LAUSR

Abstract Empirical critical loads are widely used to quantify and manage the ecological impacts of reactive nitrogen (N) deposition. Critical load values aim to identify a level of N deposition below which significant harmful effects do not occur according to present knowledge. Critical loads have been primarily based on experiments, but these are few in number and have well-known limitations, so there is a strong imperative to test and validate values with other forms of evidence. We assembled data on the spatial variability in vegetation communities in the United Kingdom and used Threshold Indicator Taxa Analyses (TITAN) to investigate linkages between species changes and modelled current and cumulative N deposition. Our analyses focused on five datasets: acid grasslands, alpine habitats, coastal fixed dunes, dune slacks and wet grasslands. In four of these habitats there was evidence for a significant decline in the cover of at least one species (a ‘species-loss change-point’) occurring below the critical load, and often at very low levels of N deposition. In all of the habitats there was evidence for clustering of many individual species-loss change-points, implying a community change-point analogous to an ecological threshold. Three of these community change-points occurred below the critical load and the remaining two overlapped with the critical load range. Studies using similar approaches are now increasingly common, with similar results. Across 19 similar analyses there has been evidence for plant species loss change-points below the critical load in 18 analyses, and community-level species loss change-points below the critical load in 13 analyses. None of these analyses has shown community change-points above the critical load. Field data increasingly suggest that many European critical loads are too high to confidently prevent loss of sensitive species.