Competition for nutrients and light: testing advances in resource competition with a natural phytoplankton community

A key challenge in ecology is to understand how nutrients and light affect the biodiversity and community structure of phytoplankton and plant communities. According to resource competition models, ratios of limiting nutrients are major determinants of species composition. At high nutrient levels, however, species interactions may shift to competition for light, which might make nutrient ratios less relevant. The "nutrient-load hypothesis" merges these two perspectives, by extending the classic model of competition for two nutrients to include competition for light. Here, we test five key predictions of the nutrient-load hypothesis using multispecies competition experiments. A marine phytoplankton community sampled from the North Sea was inoculated in laboratory chemostats provided with different nitrogen (N) and phosphorus (P) loads to induce either single resource limitation or co-limitation of N, P, and light. Four of the five predictions were validated by the experiments. In particular, different resource limitations favored the dominance of different species. Increasing nutrient loads caused changes in phytoplankton species composition, even if the N:P ratio of the nutrient loads remained constant, by shifting the species interactions from competition for nutrients to competition for light. In all treatments, small species became dominant whereas larger species were competitively excluded, supporting the common view that small cell size provides a competitive advantage under resource-limited conditions. Contrary to expectation, all treatments led to coexistence of diatoms, cyanobacteria and green algae, resulting in a higher diversity of species than predicted by theory. Because the coexisting species comprised three phyla with different photosynthetic pigments, we speculate that niche differentiation in the light spectrum might play a role. Our results show that mechanistic resource competition models that integrate nutrient-based and light-based approaches provide an important step forward to understand and predict how changing nutrient loads affect community composition.