Finding an optimal balance between livestock production and grazing impact on plant biodiversity is an important issue in the design of grazing systems. This paper describes a study of the effect of two levels of grazing intensity combined with different animal breeds on plant biodiversity and sward structural diversity of semi-natural grasslands conducted over 3 years in the United Kingdom (UK), France (F), Germany (D) and Italy (I). There were three treatments at each site: moderate grazing intensity with a commercial breed (MC), lenient grazing intensity with a commercial breed (LC) and lenient grazing intensity with a traditional breed (LT). Grazing livestock were cattle in the UK, F and D, and sheep in I. Only grazing intensity had clear effects on vegetation diversity, the effects depending on site-specific vegetation characteristics. In more diverse pastures with finer scale heterogeneity in F and I, there was little effect of the treatments. In less diverse swards with higher grass dominance and coarse-scale heterogeneity, the number of species decreased from initial levels on the lenient grazing intensity treatments (treatments LC and LT), because of increased dominance of grasses reducing the cover of other species, particularly legumes. The structural diversity of the pastures was more affected by level of grazing intensity than botanical diversity. In more diverse, semi-natural grasslands, structural heterogeneity was not affected by the treatments. In less diverse, more productive swards, structural heterogeneity decreased where large patches were initially found on the UK site, and increased where smaller patches were initially found on the D site. It is concluded that management systems to conserve diversity need to consider carefully the background environment, and that more knowledge of the mechanisms involved is needed at farm and landscape scales.

The leaf miner Cameraria ohridella causes premature defoliation of Aesculus hippocastanum trees. In order to assess the whole-plant loss of productivity caused by the parasite, we monitored seasonal changes of leaf gas exchange and leaf area losses in horse chestnut trees freely infested or chemically treated to prevent moth infestation (controls). Data were integrated in a model and the annual loss of net primary productivity (NPP) was calculated for infested trees with respect to controls. Measurements showed marked vertical stratification of C. ohridella attacks, with lower crown strata being more infested than higher ones. Leaf gas exchange was maximum between May and early June, but it strongly decreased starting from mid-June even in controls. Model calculations showed that NPP loss of infested trees was about 30% on an annual basis (when the first moth attack is recorded at the end of April). Model simulations showed that postponing the start day of attack would have important positive effects on plant's NPP. For example, if the start day of attack was postponed to 20 May, the annual loss of NPP would be about 15%. Our study suggests that A. hippocastanum trees attacked by C. ohridella are not facing serious risks of decline, especially if methods are adopted to postpone the start day of attack (e.g. removal of fallen leaves in autumn). Our data do not support the view that plants need to be totally protected from the parasite by application of insecticides.

In this paper a desertification risk index (DRI) based on the integration of climatic data and the normalised difference vegetation index (NDVI), obtained from National Oceanic Atmospheric Administration advance very high‐resolution radiometer (NOAA‐AVHRR) images, is discussed at the light of the aridity index and some eco‐physiological parameters. The good correlation between DRI, the aridity index and the eco‐physiological parameters suggests that DRI could be useful to measure the desertification risk. One advantage of DRI is that, with the help of a geographic information system (GIS), DRI maps can be easily obtained in short time and at relatively low costs.

No abstract