A comparison of an oilseed radish cover crop in a corn rotation to a corn rotation without a cover crop in Wisconsin determined that a radish cover crop appears to have several benefits for some producers. Although fall radish nitrogen (N) uptake can be quite substantial, no N credit was determined. However, spring soil nitrate samples did suggest a N credit. Data gathered from a soil penetrometer indicated differences in soil compaction in the upper profile early in the season, as well as differences in the deeper profile later in the season.
Radish has become a popular cover crop option throughout the United States within the past decade. Among its touted benefits are N scavenging, compaction reduction, and pest suppression. To date, much of the radish cover crop research has been conducted in the Mid-Atlantic region of the United States. Weil and Kremen (2007) demonstrated that brassicaceous cover crops took up more N in the fall than rye, which was the standard N capture cover crop in Maryland. Brassicaceous cover crops also rapidly depleted soluble N from soil profiles in the fall which reduced spring N leaching. These results were confirmed by other studies (Kristensen and Thorup-Kristensen, 2004; O’Reilly et al., 2012; Dean and Weil, 2009). Kristensen and Thorup-Kristensen (2004) proposed that radish was effective at capturing N from deep soil layers due to its deep root system that increased in root intensity from the soil surface to 1.5 m depth. Other common cover crops, such as ryegrass and rye have shallow roots concentrated at the soil surface. It is important to note that in most of the University of Maryland experiments, the authors applied 56 kg ha-1 of N to ensure adequate growth on sandy textured soil. It is unclear if this extra N is necessary for non-sandy soils.
A seminal paper by Dean and Weil (2009) demonstrated the benefit of radish on N uptake and reduction of residual nitrate. Averaged across 3 years, radish had greater dry matter production and captured more N in fall than rye shoots, a popular and widely-used cover crop. However, in this study, the authors applied 56 kg ha-1 of N at the same time of radish planting to ensure adequate growth on sandy textured soil. It is unclear on non-sandy soils if extra N is required for adequate growth, although in most Wisconsin systems, radish will be applied after a manure application. A more recent paper from Chen and Weil (2011) showed that radish planted in the fall prior to corn planting resulted in a significant yield increase compared to plots with no cover crop. Similarly, results reported by O’Reilly et al. (2012) showed that planting radish as a cover crop increased yield in sweet corn when compared to no cover crop, especially in plots where less than optimal nitrogen rates were applied. This may also suggest there is a N credit for radish. Dean and Weil (2009) provide a summary of eight other studies that have shown that radish has at least some level of positive impact on capturing soil nitrate or improving crop yields.
The effects of radish were also seen during the subsequent crop’s growing season. Both corn and soybean plants produced more dry matter and had greater tissue N when following radish as compared to no cover crop (Weil and Kremen, 2007; O’Reilly et al., 2012). In late summer, soil moisture sensors showed more rapid infiltration of water into the soil after rains in the subsoil of plots previously planted in radish (Weil and Kremen, 2007). Williams and Weil (2004) suggested that the root channels left behind by radish provided the subsequent crop (soybean, in this case) roots with low resistance paths to water contained in the subsoil. Radish planted in the fall before corn resulted in a significant corn yield increase compared to plots with no cover crop (Chen and Weil, 2011). These results are similar to those reported by O’Reilly et al. (2012) who demonstrated that planting radish as a cover crop increased yield in sweet corn, especially in plots with less than optimal N; however, a N credit was not determined. Dean and Weil (2009) showed that radish decreased nitrate-N concentrations in soil pore water on fine textured soil compared to the control. Soybean yields were also significantly greater following radish cover crop treatments when compared to other brassicaceous cover crop treatments as well as the no cover crop treatment (Williams and Weil, 2004). Other than O’Reilly et al. (2012), there was a lack of studies that evaluated a potential N credit for radish, which is possible based on the amount of N in the whole plant biomass (119 kg N ha-1 according to Dean and Weil, 2009) and a favorable C:N ratio for net mineralization (20:1 according to Clark, 2007).
Researchers have investigated the impact of radish on soil resistance and soil compaction as well. In 2007, Weil and Kremen used minirhizotron images to confirm that soybean roots penetrate hard plow pans by following channels made by radish the previous fall. Chen and Weil (2010) found that under high and medium soil compaction, radish had significantly more roots penetrate the soil than either rapeseed or rye. This conclusion was found to be especially true in soil with high clay content, which has significant implications for clay type soils found in Wisconsin. Corn planted after radish also had more deep roots under high soil compaction than rye or no cover crop (Chen and Weil 2011). Weil and Kremen (2007) proposed that this could be due to the fact that soil cores taken in late summer revealed about 10 times more corn roots in plots previously in radish than in plots previously in no cover crop. A recent study by Chen et al. (2014) showed that radish also had greater air permeability than other cover crops as well as the no cover crop treatment due to an increase in roots into compacted soils.
Root lesion nematodes (Pratylenchus spp.) (RL) are the third most economically damaging nematode in the world for agricultural crops, behind root-knot and cyst nematodes. This is due to their wide host range (more than 400 crop plant species), as well as their wide environmental distribution (Davis and MacGuidwin, 2000). Pratylenchus spp. are the most common nematode pest of corn in the Midwest (MacGuidwin and Bender, 2012; Tylka et al., 2011). In 2012 the Wisconsin Soybean Marketing Board expanded their testing program to include other pest nematodes at no charge so that growers would be able to monitor their total nematode populations. Of the 315 samples collected, 96% tested positive for RL nematode. Out of the samples that tested positive, 20% were above the damage threshold and were distributed throughout WI (MacGuidwin, 2013). Root lesion nematode damage is often misdiagnosed as nutrient deficiencies and can even cause yield loss without any visible aboveground symptoms (A3646; Davis et al., 2015).
Nematode suppression from brassicaceous plants occurs due to the glucosinolate compounds (organic compounds that contain sulfur and nitrogen) and the enzyme myrosinase contained in their tissues (Brown and Morra, 1997). It is the degradation products of glucosinolates reacting with myrosinase that are toxic to soilborne organisms (Donkin et al., 1995). These degradation products include isothiocyanates (ITCs), thiocyanates, and nitriles, although ITCs are considered the most toxic (Brown et al., 1991). Glucosinolates and myrosinase are physically separated in Brassica plants; therefore, it takes some form of physical disruption of the plant tissue to bring the two compounds together. Since this is a hydrolysis reaction, water is required for the reaction to become active and release ITCs (Matthiessen and Kirkegaard, 2006). Several studies have shown that ITCs can suppress plant-parasitic nematodes (Mojtahedi et al., 1991; Zasada and Ferris, 2003, 2004). It is believed that this process is part of plant defense against insects and pathogens (Matthiessen and Kirkegaard, 2006).
Brassicaceous cover crops are well known for their potential for pest nematode suppression. Both Brussels sprout and horseradish plant material amendments reduced citrus nematode survival by 39 and 59%, respectively, as compared to the control (Zasada et al., 2003). Rapeseed grown for two months, then incorporated into the soil, was more effective at reducing nematode population density than the control (Mojtahedi, et al. 1991). In a field experiment, both radish and radish intercropped with oats demonstrated lower reproduction rates for Pratylenchus brachyurus (Chiamolera et al., 2012). Radish reduced plant-parasitic nematode populations by 55.7% when compared to a no cover crop control (Wang et al., 2009). Out of the 11 radish varieties tested, oilseed radish had the highest average content of glucosinolates (Ciska et al., 2000). Zasada and Ferris (2004) demonstrated that brassicaceous amendments added to the soil based upon glucosinolate profiles could be applied to achieve consistent and repeatable nematode suppression. High concentrations of ITCs in the soil after brassicaceous biomass incorporation indicated that brassicaceous plants would have a higherpotential to suppress nematodes (Hansen and Keinath, 2013).
Utilizing brassicaceous cover crops for nematode suppression is a broad-spectrum tactic; all soilborne nematodes are targeted. Therefore, plant-parasitic nematodes such as soybean cyst nematodes (Heterodera glycines) (SCN), which are not corn pests, would also be suppressed. Only soybean is a host for SCN, but the eggs remain viable and susceptible to nematode suppression even when soybean is not present (such as when the field has rotated from soybean to corn). Soybean cyst nematodes, like RL nematodes, can also greatly reduce crop yield; SCN decreased soybean yield in the U.S. more than any other pathogen from 1996 to 2007. In 2007, SCN infection led to a reduction in soybean yield of 94 million bushels (Wrather and Koenning, 2009).
Currently, no published papers investigate the biofumigant effect of Brassicaceous cover crops on soybean cyst or root lesion nematodes specifically. A paper by Gruver et al. (2010) examined the effect of radish on free-living nematode community composition, but the results were inconclusive. The authors state that while radish had unique impacts on nematode communities, these impacts appeared to be associated more with quality of organic matter inputs rather than biofumigation. Thus, there is enough evidence to support further research of radish as a cover crop in Wisconsin.
Some research has been conducted in the North Central region, including the first two years of this project. The beginning of this project focused solely on the uptake and release of N in the cropping system. Moving forward into the third year, it has become evident that important information about radish as a cover crop is missing, and this project hopes to fill in some of these gaps. Regardless, the bulk of the research on radish as a cover crop has been performed in the Mid-Atlantic region. While the results have demonstrated significant benefits of using radish as a cover crop, these data are specific to the Mid-Atlantic region, their management practices, and their soil. A greater range of research is needed in the North Central region on our soils under our management practices in order to provide more conclusive results. This expanded research would further inform North Central farmers and land managers on how to protect their soil and water in the future.
There is a great need for further research on cover crops in Wisconsin in particular for soil and water conservation. In some cropping systems, there is a window of opportunity for a cover crop to be planted from mid-to-late summer to early fall. Radish as a cover crop would fit into these summer time slots well, providing soil and nutrient conservation benefits. For example, in these cropping systems, the soil is typically left bare after the winter wheat is harvested. Planting radish in this time slot would therefore help prevent soil erosion. Also, even though it is important to note that radish will most likely not establish as quickly as a rye (a popular cover crop), radish does produce a rather dense canopy. This canopy reduces raindrop impact on the soil surface. But in order for radish cover crops to be utilized successfully, they need to be fully researched to understand their proper use across many different conditions. There is clearly not enough data on radish as a cover crop to make recommendations for use, and this puts the agronomic community at a disadvantage. There is a need for smart cover crop use here in the Midwest as many growers are participating in NRCS conservation programs.
Objective #1. Determine the potential nitrogen credit from radish as a cover crop.
Objective #2. Discern the effect of radish on soil resistance.
Objective #3. Determine the effect of radish on nematode populations.
The experiment was conducted at three different locations in Wisconsin. Three of the fields were located at Rock County Farm in Janesville, WI within 1 km of each other. The Rock County Farm had a long history of no-tillage management practices, 5 years or more, depending on the field. The experimental design was a randomized complete block, split plot with four replications. Each block contained three whole plot treatments: (i) no cover crop (NR), (ii) radish with no N added at planting (RAD), and (iii) radish with 67 kg N ha-1 added at planting (RAD+67). The whole plot dimensions were 3 × 46 m. Each whole plot contained six subplots, which were N rates of 0, 45, 90, 134, 179, and 224 kg N ha-1. The subplot dimensions were 3 m × 8 m. This site was used for all three objectives. Two of the fields were located at a farm in Sheboygan Falls, WI within 1 km of each other. Both fields were chisel-plowed in the fall before radish planting in the no cover crop whole plot treatments only. In the spring, before corn planting, the same no cover crop whole plot treatments were subjected to one trip with a field cultivator. The experimental design was a randomized complete block, split plot with four replications. Each block contained two whole plot treatments: no cover crop (NR) and radish (RAD). The whole plot dimensions were 9 × 49 m. Each whole plot contained eight subplots, which were N rates of 0, 34, 67, 101, 134, 168, 202, and 235 kg N ha-1. The subplot dimensions were 5 × 12 m. This site was used for Objective #3. One of the fields was located at a farm in West Bend, WI. The field used in this experiment was under no-tillage management. The experimental design was a randomized complete block, split plot with four replications. Each block contained two whole plot treatments: no cover crop (NR) and radish (RAD). The whole plot dimensions were 9 × 27 m. Each whole plot contained six subplots, which were N rates of 0, 45, 90, 134, 179, and 224 kg N ha-1. The subplot dimensions were 5 × 9 m. This site was used for Objective #3. The crop rotation for the experiment included one year of corn, followed by one year of soybean, and then followed by one year planted first with winter wheat and then radish following winter wheat harvest.
At Janesville, WI, radish (Tillage Radish®, Cover Crop Solutions LLC) was planted on 15 August 2011, 30 August 2012, and 28 August 2013 following the harvest of winter wheat using a no-till drill with 17.8 cm row spacing, with a seeding rate of 11 kg ha-1. On the planting date, 67 kg N ha-1 as ammonium nitrate was broadcast surface applied to RAD+67. Volunteer winter wheat and summer annual weeds were controlled before planting with a broadcast application of glyphosate (210 g a.i. ha-1). The radish was frost-killed in the winter when air temperature dropped below -4°C for several nights in a row. Four rows (76 cm spacing) of corn (O’Brien hybrid 1155) were planted per plot with a no-till planter at a seeding rate of 79,000 seeds ha-1 on 15 May 2012, 16 May 2013, and 30 May 2014. Fertilizer was applied at planting as ammonium nitrate hand applied in row in bands for the subplot N rates. Post emergence herbicides were applied as needed for weed control.
At Sheboygan Falls, WI, radish (Tillage Radish®, Cover Crop Solutions LLC) was planted on 15 August 2012 and 12 August 2013, following the harvest of winter wheat using a no-till drill with 17.8 cm row spacing, with a seeding rate of 11 kg ha-1. Before radish planting, 45 kg of urea (21 kg of N) was applied across the whole field, including all radish and no radish plots. The radish was frost-killed in the winter when air temperature dropped below -4°C for several nights in a row. Six rows (76 cm spacing) of corn were planted per plot with a no-till planter at a seeding rate of 79,000 seeds ha-1 on 12 June 2013 and 28 May 2014. Fertilizer was broadcast applied at planting as urea with Agrotain (Koch Agronomic Services, LLC, Wichita, KS). Post emergence herbicides were applied as needed for weed control.
At West Bend, WI, radish (Tillage Radish®, Cover Crop Solutions LLC) was planted on 8 August 2013, following the harvest of winter wheat using a no-till drill with 17.8 cm row spacing, with a seeding rate of 11 kg ha-1. Before radish planting, 45 kg N ha-1was applied to the whole field, including all no radish and radish plots. The radish was frost-killed in the winter when air temperature dropped below -4°C for several nights in a row. Six rows (76 cm spacing) of corn (Great Lakes hybrid 4875) were planted per plot with a no-till planter at a seeding rate of 79,000 seeds ha-1 on 24 May 2014. Pop-up and starter fertilizer were applied at planting. Fertilizer was broadcast applied when the corn was six inches tall as urea with Agrotain (Koch Agronomic Services, LLC, Wichita, KS). Post emergence herbicides were applied as needed for weed control.
Biomass from radish and volunteer winter wheat was sampled before a hard freeze in the fall, usually in mid to late November, from all three field sites. Samples were collected in a 0.5 m2 area at one location within each whole plot. Aboveground biomass (AGB), including all the leaves and stems was collected for both radish and volunteer winter wheat, while belowground biomass (BGB), including the main taproot but no other roots, was collected for radish only. Radish biomass was collected and analyzed as whole plants in 2011 and collected and analyzed as aboveground and belowground parts in 2012 and 2013. All samples were dried in a forced air oven at 60°C; radish belowground biomass was rinsed with water to remove soil prior to drying. Volunteer winter wheat and radish plant total N and total carbon were determined by dry combustion using a Flash EA 1112 CN Automatic Elemental Analyzer (Thermo Finnigan, Milan, Italy).
At Janesville, WI, soil samples (with a diameter of 2.2 cm) for nitrate analysis were collected at two depths (0-30, 30-60 cm) in the fall, both at radish planting and at the same time as biomass collection. At planting, six soil cores were collected and combined into a composite sample in each block. At biomass collection, five soil cores were collected in each whole plot and combined into a composite sample. Soil samples were collected at two depths (0-30, 30-60 cm) prior to corn planting in the spring. Eight soil cores were collected within each whole plot and combined into a composite sample. In 2013 and 2014, soil samples were also collected (0-30 cm) in the 0N subplots (four cores per subplot, combined into a composite sample) at the following four corn physiological stages; V4, V10, tassel, and harvest (Ritchie et al., 1989). Nitrogen credits for corn were calculated according to the University of Wisconsin-Extension publication Nutrient Application Guidelines for Field, Vegetable, and Fruit Crops in Wisconsin (A2809; Laboski and Peters, 2012). Analysis of soil NO3–N concentration was conducted by the University of Wisconsin-Madison Soil and Plant Analysis Laboratory. Soil NO3–N concentration of soil samples were determined after a 2N KCl extraction with colorimetric analysis using flow injection (Lachat Quikchem 8000; Lachat Instruments, Milwaukee, WI).
At Janesville, WI, soil penetration resistance measurements were collected to a depth of 45 cm using a portable cone penetrometer (FieldScout SC 900 Soil Compaction Meter; Spectrum Technologies, Inc., Aurora, IL). Soil moisture was measured to a depth of 20 cm at the same time that penetrometer readings were taken with a portable volumetric soil water content meter (FieldScout TDR 300; Spectrum Technologies, Inc., Aurora, IL). At corn planting, 12 penetrometer (two side-by-side measurements collected at six locations) and 18 soil moisture (three side-by-side measurements collected at six locations) measurements were collected in each whole plot. During the corn growing season in 2013 and 2014, soil penetrometer and soil moisture measurements were collected at two corn physiological stages: V4 and tassel, and only in the 179 kg N ha-1 subplot treatment. Although 12 penetrometer measurements were still collected in the subplot, three side-by-side measurements were collected at four locations. Soil moisture data collection remained the same.
Analysis of variance was conducted to determine the effect of cover crop treatment, aboveground and belowground plant parts, and their interaction on total N uptake and C:N ratio, using Proc Mixed in SAS (SAS Institute Inc., Cary, NC) with block as a random effect, and analyzed separately for each year. Analysis of variance was conducted to determine the effect of cover crop treatment on soil NO3–N concentration and soil resistance, using Proc Mixed in SAS with block as a random effect, and analyzed separately at each depth for each year. All the effects were analyzed separately for each sampling date. Analysis of variance was conducted to determine the effect of cover crop treatment, N rate treatment, and their interaction on corn grain yield using Proc Mixed in SAS with block as a random effect, and analyzed separately for each year. Linear regression was conducted to characterize corn grain yield response to N using PROC REG in SAS. Four models were run separately for each treatment each year: linear, quadratic, linear plateau, and quadratic plateau. Ultimately, only one model was reported per treatment based on best fit as determined by R2 values. The Tukey–Kramer adjusted P value was used to determine the differences in treatment means at α = 0.10 significance level.
Soil samples (with a diameter of 2.2 cm) for nematode analysis were collected (0-30 cm) at corn planting as well as at the following four corn physiological stages; V4, V10, tassel, and harvest during 2013 and 2014 from all three field sites (Ritchie et al., 1989). Samples were collected in each whole plot treatment at planting; eight soil cores combined into one composite sample. After planting, samples were collected in both the 0N and the 168N or 179N subplot treatments; four soil cores combined into one composite sample. The soil samples were collected in pairs; one sample near a relic radish hole and the other 15 cm away in-between the radish rows. In the NR plots, the same spacing between the paired samples was used, except the locations were random. Samples were transported to the laboratory in coolers and kept at 4 °C until nematode extraction. Nematode soil samples, sealed in plastic bags, were gently mixed before a 100 cm3 subsample was removed. RL nematodes were extracted from the soil using the Baermann funnel incubation technique (Jenkins, 1964). Both RL nematodes and SCN were extracted from the soil using a sieving-centrifugation method with stacked 60 and 400 mesh sieves (0.250 mm and 0.038 mm, respectively). Only RL nematodes and SCN were counted because of their well-documented pest status in Wisconsin. The juvenile and adult nematodes were identified and counted using a stereomicroscope. Each sample was counted in its entirety for both RL nematodes and SCN. Total nematode counts were the sum of the funnel and centrifugation samples.
Analysis of variance was conducted to determine the effect of cover crop, N rate, and their interaction on SCN and RL nematode counts using Proc GLIMMIX in SAS (SAS Institute Inc., Cary, NC), analyzed separately for each year. The random effects were block and the interaction effect between block and cover crop treatment. All the effects were analyzed separately for each sampling date. The Tukey–Kramer adjusted P value was used to determine the differences in treatment means at α = 0.10 significance level.
In 2011, there was no significant difference in N uptake between the radish treatments (Table 1). In 2012, all treatments exhibited lower N uptake than in 2011. However, the radish treatments did show a statistically significant difference; RAD+67 had greater N uptake than RAD. The N uptake in 2013 was even less than that observed in 2012; however, RAD+67 was again significantly greater than RAD. C:N ratio was significantly different between the radish treatments in 2012 only; RAD had a larger C:N ratio than RAD+67.
The lack of significant differences in N uptake between radish cover crop treatments in 2011 indicated that in ideal growing conditions extra N may not be necessary for good establishment and growth of radish. However, in both 2012 and 2013, the growing conditions were far from ideal. The N uptakes were lower in 2012 than in 2011, most likely due to drought conditions throughout the summer and fall of 2012. In 2013, N uptakes were even lower than those seen in 2012, as a result of cool, dry growing conditions. In both 2012 and 2013, RAD+67 did have significantly more N uptake than RAD, suggesting that extra N may be beneficial for establishment and growth in dry, stressed environmental conditions. Thorup-Kristensen (1993) found that N uptake by cover crops increased with increasing amounts of soil mineral N which supports N uptake results from 2012 and 2013. In multiple studies, the N uptake of radish was found to be significantly larger than the no cover crop treatments as well as other cover crops (Dean and Weil, 2009; Kristensen and Thorup-Kristensen, 2004; O’Reilly et al, 2012), ranging from 119 kg N ha-1 to 158 kg N ha-1 to 240 kg N ha-1, respectively. All of these N uptake values are similar to the N uptake values measured in 2011.
In 2012 and 2013, radishes harvested for biomass were separated into two parts: aboveground biomass (AGB) and belowground biomass (BGB). These two parts were analyzed separately for N uptake and C:N ratio (Table 2). In 2012, the AGB and BGB N uptakes were significantly different from each other when compared between AGB in RAD+67 and BGB in RAD, but not between AGB in RAD and BGB in RAD+67. In 2013, the AGB N uptake in RAD+67 was significantly greater than the rest of the treatments. AGB N uptake was significantly greater than the BGB N uptake for RAD+67. The AGB and BGB N uptakes were significantly different from each other when compared between AGB in RAD+67 and BGB in RAD. In 2012, the C:N ratio of BGB in RAD was significantly greater than the rest of the treatments. In 2013, the C:N ratios of BGB in both treatments were significantly greater than the C:N ratios of AGB in both treatments.
It is possible that the significant differences in radish N uptake between AGB and BGB were not an effect of a difference between AGB and BGB composition, but rather an effect of the growing season conditions. In both 2012 and 2013, RAD+67 had a significantly higher N uptake than RAD as a whole plant, which was reflected by significantly higher N uptakes in the ABG and BGB of RAD+67 compared to RAD. The lack of significant differences between AGB and BGB in either radish treatment suggests that neither radish part grew better than the other nor was better at N storage.
Radish differed in C:N ratio when compared between AGB and BGB treatments (when treatments were analyzed as a whole plant, these differences were concealed, except for 2012). This indicated that C:N ratios of radish AGB and BGB were inherently different from each other. The AGB had a lower C:N ratio than the BGB in each respective radish treatment (except for RAD+67 in 2012). A similar trend was reported by Clark (2007); AGB had a C:N ratio from 10-20, while BGB had a C:N ratio from 20-30. It is important to note that differences in this study were not due to a difference in plant variety since the same radish seed was used each year. Also, all of the C:N ratios reported were low, indicating that decomposition of radish material and N release would occur relatively quickly.
Soil NO3-N concentrations were analyzed in November of each year of the radish growing season. Only one depth at one November sampling date showed significant differences among the cover crop treatments (Table 3). At 0-30 cm in November of 2012, RAD had a significantly greater soil NO3-N concentration than NR, but RAD+67 was not significantly different from the other two treatments.
Pre-plant nitrate test (PPNT) soil NO3-N concentrations among cover crop treatments only differed significantly in 2012 (Table 4). In April of 2012, RAD+67 had significantly greater soil NO3-N concentrations at both sampling depths. However, at 0-30 cm, RAD+67 was only significantly different from NR (RAD was not significantly different from either treatment), while RAD+67 was significantly different from both of the other treatments at 30-60 cm. In May 2012, RAD+67 and RAD had significantly greater soil NO3-N concentrations than NR at 0-30 cm. At 30-60 cm, RAD+67 was significantly greater than RAD, and RAD was significantly greater than NR. Nitrogen credits were calculated based on the PPNT values; only April and May of 2012 had large enough PPNT values to determine a N credit (Table 4). RAD+67 had the largest N credit in both April and May. NR had the smallest N credit for both dates, with RAD falling in-between the other two treatments. It is important to note that as the corn growing season progressed, differences in N credits between the radish treatments increased. The N credit almost doubled for both radish treatments from April to May, while the N credit for NR stayed almost the same.
In 2013 and 2014, no significant differences were seen among the cover crop treatments for soil NO3-N concentration during the corn growing season (Table 5). N credits were calculated based on the pre-sidedress nitrate test (PSNT) values; only the radish treatments in June of 2013 had large enough PSNT values to determine a N credit. Both radish treatments had N credits of 10 kg ha-1.
Large radish biomass production and N uptake resulted in greater soil NO3-N concentrations in the spring. In the spring of 2012, soil NO3-N concentrations increased from April to May in both radish treatments, but not in NR. This indicated that these increases in soil NO3-N concentrations could be due to contributions from decomposing radish material; soil NO3-N concentrations in NR did not increase because there was no radish to provide extra N. This result is confirmed by Dean and Weil (2009) who found that N was released into the surface soil in the spring as radish decomposed, as well as O’Reilly et al (2012) who saw higher plant available N in radish than no cover crop. It appears as though April was still too early in the season for full radish decomposition, as seen by the increase in soil NO3-N concentrations from April to May in 2012. Another interesting point was the fact that RAD+67 had higher soil NO3-N concentrations than RAD at the second depth (30-60 cm) in spring of 2012 even though the radish treatments took up the same amount of N in the previous fall. The extra soil NO3-N in the RAD+67 treatment could be from the extra N applied at radish planting. None of these trends were observed during in the spring of 2013 or 2014. The radishes in these experiments were too small to take up enough N in the fall to affect soil NO3-N concentrations in the following spring. The radishes in 2011 were large enough to release enough N in the spring to calculate potential N credits that can be applied to the corn crop.
In 2012, all of the effects used to analyze corn yield were not significant (Table 6). It is important to note that corn smut (Ustilago maydis) was observed in the field in 2012, and may have damaged the corn enough to contribute to the lack of significant differences. In 2013, both cover crop and N rate effects were statistically significant, but their interaction effect was not. RAD+67 had a significantly greater yield than RAD, but not NR. RAD and NR were also not significantly different in yield.
The corn grain yields in 2012 did not follow a typical N response curve (Figure 1). However, the corn grain did exhibit a response to N in 2013, and regression analysis was performed (Figure 2). The models for each treatment were chosen based on the highest R2 values. A plateau model would only be used if its R2 value was at least 0.02 greater than the linear or quadratic R2 values. A quadratic model would only be used if its R2 value was at least 0.01 greater than the linear R2 value. NR was best described with a quadratic model (R2 = 0.978), while both RAD (R2 = 0.960) and RAD+67 (R2 = 0.980) were best described with a linear model. Equations were reported in Figure 2. The response curves were different for all three treatments. The difference between the linear equation used for RAD and the linear equation used for RAD+67 further emphasizes a potential yield gain by RAD+67.
Radish did not influence corn response to N in either 2012 or 2013. The corn growing season in 2012 was hot and dry, but soil tests showed that NO3-N concentrations were high in all treatments and substantial N credits were calculated. Therefore, since corn yields were high, there was enough heat and soil N to facilitate corn growth without demonstrating a response to different applied fertilizer N rates. Also, even though there were differences in potential N credits among the treatments, corn yields failed to reflect or confirm any of these differences. In 2013, the significantly greater yield of RAD+67 over RAD indicated that the extra N applied at radish planting may have had an effect on yield under ideal growing conditions. The 2012 results contradict other studies, but the 2013 results are almost confirmed. Both soybean and corn were found to have significantly greater yields after radish than after no cover crop (Williams and Weil, 2004; Chen and Weil, 2011). Williams and Weil (2004) stated that soybean roots were able to find subsoil water easier in soil following radish because the radish had left behind sizable root channels. In 2012, there may not have been enough soil water to make a difference. In 2013, it is possible that only the radish in RAD+67 were large enough to create sufficiently sized root channels for the subsequent corn crop to find water.
All three treatments exhibited the same pattern of soil resistance in June 2013 (Figure 3). At 5 cm and 7.5 cm, RAD was statistically different from NR, while RAD+67 was not significantly different from either. At 10 cm, RAD+67 was different from RAD, but NR was not different from either RAD or RAD+67. At 32.5 cm, NR and RAD+67 were different from each other but neither was different from RAD. The average moisture for the soil, measured as volumetric water content (VWC), in June 2013 was 37%. The differences between treatments were less than 2% VWC; therefore, differences in cone index (penetration resistance) between treatments can be attributed to differences in bulk density.
In August 2013, however, the pattern of resistance among treatments did separate from each other deeper in the soil profile (Figure 4). It is important to note that although it appears as though there was a difference in soil resistance values between the two time points in 2013, this was mostly due to differences in soil moisture, not soil compaction. The soil in August was much drier than in June, and this naturally increased soil resistance. The average moisture for the soil in August was 24.6% (differences between treatments were less than 2% VWC). At 17.5 cm, NR was statistically different than RAD+67, but neither were different from RAD. From 20 cm through 30 cm, NR was different from both radish treatments; however, the radish treatments did not differ from each other.
In June 2014, soil resistance values were similar to values seen in June 2013 (Figure 5). At 0 cm, NR was statistically different from RAD, but neither were different from RAD+67. From 2.5 cm through 10 cm, both of the radish treatments were different from NR, but not from each other. At 12.5 cm and 15 cm, NR and RAD were again different from each other, but neither was different from RAD+67. At 17.5 cm, RAD and RAD were different from each other, but neither was different from NR. The average moisture for the soil in June 2014 was 52.5% (differences between treatments were less than 2% VWC).
The overall soil resistance values were higher in July 2014 than in June 2014, largely due to a decrease in soil moisture (Figure 6). The average moisture for the soil in July 2014 was 45.3% (differences between treatments were less than 3.5% VWC). At 2.5 cm, both NR and RAD were not significantly different from each other, but both were different from RAD+67. From 5 cm through 10 cm, RAD was significantly different from both NR and RAD+67, but these were not different from each other. From 12.5 cm through 35 cm, both of the radish treatments were not significantly different from each other, but both were different from NR. At 37.5cm, RAD+67 was significantly different from NR, but RAD was not significantly different from either.
The soil resistance values increased again as the growing season continued; the values in August 2014 are larger than the values seen in July 2014 (Figure 7). At 0 cm, and from 7.5 cm through 12.5 cm, RAD differed significantly from RAD+67, but neither differed from NR. At 2.5 cm, 5 cm, 17.5 cm, 20 cm, and from 25 cm through 30 cm, RAD was different from NR and RAD+67, but these were not different from each other. At 15 cm, 22.5 cm, and from 32.5 cm through 40 cm, NR and RAD were statistically different from each other, but neither were different from RAD+67. The average soil moisture across all treatments in August 2014 was 44.7% (differences between treatments was less than 2% VWC).
The effect of radish on soil penetration resistance was not demonstrated until late in the corn growing season. Early in the spring of both years (2013 and 2014), there were few differences in soil resistance between treatments and mostly in the upper 12.5 cm of the soil profile. In June of 2013, NR had a higher soil resistance than the radish treatments in the first 7.5 cm of the profile, indicating that the radish taproots were able to decrease soil resistance at these depths. In June 2014, the opposite was true; both radish treatments had higher soil resistance than NR. Later in the growing season in both 2013 and 2014, differences between treatments were more apparent and occurred deeper in the soil profile. The radish treatments in August 2013 had a higher soil resistance than NR deeper in the soil profile. It is possible that the radish pushed the soil outward as it grew, thereby compacting the nearby soil. However, in July of 2014, NR had a significantly higher soil resistance than both radish treatments for the majority of the profile. In August of 2014, NR and RAD+67 had a significantly higher soil resistance than RAD for most of the soil profile. Therefore, even though RAD+67 had a higher soil resistance than RAD, the radish taproot appears to affect soil resistance and soil penetrometer readings deeper in the soil profile. Chen and Weil (2011) found that under both medium and high soil compaction, corn planted after radish had more roots at almost all depths (0-45 cm) than no cover crop, indicating that radish was better able to penetrate consistently compacted soils. However, the only time point in our study that showed root-restrictive penetration resistance, or compaction, classified at > 2000 kPa (Bengough and Mullins, 1990), was August 2013 due to drier soil conditions. Therefore, it is possible that the fields used in this study were not compacted enough to show consistent differences among the treatments. Also, while the bulk densities of the three compaction treatments used in Chen and Weil (2011) were significantly different from each other, the penetrometer was not as adept at identifying significant differences between the compaction treatments. Their results suggest that the penetrometer was not as sensitive to changes in the soil. This viewpoint is actually supported by Weil and Kremen (2007) and Chen and Weil (2011), which suggested that the penetrometer tip was too large to detect the small root channels left behind by radish. Chen and Weil (2011) also suggested that planting cover crops for at least four years would be necessary to see the effects of radish roots on soil resistance.
Radish total biomass was the lowest at Site 1 in 2012; the RAD treatment had 1718 kg ha-1, while the RAD+67 treatment had 2597 kg ha-1 (Table 7). Site 2 in 2013 had the next greatest amount of radish total biomass at 5433 kg ha-1. Site 3 in 2013 had a radish total biomass of 6239 kg ha-1, followed closely by Site 2 in 2012 with a radish total biomass of 6855 kg ha-1.
Even though the sites demonstrated a range of radish biomass, larger amounts of biomass did not result in more nematode suppression. In fact, three out of four of the significant differences among treatments for RL nematode counts demonstrated that RAD had more nematodes than NR, even for Sites 2 and 3 which had the largest amounts of radish biomass. Wang et al. (2009) found that a radish biomass of 5920 kg ha-1 was sufficient to reduce plant-parasitic nematode populations by 55.7% when compared to the no cover crop control. While two of the sites in our experiment exhibited biomass values above this value (Table 1), no nematode suppression was determined. However, despite the conclusive results seen in Wang’s study, the bulk of the literature indicates that the amount of brassicaceous biomass necessary for nematode suppression is unclear. Matthiessen and Kirkegaard (2006) wrote a review that suggested that brassicaceous biomass values between 3,000 and 17,000 kg ha-1 (B. napus and B. juncea) were successful at suppressing nematodes compared to the control. But other studies in the same review, with the same amount of brassicaceous biomass, did not report suppression. Therefore, it is not clear if the lack of nematode suppression seen in our study was caused by insufficient radish biomass production.
At Site 1 in 2013, RL nematode counts were similar at all of the sampling dates at the 0N rate except for the low number observed at the last sampling date in October (Figure 8). For Site 1 at the higher N rate, July 2013 had low RL nematode numbers. In 2013 at Site 2, all of the sampling dates at the 0N rate had similar RL nematode counts except for June, which was very low. In 2014, RL nematode counts were very low at all sampling dates for Site 2 at both N rates (Figure 9).
Early in the growing season at Site 1 in May 2013, both NR and RAD+67 had significantly greater amounts of RL nematodes than RAD; however, NR and RAD+67 were not significantly different from each other (Figure 10). In July 2013, the N rates were statistically different but the treatments within a N rate were not (Figure 11). Each treatment in the 0N rate had significantly more nematodes than its corresponding treatment in the 179N rate except for RAD+67 which showed significantly less nematodes in the 0N rate. Later in the growing season, in August 2013, differences among treatments were observed in the 0N rate; RAD had a significantly greater amount of nematodes than NR, but not RAD+67 (Figure 12). Differences between N rates were also seen. NR and RAD+67 in the 0N rate had significantly less nematodes than NR and RAD+67 in the 179N rate; however, RAD was not different between N rates. A similar pattern was detected in October 2013, except NR was the treatment that was not significantly different between N rates (Figure 13). RAD and RAD+67 in the 0N rate had significantly less nematodes than RAD and RAD+67 in the 179N rate. The lowest numbers of nematodes were seen in May 2013 and October 2013; July 2013 had higher numbers than both May and October, while August had the highest numbers of all sampling dates.
At Site 2 in 2013, RAD had significantly more nematodes than NR early in the growing season in May (Figure 14). There were no differences between treatments for root lesion nematodes in June of 2013 (Figure 15). In July 2013, there were no differences between treatments at the same N rate, but both NR and RAD showed differences between the N rates (Figure 16). NR at the 0N rate had significantly less nematodes than NR at the 168N rate. RAD at the 0N rate had significantly more nematodes than RAD at the 168N Rate. No differences were observed in August 2013 (Figure 17). The order of sampling dates in increasing number of nematodes was as follows; June 2013, August 2013, May 2013, and July 2013.
In 2014, the numbers of root lesion nematodes at Site 2 were very low (Figures 18, 19, and 20). Several of the nematode counts for a given treatment were zero. July 2014 showed the highest numbers of nematodes at around 6 nematodes in 100 cm3 of soil (Figure 19). No statistically significant differences were seen at any sampling date.
At Site 3 in 2014, RAD had significantly more nematodes than NR in June (Figure 21). There were no significant differences seen in July 2014 (Figure 22). In September 2014, statistically significant differences were observed between treatments at different N rates (Figure 23). Both NR and RAD in the 0N rate had significantly less nematodes than their counterparts in the 179N rate. Out of the three sampling dates, June 2014 had the highest number of nematodes, followed by September 2014 and July 2014.
The numbers of soybean cyst nematodes at both sites in 2013 were very low (Table 8). Only two sampling dates at Site 2 had numbers higher than 1, 8 May 2013 and 14 June 2013. In May 2013, NR had an average SCN count of 2.25, and RAD had an average SCN count of 1.75. In June 2013, both NR and RAD had an average SCN count of 1.25. No soybean cyst nematodes were observed at either site at any time in 2014 (Table 9).
Nematode suppression was difficult to detect due to low nematode populations. The fields used in these experiments were not infested with SCN, as exemplified by the SCN counts. After counting the first batch of soil nematode samples, cysts were extracted from the same soil samples to establish if SCN were present in that form. However, the only cysts found in the soil samples were dried and empty. If the field had been infested in the past, it had a very low population or the SCN population had already been eliminated. With these results, it is unclear if nematode suppression will occur below critical levels. All of the samples collected at all of the sites in both years were well below the damage threshold for RL nematodes, set at 200 nematodes per 100 cm3 of soil for spring samples in WI (MacGuidwin, 2013). Therefore, none of these fields suffered damage due to RL nematode infestation. It might seem as though performing research on fields with low populations was not useful; however, since one of the best nematode management practices is prevention of population increase, this experiment was still helpful, especially prior to planting a crop favorable as a food source such as corn.
The timing of radish termination decreased the effectiveness of nematode suppression. At all of the sites in both years, the radish treatments were allowed to grow until the first frost- usually about 3 months of total growth. Therefore, the radish plots at all the field sites were at relatively the same mature growth stage when they winter-killed. It has been documented that glucosinolate concentrations are the lowest at maturity in brassicaceous plants (Malik et al., 2010; Matthiessen and Kirkegaard, 2006; Mojtahedi et al., 1991). Two studies observed that glucosinolate levels actually peaked at the onset of the flowering stage, and then decreased until maturity (Malik et al., 2010; Matthiessen and Kirkegaard, 2006). Rapeseed that was grown for 2 months and then incorporated into the soil resulted in the highest nematode suppression (Mojtahedi et al., 1991). Therefore, the lack of significant differences in RL nematodes among treatments may be due to the advanced stage of maturity of the radish.
Radish had significantly more RL nematodes than the control early in the corn growing season; however, this effect was not sustained throughout the rest of the season. At the first sampling date for both Site 2 in 2013 and Site 3, RAD had significantly more nematodes than NR. It is possible that radish encouraged nematode population increases during its growing season (since radish is a host for RL nematodes) which were still seen at corn planting (Matthiessen and Kirkegaard, 2006). At the first sampling date for Site 1, RAD+67 had significantly more nematodes than RAD, but not NR, which indicates that there might not have been more nematode growth overall, but that larger radish (due to addition of N at radish planting) could have encouraged more nematode growth. The fact that radish had significantly higher nematode counts at the start of the season, but not at the end, indicates that other factors affected nematode populations and suppression since yield responses are most strongly related to initial nematode populations.
The lack of tillage in the radish treatments may have decreased the effectiveness of nematode suppression. Two field studies have shown that ITC concentrations increased rapidly in the soil immediately after brassicaceous tissue disruption by tillage (Morra and Kirkegaard, 2002; Gardiner et al., 1999). According to Morra and Kirkegaard (2002), maximum ITC release into the soil was from 2 to 24 h after incorporation using tillage. Gardiner et al. (1999) found that maximum ITC concentrations were observed in the soil 30 h after the material was plowed. Morra and Kirkegaard (2002), also found that ITC concentrations decreased rapidly, reaching a minimum 48 h after incorporation. Therefore, nematode populations in radish treatments should decrease early in the spring after radish decomposition, but this trend was not shown by our data. In our experiment, radish was left to decompose without any incorporation or tillage. Therefore, only rainfall was responsible for moving ITCs into the soil, which indicates that ITCs may not have been integrating into the soil effectively or even lost to volatilization. Both temperature and rainfall affect volatilization; higher temperatures result in more volatilization (Zasada and Ferris, 2004), while more rainfall results in less volatilization (Brown et al. 1991). As a result, high rainfall and low temperatures around the time of radish decomposition, such as in 2013 and 2014, should decrease volatilization and increase nematode suppression, but our data showed the opposite. Several other factors also affect volatilization; increases in clay, organic matter, and pH also result in less volatilization (Brown et al. 1991). Several studies have also shown that ITC can be lost to the soil through sorption (Brown et al. 1991; Matthiessen and Kirkegaard, 2006; Matthiessen and Shackleton 2005; Gardiner et al. 1999). It is possible that not incorporating brassicaceous material into the soil greatly reduces both the amount and efficacy of ITCs.
It is also important to note that radish did not increase nematode populations. While radish did have significantly more RL nematodes than the control early in the corn growing season, this effect was not seen during the rest of the season. It has been well documented in the literature that while radish contains nematode suppressive properties, radish can also serve as a host for nematodes and increase nematode populations (Zasada et al., 2010; Matthiessen and Kirkegaard, 2006; Melakeberhan, et al. 2008; Monfort et al., 2007). Melakeberhan, et al. (2008) found that all Meloidogyne hapla (plant-parasitic nematodes classified as a root-knot nematode) completed their life cycle on radish. Monfort et al. (2007) determined that Meloidogyne incognita (root-knot nematode) was able to reproduce on radish during its growing period. The same authors observed that radish significantly reduced nematode populations from cover crop harvest to planting of the subsequent crop compared to the other brassicaceous cover crop treatments; however, radish still had the greatest net increase in nematode population because of the population increase during radish growth. Corn is also a host for root lesion nematodes, which had a significant effect on our nematode counts. Throughout the corn growing season for Site 1, RAD+67 had significantly more nematodes in the 179N rate than the 0N rate. Late in the season for Site 3, both NR and RAD had significantly more nematodes in the 179N rate than the 0N rate. This suggests that the larger corn at the higher N rate encouraged more RL nematode reproduction because it served as a food source. Therefore, it is noteworthy that nematode populations did not increase at these field sites.
Educational & Outreach Activities
Data from this project was presented at several field days throughout the duration of the project. In October 2013, a field day was organized by the Sheboygan County UW-Extension County Agents. All of the current data from the Sheboygan County site was presented at this field day, which consisted of a tour of several farms in the area, allowing local growers to talk about their decisions that led them to use cover crops and their personal experiences with them. Another UW-Extension Cover Crop Meeting was held in March 2014 to present current data from both of the sites at Sheboygan and Washington counties. In August 2014, an Organic Agriculture Field Day at the Arlington Research Station was organized; data from this project was presented alongside other cover crop projects. In September 2014, the Michael Fields Agricultural Institute hosted the 2014 Cover Crops Field Day which included data from the Rock County field site. The audience at these events consisted of growers, crop consultants, and government agency professionals.
This project was also represented at several conferences. In 2013, a poster was presented at the International Soil Science Society of America Conference; the following year, a scientific talk was created for the same conference. Posters were also exhibited at the MOSES Conference in February 2014, the APS North Central Division Meeting in June 2014, and the NC Extension-Industry Soil Fertility Conference in November 2014. Another scientific talk was presented at the SWCS Annual Conference in July 2014. The audience at these events generally consisted of fellow university researchers, crop consultants, and government professionals.
A few publications, both completed and in progress, are emerging from this project. A completed master’s thesis was written based on the data from this project. Two peer-reviewed publications are in progress, intended for publication in a journal with an agronomic focus (e.g. Agronomy Journal or Nutrient Cycling in Agroecosystems). At least one peer-reviewed extension publication is also in progress, and will detail the recommended agronomic usage and effects of radish as a cover crop.
Thus far, we have significantly increased our knowledge of how a radish cover crop benefits cropping systems. The project has produced data on the amount of nitrogen taken up by radish, the effect that radish has on corn yield and corn yield response to N fertilizer, as well as the effect that radish has on pest nematodes and soil compaction. With these results, farmers can decide whether or not to utilize radish as a cover crop in their own system to fulfill their own personal needs. From this project, we have been able to create a more comprehensive technical protocol on how to incorporate radish into a cropping system as a cover crop which would greatly benefit both farmers, crop consultants, seed dealers, and the NRCS.
In the future, we expect to see an increase in the adoption of cover crops (radish or others) in Wisconsin cropping systems. We have proven that radish is capable of taking up a substantial amount of N, as well as release enough N in the following spring to result in a possible N credit. This will allow farmers to use less N fertilizer and increase nitrogen use efficiency in their systems. The alleviation of compaction reduces the need for tillage in conventional management systems, as well as benefits conservation and no tillage management systems.
According to the North Central Region SARE Cover Crop Report, 55% of participants planted Brassicas such as oilseed radish, mustards, rapeseed, and turnips as a cover crop in 2013-2014. In the following year, the same report documented that 61% of participants had planted a Brassica cover crop in 2014 and/or had plans to do so in the 2015 season. In general, these reports demonstrate that more farmers across the U.S. have adopted radish as a cover crop in the last three years.
Through field days and conferences, a large number of farmers have been educated about the use of radish as a cover crop in their cropping systems. At each field day, about 20 to 30 farmers were in attendance, and at each conference, anywhere from 1500 to 4000 farmers, crop consultants, agricultural industry professionals and educators were in attendance.
For day-to-day operations for farmers, we recommend planting radish as a cover crop in mid to late August at the latest in Wisconsin to achieve adequate biomass before the first frost. Also, we have found that dry and hot growing conditions may decrease radish biomass and nitrogen uptake, so extra nitrogen fertilizer applied to radish at planting may be necessary to maintain radish growth.
Areas needing additional study
It is important to continue the nitrogen (N) uptake research for more site years to gain a better understanding of how different environmental conditions affect radish growth. It would also be worthwhile to investigate possible mitigation techniques that farmers could employ to encourage radish growth, like additional fertilizer applications or irrigation.
The C:N ratios of radish indicate that decomposition should occur readily and quickly in the spring after radish planting. It is possible that decomposition happens too quickly or too soon in the season to synchronize with corn N needs. Therefore, it could be beneficial to use labeled N in the radish to better determine the timing of N release.
In order to confirm the N credits seen in 2012, this research needs to be continued for more site years. It would also be important to further investigate the role that N applied at radish planting has on corn yield, as seen in 2013, by applying increasing rates of N fertilizer to radish.
This research should be expanded onto fields with demonstrated compacted problems, or compaction treatments should be artificially applied, like those used by Chen and Weil (2010, 2011). Our experiment did not investigate the cumulative effects of radish over several rotations in the same field. Chen and Weil (2010) observed more cover crop roots after two years of cover crop establishment than after one year because more root channels had been created. Chen and Weil (2011) suggested that planting cover crops for at least four years would be necessary to see the effects of radish roots on soil resistance. Therefore, this study should be continued on the same field for at least four years, if not many more, since radish is planted only once every three years in this crop rotation.
It would also be beneficial to pair radish with legume cover crops, such as red clover or hairy vetch which have already been studied extensively in Wisconsin farming systems. Combining N-fixing legume cover crops with N-scavenging radish could greatly improve N uptake and release.
This study should be conducted again on fields that demonstrate root lesion nematode or soybean cyst nematode infestation, in order to examine the true nematode suppressive potential of radish. Studies should be performed to investigate the proper amount of biomass necessary for nematode suppression, as well as management strategies to achieve those biomass goals.