Document: GAR-3-55-11

Inference from the spatial pattern of disease: Evidence for spore showers of Phytophthora infestans in the highland tropics.

GARRETT, K.A.* ¹, G.A.FORBES ², A.M.SMITH ³, C.C.MUNDT ⁴, L.V.MADDEN ⁵ and R.J.NELSON ⁶

Kansas State University, Manhattan, KS 66506 USA ¹
International Potato Center, Quito, Ecuador ²
University of California, Berkeley CA 94720 USA ³
Oregon State University, Corvallis, OR 97331 USA ⁴
Ohio State University, Wooster, OH 44691 USA ⁵
International Potato Center, Lima, Peru ⁶

Abstract:
Suppose we observe a population of plants in which each plant has approximately the same low level of infection by an airborne pathogen. Suppose we also know that there were few pathogen propagules originating from 'within' the area of this particular host population. The combination of few local pathogen propagules and low infection levels evenly distributed may indicate an epidemic initiated by a large number of propagules from 'outside' the area of the host population. If that is a reasonable conclusion, then field surveys of the spatial pattern of a particular disease may be useful for gathering information about how epidemics of that disease begin. But how safe is it to assume that a uniform pattern of disease implies a large number of initial propagules? To consider this question, we designed a simulation model of the dispersal of pathogen propagules through a field of host individuals. The model includes as parameters the number of propagules that initiate an epidemic, the size of the population of host individuals, the shape of the propagule dispersal gradient around an infection, and the number of pathogen generations. Propagules are distributed randomly through the host population to initiate an epidemic and in the next generation new propagules are dispersed with the probability of infection of any given host individual determined by the parameters of the dispersal gradient. A range of parameter values was considered in a series of simulations, with particular emphasis on parameter values appropriate for the pathogen Phytophthora infestans. After the designated number of pathogen generations, the level of spatial aggregation of disease was measured for each simulation run. Though it was possible to generate a uniform pattern of low disease levels from a small number of initial propagules, this was an unlikely result for the parameters of P. infestans under the assumptions of our model. This is of interest because the spatial pattern of infection by P. infestans in Ecuador and Peru has been observed to be uniform at low levels, in contrast to the aggregated disease observed in temperate regions. Disease management practices that depend only on reducing P. infestans spore production within a field may be less useful in the highland tropics because of spore showers.

Keywords: spatial pattern, epidemiology, dispersal

This abstract is being presented at: 5:00 PM in session:
Oral Session #15: Parasitoids and Diseases.