Progress report for GNC20-295
Kura clover can be integrated into current forage and row-cropping systems as a perennial cover crop, also known as a living mulch. Kura clover living mulch (KCLM) systems are an effective soil and water conservation tool and can increase the flexibility of Midwestern farming systems; as a perennial, it can be used to produce high protein forage one year, and the next year as a living mulch for annual row-crop production. While there are numerous environmental and practical benefits associated with continuous living cover, there are still agronomic and economic challenges in establishing and managing a KCLM system that limits adoption by producers who could otherwise benefit from this innovative conservation cropping system.
The main barriers to producer adoption of KCLM systems are 1) the long establishment period to develop a vigorous stand, 2) working width, horsepower requirements, and availability of equipment to effectively manage the living mulch system, and 3) the limited understanding of nitrogen (N) cycling in these systems, how it is affected by management practices, and how that translates to in-season fertilizer N recommendations.
Addressing these limitations is necessary to mitigate the risks taken on by early adopters and to promote its use in Midwestern commodity production systems. To address the lengthy establishment period, we will test the proposition that seeding kura clover into a mature alfalfa stand will both bolster forage yields and reduce establishment costs by allowing for forage harvest during the transition from a pure alfalfa stand to a pure kura stand. To reduce management barriers after the clover is well established, we will test the effectiveness of newly available cover crop residue management tools and strip tillage implements for maize production in KCLM. To strengthen fertilizer N management recommendations for maize production in KCLM, we will investigate biomass decomposition and nitrogen cycling affected by residue incorporation and fertilizer N application.
These projects fall under the title Removing the Barriers to Adoption of Kura Clover Living Mulch Systems. This work will be included in my dissertation, academic publications, extension articles, and monthly blog posts on the online AgTalk message board. This project will produce important management information for innovative stewards, producers, and conservationists and reduce the risks taken on by early adopters. The outreach efforts included in this research will introduce this system to the Midwestern landscape and this research will provide relevant information for farmers to make educated management decisions.
Through this research and dissemination of its results through scientific publications, popular literature, field days, farm shows, and agricultural forums, producers will gain practical management information required to manage KCLM systems for maize and forage production. This research will improve the awareness of KCLM systems as a potential option for conservation management in the Midwest. We hope that this research will shift the attitudes regarding KCLM systems from a niche conservation tool to a practical agronomic solution. In doing so, this system has the potential to benefit a wide audience of producers, and with its adoption, provide valuable environmental benefits.
With the removal of the barriers to adoption of KCLM systems, producers will more easily and more confidently shift their management to this system or other holistic conservation management strategies. Even if a small number of producers are willing to make this dramatic shift, other farms in their neighborhood will observe these practices in action and be more likely to consider them for their operations. Reflection on the concepts and approaches that this research investigates and utilizes will result in greater adoption of KCLM and other conservation management strategies for Midwestern row-crop production systems.
There were three field studies conducted as part of this research project. The first involved overseeding kura clover into an existing alfalfa stand as a means of increasing forage productivity during the establishment years while transitioning to a pure clover stand that could be used as a living mulch.
To test this method of establishment, inoculated kura clover was seeded in April 2020 and 2021 and in August 2020 at 10 lb live seed acre-1 into an existing alfalfa stand that was established in 2018 at the Rosemount Research and Outreach Center in Rosemount, MN. Forage samples were collected before each hay cutting (4 cuttings yr-1) and samples were processed to determine alfalfa stand count and the relative proportions of alfalfa, weeds, and kura clover biomass.
The second experiment of this project was designed to determine the interactions between row establishment and fertilizer N requirements for maize production in a kura clover living mulch (KCLM).
Field experiments were conducted during the 2021 growing season at the University of Wisconsin - Arlington Research Station in Arlington, WI, and the University of Minnesota - Rosemount Research and Outreach Center in Rosemount, MN. Soils at the Arlington site were a Saybrook silt loam (Fine-silty, mixed, superactive, mesic Oxyaquic Argiudolls), and soils at the Rosemount site were a Waukegan silt loam (Fine-silty over sandy or sandy-skeletal, mixed, superactive, mesic Typic Hapludolls). Kura clover stands were more than 3 years old at each site and the clover was managed as a living mulch for maize production in the previous growing season. While each site was managed under the same cropping system in the previous year, the specific management of the plot area differed, where the Rosemount site received shank strip tillage, broadcast glyphosate application at 2.2 kg ha-1 and no fertilizer N, and the Arlington site received broadcast glyphosate at 4.5 kg ha-1, broadcast glufosinate at 2.5 kg ha-1, and broadcast urea at 240 kg N ha-1.
The experiment investigated two factors, row establishment method and fertilizer N rate, and their effects on maize yield and nitrogen use efficiency. At both sites, individual plots measured 3 by 12 m and were laid out in a split-plot design with four replications, with row establishment as the main plot and fertilizer N rate as the split-plot treatments. Row establishment treatments included rotary zone tillage (RZT), shank strip-tillage (ST), and banded herbicides (BH), and were chosen based on previous work and their practicality for farmers managing a KCLM (Affeldt et al., 2004; Alexander et al., 2019a). Within row establishment treatments, fertilizer N was applied at 0, 45, 90, 135, 180, and 225 kg N ha-1 as SuperU, urea fertilizer with nitrification and urease inhibitors, banded in the center of the row with an EarthWay garden seeder at a 5 cm depth in the RZT and ST treatments and on the surface in the BH treatment (Koch Agronomic Services, Wichita, KS, USA; EarthWay, Bristol, IN, USA). Row establishment and fertilizer treatments were applied on 7 May at the Arlington site and 11 May at the Rosemount site. Dekalb 47-54RIB was seeded on 76 cm rows at 89,000 seeds ha-1 on 8 May and 13 May at the Arlington and Rosemount sites, respectively. Glyphosate herbicide was broadcast at 1.4 kg ha-1 on each plot area on all treatments one week after planting to kill weeds in the row zone and to chemically suppress clover growth during the establishment period. Each site received 100 kg P ha-1, 114 kg K ha-1, and 39 kg S ha-1 as broadcast triple-superphosphate and sulfate of potash 10 d after planting.
Grain, cobs, and stover were harvested on Oct 7 and Oct 21 at the Arlington and Rosemount sites, respectively. Maize ears were picked and stover was cut from a 3 m long row in each plot. Maize stover was weighed and subsampled before the subsamples were chopped into 2-5 cm pieces, bagged, and weighed at the field moisture level. Maize ears and stover subsamples were dried at 60° C until reaching a constant mass. Grain was shelled from the cob and each product was weighed before being subsampled and pulverized. Dry stover subsamples were weighed to determine field moisture content that could be related to the whole sample mass before a subsample was collected and pulverized. Pulverized grain, cob, and stover subsamples were analyzed for N content with the Dumas dry combustion method in an elemental combustion analyzer (Elementar GmbH, Langenselbold, DE) (Bremner and Mulvaney, 1982). Yield and N content data were used to calculate total crop N uptake, N balance (NB = total crop N uptake - fertilizer N) and nitrogen use efficiency (NUE = total crop N uptake/fertilizer N), which have potential to indicate differences in in-season soil N contributions from the living mulch (Congreves et al., 2021).
Dependent variables were tested for normality and homogeneity of variance by visual inspection of QQ plots and residual vs fitted plots, respectively (Kutner et al., 2004). When data from both sites were tested, several of the dependent variables violated one or more assumptions of ANOVA, and data transformations were unable to satisfy the required conditions. Because of this, data were separated between sites, retested, and when needed, appropriate data transformations were made to satisfy assumptions of normality and homogeneity of variance. Data were then fit with linear mixed-effects models in the lme4 package (Bates et al., 2015) in R (R Core Team, 2020), where row-establishment method and fertilizer N rate were fixed effects, and the main plot nested within block was the distribution of random error (Langhans et al., 2005). Models were tested for and met the assumption of homogeneity of variance by visual inspection of residual vs fitted plots (Kutner et al., 2004). Mixed models were then tested with the type III F test with Kenward-Roger estimated denominator degrees of freedom using the car package in R. When P ≤ 0.05, treatment means were separated using the emmeans (Lenth et al., 2018) and multcomp (Fox and Weisberg, 2019) packages in R, and results were presented using compact letter display (Piepho, 2018).
Weather data were collected from the UW Arlington Farm weather station of the National Weather Service and the US-Ro6 weather station of the AmeriFlux network, for the Arlington and Rosemount sites, respectively. Drought severity and coverage index (DSCI), which represents an integrated value of drought severity for a region ranging from 0 to 500, was collected from the US Drought Monitor for Columbia County, WI, and Dakota County, MN, for the Arlington and Rosemount sites, respectively.
Finally, research was conducted to track biomass turnover, N contributions from the living mulch, and N species pathways in a KCLM managed for maize production.
An experiment was conducted during the 2020 and 2021 growing seasons to investigate N cycling within a KCLM system for maize production. The site was located at the University of Minnesota’s Rosemount Research and Outreach Center on a Waukegan silt loam (fine-silty over sandy or sandy-skeletal, mixed, superactive, mesic Typic Hapludolls). Kura clover was seeded at the site in the spring of 2018 and the stand was managed as a forage crop during the 2019 growing season. The site was prepared by mowing clover to a 5 cm height on 11 May 2020 and 10 May 2021 before rows were established with a shank-type strip tillage tool (Orthman Manufacturing, Lexington, NE, USA). Fertilizer N was applied on the same day as strip tillage as SuperU, urea fertilizer with nitrification and urease inhibitors (Koch Agronomic Services, Wichita, KS, USA), banded within the tilled strip at a 5 cm depth with an EarthWay garden seeder (EarthWay, Bristol, IN, USA). Maize hybrid Dekalb 47-54RIB was seeded at 79,000 seeds ha-1 on 14 May 2020 and 13 May 2021, and glyphosate was broadcast onto the plot area at 1.4 kg a.e. ha-1 on 30 May 2020 and 25 May 2021.
Clover biomass and decomposition
Aboveground clover biomass was collected before mowing, after mowing, and weekly after treatment application. Clover was manually cut from a 0.5 m2 quadrat that spanned one row width before samples were dried at 60° C until they reached a constant mass. Samples were weighed and adjusted to the dry kg ha-1 basis.
Litter bags were prepared by cutting 6 x 12 cm segments of 500 μm nylon mesh material that was folded in half on the long edge before two of the edges were sealed with a soldering iron. The empty mesh bags were weighed before ~5 g of dry clover biomass was added, and the final edge was sealed. Bags were placed onto a reciprocal shaker for 2 h at 240 rpm to sieve and remove residue particles that were <0.5 mm. Finally, the prepared litter bags were weighed to determine the initial clover residue mass. Litter bags were placed at 0, 10, and 20 cm depth within the tilled row and collected after 1, 2 , 3, 4, 6, and 8 wk after installation. Bags were dried at 60° C until they reached a constant mass before they were weighed, cut open, and the partially decayed clover residue was removed. Empty bags were weighed again to determine the amount of clover residue remaining after deployment. Small root fragments were removed from the decayed clover residue and weighed before samples were pulverized and homogenized with a mortar and pestle. Decayed litter samples were subsampled and analyzed for total C and N using a combustion analyzer (vario MAX cube, Elementar, Langenselbold, DE) (Stevenson, 1996). The remaining residue was weighed and combusted in a muffle furnace at 900° C for 4 h and the remaining material was weighed to determine the amount of ash introduced from the mineral portion of the clover residue and soil contamination during deployment.
Clover residue burial and distribution was determined by tracer analysis (Allmaras et al., 1996). 500 plastic beads were placed on the soil surface in the area that is intercepted by the strip tillage tool. The beads were tilled under the same conditions as the plot area, and after tillage, beads were recovered from the surface, 0-5cm zone, and 5-10 cm zone in 5 cm increments on each side of the tilled strip. Bead tracers were recovered from four replications and bead location was used to determine a probability matrix of the spatial distribution of buried clover residue following strip tillage.
After treatment application, soil was sampled weekly in the center of the maize row and the center of the interrow zone from the 0-15 and 15-30 cm depths with a 2 cm i.d. corer. Samples were weighed before 8-10 g was subsampled and added to 38 mL of 2M KCl solution and shaken for 1 h at 120 RPM (Mulvaney, 1996). The extractant was filtered with 1 μm filter paper and analyzed for NO3-N and NH4-N with the Griess-Ilosvay with Cd reduction and sodium salicylate-nitroprusside method, modified for flowthrough injection analysis, respectively (Lachat, Loveland, CO, USA) (Mulvaney, 1996). A second 5 g subsample was collected, weighed, and dried at 105° C for 24 h, and weighed to determine the mass water content which could be related to the whole sample mass to determine soil bulk density. Soil N concentrations were adjusted to the dry soil basis using the soil moisture content and to the kg ha-1 basis using bulk density measurements.
Nitrous oxide emissions
Nitrous oxide emissions were measured from the plot area bi-weekly using semi-static chamber methods (Parkin and Venterea, 2010). Because of the distinct spatial and temporal distributions of N between row and interrow zones in the KCLM system, each were measured separately. Chamber bases with a 29 x 50 cm or a 15 x 30 cm footprint were inserted at least 50 mm into the ground in the center of the interrow and 3 cm to one side of the maize row, respectively. On each sampling date, three 12 mL gas samples were collected near the soil surface with a syringe. Samples were immediately transferred to glass vials sealed with a butyl rubber septum and used as the T0 gas measurement. The interrow zone contained clover that grew and died back throughout the sampling period. Biomass that exceeded chamber top height was folded into each top upon chamber placement and the displaced chamber volume was estimated using clover biomass measurements and constant values for clover moisture content and biomass density (Alexander et al., 2019). Insulated and vented chamber tops were placed onto chamber bases and secured using binder clips in 60 s intervals. After installation, 12 mL samples were collected from each chamber after 30, 60, and 90 minutes. Gas samples were analyzed for N2O and CO2 concentration with a 5890A Gas Chromatography analyzer (Hewlett–Packard, Palo Alto, CA, USA) in conjunction with a 7000 Headspace Autosampler (Teledyne Tekmar, Mason, OH, USA). Measured gas peak areas were adjusted by subtracting the peak area of vials that contained ambient concentrations of gasses during vial preparation and adjusting the sampled volume by the air temperature during sampling relative to the ideal gas constant. Gas fluxes were calculated using the restricted quadratic method, where the first derivative of the quadratic regression between N2O concentration and time is used unless the second derivative of the resulting quadratic function is greater than 0, in which case a linear function is used (Parkin et al., 2012).
Overseeding kura clover into the established alfalfa stand did not work as intended. The clover was outcompeted and soil and environmental conditions did not allow for clover germination. No kura clover was found in the processing of samples from this project and data collection ceased after the first cutting in 2021.
The interactions between row establishment method and fertilizer N rate offered several insights into the performance of KCLM systems under varying intensity of drought.
Growing conditions at the two sites were challenging during the 2021 growing season. The Arlington site received 64% of the 1991 – 2020 average rainfall between 20 Apr and 31 Oct resulting in moderate drought conditions for most of the growing season (Figure 1). Drought conditions were characterized using the Drought Severity and Coverage Index (DSCI) (National Drought Mitigation Center et al., 2021). Despite this, rainfall was relatively well distributed, and temperatures were near the 1991 – 2020 average, allowing maize plants to mature normally (Figure 2). The Rosemount site received very little (< 16 mm) rainfall in the 23 days before and 4 days after planting and entered a severe drought (DSCI > 200) on 27 July that ended on 24 August. Air temperatures were higher than the 1991 – 2020 average beginning in June until the end of the growing season. Underdeveloped maize plants nearly died in two of the four blocks at the Rosemount site.
At Arlington, row establishment treatments influenced grain yield, grain N, stover N, total N, NB, and NUE (Table 2). Grain yield and Grain N were greatest in the RZT treatment and lowest in the BH treatment, while ST was not significantly different than RZT or BH treatments (p < 0.05). Stover N was greatest in the RZT treatment, lowest in the ST treatment, and the BH treatment was not significantly different than the RZT and ST treatments. Total N uptake, NB, and NUE was greater in the RZT treatment than the ST and BH treatments.
Grain N, Stover yield, Stover N, Total N, NB, and NUE were significantly affected by fertilizer N rate. Grain N was lowest in the unfertilized treatment and greatest in the 90 – 180 kg N ha-1 treatments. Stover yield was lowest in the unfertilized treatment and greatest in the 135 and 225 kg N ha-1 treatments. Total N was greater in the 90 – 225 kg N ha-1 treatments than the unfertilized treatment. N Balance and NUE were maximized in the low N rate treatments and followed a downward trend with increasing fertilizer N rates.
At Rosemount, row establishment treatments significantly affected Grain yield, Grain N, Total N uptake, and NUE. For each of these response variables, the BH treatment was significantly greater than the RZT treatment and the ST treatment was not significantly different than the BH or RZT treatments at the p < 0.05 level. Nitrogen Balance was also affected by the row establishment treatment, where the BH and ST treatments took up a larger fraction of the applied fertilizer N than the RZT treatment.
Stover yield, Stover N, Total N uptake, NB, and NUE were affected by fertilizer N rate treatments. Stover yield was greatest in the 135 and 180 kg N ha-1 treatments, lowest in the unfertilized plots, and the 45, 90, and 225 kg N ha-1 treatments were not significantly different than the highest and lowest yielding treatments. Stover N was greatest in the 135 - 225 kg N ha-1 treatments, lowest in the unfertilized treatment, and the 45 – 90 kg N ha-1 treatments were not different than the 0 and 180 kg N ha-1 treatments. Similarly, total N uptake was lowest in the unfertilized treatment, greatest in the 135 – 225 kg N ha-1 treatment, and the 45 – 90 kg N ha-1 treatments were not significantly different from high and low uptake treatments. Nitrogen Balance and NUE were maximized at the low N rate treatments and followed a negative trend with increasing fertilizer N application. The NB was positive in the 0 - 45 kg N ha-1 treatments and reached -149 at the highest N rate.
Results from these experiments highlight the risks of utilizing a KCLM system for maize production in rainfed cropland in the upper Midwest. While the Arlington site had acceptable grain yields at all fertilizer N rates despite moderate drought conditions during some periods of the growing season, the Rosemount site nearly experienced a total crop failure. Differences in crop performance between sites is almost certainly a result of crop moisture stress. While the Arlington and Rosemount sites received a similar amount of total rainfall, the distribution varied. Rainfall was relatively well distributed over the growing season at the Arlington site, but the Rosemount site received most of its precipitation in May and early September. Research into the distribution of irrigation has found that irrigation in July has a positive correlation with grain yield while irrigation distributed mostly in August or September is neutral or negatively correlated with grain yield, respectively (Payero et al., 2009). These findings are well aligned with the observed conditions during the 2021 growing season at Arlington and Rosemount, which despite a similar amount of total rainfall, resulted in drastically different crop performance.
Moisture stress may have also been influenced by soil quality and site history. While grain yields across all plots at Rosemount yielded 3.6 Mg ha-1, the plots that were less affected by drought yielded 5.3 Mg ha-1, which is not significantly different than conventionally managed plots adjacent to the study area, which yielded 5.2 (SD = 0.6) Mg ha-1 (unpublished data). Differences in crop development and maturity between blocks within the plot area were apparent early in the growing season, and without rainfall in early September, it is likely that the most severely affected plots would not have made grain. The highest and lowest yielding plots at the Rosemount site were spatially correlated and differences in crop performance are likely attributed to soil depth and texture, which are well known to exacerbate the effects of drought (Jia et al., 2013). Because of the variability in soil quality across the plot area, we cannot say whether the KCLM exasperated or helped the maize crop respond to drought conditions.
Differences in moisture stress between the two sites may have also been influenced by initial clover stand health. Kura clover growth is most vigorous from mid-April to late-May and typically accumulates ~2 Mg biomass ha-1 during that time. Ochsner et al. 2010, analyzed soil water content and estimated evapotranspiration and found that a KCLM system imposed a water deficit of 37 – 50 mm in May relative to monocrop maize (Ochsner et al., 2010). Living mulch management imposes prolonged stress on kura stands and depletes belowground clover biomass reserves (manuscript in preparation). While both sites were used as a living mulch in the previous year, the management of the sites may have favored increased relative clover health at the Rosemount site. Specifically, the Arlington site received broadcast glyphosate at 4.5 kg ha-1 and glufosinate at 2.5 kg ha-1 while the Rosemount site was managed with strip tillage and broadcast application of glyphosate at 2.2 kg ha-1. The additional broadcast herbicide application may have increased the suppression of clover in the interrow and reduced the spring vigor and water use of the clover stand at the Arlington site.
Row establishment treatments influenced maize performance at both sites. At Arlington, the RZT treatment outperformed the ST and BH treatments, and at Rosemount, the BH treatment outperformed the ST and RZT treatments. While this experiment did not collect in-season soil moisture, clover biomass, or crop development data, the literature suggests that these observed treatment differences could be attributed to moisture conservation, inter-species competition, and unique weather conditions during the study period. Conservation tillage management, including no-till, is known to increase soil moisture availability and reduce evapotranspiration during the early growing season (Blevins et al., 1983). This period was critical at the Rosemount site because of dry planting conditions, whereas Arlington received rain the day before tillage and planting. Because soil moisture at Arlington was not limiting as seedlings emerged, the more intensive mechanical suppression of clover from the RZT treatment likely sped the emergence and development of maize at this site (Dobbratz et al., 2019). The Arlington site entered a moderate drought on 8 June that lasted until 20 July, and the better developed and more deeply rooted plants may have been able to better withstand the moderate moisture stress during that period. The Rosemount site did not receive significant rainfall in the 23 days preceding and 7 days after planting. The low moisture in the seed zone of the RZT treatment likely slowed maize emergence and development relative to ST and undisturbed BH treatments due to the water conservation of less intensive tillage. The Rosemount site entered a moderate drought on 8 June and a severe drought on 27 July that lasted until 24 August. Crops that were not well established before this time nearly died.
Fertilizer N rate treatments did not affect maize performance at either site. KCLM systems have been found to reduce fertilizer N requirements for maize in previous research, however, because of the unique environmental conditions during the 2021 growing season, conclusions regarding the N management of KCLM-maize systems will not be made based on the observed data. Nitrogen efficiency metrics were unable to reveal differences in the N contributions from row-establishment treatments. This result is unsurprising, as grain yields at both sites were unresponsive to fertilizer N rate treatments. These metrics might be useful in future N rate research to determine estimates of in-season N contributions from the KCLM or other legume inter-cropping systems under N limited conditions.
Results from the experiment detailing biomass turnover and N cycling are still being processed.
Educational & Outreach Activities
In the last year, I presented my experimental results with farmers and agricultural professionals on the AgTalk message board. My research summaries were viewed hundreds of times and received several comments and questions. Roughly half of my outreach with farmers was conducted with these summaries on the AgTalk forum.
I also had the opportunity to present my preliminary results with the scientific community at the 2020 Soil Science Society of America conference. This conference was online-only, and so I do not have high confidence in my estimated attendance, however, these presentations are now archived and available online for anyone to view.
In February 2022, I presented my research at the 8th Annual Nitrogen: Minnesota's Grand Challenge and Compelling Opportunity Conference, which was attended by 130 producers and agricultural professionals. There was high interest in my project and I made connections with people in attendance who I've stayed in contact with for personal consultation.
There are several activities that are in progress and will be completed before the end of the year. I am ~2 months away from defending my dissertation and two of the chapters focus on projects partially funded by the SARE Graduate student grant. One of these chapters was recently submitted for publication in Agronomy Journal and I will be presenting my results to the faculty and students in the Department of Soil Water and Climate during my dissertation defense.
This research has developed revised management recommendations for corn production in KCLM, identified management practices that do not work well for establishing new clover stands, and led to new hypotheses that will be tested in future funded work. The insights that were gained with this research will follow me into the next part of my career as a post-doc, where I will investigate agronomic, economic, and environmental aspects of managing a KCLM system for corn production. The support for this work is gaining in the agricultural community and the continued development of this cropping system has the potential to increase economic and environmental sustainability in the upper Midwest.
We each learned important information regarding kura clover living mulch systems for corn production, and with that, our general knowledge regarding conservation agriculture has grown. Our research was very direct and our findings were explained by our existing conceptual view of the cropping system. We feel more confident in our understanding and management of KCLM systems and are more comfortable answering questions about the cropping system and promoting its use to farmers and agricultural professionals.