| dc.description.abstract | The size of the Norwegian red deer population is historically high and typical of the trend seen over much of Europe. Dense populations may cause damage to agricultural crops, and crop yield is drastically decreased by red deer grazing in certain areas. We know that red deer select actively managed meadows, i.e., frequently renewed by fertilisation and re-seeding, over other agricultural meadows. Despite its importance, information regarding spatial grazing patterns by red deer on agricultural meadows is limited. In this study, I aim to quantify how grazing on agricultural meadows by red deer varies across spatial scales in southwest Norway. I hypothesise that grazing on agricultural meadows is determined by three major effects: (H1) Factors affecting forage quality and availability in meadows relative to natural habitats, such as population density and seasonal change, (H2) meadow management, such as renewal of meadows, and (H3) perceived predation risk and human disturbance, such as distance to settlement and forest edge. Grazing levels were assessed across meadows in a hierarchical study design, and I analysed the data using a binary logistic regression that model absence of grazing and a beta regression that model the level of grazing given grazing occurred. This enabled me to quantify both variation among spatial hierarchical units and the mechanisms behind spatial grazing patterns. I found that the grazing variation was largest between meadows in the local area and smallest on broader scale. High red deer density areas received more grazing relative to low-density areas, more grazing occured when meadow grass was shorter, and early in summer relative to late, suggesting that red deer select meadows over natural habitat when the difference in quality and availability of forage are large enough. Newly refreshed meadows received more grazing than the older ones, implying that a large part of the local site effect was caused by meadow management. Evidence of trade-off effects also appeared important as spatial grazing patterns changed near roads, houses, and forests. Broad-scale variation in red deer density explained some of the variations in grazing. However, since the largest variation in grazing was found locally, population reduction at broad scales may not effectively lower damages. These results may affect the scale at which management should target mitigation efforts. | eng |