Structure and dynamics of grazed vegetation

The relationship between vegetation and grazing herbivores is dynamic; the structure and quality of vegetation affect the diet of grazing herbivores and, in turn, the components of grazing (defo- liation, excretal return and treading) affect the structure and species composition of the vegetation. Both the vertical and horizontal distribution of vegetation are important in understanding plant-animal interactions in grazed plant communities. Mechanisms which confer grazing resistance, through avoidance and tolerance strategies, are discussed. The difficulties in addressing scaling issues at the plant-animal interface are addressed; measurements or events at a small scale cannot be directly extrapolated to a larger scale. Moreover, integration of small-scale functions could result in counter intuitive results at the larger scale. The role of modelling in exploring spatio-temporal heterogeneity in plant-animal interactions is developed. Such models allow the exploration of a large number of hypotheses which cannot be tested under experimental conditions because of logistical, time and money constraints, and provide a framework in which to better understand grazing systems. As a man- agement tool, models have value in their predictive capacity and ability to simulate the complex sit- uations in grazed plant communities. @ Elsevier / Inra grazing resistance / heterogeneity / plant competition / plant-animal interactions / scaling Rsum — Structure et dynamique de la vgtation pture. En conditions ptures, la relation entre la vgtation et l'animal est dynamique. La structure et la qualit de la vgtation affecte l'ingestion par l'animal et, en retour, les composantes du pturage (dfoliation, restitutions et pitinement) modifient la structure et la composition floristique du couvert vgtal. En consquence, la prise en compte des distributions verticale et horizontale de la vgtation sont dterminantes pour une meilleure comprhension des interactions entre la plante et l'animal. Les mcanismes qui confrent une rsistance au pturage, travers des stratgies d'vitement et de tolrence sont discuts, de mme que la difficult d'identifier des chelles d'tude pertinentes pour chaque fonction de l'inter- face plante-animal. En effet, les mesures ralises ou les vnements ayant lieu une chelle fine (< in 2) ne sont pas directement extrapolables aux chelles plus larges. De plus, l'intgration des fonctions identifies une chelle fine peuvent conduire des rsultats contre-intuitifs des chelles plus larges. Le rle de la modlisation comme instrument d'exploration de l'htrognit spatio-