Introduction

Roots of cover crops provide several ecosystem services in arable soils. Griffiths et al. (2022) grouped these services into the main categories: soil structural remediation, soil resource capture, maintenance of the rhizosphere, and building soil organic matter content. To maximise the ecosystem services provided by roots, intercropping has been suggested as a potentially promising approach. Intercropping is defined as growing two or more different crops on the same field for at least some time during the growing period. Aboveground, intercropping has the potential to outperform sole crops due to overyielding (Wendling et al. 2017) and complementarity by niche differentiation (Justes et al. 2021) and to insure against crop failure by compensation (Döring and Elsalahy 2022). It is reasonable to assume that such mechanisms exist belowground as well. Thus, the agroecosystem services of roots might be enhanced in mixtures. However, in their review, Homulle et al. (2022) underlined that there is still a knowledge gap in root-root interactions in intercropping systems. For mixtures of cover crops, Lavergne et al. (2021) reported that there are inconsistent results of studies comparing mixtures with pure stands, and advocate more research on root traits of cover crop mixtures.

Root mass of cover crops is important for carbon storage (Kätterer et al. 2011; Poeplau and Don 2015) and root length is relevant for aggregate stability (Hudek et al. 2022) and nutrient uptake (Sainju et al. 1998). Thus, the cover crop mixtures were investigated for enhanced architectural root traits root length density (RLD) and root mass density (RMD) which are root length per soil volume and root mass per soil volume, respectively (Freschet et al. 2021). Besides, the vertical RMD profiles of sole and mixed cover crops need to be compared to test whether mixtures achieve higher soil exploitation. Moreover, we analysed the morphological root trait specific root length (SRL), which is root length per root mass and represents belowground economics (Freschet et al. 2021). SRL is also considered as an acquisitive root trait (Isaac and Borden 2019) that is discussed to be linked with nutrient uptake efficiency (Eissenstat 1992).

Belowground interactions have been well-studied for cereal-legume intercrops (Esnarriaga et al. 2020; Hauggaard-Nielsen et al. 2001; Hauggaard-Nielsen and Jensen 2005), relay-strip-intercropping systems (Li et al. 2006) and perennial cropping systems like clover-ryegrass mixtures (Dirks et al. 2021) or grasslands (Mommer et al. 2010; Oram et al. 2018). However, cover crops differ from relay-strip-intercrops in temporal niche differentiation and from main crop or perennial crops in the cropping duration. For cover crop mixtures, less research on roots has been conducted so far. Heuermann et al. (2019) attempted to quantify root biomass distribution of a four-species cover crop mixture and found root mass overyielding but less clear results for niche differentiation. Amsili and Kaye (2021) tried to compose a cover crop mixture with complementary root traits and complementary nitrogen acquisition strategies, but did not find more root carbon in the mixture compared to the best sole crop.

In a previous study (Kemper et al. 2020), we screened RLD and vertical root distribution of several cover crops. Based on these findings we suggested intercropping differently rooting cover crops to maximize root exploitation throughout the soil profile. In a further study (Kemper et al. 2022), we analysed root traits in cover crop mixtures of tap-rooted blue lupin (Lupinus angustifolius L.) and fibrous rooted winter rye (Secale cereale L.) and found weak evidence for shoot overyielding and root mass complementarity and no vertical niche differentiation of root mass density (RMD) between sole crops and mixtures, but positive mixture effects for specific root length (SRL).

Our cover crop root type screening (Kemper et al. 2020) revealed that oil radish (Raphanus sativus L. var. oleiformis Pers.) and winter rye had very different rooting patterns. Winter rye had a high RLD in topsoil and a shallow cumulative root distribution, whereas tap-rooted brassicas like oil radish had high subsoil RLD, deep cumulative root distribution and intensive use of biopores. From other studies on roots of cover crops, it is known that legumes with larger root diameters and low root length allocation to deep layers strongly differ from brassicas with highly branched and dense rooting systems (Bodner et al. 2019). Thus, we expected more intensive root complementarity when intercropping rye with oil radish than with lupin. Hence, in the present study, we investigated mixture effects on root traits of two cover crop mixtures with complementary root traits: deep-rooting tap-rooted oil radish and shallow-rooting fibrous rooted winter rye (OR-WR) as well as oil radish, the leguminous shallow-rooting crimson clover and winter rye (OR-CC-WR). Based on our findings of vertical root distribution screening (Kemper et al. 2020) and the lupin-rye mixture (Kemper et al. 2022) we aimed to test the following hypotheses:

  1. (1)

    In both mixtures, mixture effects of shoot biomass and nutrient uptake are positive.

  2. (2)

    Oil radish and mixtures containing oil radish have higher RLD and RMD in the subsoil compared to rye and crimson clover.

  3. (3)

    As a result of complementary root architecture, positive mixture effects for RLD and RMD occur in topsoil and subsoil.

  4. (4)

    SRL is increased in mixtures compared to sole cover crops.

  5. (5)

    Vertical root niche differentiation occurs for rye and oil radish as rye has higher RMD than expected from sole crops in topsoil, and oil radish has a higher RMD than expected from sole crops in subsoil.

Materials and methods

Experimental site

The field experiments were performed at the experimental farm for Organic Agriculture ‘Wiesengut’ of the University of Bonn in Hennef (50°48′ N, 7°17′ E), Germany. At this site mean (1981–2010) annual precipitation is 840 mm and the mean air temperature is 10.3 °C. Precipitation and mean temperature were 714 mm and 11.8 °C in 2018 and 814 mm and 11.4 °C in 2019. The soil of the experimental field is a Fluvisol with a silt loam texture on gravel layers with a soil thickness of 0.6 to 2.0 m. More details about climate and soil are given in a previous study (Kemper et al. 2020). Soil mineral nitrogen content in 0–90 cm at the time of sowing was 23.7 and 91.7 kg/ha in 2018 and 2019, respectively.

Experimental setup and management

Field trials were performed in a randomized complete block design with four replicates, three of which were investigated for root traits. The same three replications were used for the calculation of root-shoot-ratios. The size of the experimental plot was 1.5 m × 12.5 m in 2018 and 1.5 m × 20 m in 2019.

Sole cover crops of winter rye (WR, Secale cereale L., cv. Bonfire), crimson clover (CC, Trifolium incarnatum L., cv. Linkarus) and oil radish (OR, Raphanus sativus L. var. oleiformis Pers., cv. Siletina) were sown with a sowing density of 120 kg/ha, 30 kg/ha and 25 kg/ha, respectively. Cover crop mixtures of oil radish, crimson clover and winter rye were sown as mixed and substitutive intercrops (Hauggaard-Nielsen et al. 2008). There was a 50:50 mixture of oil radish and winter rye (OR-WR) and a 33:33:33 mixture of oil radish, crimson clover and winter rye (OR-CC-WR). Sowing of cover crops was performed with a plot seeder (Hege, Waldenburg, Germany) with a 17 cm row distance on 12th and 14th August in 2018 and 2019, respectively. Pre-crop was spring-sown faba bean (Vicia faba L.) in both years. Tillage after faba bean harvest was carried out with a mouldboard plough (23 cm deep) and a rotary harrow. The trial was irrigated to ensure crop establishment. More details on the pre-crop history of the experimental field, tillage, weeding, irrigation and soil moisture are given in Kemper et al. (2022).

Shoot biomass, nutrient content and soil mineral nitrogen

Shoot biomass samples of the cover crops were collected on 15th and 13th November 2018 and 2019, respectively. Samples were cut on 1.0 m² per plot, and dried at 105 °C until constant weight. After milling, samples were analysed for their nitrogen (N), phosphorous (P) and potassium (K) content.

On 8th and 9th November 2018 and 13th November 2019 soil samples were taken for analysis of soil mineral nitrogen (Nmin) using a Puerckhauer auger (inside diameter 18 mm). In 0–30 cm and 30–60 cm, soil samples were taken in both years and additionally in 60–90 cm in 2019. Eight samples per plot were mixed for the topsoil layers, and for the subsoil layers four samples per plot were mixed. For further information on milling and chemical analysis of plant and soil samples see Kemper et al. (2022).

Monolith method

To gain information about root mass and root length the monolith method by Böhm (1979) was used. From a trench, dug in the transverse direction to the plots (Fig. S1, supplementary material), monoliths were extracted in 10 cm intervals of soil depth. Monolith size was 10 cm × 10 cm × 20 cm in 2018 and 10 cm × 10 cm × 17 cm in 2019. These root extractions were done in October 2018 and 2019. Subsequently, roots were washed from the soil, separated manually from organic residues, scanned at 800 dpi (photo scanner, Expression 12000XL, Epson, Suwa, Japan) and analysed with WinRHIZO (Version WinRhizoPro 2019a 64bit, Regent Instruments, Quebec, Canada) for root length and dried for root dry mass determination. For further details on the washing process, sieve sizes, separation from non-root material, scanning and drying please consider the detailed method description in Kemper et al. (2022).

Fourier Transform Infrared Spectroscopy

The roots of OR, CC and WR collected in 2018 and 2019 were used to evaluate the root mass proportion of the different species in the cover crop mixtures. The dried root samples of sole and mixed cover crops harvested with the monolith method were ground to 0.2 mm (ultra-centrifugal mill, ZM 200, Retsch, Haan, Germany). Absorption spectra were measured by the FTIR-ATR spectrometer (Alpha-P with a diamond crystal attenuated total reflection (ATR) device, Bruker Optics, Ettlingen, Germany) with a resolution of 4 cm−1 and 32 scans in the spectral range of 4,000-400 cm−1. Each sample was measured five times. The evaluation of the FTIR-ATR spectra was conducted with the Opus software Quant 2 (version 7.2, Bruker Optics, Ettlingen, Germany).

The FTIR spectra of the single sample species were used for cluster analysis (Opus software, version 7.2, Bruker Optics) to confirm the species separation. For the quantification of the proportion of the species in the mixed species root samples, the FTIR spectra of the single species samples were used to generate a model. With the chosen model for each year, the FTIR spectra of the mixed species samples were evaluated. The output of this evaluation was the percental share of root mass of each species within the root samples of the cover crop mixture. The percentage of the single species was used for further calculations on the mixtures root mass. For details about establishing and evaluating the model see supplementary material and Kemper et al. (2022). Information about the species-specific proportions of root length in the cover crop mixtures could not be derived.

Calculations

Based on the root length and root mass obtained from the samples, root length density (RLD), root mass density (RMD) and specific root length (SRL) were calculated for the monoliths. RLD is root length per soil volume in cm cm−³, RMD is root mass per soil volume in mg cm−³ and SRL is root length per root mass in m g−1. Statistical comparisons of RLD, RMD and SRL were done in combined soil layers of 0–10 cm, 10–30 cm, 30–60 cm and 60–100 cm.

The land equivalent ratio (LER) (Mead and Willey 1980) was used to analyse the land use efficiency of cover crop mixtures for root and shoot mass and shoot nutrient uptake. An LER > 1 indicates a higher land use efficiency of intercrops compared to sole crops. LER is calculated by adding up partial land equivalent ratios pLER (Eq. 1). Yi and Mi are biomass (g/m²) or nutrient uptakes (g/m²) of intercrop and sole crop of species i, respectively.

$$\text{LER}=\sum{\text{pLER}}_\text{i}=\sum\frac{Y_\text{i}}{M_\text{i}}$$
(1)

The mixture performance of root and shoot traits was assessed by computing relative mixture effects (RME). RME is the relative gain observed in the mixture compared to the expected value, which is the sole crops average weighted by relative density pi (Eq. 2). Relative density is the ratio of mixture and sole crop density. Here, the values for pi were 1/2 and 1/3 for two-species and three-way mixtures, respectively.

$$\text{RME}=\frac{\sum(Y_\text{i}-p_\text{i} \ast M_\text{i})}{\sum p_\text{i} \ast M_\text{i}} \ast100$$
(2)

The overall diversity effects (g/m²) can be interpreted as a net biodiversity effect (NE), which can be decomposed into the complementarity effect (CE) and the selection effect (SE), according to Loreau and Hector (2001). Following this concept, we used the notation of Elhakeem et al. (2019). NE is the difference between observed and expected yield of the mixture and the sum of CE and SE (Eqs. 3 and 4). CE indicates facilitation or niche differentiation in mixtures whereas SE shows species dominance. Details about the interpretation of complementarity and selection effect are discussed in Kemper et al. (2022). Yi, Mi and pi are used as in the equations above.

$$\text{NE}=\sum(Y_\text{i}-p_\text{i} \ast {\text{M}}_\text{i})$$
(3)
$$\text{NE}=\text{CE}+\text{SE}$$
(4)

CE is the species number (N) multiplied by the average relative yield gain (∆R̅Y) and the sole crops average yield M̅ (Eq. 5).

$$\text{CE}=\text{N}\ast \Delta \overline{\text{R}}\text{Y}\ast \overline{\text{M}}$$
(5)

∆RY, the relative yield gain, is the observed minus the expected relative yield (Eq. 6).

$$\Delta {\text{RY}}_\text{i}={\text{RY}}_\text{i}-p_\text{i}=\frac{Y_\text{i}}{M_\text{i}}-p_\text{i}$$
(6)

SE is the species number multiplied by the covariance between relative yield gains and sole crop yields (Eq. 7).

$$\text{SE}=\text{N} \ast \text{cov}(\Delta{\text{RY}}_\text{i},{\text{M}}_\text{i})$$
(7)

Statistical analysis

Data analysis was performed with R (version 3.5.2) and R studio (version 1.1.463) (R Core Team 2018). A one-factorial analysis of variance was conducted for shoot biomass, root biomass, root-shoot-ratio and mean of RLD, RMD, SRL in combined soil layers. Subsequently, mean comparisons were done with a Tukey test at a significance level of α = 0.05 with the agricolae package (Mendiburu 2019). The prerequisites for analysis of variance, homogeneity of variances and normal distribution, were visually checked with a plot of residuals versus fitted values and a normal quantile-quantile plot, respectively (Kozak and Piepho 2018). If necessary, data were square root or logarithm transformed to meet the above-mentioned prerequisites. A one-sample t-test against 1 was used to test the significance of LER and one sample t-test against 0 was used to test the significance of RME and diversity effects. With a two-sample Welch t-test the percentage share of root mass and the difference between expected and observed shoot biomass, root biomass, RMD and pLER were tested.

Results

In 2018 and 2019, the months June to September were characterized by above-average temperatures as well as below-average precipitation and thus dry soil conditions. However, drought intensity in 2018 was stronger than in 2019, as not only the summer but also the autumn was dry (Kemper et al. 2020).

Shoot biomass and nutrient uptake in sole and mixed cover crops

Total shoot dry mass (DM) did not differ significantly between cover crops in 2018 (Fig. 1). In 2019, shoot DM of crimson clover was significantly lower compared to cover crops containing oil radish. Shoot DM of rye was significantly lower than those of oil radish as sole crop and the OR-CC-WR mixture. When mixture proportions changed compared to expected values, proportions of clover and rye were reduced whereas proportions of oil radish increased (Fig. 1), indicating differences in competitiveness among these species. Partial LER values were mostly in line with these results; in both mixtures, pLER of rye was significantly increased in 2018 and decreased in 2019, while pLER of crimson clover was significantly reduced in both years and pLER of oil radish was higher than expected in the OR-CC-WR mixture in 2019 (Tables S1, S2). High pLERs in stronger partners were partly associated with low pLERs in weaker partners, especially in 2019. As a consequence, the LER values did not differ significantly from 1 for shoot biomass or shoot nutrient uptake (Tables S1, S2).

Fig. 1
figure 1

Mean values ± standard deviation of shoot (n = 4) and root (n = 3) dry mass [g/m²] and mean values of root-shoot-ratio (n = 3) of sole cover crops oil radish (OR), winter rye (WR), crimson clover (CC), 50:50 cover crop mixtures of oil radish and winter rye (OR_WR_50_50), 33:33:33 cover crop mixtures of oil radish, crimson clover and winter rye (OR_CC_WR_33_33_33) and of expected mixtures (corresponding proportions of sole crops) in 2018 (A) and 2019 (B). Bars above 0 represent shoot dry mass, bars below 0 show root dry mass. Bar colours refer to species: violet for radish, green for rye, red for clover. Transparent bars represent expected values. Red numbers below bars represent root-shoot-ratio. Different lowercase letters indicate significant differences (p ≤ 0.05, HSD Tukey-Test) between different cover crops for total shoot or total root dry mass or root-shoot-ratio within one year. Different uppercase letters indicate significant differences (p ≤ 0.05, t-test) between observed and expected dry mass for shoot or root of intercrop partner within one year

Full size image

The analysis of the mixture effects into their components (complementarity and selection effects) yielded variable results. For the shoot dry matter and nutrient uptake, CE values were not significantly different from 0, except for N uptake which was positive in the two-species-mixture in 2018; significant SE values were found for both mixtures in 2019 (Table 1, Table S4). Mixtures with non-legumes significantly reduced soil mineral N in the soil before winter compared to legume sole crops and bare fallow (Fig. S2).

Table 1 Mean values ± standard deviation of net biodiversity effect (NE), complementarity effect (CE) and selection effect (SE) of a 50:50 mixture of oil radish and winter rye (OR-WR) and a 33:33:33 mixture of oil radish, crimson clover and winter rye (OR-CC-WR) for shoot dry matter (DM) [g/m²] (n = 4) and root dry matter (DM) [g/m²] (n = 3). Effects significantly different from 0 (p ≤ 0.05, t-test) are given in bold
Full size table

Root biomass in sole and mixed cover crops

Total root DM did not differ significantly among cover crops in 2018 (Fig. 1). In 2019, root DM of crimson clover was significantly lower than for all other cover crops.

When mixture proportions changed compared to expected value, proportions of clover were reduced whereas proportions of oil radish increased. LER values for root mass were < 1, except in the OR-CC-WR mixture which reached 1.35 in 2019 (Tables S1 and S2). However, LER values for root mass were not significantly different from 1. pLER of rye root mass was significantly reduced in both mixtures in 2019. As for shoot, complementarity effects (CE) of root mass were not significantly different from 0.

Root-shoot-ratio did not differ significantly among cover crops in 2018 (Fig. 1). In 2019, rye had the highest root-shoot-ratio and differed significantly from oil radish, crimson clover and mixtures.

Root length density, root mass density and specific root length

Topsoil RLD in both years was highest for rye as a sole crop (Fig. 2A, B). In line with the root mass results, the RLD of crimson clover as a sole crop was generally low. In the lower subsoil (60–100 cm) RLD of crimson clover and rye as sole crops was lower than for the cover crop treatments with oil radish.

Fig. 2
figure 2

Mean values ± standard deviation (n = 3) of root length density [cm/cm³] (RLD) (A, B), root mass density [mg/cm³] (RMD) (C, D) and specific root length [m/g] (SRL) (E, F) of sole cover crops oil radish (OR), winter rye (WR), crimson clover (CC), 50:50-cover crop mixtures of radish and rye (OR_WR_50_50), 33:33:33-cover crop mixtures of radish and rye (OR_CC_WR_33_33_33) in 2018 (A, C, E) and 2019 (B, D, F) at different soil depth levels. Colours refer to cover crops: green for rye, yellow for 50:50-mixture, orange for 33:33:33-mixture, red for clover and violet for radish. Different lowercase letters indicate significant differences (p ≤ 0.05, HSD Tukey-test) between different cover crops within the different soil depth layers (0–10 cm, 10–30 cm, 30–60 cm and 60–100 cm) within one year

Full size image

RMD of rye as a sole crop was highest between 10 and 30 cm in both years (Fig. 2C, D). In the lower subsoil (60–100 cm) RMD of sole crop crimson clover was lowest and RMD of sole crop oil radish and the OR-CC-WR mixture were highest.

Generally, SRL of rye tended to be lower than SRL of crimson clover, oil radish and the mixtures below 10 cm. SRL of oil radish was significantly higher than that of rye and crimson clover in the 10–30 cm soil layer in 2018 (Fig. 2E). In 2019, SRL of rye was significantly lower than that of oil radish and both mixtures in the 30–60 cm soil layer (Fig. 2F).

Belowground mixture effects

Relative mixture effects of RLD in both mixtures were negative or positive in topsoil and total profile (Fig. 3). With increasing soil depth, RME of RLD increased. A significantly positive RME of RLD was found in the lower subsoil for the OR-CC-WR mixture in 2019.

Fig. 3
figure 3

Mean values ± standard error (n = 3) of relative mixture effect [%] for root length density (RLD), root mass density (RMD) and specific root length (SRL) for 50:50 cover crop mixtures of radish and rye (OR_WR_50_50) and 33:33:33 cover crop mixtures of radish, clover and rye (OR_CC_WR_33_33_33) in 2018 (A) and 2019 (B) at different soil depth levels and for the total soil depth profile. Colours refer to cover crop mixtures: yellow for 50:50-mixture and orange for 33:33:33-mixture. Asterisks indicate a significant difference in RME from 0 (p ≤ 0.05, t-test)

Full size image

RME of RMD was partly negative in the topsoil and total profile, whereas it was positive in the lower subsoil in both mixtures and years (Fig. 3). RME of RMD between 10 and 30 cm was significantly negative for the OR-WR mixture in 2019.

Relative mixture effects of SRL were positive for both mixtures and years in all soil layers except in 0–10 cm in 2019 (Fig. 3). RME of SRL between 10 and 30 cm were significantly positive for the OR-WR mixture in 2019.

Species-specific root mass proportions and vertical niche differentiation of RMD between sole and mixed cover crops

FTIR results revealed that the root mass density was dominated by oil radish below 30 cm in both mixtures and years (Fig. 4). In the topsoil winter rye reached shares of RMD comparable to oil radish except for the OR-CC-WR mixture in 2019. RMD shares for crimson clover were low across all depths in both years.

Fig. 4
figure 4

Mean values (n = 3) of root mass share [%] of oil radish (violet), crimson clover (red) and winter rye (green) at different soil depth in 2018 (A, B) and 2019 (C, D) in a 50:50 cover crop mixture of oil radish and winter rye (A, C) and in a 33:33:33 cover crop mixture of oil radish, crimson clover and winter rye (B, D). Different uppercase letters indicate significant differences (p ≤ 0.05, t-Test) between proportions of root mass of mixture partners within one soil depth layer

Full size image

In the analysis of niche differentiation, observed and expected RMD of oil radish did not differ significantly in both mixtures and years (Fig. 5B, E, G, J). Observed RMD of winter rye was lower than expected RMD in single topsoil layers in both years in the OR-WR mixture (Fig. 5A, F) and in 2019 in the OR-CC-WR mixture (Fig. 5H). Crimson clover had a lower than expected RMD in topsoil in both years (Fig. 5D, I).

Fig. 5
figure 5

Mean values ± standard deviation (n = 3) of observed (black) and expected (grey) RMD of winter rye (A, F) and oil radish (B, G) in a 50:50 cover crop mixture of oil radish and winter rye and of winter rye (C, H), crimson clover (D, I) and oil radish (E, J) in a 33:33:33 cover crop mixture of oil radish, crimson clover and winter rye in 2018 (A, B, C, D, E) and 2019 (F, G, H, I, J). Different lowercase letters indicate significant differences (p ≤ 0.05, t-test) between observed and expected RMD within one soil depth layer for one crop partner

Full size image

Discussion

Biomass in sole and mixed cover crops

Unlike winter rye and crimson clover, oil radish completed its transition from vegetative to generative development before winter which is usually accompanied by a higher growth rate. This resulted in significantly higher shoot dry mass of cover crop treatments containing oil radish compared to the winter hardy sole crops rye and crimson clover (Fig. 1). Furthermore, crucifers have a high competitiveness for abiotic resources (Couëdel et al. 2018) in cover crop mixtures. Moreover, oil radish is known to be dominant due to its early development and horizontal leaf orientation, leading to a high light interception and shading of other mixture partners (Elhakeem et al. 2021b). Crimson clover was not able to compete with oil radish or winter rye, as its biomass share was consistently lower than expected from sole crops (Fig. 1). Crimson clover was not only weak in the mixture, root biomass was also low in the sole crop. This observation is in line with the results obtained by Amsili and Kaye (2021) in a five-species mixture with the lowest root biomass carbon found for crimson clover. In our study, winter rye had the highest root biomass in both years and the highest root-shoot-ratio (Fig. 1). Similarly, Amsili and Kaye (2021) observed the highest root biomass and root-shoot-ratio in triticale.

The LER values underlined the above-mentioned results. In the OR-WR mixture, root LER was below 1 in both years and shoot LER was only greater in 2018 than 1 (Table S1). In this year, there were two strong partners (oil radish and winter rye with significantly enhanced pLER), contributing equally to shoot performance of the mixture. In contrast, in 2019 oil radish dominated the mixture, while winter rye growth was weak, yielding an overall shoot LER < 1. Yu et al. (2016) discussed in their meta-analysis that a competitive balance in species mixtures is important for the performance of the mixture. In the OR-CC-WR mixture, pLER values for crimson clover were always low (Table S2). Thus, this mixture consisted of two strong (oil radish and winter rye) and one weak partner in 2018, and of one dominating (oil radish) and two weak partners in 2019. Overall, positive but non-significant mixture effects of shoot parameters occurred in the three-species mixtures. Our first hypothesis that there are positive mixture effects on shoot biomass and nutrient uptake was confirmed only for conditions resulting in balanced species composition.

Complementarity effects (CE) for N uptake were positive in 2018 (Table S3), showing that two strongly competing partners increased their N uptake by complementarity. The strong oil radish growth resulted in a positive and significant selection effect (SE) for shoot mass in both mixtures in 2019 illustrating the imbalance in mixture. The gain of oil radish was higher than the loss observed of winter rye and crimson clover. Likewise, other studies emphasised the importance of species identity and a concomitant high selection effect in cover crop mixtures (Freund et al. 2020; Wendling et al. 2019). In line with our results, complementarity effects in the literature are variable, but CE for root mass are mostly missing. In another study with cover crop mixtures (Elhakeem et al. 2019) positive CE were found for aboveground biomass. Grassland studies examined aboveground CE and the influence of root traits (Bakker et al. 2016; Felten and Schmid 2008). Felten and Schmid (2008) showed that rooting depth does not necessarily enhance CE of aboveground biomass in a short-duration grassland species pot experiment. Bakker et al. (2016) observed no relationship between root trait diversity and aboveground CE in a grassland biodiversity experiment.

Other grassland studies (Mommer et al. 2010; Oram et al. 2018) also reported positive CE for belowground biomass. Oram et al. (2018) showed that root mass CE increased with species richness, but there was no evidence that functional diversity of rooting patterns was relevant for root mass CE. Instead, a weak but significant relationship between CE and deep rooting was found. However, it needs to be considered that they examined roots only until 40 cm soil depth. In our previous experiment, differences in rooting depth were only found for crimson clover and oil radish grown as sole crops in 2018 (Kemper et al. 2020). Detecting hardly positive CE for belowground biomass in cover crop mixtures in the current study compared to grassland studies (Dirks et al. 2021; Mommer et al. 2010; Oram et al. 2018), may be due to the shorter growth period of cover crops compared to grassland species, as previously discussed by Kemper et al. (2022).

Root traits and their mixture effects

In our root trait analysis, the highest RLD values were observed for winter rye within the topsoil (0–30 cm, Fig. 2). RLD and RMD for crimson clover were low in top- and subsoil. For both RLD and RMD, high values were found in the subsoil for oil radish and mixtures containing oil radish. High values for RLD, root intensity and root growth rate of oil radish were previously shown by Wendling et al. (2016), Kristensen and Thorup-Kristensen (2004) and Thorup-Kristensen (2001), respectively. Based on these results, we can confirm the second hypothesis, that oil radish and mixtures containing oil radish differ in RLD and RMD in the subsoil compared to rye and crimson clover. Thus, positive RME for RLD and RMD were mainly found in the subsoil (Fig. 3), due to the taproot architecture of oil radish with a strong proliferation of roots in deeper soil layers (Bodner et al. 2019). In a previous experiment (Kemper et al. 2020), it was shown that oil radish intensively occupied biopores in the subsoil. This further explains higher root quantities observed in the subsoil in mixtures containing oil radish. Thus, mixtures with oil radish ensure rooting into the subsoil which may lead to enhanced nutrient uptake from the subsoil and prevention of leaching. This can be seen from the soil mineral N data in Fig. S2. Hence, our third hypothesis, that positive mixture effects for RLD and RMD occur for two- and three-species mixtures in topsoil and subsoil, can only be accepted for subsoil, not for topsoil.

Our results raise the question of why there is no strong mixture effect for RMD and RLD in the topsoil. For shoot biomass, Elhakeem et al. (2021a) detected that the average shoot biomass was greater in mixtures than in sole crops, but it was similar when mixtures were compared to the highest yielding pure stands. In another study, Elhakeem et al. (2021b) showed that mixture performance is governed by radiation interception of the dominant species. The more competitive partner was faster in ground cover and consequently contributed more to ground cover. Similar mechanisms of dominance may be true for belowground interactions, as oil radish has a higher rooting depth penetration rate than winter rye (Thorup-Kristensen 2001). This might explain the higher soil volume and root mass dominance of oil radish. Thus, winter rye and crimson clover were outcompeted for root growth in mixtures, which become similar to pure stand of the dominant species.

In both years and mixtures, the RME of SRL was positive, except 0–10 cm in 2019. This confirms our fourth hypothesis and is in line with the results observed for lupin-rye mixtures (Kemper et al. 2022). In a situation with a competing mixture partner plants may increase root length for better nutrient and water acquisition without changing investment in root dry matter, as SRL is associated with belowground economics (Freschet et al. 2021).

Vertical root niche differentiation

The root mass percentages, revealed by the FTIR analysis, did not show a clear pattern with homorhizous winter rye dominating the topsoil and allorhizous oil radish dominating the subsoil (Fig. 4). Oil radish dominated top- and subsoil rooting in 2019, while in 2018 winter rye was present to the same extent as oil radish at least in the topsoil. In both mixtures, there was no difference between expected and observed RMD for oil radish (Fig. 5), thus no indication for niche differentiation was found. Winter rye was outcompeted for its niche in upper soil layers by oil radish in the OR-CC-WR mixture in 2019 and in the OR-WR mixture in both years. The same was detected for crimson clover in both years. Thus, our data lead to the rejection of our fifth hypothesis regarding vertical root niche differentiation.

Likewise, in a mustard-phacelia-oat-clover mixture, the weaker components clover (Trifolium alexandrinum L.) and oat (Avena strigosa Schreb.) were outcompeted for their niche by phacelia (Phacelia tanacetifolia Benth.) and mustard (Sinapis alba L.), which occupied the same vertical niche as in sole crops (Heuermann et al. 2019). For white clover and perennial ryegrass, Dirks et al. (2021) found the same vertical niche of clover and grass root mass in two of three years. Mommer et al. (2010) reported that belowground complementarity and overyielding were not due to vertical nice differentiation, but rather resulted from enhanced rooting in topsoil of one species.

Vertical niche differentiation and mixture effects might not have been established pronouncedly in our experiment due to a certain imbalance between the mixture partners. Murrell et al. (2017) showed that the productivity of mixtures is not higher than the best performing monoculture, especially when the mixture is dominated by this species. To achieve diverse and well-balanced mixtures, they suggested the following two strategies: (1) sow early and (2) reduce the sowing density of aggressive species. Elhakeem et al. (2021b) suggested that if species with slow initial growth or low radiation use efficiency are used as mixture components, the proportion of these species should be increased to facilitate their competitiveness. Bybee-Finley et al. (2022) provided a framework to adjust seed density ratios of cover crop mixtures with the objective to avoid dominating competitive species in the mixture and identifying relative seeding rates that produce high biomass with high species evenness. A well-designed mixture might more likely show vertical niche differentiation and mixture effects.

Belowground interactions in cover crop mixtures of oil radish, winter rye and crimson clover

Different root traits represent specific functions of cover crop roots. Root intensity for improvement of aggregate stability can be represented by topsoil RLD (Hudek et al. 2022) and root intensity for nutrient uptake by subsoil RLD (Thorup-Kristensen 2001; Sainju et al. 1998). Nutrient uptake efficiency can be linked to SRL (Eissenstat 1992; Freschet et al. 2021). Carbon storage potential is determined by root mass (Kätterer et al. 2011; Poeplau and Don 2015) and rhizodeposition (Pausch and Kuzyakov 2018). The latter was not measured in the current study. N provision is linked to N2-fixation. Having measured these traits in cover crop mixtures, we can assess the effects of belowground functional diversity. In particular, there are potential effects of species identity and of mixing (Table 2). The mixture effect is the comparison of the mixture with the average of sole crops, called overyielding when positive. The identity effect shows that the performance of the mixture is determined due to the identity of the species in the mixture. Mixture effects were found for SRL and subsoil RLD. For both traits there was also an identity effect of oil radish, which was dominating the mixtures. The mixture effect of subsoil RLD occurred due to intensive subsoil rooting and biopore use of oil radish (Kemper et al. 2020). Topsoil structure may be improved by winter rye. N can be provided by N2-fixing crimson clover, yet this species was suppressed in mixtures because of its low growth rate. Both winter rye and oil radish achieve high root biomass, but no root biomass overyielding in mixtures was found. Carbon storage potential is not only represented by root biomass, but also by rhizodeposition; while this was not measured in the current study, it might be important in cover crop mixtures (Gentsch et al. 2020).

Table 2 Identity and mixture effects of sole and mixed cover crops for different root traits representing specific functions of cover crop roots. The identity effect is given due to the species contained in the mixture: WR = winter rye, OR = oil radish, CC = crimson clover. The mixture effect, which is comparing the mixture with the mean of the sole crops, can be negative (-), positive (+), not existing (0) or was not analysed (?)
Full size table

Contrary to our assumptions, the mixing effects of the architectural root traits RLD and RMD were only found in the subsoil, whereas root biomass overyielding and a change in root allocation patterns (vertical niche differentiation of RMD between sole and mixed crops) were not detected. In contrast, there was a clear positive mixing effect for the morphological root trait specific root length (SRL). An increase in this acquisitive root trait SRL was previously found for lupin-rye mixtures (Kemper et al. 2022), faba bean-triticale mixtures (Esnarriaga et al. 2020) and vetch-rye mixtures (Bukovsky-Reyes et al. 2019). Changes in SRL might also result in altered root longevity and root turnover (Ostonen et al. 2007; Ryser 1996). For trees it was found that root exudation was increased with higher SRL (Meier et al. 2020; Yin et al. 2023). For herbaceous crop species, such a relationship has also been proposed, but could not be confirmed in a first attempt comparing whole root systems of different plant species (Roumet et al. 2008). To our knowledge, to date there are no studies attempting to relate root exudation to roots of different sizes or orders within the same species. Gentsch et al. (2020) found higher rhizodeposition in diverse cover crop mixtures compared to sole cover crops. Thus, it might be that for short term grown cover crops mixture effects are less found in change of root allocation patterns but mixture effects are rather shaped by morphological adaptions which might push rhizodeposition and microbial processes.

Conclusion

In mixtures, oil radish was dominant in shoot and root underlining the importance of species identity in cover crop mixtures. Positive mixture effects for RLD and RMD were more frequent in the subsoil due to a strong rooting of oil radish if integrated as a mixture partner. Consistently positive mixture effects were found for SRL, i.e. mixing makes roots thinner. Root mass complementarity, vertical root niche differentiation and mixture effects in topsoil might have not occurred pronouncedly because partners were not sufficiently balanced in mixtures. Thus, future research should focus on the balance when mixing different cover crop species. More conceptual planning of seed density ratios and mixture composition is advised to design well-balanced cover crop mixtures to exploit mixture effects above- and belowground more effectively. Mixture effects seem to be more prominent for morphological than architectural root traits. Enhanced SRL in mixtures might be linked to higher root turnover and rhizodeposition which would influence cover crop functions such as carbon storage and microbial activity. Future research on roots of cover crop mixtures should consider root turnover, rhizodeposition, microbial processes as well as other acquisitive root traits such as root tips, branching and root N content.

We have given first insights into belowground interactions of mixed cover crops and mixture effects on their root traits in autumn. Future root research will profit from methods that allow an understanding of belowground interaction throughout the crop growth period.