Skip to main content
ARS Home » Southeast Area » Fort Pierce, Florida » U.S. Horticultural Research Laboratory » Subtropical Plant Pathology Research » Research » Publications at this Location » Publication #338214
Title: Risk based management of invading plant disease

Author
item HYATT-TWYNAM, S - University Of Cambridge
item PARNELL, S - University Of Salford
item STUTT, R O J - University Of Cambridge
item Gottwald, Timothy
item GILLIGAN, C - University Of Cambridge
item CUNNIFFE, N - University Of Cambridge

Submitted to: New Phytologist
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/19/2017
Publication Date: 3/28/2017
Citation: Hyatt-Twynam, S.R., Parnell, S., Stutt, R.H., Gottwald, T.R., Gilligan, C.A., Cunniffe, N.J. 2017. Risk based management of invading plant disease. New Phytologist. 214:1317-1329.

Interpretive Summary: Effective control of newly introduced plant diseases is challenging. Attempts to eradicate diseases often involves removal of all plants within a certain distance from any infected plant. This results in not only removing infected plants but also numerous other plants that may or may not be infected and have no symptoms of infection. This can be objectionable to commercial and residential growers alike that do ot want to remove seemingly healthy plants.. We have developed a suite of mathematical/statistical models for disease spread and that measures the effects of various control strategies based on data from disease outbreaks. For this study we used data from the epidemic of citrus canker disease in Florida in the 1990s-2000s. Rather than just considering any plant near and infected plant could also be infected, we use a risk ranking of the probability for each host tree in the area being infected. This new model significantly outperforms our prior model that simply uses a constant distance to prescribe what trees to remove. We use a simple and eloquent equation call the “rule of thumb” within the model and to demonstrate to stakeholders how the new model requires less trees to be removed during an epidemic. Using the new model results in improved control at a lower impact, i.e. less trees that needs to be removed. Thus, this model can be used to aid in management decisions by regulatory agencies, growers and residential citrus growers.

Technical Abstract: Effective control of new and emerging plant disease remains a key challenge. Attempts to eradicate pathogens often involve removal of all plants within a fixed distance of detected infected hosts, targeting asymptomatic infection. Here we develop and test potentially more efficient, epidemiologically-motivated management strategies, using a spatially-explicit, stochastic, epidemic model, previously fitted to data on the spread of citrus canker in Florida. A risk based strategy, ranking hosts for potential removal based on the number of secondary infections they are expected to cause, significantly outperforms constant radius removal. However, the risk based strategy is not transparent to stakeholders, since removals depend in a complex fashion on past patterns of pathogen spread. This motivates a variable radius strategy, in which removal radii are fixed at the start of the epidemic, but vary for each host according to a “rule of thumb”. This strategy has intermediate performance. Both strategies are robust to changes to disease spread parameters and patterns of susceptible hosts. However, improved control has a cost, with efficiency degrading if epidemiological parameters are incorrectly characterised. This focuses attention on gaining maximal information from past epidemics, understanding model transferability between locations and adaptive management strategies that can change over time.