Temporal and spatial dynamics of watermelon gummy stem blight epidemics

Abstract

The effect of the distance of initial inoculum on the intensity of watermelon gummy stem blight, caused by Didymella bryoniae, was studied in a naturally-infected rainfed commercial field. The shorter the distance from the focus, the sooner was disease onset and the earlier maximum disease levels were achieved. Maximum disease incidences were reached earlier than maximum severities, but eventually destructive levels were observed for both disease incidence and severity. Disease progressed at similar rates, irrespective of the radial distance from the focus. A detailed study of the disease temporal progress was conducted in inoculated rainfed experimental fields with commercial genotypes Crimson Sweet (susceptible, S) and Riviera (moderately resistant, R). The Gompertz model best described the disease progress curves, and estimated apparent infection rates were 0.049 and 0.020 respectively for S and R genotypes. In addition, spatial pattern studies were conducted during the dry season in overhead irrigated experimental plots, inoculated with point-source foci. Disease intensity gradients were better explained by the Exponential model than by the Power Law model. Gummy stem blight distribution was classified as aggregated by the Ordinary Runs procedure. Two different spatial autocorrelation methods (2DCorr and LCOR) revealed strong short distance spatial dependencies. Long distance positive correlations between quadrats were observed along with periods of higher progress rates. The dynamic patterns of the epidemics of gummy stem blight in watermelon described here are consistent with epidemics of polycyclic diseases with splash-dispersed spores.