Nitrogen limits plant growth in almost all terrestrial ecosystems. However, nitrogen fertilization experiments in water-limited ecosystems showed contradictory results. Our objective was to quantify plant growth response to N addition and to assess changes in this response along a precipitation gradient. The specific questions that guided our study were: 1. Does nitrogen limit plant growth in water-limited ecosystems and how large is the effect overall? 2. Is plant response to N addition similar in arid, semiarid and subhumid ecosystems? 3. How does plant response to N addition change along temperature gradient? 4. Are vegetation responses of grass-dominated sites similar as those of shrub-dominated sites?, and 5. How do fertilizer type (i.e. urea vs. ammonium nitrate) and doses affect plant responses? To answer these questions, we conducted a meta-analysis which offers formal statistical methods for comparing and integrating results of previously published N-fertilization studies. Mean effects were estimated as the response ratio, the ratio of some biomass quantity measured in experimental and control groups (relative yield). We assessed changes in mean effects along mean annual precipitation (MAP), mean annual temperature (MAT), latitude, primary production, and the rate of  N addition.

Results/Conclusions

Nitrogen addition increased plant growth across all studies by 60%. Nitrogen effect on plant growth increased significantly along the 50-550 mm precipitation gradient (Qw=8.94; P=0.002). Precipitation seasonality did not show a significant response when evaluated overall (P=0.18), but when subhumid ecosystems (sites with MAP higher than 400) were excluded, ecosystems with summer precipitation showed significantly higher (P=0.016) responses to N addition than those with winter precipitation. Across all studies, correlation with latitude and MAT was not significant (P=0.75 and P=0.70 respectively). However, when studies of subhumid ecosystems were excluded, the response ratio decreased significantly with mean annual temperature (P=0.04). Vegetation physiognomy had a significant effect (P=0.036) on the response ratio as ecosystems dominated by grasses showed significantly higher responses than those co-dominate or dominated by shrubs. Ecosystems sowed with grasses (cultivated) showed significantly higher (P=0.018) responses than natural ecosystems. The fertilizer type did not show a significant effect (P=0.71) on the biomass response to N addition. We showed that N limits plant growth in arid, semiarid and subhumid ecosystems and the limitation increases with mean annual precipitation. Increasing N availability is a major component of global change and the impacts on plant growth may be important, even in arid ecosystems.