Opportunities exist to improve decision support systems through the use of dispersal information gained from epidemiological research. However, dispersal and demographic information is often fragmentary in plant pathology, and this uncertainty creates a risk of inappropriate action whenever such information is used as a basis for decision making. In this article, a scenario-based simulation approach is used to evaluate crop and economic risks and benefits in the use of dispersal information for decision making using the potato late blight pathosystem (Phytophthora infestans–Solanum tuberosum) as a case study. A recently validated spatiotemporal potato late blight model was coupled to submodels for crop growth, tuber dry matter production, and fungicide efficacy. The yield response of a range of management scenarios to a single influx of primary inoculum (the initial spore load) was calculated. Damage curves (relative yield loss versus initial spore load) from a range of combinations of varietal susceptibility and fungicide treatments were used to classify the various management scenarios as either sensitive to initial spore load or tolerant to initial spore load, thus identifying where a high degree of accuracy would be required in dispersal information for appropriate decision making, and where a greater degree of uncertainty could be tolerated. General epidemics, resulting from spatially homogeneous initial spore loads, responded more strongly to the size of the initial spore load than focal epidemics, resulting from an initial spot infection. Susceptible cultivars responded with sizeable yield losses even at low levels of initial spore load, regardless of the fungicide management regime used. These results indicated that, for susceptible cultivars (late cultivars in particular), the degree of accuracy that would be required in dispersal information for appropriate decision making is unlikely to be practically attainable. The results also indicated that, contrary to “folk wisdom,” spore loads of a few hundred spores per square meter do not lead to appreciable crop loss in resistant cultivars and are therefore acceptable. We conclude that scope exists for including dispersal information in decision making for potato late blight with resistant potato cultivars but not for susceptible cultivars. The modeling framework used in this study can be extended to investigate the scope for inclusion of dispersal information in decision support for other aerially transmitted pathogens.
Analysing social-ecological interactions in disease control: An agent-based model on farmers’ decision making and potato late blight dynamics
