smpsaoc.gif (8026 bytes)
The chemical management module also functions at several levels. The module is subdivided into three submodules: 1) the chemical selection module; 2) the chemical rate module and 3) the spray interval module.

The overall output of the module is the determination of the chemical, its rate and the spray interval until the next application.

 

The objective of this module is to determine which insecticides, acaricides and fungicides registered for use on apples would be effective in the control of the pests determined to be a problem by the expert system. The control and use characteristics of each chemical is described within the dependency networks. Critical factors which contribute to the selection of a chemical are: current phenology of the orchard, the array of insects over threshold and disease levels in the orchard, estimation of pathogen infection potential since the last spray, pesticides applied in the last application, known fungicide or insecticide resistance in the orchard, and the number of days until the orchard will be harvested. The appropriate use of benomyl for the control of apple scab is described as an example and is shown in Figure 8. There are two basic scenarios which describe the use pattern for benomyl. One describes benomyl's use as a protectant fungicide, the other describes benomyl's use in a post-infection situation. As a protectant benomyl is recommended from green tip to second cover, when apple scab potential levels are from severe to moderate, there is no infection potential since the last application of fungicide, a sterol inhibitor was not applied in the last application as the first application following an infection period, no benomyl resistance is known in the orchard, and the harvest date for the orchard is greater than 7 days from the present. The post-infection scenario described by the network is the same except that the apple scab disease potential as determined by the system is high or low post-infection and the beginning of the infection period which occurred since the last application is not more than 24 hours from the current time. The output of the chemical module is a list of insecticides, acaricides and fungicides which will control the pests determined by the system under the current conditions.

 

Once the chemicals have been selected the system must determine which rate of chemical to recommend. This again is described in dependency networks. For example, fungicides recommended for apple scab control early in the season are generally recommended in combination to prevent resistance and to take advantage of two modes of action. The rate for in-combination use are high, moderate and low. Later in the season fungicides, generally protectants, are recommended alone. The high fungicide rate is recommended from tight cluster through prepink, when the apple scab severity level is determined to be high, and the chemical is to be recommended in combination with another fungicide. A table containing the actual rates of each chemical is built into the system. Within the system, Nova (myclobutanil) is an exception to this method of rate determination. Since Nova is labelled according to tree row volume calculations, it is recommended within the system according to tree row volume of the orchard and the disease potential in the orchard. Tree row volume is automatically calculated from information in the profile.
The system also recommends the date of the next pesticide application. The factors used to adjust the spray intervals are phenology of the orchard, disease potential rating, occurrence of an infection period since the last spray, grower's spray interval preference, rainfall since the last application and the chemical applied in the last application. Early in the season, the application timing is mandated by the requirements to control apple scab. After second cover, insect control requirements and weather conditions which affect summer apple diseases dictate application timing. Grower application preferences are also considered due to the difference in the ability of growers to spray an orchard at short notice and the necessity for some growers to plan spray applications on a routine schedule in order to manage labor. This option allows some growers to follow completely variable spray intervals based on pest and weather circumstances and other growers to plan applications and be able to schedule labor requirements. Even if a routine schedule is selected, spray intervals will be adjusted if post-infection situations occur. The effect of rainfall on chemical residue is included in the system based on heuristics. Growers have for a long period of time used the rule-of-thumb that less than 1" of rain will not affect spray residue, 1 to 2" will reduce spray residues by one half and 2" or more of rain will remove all of the spray residue . As an illustration, the next pesticide application is recommended within 5 days of the last application if the orchard is green tip to petal fall, the apple scab severity rating level is severe to low, there has been no infection potential since the last appication, there has been 1 to 2" of rain since the last application, a sterol inhibitor was not applied in the last application as the first application following an infection period, and the last application was applied as an alternate side application. All possible scenarios which describe situations which lead to spray intervals are described within dependency networks.

 

Once the pesticides have been selected they are checked for compatibility. Every chemical included within this system is compared before the recommendation is given. Dependency networks are used to describe which chemicals should not be mixed. Likewise, recommended chemicals are checked against the computer's calendar to determine if it is too close to harvest for application.

Back to Expert Systems Development Group Main Page