DOT Project Number:  90-00-LRTF-507

Fiscal Year:  2005

Award:  $17,000

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

Other Project Participants:  Andrea Blong, Department of Ecology, Evolution and Organismal Biology, Iowa State University

Summary Report:

NATIVE COVER CROPS: EFFECTS ON WEED INVASION AND PRAIRIE ESTABLISHMENT

INTRODUCTION

The objective of our project is to test whether an early establishing native species can be used to facilitate prairie establishment (Figure 1).  Native cover crops are predicted to facilitate prairie establishment through an indirect pathway by reducing weeds.  This is predicted to occur only if the indirect effect is larger than any direct negative effect of native cover crops on prairie species.  Here we present data from the second year (for cover crops) of a native cover crop study.  Prairie seeds were added in December 2004, and thus, 2005 was only the first year for seedling emergence.  Data on prairie establishment from seed will be presented in the 2006 annual report in June 2007.  The questions that we address are:

1.  Do native cover crops reduce the abundance of weeds?
Smooth brome (Bromus inermis) and crown vetch (Coronilla varia) were the most problematic weeds, so we will report on them.  Late spring fires are often used to control weeds in prairie plantings, and plots were burned in April, 2006. 

2.  Do native cover crops have direct effects on prairie plants? 
This was tested in an experiment with transplants.  Weeds were removed so that the direct effect could be quantified.

3.  Do native cover crops have an overall positive effect on prairie plants?
This is in progress.  We are currently collecting data on prairie plant establishment from seed and will present these data in June, 2007.

4.  Do native cover crops and timing of seeding interact?
This question is being tested in a separate experiment that varies the timing of seeding (spring or summer) and the presence of cover crops (side-oats grama, Canada wildrye, black-eyed susan, partridge pea, or all four species together, or no cover crop as a control).  This experiment was set-up during Fall 2005 and Spring 2006.  It will be reported on during June 2007.

Figure 1.  Model for native cover crop effect.  Native cover crops are predicted to facilitate prairie establishment through an indirect pathway by reducing weeds.  This is predicted to occur only if the indirect effect is larger than any direct negative effect of native cover crops on prairie species.

Study sites

Experimental plots were set up on slopes near roadsides at the Iowa State University farms near Ames (Horticulture Farm, a mesic site) and at the Western Research and Demonstration Farms (a dry site) near Castanea in Monona County.

Experimental design

Seed mixes (see more detailed section below) were added to plots that contain one of 6 cover crop treatments in November 2004.  Cover crop treatments include:

1.  No cover crop (control)

2.  Canada wildrye (Elymus canadensis)

3.  Partridge pea (Chamaecrista fasciculata)

4.  Illinois bundleflower (Desmanthus illinoensis)

5.  Black-eyed susan (Rudbeckia hirta)

6.  Side-oats gramma (Bouteloua curtipendula)

These species are all short-lived and range from annual (Partridge pea) and biennial (Black-eyed susan) to perennial (Illinois bundleflower, Canada wildrye, and Side-oats gramma).  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) 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.

These 6 treatments were applied to experimental plots at each of the two sites (Story and Monona Counties).  These two sites were selected because 1) they represent a mesic and a dry site and broader generalizations can be made as a result, and 2) they represent sites that are conveniently located for the PI, with the Ames site close to ISU and the Monona County site close to other projects.  Six replicate plots were established for each treatment, for a total of 72 plots (2 sites x 6 treatments x 6 replicates).  Plots are 5 m x 5 m in size and were planted in former brome grass fields.  A 2 m mowed corridor is located between each experimental plot to maintain a constant background.  All plots were placed on slopes so that they are more relevant to roadside plantings.

Each cover crop was seeded on April 9, 2004 (Monona County) and April 14, 2004 (Story County) at a rate of 10 lbs/acre and was allowed to establish for the entire 2004 growing season.  Some mowing above the cover crop seedlings of weeds during the first month was done to help in their establishment.   After the first month, weeds were allowed to come in to collect data on weed establishment among cover crop treatments.  The ability to resist weed invasion is the first important part in how useful a species will be as a cover crop (the second part being how well the cover crop lets in prairie species).   Twenty-nine prairie species (see Table in proposal for species list) were added in a seed mix in November 2004 and in December 2005 by adding seed to the snow.  Seed was supplied at a rate of 10 lbs. per acre in each year.  Thus, together with the cover crop seed, there is a total of 30 lbs. seed per acre.  Comparisons of establishment among species are difficult if the same numbers of seeds are 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 Heyne Custom Seed Services (Pottawattamie County) for the Monona County site. 

Cover crop establishment in 2005

Establishment remained high at the Story County site during early 2005 with the exception of partridge pea.  Partridge pea is an annual, and it failed to reseed itself during early 2005.  This will help us to interpret cover crop effects in future years as the prairie species establish from seed.  If partridge pea plots turn out to have different species compositions than other cover crop treatments, then we will know it is from the first year suppression of weeds rather than from facilitation by plants in later years.

Establishment was much higher in spring 2005 at the Monona County site than it was during 2004 (Figure 2, bottom right panel).  Black-eyed susan had become the dominant species in its respective plots.  Canada wildrye and side-oats grama are also coming in fairly well as of June 2005.  Partridge pea and Illinois bundleflower had very poor establishment and would probably not make good cover crops in western Iowa.  Thus, it appears that establishment at the Monona county site was delayed during the first year but is rapidly catching up to the Story County site for all species except the legumes.

Weed invasion into plots during 2005

There are two important characteristics that a cover crop should have for it to be useful in prairie restoration: 1) it should keep out weeds, and 2) it should increase the establishment of target prairie species.  Both of these issues are important, and both will be addressed before recommending a cover crop because a cover crop species that is especially good at keeping out weeds may also keep out prairie species.  The best cover crop species may be one that allows in an intermediate amount of weeds and also allows an intermediate amount of prairie establishment.  We are measuring both characteristics in this study; the first part will be presented here, and the second part is ongoing and will be presented in future final reports.

The most problematic weeds in 2005 were smooth brome (Bromus inermis) and crown vetch (Coronilla varia).  Smooth brome was common at both sites, whereas crown vetch was found only at the Ames site only.  These variables were combined and analyzed for both sites as a general measure of weed invasion.  Crown vetch was analyzed for the Ames site only.          

Cover crops did not affectively keep out these weeds overall during 2005.   This situation may have reversed during 2006 (personal observations, Blong and Wilsey).  The control plots had annual species dominating in 2005; this situation has reversed in 2006, with control plots containing a lot of perennial weed cover.  This will be reported on in 2007.   However, there were interesting differences among species in 2005, and these differences were variable between the two sites (site x cover crop treatment ANOVA, p < 0.01).  Partridge pea, which kept out almost all weeds during 2004 (first year of growth), now has the greatest biomass of smooth brome and crown vetch (Figures 3 and 4), especially at the Ames site.  Canada wildrye had the lowest weed biomass overall during 2005.  Black-eyed susan, which was relatively weed free in 2004, had more weed biomass by August 2005.

In April 2006, we burned the plots.  Interestingly, we found that plots with side-oats grama and Canada wildrye burned at higher temperatures than did plots with other species or control (p < 0.01).  (Fire temperature was measured with paints that melt at known temperatures).  Plots with these two species also had a greater proportion of the plot that was burned than plots with other species or control (p < 0.01).   As of June 2006, it appears that plots with these two species are more weed free than plots with other cover crop species planted.

Prairie establishment from the seed mix

Seedlings of big bluestem, compass plants, ox-eye, tall dropseed, switchgrass, purple prairie clover, and purple coneflower have been observed in the plots.  Compass plant is especially abundant in some plots.  We are sampling establishment of these species and will report our results in the June 2007 report.

Sampling in progress

Surface soil moisture has been measured since 2005 and we will be continue to measure it monthly in each plot during 2006 and beyond using time domain reflectometry rods which have been calibrated with gravimetric soil moisture measurements.  Light availability at the soil surface has been measured with a 1 m long Decagon light in 2005/2006, and will be measured again at least bimonthly during the 2006 growing season.  Light and water capture by the cover crop are predicted to be major factors in either promoting or inhibiting other prairie species.  We are predicting that an intermediate amount of light and water at the soil surface will be associated with the greatest number and cover of other prairie species (Wilsey and Polley 2003), with weed invasion occurring at the high light end, and suppression of the prairie species by the cover crop occurring at the low light end.  Measures of soil water, light at the soil surface, and plant biomass and seedling density will enable us to test this hypothesis.

SUBPLOT EXPERIMENT

Methods

Experimental Design:

At each of the sites, replicated 5 x 5 m plots were seeded with one of five, functionally different, early-emerging species at the rate of 11 kg/ha, or kept as non-vegetated controls (Table 1). Equal masses of seeds were added to the treatment plots in order to place the cover crop species upon equal footing.

We chose these species as potential cover crops because they are among the first to germinate and establish in tallgrass prairie restorations. Dornbush and Wilsey (unpublished data) found that in first-year prairie restoration experiment started with seeds, B. curtipendula, D. illinoensis, and R. hirta were the most numerous seedlings. (Wilsey and Stirling in review).

Non-vegetated controls were also established to test if the presence of an early-emerging species affected the species composition of a developing community. Six replicate plots were established for each treatment, for a total of 36 plots at each site, and 72 plots total in the experiment. Treatments were assigned to plots using a completely randomized design. Each treatment was applied by adding seeds on bare ground a week after tilling in April 2004. Corridors of 3 m between plots were periodically mowed to maintain a constant background.

In June 2004, two 1m x 1m subplots, at least 20 cm from the main plot edge, were cleared of all non-cover crop plants (see Figure 6).

Two plants of each of the following species were be added into each subplot late June/early July in 2004:

1.      Dalea purpurea, a perennial, N-fixing legume

2.      Lespedeza capitata, a perennial, N-fixing legume

3.      Dicanthelium oligosanthes, a perennial, cool-season grass

4.      Stipa spartia, a perennial cool-season grass (only 1 plant in each subplot, due to restricted supplies)

5.      Schizachyrium scoparium, a perennial warm-season grass

6.      Andropogon gerardii, a perennial warm-season grass

7.      Monarda fistulosa, a perennial forb

8.      Ratibida pinnata, a perennial forb

 Transplants were placed in each subplot in random locations, with the restriction of transplants always being at least 10 cm away from each other. When the random locations landed upon a cover crop plant, transplants were placed directly to the side of the cover plant. Subplots were weeded throughout each summer, so that plant-plant interactions would only occur between the cover crop and the transplants.

In 2004, a random subplot was harvested from each plot in October (Ames) and November (Castana). The second subplot was harvested in 2005, in September (Ames), and October (Castana). At the Horticulture Farm, roots of one transplant of each species were excavated to a depth of 15 cm, and the aboveground biomass of both transplants and the cover crop collected. At Western Research Farm, all transplants and their crowns were harvested, and the aboveground biomass of the cover crop plants.

Results

Total Biomass

At the Ames site, the presence of a cover crop decreased total prairie transplant biomass; at Castana, the presence of a cover crop had no significant effect upon prairie transplant biomass. The research sites differed in their amount of transplant biomass. In both 2004 and 2005, the Ames site had significantly higher prairie transplant biomass than Castana (p<0.0001). Restorations at sites similar to Castana will likely not be as productive as sites more like the Ames research area. In addition, at the Ames site, the Control (p<0.01) and Desmanthus (p<0.02) treatments had higher prairie transplant biomasses in 2004, and only the Control treatment in 2005 (p<0.01).

Species composition: Aboveground data

 In 2004, species composition of transplants (aboveground biomass) differed among cover crop treatments and between study sites, most notably in the forb and cool-season grass treatments. Monarda (p<0.01) had higher relative aboveground biomass at the Ames site than at Castana. Conversely, Dalea (p<0.03) and Stipa (p<0.04) had higher relative aboveground biomasses at Western Research Farm. Ratibida (p <0.01) was higher in the Control treatment, and Stipa (p< 0.02) was also higher the Rudbeckia treatment. Other species showed no significant differences among cover crop treatments. At Castana, transplants of Dicanthelium (p<0.03) had higher relative aboveground biomasses in the Desmanthus and Bouteloua treatments; in contrast, at the Ames this species showed no significant differences among treatments. These results show that the same cover crop can have differing effects on community composition depending on the nature of the site.

In 2005, species composition of transplants was skewed more towards the warm season grasses, and although one legume had significant differences. Transplants of Schizachyrium (p<0.03) and Andropogon (p<0.03) had higher relative aboveground biomasses at Castana than at Ames in 2005. Control treatments resulted in higher Schizachyrium transplant relative aboveground biomass (p<0.02), while Andropogon transplants had higher biomasses in Elymus treatments (p<0.01); this was significant only at the Castana site (p<0.03 for Schizachyrium and p<0.02 for Andropogon). At the Ames site, transplants of Lespedeza (p<0.04) showed higher relative biomasses in the Elymus treatments, but at the Castana site, Lespedeza transplants were not significantly different amongst treatments. These results also support the fact that depending on the conditions of a site, a cover crop can have different effects on community composition.

Species composition: Belowground Biomass

The relative root biomasses generally follow the same pattern as the aboveground biomass data. In 2004, Lespedeza (p<0.01) and Monarda (p<0.01) had higher relative root biomasses at the Ames site, while Dalea (p<0.010) had higher relative root biomasses at Castana. Transplants of Ratibida (p<0.01) had significantly higher relative belowground biomasses in Control treatments.

In 2005, only one transplant species showed a significant difference in relative root biomass. Transplants of Lespedeza (p<0.04) had significantly higher relative root biomasses in the Control treatment; this result was significant at the Ames site only (p<0.01).

Species composition: Total Biomass

Relative total biomass data also generally mirror the results of the total biomass data for both years. In 2004, Monarda transplants had significantly higher relative total biomasses at the Horticulture Farm (p<0.01), while transplants of Dalea were significantly higher at Castana (p<0.01). In Control treatments, transplants of Ratibida had higher relative total biomasses (p<0.01), significant only at the Horticulture Farm (p<0.02). Transplants of Stipa had higher relative biomasses in Rudbeckia treatments (p<0.03), but this was significant only at Castana (p<0.04). Transplants of Dicanthelium also showed a significant Site*Crop interaction (p<0.02), having higher relative total biomasses in Western Research Farm Bouteloua treatments.

In 2005, the warm-season grass transplants showed significant responses to two treatments at Castana. Schizachyrium and Andropogon transplants had higher positive responses at Castana (p<0.03 and p<0.04, respectively). The transplants of Schizachyrium had a higher relative total biomass in the Control treatments (p<0.05), while Andropogon transplants were significantly higher in Elymus treatments (p<0.03). Andropogon had a significant Site*Crop interaction (p<0.04), having higher relative total biomass at Castana in the Elymus treatments.

Figure 2: Total above- and belowground biomass in prairie transplant subplots in 2004 and 2005at Horticulture Farm, Ames, Iowa, at Western Research Farm, Castana, Iowa. Please note the differences in the scale of the Y-axis. Please note that legend should be read from bottom to top.

Figure 3: Species composition (relative aboveground biomass) in prairie transplant subplots in 2004 and 2005 at Horticulture Farm, Ames, Iowa, and at Western Research Farm, Castana, Iowa. 

Figure 4: Relative belowground biomass in prairie transplant subplots in 2004 and 2005 at Horticulture Farm, Ames, Iowa, and at Western Research Farm, Castana, Iowa.

Figure 5:  Relative above- and belowground biomass in prairie transplant subplots in 2004 and 2005 at Horticulture Farm, Ames, Iowa, at Western Research Farm, Castana, Iowa.

NEW EXPERIMENT ON SEEDING TIMES/COVER CROPS

 We have established a new set of plots during Fall 2005/ Spring 2006 at each of the two sites.  These new plots will provide a second year of establishment (replicating initial growing conditions) data, as well as test whether spring planting of cover crops differs from fall planting.  A split-plot experimental design is being used.  New plots, again 5 x 5 m, were marked out at each of the two study areas.  Each plot was split up to four 2 x 2 m subplots, with 1 m corridors between subplots.  Each subplot received one of four treatments: 1) spring 2005 planting of cover crop with prairie seed mix, 2) spring 2005 planting of cover crop with prairie seed mix added one year later (spring 2006), 3) fall 2005 planting of cover crop with prairie seed mix, or 4) fall 2005 planting of cover crop with prairie seed mix added the following spring (spring 2006).  The cover crops remained the same with the exception of Illinois bundleflower, which was dropped due to poor establishment in the first experiment.  Canada wildrye, Side-oats gramma, Black-eyed susan, and partridge pea are being used again in the new experiment.  These cover crops are being compared to control plots that will receive the prairie mix only.   However, a sixth treatment was added to this new design: a mixture of all cover crops (but with the same overall seed mass).  This will test the idea that having all of the early emerging species included as a cover crop will be better than having only one species.  The same response variables are being measured on these new plots (except for soil moisture due to budgetary constraints), as explained above.  These plots were established during Spring 2005 and were seeded with their respective treatments in April 2005 or 2006. Many cover crop and other prairie seedlings are emerging as of June, 2006, and preliminary results suggest that soil temperatures are lower and soil moisture levels are higher with spring seeding than with fall seeding.  Spring seeded plots have a different species composition than do fall seeded plots.  Results will be discussed in the 2007 final report.

CONCLUSIONS

1.  The length of time that a cover crop species will suppress weeds is a function of its lifespan.  The annual (partridge pea) and biennial (black-eyed susan) that we used greatly suppressed weeds in year 1 (partridge pea) and years 1 and early in year 2 (black-eyed susan).  Both species failed to reseed themselves, and both species senesced at a time period that weedy perennials like smooth brome and crown vetch tend to establish in the Fall.  Partridge pea was more weedy than any other treatment by 2005 (year 2).  Given that prairie plants take several years to establish, we suggest that perennial cover crops (e.g. Canada wildrye) will provide the better overall long-term weed suppression.  Not using a cover crop, and merely treating the weeds before seeding remains a viable alternative.

2.  Cover crops competed with prairie transplants at the more mesic Ames site.  Cover crops had a nuetral effect on prairie transplants at the more xeric Castana site.  

3.  The cover crops side-oats grama and Canada wildrye produced abundant fuel that led to plots being burned more thoroughly and at much higher temperatures than control plots or plots seeded with other putative cover crop species.    Higher fire temperatures appear to have knocked back smooth brome and crown vetch.  This interesting aspect of cover crop ecology will be reported on in the 2007 annual report.  

Literature cited

Christianson, P. and M Müller.  1999.  An illustrated guide to Iowa Prairie plants.  University of Iowa Press, Iowa City, IA, USA.

Huston, M.A.  1994.  Biological Diversity.  Cambridge University Press, Cambridge, UK.

Jonasson, S.  1988.  Evaluation of the point intercept method or the estimation of plant biomass. Oikos 52:101-106

Packard, S. and C.F. Mutel.  1997.  The tallgrass prairie handbook: for prairies, savannas and woodlands.  Island Press, Washington, D.C.

Mlot, C.  1990.  Restoring prairie.  Bioscience 40:804-809. 

Novecek, J.M., D.M. Roosa, and W.P. Pusateri.  1985.  The vegetation of the Loess Hills landform along the Missouri River.  Proc. Iowa Acad. Sci.  92:199-212.

Rabinowitz, D. and J.K. Rapp.  1980.  Seed rain in a North American tallgrass prairie.  Journal of Applied Ecology  17:793-802.

Shirley, S.  1994.  Restoring the tallgrass prairie.  An illustrated manual for Iowa and the Upper Midwest.  University of Iowa Press, Iowa City.

Smith, D.  1998.  Iowa Prairie: original extent and loss, preservation and recovery attempts. Journal of the Iowa Academy of Sciences  105:94-108.

Wilsey, B.J. and H.W. Polley.  2003.  Effects of seed additions and grazing history on diversity and aboveground productivity of sub-humid grasslands.  Ecology  84:920-932

Wilsey, B.J. and H.W. Polley 2002.  Reductions in grassland species evenness increases dicot seedling invasion and spittle bug infestation.  Ecology Letters  5:676-684.