Pollution from atmospheric nitrogen deposition is a major threat to biodiversity. The 160-year-old Park Grass experiment has uniquely documented this threat and demonstrated how nitrogen reductions lead to recovery. See Letter p.401
Although greater availability of a scarce nutrient might seem beneficial for all plant species, this is not so. Even in seemingly pristine and protected ecosystems, large losses of plant diversity can be caused by the addition of a nutrient that limits plant growth1. One documented example is the effect of deposition on land of nitrogen that was released into the atmosphere by fossil-fuel combustion and agriculture. However, it has been unclear whether plant diversity will recover when nitrogen emissions are reduced or whether additional restoration practices are required. On page 401 of this issue, Storkey et al.2 use the unparalleled long-term data of the Park Grass experiment to show that plant diversity recovers as nitrogen deposition decreases.
The Park Grass experiment at Rothamsted Research in Harpenden, UK, was started in 1856 and is the longest-running study of grassland in the world (Fig. 1). By comparing fertilized and control (never fertilized) plots, the authors observed that plant diversity declined to about 30% of its original level during 135 years of nitrogen fertilization, but returned to about 70% of its original level two decades after fertilization was halted. Moreover, plant diversity declined in unfertilized control plots to about 50% of its original level as atmospheric nitrogen deposition increased from 1950 to 1985. Then, when the introduction of cleaner technologies greatly decreased nitrogen deposition from 1985 to 2012, plant diversity increased to about 80% of its original level. In both recoveries, plant communities tended to regain their former species compositions.
These observations contrast with results of a grassland experiment in Minnesota in which little, if any, recovery had occurred two decades after cessation of high rates of nitrogen fertilization3. Storkey and collaborators suggest the intriguing possibility that this difference is due to the fact that the Park Grass plots have been hayed (the grass cut, dried and removed) twice each year since the experiment started, whereas the Minnesota plots were never hayed. Why might this matter? Haying removes biomass and its nitrogen. If not removed, excess nitrogen that had accumulated in an ecosystem would recycle within that system, thereby retaining its ecological impacts long after nitrogen addition slowed or ceased.
Park Grass hay contains 1.5–2% nitrogen (see Fig. 2 of the paper2), and so annual removal of around 2 and 5 tonnes per hectare of hay from the control and fertilized plots, respectively, probably removed around 35 and 90 kilograms of nitrogen per hectare per year. This removal occurred alongside a reduction in nitrogen deposition of around 25 kg ha−1 yr−1 from its peak, for the control plots, and of an additional reduction of 96 kg ha−1 yr−1 for the plots that stopped receiving fertilizer. The removals of nitrogen through haying possibly hastened the plots' recovery.
The reason why increased availability of limiting nutrients can cause biodiversity losses lies in the evolutionary trade-offs that cause species to be specialists. Adaptations that increase the ability of a given plant species to compete for one limiting resource come at a cost to the species' capacity to deal with limitation by another resource or factor4. Thus, an increased supply of one resource should lead to the competitive displacement of those species that are superior competitors for the enriched resource, because they are among the poorer competitors for the new limiting factor. In theory, if enrichment led to accumulation of a formerly limiting nutrient, both cessation of its addition and decreases in its stores would be needed for that nutrient to again become limiting and for biodiversity to begin recovering.
Although further tests are needed to confirm that haying helped the Park Grass plots' recovery, the idea is supported by several other examples. Human activities release more available nitrogen and phosphorus than all natural terrestrial processes combined5,6,7,8, and accumulation of these nutrients can cause dramatic shifts in species compositions and biodiversity in terrestrial and aquatic ecosystems. For instance, high-diversity heathlands in the Netherlands and Germany were replaced by low-diversity grasslands as nitrogen deposition reached higher rates than those in Britain9,10. Successful restoration of these heathlands often required physical removal of both vegetation and topsoil11,12,13,14. Similarly, 40 square kilometres of vegetable farmland in south Florida became a virtual monoculture of Brazilian pepper trees after it became part of Everglades National Park and agriculture ceased in 1975. Attempts to restore the pre-agricultural ecosystem were futile until both the invasive trees and the fertilized agricultural soil were removed15,16,17. For phosphorus-limited lake ecosystems, reduction of phosphorus inputs can be insufficient for lake recovery if excess phosphorus inputs from agriculture are retained and recycled18.
These cases suggest that both reduction of nutrient inputs and removal of any large stores of accumulated nutrients may be required for restoration of native ecosystems. Some terrestrial restorations also require liming to overcome soil acidification, and seed addition when formerly abundant plant species are absent2,11,13. However, it is not yet clear how the magnitude of increases in nitrogen stores influences the recovery of grassland diversity after nitrogen addition decreases or ceases3,19,20.
The insights from the Park Grass experiment, together with results from earlier studies, show that biodiversity can recover even after chronic high rates of nutrient pollution, and suggest that this recovery may be hastened by, or perhaps require, management practices that reduce accumulated nutrient stores. Moreover, it suggests that haying, a much gentler practice than destructive removal of both vegetation and soil, may reduce nutrient stores sufficiently to allow grassland diversity to recover. Finally, Storkey and colleagues' work demonstrates the great value that long-term studies can provide in identifying solutions to environmental problems.