The uncertainties and sources of variation in projected impacts of climate change on agriculture and terrestrial ecosystems depend not only on the emission scenarios and climate models used for projecting future climates, but also on the impact models used, and the local soil and climatic conditions of the managed or unmanaged ecosystems under study. We addressed these uncertainties by applying different impact models at site, regional and continental scales, and by separating the variation in simulated relative changes in ecosystem performance into the different sources of uncertainty and variation using analyses of variance. The crop and ecosystem models used output from a range of global and regional climate models (GCMs and RCMs) projecting climate change over Europe between 1961–1990 and 2071–2100 under the IPCC SRES scenarios. The projected impacts on productivity of crops and ecosystems included the direct effects of increased CO2 concentration on photosynthesis. The variation in simulated results attributed to differences between the climate models were, in all cases, smaller than the variation attributed to either emission scenarios or local conditions. The methods used for applying the climate model outputs played a larger role than the choice of the GCM or RCM. The thermal suitability for grain maize cultivation in Europe was estimated to expand by 30–50% across all SRES emissions scenarios. Strong increases in net primary productivity (NPP) (35–54%) were projected in northern European ecosystems as a result of a longer growing season and higher CO2 concentrations. Changing water balance dominated the projected responses of southern European ecosystems, with NPP declining or increasing only slightly relative to present-day conditions. Both site and continental scale models showed large increases in yield of rain-fed winter wheat for northern Europe, with smaller increases or even decreases in southern Europe. Site-based, regional and continental scale models showed large spatial variations in the response of nitrate leaching from winter wheat cultivation to projected climate change due to strong interactions with soils and climate. The variation in simulated impacts was smaller between scenarios based on RCMs nested within the same GCM than between scenarios based on different GCMs or between emission scenarios

No abstract

The inefficient use of phosphorus (P) in the food chain is a threat to the global aquatic environment and the health and well-being of citizens, and it is depleting an essential finite natural resource critical for future food security and ecosystem function. We outline a strategic framework of 5R stewardship (Re-align P inputs, Reduce P losses, Recycle P in bioresources, Recover P in wastes, and Redefine P in food systems) to help identify and deliver a range of integrated, cost-effective, and feasible technological innovations to improve P use efficiency in society and reduce Europe’s dependence on P imports. Their combined adoption facilitated by interactive policies, co-operation between upstream and downstream stakeholders (researchers, investors, producers, distributors, and consumers), and more harmonized approaches to P accounting would maximize the resource and environmental benefits and help deliver a more competitive, circular, and sustainable European economy. The case of Europe provides a blueprint for global P stewardship.

The procedure for applying phosphorus (P) fertilizer to soil can be divided into three consecutive steps: (i) Measurement of soil‐P availability, (ii) calibration of the soil‐P fertility level and (iii) estimation of the recommended P dose. Information on each of these steps was obtained for 18 European countries and regions with the aim of comparing P fertilizer recommendation systems at the European scale. We collected information on P fertilizer recommendations through conventional or grey literature, and personal contacts with researchers, laboratories and advisory services. We found much variation between countries for each of the three steps: There are more than 10 soil‐P tests currently in use, apparent contradictions in the interpretation of soil‐P test values and more than 3‐fold differences in the P fertilizer recommendations for similar soil‐crop situations. This last result was confirmed by conducting a simple experimental inter‐laboratory comparison. Moreover, soil properties (pH, clay content) and crop species characteristics (P responsiveness) are used in some countries in the calibration and recommendation steps, but in different ways. However, there are also common characteristics: soil‐P availability is determined in all countries by extraction with chemical reagents and the calibration of the soil‐P test values, and the fertilizer recommendations are based on the results from empirical field trials. Moreover, the fertilizer recommendations are nearly all based on the amount of P exported in the crops. As long as rational scientific and theoretical backgrounds are lacking, there is no point in trying to synchronize the different chemical methods used. We therefore call for a mechanistic approach in which the processes involved in plant P nutrition are truly reproduced by a single standard method or simulated by sorption‐desorption models.

than 50% of the yearly contribution of P to Danish freshwaters (Kronvang et al., 1995). In spite of these Strongly sorbing compounds such as P, pesticides, and heavy metals significant reductions in discharge, P is still problematic can be transported through soils while being adsorbed to mobile colloidal particles. While the rapid leaching of nonadsorbing chemicals