Meso- and Microscale Patterns of Fusarium Head Blight in Spring Wheat Fields in Minnesota

Several techniques were used to study the spatial distribution of Fusarium head blight (FHB) in spring wheat fields in the Red River Valley of Minnesota in 2001 and 2002. Four scales of study were analyzed, ranging from countywide comparisons down to adjacent heads. Agreement to frequency distributions and the departure from randomness, along with the autocorrelation of FHB clusters, were calculated at each scale. Regression analysis also was used to describe any edge affects that might exist. Finally, the optimal number of samples to adequately assess a 32.4-ha (80-acre) field was determined to a precision of ±5% disease incidence (DI). The analysis showed that DI had a random pattern at scales smaller than 32.4-ha fields, including quadrats of 6,561, 729, and 9 m², as well 0.8-m transects involving consecutive heads within drill rows. There was no difference in DI associated with edges of fields compared with the incidence of FHB in transects within fields. Analysis of the spatial distribution at the mesoscale (county and township) showed no association in FHB DI between fields based on proximity. Differences were attributed to previous crop and in-field residue. Disease incidence was most accurately assessed on a field-by-field basis. The optimal sampling size (with <5% error) for surveying for the incidence of FHB in a standard commercial field was seven 20-head samples. Regional disease estimates would improve from including more fields at the expense of fewer samples per field.