AIRBORNE
Uncorking Technology
Better wine through remote sensing
By Joy A. Colucci, Ph.D.

Scientists at NASA Ames Research Center, Mountain View, California, are helping to bring space-age remote sensing technology to the ancient art of wine making through a joint effort with Robert Mondavi Winery in California's Napa Valley. Although color infrared (CIR) aerial photographs have been widely employed by winegrowers, the use of precision agriculture methods and digital remote sensing data in vineyard operations is still relatively rare.
    This unique collaboration came about through funding from the NASA Ames Commercial Technology Office for Technology Transfer Projects, coupled with a significant commitment on the part of Mondavi in terms of staff time and matching funds. California wine growers, and the agricultural community in general, tend to be conservative in the use of new technology. However, Daniel Bosch, Mondavi Vineyard technical manager, says that Mondavi sees its involvement in this program as part of the company's tendency to "look a little further ahead" in the adoption of new and innovative practices. As a result of this commitment, what started out as a small research project now forms the basis of a growing market for remote sensing data in the Northern California wine industry.

Pests in the Vineyard
Mondavi initially sought help from NASA in 1993 with its battle against the phylloxera infestation that affected much of the California wine country. Phylloxera are a type of plant lice that attack the root system, thereby depriving the plant of water and nutrients. The exponential growth of the pest is potentially devastating; worse yet, it cannot be cured. Pesticides are not effective, and the only way to stop the infestation is to remove damaged vines and replant with more resistant rootstock.
    The GRAPES (Grapevine Remote-sensing Analysis of Phylloxera Early Stress) project was a collaborative effort involving NASA Ames and Mondavi, along with the University of California-Davis, California State University-Chico, and the University of California Cooperative Extension (Napa). Lee Johnson, senior remote sensing scientist at NASA Ames, explained the goals of the program: "We knew that phylloxera could not be stopped from destroying vines. The usefulness of the imagery is in determining exactly where the infestation is, and at what stage." Given that information, the vineyard managers could make decisions regarding the care and replanting of affected vines.
    During the 1993 and 1994 growing seasons, Johnson and his colleagues collected data in the visible and near-infrared ranges using three different airborne sensors. The data were geolocated using GPS, and had spatial resolutions of 2 to 3 meters. A normalized difference vegetation index (NDVI) was generated from the data using standard methods. As an indicator of photosynthetically active vegetation, NDVI can be related to plant stress, as well as the structure of the canopy.
    In comparing NDVI data from one growing season to those of the previous year, GRAPES researchers were able to identify declines in canopy density that resulted from the phylloxera infestation. "We could detect phylloxera infestation up to one year in advance of when the symptoms became obvious in conventional CIR photographs," said Johnson.
    By 1995, Mondavi's phylloxera problem was under control with most of the susceptible root stock removed and replanted. In addition to aiding in the early detection of infestation, Bosch noted that "the greatest value of GRAPES was to generate interest in remote sensing and GIS." Since then, the seeds of the GRAPES research have led to broader use of remote sensing data in Mondavi's vineyard operations.

Harvest Optimization is the Goal
In 1997, the NASA Ames Commercial Technology Office funded a one-year follow-on project to GRAPES called CRUSH (Canopy Remote-sensing for Uniformly Segmented Harvest). While GRAPES was more of a scientific research program with university collaboration, the focus of CRUSH is on technology transfer, and the partners are commercial companies. NASA and Mondavi were joined on CRUSH by two vineyard consulting firms, Terra Spase (Napa) and Winegrow (Healdsburg). Terra Spase president, Paul Skinner says that his company also invested significant resources into the CRUSH project. "We see it as a long-term investment in the technology," said Skinner.
    The objective of CRUSH is to demonstrate precision agriculture practices using remote sensing data to the California wine industry. NASA contracted with Crop Image (Salinas, California) to collect imagery over Napa and Sonoma vineyards in late July 1997 using an ADAR 5500 digital camera system flown at 14,000 feet, and measuring 4 spectral bands (blue, green, red and near infrared) at 2 meter resolution. In collaboration with Ames researchers, the data were processed by Terra Spase using the ENVI image processing system developed by Research Systems, Inc., Boulder, Colorado. As part of the ground study component of CRUSH, Mondavi and NASA made field measurements that were compared to the remote sensing data in two study blocks: one for chardonnay and one for pinot noir. These included sugar, malate, and pH in the grapes, as well as canopy light interception, chlorophyll and leaf water potential readings in selected locations within the study blocks.
    As in the GRAPES project, CRUSH is based primarily on NDVI data. "The NDVI has the advantage of using relatively narrow spectral bands centered in regions that are perfect for looking at vegetation," explained Johnson. The vegetation index uses one measurement in the near infrared "plateau," and one in the visible red region, and this emphasizes the information on the canopy density. "The data are then combined into groups of similar value and color-coded to make it visually easier to see what's going on," Johnson continued in discussing the CRUSH image processing methods. "In the end, instead of looking at a continuous tone image, like CIR, you're looking at polygons with colors indicating density or vigor, which helps the user to filter through the information."
    CRUSH participants have found several benefits of using the digital imagery over the traditional CIR photos. Primarily, digital data allow growers to use the new geospatial tools that are becoming available, including GPS and sophisticated GIS applications. Of course, CIR photos can be scanned and converted into digital data. However, the scanning process is time-consuming, may introduce inconsistencies, and may produce data at reduced dynamic range.
    Mondavi growers use the NDVI images to see what parts of the field have the similar canopy density or vigor as others, and what proportions of the field are above or below a certain threshold. This allows the field to be divided into management units that can be sampled for maturity and then harvested separately. Traditionally, Mondavi harvested an entire field at once, based on ripeness. But this method resulted in some grapes being over ripe when harvested, and others under ripe. Separation of the field into discrete units based upon remote sensing characterization allowed Mondavi to harvest segments of its fields at different times to better coincide with optimal ripeness, and also to avoid mixing higher quality with lower quality grapes in the same wine lot for fermentation.

A Growing Marketplace?
In addition to Mondavi, Terra Spase is processing and distributing CRUSH imagery to an additional 25 clients in the Napa/Sonoma region. CRUSH adds onto what it's already providing with Terroir, a proprietary GIS tool that was developed to visualize soil spatial variability data. It also aids winegrowers in selecting rootstock based on soil type and in diagnosing problems in their vineyards. Several Terra Spase clients have found the NDVI images useful. "We had feedback that it was helpful in a number of different ways, and that's what we're building on this year," said Skinner. "We're trying to redefine our strategy and take advantage of what we learned last year to go forward and make it an even more successful tool." Terra Spase plans to cover more area more frequently with new data acquisitions this year. It also hopes to bring the technology to its clients at a lower cost, having already made a major investment last year in software development and training.
    Skinner, however, is cautious. "It takes a lot of effort to bring a new tool into the marketplace. There is a lag time, growers have to get comfortable with it and we have to figure out better and better ways to do it. So I see this as a long-term development effort." Skinner further noted that many past efforts to market remote sensing products to agriculture have failed because they didn't provide users with the background they needed to really understand how to use the information. Skinner bases his business on having long-term relationships with his clients, and on a thorough understanding of their needs. He views the remote sensing data as one more in a wide array of tools that he offers his customers to help manage their vineyards.
    In funding CRUSH, NASA hopes to boost demand for the technology by having Mondavi relate its experience to other growers, as well as increase the supply capability by transferring remote sensing expertise to established crop consultant companies like Terra Spase. "What's going to cause these markets to happen is for the early innovators to stand up and say, 'This was my experience, and this is what it did for me on the bottom line,'" said Johnson in discussing the technology transfer goals of the CRUSH project. "If people see that Mondavi was able to boost [its] crop value by so many dollars in doing this, then they can make an informed decision about whether or not to avail themselves of the technology."

The Proof is in the Wine
Mondavi has just finished tasting the wine from the CRUSH project, and Bosch reports that there are clear differences. "There were portions of the vineyard that, when we picked before, could not be used for our higher quality wine, and now some of the blocks that were part of CRUSH will be used in the higher quality wine. That's a big step for us." The clearest differences in the CRUSH study blocks were seen in the chardonnay grapes. The vines were categorized into three groups of vigor: weak, moderate and strong. Whereas before, none of the chardonnay grapes in the moderate vigor category made it into Mondavi's reserve program, as a result of CRUSH those grapes will now be used for the higher quality wine.
    More importantly, use of the CRUSH technology is leading to a better understanding of the nuances of the vineyard. "What we've done with the chardonnay grapes is separate the wheat from the chaff," explained Bosch. "So although some of it is better, some of it is not. It's forced us to look more closely at those areas, so now we're going about the process of fixing them. This is leading us to a much better understanding of what's going on." Thus, the initial results of the CRUSH project have led Mondavi to begin managing subsections of its vineyard differently that before were treated more uniformly. This brings the California winegrower a little closer to European practices. Vintners in Europe have had hundreds of years to understand all the physical factors that exist in their vineyards, and how the grapes vary accordingly. Many California vineyards have only been established for 20 to 30 years, and growers are just beginning to understand the subtle nuances of their vineyards. Skinner confirmed that many Terra Spase clients who had previously managed blocks of their vineyards differently were now modifying those boundaries based upon the remote sensing data and were pleased with the results.
    In wrapping up the CRUSH project, Mondavi has agreed to conduct a cost-benefit analysis of the results. Although the costs are known, this task cannot be completed until it can place a value on the wines. In the meantime, Mondavi already plans to incorporate the technology into their operations next year. Bosch reports that they will continue to use the NDVI images to optimize harvest of their chardonnay grapes, and will explore other ways in which to use the data to better manage their vineyard.

About the Author:
Joy A. Colucci, PhD., is a freelance science writer. She also writes for TerraComm, a scientific communications company in Mountain View, California. She may be reached at 650-961-3917.

References:
Johnson, L. et al., 1996, Airborne imaging aids vineyard canopy evaluation. California Agriculture, v. 50, no. 4, p. 14-18. Oltman, D., 1997, Building an image. California Farmer, September 1997. Bosch, D., 1997, Vine and Vigor: Art and Science in the Vineyard. Precision Farming, December, 1997.

Back