1996 Annual Report for LNE96-072
Sustaining Grape Production in the Northeast through Farm-tested Information Technologies
Summary
Summary
Management of pests on grapes in the northeastern United States has historically been dependent on routine pesticide applications. Growers generally cannot access hourly weather information and disease predictive models that help to time applications more precisely and in a way that would reduce pesticide use. This project has used the knowledge of grape growers in Pennsylvania and New York to direct the development of the VITIS expert system and daily observed and forecast SkyBit weather reports. VITIS, Latin for “vine,” works with weather data the help growers in make disease management decisions that sustain grape production.
Objectives
The primary goal of the VITIS expert system is to replace fungicides with knowledge. Fungicides are often not required if it can be demonstrated, through consideration of all the factors, that there is no disease risk. In addition, even when there is a disease risk, there are options to select less harmful fungicides or reduce application rates. The specific project objectives were:
Use grape-industry grower organizations to develop and farm-test new information tools.
Evaluate the usefulness, reliability, cost, and acceptance of weather information sources for grower use in sustainable vineyard management decision-making.
Incorporate sustainable vineyard management tools such as predictive pest models, site-specific weather information, and forecasts into the VITIS expert system, and then evaluate their impact on decision-making in commercial vineyards.
Methods
The prototype of VITIS, developed before this project began, offered growers specific disease management recommendations. During meetings connected with this project, growers requested that VITIS be redesigned to become a decision support tool. These growers were interested in having the most important disease management factors presented, along with an interpretation of the disease risk. The farmers could then use the information themselves in making a disease management decision. VITIS no longer provides a specific recommendation, but instead has been developed to assist growers in making decisions.
Meetings with Pennsylvania and New York growers and extension personnel were held, both in groups and individually, to provide feedback on the VITIS format and operation. In addition, plant pathology specialists from Ohio, Michigan, New York, Pennsylvania, and Canada were involved in the development and representation of the disease management logic in VITIS.
Five Grower Priorities for VITIS Operation
Windows version of VITIS. The initial meeting after this project was funded brought the entire grower group together to discuss the direction of VITIS development. Growers quickly observed that the program should be translated from a DOS to a Windows platform. This was a necessary step, but was also time consuming—the program format and screen displays were completely rebuilt for Windows operation. This took up the first year of the project, using both specialist and programmer time. The Windows version of VITIS will be discussed more fully later on, but the conversion to Windows allowed more effective weather graphing and user interface.
Incorporate vineyard description and disease history in decision support. The only way to arrive at a site-specific disease management decision is to consider all relevant vineyard and disease factors that contribute to disease risk. Growers requested that these factors be included in the VITIS expert system, and we did include them; their importance to grape disease management will be discussed in more detail later in this report. Plant pathology specialists from the eastern United States participated in the revision and development of disease management logic over the course of the project.
Predictive, site-specific weather input. The growers recognized the need to input observed weather from multiple sources into the VITIS expert system to use in daily decision-making. The growers also needed a way to input hourly weather forecasts into VITIS to predict disease risk more accurately. So, in addition, they asked for site-specific, predictive weather information. In most disease predictive models, historical weather information is used to predict the current disease risk; here, growers wanted a truly predictive capability that used future, one-to-three-day, hourly-predicted weather parameters to predict disease risk over the next several days. In addition, they wanted the predictive weather information to be specific for their own vineyards.
Ground station weather monitors provide current weather information from individual vineyards to the VITIS expert system. The VITIS expert system has also been programmed to accept weather inputs from the major ground station weather monitors (Sensor Field Monitor, Metos, Campbell Scientific, NE Weather Association, and Watch Dog). The weather data-keeping system within VITIS can be easily programmed to accept data from new weather information sources as they become available.
Site-specific weather forecasts. A private company, SkyBit, Inc., worked with the principal investigators over a three-year period to develop and refine the input of hourly historical and predictive weather information into VITIS. Once the format and input parameters were developed, the predictive weather inputs were evaluated at the field level for accuracy over a two-year period. Several revisions of the original predictive weather models were used to improve the final product. Site-specific predictive models make use of latitude, longitude, and elevation to calculate the information for each farm.
At the growers’ request, VITIS now also accepts both ground station and site-specific weather information in the same weather-operating file. This capability allows the grower to use weather information that was recorded by a ground station monitor in his or her own vineyard, but there is also capacity to use information that has been calculated for the site and transferred to the user over the Internet. Ground station data is entered into the grower’s PC based on the output method used by the ground station equipment; predictive weather data is received and entered through e-mail. Once the weather data is in the grower’s computer, it can be quickly loaded into the VITIS program, where it is summarized and used in decision support. The use of this type of weather data in VITIS will be explained more fully later in this report.
VITIS “light” or “full.” The growers felt it was important to have VITIS operate in two different consultation modes, light and full. The light mode satisfied the growers’ wish that VITIS offer information within, as the growers put it, the “time it takes to drink a cup of coffee.” This quick-access feature within VITIS has been developed, and it lets growers quickly get a summary of current, site-specific, and predicted weather parameters, have the weather information interpreted through disease predictive models, have VITIS calculate estimated disease risk if no action is taken, and, finally, weigh management options. These tasks can be accomplished within the cup-of-coffee time requirement.
The full access mode allows these same functions to be performed, but also provides additional information on grape disease cycles, management options, and disease risk assessment that takes into consideration specific vineyard circumstances, weather, and disease factors.
VITIS Operation Manual. A manual was developed that provides a step-by-step description of how to operate VITIS. Since VITIS is designed to operate on the growers’ own PC, it provides instruction in loading the program, setting it up, running it, and solving problems.
Results
VITIS is designed to operate on a PC and provide support to the grape grower in making site-specific decisions. The program is available on floppy disks, zip disks, or CDs.
The light VITIS expert system. As stated previously, growers wanted VITIS to offer two operating modes, light and full. The light mode was designed for quick access and the full mode was designed to give a broader presentation of decision support information. In the light mode, the grower can load current and predicted weather—weather data is either measured from a ground station monitor or is remotely sensed and received via e-mail. Data from seven different types of acquisition systems can be loaded and translated; VITIS uses hourly weather inputs as a basis for daily weather summaries, pathogen life stage development, disease predictive models, and calculation of remaining fungicide residues since the last pesticide application.
Because VITIS allows both ground station and remotely sensed weather information to be loaded into the same weather file, it provides the grower with on-site observed information as well as site-specific predictive information. A quick summary of the current weather, predicted weather, and disease model can be output as color graphs or obtained the light mode. Also in this light mode, the user can change the current date and time within VITIS without changing the PC’s date and time; this offers the user the option of running what-if scenarios of future or even past disease events and to review VITIS outputs.
In the light mode, it is also possible for the user to run just the downy mildew model, a complex disease predictive model that was developed by Cornell University. It will be explained in more detail in the explanation of the full system mode; from the light mode the user can switch to the full mode by selecting specific vineyard recommendation.
There are two grayed-out buttons on the light mode input screen: scan vineyard alerts and demonstration. These are ideas for future VITIS development. Right now, VITIS considers one vineyard management block at a time; the scan vineyard alerts function will allow a grower to scan all the vineyard blocks entered into VITIS to determine which blocks may be in immediate need of attention. This will save the grower time, since the he or she will not have to run each block individually through VITIS to determine its status. The demonstration button will link new users to a demonstration of all parts of the VITIS expert system.
The full VITIS expert system. The full VITIS mode is accessed through the specific vineyard recommendation on the light screen. The full expert system uses more vineyard culture and disease information and accesses deeper logic to provide more management options than the light system. The first question asked of the user upon entering the full expert system is whether the user would like to load a profile or a weather file: These are time-saving features that allow the user to load a large amount of information quickly about the specific vineyard block. The profile provides VITIS with vineyard-specific information, most of which can be created and entered into VITIS during the off-season so that there is no need to re-enter the information each time VITIS is consulted.
This profile contains specific information like vineyard location. Since VITIS has been developed to operate in Michigan, Ohio, Pennsylvania, and New York, it must know where the vineyard is. Pesticide labels vary by state, and VITIS adjusts the potential list of fungicides based on the state.
The profile also contains the vineyard age—VITIS is designed to consider the management of disease in mature bearing vineyards—and vineyard use. The management of disease on processing grapes verses wine grapes or fresh grapes is very different due to the greater susceptibility and value of wine and fresh grapes. Also included is the pruning date, a critical piece of information in preventing Eutypa dieback on grapevines, since infection occurs through fresh pruning wounds. This module is under development but not yet functional.
The profile also has information on vineyard infection resistance: The gray mold fungus, Botrytis, has developed resistance to Rovral in some vineyards, and VITIS will not recommend Rovral if Botrytis resistance is known to occur in a vineyard. The phenology profile is the developmental growth stage of the grape vine, and is used in many decision support functions in VITIS such as disease risk thresholds for specific pathogens, host susceptibility, and fungicide application limitations. Downy mildew dates calculate the risk of infection by developing a complex model of incubation, sporulation, and risk of infection on an hourly basis throughout the season. Bud break and the five to six leaf stage dates are critical to accurate modeling of the disease. The profile data of the harvest date accounts for fungicide days-to-harvest limitations, thus preventing illegal residues at harvest. VITIS stores the days-to-harvest limitations of all the fungicide labeled for use on grapes, along with any differences between the states, and will not include any fungicide on its list of options that would violate the days-to-harvest limitations. The harvest date can be adjusted throughout the season as estimates change.
Another component of the profile is the pruning method. Processing grapes are pruned either by hand or mechanically, and mechanically pruned grapes have a much higher potential for disease development. VITIS adjusts its management strategies based on pruning method. VITIS also collects the disease history, which can determine the likelihood of additional disease development and the rigor of disease management tactics required. VITIS requests information on the occurrence of all the major fruit and foliar diseases of grapes for the past and current season.
Profile information is collected on which fungicides have been sprayed and the fungicide application dates. VITIS calculates the potential activity of the fungicide since the last application date based on plant growth, weather conditions, disease risk factors and specific fungicide properties. Another factor is grape varieties—there is great variability of disease susceptibility between cultivars, and VITIS has a rating of each cultivar’s susceptibility to each disease.
VITIS uses this profile information, along with other factors collected from the user, to estimate disease risk for the vineyard. An example of one of these other factors is the affect of berry moth damage on Botrytis infection. VITIS will automatically ask if berry moth damage exists in the vineyard as Botrytis risk increases near harvest. If berry moth damage has occurred, VITIS will increase the disease risk potential above the foundation risk level provided by all other factors. All possible vineyard, management, disease and weather factors are considered when assigning individual risk levels to the fruit and foliar diseases managed within VITIS.
Once the profile has been loaded into VITIS, the grower has the option of loading a weather file and acquiring the same weather summaries as described under the light access mode. It is also possible to view, edit, graph, and print the weather files or the summary grapes.
The current disease risk for the vineyard can be obtained by selecting individual diseases or query all diseases from the diseases menu. VITIS then displays a graph of the potential disease level in the next two weeks if no action taken. Potential risk levels range from none to high., and the graph provides a summary of the current risk level of all the potential diseases.
Impacts
VITIS has several potential impacts on disease management and fungicide use in vineyards.
Most importantly, VITIS supports the replacement of fungicides with knowledge. Without the careful consideration of vineyard, weather, and disease factors, a grower will often spray a fungicide to prevent the imagined threat of disease even if none exists.
VITIS warns of disease risk. If there is a potential for disease, the grower is warned in time to take appropriate management actions to prevent disease buildup in the vineyard. The grower can also be advised about the potential fungicides that may be used. There are many possible cultural strategies growers can use to limit disease, but when a disease epidemic is predicted by VITIS a grower has few options but to rely on the use of a fungicide.
VITIS assists the grower in the wise use of a fungicide. Not every circumstance requires the most effective, most specific, and often the most expensive fungicide available, and VITIS displays all possible fungicides that would be effective against the current risk. When a fungicide is selected, the other fungicides on the list that have the same mode of action are deselected so that two materials that have the same mode-of-action on the same fungicide are not needlessly applied at the same time. VITIS lists only those fungicides that are effective in controlling the diseases at the current disease risk levels. Often, less expensive fungicides like sulfur or copper can be applied with very satisfactory results; it is also true that these same fungicides can be applied uselessly, in the wrong situation, with crop loss as a result. When VITIS provides a list of the appropriate fungicides for the situation, the grower is then free to select their fungicide of choice. Another risk in selecting a fungicide is that, unless there is complete knowledge of the modes of action, multiple fungicides with the same mode of action may be chosen, wasting the duplicate material. In addition, since fungicides vary greatly in their ability to control specific diseases, it is possible that important diseases will not be controlled even by a combination of fungicides. VITIS directs the grower to combine fungicides as a fungicide resistance management strategy.
VITIS calculates the rates appropriate for the disease risk. When the grower has made the fungicide selection, VITIS bases the rate calculation on disease risk factors and fungicide activity and longevity. Rates that are too low will not satisfactorily control the disease and may stimulate resistance to the fungicide by the targeted fungus. Fungicide rates that are too high for the disease risk situation waste material and may needlessly put pressure on the fungal population to become resistant. Both low and high rates will result in additional fungicide applications to the vineyard.
The grower and VITIS form an integrated disease management team. One of the most common comments made by growers after using VITIS is that they eventually start to understand and anticipate the logic used by VITIS in making decisions. Over time, they rely less on the logic provided in VITIS and more on their own logic and their own far more complex understanding of the situation than is provided in VITIS. The final end product is a grower who uses the computational strengths of a computer system to summarize weather data and run mathematical disease predictive models combined with their own understanding and logic in managing all aspects of a vineyard.
Economic Analysis
Weather information costs. General weather information is available to growers from several sources—radio, TV, newspapers, and the Web. However, none of these sources of information can provide hourly, site-specific weather information that is needed by growers to make effective disease management decisions. To be useful, weather information must be specific to individual farms. To address one of the objectives of this study, commercially available environmental monitoring equipment was evaluated for accuracy, dependability, and proficiency in putting the data into VITIS. The results of this evaluation indicate that ground station weather monitors can be helpful to provide historical weather information. There are concerns over the accuracy of some of the commercially available units.
An alternative source is remotely sensed weather information. Remotely sensed information is calculated for specific locations using satellite, radar, and national weather service environmental data inputs. SkyBit cooperated with project to develop and test remotely sensed weather information in VITIS and to evaluate its usefulness to growers. The weather information is much more useful for the growers; the accuracy of the information is comparable to a ground station unit placed in the vineyard.
Ground station units cost about $1,500 to $7,000 to serve one site. They require about three hours of maintenance and data collection time each week. These units are used by growers over a seven-month period during the growing season. The cost to the grower of operating a ground station unit, after the initial purchase, is about $900 a season, or three hours over thirty weeks at $10 an hour.
Remotely sensed SkyBit weather information is received over the Internet each day and is specific for the site. SkyBit costs the grower $75 a month; there is little additional labor required for SkyBit, since it is placed directly from the Internet into VITIS. The cost for SkyBit is about $525 a season, or $75 a month over seven months.
A combination of ground station and SkyBit information can be combined in the same VITIS weather file, allowing the grower to use their own ground station information and SkyBit weather forecasts for their site. VITIS has been designed to operate on either system or on both. Operating costs would be a combination of the costs of both options.
Fungicide reduction savings. VITIS can save the grower money through reduced fungicide application. One of the chief problems faced by a grape grower is knowing when a wetting period increases the risk for disease development and when disease risk is low. Because VITIS estimates risk for individual vineyards, growers know when to ignore wetting periods, which reduces fungicide application. Often, when a grower is uncertain about a disease risk potential, a fungicide is applied to avoid crop loss:
If a grower saves only two fungicide applications per season, the savings on a 50 acre vineyard would be about $5,000 a season, or $50 and acre times two applications times 50 acres. Two fungicide applications saved per season is a reasonable expectation according to research trial results and grower feedback.
Reduction in crop loss. Crop loss in grapes is not uncommon, even when the grapes are repeatedly sprayed with fungicides. Grape diseases are highly destructive to the fruit, with diseases infecting both green and mature berries. Through the wise use of weather information, disease predictive models, and vineyard susceptibility evaluations, VITIS has the potential to reduce crop loss in wine grapes by 10 percent per season. Processing grape losses are normally less but do occur. Both wine and processing grapes are valued at about $1000 an acre; if we consider the average grape vineyard acreage to be 50 acres, savings due to reduced crop loss from the use of VITIS can be estimated at about $5,000 a season, or 10 percent crop loss times the $1000 per acre crop value times 50 wine grape acres.
Overall savings. The potential yearly savings for a 50-acre vineyard using VITIS and growing wine grapes is $9,475, or the savings of two fungicide applications ($5,000), the anticipated 10 per cent crop increase ($5,000), less the cost of SkyBit ($525) for the season.
The peace-of-mind factor. An additional benefit for the grower is less tangible but perhaps the most important. All the growers involved in this study have said that spraying fungicides is the least rewarding aspect of growing grapes. They all agreed that they would not spray a fungicide if they could feel comfortable that disease would not destroy their crop, and VITIS provides that peace of mind. VITIS considers all aspects of their vineyard and its current susceptibility to disease; it clearly displays disease risk for their vineyard; it uses current and predicted weather information and the best mathematically available disease predictive models. VITIS uses knowledge to replace fungicides, and does it in a way that reduces the use of fungicides. For growers, this peace-of-mind factor appears to be priceless.
Areas Needing Further Study
A decision support system for growers is only begun, it is never finished, and the growers cooperating in this study have many more ideas that would make the VITIS expert system even more useful to them. These ideas include integrating record keeping, web delivery, and continual updating, and economic cost estimates of disease management options. As the VITIS system becomes widely adopted by grape growers, its function will be expanded as time and resources permit. New disease predictive models are continually being improved and will be incorporated into VITIS once they have been validated in vineyards. The developers of VITIS—grower cooperators and university specialists—have decided to field-test VITIS one more season. Beta-test grower groups will be formed in Pennsylvania and New York to test VITIS’ output recommendations, and VITIS is scheduled for commercial release in 2002.
Outreach
VITIS users manual. As part of the objectives of this grant, a users manual for VITIS has been developed. A copy of the manual is included with this report. Revisions to the manual will continue as growers identify areas for improvement.
Workshops. Training workshops will be provided to new users to acquaint them with VITIS; these will begin with the Beta test group during the winter of 2001.There will be an initial training session in March, with follow-up training in April and May. Feedback sessions will be held at the middle and end of the growing season.
Local User Groups. The growers cooperating in the Beta tests this season will be asked to form and coordinate local user groups in their area when VITIS is released in 2002. This will provide new users with the opportunity to interact with experienced users and other new users as they learn to fully use the system. To raise awareness of VITIS, there will be presentations at grower educational meetings this winter. Some of these meetings include the multi-state fruit meeting being held in Hershey, Pennsylvania in late January 2001, at the Partnerships for Sustaining California Agriculture: Profit, Environment, and Community conference to be held in Woodland, California in March 2001, and at smaller grower meetings held across the region.
Reported January 2001
Collaborators:
NY