Influence of cropping systems on foot and root diseases of winter wheat: Fitting of a statistical model

In a series of trials carried out in France from 1986 to 1993, 503 winter wheat plots were assessed at heading for eyespot (Pseudocercosporella herpotrichoides), take-all (Gaeumannomyces graminis var. tritici) and sharp eyespot (Rhizoctonia cerealis). For each disease, a model was developed to estimate disease risk as a function of environment and cropping system. Eyespot depended on environment, crop succession, soil tillage (inversion vs. non-inversion), sowing date and tiller number per plant; take-all depended on environment, crop succession, sowing date and total nitrogen supply from soil plus fertiliser; and sharp eyespot depended on environment, crop succession, soil tillage and sowing date. Interactions between cultivation techniques were integrated by using a multiplicative model. The hierarchy of techniques is discussed. These models allow us to classify cropping systems according to their inherent disease risk and to foresee the impact of changes in system strategies. Comparison with previous studies also suggested new conclusions on the effect of ''risk amplifying non-host crops''. (C) 1997 Elsevier Science B.V.