Within-Field Spatial Distribution of Stink Bug (Hemiptera: Pentatomidae)-Induced Boll Injury in Commercial Cotton Fields of the Southeastern United States

Abstract

Spatial distribution of boll injury caused by stink bugs to developing cotton (Gossypium hirsutum L.) bolls was studied in five commercial fields (≍22 ha each) in 2011 and 2012 to understand variability in boll injury dynamics within fields. Cotton bolls and stink bugs were sampled weekly from a georeferenced grid of sampling points (one sample per 0.40 ha) in each field, but no samples were taken within 30 m of field edges. The inverse distance weighted interpolation, variogram analysis, and Moran's I were used to describe spatial variability of boll damage within the fields. Boll injury was found to be spatially associated at distances ranging from ≍75 to 275 m with an average distance ≍150 m. An exponential variogram model was selected as the best fitting model to describe the spatial association in four of the five fields. Moran's I indicated that spatial association was significant in three of the five fields. The spread of boll injury from stink bugs was gradual in most fields and always exceeded the treatment threshold during the fourth or fifth week of bloom. Capture of stink bugs using a sweep net was inefficient, strongly suggesting that quantifying boll injury is a better sampling method and predictor of stink bug activity when sampling all but the edges of the field. These data suggest that scouts need to sample boll injury from sample locations separated by at least 150 m to assure independence in the central part of large fields. Second, future researchers who plan to use parametric statistical methods could use a 150-m grid, as opposed to a denser grid that would require greater time and effort.

Sampling Optimization and Crop Interface Effects on Lygus lineolaris Populations in Southeastern USA Cotton

“…In this study, defined sampling quadrants explained the second-highest level of variability in L. lineolaris densities, suggesting a non-random, aggregated spatial distribution within fields. This result is not surprising, given previous work suggesting non-random, clustered distributions of closely related Lygus species in cotton [51,52] and other heteropteran species in field crop production systems [53][54][55][56]. In addition to field-specific scouting, intensive L. lineolaris sampling within fields is likely important to minimize the probability of false negative or positive assessments when making management decisions using economic thresholds.…”

Spatio‐temporal distribution of Anthonomus grandis grandis Boh. in tropical cotton fields

“…Previous research on adult boll weevil in‐field distributions used whole‐plant sampling 14 and pheromone‐based traps both possessing different dimensions for analyzing dispersion compared with the current approach 8,15,41 . Alternatively, application of kriging and other techniques to characterize in‐field cotton pest distributions were used to describe spatial distribution of pests such as the pink bollworm, 42 cotton bollworm, 43,44 pentatomid stink bugs 45–48 and thrips 49 . Our study is the first to describe spatial distribution of the boll weevil with ordinary kriging.…”

Evaluation of macadamia felted coccid (Hemiptera: Eriococcidae) damage and cultivar susceptibility using imagery from a small unmanned aerial vehicle (sUAV), combined with ground truthing

“…To visualize ground-based damage assessment scores, spatial data visualization tools such as a proportional bubble map and Inverse Distance Weighted interpolation map were used. 39,40 A proportional bubble map was used to visualize raw data, while we used Inverse Distance Weighted interpolation map after assigning a spatial weight based on the distances between trees. Moran's I was used for statistical hypothesis testing of the spatial distribution of infestation.…”

“…For the same dataset, we used the PROC GLIMMIX procedure 38 to compare the mean percent damage among rows. To visualize ground‐based damage assessment scores, spatial data visualization tools such as a proportional bubble map and Inverse Distance Weighted interpolation map were used 39,40 . A proportional bubble map was used to visualize raw data, while we used Inverse Distance Weighted interpolation map after assigning a spatial weight based on the distances between trees.…”