Submitted to: Weed Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/7/1998
Publication Date: N/A
Interpretive Summary: Sometimes, herbicide drift or herbicide residue carryover from one growing season to the next damages crops and reduces yield. Although economic benefits to farmers from using herbicides to control weeds can be estimated for most major field crops, the extent and impact of negative effects, such as crop damage from herbicides, are harder to estimate from crop injury symptoms. One goal of this research was to determine whether crop damage measured early in the growing season soon after treatment could be used to estimate crop yield loss at harvest. Soybean projected ground cover from video photographs was measured using computerized image analysis software. These percent cover measurements were compared with relative soybean stunting which was estimated by eye compared with untreated plants. Both types of measurements were made early in the growing season, soon after treatment with a damaging herbicide mixture at various rates. The research hshows that percent ground cover measured with video camera and computer technology has potential for estimating later yield loss compared with rated stunting. If later research shows that this approach can be automated, this information might help farmers improve crop management decisions when herbicide damage occurs.
Technical Abstract: The goal of this research was to determine whether crop damage from herbicides measured early in the growing season soon after treatment could help estimate relative crop yield loss at harvest. Two to four weeks after spraying soybeans with an unregistered, phytotoxic mixture of thifensulfuron plus sethoxydim at various rates plus crop oil concentrate, visually evaluated (rated) stunting and percent soybean ground cover from video photographs were measured. In each of three years, relative percent soybean yield was a negative linear function of relative herbicide rate from 0.25x to 2x the registered rates. The 1x rates of thifensulfuron and sethoxydim were 17.5 and 420 g ai/ha, respectively. Relative soybean yield was also positively linearly related to soybean ground cover and negatively linearly related to rated stunting over three years. Linear regression models of relative soybean yield versus percent soybean ground cover explained more data variability (r**2 = 0.45 to 0.62) than did either relative herbicide rate or rated stunting over three years. Year-to-year variation in relative soybean yield and soybean damage caused by thifensulfuron plus sethoxydim mixtures was more closely related to total monthly growing season precipitation and the distribution of monthly precipitation than to either average monthly high or low temperatures.