DOT Project Number:  90-00-LRTF-706

Fiscal Year:  2007

Award:  $9,422.00

Principal Investigator:  Dr. Brian Wilsey, Department of Ecology, Evolution and Organismal Biology, Iowa State University, bwilsey@iastate.edu

Other Project Participants:  Andrea Blong, Graduate Assistant

Summary Report:

NATIVE COVER CROPS AND TIMING OF PLANTING: EFFECTS ON WEED INVASION AND PRAIRIE ESTABLISHMENT

Introduction

Planting cover crops to simultaneously establish native prairie seedlings and prevent weed invasion is an increasingly common management practice for prairie plantings.  The idea is based on the assumption that the cover plant will act as a nurse plant to prairie seedlings, and will have a positive effect on seedling recruitment by increasing weed suppression and by lowering the harmful effects of high evaporation and light availabilities (Figure 1).  Cover crops could also potentially reduce the amount of soil erosion that occurs during planting.  This is predicted to lead to reduced weed biomass and increased prairie establishment in restoration plantings. 

However, the evidence supporting these benefits is mostly anecdotal and has been challenged by some.  Clearly, further scientific evidence is needed on the efficacy of cover plants, whether they reduce weed biomass, and whether cover plants have a facilitative or competitive effect on prairie establishment.

As part of an ongoing project, we have varied cover crop identity and timing of seeding to determine whether prairie establishment will be affected by treatments (Figure 1).

Materials and Methods

There are several native species that have potential as cover crops.  During the 2005 growing season, we established five native species as cover crops at two separate sites (Story County at the Horticulture Farm and Monona County at the Western Research Farm) and measured weed and prairie establishment from a seed mix. 

Experimental plots were set-up in a split-plot design (Figures 2 and 3).  Seed mixes containing 29 prairie species were added to main plots that contain one of 6 cover crop treatments.  Cover crop treatments include:

     1.  Canada wildrye (Elymus canadensis, Ec)

     2.  Partridge pea (Chamaecrista fasciculata, Ca),

     3.  Black-eyed susan (Rudbeckia hirta, Rh)

     4.  Side-oats grama (Bouteloua  curtipendula, Bc)

     5.  No cover crop (Control)

     6.  All four cover crop species combined (Mixed)

These species are all early emerging species that have the potential to reduce weed establishment and help to facilitate establishment of later emerging prairie species.  Five replicate main plots were established for each treatment at each of the two sites.  Plots are 5 x 5 m and were established on tilled areas that were formerly dominated by smooth brome (Bromus inermis).  Within each main plot, four sub-plots (2 x 2 m) were established to receive one of four seed timing treatments: 1) spring-seeded with prairie mix added at the same time that cover crops were established, 2) spring-seeded with prairie mix added the following growing season in the spring, 3) fall-seeded with prairie mix added at the same time that cover crops were established, or 4) fall-seeded with prairie mix added the following growing season in the spring.  Seeding time treatments were established to determine if prairie establishment and cover crop effects would vary between times of the year based on expected peak growth periods (Figure 4).  Biomass of prairie and weed species was estimated with point intercept sampling, which involved counting all plant contacts with a metal pin dropped through the canopy in the middle of each plot during July, 2006 (year 2) and 2007 (year 3).

Cover crop species are all short-lived and range from annual (Partridge pea) and biennial (Black-eyed susan) to perennial (Canada wildrye and Side-oats grama) species.  They all are found in disturbed areas and along roadsides, have high germination rates and are fast growing (Christianson and Müller 1999).  All species are cool season (C₃) plants (with the exception of side-oats gramma, which is warm season or C₄) that have been observed emerging early during prairie plantings.  The inclusion of annuals and perennials will enable us to compare the longer-term effectiveness of the perennial species as cover crops.  Furthermore, by including several species, we will be able to answer the question of whether it is the presence of just any plant that is important, or whether it is a specific species that is important. 

Canada wildrye is a relatively short-lived perennial grass that reaches 3-4 feet tall and has been used as a cover crop plant at Neal Smith National Wildlife Refuge (P. Drobney personal communication).  It establishes well during the first year or two of planting at Neal Smith NWR (personal observation).  It is included here because of its previous use as a cover crop in previous reconstructions (Packard and Mutel 1994).  Thus, its effectiveness can be used as a benchmark in comparisons with other species.  Partridge pea is a N-fixing legume.  Partridge pea achieves a height of 1.5-3 feet tall (Christianson and Müller 1999) in Iowa.   This species is included here because of its potential to aid in the establishment of prairie seedlings due to N fixing capabilities (Hooper 1998).  Black-eyed susan (1-2 feet) is a biennial forb with high germination and quick establishment rates.  This species is included because they have attractive flowers, which would make them popular candidates as native cover crops.  Side-oats gramma is a short perennial warm season grass with high germination and quick establishment rates.  It has been observed to establish very quickly in restoration plots, and seems to be relatively non-aggressive for a warm season grass.

Seed mixes

Seed mixes of 29 species were hand broadcasted as described above.  Seed mixes consist of common warm and cool season grass and forb species of both mesic prairies (Ames, Story County) and dry prairies (Castana, Monona County) (Table 1).  Comparisons of establishment among species are difficult if the same number of seeds is not added to each plot.  For this reason, seed mixes were created with the same number of seeds for each species (based on a previously determined seeds/g value).  Seeds were obtained from Allendan Seed Company (Madison County) for the Story County site and from Custom Seed Services (Pottawattamie County) for the Monona County site.

RESULTS

No prairie species established in 2005, and spring-seeded plots were dominated by annual weed species, whereas fall seeded plots were dominated by biennial weed species as initially predicted (Figure 4).  Weeds in the spring-seeded plots were annuals such as witchgrass (Panicum capillare) and ragweed (Ambrosia artemisiifolia), whereas in the fall-seeded plots biennials were more common (e.g. mullien Verbascum thapsus).  The identity and effects of these weed species, and the changes in soil moisture and temperature that we saw in 2005 and 2006 was predicted to result in a different prairie community, and this prediction has been supported so far based on 2007 sampling. 

By 2007, cover crop treatments had different amounts of weed biomass in both spring and fall plantings (Figure 5, ANOVA, cover crop F_5,48 = 3.3, P = 0.01), and reductions were greatest when treatments matched the optimal growing period of each cover crop.  Canada wildrye had significantly less exotic weed establishment than did control plots (Tukey’s test, treatments with a different letter were significantly different from one another, Figure 5).  None of the other cover crops significantly lowered exotic weed establishment.    

Prairie species from the seed mix tended to be less abundant in the cover-crop plots than in control plots (Figure 5, ANOVA, cover crop F_5,48 = 2.9, P = 0.02), although this was statistically significant only for side-oats grama.  Side-oats grama plots had significantly less prairie species abundance than control plots (Figure 5, different letters denote significant differences according to a Tukey’s test).  Other cover crop treatments were not different from the controls.  The lower prairie establishment with side-oats grama is opposite of what would be predicted if this species was acting as a nurse plant for prairie species.

There were even larger differences among the timing and priority effect treatments. (Figure 6).  For exotic weed abundance, the lowest weed biomass was found in the spring seeded plots with seeds added at the beginning (timing x priority interaction, F_1,144 = 9.7, P < 0.01).  This was also the treatment with the highest prairie establishment from the seed mix (site x timing x priority interaction, F_1,144 = 19.8, P < 0.01), which suggests that prairie species were helping to resist invasion.  Prairie establishment from the seed mix also varied across seeding date and priority effect treatments.  Spring planting supported much higher establishment of both warm season species and prairie forbs which are cool season species.  This was not what we initially predicted.  We originally expected that forbs would establish best in the fall plantings because they are cool season species.  However, we did not find evidence for this.  What this means is that all native prairie species benefitted by the spring planting date, including both forbs and warm season grasses.  

Lower establishment of prairie species in fall seeded plots was associated with high abundance of smooth brome (Bromus inermis) and other introduced cool season grasses.  These species were favored over native prairie species by the fall planting dates, and they were associated with lower prairie establishment.  These species were associated with low overall species diversity as well (data not shown).   

The priority effect treatment, where the prairie mix was added in the growing season after the cover crops were seeded, was meant to mimic a management regime where weeds were controlled initially with the cover crop species, and then the prairie mix was over-seeded onto the cover crop at a later date.  If the cover crops were able to control weed invasion, and if prairie over-seeding of intact vegetation was in fact possible, this management regime would be useful in some cases.  However, we found that only Canada wildrye controlled weeds, and that we found little or no prairie establishment with overseeding during the year after cover crop establishment.   Much more prairie had established when the seeds were added with the cover crop than when it was over-seeded the year afterwards (“Mix” vs. “Later” in Figure 6).  This suggests that this management regime is to be avoided. 

Abundant prairie species that have established and set seed included big bluestem, Indian grass, switchgrass, compass plant, narrow-leafed purple coneflower, butterfly milkweed, porcupine grass, and oxe-eye (Appendices 2 and 3) in spring planted plots, and primrose in fall planted plots.  Fall planted plots are largely dominated by exotic cool season grasses such as smooth brome and Kentucky bluegrass.   Species composition has greatly diverged across the priority and seeding date treatments (Appendix Figure 7).  This suggests that seeding times and priority effects can lead to completely different species compositions, and it will be interesting to see if these differences persist over longer time periods, or if species compositions will eventually converge.  If they persist, then it suggests that alternative stable states can emerge during early establishment phase that will have long-term consequences. 

CONCLUSION

Cover crops had some benefits by reducing weed biomass, but this was counteracted by their tendency to inhibit the establishment of prairie species.  Control plots had as high or higher prairie establishment than cover-crop plots, which is opposite of what you would expect to see if cover crops were acting as nurse plants.  If a cover crop is to be used, we recommend using Canada wildrye.  It reduced weed establishment and it did not significantly lower prairie species establishment.  The worst cover crop species was side-oats grama.  It reduced prairie species establishment significantly below what was seen in control plots, and it had marginal resistance to weeds.

Prairie establishment was highest when prairie seeds were added at the beginning of the project during spring.  Very little prairie establishment occurred when cover crops were over-seeded during the following growing season.  The quick establishment of prairie in spring-planted plots was associated with reduced weed biomass in year 3.  This suggests that quick establishment of prairie species will help plantings resist exotic weed invasions.  As a final note, our work was done in disked brome fields and not in former annual crop fields.  Annual crop fields usually have fewer perennial weeds, and seeding in the fall in that case may lead to greater establishment of prairie forbs than what we found.  Further research is needed to determine how previous land use affects restoration outcomes.  In any case, we suggest that when plantings are done in former brome fields, prairie establishment will be highest with spring plantings without a native cover crop, or with spring plantings with Canada wildrye as a cover crop. 

Finally, plots were burned for the first time during late April and early May, 2008.  Responses to spring burning will be reported on in the 2008 final report.

Appendix 1. Measurements of maximum growth rates, stem production, litter mass production and litter decomposition were made in 2006.  There were large differences in these traits observed among cover crop species (Table 2).  Grasses had narrower leaf widths, and greater stem production than forbs.  Litter mass in spring (when most prescribed fires take place) was similar among species.

Appendix 2.  Frequency (number of plots that each species was found in) of plant species at the Story County site (Horticulture Farm).  Treatments were spring seeded with prairie mix, spring seeded with prairie mix added the following spring, fall seeded with prairie mix, fall seeded with prairie mix added the following spring.

Appendix 3.  Frequency (number of plots that each species was found in) of plant species at the Monona County site (Western Research Farm).  Treatments were spring seeded with prairie mix, spring seeded with prairie mix added the following spring, fall seeded with prairie mix, fall seeded with prairie mix added the following spring.