Jump to main content.

Pesticide Risk Reduction Utilizing the PEET Multiple-Objective Decision-Support System

D.L. Nofziger; Don Murray, Case Medlin, Jinquan Wu

Oklahoma Agricultural Experiment Station
Oklahoma State University

Department of Plant and Soil Sciences
368 Agriculture Hall
Stillwater, OK 74074

(405)744-9592

(405)744-5269

dln@okstate.edu

Executive Summary

Pesticide environmental stewardship requires decision-making. Farmers and pest management consultants generally make decisions about whether to apply a pesticide, what pesticide to apply, and when to apply it with little regard to the risk of these pesticides degrading groundwater quality. The Pesticide Economic and Environmental Tradeoffs (PEET) decision-support system can assist in this decision-making process (Hoag et al., 1994). PEET evaluates the economic return for using different labeled herbicide treatments for the soil and weed densities in the field of interest. It also simulates the leaching and degradation of the herbicides in the soil and calculates a groundwater hazard index as the ratio of the estimated concentration of the active ingredient to a critical concentration for that chemical. PEET enables a user to select a herbicide after considering the risk that product poses to groundwater and the economic benefit of the application. PEET was developed for use in any crop and geographic area. It is being used for peanuts and cotton in Oklahoma. Farmers and consultants are requesting PEET for other crops. Scientists in other states have expressed interest in using PEET as a weed control decision-aid, but more importantly for a groundwater protection tool.

This project will enable PEET to be used in an expanded area by (1) improving the software to enable real-time calculation of the groundwater hazard index so results match the practices of each farmer, (2) developing database management software to assist specialists entering and maintaining required data, (3) conducting a workshop for scientists interested in using PEET in new locations or crops, and (4) implementing PEET for additional crops.

The use of PEET can reduce pesticide risk. It can lead to fewer applications of herbicide, to the use of herbicides that are less likely to leach, and to irrigation practices that reduce leaching. This project will expand the use of PEET so these benefits will occur on a wider area.

Objectives

  1. Expand PEET multiple objective decision-support system to calculate groundwater hazard values associated with different weed control practices and management systems in real time.
  2. Develop database management software and documentation to facilitate management and rapid expansion of PEET databases for additional crops
  3. Expand PEET databases for use in additional cultivated crops and geographical areas of Oklahoma
  4. Conduct a workshop for specialists in other states interested in developing databases needed for using PEET in their states and crops

Justification

  1. The probability of a pesticide leaching to groundwater in sufficient quantity to pose a hazard to organisms consuming that water depends upon the soil and management practices used as well as the pesticide and organism of interest. In the current version of PEET, these hazards are calculated using Monte Carlo techniques for a small set of management options. The calculated groundwater hazards are stored in a database for use in the interactive program. Current personal computers have the speed to make these calculations in real time. Implementing this real-time calculation will enable the simulation to be carried out for the specific management practices employed by the user in the field of interest. Thus the results will provide better risk estimates for the actual practices used. A secondary benefit of this approach is that the area specialist developing the decision-support system for a new area, crop, or herbicide will not need to perform all of the calculations before releasing PEET for the new area and crop. This will decrease the time required to release the decision-support system for the new crop or area. Also the amount of data to be transferred from the internet server to the user’s computer will be reduced.
  2. The database for PEET that underlies its use in cotton and peanuts in Oklahoma contains 25 related tables and occupies more than 80 MB in a relational database management package. There are many relationships between the data tables and many requirements must be met to assure integrity in the values stored. Up to this point, we as developers have also managed this database. PEET was written to be usable in different areas by entering area-specific data on weeds, management practices, soils, and weather. To make the package truly usable in other states, additional software must be developed to ease the construction and testing of these databases. That software and associated documentation will be developed in the first six months of this project. The software will also enable specialists to keep the data current by adding or deleting herbicide treatments, modifying rates, or adding warning messages that accompany each treatment.
  3. PEET is a useful decision aid for several reasons. (1) Currently farmers select herbicides with little, if any, information about the probability of that chemical leaching through the soil and degrading groundwater quality. (2) Research indicates that farmers often apply herbicides to soils when they may not be needed and the economic benefits are minimal. (3) Farmers have a choice of herbicides available for their use. They have requested information to guide them in selecting appropriate herbicides for the combination of weeds in their field. PEET provides information in all three of these areas so decision-makers can make more knowledgeable decisions. In many cases, improved economic returns and reduced risk of degrading groundwater quality can both be achieved by the proper selection of herbicide. We have implemented PEET for cotton and peanuts in Oklahoma. In carrying out this objective we will implement it for other crops including sorghum and wheat and for areas of Oklahoma in which they are grown. Thus the benefits of PEET will extend to a much larger group of decision-makers. A secondary benefit of this development is that regulatory agencies can use PEET to identify sensitive crop-soil-management systems to be monitored more intensely.
  4. As stated above, PEET was written to be usable in different areas by entering area-specific data on weeds, management practices, soils, and weather. The amount of data required is substantial, but it is generally available. Specialists in other states have expressed interest in using PEET in their crops and regions. To facilitate this, we will conduct a workshop to assist them in developing the databases required for their applications. We will provide additional support to them as they develop databases and implement PEET in their areas. At the end of the development phase, they will be able to publish PEET for their crops and geographic areas on internet servers in their states, or if necessary on our server here at OSU. The decision-support system is available without charge to the developer or end-user. This approach will result in the availability of PEET to more decision-makers and a broader geographical area. Scientists in other states have expressed interest in attending such a workshop.

Overall Summary

Groundwater is an important source of exposure of humans to pesticides. Thus reducing the risk of degrading groundwater quality can result in a reduced risk of exposure. Reducing the risk of groundwater quality degradation can be achieved by improved decision-making with regard to the choice of pesticide, the rate of application, and related soil – crop management practices. Farmers use pesticides to control pests and to improve economic return. Hoag (1990) demonstrated that in some cases reduced risk to groundwater could be achieved with the same practices that improve economic return. In other cases there is a tradeoff between these objectives. The Pesticide Economic and Environmental Tradeoffs (PEET) decision-support system (Hoag et al, 1994; Nofziger et al., 1998, Hornsby et al., 1998) was developed to assist in pesticide selection by providing both the economic impact and the risk of degrading groundwater quality for each potential herbicide treatment. This combination of objectives into one decision-support system makes PEET unique.

Numerous models and decision-support systems have been developed to assist farmers in selecting herbicides for weed control(Mortensen and Coble, 1991; Coble and Mortensen, 1992; Kropff et al., 1992;). Early work at North Carolina State University formed the basis for the HERB software (1986) and its successor HADSS (Herbicide Application Decision-Support System (1997), both from North Carolina State University. All of these tools focus only on pesticide selection without regard to risk of degrading groundwater quality. The economic component of PEET is based on this same concept as that used in HERB and HADSS. That approach has been tested by many weed scientists for different crops (Wilkerson et al, 1991; Monks et al, 1995; Rankins et al, 1998; Murdock and Murray, 2000; Murdock et al, 2001). Currently it is being used in numerous states including Alabama, Arkansas, Florida, Georgia, Louisiana, Mississippi, North Carolina, Oklahoma, South Carolina, Tennessee, and Texas.

Just as there are numerous models to aid in selecting herbicides for weed control, there are also numerous models and decision-aids for predicting the risk that pesticides pose to groundwater (without regard to their efficacy in controlling specific weeds). These range from basic screening tools (Goss and Wauchope, 1991; Hornsby et al, 1991; Weber and Warren, 1991) to full-fledged site-specific simulation models such as PRZM (Carsel et al., 1984), LEACHM (Wagenet and Hutson, 1986), CMLS (Nofziger and Hornsby, 1986), GLEAMS (Knisel et al., 1989) and PESTLA (Boesten and van der Linden, 1991).

In Oklahoma and Florida, the CMLS management model (Nofziger and Hornsby, 1986) was used to simulate the movement of each active ingredient through each soil on which the crop was grown. Since the amount leached is highly dependent upon weather sequences at a single site (Haan et al, 1994), historical weather records or stochastically generated daily weather data are used with Monte Carlo techniques to determine the probability distribution of amount leached to groundwater for each soil-chemical-management system. These results are stored in a database for use in the interactive program. The groundwater hazard or ratio of the predicted concentration of the active ingredient in groundwater to the critical concentration of the chemical is used to assess the impact of each treatment on groundwater quality (Hoag and Hornsby, 1992).

The CMLS model uses a daily time step to calculate water and chemical movement in the soil. Sorption is assumed to be linear, reversible, and instantaneous. First-order degradation of the chemical is assumed to hold. Soil and chemical properties can change with depth. Values of the organic carbon sorption coefficients and degradation rates for most herbicides can be found in Hornsby et al. (1996). Soil properties required in the model can be readily obtained from soil survey data. The CMLS predictions of chemical transport and fate have compared reasonably well to experimental results (Pennell et al., 1990, Macur et al., 2000).

Approach and Outcomes

Objective 1.
The CMLS model of Nofziger and Hornsby (1986) will be used for the real- time calculation. PEET is written in the Java programming language. Recently CMLS has been totally rewritten in Java as well. To achieve real-time calculation in PEET, the computational part of the interactive version of CMLS will be extracted and merged with PEET. The user interface will be modified slightly to allow the user to define the management practices and irrigation methods used in the field of interest. The simulation model will be run for each active ingredient in the collection of potential herbicides. Since the future weather at the site is not known at the time the decision must be made, the simulations will be carried out for many weather realizations for the site of interest. From this, the probability of exceeding different groundwater hazard values will be determined for each potential treatment. These values will then be summarized graphically as done in the present version of PEET. Databases of chemodynamic properties, weather data, and soil properties for each county will be prepared.

Objective 2.
Expanding PEET for use in additional crops and regions requires weed scientists to identify potential weeds that may infest the crop, the yield loss due to a unit density of each weed species, labeled weed control practices in that crop, and the efficacy of each practice for each weed species. These values are obtained using controlled experiments and knowledge gained over many years of experience. People using HADSS already have these data. In addition, the soil scientist must determine the soil map units on which the crop is grown, the properties of each soil, and daily weather databases for the areas of interest. The pesticide databases must also be updated to include all of the active ingredients for the potential treatments. This includes their sorption coefficients, degradation rates, and toxicities. These data are then stored in appropriate databases for use in the PEET software. In related research, experimental evaluations of treatments ranked highly by PEET will be compared with traditional treatments.

The data management software required must facilitate data entry, quality control, and compression for distribution. It will be centered on a well-designed database with customized interactive software to enter and edit the data. The current structure of this database is shown in Table 1. By using this approach, individuals with data already in electronic form can import their data into the database without manually entering it again. To insure quality control, additional software will be developed to examine each piece of data to see if it is within a reasonable range of values. Software will also be developed to insure that all required data are included for each soil and each potential treatment. Finally, custom software will be developed to convert the data in the database format into a compressed form for use in the interactive PEET program.

Objective 3.
Weed competition and herbicide efficacy studies have been an ongoing part of research conducted by weed scientists here at Oklahoma State University. Those data for wheat and sorghum in Oklahoma will be collected and entered into the database shown in Table 1. Soil properties for each cultivated map unit will be obtained from the Oklahoma Soil Survey. Daily rainfall and air temperature data from at least one weather station will be selected for each county in which the crops are grown. Extension specialists will inform farmers and consultants on the availability and use of the software.

Table 1. Overview of database tables required for implementation of PEET
Table Contents
ApplicationTypes List of all possible types of applications for all crops
ApplicationTypes ForCrop List of application for specific crops
AIproperties Chemical properties for each active ingredient
Cost Cost of each potential herbicide
Counties Names of counties in which any crop is grown
CountiesWithCrop List of counties in which each crop is grown
Crops Crop name and related data for each crop in database
EfficaciesPost Efficacy values for each weed for post emergence treatments
EfficaciesPPI Efficacy values for each weed for preplant incorporated and pre-emergence treatments
GroundwaterHazards Calculated groundwater hazards for each soil, treatment, crop, irrigation practice, weed size, application date, and 5 probability levels. (This table will not be needed when real-time calculation is implemented.)
Herbicides Herbicide names and application rates for all treatments in all crops
IrrigationTypes List of potential irrigation types
OrganicCarbonClasses Names and ranges for soil organic carbon classifications
SoilMoistureLevels Names and abbreviations for soil moisture classes
SoilNames Names and generalized properties of each soil in each county
SoilProperties Soil properties needed by CMLS model for each layer of each soil
TextureTypes List of soil texture descriptions and abbreviations
TillagePractices List of potential tillage practices used in crops
TradenameDictionary Active ingredients and their concentrations for each tradename
Treatments Details of application date, depth, and efficacy index for each treatment for all crops
TreatmentsForCrop List of treatments used on a specific crop
Units Conversion factors for different pairs of units
WarningCodes Warning message codes and treatments for which that message is to be displayed.
WarningMessages Message associates with each warning code
Weeds List of weeds grown in any crop
WeedsInCrop List of weeds grown in a specific crop and its competitive index in that crop
WeedSizes Abbreviations and associated weed sizes
Weather Names of weather stations that can be used in simulating chemical movement and associated data files

Objective 4.
A training workshop will be conducted for specialists interested in utilizing PEET in their areas and for their crops. The workshop will be organized in two parts. The first part will explain the underlying concepts used in PEET, the nature of the data required, and how the data can be obtained. Material presented will be developed in a form that can be placed on the internet for use by an expanded audience after the workshop. The second part of the workshop will focus on how the data are entered into the computer, how to edit existing databases, and deployment of the software and databases on the internet. Sensitivity analyses will be carried out and incorporated into the training to enable the scientists to understand the accuracy required for each parameter. Modern computer laboratories at Oklahoma State University will be used for the hands-on component of the workshop. The workshop will be limited to twenty participants due to limited computer space. Participants will likely come from states in which the competitive load model used in HERB, HADSS and PEET has been tested so the weed data will already be available.

Impact Assessment

Several measures of the impact of this project can be used. These include the increase in the number of crops and land area on which PEET can be used, the number of scientists implementing PEET in their areas, and the increase in the number of times the software is downloaded from the internet. As the decision-support system becomes more widely used, qualitative information on its impact can be obtained from reports of interactions of extension personnel with farmers. The databases created for use in PEET can be used to identify soil-crop-herbicide-management systems that are particularly likely to degrade groundwater quality so special research and educational efforts can be directed to those areas to reduce these risks.

Literature Cited

Coble, Harold D. and David A. Mortensen. 1992. The threshold concept and its application to weed science. Weed Technology 6:191-195.

Haan, C.T., D.L. Nofziger, and F.K. Ahmed. 1994. Characterizing chemical transport variability due to natural weather sequences. J. Environ. Qual. 23:349-354.

Hoag, D. and A. G. Hornsby. 1992. Coupling groundwater contamination with economic returns when applying farm chemicals. J. Environ. Qual. 21:579-586.

Hoag, D. L., A.G. Hornsby, and D.L. Nofziger. 1994. PEET: Pesticide Economic and Environmental Tradeoffs. Proceedings of the 5th International Conference on Computers in Agriculture, Orlando FL.

Hornsby, A. G., R.D. Wauchope, and A.E. Herner. 1996. Pesticide Properties in the Environment. Springer. New York. 227 p.

Hornsby, A.G., D.L. Hoag, and D.L. Nofziger. 1998. Pesticide Economic and Environmental Tradeoffs (PEET): Users Perspective. pp. 75-81. In El-Swaify, S. and D. Yakowitz (eds.) Multiple Objective Decision making for Land, Water, and Environment. St. Lucie Press Corporation, Delray Beach, FL. 743 pages.

Kropff, M. J., S.E. Weaver, and M. A. Smits. 1992. Use of ecophysiological models for crop-weed interface: Relations amongst weed density, relative time of weed emergence, relative leaf area, and yield loss. Weed Sci. 40:296-301.

Macur, Richard E., Hesham M. Gaber, Jon M. Wraith, and William P. Inskeep. 2000. Predicting solute transport using mapping-unit data: Model simulations versus observed data at four field sites. J. Environ. Qual. 29:1939-1946.

Monks, C. Dale, David C. Bridges, John W. Woodruff, Tim R Murphy, and Daniel J. Berry. 1995. Expert system evaluation and implementation for soybean (Glycine max) weed management. Weed Technol. 9:535-540.

Mortensen, David A. and Harold D. Coble. 1991. Two approaches to weed control decision-aid software. Weed Technol. 5:445-452.

Murdock, S.W. and D.S. Murray. 2000. Adaptation of a computer decision support system (DSS) to Oklahoma peanut production. Proc. South. Weed Sci. Soc. 53:153.

Murdock, S.W., D.S. Murray, and J.W. Moore. 2001. Weed control and net returns with transgenic cotton using DSS and human recommendations. Proc. South. Weed Sci. Soc. 54:31.

Murdock, S.W. and D.S. Murray. 2002. Obtaining weed populations for computerized decision support system (DSS) inputs, counts versus estimations. Proc. South. Weed Sci. Soc. 55:130

Nofziger, D.L. and A.G. Hornsby. 1986. A microcomputer-based management tool for chemical movement in soils. Appl. Agr. Res. 1:50-56.

Nofziger, D.L., A.G. Hornsby, and D.L. Hoag. 1998. Pesticide Economic and Environmental Tradeoffs (PEET): Developers Perspective. pp. 83-92. In El-Swaify, S. and D. Yakowitz (eds.) Multiple Objective Decision making for Land, Water, and Environment. St. Lucie Press Corporation, Delray Beach, FL. 743 pages.

Pennell, K.D., A.G. Hornsby, R.E. Jessup, and P.S.C. Rao. 1990. Evaluation of five simulation models for predicting aldicarb and bromide behavior under field conditions. Water Resources Res. 26:2679-2693.

Rankins, Alfred Jr., David R. Shaw, and John D. Byrd, Jr. 1998. HERB and MSU-HERB field validation for soybean (Glycine max) weed control in Mississippi. Weed Technol. 12:88-96.

Wilkerson, G.G., S.A. Modena, and H.D. Coble. 1991. HERB: Decision Model for Postemergence Weed Control in Soybean. Agron. J. 83:413-417.

Timetable

December 31, 2003:

June 30, 2004:

December 31, 2004:

June 30, 2005:

Major Participants

Scientist Role in Project

David L. Nofziger, Professor
Department of Plant and Soil Sciences Oklahoma State University
Speciality: Soil Physics
Coordinate overall project
Design and test software for real-time simulation
Design and test data management software
Write data management documentation
Develop and present portions of workshop materials

Don S. Murray, Professor
Department of Plant and Soil Sciences Oklahoma State University
Speciality: Weed Science
Lead weed science component of database development and testing for sorghum
Develop portions of workshop materials and lead workshop

Case R. Medlin, Assistant Professor
Department of Plant and Soil Sciences Oklahoma State University
Speciality: Weed Science
Lead weed science component of database development for wheat
Develop and present portions of workshop material

Lead educational activities to PEET end-users

Jinquan Wu, Research Associate
Department of Plant and Soil Sciences Oklahoma State University
Speciality: Soil Physics

Lead development and testing software
Supervise other software developers
Lead in the development of soil databases required for sorghum and wheat
Perform sensitivity analysis
Prepare manuscript of sensitivity results

Project Budget

Funding Request
Funding Requested Other Funding Total Funding
$39,935
$31,023
$70,958

Local Navigation

Jump to main content.