Search

Abstract Spatially detailed information on agricultural nitrogen (N) budgets is relevant to identify regions where there is a need for a reduction in inputs in view of various forms of N pollution. However, at the scale of the European Union, there is a lack of consistent, reliable, high spatial resolution data necessary for the calculation of regional N losses. To gain insight in the reduction in uncertainty achieved by using higher spatial resolution input data. This was done by comparing spatially disaggregated agricultural N budgets for Denmark for the period 2000–2010, generated by two versions of the European scale model Integrator, a version using high spatial resolution national data for Denmark (Integrator-DK) and a version using available data at the EU scale (Integrator-EU). Results showed that the national N fluxes in the N budgets calculated by the two versions of the model were within 1–5% for N inputs by fertilizer and manure excretion, but inputs by N fixation and N mineralisation differed by 50–100% and N uptake also differed by ca 25%, causing a difference in N leaching and runoff of nearly 50%. Comparison with an independently derived Danish national budget appeared generally to be better with Integrator-EU results in 2000 but with Integrator-DK results in 2010. However, the spatial distribution of manure distribution and N losses from Integrator-DK were closer to observed distributions than those from Integrator-EU. We conclude that close attention to local agronomic practices is needed when using a leaching fraction approach and that for effective support of environmental policymaking, Member States need to collect or submit high spatial resolution agricultural data to Eurostat. Graphical abstract Unlabelled Image Highlights • Spatially disaggregated agronomic nitrogen (N) budgets for Denmark are compared. • Effects of using national rather than EU input data are analysed. • Detail of data causes large differences in N excretion and N losses to air and water. • Results obtained with detailed Danish input data were closer to observed distributions. • Good policy support requires high spatial resolution input data. [ABSTRACT FROM AUTHOR]