Responses of terrestrial ecosystem phosphorus cycling to nitrogen addition: A meta-analysis

Aim: Anthropogenic additions of nitrogen (N) are expected to drive terrestrial ecosystems toward greater phosphorus (P) limitation. However, a comprehensive understanding of how an ecosystem's P cycle responds to external N inputs remains elusive, making model predictions of the anthropogenic P limitation and its impacts largely uncertain. Location: Global. Time period: 1986-2015. Major taxa studied: Terrestrial ecosystems. Methods: We conducted a meta-analysis including 288 independent study sites from 192 articles to evaluate global patterns and controls of 10 variables associated with ecosystem P cycling under N addition. Results: Overall, N addition increased biomass in plants (134%) and litter (115%) as well as plant P content (117%), while decreasing P concentrations in plants and litter (28% and 211%, respectively). N addition did not change soil labile P or microbial P, but enhanced phosphatase activity (124%). The effects of N addition on the litter P pool and soil total P remained unclear due to significant publication biases. The response of P cycling to N addition in tropical forests was different from that in other ecosystem types. N addition did not change plant biomass or phosphatase activity in tropical forests but significantly reduced plant P and soil labile P concentrations. The shift in plant P concentration under N addition was negatively correlated with the N application rate or total N load. N-induced change in soil labile P was strongly regulated by soil pH value at the control sites, with a significant decrease of 14% only in acidic soils (pH < 5.5). Main conclusions: Our results suggest that as anthropogenic N enhancement continues in the future it could induce P limitation in terrestrial ecosystems while accelerating P cycling, particularly in tropical forests. A quantitative framework generated on the basis of this meta-analysis is useful for our understanding of ecosystem P cycling with N addition, and for incorporating the anthropogenic P limitation into ecosystem models used to analyse effects of future climate change.