Weed Management Decision Support Software

Making weed management decisions that are environmentally and economically sound is a complex task. To help farmers, consultants, and extension personnel in making these decisions, lab personnel have developed three computer programs: the Windows-based program HADSS^TM , Herbicide Application Decision Support System, the Windows CE-based Pocket HERB that runs on small palmtop computers, and WebHADSS which can be accessed over the Internet. Weed scientists who have an interest in developing a version of these programs for another crop or location should contact Dr. Wilkerson.

Implementing a Weed Management Decision Support System Across the South

With funding from USDA, Dr. Wilkerson has coordinated a project to implement HADSS in 10 Southern states. Weed scientists in each state have modified program databases as necessary to reflect regional differences in weed problems and weed management strategies. Field validation and demonstration trials were conducted across the South. HADSS is now made available to farmers, consultants, and extension personnel in each participating state. Right: A training session conducted by Wilson Faircloth at Auburn University in Spring 2001. Bridget Robinson serves as project director.

Site-Specific Weed Management

Numerous studies have shown that weed distribution is not uniform across a field; weeds tend to be clumped together in patches. This variability provides tremendous potential for improving overall weed management, increasing economic returns, and reducing application of herbicides through site specific management. As part of a mid-Atlantic regional project funded by the Foundation for Agronomic Research and the United Soybean Board, Dr. Wilkerson has been conducting research on site-specific weed management. Left: David Krueger (Ph.D. student) maps weed populations at one research site.

Lab personnel have developed a system for collecting weed scouting data, creating maps of weed populations across the field, determining the best herbicide treatment for each section, and generating a treatment map that can be used to direct a variable rate herbicide applicator. Right: This system has been field tested at the Center for Environmental Farming Systems in Goldsboro, NC and at the Caswell Farm in Kinston, NC, as well as on two growers farms.

Soybean Production Aids

In cooperation with Dr. Jim Dunphy , programmers in Dr. Wilkerson's lab have developed two Web-based applications for soybean extension agents and producers. The first, SoyVar, is designed to present information about soybean cultivars commonly grown in North Carolina. The user can search for varieties with particular characteristics.

Modeling Soybean Response to Variability in Soil Series Characteristics

Modeling soybean yield range for a taxonomically defined soil series first requires numerical definition of a series  range of characteristics (graph at left). Taxon ranges may be generated from the rules of soil taxonomy and data available from NRCS databases. In this project, thousands of soil profiles representing the typical and series-defined range of profile characteristics have been generated for several soil series. The CROPGRO-Soybean model has been used to estimate yield potential for each of these profiles over 30 years of recorded weather in North Carolina and Iowa. This cooperative project with Dr. Stanley Buol in the Soil Science Department should provide information to aid in soil-specific management decisions such as variety selection, planting date, and irrigation.

Database Management for Crop Simulation Model Files

Crop growth models are being used around the world to help explain and predict crop response to soil characteristics, weather patterns, pest infestations, and management strategies. Due to the complexity of these models, most require large amounts of information in order to correctly simulate crop growth throughout the season. Lab personnel have been working for a number of years on ways to facilitate creation, organization, and maintenance of the data files required by these models. Work has focused on developing relational databases to store large amounts of data, and computer programs to access and use the data. These programs help to organize data as needed by the crop simulation models, filter information to create simulation files as needed, automate simulations, analyze results of these simulations, and display results in tabular and graphical form.

Current Project Group

Gregory S. Buol, Applications Programmer, development of database management tools and crop models.

Warren D'mello, Research Assistant development of Web-based applications.

Jenifer Jordan Web Applications Engineer, Development of Turffiles web site.

David W. Krueger, Ph.D. graduate student, development of scouting protocols for weed management decision-making.

Bridget L. Robinson Research Assistant, project leader for implementation of weed management decision support system across the southern region.

Michael C. Sturgill, Research Assistant, development of weed management decision aids.

Zhengyu Yang, graduate student, nitrogen application decision support model for corn.