Plant diversity is crucial for a successful ecosystem

Research from the Department of Plant Sciences has contributed to a huge international study that has found that for ecosystems to function well it is important to conserve diverse species both at single locations and across landscapes.

It was already known from small-scale experiments that low diversity communities and monocultures are often less productive and less stable than more diverse alternatives that contain more species.

However this new study, published today Nature Ecology and Evolution, is the first to prove that the same principles apply to landscapes at a bigger, more realistic scale.

The study used an international network of grassland sites called the Nutrient Network, that is maintained and monitored by plant scientists across the globe. At each site (across 15 countries and five continents!), researchers counted the local diversity as well as the species composition across the landscape, and related these to the functioning of grassland ecosystems. They found that ecosystems function best when diversity is high at both scales simultaneously.

Professor Andy Hector from the University of Oxford said: “These new results show the danger that short-term and small-scale research may underestimate the importance of biodiversity for keeping our ecosystems stable and functioning as we’d like them to.”

Lead author, Professor Yann Hautier of Utrecht University, added: “The effects of one boost those of the other probably due to a greater range of ecological traits of plants including height, leaf area and rooting depth, that complement one another when diversity is high both locally and across the landscape.”

The results have implication for grassland management and restoration. For example, rather than seeding an entire field with a single mixture (high local diversity, low landscape diversity), these results suggest that it would be valuable to use different seed mixtures in different parts of the field (high local and landscape diversity).

To read the full paper: