The gap between atmospheric nitrogen deposition experiments and reality

Abstract Anthropogenic activities have dramatically altered the global nitrogen (N) cycle. Atmospheric N deposition, primarily from combustion of biomass and fossil fuels, has caused acidification of precipitation and freshwater, and triggered intense research into ecosystem responses to this pollutant. Experimental simulations of N deposition have been the main scientific tool to understand ecosystem responses, revealing dramatic impacts on soil microbes, plants, and higher trophic levels. However, comparison of the experimental treatments applied in the vast majority of studies with observational and modelled N deposition reveals a wide gulf between research and reality. While the majority of experimental treatments exceed 100 kg N ha−1 y−1, global median land surface deposition rates are around 1 kg N ha−1 y−1 and only exceed 10 kg N ha−1 y−1 in certain regions, primarily in industrialized areas of Europe and Asia and particularly in forests. Experimental N deposition treatments are in fact similar to mineral fertilizer application rates in agriculture. Some ecological guilds, such as saprotrophic fungi, are highly sensitive to N and respond differently to low and high N availability. In addition, very high levels of N application cause changes in soil chemistry, such as acidification, meaning that unrealistic experimental treatments are unlikely to reveal true ecosystem responses to N. Hence, despite decades of research, past experiments can tell us little about how the biosphere has responded to anthropogenic N deposition. A new approach is required to improve our understanding of this important phenomenon. First, characterization of N response functions using observed N deposition gradients. Second, application of experimental N addition gradients at realistic levels over long periods to detect cumulative effects. Third, application of non-linear meta-regressions to detect non-linear responses in meta-analyses of experimental studies.