Competition or complementarity between associated plants due to belowground interactions has been observed in alley-cropping systems (ACs), but the initialization of these processes remains poorly investigated. Here, we used the core-break and soil coring method to quantify the vertical tree and wheat fine root abundance and biomass down to 120 cm in a 4-year-old temperate AC. Fine roots were measured at 2 m from a reference tree (hornbeam, wild cherry or willow) in tree–wheat AC, pure-forest associated with ryegrass (FC) and wheat sole-crop (CC) plots at the Ramecourt experimental site. The mean wheat fine root abundance (WFRA) was twice as high in the CC plot (874 ± 152 m− 2) as in the AC plot (437 ± 47 m− 2). It was significantly higher for wheat associated with hornbeam than for willow, particularly at the 10 cm depth. Tree fine root abundance (TFRA) was linearly correlated with tree fine root biomass (TFRB) for hornbeam (R² = 0.79***), willow (R² = 0.77***) and wild cherry (R² = 0.54***). Using TFRA, the van Noordwijk's equation gave a better prediction of the TFRB for willow and wild cherry than for hornbeam. The mean value of the TFRA was seven times higher in the FC plot (1116 ± 97 m− 2) as in the AF plot (146 ± 24 m− 2) for all soil depths and all tree species due to the lack of nutrients from the absence of fertilization. At 4 years old, willow and hornbeam fine roots cohabited with wheat in the upper soil layer, whereas wild cherry had already developed deep fine roots under the crop rooting zone.
Research Article
Tree–wheat vertical fine root distribution in a four-year-old temperate alley-cropping system
https://doi.org/10.21203/rs.3.rs-3143059/v1
This work is licensed under a CC BY 4.0 License
Journal Publication
published 27 Jan, 2024
You are reading this latest preprint version
Tree fine root biomass
wheat fine root biomass
alley cropping
core-break method
soil coring
tree species
In alley-cropping systems (ACs) where tree rows are intercalated with annual crops in arable land, the competition or complementarity between associated plants is often determined by interspecific root overlap (Isaac and Borden, 2019) or niche differentiation along the soil profile, respectively. It has been widely documented that intercropping systems modify the spatial root distribution of trees. Andrianarisoa et al. (2016) showed in an adult walnut–wheat AC that tree fine roots grew deeper than in a forestry control plot, indicating the plasticity of trees. Bouttier et al. (2014) also found an adaptation of oak root distribution that occurs through the reduction of density in the upper soil layer, resulting in a spatial separation with the intercropped hay root system. Trees in an AC have been found to develop deeper root systems than those in monocrop stands and to have a higher root density in the tree rows than in the alleys (Cardinael et al. 2015). All of these studies showed a complementary spatial distribution between trees and crop fine roots in ACs; however, this was not always observed. According to Mulia and Dupraz (2006), the poplar plate root system overlapped with that of wheat despite intercropping. Similar results were observed by Bouttier et al. (2014) and Swieter et al. (2022), who reported a cohabitation between poplar and crop fine roots within a transition zone outside of the row.
Different methods with varying assets and constraints have been reported in the literature for studying plant root distribution in systems as representative as possible of reality. The core-break method was proposed by van Noordwijk et al. (2001) and was used to characterize the spatial variability of roots in ACs (Moreno et al. 2005; Mulia and Dupraz 2006; Andrianarisoa et al. 2016). In the soil coring method, volumetric soil samples are obtained using an auger and washed to separate roots from soil particles (Coleman 2007; Ravenek et al. 2016; Andrianarisoa et al. 2017). With the development of high-resolution cameras, minirhizotrons have been installed at some experimental sites to continuously monitor root growth (Jose et al. 2001; Germon et al. 2016). Then, a root profile can be determined by counting roots on a grid (Bouttier et al. 2014) or by root observation on a profile wall (Battie-Laclau et al. 2020). Finally, root mapping can be performed using a plastic overlay (van Noordwijk et al., 2001). All these methods have advantages as well as disadvantages. The core-break method may be used for the rapidity of its measurements, but it presents the main disadvantages of (i) the spatial variability of samples, which is higher than that of the other methods; (ii) the effect of preferential root orientation; (iii) the random variation in the number of roots intersecting the plane of observation; and (iv) the counting errors (van Noordwijk et al., 2001). These issues may be compensated by increasing the number of samples collected. In addition, the core-break method evaluates tree fine root abundance on a surface basis, but the conversion of tree fine root abundance into biomass per volume of soil is not always possible. An equation has been proposed by van Noordwijk et al. (2001) to determine the tree fine root biomass from the number of roots counted with the core-break method; however, the calibration factor is not always available, and its validity for young stands is questionable. In contrast, the soil coring method allows the accurate quantification of fine root biomass within a defined volume of soil, but it is time-consuming, and root recovery efficiency is operator dependent.
Studies on tree and crop fine root distribution in very young AC stands are scarce. Recent studies (O’Connor et al. 2023) showed that before 5 years old, young trees developed mainly shallow fine roots (< 30 cm) close to tree rows and beyond this age started to produce deeper roots in the tree row. Zhang et al. (2015), using soil coring, measured the root length density, root diameter and specific root length of very young (< 2 years old) intercropped Persian walnut trees and showed that tree roots spread under the crop, which in return tended to have a shallower distribution in the soil profile. In 2-, 4- and 6-year-old jujube tree/wheat intercropping systems, Zhang et al. (2013) used the same method and concluded that wheat and jujube had less root length density than monocrop systems. Most studies on very young AC stands are focused on tree effects on resource sharing (Zamora et al. 2009), early changes in soil microbial activity (Clivot et al. 2020) or the safety-net role (Bergeron et al. 2011). However, root biomass quantification at the early stage of tree establishment in plots remains relevant to producing knowledge about the initialization of roots and to anticipating the management of potential future competition.
The aim of this work was to quantify the tree and crop fine root distribution within a very young temperate alley cropping system. We compared core-break and soil coring methods to verify the reliability of each method for assessing the vertical distribution of young tree fine roots. We hypothesized (1) that trees at a young age should develop fine roots in the crop rooting zone and share a transition zone with the crops; (2) that vertical tree fine root distribution is significantly affected by the type of system within which the tree grows; (3) that intercropping changes the crop fine root distribution early after tree establishment; and (4) that core-break methods allow us to effectively estimate tree fine root biomass.
2.1. Experimental site description
This study was carried out at the alley-cropping experimental site of Ramecourt located 77 km from Lille city in Northern France (50°22’N, 2°17’E) (Andrianarisoa et al. 2019). The climate is oceanic with an average annual temperature of 10.6°C and rainfall of 859 mm, as recorded from 2010–2020 at the Humières weather station situated 6.7 km from Ramecourt. The upper layer soil texture is a silt loam with 73% silt, a pH of 7.9 and an organic matter content of 2.1%. The experimental plot was a randomized block design with three replications. Within each block, the AC plot was compared with sole-crop (CC) and pure-forest (FC) control plots. In November 2018, one-year-old trees were planted in SW‒NE oriented rows spaced from 38 m and 7 m apart in the AC and FC plots, respectively. Within each row, tall trees were planted two by two at 1 m apart, with 8 m between pairs, and were intercalated every 1 m by 9 species of shrubs. The tall trees were Quercus robur (oak), Carpinus betulus (hornbeam), Juglans regia x regia (hybrid walnut), Alnus glutinosa (alder), Prunus avium (wild cherry) and Robinia pseudoacacia (black locust). The shrubs were Castanea sativa (chesnut), Cornus sanguinea (dogwood), Acer campestris (field maple), Euonymus europaeus (European spindle), Corylus avelana (common hazel), Tilia cordata (linden), Ligustrum vulgaris (wild privet), Salix alba (willow) and Viburnum lantana (wayfarer). Details of tree and shrub sequencing within rows in the AC plot are available in Andrianarisoa et al. (2019). In the FC plot, a cover crop of sown Lolium multiflorum (hereinafter referred to as ryegrass or RG) dominated the alleys and was mechanically cut twice a year. When considering only tall trees, the density of plantations was 66 trees ha− 1 and 357 trees ha− 1 in the AC and FC plots, respectively. In the CC and AC alleys, winter wheat (variety ‘Extase’) was sown in October 2020 following a previous chicory witloof crop and was harvested in August 2021.
2.2. Tree and crop fine root abundance and biomass measurements
Three tree species with contrasting growth rates—hornbeam, wild cherry and willow—were selected from the AC and FC plots. The willow grew faster and was taller than the other species except in the FC plot, where its growth was not significantly different from that of the other species (Table 1). In 2021, the tree height was on average 2.7 ± 0.7 m. In the AC plot, the tree height was significantly higher for willow (4.9 ± 0.3 m) than for wild cherry (3 ± 0.3 m) and hornbeam (2.3 ± 0.3 m). The growth rate was the same for all tree species in the FC plot, whereas it was significantly higher for willow (130 ± 30 cm y− 1) than for wild cherry (80 ± 20 cm y− 1) and hornbeam (60 ± 10 cm y− 1) in the AC plot (Table 1).
| Height (m) | Growth rate (m y− 1) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Tree species | AC | FC | #p value | AC | FC | #p value | ||||
| Hornbeam | 2.3 (± 0.3) a | 1.9 (± 0.2) a’ | ns | 0.6 (± 0.1) a | 0.4 (± 0.1) a’ | 0.04 | ||||
| Wild cherry | 3 (± 0.4) a | 2.1 (± 0.1) a’ | < 0.001 | 0.8 (± 0.2) a | 0.4 (± 0.05) a’ | < 0.001 | ||||
| Willow | 4.9 (± 0.3) b | 2.9 (± 0.3) b’ | < 0.001 | 1.3 (± 0.03) b | 0.4 (± 0.1) a’ | < 0.001 | ||||
| Mean | 3.2 (± 1.1) | 2.4 (± 0.5) | 0.008 | 0.8 (± 0.3) | 0.5 (± 0.2) | < 0.001 | ||||
| # For a given tree species, p values < 0.05 indicate a significant difference between AC and FC. ‘ns’ indicates not significant. | ||||||||||
Tree and crop fine root abundance and biomass were measured using “core break” and “soil coring” methods, respectively (van Noordwijk et al. 2001). Soil coring was carried out in November 2021. Soil cores down to a 1.2 m depth were collected 2 m from a reference tree using a portable electric core drill consisting of a gouge (60 cm length and 85 mm diameter) connected to an electrical percussion hammer (BOSCH GSH 27 VC, Apageo). In each block, two soil cores were collected per tree species in the AC plot but only one per tree species in the FC plot because of the small size of the plot. In the CC plot, two soil cores were extracted in each block in an area randomly chosen in the middle of the plot. Each 1.2 m soil core was divided into 0.2 m long subcores. Each subcore was broken by hand, close to the middle, and the number of living fine roots (diameter < 2 mm) visible on both horizontal surfaces was counted. The crop root assessment was achieved by means of the soil core collected in the CC plot, whereas those for trees were carried out from the soil core taken at the bottom of the trunk. For the RG, the counter was further trained to recognize fine roots by uprooting some plants taken in the middle of the FC alley. In comparison with crops, tree roots were more lignified, hairy and often brownish. Despite our identification experience, tree root counts might have been slightly overestimated in rows due to the presence of weeds in the AC plots. A single person carried out root counting for all samples to avoid bias from the counter. The mean of the number of fine roots counted on both sides of each subcore was expressed on a square meter basis and was called tree fine root abundance (TFRA; m− 2), wheat fine root abundance (WFRA, m− 2) or ryegrass fine root abundance (RGFRA). All subcores were recovered, transported to the laboratory and stored at 4°C.
In the laboratory, fine roots within each subcore were manually collected, sorted according to their affiliation (trees, crops, RG and weeds), dried (65°C, 48 h), cleaned and weighed for biomass measurement. Tree, crop, RG and total fine root biomass, hereafter referred to as TFRB, WFRB, RGFRB and TTFRB, respectively, were expressed in kilograms of dry matter per hectare (kg DM ha− 1). For a given soil depth, the percentage of tree or wheat fine root biomass and the root equivalent ratio (RER) were calculated according to the following formulae:
\({\%TFRB}_{i}=\frac{{TFRB}_{i}*100}{\left({TTFRB}_{i}\right)}\) (Eq. 1)
\({\%WFRB}_{i}=\frac{{WFRB}_{i}*100}{\left({TTFRB}_{i}\right)}\) (Eq. 2)
\(RER \left(i\right)=\frac{{TFRB}_{AC\left(i\right)}}{{TFRB}_{FC\left(i\right)}} + \frac{{WFRB}_{AC\left(i\right)}}{{WFRB}_{CC\left(i\right)}}\) (Eq. 3)
where i is the depth of the measurement and TFRBAC(i), WFRBAC(i), TFRBFC(i) and WFRBCC(i) are TFRB and WFRB in the AC, FC and CC plots at depth (i).
2.3. Statistical analyses
The variability of tree height and growth rate was analyzed using two-way ANOVAs, with the tree species or the type of system as explanatory variables together with the block, followed by a post hoc Tukey’s test.
For comparisons, our studied individuals were designated as follows: (i) the wheat in the CC and AC plots were referenced as Wheat-CC and Wheat-AC, respectively; (ii) the wheat close to alder, hornbeam or wild cherry in the AC plot was renamed Wheat-AC-Alder, Wheat-AC-Hornbeam and Wheat-AC-Wild cherry, respectively.
Two-way ANOVAs were performed to analyze the variability of the TFRA with the soil depth and the block as explanatory variables. A post hoc Tukey’s test was performed for each ANOVA. For fine root abundance, the comparison between the AC and CC plots for wheat and between the AC and FC plots for trees was performed using Student’s t test. A Kruskal–Wallis test was used to analyze the wheat and RG fine root abundance and the wheat, tree and RG fine root biomass variability according to the soil depth or the associated plant species or the type of system. This was followed by post hoc Mann–Whitney tests because the normality of the ANOVA residuals model was not confirmed. The comparison between Wheat-AC and Wheat-CC for the WFRB and between the AC and FC plots for the TFRB was performed using a Mann‒Whitney test. Finally, a Kruskal–Wallis test was used to analyze the TFRA and TFRB variation according to the tree species, followed by a Dunn's multiple comparisons test (p < 0.1). The WFRA and WFRB were analyzed with a two-way ANOVA with the associated tree species and the block as explanatory variables, followed by a post hoc Tukey’s test.
Correlation analysis was performed between the TFRA and TFRB for each tree species, and a second-degree polynomial equation was calculated between the WFRA and WFRB. Finally, the model proposed by van Noordwijk et al. (2001) was used to estimate the tree fine root biomass using TFRA, according to Eq. 4:
\(Simulated TFRB \left(kg {DM ha}^{-1}\right)=\left[\frac{10\text{*}\left(0.002\text{*}\text{X}\text{*}\text{T}\text{F}\text{R}\text{A}\text{*}h\right)}{\text{S}\text{R}\text{L}}\right]*1000\) (Eq. 4)
where h is the height of the subsoil core in m (h = 0.2 m); SRL is the specific root length of the tree species (SRL = 15 m g− 1 for the hornbeam (Kubisch et al. 2015); SRL = 27 m g− 1, for the wild cherry (Weemstra et al., 2020) and SRL = 66.8 m g− 1 for the willow (McIvor et al., 2014; Weemstra et al., 2020)) and X is a calibration factor depending on tree species. This factor was calculated using a nonlinear generalized reduced gradient with an Excel solver. The retained X value was the one that minimized the root mean square error (RMSE) between the simulated TRFB and observed TFRB according to the following formula:
\(RMSE \left(kg {DM ha}^{-1}\right)= \sqrt{\frac{1}{n}\sum _{i=1}^{n}{\left({Simulated TFRB}_{i}-{Observed TFRB}_{i}\right)}^{2}}\) (Eq. 5)
where n is the number of observed values.
The efficiency of the retained model was calculated with the following formula:
\(EF=1-\frac{\sum _{i=1}^{n}{\left({Simulated TFRB}_{i}-{Observed TFRB}_{i}\right)}^{2}}{\sum _{i=1}^{n}{\left({<Observed TFRB}_{i}>-{Observed TFRB}_{i}\right)}^{2}}\) (Eq. 6)
where < Observed TFRBi > is the mean value of Observed TFRBi.
For all the regression analyses, asterisks indicate a significant determination coefficient: *** for p < 0.001, ** for p < 0.01 and * for p < 0.05 level.
3.1. Wheat and RG fine root abundance and biomass
The mean value of WFRA was twice as high in the CC plot (874 ± 152 m− 2) as in the AC plot (437 ± 47 m− 2) (Fig. 1A). These differences were observed down to a 70-cm depth. The WFRA significantly decreased with depth in both the CC and AC systems. The WFRA was significantly different between the CC and AC plots down to a 30 cm depth for wheat associated with hornbeam and willow and down to a depth of 70 cm for wild cherry. In the AC plot, the mean WFRA was higher for wheat associated with hornbeam than for wheat associated with wild cherry and willow (Fig. 1B), particularly at a depth of 10 cm (Fig. 1D and 1F). Below a 30 cm depth, there was no longer a significant difference in the WFRA between tree species. Additionally, the RGFRA in the FC plot was significantly lower than the WFRA, regardless of the soil depth. It was evenly distributed along the vertical soil profile down to a 110 cm depth.
In contrast to the WFRA, no significant difference in WFRB was noticed between the CC plot (45 ± 17 kg ha− 1) and AC plot (21 ± 5 kg ha− 1), regardless of the associated tree species and the depth (Fig. 2A and 2B). In the AC plot, the WFRB significantly decreased with depth down to 50 cm depth and was stable beyond, whereas in the CC plot, the WFRB was uniformly distributed along the soil profile. For the AC plot, when data were analyzed by tree species, the WFRB decrease with depth was significant only for wheat associated with wild cherry (Fig. 2E). For RG, fine root biomass was particularly high at 10 cm depth (4983 ± 545 kg DM ha− 1), sharply decreased at 30 cm depth (100 ± 21 kg DM ha− 1) and was low beyond 70 cm depth.
The WFRA and WFRB were significantly correlated (R² = 0.58***) with a polynomial regression (Fig. 5G). On average, along a soil profile down to 120 cm depth, the WFRB represented 53% of the total fine root biomass in the AC plot, whereas it was almost 100% in the CC plot. The %WFRB was evenly distributed along the vertical profile (Table 2).
3.2. Tree fine root abundance and biomass
The mean TFRA value was seven times higher in the FC plot (1116 ± 97 m− 2) as in the AC plot (146 ± 24 m− 2) for data including all soil depths (Fig. 3A). This difference was observed at all soil depths down to 110 cm (Fig. 3C) and for all tree species (Fig. 3B). The TFRA was the highest at 10 cm depth and significantly decreased with depth in both the AC and FC plots except for in the case of the hornbeam in the AC plot. In the AC plot, the wild cherry and willow TFRA significantly decreased down to 50 cm depth (from 529 ± 87 m− 2 to 59 ± 33 m− 2) and then became null or stable beyond that depth (Fig. 3E and 3F). Hornbeam TFRA was evenly distributed throughout the soil profile (Fig. 3D). The TFRA was not significantly different between tree species at any depth in the AC plot except at 10 cm, where the TFRA was significantly lower for hornbeam than for willow (Fig. 3G). In the FC plot, the TFRA was the highest at 10 cm and then globally decreased with depth (Fig. 3C). For wild cherry, the TFRA at 10 cm was significantly higher than in all other soil layers, whereas for willow and hornbeam, it progressively decreased with depth (Fig. 3D, 3E and 3F). No significant difference was observed between tree species for the TFRA at each soil depth (Fig. 3G).
Regarding the TFRB, the same trend was observed as for the TFRA: the mean value of the TFRB was 12 times higher in the FC plot (496 ± 52 kg ha− 1) as in the AC plot (41 ± 18 kg ha− 1) (Fig. 4A). This difference was noticed for all tree species (Fig. 4B) at each soil depth except at 10 cm depth (Fig. 4C). The TFRB comparison between tree species showed a significant difference only at 10 cm depth: the values for willow were significantly higher compared to those for hornbeam in the AC plot and to those for wild cherry in the FC plot. The TFRB gradually decreased with depth in the AC plot (Fig. 4C).
Throughout the soil profile, the %TFRB was on average 55% and 37% in the AC and FC plots, respectively. In the FC plot, the %TFRB was very low at 10 cm depth (11%) and was on average 95% below this depth. In the AC plot, the %TFRB was evenly distributed along the soil profile (Table 2). Finally, in the AC plot, the TFRB was 19 times greater than the WFRB at 10 cm depth and on average 4.7 times higher from 30 to 50 cm depth. Finally, between 70 and 110 cm depth, the WFRB was on average 2.5 times higher than the TFRB. For hornbeam, the WFRB was always higher than the TFRB, but below 70 cm depth, the TFRB to WFRB ratio was null. For wild cherry, there were more tree than wheat fine roots down to 30 cm depth, below which the quantity was low. Additionally, the TFRB of willow was higher than the WFRB down to 50 cm depth, whereas at 70 cm depth, the WFRB was 1.25 greater than the TFRB, and no more roots were observed below.
The mean RER was equal to 0.6 in the topsoil and decreased to 0.3 from 30 to 90 cm depth (Table 2). The mean RER was higher than 1 at 110 cm depth, particularly for the willow, due to abundant crop roots observed in the CC plot (Fig. 2F). For the hornbeam, the RER progressively decreased with depth, whereas it was homogenous for the willow. For the wild cherry, the RER was the highest at the 10 cm depth and decreased strongly below the 30 cm depth.
3.3. Predicting tree fine root biomass based on core-break measurements
The linear regression analyses showed that the TFRA (m− 2) measured with the core-break method was linearly correlated with the TFRB (kg DM ha− 1) measured with soil coring for each tree species (Fig. 5A, 5B and 5C). The regression coefficient of determination was R² = 0.79***, 0.77*** and 0.54*** for hornbeam, willow and wild cherry, respectively. The calibration of the van Noordwijk et al. (2001) equation gave a better prediction of the TFRB for willow (EF = 0.6; Fig. 5F) and wild cherry (EF = 0.5; Fig. 5E), as the linear regression between simulated and observed could almost be confused with the bisector compared to hornbeam (EF = 0.4; Fig. 5D). With the SRL values proposed in the literature, the calibration factor for the conversion of the TFRA measured with the core break method into fine root biomass was X = 1.3 (RMSE = 211 kg DM ha− 1), 2.1 (RMSE = 166 kg DM ha− 1) and 8.1 (RMSE = 253 kg DM ha− 1) for hornbeam, wild cherry and willow, respectively.
4.1. Vertical distribution of wheat and RG fine roots
Wheat fine roots
Although we recorded a higher WFRA on average in the CC than in AC plots, we did not find a significant difference in the WFRB between the two systems, probably due to an underestimation of the WFRB. Indeed, soil coring was carried out in November when wheat fine roots started to decompose. These roots were sufficiently visible to be counted but were very delicate and breakable; therefore, after removing the soil particles sticking to the wheat roots, some of the roots could not be recovered or were poorly detected by the precision weighing scale. This was consistent with the values for the %WFRB, which was surprisingly low (50%), whereas wheat roots should have dominated the soil profile from March to July, according to Battie-Laclau et al. (2020), as the trees were still young. We assumed that during the time lapse between crop harvest and soil coring, part of the roots were decomposed, especially in the upper soil layer.
Regarding the WFRA, the lower values observed in the AC plot compared to the CC plot may be due (i) to higher levels of wheat root decomposition close to understory vegetation strips through the activity of living tree or weed roots; (ii) to lower wheat growth close to tree rows in the AC plot compared with the CC due to light competition from trees or to nutrient competition from weeds; or (iii) to a reduction in crop growth caused by border effects on this zone of tree/crop root cohabitation. Indeed, due to the water absorption by trees in the AC plot, water may have been shared between associated plants leading to a lower crop root abundance in the AC plot than in the CC plot. Many studies have shown a reduction in crop roots close to tree rows (Livesley et al. 2000; Mulia and Dupraz 2006; Zhang et al. 2013; Cardinael et al. 2015; Duan et al. 2019; Battie-Laclau et al. 2020). Tree and crop roots can overlap in the upper soil layer despite strong competition (Moreno et al. 2005). For instance, Duan et al. (2019) found a low wheat root density close to rows of 1- or 2-year-old walnut trees and attributed it to the overlap of wheat and walnut roots. Zhang et al. (2013) studied the root density in intercropped 2-, 4- and 6-year-old jujube tree and wheat, in monocropped wheat and in 2-, 4- and 6-year-old jujube only down to 100 cm depth and concluded that intercropped wheat and jujube had less root length density at all soil depths than monocropped wheat had and added that inhibition in wheat root growth increased with increasing jujube age. Contrary to our results, their study showed that intercropped wheat had shallower roots than the monocropped wheat crop. The competitiveness of roots in the cohabitation zone is the main reason for this reduction in growth. Species with high root length density are more competitive than others. Zhang et al. (2013) showed that 6-year-old jujube trees were more competitive and contributed to a decrease in crop root growth. The opposite results were found in a dryland AC, where the sorghum root system exploited a larger soil volume, utilizing soil resources more efficiently than in monocrops (Lehmann et al. 1998).
However, even though a reduction in crop root development was observed close to tree rows, in contrast to what was observed in a CC system, the sum of total roots observed close to the tree row was higher than the total roots in the CC system. Compared to monocropping, this cohabitation may contribute to improving water resources in agroforestry systems, increasing nutrient and water use efficiency. Even though the increase in resource use efficiency was largely documented in agroforestry systems (Rowe et al. 2001; Link et al. 2015; Zhang et al. 2022), our results showed that the same effect occurred for very young trees. Cohabitation between trees and crops was more pronounced in the topsoil, particularly when the wheat was associated with willow rather than with hornbeam. From 50 to 70 cm depth, the wheat in the CC plot was as well developed as it was in association with hornbeam, whereas the wheat associated with the wild cherry was less developed. These results showed that beyond a depth of 50 cm, hornbeam no longer had an impact on wheat fine root growth. This was in contrast to what occurred with wild cherry, which reduced wheat fine root growth beyond a depth of 50 cm. These results suggested that the hornbeam fine roots were mainly developed in the upper soil layer, whereas those of the wild cherry grew deeper. These observations were consistent with the findings of shallow fine root distribution in hornbeam reported by Abdi et al. (2009) and with the ability of wild cherry roots to develop deeply, as reported by Bibalani et al. (2008). Moreover, throughout the soil profile, the proportion of hornbeam fine roots was particularly lower than that of wheat (Table 2), supporting that wheat was dominant over trees. This shallow root system of the hornbeam was more exposed to plowing destruction than were the root systems of the other tree species.
RG fine roots
In the FC plot, we found a very high RGFRB value in the top 10 cm of soil, whereas the RGFRA value was very low. Such a difference may be due to an underestimation of RG fine root counted by the observer in the field due to a very high number of roots to count on a small surface. Below 10 cm depth, the RG fine root abundance or biomass was evenly distributed along the soil profile. Moreno et al. (2005) also showed shallower herbaceous root development in grazed or intercropped systems. A high and dense root system in the top 10 cm of soil is a common pattern for herbaceous plants (Jackson et al. 1996). To reduce competition with RG for belowground resources, and due to their ability to access deep soil layers, trees would be expected to develop a deep root system (Schroth 1995) with little root proliferation near the topsoil. This different spatial root distribution has been found to allow RG to use resources from the upper soil layer. Fernández et al. (2008) showed a complementary use of soil water between pines and grasses in a temperate semiarid AC. Cheng and Bledsoe (2004) reported on the high competitiveness of grass against blue oak seedlings for nitrogen due to the high rooting density and rapid growth rate of grass, as also reported by Ravenek et al. (2016).
4.2. Vertical fine root distribution of monocropped and intercropped trees
Globally, the tree fine roots were more developed in the FC plot than in the AC plot throughout the soil profile and for all tree species. Andrianarisoa et al. (2016) also found more fine roots in forest systems than in ACs down to 1 m depth. Indeed, due to the absence of fertilization, trees in the FC plot needed to develop more roots to acquire nutrients, whereas trees in the AC plot benefited from external input in rows during fertilizer spreading. Moreover, trees in the FC plot competed with the RG fine roots, particularly those that developed in the topsoil (Fig. 3D and 4D, Table 2). This finding corroborates the conclusion of Moreno et al. (2005) concerning the high abundance of herbaceous roots in topsoil when herbaceous plants were associated with trees. Unlike trees in the FC plot, where fine roots completely dominated the deep soil layer (< 20 cm), tree fine roots in the AC plot were evenly distributed along the soil profile (Table 1). Species such as hornbeam were completely dominated by wheat throughout the soil profile, whereas willow dominated the upper soil layer down to 50 cm depth and wild cherry dominated both the top (< 40 cm) and the deep soil layers (> 100 cm). Duan et al. (2019) also showed that intercropping significantly influenced walnut root density both vertically and horizontally. Some tree species present different root plasticity when intercropped due to crop management and stand age (O’Connor et al. 2023) and due to the ecophysiological traits of each species (Mulia and Dupraz 2006). Zhang et al. (2013) showed that young jujube dominated over wheat root plants in the AC plot. Duan et al. (2019) found that wheat reduced the root density of young walnut trees compared with that of monocropped 2-year-old walnut trees. In a young wheat–walnut AC, Zhang et al. (2015) showed that the roots of intercropped trees remained in shallow soil in the same layer as wheat roots. O’Connor et al. (2023) confirmed that before 5 years old, tree roots mainly colonized the rows and nearby areas in the upper soil layer. Moreover, the fact that our mean RER values were less than 1 suggested that the root development of associated plants in the AC plot was lower than in monocrop plots, certainly due to the young age of the tree (Zhang et al. 2013, 2019). However, it was particularly high for willow at 110 cm depth, suggesting that there was more competition in the upper soil layers forcing the wheat to root deeply. In the following years, the RER would be expected to be higher in depth as the trees age (Duan et al. 2019).
4.3. Comparison between core-break and soil coring methods
Core-break and soil coring are two methods largely used in agroforestry (van Noordwijk et al. 2001; Moreno et al. 2005) and arable land (van Noordwijk et al. 1995; Mulia and Dupraz 2006) to quantify the fine root abundance, biomass or length density for trees and crops. We showed a linear regression between the TFRA estimated with the core-break method and the TFRB evaluated with soil coring for each tree species. For wheat, a polynomial regression relationship was identified. These regression relationships attest to the coherence and reliability of our data for the two methods despite the incertitude related to the root counting bias and the early decomposition of wheat roots at the autumn sampling date. From the core-break method, Moreno et al. (2005) also showed a linear relationship between root abundance counted in small soil surface areas and root length density (km m− 3) obtained after root washing, scanning and length measurement. Unlike their findings, in our results, the intercept of the linear regression was positive for wild cherry (Fig. 5B) and negative for hornbeam (Fig. 5A) and willow (Fig. 5C). A positive intercept indicates that the observer in the core-break method counted zero tree fine roots on the soil surface, whereas fine roots existed within cores; conversely, a negative intercept means that the observer counted tree fine roots on the soil surface, whereas roots did not exist within cores. This is illogical and contradictory to the observations; in most cases of low root biomass obtained with the soil coring method, no roots could be counted on the broken faces. To address this error, Kücke et al. (1995) used the quadratic regression y = (c + dx)², where c > = 0; however, in our case, the polynomial model did not improve the prediction quality (results not shown). There are difficulties in both methods concerning the identification of fine roots during counting in core breaks or during root sorting after recovery in the soil coring method.
Using the equation proposed by van Noordwijk et al. (2001) and Mulia and Dupraz (2006), we succeeded in proposing calibration factors varying from X = 8.1, 2.1 to 1.3 for willow, wild cherry and hornbeam, respectively. The calibration factor for wheat was not calculated. As suggested by van Noordwijk et al. (2001), X should differ from 1 for preferentially oriented roots. Our values are not in the range of values proposed by Bennie et al. (1987) or van Noordwijk et al. (1995), but the incertitude remains about the values of SRL that we used: SRL = 15 m g− 1 for hornbeam (Kubisch et al. 2015), SRL = 27 m g− 1 for wild cherry (Weemstra et al., 2020) and SRL = 66.8 m g− 1 for willow (McIvor et al., 2014; Weemstra et al., 2020). Further work is needed to accurately determine the SRL by washing, scanning and measuring root length relative to root mass. The X values that we proposed were close to those of Drew and Saker (1980) except for willow, the X value of which was surprisingly high. Our equation allowed a better prediction of fine root biomass for willow and wild cherry but underestimated fine root biomass for hornbeam. Finally, for our fine root biomass estimation, roots were not washed, contrary to the method advised in (van Noordwijk et al. 2001). Soil particles adhering to fine roots were removed by hand after oven drying, but despite our attentiveness, some might have evaded our vigilance.
Our findings showed that intercropping changes the root distribution of associated plants due to belowground competition. Young trees reduced the crop root abundance and biomass close to tree rows and cohabited with them in the upper soil layer. This was particularly true for willow and hornbeam, whereas wild cherry had already developed deep fine roots under the wheat rooting zone. At 4 years old, different growth rates between species were evidenced with taller and more developed fine roots in the topsoil for willow than for hornbeam in the AC plot. In pure-forest control plots where no fertilizer was applied after planting, tree root development down to 1 m depth was very high compared to trees in the AC plot, certainly due to the lack of nutrients. The core-break method allowed an accurate estimation of crop and tree root abundance but suffered from potential bias in root identification, especially when roots were abundant or when they were already in decomposition at the time of counting. Further studies are needed to determine the evolution of tree and crop root distributions along the soil profile with increasing age. Such studies will indicate if the fine root growth of each tree species will continue in the same direction as observed at 4 years old, or if they will spread in the topsoil as they are getting older, particularly for willow and hornbeam. Analyses of water and nitrogen content could also be investigated to quantify the exact effect of trees on water and nutrient competition.
Acknowledgments
This study was financed by “La Fondation de France”, the French water agency “l′Agence de l′eau Artois – Picardie”, the French Environment and Energy Management Agency or ADEME and the Hauts-de-France Region. We thank Antoine Dequidt for providing us with this experimental plot for this investigation and Fabien Monte for all the agricultural work. We also thank the students of “Institut d’Hazebrouck” for the tree growth measurement. We are grateful to Pierre Saillard and Marianne Boilland for their valuable help during soil sampling in the field and for the very time-consuming root counting and sorting work in the laboratory.
- Abdi E, Majnounian B, Rahimi H, Zobeiri M (2009) Distribution and tensile strength of Hornbeam (Carpinus betulus) roots growing on slopes of Caspian Forests, Iran. J For Res 20:105–110. https://doi.org/10.1007/s11676-009-0019-x
- Andrianarisoa KS, Dufour L, Zeller B, Bienaimé SS, Zeller B, Dupraz C (2016) The introduction of hybrid walnut trees (Juglans nigra × regia cv. NG23) into cropland reduces soil mineral N content in autumn in southern France. Agrofor Syst 90:193–205. https://doi.org/10.1007/s10457-015-9845-3
- Andrianarisoa KS, Zeller B, Ranger J, Bienaimé S, Dambrine E (2017) A reciprocal soil exchange experiment highlights tree root colonization effects on soil nitrification. Soil Use Manag 33:45–53. https://doi.org/10.1111/sum.12326
- Andrianarisoa S, Delbende F, Pruvot C, et al (2019) Agroforestry in the Hauts-de-France - A Research and Demonstration Experimental Site in Ramecourt. In: 4th World Congress on Agroforestry, 20-22 mai. Montpellier, 2019, Montpellier, France, p 1
- Battie-Laclau P, Taschen E, Plassard C, et al (2020) Role of trees and herbaceous vegetation beneath trees in maintaining arbuscular mycorrhizal communities in temperate alley cropping systems. Plant Soil 453:153–171. https://doi.org/10.1007/s11104-019-04181-z
- Bennie ATP, Taylor HM, Georgen PG (1987) An Assessment of the Core-Break Method for Estimating Rooting Density of Different Crops in the Field. Soil Tillage Res 9:347–353
- Bergeron M, Lacombe S, Bradley RL, Whalen J, Cogliastro A, Jutras MF, Arp P (2011) Reduced soil nutrient leaching following the establishment of tree-based intercropping systems in eastern Canada. Agrofor Syst 83:321–330. https://doi.org/10.1007/s10457-011-9402-7
- Bibalani G, Bazhrang Z, Mohsenifar H, Shibaei N, Joodi L (2008) Effect of Prunus avium roots on river bank strong. Pakistan J Biol Sci 11:1126–1131
- Bouttier L, Paquette A, Messier C, Rivest D, Olivier A, Cogliastro A (2014) Vertical root separation and light interception in a temperate tree-based intercropping system of Eastern Canada. Agrofor Syst 88:693–706. https://doi.org/10.1007/s10457-014-9721-6
- Cardinael R, Mao Z, Prieto I, Stokes A, Dupraz C, Kim JH, Jourdan C (2015) Competition with winter crops induces deeper rooting of walnut trees in a Mediterranean alley cropping agroforestry system. Plant Soil 391:219–235. https://doi.org/10.1007/s11104-015-2422-8
- Cheng X, Bledsoe CS (2004) Competition for inorganic and organic N by blue oak (Quercus douglasii) seedlings, an annual grass, and soil microorganisms in a pot study. Soil Biol Biochem 36:135–144. https://doi.org/10.1016/j.soilbio.2003.08.024
- Clivot H, Petitjean C, Marron N, Dallé E, Genestier J, Blaszczyk N, Santenoise P, Laflotte A, Piutti S (2020) Early effects of temperate agroforestry practices on soil organic matter and microbial enzyme activity. Plant Soil 453:189–207. https://doi.org/10.1007/s11104-019-04320-6
- Coleman M (2007) Spatial and temporal patterns of root distribution in developing stands of four woody crop species grown with drip irrigation and fertilization. Plant Soil 299:195–213. https://doi.org/10.1007/s11104-007-9375-5
- Drew MC, Saker LR (1980) Assessment of a rapid method, using soil cores, for estimating the amount and distribution of crop roots in the field. Plant Soil 55:297–305. https://doi.org/10.1007/BF02181809
- Duan ZP, Gan YW, Wang BJ, Hao XD, Xu WL, Zhang W, Li LH (2019) Interspecific interaction alters root morphology in young walnut/wheat agroforestry systems in northwest China. Agrofor Syst 93:419–434. https://doi.org/10.1007/s10457-017-0133-2
- Fernández ME, Gyenge J, Licata J, Schlichter T, Bond BJ (2008) Belowground interactions for water between trees and grasses in a temperate semiarid agroforestry system. Agrofor Syst 74:185–197. https://doi.org/10.1007/s10457-008-9119-4
- Germon A, Cardinael R, Prieto I, Mao Z, Kim J, Stokes A, Dupraz C, Laclau JP, Jourdan C (2016) Unexpected phenology and lifespan of shallow and deep fine roots of walnut trees grown in a silvoarable Mediterranean agroforestry system. Plant Soil 401:409–426. https://doi.org/10.1007/s11104-015-2753-5
- Jackson RB, Canadell J, Ehleringer JR, Mooney HA, Sala OE, Schulze ED (1996) A global analysis of root distributions for terrestrial biomes. Oecologia 108:389–411. https://doi.org/10.1007/BF00333714
- Jose S, Gillespie AR, Seifert JR, Pope PE (2001) Comparison of minirhizotron and soil core methods for quantifying root biomass in a temperate alley cropping system. Agrofor Syst 50:161–168. https://doi.org/10.1023/A:1010667921970
- Kubisch P, Hertel D, Leuschner C (2015) Do ectomycorrhizal and arbuscular mycorrhizal temperate tree species systematically differ in root order-related fine root morphology and biomass? Front Plant Sci 6:1–12. https://doi.org/10.3389/fpls.2015.00064
- Kücke M, Schmid H, Spiess A (1995) A comparison of four methods for measuring roots of field crops in three contrasting soils. Plant Soil 172:63–71. https://doi.org/10.1007/BF00020860
- Lehmann J, Peter I, Steglich C, Gebauer G, Huwe B, Zech W (1998) Below-ground interactions in dryland agroforestry. For Ecol Manage 111:157–169. https://doi.org/10.1016/S0378-1127(98)00322-3
- Link CM, Thevathasan N V., Gordon AM, Isaac ME (2015) Determining tree water acquisition zones with stable isotopes in a temperate tree-based intercropping system. Agrofor Syst 89:611–620. https://doi.org/10.1007/s10457-015-9795-9
- Livesley SJ, Gregory PJ, Buresh RJ (2000) Competition in tree row agroforestry systems. 1. Distribution and dynamics of fine root length and biomass. Plant Soil 227:149–161. https://doi.org/10.1023/A:1026551616754
- McIvor IR, Sloan S, Pigem LR (2014) Genetic and environmental influences on root development in cuttings of selected Salix and Populus clones - a greenhouse experiment. Plant Soil 377:25–42. https://doi.org/10.1007/s11104-013-1770-5
- Moreno G, Obrador JJ, Cubera E, Dupraz C (2005) Fine root distribution in Dehesas of Central-Western Spain. Plant Soil 277:153–162. https://doi.org/10.1007/s11104-005-6805-0
- Mulia R, Dupraz C (2006) Unusual fine root distributions of two deciduous tree species in southern France: What consequences for modelling of tree root dynamics? Plant Soil 281:71–85. https://doi.org/10.1007/s11104-005-3770-6
- O’Connor C, Zeller B, Choma C, Delbende F, Siah A, Waterlot C, Andrianarisoa KS (2023) Trees in temperate alley-cropping systems develop deep fine roots 5 years after plantation: What are the consequences on soil resources? Agric Ecosyst Environ 345:. https://doi.org/10.1016/j.agee.2022.108339
- Ravenek JM, Mommer L, Visser EJW, van Ruijven J, van der Paauw JW, Smit-Tiekstra A, de Caluwe H, de Kroon H (2016) Linking root traits and competitive success in grassland species. Plant Soil 407:39–53. https://doi.org/10.1007/s11104-016-2843-z
- Rowe EC, Van Noordwijk M, Suprayogo D, Hairiah K, Giller KE, Cadisch G (2001) Root distributions partially explain 15N uptake patterns in Gliricidia and Peltophorum hedgerow intercropping systems. Plant Soil 235:167–179. https://doi.org/10.1023/A:1011961409353
- Schroth G (1995) Tree root characteristics as criteria for species selection and systems design in agroforestry. Agrofor Syst 30:125–143. https://doi.org/10.1007/BF00708917
- Swieter A, Langhof M, Lamerre J (2022) Competition, stress and benefits: Trees and crops in the transition zone of a temperate short rotation alley cropping agroforestry system. J Agron Crop Sci 208:209–224. https://doi.org/10.1111/jac.12553
- van Noordwijk M, Brouwer G, Meijboom FG, do Rosário Oliveira M, Bengough AG (2001) Trench Profile Techniques and Core Break Methods. In: Smit AL, Bengough AG, Engels C, et al. (eds) Root methods. Springer, Berlin, Heidelberg, pp 211–233
- van Noordwijk M, van Driel W, Brouwer G, Schuurmans W (1995) Heavy-metal uptake by crops from polluted river sediments covered by non-polluted topsoil - II. Cd-uptake by maize in relation to root development. Plant Soil 175:105–113. https://doi.org/10.1007/BF02413015
- Weemstra M, Kiorapostolou N, van Ruijven J, Mommer L, de Vries J, Sterck F (2020) The role of fine-root mass, specific root length and life span in tree performance: A whole-tree exploration. Funct Ecol 34:575–585. https://doi.org/10.1111/1365-2435.13520
- Zamora DS, Jose S, Napolitano K (2009) Competition for 15N labeled nitrogen in a loblolly pine-cotton alley cropping system in the southeastern United States. Agric Ecosyst Environ 131:40–50. https://doi.org/10.1016/j.agee.2008.08.012
- Zhang C, Ren Y, Yan M, Sang T, Li H, Chen Y, Zhang S (2022) Stable isotopes δ18O and δ2H reveal differential water uptake from intercropped maize and soybean soil profiles. F Crop Res 288:108695. https://doi.org/10.1016/j.fcr.2022.108695
- Zhang W, Ahanbieke P, Wang BJ, Gan YW, Li LH, Christie P, Li L (2015) Temporal and spatial distribution of roots as affected by interspecific interactions in a young walnut/wheat alley cropping system in northwest China. Agrofor Syst 89:327–343. https://doi.org/10.1007/s10457-014-9770-x
- Zhang W, Ahanbieke P, Wang BJ, Xu WL, Li LH, Christie P, Li L (2013) Root distribution and interactions in jujube tree/wheat agroforestry system. Agrofor Syst 87:929–939. https://doi.org/10.1007/s10457-013-9609-x
- Zhang W, Xu WL, Ahanbieke P, Wang BJ, Hao XD, Zhu Y, Li LH (2019) Competition for 15N-labeled nitrogen in a jujube tree (Zizyphus jujuba Mill.)/wheat (Triticum aestivum L.) agroforestry system in northwestern China. Agrofor Syst 93:2097–2110. https://doi.org/10.1007/s10457-018-0332-5
No competing interests reported.