Nik Cunniffe, Richard Stutt and Chris Gilligan have published work in PLoS Computational Biology showing how control of plant disease via removal of host plants can be optimised. The focal example in their paper is citrus canker, a bacterial disease of citrus trees. The US government spent an estimated $1 billion on survey, control and compensation costs during a highly controversial campaign that attempted to eradicate citrus canker from Florida after its initial introduction near Miami in around 1995. However, after removal of at least ten million trees, the campaign was deemed to have failed in 2006.
The work shows how this type of control can only be optimised by matching the spatial scale of control with the spatial scale of epidemic spread. It also shows how effective control should depend on the level of risk that is tolerable to the decision maker, and how it can account for the potential larger-scale impacts of a small-scale outbreak.
The underlying model is available as an interactive, user-friendly interface at http://www.webidemics.com/, intended to illustrate the sometimes counter-intuitive epidemiological principles that underlie successful disease control.