Abstract Implementing an extensive pest survey is expensive and time-consuming, so optimizing site selection is critical to success. Efficiently allocating resources and maximizing survey coverage in areas where crops are at risk from pests could alleviate economic losses and threats to food security. Therefore, we developed a “site score metric” to quantitatively assess the value of each existing site to the survey. The site score metric is based on (i) historic corn production intensity, (ii) pest count per site, and (iii) distance from the closest site. By evaluating each site based on these factors, we were able to identify which sampling locations are most valuable to the survey. A k-nearest neighbor model was developed to interpolate pest populations from individual trap counts corn fields across the primary corn producing region in the state of Wisconsin. By evaluating the interpolated map of Wisconsin with the site scoring metric, the value of hypothetical additional survey locations can be assessed. An additional k-nearest neighbor model was used to classify similarly scoring regions to identify clusters where additional surveillance would be most valuable. This scoring metric can be used to guide new survey efforts, thereby improving the quality of data collected. Additionally, this could save costs by eliminating low scoring locations.
Robbins et al. (Sat,) studied this question.