Abstract

Citrus greasy spot (CGS), caused by Zasmidium citri, induces premature defoliation and yield loss in Citrus spp. CGS epidemiology is well understood in areas of high humidity such as Florida (USA), but remains unaddressed in Brazil, despite differing climatic conditions and disease management practices. We characterize the spatiotemporal dynamics of CGS in the Reconcavo of Bahia, Brazil, focusing on four hierarchical levels (quadrant, plant, grove and region). A survey conducted in 19 municipalities showed that disease is found throughout the entire region with a prevalence (i.e. proportion of affected sampling units) of 100% in groves and plants, and never lower than 70% on leaves. Index of dispersion (D) values suggest the spatial pattern of symptomatic units lies somewhere between random and regular. This was confirmed by the parameters of the binary power law for plants and their quadrants (log(A)<0 and b<1). Variability in disease severity at different plant heights (0.7 m, 1.3 m and 2.0 m) was tested, but no consistent differences were observed. We introduce a simple compartmental model synthesising the epidemiology of the disease, in order to motivate and guide further research. The data we have collected allow such a model to be parameterised, albeit with some ambiguity over the proportion of new infections that result from inoculum produced within the grove vs. external sources of infection. By extending our model to include two populations of growers - those who control and those who do not - coupled by the spread of airborne inoculum, we investigate likely performance of the type of cultural controls that would be accessible to citrus growers in Northeastern Brazil. Our model shows that control via removal of the key source of inoculum - i.e. fallen leaves - can be very effective. However, successful control is likely to require area-wide strategies, in which a large proportion of growers actively manage disease.