Cover crop cocktails: Harnessing diversity to enhance nitrogen retention in agroecosystems

Cover crop biomass and N content

Cover crop biomass production in year 1 (2011-12) ranged from 100 to 8000 kg ha^-1 (figure 1A). We were not successful in establishing soybean and sunnhemp as winter killed legumes in monoculture or mixtures. As shown in figure 1, we successfully established 4 and 8 species mixtures in year one, though representation by winter killed legumes was nominal. Additionally, oats tended to dominate fall mixtures when present. In year two (2012-2013), we observed higher biomass production ranging from 500 to 10,000 kg/ha (figure 1B). We believe this was due in part to higher levels of fall production promoted by an earlier planting date and warmer fall in year two compared to year one. We observed lower production by the winter hardy legumes (red clover and hairy vetch) in both monoculture and mixture in year two compared to year one. We also saw a dominance of rye in mixtures in spring of year two. These differences in performance are likely due to environmental factors such as climate and soil that mediate not only individual species growth but also competition between species in mixtures. As shown in figure 2, the contributions of cover crop groups varied between years, particularly for those mixtures containing hairy vetch and red clover.         

As we had predicted, aboveground biomass N in mixtures containing non-legumes and legumes, specifically those with winter hardy legumes (red clover and hairy vetch), was similar to biomass N in winter hardy legume monocultures and higher than in non-legume monocultures in year 1 (figure 3A; expected outcome 1). This trend was much less pronounced in year 2 due to the poor performance of hairy vetch and red clover (figure 3B). A comparison of the seeding rate ratio (seeding rate in mixture:seeding rate in monoculture) to the biomass N ratio (biomass N in mixture: biomass N in monoculture) for non-legume cover crops suggests that some species benefited from growing in a mixture with legumes (known as facilitation). A biomass N ratio that is higher than the seeding rate ratio would suggest a facilitation effect. For example, the seeding ratio for rye was 0.5. For mixtures in which rye was grown with a legume, the biomass N ratio was 0.5 or higher (Table 3). We see evidence of a similar facilitation effect for oats grown with soybean and sunnhemp in year 1 and with hairy vetch and red clover in year 2. Forage radish also benefitted from legume facilitation when grown with red clover and hairy vetch in both years and with sunn hemp and soybean in year 2.

Microbial N uptake

Measurements of N contained in the soil microbial biomass do not indicate any effects of cover crops or cover crop diversity on soil microbial N uptake and retention (figures 4 A&B). This in contrary to our expectation that diverse cover crop mixtures would promote greater microbial N uptake during the cover crop season (expected outcome 2) and previous studies that have shown increased microbial biomass (and, consequently, increased microbial N uptake) in the presence of cover crops (17). Our result may have been to do the short-term nature of our study in that cover crop effects on microbial communities may take more than one season to manifest. Additionally, our assessments of soil microbial biomass were performed on bulk soil. Changes in the microbial community, particularly during a single season may be detectable in the rhizosphere of cover crops (the soil within 1mm of cover crop roots), but not in bulk soil.

Nitrate leaching

In year one, legume and oat monocultures experienced N leaching losses similar to the no cover crop control (figure 5A). This is consistent with previous findings that legume cover crops do not perform as well at retaining N as non-legumes (2). Nitrogen losses from oat were likely due to the fact that oats winter kill, make fields susceptible to N leaching in the spring. Monocultures of canola and rye and 4-species mixtures containing these species reduced N leaching by 85-97% compared to the control. The 8-species mix reduced leaching by 65%. The highest N losses were observed in the forage radish monoculture. We attribute this outcome to periods of above average spring temperatures that led to early mineralization of forage radish residues.

Soybean, sunnhemp, and red clover monocultures experienced N losses similar to those in the no cover crop control in year two (figure 5B). Though losses from hairy vetch were higher in year two than in year one, it reduced losses by approximately 50% compared to no cover crop in year two. Losses from forage radish were lower in year two than year one, supporting our spring mineralization hypothesis to explain high losses in year one. With the exception of one 4-species mix, all mixtures reduced nitrate leaching compared to no cover crop, with the greatest reductions (90-99%) by mixtures containing canola and cereal rye. These mixtures, however, did not lead to significantly greater reductions in N leaching compared to canola and cereal rye monocultures.

N provision to corn

As expected, the highest yields (and, therefore, the greatest N provision) were achieved in the red clover (9.7 Mg ha^-1 and 10.2 Mg ha^-1) and hairy vetch (9.8 Mg ha^-1 and 10.8 Mg ha^-1) monocultures in both years (figures 6A & 6B). In year one, three mixtures produced yields similar to the legume monocultures and higher than the no cover crop control (7.0 Mg ha^-1): canola, cereal rye, hairy vetch, and red clover (9.2 Mg ha^-1); forage radish, oat, red clover, and hairy vetch (10.2 Mg ha^-1); and the 8-species mix (9.9 Mg ha^-1). This outcome suggests that N provisioning by mixtures is largely dependent on the presence of winter hardy legumes. Data from year two further show that N provisioning is also influenced by the percent of total biomass contributed by winter hardy legumes. In year one, legumes contributed 40% of the cover crop biomass to the mix of canola, cereal rye, red clover, and hairy vetch; in year two legumes contributed only 15% of biomass and yields were 6.1 Mg ha^-1, lower than the no cover crop control (8.7 Mg ha^-1). Similarly, legumes contributed 50% of cover crop biomass to the 8-species mix in year 1, but less than 10% in year two. This mix yielded 5.9 Mg ha^-1 in year two. In both of these mixes, we concluded that the high proportion of non-legume biomass, particularly cereal rye, led to N immobilization that reduced yields. The final high-yielding mixture in year one combined red clover and hairy vetch with forage radish and oats, and legumes contributed 70% of biomass to this mix. In year two, legumes contributed only 25% of biomass, but this mixture yielded 10.1 Mg ha^-1, similar to the legume monocultures and higher than the no cover crop control. The fact that a mixture combining forage radish, oats, soybean, and sunnhemp produced an equivalent yield (10.2 Mg ha^-1) suggests that N provisioning in both of these mixtures was not due primarily to legumes, but likely a combination of soil N reserves and N available from winter killed non-legumes. Strong N provisioning by the mixture of red clover, hairy vetch, forage radish, and oat in both years despite significant changes in biomass composition suggests that the combination of species in this mix provide an insurance effect for N provisioning.

N services provided by cover crop mixtures

It is clear from our findings that the N services of retention and provision are species and/or functional group dependent – canola and cereal rye (winter hardy non-legumes) lead to strong N retention in monoculture and when included in a mix; hairy vetch and red clover (winter hardy legumes) are strong N providers in monoculture and when they contribute at least 40% or more to mixture biomass; oats and forage radish (winter killed non-legumes) appear to contribute to both N retention and provision under certain conditions. Given that mixtures provide the opportunity to combine these species, it follows that mixtures will exhibit “multifunctionality”, the ability to provide not just one N service, but multiple N services. Using N leaching data as an indicator of N retention and corn yield as an indicator of N provision, our research shows trade-offs between N retention and N provision by monocultures (figures 7 A&B). Certain mixtures, however, were able to both retain and provide N. In year one, these mixtures were those that combined hairy vetch, red clover, canola, and cereal rye. In year two, the multifunctional mixtures were those that combined forage radish and oats with legumes (red clover and hairy vetch or sunn hemp and soybean).

References
2. Tonitto C, David MB, Drinkwater LE (2006) Replacing bare fallows with cover crops in    fertilizer-intensive cropping systems: A meta-analysis of crop yield and N dynamics. Agriculture, Ecosystems & Environment 112:58–72.
17. Buyer JS, Teasdale JR, Roberts DP, Zasada IA, Maul JE (2010) Factors affecting soil microbial community structure in tomato cropping systems. Soil Biology and Biochemistry 42:831–841.

• Figure 1
• Figure 2
• Figure 3
• Figure 4
• Figure 6
• Table 3
• Figure 5
• Figure 7