Résumé
Disease and pest attack intensities are mainly determined at the plot level through interactions between the host, noxious organism, environment and agricultural management (Zadoks and Schein, 1979). However, the immigration of noxious and beneficial organisms from outside may also affect pest and disease incidences at the plot scale. Successful immigration is facilitated in landscapes with greater connectivity between resources patches (Zadoks, 1999). Functional connectivity of landscapes depends on the distribution and density of specific land uses, on how these are perceived (hostile or not) by specific organisms, and on organisms' dispersal ability to move across non-habitat areas. In a given landscape context, higher connectivity is therefore expected for generalist noxious organisms with high dispersal abilities. Here, we study the relationship between coffee pest and disease incidence in coffee farms and landscape context. We hypothesize that greater coffee cover within the local context (<1500 radius) will increase pest and disease incidence whereas greater forest cover will decrease it. We use three focal organisms to test these hypotheses: (1) coffee rust (#Hemileia vastatrix#), (2) coffee berry borer (#Hypothenemus hampei#) and (3) the root-knot nematodes (#Meloidogyne# spp.). These noxious organisms differ by their host specificity and dispersal ability. Coffee rust is coffee specific. Its uredospores are spread by wind over great distances and can even cross oceans (Bowden, 1971). The coffee berry borer is very specific to coffee, however it has been found to reproduce in several plant species (Damon, 2000). The females are able to fly, and probably can be transported by convection winds, over a few hundred meters (Baker, 1984). Root-knot nematodes are able to infect different plant species and, when not dispersed through human activities, can be considered nearly immobile. (Résumé d'auteur)