Spatio-temporal modelling of soil-borne diseases

Soilborne diseases are difficult to control by fungicides or resistant cultivars; they are characterized by processes and spatio-temporal scales that can be different from those of aerial pathogens. In particular, the dispersal processes of soilborne pathogens result in the appearance and persistence of strong spatial patterns, namely disease foci. Aggregation influences both the temporal dynamics of diseases and their relation to crop losses, thus making spatial pattern an important factor when assessing control methods. The objectives of this study were to understand and model the spatio-temporal development of soilborne epidemics in order to determine control strategies that may reduce disease risk through an effect on spatial pattern. To that end, two models were developed. The first one is spatially explicit and biologically realistic; its parameters were measured for take-all disease of wheat. The model was tested with field data; this showed that the model is precise but overestimates disease development. This test also provided possible avenues for improving the predictive ability of the model. The second model is simpler and more generic; it uses hierarchy theory in order to simulate disease spread across multiple spatial scales. Hypotheses concerning the behaviour of epidemics and the relationships between disease incidences at different spatial scales could be tested using this model, for example the effect of the host and inoculum spatial structures on disease dynamics and aggregation. In the case of take-all disease of wheat –for which we have demonstrated the importance of the spatial pattern of primary inoculum– the simulation results advocate for different sowing patterns according to the rank of consecutive wheat crops.