Ecological restoration practitioners have debated the use of ecosystem models and reference ecosystems as targets for achieving ecological restoration of plant communities.

We used an ecosystem model based on nine tallgrass prairie remnants to establish high-quality tallgrass prairie plots in a prairie restoration. We planted 82 of the most common remnant species in five 10 x10m plots, aiming to replicate remnant-like species composition.

Most species required one to two planting attempts and established within the first few years. Within 5 years, we had planted all 82 species, most persisted and were still present in year six.

The most common clonal species in remnants became most abundant and spread rapidly beyond plot boundaries, while individual plot richness (64-72 species, mean 68.8) was similar to that found in remnant reference ecosystems (82 species).

Two ubiquitous species, Schizachyrium scoparium and Eryngium yuccifolium, showed effects of increased competition (height, biomass, density) due to different seed mixes, with the remnant seed mix showing the greatest effects and the buffer seed mix showing the least effect.

Practical implication: Our results demonstrate the feasibility of replicating the plant composition and species richness of mesic prairie remnants through ecological restoration.

All data were collected between 2018 and 2023. We made annual visual percent cover estimates of all species >1% (1m²) in the plots from 2018–2021, after the initial abundance of agricultural weeds declined and we could accurately estimate which species had become established. We defined an “established” species as one that was observed in 2018 and still present in 2021. In 2021, we made cover estimates of all species in five randomly located, nested 0.25m² and 1m² quadrats (a 0.25m² quadrat nested in the center of each 1.0 m² quadrat) in each 10 x 10m plot. Total plot richness was determined using meander searches throughout the growing season. To assess dispersal and spread beyond plot boundaries, we conducted meander searches in the buffer around each plot for the 42 species planted in the INAI plots but not in the buffer, and recorded the maximum distance that species unique to the INAI model seed mix had spread from each side of each plot. In 2023, two ubiquitous species, Schizachyrium scoparium and Eryngium yuccifolium, were measured for height, biomass, and density in plots and the buffer.

Statistical analysis

We compiled species richness data at four scales: 0.25m², 1m², five 0.25m² quadrats combined, and entire plots (100m²), comparing them to our INAI model. For 0.25m² and 1m² quadrats, species richness counts were made from the five quadrat cover estimates in each plot. We used the five nested quadrat samples from each plot to calculate cumulative species richness at the 0.25m² and 1m² scales. We used cumulative 0.25 m² instead of m² because INAI data use 0.25m², making it more comparable. To do so, we combined the species from the five quadrat samples into a single list. The INAI data consist of 20-0.25m² quadrats so we used a rarefaction curve to estimate richness for 5-0.25m² quadrats. The rarefaction curve is a species-accumulation curve using all possible combinations of the 20-0.25m² quadrats and resulting confidence intervals.