Submitted to: 1999 Georgia Cotton Research and Extention Report
Publication Type: Experiment Station
Publication Acceptance Date: 8/1/1999
Publication Date: N/A
Citation: Webster, T.M., Culpepper, A.S., Bennet, A., Wilkerson, G.G. 1999. Validation of cotton-herb in Georgia - year 1. 1999 University of Georgia Cotton Research and Extension Report. p.38-42. Interpretive Summary: There is a great wealth of information that currently exists concerning weed management and weed biology, however it is in a format that is largely inaccessible to growers. HERB is a weed management decision support system that transfers this information to the growers. HERB is designed to provide a grower assistance in making weed management decisions. HERB asks for minimal user inputs: expected crop yield and selling price, weeds present and estimated densities. HERB calculates the expected yield losses based on the weeds present and attaches a dollar value to that loss. Next HERB evaluates the most economical management options. A field study confirmed that HERB-Cotton recommended weed control programs that effectively controlled weeds and resulted in yields similar to those in systems recommended by Dr. Culpepper. To further test Cotton-HERB, 60 scenarios were provided to our Georgia extension weed specialists. The recommendation of HERB was the same as one of our two experts in 42% of the scenarios. Upon further inspection, we found that our experts advised a tank mix application of MSMA with each of their control recommendations. HERB recommended MSMA in only 40% of the scenarios. If the HERB database was altered to include MSMA applications, then the agreement rate between HERB and our experts approached 72%. Our goal is to continue to make subtle changes to the database that will allow for greater similarity between our experts and HERB, which will be the true validation of this system.
Technical Abstract: HERB is a computer program that allows growers to access and utilize weed biology and weed management data that is largely unavailable to them in any other format. A field study was conducted in Tifton, GA to validate Cotton-HERB, a weed management decision support system. Three cotton management systems were utilized, glyphosate-tolerant cotton, bromoxynil-tolerant cotton, and non-transgenic cotton. In each system there were five treatments: 1. standard preemergence (PRE) herbicide (pendimethalin plus fluometuron) followed by (fb)HERB recommendation postemergence (POST), 2. Standard PRE fb Dr. Culpepper's POST recommendation, 3. Standard PRE fb HERB POST, 4. Weed-free control, and 5. Non-treated control. Weed control was good in all systems and cotton yields were not different among the treated plots. To increase the number of comparisons, 20 weed scenarios were posed to the Georgia weed extension specialists and HERB in each of the three cotton systems. The recommendation of HERB was the same as one of our two experts in 42% of the scenarios. Our experts advised a tank mix of MSMA with each of their recommendations. HERB included MSMA in only 40% of the scenarios. If the HERB database was altered to include MSMA applications, then the agreement rate between HERB and our experts approached 72%. Our goal is to continue to make subtle changes to the database that will allow for greater similarity between our experts and HERB, which will be the true validation of this system.