- Harpole, W Stanley;
- Sullivan, Lauren L;
- Lind, Eric M;
- Firn, Jennifer;
- Adler, Peter B;
- Borer, Elizabeth T;
- Chase, Jonathan;
- Fay, Philip A;
- Hautier, Yann;
- Hillebrand, Helmut;
- MacDougall, Andrew S;
- Seabloom, Eric W;
- Williams, Ryan;
- Bakker, Jonathan D;
- Cadotte, Marc W;
- Chaneton, Enrique J;
- Chu, Chengjin;
- Cleland, Elsa E;
- D'Antonio, Carla;
- Davies, Kendi F;
- Gruner, Daniel S;
- Hagenah, Nicole;
- Kirkman, Kevin;
- Knops, Johannes MH;
- La Pierre, Kimberly J;
- McCulley, Rebecca L;
- Moore, Joslin L;
- Morgan, John W;
- Prober, Suzanne M;
- Risch, Anita C;
- Schuetz, Martin;
- Stevens, Carly J;
- Wragg, Peter D
Niche dimensionality provides a general theoretical explanation for biodiversity-more niches, defined by more limiting factors, allow for more ways that species can coexist. Because plant species compete for the same set of limiting resources, theory predicts that addition of a limiting resource eliminates potential trade-offs, reducing the number of species that can coexist. Multiple nutrient limitation of plant production is common and therefore fertilization may reduce diversity by reducing the number or dimensionality of belowground limiting factors. At the same time, nutrient addition, by increasing biomass, should ultimately shift competition from belowground nutrients towards a one-dimensional competitive trade-off for light. Here we show that plant species diversity decreased when a greater number of limiting nutrients were added across 45 grassland sites from a multi-continent experimental network. The number of added nutrients predicted diversity loss, even after controlling for effects of plant biomass, and even where biomass production was not nutrient-limited. We found that elevated resource supply reduced niche dimensionality and diversity and increased both productivity and compositional turnover. Our results point to the importance of understanding dimensionality in ecological systems that are undergoing diversity loss in response to multiple global change factors.