Nitrogen (N) and phosphorus (P), either individually or in combination, have been demonstrated to limit biomass production in terrestrial ecosystems. Field studies have been extensively synthesized to assess global patterns of N impacts on terrestrial ecosystem processes. However, to our knowledge, no synthesis has been done so far to reveal global patterns of P impacts on terrestrial ecosystems, especially under different nitrogen (N) levels. Here, we conducted a meta‐analysis of impacts of P addition, either alone or with N addition, on aboveground (AGB) and belowground biomass production (BGB), plant and soil P concentrations, and N : P ratio in terrestrial ecosystems. Overall, our meta‐analysis quantitatively confirmed existing notions: (i) colimitation of N and P on biomass production and (ii) more P limitation in tropical forest than other ecosystems. More importantly, our analysis revealed new findings: (i) P limitation on biomass production was aggravated by N enrichment and (ii) plant P concentration was a better indicator of P limitation than soil P availability. Specifically, P addition increased AGB and BGB by 34% and 13%, respectively. The effect size of P addition on biomass production was larger in tropical forest than grassland, wetland, and tundra and varied with P fertilizer forms, P addition rates, or experimental durations. The P‐induced increase in biomass production and plant P concentration was larger under elevated than ambient N. Our findings suggest that the global limitation of P on biomass production will become severer under increasing N fertilizer and deposition in the future.

The traditional paradigm is that plant growth at high latitudes is generally nitrogen (N) limited, whereas phosphorus (P) limitation occurs at low latitudes. However, this latitudinal pattern of nutrient limitation is not empirically tested and the underlying mechanisms are far from clear. Here we performed a coordinated experiment of N and/or P addition at three forest sites in China, a subtropical forest, a warm‐temperate forest and a cold‐temperate forest. By measuring relative growth rate (RGR) and leaf nutrient traits among different tree size groups, we assessed how they vary with nutrient addition and tree sizes and uncovered the likely mechanisms underlying these observed responses. Our results revealed that P addition enhanced the RGR of small trees (DBH < 15 cm) by 41% in subtropical forest and 114% in warm‐temperate forest, but reduced it by 57% in cold‐temperate forest. Moreover, small tree RGR increased linearly with soil available P at subtropical and warm‐temperate sites, while it followed a quadratic relationship with soil available N:P ratio at a cold‐temperate site. N addition only affected small tree RGR at the cold‐temperate site. In contrast, the RGR of large trees (DBH > 15 cm) was not impacted by any nutrient addition treatment or soil nutrient variations at any site. Leaf P concentration and resorption efficiency in both small and large trees mostly showed a linear response to soil available P at subtropical and warm‐temperate sites, while leaf N:P ratio in small trees elevated linearly with soil available N:P ratio at a cold‐temperate site. Overall, this study presents robust experimental evidence that growth in small trees, not large trees, is primarily limited by P in subtropical and warm‐temperate forests, but is co‐limited by N and P in cold‐temperate forests. This size‐dependent nutrient limitation highlights the importance of considering tree size classes when assessing nutrient limitation in forest. A plain language summary is available for this article.