Submitted to: Plant Disease Epidemiology International Workshop
Publication Type: Abstract Only
Publication Acceptance Date: April 1, 2005
Publication Date: June 1, 2005
Citation: Gottwald, T.R., Taylor, E.L. 2005. Predicting the risk of infection in a Citrus tristeza virus epidemic and testing for the intensity of virus competition with survival analysis. Proceedings of the 9th International Workshop of Plant Disease Epidemiology. C6 p. 53. Technical Abstract: Background and objectives. Citrus Tristeza Virus (CTV) can cause several diseases of citrus and has been responsible for the death of over 50 million citrus trees in numerous citrus growing countries worldwide. The spatial and temporal changes in incidence of CTV during epidemics have been described for two virus/insect pathosystems: the CTV/Toxoptera citricida and CTV/Aphis gossypii pathosystems (2). The present study focuses on the CTV/T. citricida pathosystem. Aggregation of infected trees occurs due to local movement of viruliferous aphids. In addition, long distance spread of CTV has been document for both pathosystems. Two questions are addressed here, 1)“What threat do trees with prior CTV-infections present to healthy trees in the vicinity and how does this threat vary over distance?”, and 2)“If two isolates of CTV are present in a planting, do they interact in any way within the population or do they spread independently irrespective of prior infection by an opposing isolate?” Materials and Methods. Survival analysis has been used recently to investigate the occurrence and timing of plant disease events (1,3) and here was used to examine the probability of survival (remaining in a non-infected state) of CTV-free trees located at various distances from CTV-infected trees through time. A risk index was calculated via a modified Cox proportional hazards model to estimate the probability of survival through time of CTV-free trees when located at various distances to trees that became CTV-infected in prior years. The risk of becoming infected was related to ‘local areas of influence’ (i.e., proximity of previously infected trees), as previously reported. CTV Isolates (T30 and T36, mild and decline-inducing isolates, respectively) were differentiated serologically. Results and Discussion. The contribution of short distance transmissions and the influence they have on the overall spatial pattern of CTV that develops through time was examined using the Cox model. A substantial proportion of newly infected trees within a planting could be accounted for as trees determined to be infected 6 months previously within a ‘local area of influence’ of 8, 16, 24 or 32 m radii, and that ‘survival’ in a disease-free condition decreased significantly through time as the number of prior infected trees within these areas increased. Secondly, it was found that isolates of CTV that previously were thought to co-migrate independently, actually did have a subtle but measurable affect on subsequent infection by the opposing isolate as measured by the ‘hazard ratio’ estimated by the Cox model. Although this effect varied among the citrus plantings studied, the occurrence of prior infection by the opposing CTV isolate appeared to decrease the probability of infection by the other isolate by 10-20%, an affect that had previously gone unnoticed in field populations.