NETWORK EPIDEMIOLOGY AND PLANT TRADE
M.J. Jeger, G. Stancanelli, M. Pautasso
Effective control of a plant disease by means of prohibition, eradication, containment, or ongoing disease management, requires information on all pathways that lead to the introduction, dissemination and subsequent impact of the causal pathogen. Globally, regionally and within-country, trade networks are major pathways leading to serious disease outbreaks and epidemics. Recent examples include Phytophthora ramorum, where the pathogen was first recorded in Europe in the ornamental nursery trade and has subsequently jumped to a previously unsuspected host, Japanese larch (Larix kaempferi), with devastating consequences in the United Kingdom. More recently, ash dieback has moved inexorably across Europe and threatens Fraxinus excelsior and its associated biodiversity, whereby the trade in ash saplings facilitated long-distance dispersal of Hymenoscyphus fraxineus. In many ways this aspect is common to both pathogens and insect pests such as Dryocosmus kuriphilus which has spread from NW Italy to widely-dispersed areas in Europe. Of particular concern in Italy and Europe is the current outbreak of a strain of Xylella fastidiosa, causing decline of olives in Puglia – the first occurrence of this quarantine organism in Europe. Epidemiology has long been the basis for disease management, but until recently techniques for analysis of disease spread on plants moving in trade networks (seeds, plants-for-planting and for direct retail/wholesale markets) had been poorly developed and little used. Methods for network epidemiology are now available and increasingly used in different contexts and spatio-temporal scales. These methods are based on graph-theoretic con- cepts applicable to (trade) networks, which essentially consist of a set of nodes with linkages between them representing an overall network structure. The nodes might be, for example, nursery sites with some imposed hierarchy, i.e. producers, wholesalers, and retailers; the linkages would represent the directed flow of plants from site to site. From mathematical analysis of the network structure it is possible to identify some key attributes determining the overall likelihood of disease spread on the network: these are the degree of connectedness across the network; the node/site at which the pathogen is first introduced; the correlation between in-out linkages; and the role played by “hubs”, or highly connected sites, within the network. From these attributes it is possible to determine a threshold for pathogen establishment and dissemination, whether in the long-term the pathogen is likely to persist, and as a consequence the disease control options that could be used in prevention or mitigation. Despite these recent developments in methods for epidemiological analysis of networks, there are some major challenges still to be resolved. The horticultural industry, particularly the nursery trade is fragmented in terms of size and scale of operations and connectedness, with an enormous range of traded commodities. Additionally, plant trade is dynamic which leads to network structures that are often changing because of commercial interests, but also sometimes as a consequence of a regulated disease affecting the trade.