The cost of counting and identifying weed seeds and seedlings

Bioeconomic weed management models help growers achieve appropriate weed management with less herbicide by matching management to the weed population in a field. Growers, however, will not use bioeconomic models unless cost-effective methods to sample their weed populations are identified. Counting and identifying seeds and seedlings is the most time-consuming and costly part of sampling weed populations. The time required for this process was investigated and modeled as a first step in developing sampling plans for growers using WEEDCAM, a bioeconomic model for weed management in Zea mays L. in Colorado. The time required to count and identify seeds or seedlings was recorded for 9,405 soil cores (5 cm in diameter and 10 cm deep) and 9,726 quadrats (18-cm band over 1.52 m of crop row) collected or examined in eight corn fields in eastern Colorado. The time required to count and identify seeds was best described using a log-linear regression with time increasing with the number of seeds and species and the amount of sand in the core. The average cost of determining there are no seeds in a core is $1.07 for a core from a field with 37% sand and $4.32 if the field has 88% sand. The average cost of counting and identifying 36 seeds of four species is $2.70 and $10.88 for cores with 37 and 88% sand, respectively. The time required to count and identify seedlings was best described using a log-linear regression with time increasing with the number of seedlings and species. Classifying seedlings as grass or broadleaf did not improve the model. The average cost of determining that a quadrat is weed-free is $0.02. The average cost of counting and identifying 37 seedlings is $0.05, $0.07, $0.19, and $0.26 per quadrat for 1, 2, 4, and 6 species, respectively. The cost of identifying seeds and seedlings in eastern Colorado Z. mays fields to use WEEDCAM is estimated as $2.71 per core for the seed bank and $0.08 per quadrat for seedlings.