Sample distance and observer characteristic effects on corn rootworm beetles counts in corn and resource allocation optimization in corn rootworm insecticide trials

摘要

Counting adult rootworms has not yet met with widespread adoption by Iowa growers. It is more economical to always treat with an insecticide than to scout and treat when the rootworm population reaches economic levels. However, it is estimated that the majority of insecticide applications are not warranted. This inconsistency results from the cost and variability of scouting resulting in unrealistic thresholds;Establishing the existence or non-existence of an edge effect can reduce the error and/or cost of scouting. If there are differences at the edges, then the establishment of the sample boundaries will reduce sampling error. Sampling as close to the edge as possible without biasing the samples will save time in scouting. I looked at data collected from 34 Iowa continuous cornfields over three years to determine if counts of adult rootworms were different on the outside edges of the fields as compared to counts taken farther into the interior. I found that the counts of northern corn rootworms were higher in row 2 than in rows 12, 27, 47, and 80. Counts of the western corn rootworm were not affected by the position of the row. When analyzing combined counts, there were no differences for any row position. The relative population level also has an effect on the counts, with an increased significance of the difference in the counts of northern rootworms between low and high population levels;Another source of error is inter-sampler variation. I tested the effects of two sampling characteristics, experience and height, in four fields and found that sampler experience had more impact on beetle counts than did height. Because experience levels differ among samplers, the implication for pest management scouting is to sample fields with multiple samplers to reduce this effect when making recommendations for multiple fields;Finally, root-rating scores from five years of insecticide efficacy trials were analyzed and variance components for plants, plots, and blocks were estimated. By using these estimates I was able to calculate the number of roots to sample in insecticide efficacy trails given a desired precision level and number of blocks in the test. For a precision of 20 percent, and in tests with 4 blocks, 3 roots from each treatment in the block are adequate.