spectral imaging data that the amount of foliage on the vines is directly related to their stress levels,” says Johnson. In phylloxera infested plants, a lack of foliage—either fewer leaves or smaller leaves—usually means the plant has become stressed by the insects and is dying. The researchers reasoned that differing amounts of foliage in crops that are not infested may be a good indicator of vine vigor. Specifically, more vigorous plants will have more foliage.

NASA and the Mondavi winery teamed up again in an experiment named CRUSH (Canopy Remote sensing for Uniformly Segmented Harvest) to test whether remote sensing could delineate the plants by their vigor and ultimately by the quality and characteristics of the grapes the vines produce. Johnson explains that the multi-spectral imager they used is essentially a very precise digital camera. Unlike a hand-held camera with film, this imager has several types of digital photoreceptors on it that record very specific wavelengths (colors) of light. The ADAR System 5500 the Ames scientists employed was developed with the assistance of NASA’s Commercial Remote Sensing Program and has four different sensors. One detects only the blue light, one detects only the green light, one the red light, and one the near-infrared light. The data the sensors receive are fed into a computer where images can be produced of each band or combinations of the bands (Johnson et al., 1998).

Propeller driven aircraft similar to this one, operated by Air Flight Services, carry the ADAR 5500 instrument that imaged the Robert Mondavi vineyard. After the NASA and Robert Mondavi Winery scientists published their results, several additional California wineries used airborne remote sensing to gather information about their vineyards. (Image copyrightPositive Systems, Inc, Whitefish, MT)

For the CRUSH project, the scientists mounted this imager on a prop airplane, which was then flown 15,000 feet above the vineyard. To determine the thickness of foliage across a vineyard, they trained the imager on the amount and colors of sunlight reflected off the leaves (Johnson et al., 1998). As can be seen through a prism, many different wavelengths make up the spectrum of sunlight. When sunlight strikes objects, certain parts of this spectrum are absorbed and other parts are reflected, and some heat is emitted. In plant leaves, chlorophyll absorbs red light and other visible wavelengths from the sun for use in photosynthesis. The cell structure of the leaves on the other hand reflects near infrared light. The more foliage a plant has, the more these types of light are affected.

Since the imager measures the intensity of infrared and red light coming off the vineyard, the scientists simply gathered the data the instrument recorded on its flight and compared the intensity of the two types of light across the vineyard. In general, where the difference between infrared and red light was at a high value, then the vines had more foliage and were probably more vigorous. Where the imager recorded low values of this difference, the vegetation was less dense and the vines were probably less vigorous. All of these imaging data were then fed into a Geographic Information System, computer software that essentially matches up the raw images with the landmarks and topography on the ground. The result was a complete map of the vineyard showing general areas where the vines were vigorous, and areas where the vines were stressed.

“The remote sensing gave us a good rough outline of where the high, medium and low-vigor plants were,” says Johnson. Technicians at the winery then went around to these areas and tasted the grapes and ran a number of chemical and water tests to see if the aircraft measurements were correct. After some trial and error, they were able to get a full picture of grapevine vigor across many acres of the Mondavi winery.
 

Healthy vegetation strongly reflects infrared light, but absorbs most green light. The false color image at left is made up of the near infrared, red, and green bands from the ADAR 5500 remote sensing system in the red, green, and blue channels of the image. Vigorous vegetation, such as the stand of trees at upper right, are bright red. Stressed vegetation, like the vines running diagonally through the lower vineyard, is bluish. Roads, water, and buildings are bright white or dark blue.

The patterns of healthy and stressed vines are enhanced by comparing the near infrared and visible light data (right). Vigorous vines are colored green, and progressively stressed vines go from yellow to brown. Using these data, the vineyard managers can subdivide the vineyard into regions of similarly healthy vines. (Image courtesyCRUSH project, NASA Ames Research Center)

“It’s amazing to see how small a change in slope, for instance can affect the quality of the grapes. We found that an increase of a few inches in elevation on a hillside will make a difference," says Bosch. Over the past two years, the winery has begun to micromanage the vineyard based on these subtle differences in vigor. Given the vigor of the vine, Bosch explains they can change the amount of shoots grown by first pruning for the correct number in the fall and then increasing or decreasing the amount of irrigation water the vine receives as it grows. Though it still takes some time to get results, every acre of the vineyard that they can transform into reserve wine quality grapes increases their revenue by $800 per year. Of course in some areas they’ve found that the vigor is the same throughout and there is not much they can do.

Beyond providing the wineries with a bigger profit, the remote sensing images have educated the technicians in how to manage a vineyard. “Seeing all of it at once gives us more experience and the ability to recognize the patterns fairly quickly,” says Bosch. Only after a few years of using the remote sensing devices, he boasts he is getting to the point where he can spot the variation on his own without the imagery. In the fall, he can now clearly see the order in which the leaves turn. Within the span of a few years, they’ve essentially been able to do what took the French decades to accomplish. “This is a real revolution in how we are able to manage our vineyards,” says Bosch.

Successful Tests Lead to Ongoing Research
Johnson says that the process they developed has already been commercialized. Several remote sensing companies that work with Ames are offering remote sensing for any vineyard willing to pay the price, and a number of the wealthy larger wineries in the valley are employing the technology. While all this refinement has the potential of being of great benefit to wineries and wine connoisseur alike, their experiments with the Mondavi winery are far from over.

Ultimately Johnson would like to know how vigorous vines on a patch of land will be before they even plant the first seedling. His team is working with several models developed jointly by scientists at NASA and the University of Montana that may be able to do just that. Given the topography and soil type of a patch of land, these models should simulate how various crops will grow on uncultivated land. “In this way wineries could design a vineyard based on plant vigor from the start,” says Johnson. Experiments like the ones they performed at the Mondavi winery allow the scientists to refine the model. By first testing the model on regions that are already planted, they can get estimates on how the model works and then adjust it to work on areas without plants.

In the long run, the goal of the Ames team is not only to improve the quality of wine in California, but also to gain a better understanding of how to use remote sensing systems and irrigation methods to deal with crop stress. Farmers who raise soybean and corn crops do not want their plants to be stressed at all. By understanding what causes stress in wineries, future agriculturists may be able to look at any crop and tell farmers the best way to irrigate and farm using the least possible resources. “The study is a nice example of a NASA-funded science project that has scientific merit as well as economic benefits,” says Johnson.

• References
• Johnson, L., B. Lobitz, D. Bosch, S. Wiechers, D. Williams, and P. Skinner, 1998: Of Pixels and Paletes: Can Geospacial Technologies Help Produce a Better Wine,Proceedings 1^st International Conference on Geospacial Information in Agriculture and Forestry, 1-3 June, pp. II-469- II-475.

 It's All in the Grapes

This Robert Mondavi Vineyard manager is using the vineyard image and a personal GPS to navigate and prepare the field for harvest. He is sectioning off different parts of the field, based on vigor, with flagging tape. Grapes from different sections were fermented separately. Here, highest quality ’reserve” wine was produced from low vigor areas. (Image courtesyCRUSH project, NASA Ames Research Center)