Increasing the effectiveness of pollinator conservation grasslands within the tallgrass prairie region

Final report for GNC20-301

Project Type: Graduate Student
Funds awarded in 2020: $15,000.00
Projected End Date: 12/31/2021
Grant Recipient: University of Nebraska-Lincoln
Region: North Central
State: Nebraska
Graduate Student:
Faculty Advisor:
Dr. Craig Allen
University of Nebraska, Lincoln
Faculty Advisor:
Dr. David Wedin
University of Nebraska-Lincoln
Expand All

Project Information


Pollinators are declining in the Great Plains, largely due to loss and degradation of native habitat. Given this decline, and the beneficial impacts that pollinators can have on crop production and the resilience of native habitats, it is critical and urgent that we understand the relationship between floral resources, pollinators, and prairie restoration techniques. However, there are significant knowledge gaps regarding the relative benefits that different grassland plantings and their management offer to pollinators and the floral resources that support them. It is also uncertain how different restoration techniques and management impact the economic and ecological trade-offs made by farmers, ranchers, and the conservation community interested in conservation grassland programs such as the CRP, which purportedly support pollinators but have undergone little formal assessment. As such, we designed an experiment with the City of Lincoln, Nebraska ("City"), to compare floral resources offered by different grassland planting methods within a large, public green-ways project, the Prairie Corridor ("Corridor").


Titled "Increasing the effectiveness of pollinator conservation grasslands within the tallgrass prairie region", this project is entering it's second of three years and is comparing the floral resources available for the local bee community from two types of restoration, a high-diversity mix, and mid-diversity pollinator-specific mix. We planted 24 1-acre plots in April 2019 (eight each of high-diversity, mid-diversity, and control), and further randomly assigned half the common grassland management method of mowing, and collected data on plant community composition, the timing, quantity, and diversity of floral resources. We will add data on bee visitation rates for 2020 and 2021. 


As this proposed study is within a multi-million dollar, high visibility public project, it is ideally situated to partner with multiple non-academic stakeholders on research relevant to farmers, ranchers, and communities in the Great Plains. These include the USDA-NRCS Nebraska, the Xerces Society, Prairie Legacy, Inc., Prairie Plains Resource Institute, LLC, and the City of Lincoln. These organizations work with broad audiences, including many farmers and ranchers with land conservation goals and are ideally positioned to offer them recommendations based on the proposed research. The City also has project outreach strategies and outlets in place, and as such is the ideal partner to assist in communicating lessons learned from the proposed research.

Project Objectives:

This project was designed for two audiences: the conservation community restoring native tallgrass prairie in the eastern Great Plains (Lincoln City Parks and other parks, preserves, natural areas) and farmers and ranchers planting grasslands that provide pollinator resources (the Conservation Reserve Program or "CRP", hay meadows, etc.). In 2016, President Obama established the National Pollinator Health Strategy and pollinator habitat became a priority for working grassland programs like the CRP. Agencies spend close to $100 million per year on grassland-related conservation on private lands in Nebraska. Currently, we use lists of plants potentially favored by pollinators, particularly wild bees, when developing prairie seed mixes, but little data exists on how bees use those resources, and how seed mix and management may affect bee usage.


In other words, we poorly understand the return on society’s $100 million dollar investment in grassland conversation in Nebraska. Although high diversity prairie restorations are usually expensive (>$500 per acre), seed mix cost per acre varies widely. As such, this project will help land managers and conservation professionals balance the trade-offs between high establishment costs and high payoffs in terms of pollinator habitat, leading to powerful learning outcomes for both of our audiences. The first expected action outcome is that the City can apply this project’s new information when designing seed mixes for hundreds of acres planned for restoration within the Corridor, with other professionals following suit as information is disseminated to landowners through City outreach and collaboration with the NRCS and The Xerces Society. 


Materials and methods:

Materials and methods

-Experiment design and management

In Google Earth Pro I created 24 permanent plot boundaries within Bobcat of 0.40 ha (±5%) or 1 acre each (Google LLC 2021). On site, we established the plots with T-posts on the west border, and temporary fence posts at plot corners until plot borders were established by mowing. In reality, most plots were slightly smaller than 0.40 ha due to 1) mowed interior paths to delineate plot borders and 2) exterior two-tracks around the site boundary to facilitate access by City personnel for land management activities and lessee access for cattle grazing elsewhere on the tract.

We numbered the plots 1-24 using cattle ear tags, arranged into eight blocks of three plots each. We established eight plots each of three seeding treatments, including a high-diversity local ecotype seed mix of 156 species from Prairie Plains Resource Institute in Aurora, Nebraska (hereafter “high” or “high diversity”), a local ecotype modified “CP-42” pollinator mix of 40 species from Prairie Legacy, Inc in Western, Nebraska (hereafter “mid” or “mid diversity”), and control (hereafter “control”). Species lists for the seed mixes are in the Appendix. Seeding treatments were randomized within each block to stratify the seeding treatments across the entire site and account for heterogeneity within soils, slope, or surviving vegetation within the site.

We prepared the study site for the seeding treatments with two glyphosate treatments to weaken the pre-existing vegetation, mostly B. inermis as it grows actively later into the fall and earlier in the spring than most native species. The two glyphosate treatments were applied using a boom sprayer  to 16 plots (excluding the 8 designated as control plots), on November 18th, 2018 and April 19th, 2019. The seeding treatments were completed 2 days apart, with the mid-diversity mix being drill seeded on May 1st, 2019, and the high-diversity mix being broadcast from an ATV spreader on April 29th, 2019 onto a lightly disked surface to ensure seed-soil contact. The eight control plots were left as is and received no herbicide or seed applications.

Management treatments consisted of annual mowing or shredding in late summer (early-mid August dependent on weather) during 2019, 2020, and 2021. Half of the site (12 plots) were mowed or shredded and half were left as is. The twelve plots receiving management treatment were shredded in the first growing season (2019) due to ruderal plant growth’s marginal value as hay for cattle. The same twelve plots receiving management treatments were then hayed in 2020 and 2021. This created six treatment levels, with four replicate plots per treatment (Table 1).

Table 1. Description of unique treatment combinations and replicates on the Bobcat Prairie experiment.

Seeding treatment

Management treatment

# of plots (replicates)







CP-42 mix (40 species)



CP-42 mix (40 species)



HDLE mix (156 species)



HDLE mix (156 species)




Within each plot, we established 6 permanent cardinally oriented 1x1m subplots on a 2 x 3 grid from which to collect plant and soil data, yielding 144 unique subplots (24 per unique treatment combination as described in Table 1). The long edge of the grid aligns with the long edge of the plots (extending east-west), so the subplots are approximately 30 meters apart from each other. Each subplot was originally established at least 5 meters from the edge of the plot, though this margin varied slightly over the course of the experiment due to mowing and vehicle traffic around the edges of the plots. We marked each permanent subplot with two 6-inch nails and pink fluorescent whiskers on the southeast and southwest corners, hammering them in until they were flush with the ground so they could remain in place throughout haying or any other future management activities. Later on, we additionally marked each subplot with a temporary metal rebar or plastic step-in post that was also coated with pink fluorescent paint at the southwest corner, so subplots were more easily locatable after the vegetation grew to a significant height.

-Data collection

--Vegetation composition and floral resources

Every two weeks during the growing season in 2019, 2020, and 2021 we collected data on species richness, species percent cover, phenological stage, and number of flowering ramets (when present) from all 144 subplots. Sampling timelines for the three growing seasons of the study are summarized in Table x. Species richness was defined as the number and identity (via USDA 4-letter codes) of all unique plant species rooted within the subplot. Percent cover was defined as proportion of cover per species within the subplot, not the percent of total canopy cover (i.e., all percent covers within a subplot could sum to more than 100%). Percent cover was binned into the most accurate category: 1-2%, 2-5%, 5-10%, 10-25%, 25-50%, 50-75%, or 75-100%, and observers were instructed to round down in ambiguous situations to account for the fact that percent cover estimations are more likely to be over- rather than under-estimated (Floyd et al. 1987).

For each species present in the subplot, observers recorded the phenological stage that applied to the majority (>50%) of individuals per species. Phenological stages were vegetative (V), budding (B), partial bloom (P), full bloom (F), or senesced (S). Partial bloom was included as a category to account for those species that have multiple individual blossoms per ramet that may flower at different times. Graminoids were always recorded as vegetative (V) due to their being almost exclusively wind pollinated. All other species were assigned V if they had not yet produced reproductive parts. Finally, if there were any signs of reproductive parts, species were assigned the most appropriate remaining cover class, and the number of ramets that fit that phenological class were then counted. Ramet was defined as any stalk belonging to the same parent plant that split into its own stem at or below the ground surface (i.e., was not an aboveground branch).

At the beginning of the field season, observers worked in pairs or small groups for the purpose of 1) practicing plant species identification and 2) making independent percent cover estimates and then discussing differences until calibrated with the other observers. All plants were identified to species except for Brassica and Carex. Unknown species were either collected and brought back to the lab for identification or documented and identified from photographs if only one individual of the unknown species could be found.

--Soil organic matter, texture, and type

We sampled soils from all 144 subplots once during the study on September 1st and September 2nd, 2021. Within each m2 subplot, we pulled two 8-inch cores (starting from the soil surface) from the southeast and southwest corners, removed visible plant matter, and combined them to make one aggregated sample per subplot. Samples were bagged in plastic Ziplocs, labelled with the date and plot number and transported to Ward Labs in Kearney, Nebraska for analysis. We had all samples (n = 144) analyzed for percent organic matter, percent sand, percent silt, and percent clay. We also had two subplots (1 and 3) per plot assessed for soil texture (n = 48).

--Data entry and storage

Upon returning from the field, vegetation datasheets were first checked for abnormalities such as uncertain IDs or notes of pictures taken/samples collected of unknown species, unclear hand writing, or accidental blank spaces. These abnormalities were corrected or otherwise accounted for on the datasheets in red pen (all original data were recorded in pencil), including identification of any remaining unknowns that could not be identified in the field. Datasheets were then entered via a custom entry form into a Microsoft Access database designed for housing all the data from this experiment. Data entered into the corresponding fields for species code, percent cover class, and phenological stage had to match the pre-existing accepted values (described above) in order to be entered, to reduce errors. Each datasheet was initialed by the student performing the data entry and dated immediately after entering into the database. All vegetation data sheets were then scanned and grouped by sampling event into PDFs for back-up storage, and all vegetation data are presented in table format in the appendix. There were no datasheets created in association with the soil sampling. Soil data are also presented in table format in the appendix.

Research results and discussion:

Data analysis is not complete yet.

For plant data, preliminary results suggest that species richness and diversity are higher in both treatments than in the control plots, but comparable between treatments. Control plots that did not receive haying management were consistently the lowest diversity. Species richness increased during the first three years of the study, even though weedy and disturbance-oriented species declined. This indicates increasing establishment of the native species seeded in 2019. Preliminary reslts of floral resources for pollinators indicate that the high diversity mix has more consistent, but less numerous floral resources on which native pollinators could forage. The mid diversity mix tends to have more temporal variance in floral abundance, with fewer species blooming throughout the season but higher peaks in the floral abundance of those species. 

Preliminary results indicate that soils are more homogeneous than originally suspected, and that the subplots with highest organic matter in the soil were actually comparable to local unplowed prairie remnants. The homogeneity of the soils suggests that this is not a factor driving vegetation composition, but that has not been robustly assessed yet. 

Participation Summary
2 Farmers participating in research

Educational & Outreach Activities

3 Curricula, factsheets or educational tools
2 Journal articles
1 Published press articles, newsletters
2 Tours
1 Webinars / talks / presentations
3 Other educational activities: #1: grassland field research and management methods for undergraduate interns (2 in 2021, 7 total)
#2: PI mentorship of 1 undergraduate senior thesis
#3: field methods training for undergradute internship summer project

Participation Summary:

5 Ag professionals participated
Education/outreach description:

For the project duration (2021), COVID-19 created many limitations to the in-person outreach activities we had planned. However, we did do the following in 2021:

(3): - the PI created visual site descriptions and revised protocols for educating the undergraduate interns (described in "other activity"

Project Outcomes

1 Farmers reporting change in knowledge, attitudes, skills and/or awareness
2 New working collaborations
Project outcomes:

The outcomes of this project exceed those explicitly included in this report, since the project started in 2018. Prior to COVID-19 we were able to take ~15 conservation and agricultural landscape professionals out to tour the site and discuss the research, applications beyond the Prairie Corridor, and concerns of local land managers. Earlier stages of the project also included training and education for 5 other undergraduate interns (on top of the two involved in just 2021) and extensive photo and video (including timelapse of the restoration) documenting for the Platte Basin Timelapse project that seeks to educate on Great Plains landscapes and ecosystems through digital media. 

The more agriculture-specific outcomes of this project included a working relationship with the farmer who conducted the site preparation and site management treatments (haying/mowing) and grazed his cattle on adjacent pastures throughout the course of the experiment. Conversations the PI had with him indicated his deep personal involvement in the study site and interest in the results of the experiment, and desire to learn how to provide pollinators with appropriate forage resources. Although the experiment was still in an incomplete state at the time, anecdotal observations from the study and thoughts from the PI on how to provide pollinator habitat were well recieved. 

More broadly and upon preliminary analyses of the data, this study will likely show the amount of flexibility that farmers can have if they choose to restore pollinator habitat within their broader working landscape. Seed mixes of different species richnesses and installed with different methods (broadcast vs. drill seeding) both offer unique benefits to pollinators, including the ability to increase floral abundance on degraded hay pastures without necessitating any plowing, disking, or disruption of the soil profile, or taking the pasture land out of grazing for an extended period of time. 

Knowledge Gained:

The preliminary analyses of the experiment have highlighted the ability of grassland restoration for pollinators to work within agricultural constraints. This project was designed not with the goal of "turning the clock back" or restoring to an arbitrary point in time, but to explore methods that could be applied within agricultural systems to increase their ability to support pollinators without taking the land out of forms of economic production such as hay production or potentially grazing, although we could not assess that variable explicitly within this study. We also learned how hard it is to set up a spatially small (i.e. smaller than the foraging range of many bees) vegetation study that can also be easily repurposed for pollinator field studies. 


Combining datasets from projects like this (assessing the floral composition resulting from different restoration methods) in tandem with pollinator data, such as flower visitation, nutrition, seasonal (nesting and overwintering) habitat, potential competition between native bees and honey bees, and lethal/sublethal impacts of pesticides and other agricultural chemicals are critical research topics with large knowledge and application gaps in pollinator conservation in working landscapes. 

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.