Development of a Research-Based, User-Friendly, Rapid Scouting Procedure for the Invasive Sugarcane Aphid (Melanaphis sacchari), in Grain Sorghum

Abstract

Sugarcane aphid (SCA), Melanaphis sacchari Zehntner, is a significant economic pest in grain sorghum in the Southeastern US and Southern Great Plains. A collaborative project led by Oklahoma State University was tasked with developing an effective scouting plan for SCA. A stratified sampling protocol was used to collect SCA data from 299 fields from six states (OK, KS, TX, AR, LA, MS), over two years (2016-2017). Using a nested analysis of variance (NANOVA) and Taylor's (1961) power law the within field sampling variance and dispersion pattern was defined. Results from these analyses revealed two significantly different geographical regions: a southern and a northern. Results show that in either sampling region three consecutive plant samples should be taken per stop within 30m of one another. Additionally, whole plant enumerative sampling was used to define where within the plant canopy sugarcane aphids were distributed. Results from that study showed the middle of the canopy may be the best area to extract the two-leaf sample unit. Three enumerative sampling plans for estimating population density and classification of a threshold were developed. Due to large sample sizes, these sampling protocols gave evidence that a binomial sequential sampling plan would be the best option for a rapid scouting tool. To develop the binomial sequential sampling tool tally threshold regressions were analyzed to define the relationship between the mean SCA per leaf and proportion of plants infested. After the fitness and practicality of the model was considered, tally thresholds of 50 and 100 aphids per plant were selected. Wald's sequential probability ratio test (SPRT) was used to determine stop lines for both sampling plans, which ranged from 10-24 plant samples per sampling event, with an average of 11 plant samples per sampling event, depending on state, action threshold, and error level. The binomial sampling plans were validated using 48 externally sampled fields analyzed with resampling for validation of sampling plans (RVSP) software. An in-field sampling tool was developed using the tally threshold of 50 SCA. This sequential binomial sampling plan for SCA will enable time-efficient scouting, expeditiously determine if, and when an action threshold is reached, and prevent unnecessary insecticide applications.