Aims
The restoration of degraded ecosystems typically focuses on establishing assemblages of target species, but successful recovery should also be evaluated by the ecosystem’s functioning to guarantee long-term persistence. We investigated how the processes underlying community assembly (i.e. species loss, species gain and changes in abundance of resident species) influenced ecosystem functioning in experimental grassland communities in different restoration states.

Location
A greenhouse experiment in Northern Flanders, Belgium.

Methods
We set up a mesocosm experiment with communities of nineteen planted species, ranging from slow-growing species from poorly productive Nardus grasslands to fast-growing species from highly productive Lolium perenne grasslands. We categorized the mesocosms into different grassland restoration states based on known abiotic and biotic restoration barriers for semi-natural grassland restoration: soil phosphorus levels and soil biota communities. After two growing seasons, we used the CAFE approach, an ecological application of the Price equation, to partition the effects of plant community assembly on ecosystem functioning (here community productivity) for the different restoration states.

Results
Adding soil biota communities sampled from reference Nardus grasslands versus more intensively managed grasslands did not have a significant effect on either plant species richness or biomass productivity. Lower soil phosphorus concentrations (i.e. abiotic restoration) resulted in a higher plant species richness. However, the net effect on productivity was close to zero. The increase in productivity caused by species gains was compensated through decreases in productivity caused by species loss and by decreases in the abundance or functioning of species that are present in both abiotically degraded and abiotically restored states.

Conclusions
Not only species richness but also species identity resulted in changes in ecosystem functioning (i.e. productivity), even though the net functional effects were close to zero. More specifically, we found that species richness-driven increases in productivity were counterbalanced by resource-driven and species identity-driven reductions in productivity.

In June 2017, the harvested biomass was sorted per species, dried for 48 h at 70°C, and weighed. Summing the biomass across species gave the total community biomass, which we used as measure for the aboveground productivity. Immediately after biomass harvest, the 0-5 cm mineral soil layer of each mesocosm was sampled (n=3) with a 1 cm diameter auger. Analysis of bioavailable phosphorus (P_Olsen) was done on the dried soil samples (48h, 40°C) with an extraction with sodium bicarbonate according to Olsen et al. (1954) and colorimetric measurement with the malachite green procedure of Lajtha et al. (1999).

The readme file ('DeCock_AVS_Readme.txt') contains an explanation of each of the variables in the dataset, its measurement units, and the allowed range as well as information on the measurements.