Oral Presentation Science Protecting Plant Health 2017

Optimal surveillance strategies for invasive plant pathogens using a spatial stochastic algorithmic mapping approach (4324)

Eric Grist 1 , Nik J Cunniffe 2 , Tim R Gottwald 3 , Stephen R Parnell 1
  1. School of Environment and Life Sciences, University of Salford, Salford, Manchester, UK
  2. Department of Plant Sciences, University of Cambridge, Cambridge, UK
  3. United States Department of Agriculture, Agricultural Research Service, Fort Pierce, Florida, United States

The global rise in emerging plant disease threats has led to an urgent need for improved surveillance strategies in plant health. Improved mapping and monitoring for targeted surveillance must address spatial heterogeneity whilst simultaneously incorporating relevant epidemiological processes. At more localised spatial scales, statistical inference of disease status through empirical estimation methods may surpass that of mechanistic approaches by taking the effects of local random variation better into account. However, identification of the most effective locations for disease intervention and monitoring still poses a challenge because of the dynamic complexities presented by pathogen dispersal, transmission and current spatial availability of plant hosts. These characteristics are typically the most crucial factors determining the outcome of plant disease invasions at the scale of a plantation and ultimately beyond.

Here, we apply a stochastic algorithmic mapping (SAM) approach first developed by Parnell et al (2011)* which includes key dispersal and transmission epidemiological parameters to describe the dynamic spatial distribution of an invasive plant pathogen. We consider typical scenarios where sampling and monitoring to determine disease status of plant hosts is subject to finite resources. The performance of the SAM approach is then determined in terms of standard receiver operating characteristic (ROC) curves and compared with a geostatistical interpolative mapping approach (such as kriging). Using data collected for Asiatic citrus canker in Florida, we illustrate how in practice this may better enable control and monitoring strategies to be determined under a variety of sampling design objectives. These may range from identifying the host locations of highest risk to those of greatest uncertainty to currently target.


*SR Parnell, TR Gottwald, MS Irey, W Luo and F Van den Bosch (2011). A stochastic Optimization method to estimate the spatial distribution of a pathogen from a Sample. Phytopathology 101(10), 1184-190.