Alternative methods for herbicide spray drift detection in corn and cotton

The acceptance of herbicide-resistant crops and the subsequent increase in glyphosate use throughout the growing season has led to increased problems with herbicide drift in corn and cotton. Potential yield losses due to herbicide spray drift cannot always be accurately assessed using visual ratings or by measuring reductions in plant height. This research was conducted to determine if other methods, such as biochemical assays or remote sensing, could be used to assess potential yield reductions and visual injury associated with herbicide drift.; Other objectives included using hyperspectral and multispectral remote sensing for herbicide drift detection in corn and cotton. Classification accuracies were highest 14 DAA when distinguishing untreated from treated corn. When the data were classified by herbicides, overall accuracies of ≥74% resulted 14 and 28 DAA. Classifications based on yield reduction also resulted in highest overall classification accuracies 14 and 28 DAA, ranging from 69 to 70%. Within the glufosinate-treated corn, overall accuracies for classifying percent yield reductions was 85% at 3 DAA and decreased to 68% at 28 DAA. Overall classification accuracies for determining percent yield reductions with glyphosate were higher 14 and 28 DAA. With pyrithiobac, overall accuracies for determining percent yield reductions ranged from 73 to 82% across all evaluations. Of the nine vegetative indices tested to classify the data, Anthocyanin Reflective Index (ARI) and Chlorophyll Fluorescence Ratio of reflectance at 735 nm/700 nm (CF 735) were the most important for classifying, assessing, and detecting herbicide spray drift in corn.; Multispectral aerial imagery was also used to identify cotton affected by a simulated bromoxynil drift event. Multispectral aerial imagery collected 9 DAA in 2002 resulted in better classification accuracies for identifying bromoxynil rate, percent visual injury, and percent yield reduction than imagery collected 21 DAA in 2001. The results from this research indicate that bromoxynil spray drift could be detected with multispectral aerial imagery. However, the images would need to be collected soon (∼9 DAA) after the drift event occurred. (Abstract shortened by UMI.)...