Written by: Andrew Kaul; Reviewed by: Olga Kildisheva
Prairie loss and seed-based restoration
Tallgrass prairies have been removed from all but a few percent of their historic extent. To preserve this ecosystem, prairie restorations have been carried out for nearly a century.
It is common to compare planted prairies (restorations/ reconstructions) to remnant prairies that have never been plowed, to evaluate the success of restoration efforts in terms of species composition and diversity, vegetation structure, and ecosystem functioning (Martin et al. 2005, Barak et al. 2017, Newbold et al. 2020, Kaul & Wilsey 2021, Kaul & Wilsey 2023). Restored prairies generally have lower diversity and floristic quality and have unique compositions compared to remnant areas that were never plowed. In practice, most restoration projects rely on commercially available seed mixes and the diversity and selection of species introduced in the initial seed mix is one of the most important predictors of plant diversity in restored prairies.
The goal of research on the “seed-based restoration pipeline” is to identify those species that occur in remnants but are missing in restorations, and then to assess where these species are being lost in the pipeline. Many studies have directly compared remnant prairies to restored sites (long arrow on left), but more work is required to determine which species, or types of species (i.e. functional groups / traits) are selected for during each of the three transitions. The diagram above shows examples of studies that address these gaps in our understanding of restoration outcomes. Native species found in remnant sites may be lost if they are 1) not hand collected or sold by seed vendors 2) not included in seed mixes, or 3) if they do not establish when seeded.
Transition 1) What species are in the seed market, also known as the “restoration species pool”?
Two recent papers documented biases in the kinds of species that are sold by native plant vendors (Kaul et al. 2023a, Zinnen et al. 2025a). These studies show that there is bias towards species that are longer-lived (woody over herbaceous, perennial over annual), have larger ranges, bloom for a longer duration, and for forbs over graminoids. Both studies also outlined taxonomic/ phylogenetic signals in which species are selected from the species pool. According to these studies, multiple plant families were identified as under-represented in the commercial market, including Orchidaceae, Potamogetonaceae, Cyperaceae, Brassicaceae, Caryophyllaceae, and Euphorbiaceae.
Transition 2) What species tend to be included in restoration seed mixes?
Comparing seed mixes used in restorations to reference remnant sites, can be useful for documenting differences in species diversity, functional group ratios, and functional trait compositions (Kaul & Wilsey 2021, 2023). With the publication of large datasets on seed mixes (Zinnen et al. 2022b) and remnant prairies (Zinnen et al. 2025c) we are now able to broadly quantify these differences across the entire Eastern tallgrass prairie range. For instance, pollinator-oriented seed mixes are lacking the early-blooming forbs common in remnant prairies (Zinnen et al. 2025b).
I am currently working on a project which aims to 1) identify which kinds of species are common to remnants but are often missing from seed mixes, and 2) to quantify how seed mixes differ based on their name and description (Kaul, Zinnen, et al. In Prep). To achieve goal 1) we compared 578 commercial seed mixes from Midwestern seed vendors to species lists for 65 prairie remnants across eight Midwest states. We analyzed a subset of prairie remnants to test how the relative abundance of taxonomic and functional groups differ between remnants and seed mixes. To achieve goal 2) we codified each mix based on keywords in the name to compare mixes based on cost (high diversity vs. economy), soil type (ex. mesic vs. dry), height (ex. tallgrass vs. shortgrass) and intended value to specific animals (pollinators, birds, or wildlife). Preliminary analyses indicate that remnants and seed mixes have highly distinct compositions. We identified several “workhorse” species which are over-represented in seed mixes compared to remnants (ex. Desmanthus illinoensis) and “missing” species which are unique to remnants (ex. Dichanthelium spp., Comandra umbellata). In general, seed mix descriptions were associated with their composition, especially based on height and hydrology. However, mix compositions often overlapped significantly, suggesting that they consist of similar species despite their attributes. Adjusting future seed mix design to include important prairie taxa that are absent from restored sites may improve restoration outcomes by creating more remnant-like communities.
Transition 3) Which species establish and persist when planted?
Many, or even most of the species initially added from an establishment seed mix can be missing within the first few years of community assembly in a novel restoration site. Research on transition 3) investigates which species are “winners” and “losers” and seeks to identify the mechanisms that inhibit the successful establishment of each species (ex. Grman et al. 2015). In a recent study, I worked on a project applying this style of analysis to a seed-based restoration of an herbaceous understory in a temperate woodland (Kaul et al. 2023). We found that recruitment probability was linked to seeding rate, functional group, and conservatism – a measure of species’ fidelity to high-quality natural areas. Species sown at higher rates and with a low “ruderal” conservatism were more likely to have established after two years. This type of analysis is useful for directing future seed mix design, as some species can be sown at lower rates when they establish reliably on a per-live-seed basis. Other species may need to be increased in their relative abundance in the mix to reliably establish. Some groups are establishment-limited and do not recruit even when sown at high rates. Many mechanisms can prevent establishment including genetics, biotic factors, abiotic factors, and site conditions (De Vitis et al. 2022). Only careful research on species recruitment success and growth can inform best practices to increase the success of seed-based restoration efforts.
Literature Cited:
Barak RS, Ma Z, Brudvig LA, Havens K (2022) Factors influencing seed mix design for prairie restoration. Restoration Ecology 30:e13581
Barak RS, Williams EW, Hipp AL, Bowles ML, Carr GM, Sherman R, Larkin DJ (2017) Restored tallgrass prairies have reduced phylogenetic diversity compared with remnants. Journal of Applied Ecology 54:1080–1090
De Vitis M, Havens K, Barak RS, Egerton-Warburton L, Ernst AR, Evans M, Fant JB, Foxx AJ, Hadley K, Jabcon J, O’Shaughnessey J, Ramakrishna S, Sollenberger D, Taddeo S, Urbina-Casanova R, Woolridge C, Xu L, Zeldin J, Kramer AT (2022) Why are some plant species missing from restorations? A diagnostic tool for temperate grassland ecosystems. Frontiers in Conservation Science 3
Grman E, Bassett T, Zirbel CR, Brudvig LA (2015) Dispersal and establishment filters influence the assembly of restored prairie plant communities. Restoration Ecology 23:892–899
Kaul A, Wilsey BJ (2023) Exotic species explain plant functional trait differences between seed mixes, restored and reference prairies. Applied Vegetation Science 26:e12709
Kaul AD, Barash M, Albrecht MA (2023a) Common, showy, and perennial species dominate a restoration species pool. Restoration Ecology 31:e13969
Kaul AD, Dell ND, Delfeld BM, Engelhardt MJ, Long QG, Reid JL, Saxton ML, Trager JC, Albrecht MA (2023b) High-diversity seed additions promote herb-layer recovery during restoration of degraded oak woodland. Ecological Solutions and Evidence 4:e12202
Kaul AD, Wilsey BJ (2021) Exotic species drive patterns of plant species diversity in 93 restored tallgrass prairies. Ecological Applications 31:e2252
Ladwig LM, Zirbel CR, Sorenson QM, Damschen EI (2020) A taxonomic, phylogenetic, and functional comparison of restoration seed mixes and historical plant communities in Midwestern oak savannas. Forest Ecology and Management 466:118122
Martin LM, Moloney KA, Wilsey BJ (2005) An Assessment of Grassland Restoration Success Using Species Diversity Components. Journal of Applied Ecology 42:327–336
Meissen JC, Glidden AJ, Sherrard ME, Elgersma KJ, Jackson LL (2020) Seed mix design and first year management influence multifunctionality and cost-effectiveness in prairie reconstruction. Restoration Ecology 28:807–816
Newbold C, Knapp BO, Pile LS (2020) Are we close enough? Comparing prairie reconstruction chronosequences to remnants following two site preparation methods in Missouri, U.S.A. Restoration Ecology 28:358–368
White A, Fant JB, Havens K, Skinner M, Kramer AT (2018) Restoring species diversity: assessing capacity in the U.S. native plant industry. Restoration Ecology 26:605–611
Zinnen J, Barak RS, Matthews JW (2025a) Influence of ecological characteristics and phylogeny on native plant species’ commercial availability. Ecological Applications 35:e3070
Zinnen J, Chase MH, Charles B, Harmon-Threatt A, Matthews JW (2025b) Pollinator seed mixes are phenologically dissimilar to prairie remnants. Restoration Ecology 33:e14352
Zinnen J, Chase MH, Charles B, Meissen J, Matthews JW (2025c) RELIX: A Dataset of Vascular Plant Species Presence for 353 Prairie Remnants in the Midwestern United States, with Prairie Remnant Metadata. Natural Areas Journal 45:150–158
Zinnen J, Matthews JW (2022a) Native species richness of commercial plant vendors in the Midwestern United States. Native Plants Journal 23:4–15
Zinnen J, Matthews JW (2022b) Species Composition and Ecological Characteristics of Native Seed Mixes in the Midwest (USA). Ecological Restoration 40:247–258