Sclerotinia crown rot, caused by Sclerotinia minor and S. sclerotiorum,
is a disease of pyrethrum in Australia that may cause substantial decline
in plant density. The spatiotemporal characteristics of the disease
were quantified in 14 fields during three growing seasons. Fitting the
binary power law to disease incidence provided slope (b = 1.063) and
intercept (ln(Ap) = 0.669) estimates significantly (P ≤ 0.0001) greater
than 1 and 0, respectively, indicating spatial aggregation at the sampling
unit scale that was dependent upon disease incidence. Covariate
analyses indicated that application of fungicides did not significantly
influence these estimates. Spatial autocorrelation and spatial analysis
by distance indices indicated that spatial aggregation above the sampling
unit scale was limited to 20 and 17% of transects analyzed, respectively.
The range of significant aggregation was limited primarily
to neighboring sampling units only. Simple temporal disease models
failed to adequately describe disease progress, due to a decline in
disease incidence in spring. The relationships between disease incidence
at the scales of individual plants within quadrats and quadrats
within a field was modeled using four predictors of sample size. The
choice of the specific incidence–incidence relationship influenced
the classification of disease incidence as greater than or less than 2%
of plants, a provisional commercial threshold for fungicide application.
Together, these studies indicated that epidemics of Sclerotinia
crown rot were dominated by small-scale aggregation of disease.
Larger scale patterns of diseased plants, when present, were associated
with severe disease outbreaks. The spatial and temporal analyses
were suggestive of disease epidemics being associated with localized
primary inoculum and other factors that favor disease development at
a small scale.