Effect of soil quality and planting material on the root architecture and the root anchorage of young hybrid poplar plantations on waste rock slopes

ABSTRACT On mine waste rock slopes, trees with inadequate root development could be prone to uprooting. The anchorage of trees is mainly determined by the architecture of the root systems that drive their mechanical interactions with the soil. The objective of this study was to evaluate the effects of soil quality and of different planting materials on the architecture and resistance to shear stress of root systems of hybrid poplars four years after they were initially planted. The study was conducted in an open-pit-gold mine. A hybrid poplar plantation was established in 2013 on 33% soil-covered waste rock slopes, using a randomised complete block design, that is, 3 replicated blocks × 3 planting materials × 2 soil qualities. The stability of the hybrid poplars (resistance to uprooting) was evaluated using lateral traction tests. Complete excavations were performed to characterise their coarse root (>4 mm) architecture. Results showed no significant differences between treatments in terms of the maximum resistance force to uprooting, which varied between 7142 and 8989 N. After four growing seasons, no significant effects due to soil quality or planting material were observed in the number of lateral roots, mean root diameter, root biomass, aboveground biomass, and shoot/root ratio.


Introduction
In forested areas, the revegetation of mine sites after closure often involves planting trees on waste storage facilities that need to remain geotechnically stable over long periods of time. The resistance of trees to uprooting is, therefore, important to ensuring the stability of these waste facilities, which are often exposed to strong winds, particularly on slopes. Metal mining generates large volumes of solid wastes, and in particular, waste rocks, which consist of the uneconomic material that is extracted to reach the ore body. Waste rocks are usually stored at the surface in the form of piles of several tens of metres high called waste rock piles [1]. Typically, these structures are difficult to revegetate due to their physical and chemical characteristics [2]; specifically, waste rock piles lack the proper physical structures, nutrients, organic matter (OM), and microorganisms to support plant growth [3,4]. The geometry of sites can also be a challenge for establishment of vegetation, especially where there are higher angle slopes and overall elevations. differing planting materials in a hybrid poplar plantation on waste rock slopes (3 H:1 V, 33%) covered with 50 cm of soil. More specifically, this study examines the effects of using different planting materials (whip, cuttings, and bareroot) and soil qualities (50 cm of topsoil or 40 cm of mineral soil +10 cm of topsoil) on survival rates, aboveground growth, root architecture and mechanical resistance to uprooting of planted trees four years after planting. Lateral traction tests of hybrid poplars were conducted until uprooting as well as a complete excavation method to relate anchorage to the root architectural characteristics. The study is based on the following hypotheses: (i) Only bareroots seedlings have roots at planting. Therefore, the growth and survival of trees will be highest for bareroots. (ii) The growth and survival of hybrid poplars will decrease when the topsoil is combined with mineral soils as compared to using topsoil only. (iii) The root architecture of unrooted plants (cuttings and whips) will be less developed (number and diameter of roots; maximum rooting depth) than rooted plants (bareroots) and its development will increase with the diameter of the planting material (bareroots > whips > cuttings). Therefore, the resistance to uprooting of cuttings will be the lowest. (iv) Higher topsoil quantities will foster more developed root systems (in terms of number, length, and diameter of roots), thus, allowing for higher resistances to uprooting forces.

Site description
The study site was located at the Canadian Malartic mine in the Abitibi-Temiscamingue region of Quebec, Canada (48° 08′ 00″ N 78° 08′ 00″ W). Canadian Malartic is an open-pit gold mine (average grade 1 g/t) that exploits an orebody a rate of approximately 55,000t/day. The ore to waste rock ratio at the mine is typically between 2 and 4, and thus mining operations at Canadian Malartic generate significant quantities of solid wastes. The forest vegetation that surrounds the site is mainly comprised of stands of black spruce (Picea mariana), jack pine (Pinus banksiana), and larch (Larix spp.) mixed with white birch (Betula papyrifera) and trembling aspen (Populus tremuloides). The mean annual air temperature is 1.5°C and the mean annual precipitation is approximately 930 mm (Government of Canada, 2015). The average length of the growing season ranges between 120 and 130 days and the mean number of frost-free days is 97 [47].

Experimental design
The experimental design, shown in Figure 1, included 18 experimental plots organised in a randomised complete block design that included three replication blocks and two tested factors (2 × 3 factorial design): 1) three plant types (whips, bareroots, and cuttings of the hybrid poplar MxB 915,318); and 2) two substrates (50 cm of topsoil versus 40 cm of mineral soil +10 cm of topsoil). Each of the 18 experimental plots contained 25 trees (pseudoreplicates) with a spacing of 2 × 2 m. The plots were separated by 4-m-wide buffer zones. Two lines of fast-growing willows (Salix miyabeana Seemen, clone S×64) were planted in the upper half to limit soil erosion and water run-off.

Planting material and growing conditions
The experimental plots were established in May 2013 on 3 H:1 V (33% or 18°) slopes of mine waste rocks covered with 50 cm of soil. The topsoil was a Grey Luvisol soil [47] from a swampy area that was located above the actual open pit. The topsoil consisted of the O and A horizons; i.e., the first 30 cm, which were dark in colour and rich in organic matter (~20% OM content). The mineral soil consisted of the remaining sandy clay that was excavated down to the bedrock after the overburden topsoil had been removed. The soil texture was composed of 42% clay particles, 27% silt particles, and 31% sand particles, and contained ~1% OM. The topsoil and mineral soil were stored for 30-36 months before use in 7-m-high piles with a slope of 2.5:1. One composite sample (consisting of two samples per plot; 0-10 cm depth) was used for chemical characterisation during planting (May 2013). Soil nutrient analyses were conducted on sieved (2 mm mesh), finely ground, ovendried samples (50°C) by the Lakehead University Centre for Analytical Services (Thunder Bay, Ontario, Canada). Total nitrogen (N) and sulphur (S) were determined by the Dumas combustion method (CNS 2000, LECO Corporation, Mississauga, ON), and organic carbon (C) was determined using the thermogravimetric method (LECO TGA, Mississauga, Ontario). A conversion factor of 1.72 was used to convert organic carbon content to organic matter content. Following an HNO 3 -HCl digestion, bulk P, K, Ca, Mg, Na, Al, As, B, Cd, Co, Cr, Cu, Fe, Mn, Ni, Pb, Sr, and Zn concentrations were determined by inductively coupled plasma atomic emission spectrometry (ICP-AES, Vista PRO, Varian Canada, Mississauga, ON). Available phosphorous (P) was determined colorimetrically on sodium bicarbonate extracts of the soils (Olsen 1954). Bulk pH was determined from saturated soil pastes, while soil electrical conductivity was determined in a 1:2 (soil:water) mixture. Soil texture was determined using the hydrometer method (Bouyoucos, 1962) (supplementary materials). All measured soil and waste rock metal concentrations were below relevant regulatory thresholds.
A physical characterisation of the same soil (topsoil and mineral soil) was determined in other studies [7,48,49]. Briefly, these studies showed a greater macroporosity (15-20%) and lower density in the topsoil (0.7-0.9 g·cm −3 ) relative to the mineral soil (macroporosity: 12-16%; density: 1.1-1.2 g·cm-3). Both the topsoil and mineral soil were found to have macroporosities above the 10% threshold that would allow root growth [50]. The density of the two soils is also adequate for root growth [51].
The planting material used in this study was a semi-exotic hybrid poplar clone (Populus maximowiczii A. Henry × Populus balsamifera L. (M × B) -clone 915,318), locally produced by the Quebec Ministry of Energy and Natural Resources (MERN). This poplar selected was based on its rapid growth and establishment from vegetative material. Three planting materials were used ( Figure 2): bareroot plants (mean length above ground = 131 cm and mean diameter = 11 mm); whip plants (mean length above ground = 93 cm and mean diameter = 15 mm), and cuttings (mean length above ground = 33 cm and mean diameter = 11 mm). Whips and cuttings are generally one year-old shoots, while bareroot plants are produced from 12 to 15 cm cuttings that are grown in the field for one growing season. The mean lengths and diameters of the planting materials were initial measurements taken just after planting. These measurements exclude the 20-30 cm of stems/roots buried in the soil.

Measurements, sampling, and analysis
Initial growth measurements (maximum height and basal diameter) were conducted at the planting in spring 2013, as well as at mortality. This inventory was repeated in the fall in 2013, 2014, 2015, and 2016.
According to climate data for 2016, monthly precipitation and temperature over the growing season (May -October) was similar to normal mean values calculated for the past 30 years.

Coarse root observations
In this study, only roots with a diameter >4 mm were considered coarse, structural roots [52]. Observation trenches were dug after four growing seasons (June 2016) under the second tree of the second line from the bottom of the slope, approximately 10 cm from the stem. These 1 × 1 × 1 m 3 trenches were dug with a mechanical shovel ( Figure 3) for each plot (N = 18). The maximum rooting depth was noted for each treatment.

Root extraction
In June 2016 (fourth growing season), the root systems of 18 trees (one tree in the centre of each plot for each treatment), were excavated with a high-pressure water jet using a hydraulic pump (PUMP 2 '' MULTIQUIP, N Series: 2 H-7348). The structure of the root system was photographed, described, and schematised (360° distribution, vertical distribution). The overall description of the root systems was complemented by the following quantitative measurements: the number of main lateral roots (roots with a diameter >10 mm), the number of branches from each main root, the diameter of each main root every 50 cm (until there was no more change in the diameter), the total length of the main roots until reaching a diameter <4 mm, the angles between main lateral roots of the upslope side and the line separating upper and lower slope, and the number of sinker roots (in-depth). Measurements were performed from the collar (origin of all roots) in the direction of growth. Prior to excavation, the stem basal diameter (diameter of the base of the stem) and maximum height were measured for the 18 studied hybrid poplars. For each excavated tree, the aboveground biomass and belowground biomass was measured. The material was oven-dried at 90°C for 48 h for the aboveground parts and for 72 h for the roots. Once dried, the samples were weighed.
A root anchorage index was calculated to integrate some of the key measured root architectural parameters ( Figure 4) for each plot. These included: • Angles between the main lateral roots (A n ): the sum of the angles between each main lateral root on the upslope side and the line separating upper and lower slope was calculated for each excavated tree (one tree per plot, 12 in total). • Diameters of the lateral roots on the upslope side (d; mm): the diameter of each lateral root was measured at the collar of each excavated tree. • The length of the main lateral roots 1 (L; cm): • The maximum rooting depth (roots with a diameter >4 mm) (P; cm): measured for each excavated tree.
Two indexes were calculated and compared: I 1 (cm) normalised by the sum of the diameters of the upslope roots. I 2 (cm 2 ) multiplied by maximum root depth and divided by soil depth.
The concept is based on the balance of forces in a 2-D plane. The following hypotheses were considered: • Hypothesis 1: the traction force exerted on the lower trunk is transferred to the upslope roots as a tensile force. If we assume that the force exerted is in the direction of the y-axis, only the y component of the roots will resist the exerted force. We also consider that everything happens in a 2-D plane. • Hypothesis 2: the downslope roots are in compression and have a negligible impact on the resistance of the roots to the uprooting force exerted by the winch. • Hypothesis 3: for the first index (I 1 ), it is assumed that the influence of the roots is limited to a certain length 1 . It is also assumed that the root diameter has an influence on its ability to resist the tension exerted.

Lateral uprooting tests
In July 2016 (fourth growing season), lateral traction tests were performed for 18 hybrid poplars (one tree in the centre of each plot) using methods adapted from previous studies [53] (Sheedy, 1996;Grouard, 1995). Before the uprooting tests, maximum height and basal diameter of the stem were measured for each tree. The point of attachment was determined by preliminary tests in order to have a significant force on the root systems and not cause curvature of the stems. Our study aimed to characterise the anchoring of young trees with a test where shear was the main mechanical stress applied to the tree. Our objective was not to reproduce the effect of the wind on tree stability but to evaluate the strength of root anchorage associated to differing planting designs. The objective was to compare the anchorage of trees to select the design (s) that minimised the risks of planted trees' uprooting that could affect the integrity of revegetated mine berms. Thus, for the pullout tests performed, the cable used for the lateral uprooting tests was attached as low as possible on the stem (at the base of the tree). Therefore, the response corresponds to the maximum resistance of the root system to a shear force. An automatic winch (WARN Pro Vantage 2500SCE 4200 lb) was used. It was attached to an optimum dynamometer scale model OP 926 (Optima Led Digital Hanging Scale 2000 lb), which was in turn attached to the tree with a sling and chain around the base of the trunk. The force was measured by the dynamometer and the readings were converted to Newtons. A lateral force (0-8811N) was exerted down the slope, parallel to the slope. The applied force (N) and the displacement (cm) of the tree (displacement of the stem compared to the initial state) were noted each minute until the root system was removed from the soil.

Statistical tests
Results from the lateral uprooting tests (maximum resisting force); root architecture observations (mean root diameter, mean number of roots, mean number of root branches); biomass analyses (aboveground and belowground biomass, shoot-to-root ratio); maximum measured rooting depths; and calculated root anchorage indexes were analysed using mixed linear models created with R (ver.3.1.0). The fixed effects were planting material and soil quality, and the random effect was block. An ANOVA with repeated measures was used for tree height, basal diameter, and diameter at breast height. The normality of the response variables and the ANOVA assumptions were verified. The Tukey multiple comparison test was used when an effect was significant. A significance level of 5% was considered for the statistical analyses performed in this study. Correlations between the maximum resisting force and the stability indices (I 1 and I 2 ) were tested with Pearson's correlation analyse.

Growth, survival, and poplar biomass
After four growing seasons, survival was high (95-100%) for all treatments ( Table 1). The maximum height of the trees was lower for the cuttings compared to the other two types of planting materials until autumn 2015 for the topsoil only treatment and autumn 2014 for the topsoil + mineral soil treatment ( Figure 5). However, at the end of the fourth growing season, none of the treatments were statistically different in terms of the aboveground and belowground development of trees (maximum height, basal diameter, diameter at breast height, root biomass, aboveground biomass, aboveground/belowground ratio, aboveground/total biomass and belowground/total biomass; Table 2). No interactions were observed between planting material and soil quality. In the fourth growing season, mean tree height ranged from 327 to 471 cm and the basal diameter ranged from 50 to 78 mm. Aboveground biomass (1791-2672 g) was four times greater than the root biomass (379-566 g).

Root architecture
Four years after planting, there was no significant difference between the topsoil only and topsoil + mineral soil treatments in terms of the main quantitative variables describing the root system structure (Table 3). These included the: mean root number, mean root diameter, maximum rooting depth, and number of sinker roots. Similarly, the planting material had no significant effect on the same variables. There was no interaction between planting material and the quality of soil. Hybrid poplars showed herringbone root structure ( Figure 6) characterised by a dominant vertical axis (main shoot of the planting material that was pulled down 30 cmdeep in the soil at planting for whips and cuttings), with lateral roots distributed on this axis every 10-20 cm. For each tree, the vertical axis diameter was constant vertically. The main roots (lateral roots) were ramified and could extend to lengths of more than 2 m (horizontally). The diameter of these roots varied between 11 and 26 mm at a maximal distance of 1 m from the trunk. The diameter of the lateral roots decreased rapidly after a length of 50 cm.
For all root parameters, the results showed that the two calculated root anchorage indexes did not change significantly with planting material or soil quality. There was no interaction between the two evaluated factors (planting material and soil quality). (Figure 7).

Lateral uprooting tests
For both the aboveground and belowground parameters, soil quality and tree planting material did not significantly affect the maximum resistance to uprooting and there was no significant interaction between the two tested factors. The measured maximum resistance force to uprooting varied between 7351 and 8851 N (Figure 8), with the following mean values for the topsoil treatment  Bareroots 100 ± 0 (a) 100 ± 0 (a) 100 ± 0 (a) 100 ± 0 (a) Mineral soil Whips 100 ± 0 (a) 100 ± 0 (a) 100 ± 0 (a) 100 ± 0 (a) Cuttings 100 ± 0 (a) 100 ± 0 (a) 98 ± 2 (a) 97 ± 2 (a) Bareroots 100 ± 0 (a) 100 ± 0 (a) 95 ± 4 (a) 95 ± 4 (a)   Figure 9 shows the evaluation of the uprooting force relative to the displacement of the tree; all treatment curves had similar uprooting speeds. The trees were generally uprooted after 30 cm of displacement when the force reached approximately 8000 N. The Pearson correlation between root anchorage index 1 (I 1 ) and the maximum resisting force is 0.92 (P = <.0001). The Pearson correlation coefficient between index (I 2 ) and maximum resisting force is 0.89 (P = <.0001).

Discussion
To the best of our knowledge, this trial is one of a kind study that examines the effect of planting material and soil quality on the anchorage of trees planted on waste rock slopes.
Tree stability can be greatly influenced by the distribution of biomass between the aboveground and belowground parts. Contrary to the authors' first hypothesis, there were no statistically significant differences between the six treatments in terms of aboveground biomass, belowground biomass, and root/shoot ratios. Prior experiments by DesRochers and Tremblay [42] compared four planting materials (bareroots, rootstocks, whips, and cuttings) of hybrid poplars in a clayey soil. They observed significant differences in the root/shoot ratio that were attributed to the different planting types, but only in the first growing season. Thus, the difference between rooted and unrooted planting materials can disappear over successive growing seasons [54]. At the end of growth monitoring in the present study (autumn 2016), the trees from cuttings were the same sizes as those from the bareroots and whips, despite their height being much smaller at the beginning of the experiment. This could be explained by a higher growth rate in the cuttings [42,55]. Results from the present study also indicated similar mean diameters for the three tested planting materials after four growing seasons, despite the initial mean diameter of the whips being higher than that of the cuttings and bareroot seedlings. Similarly, previous studies did not find a positive relationship between initial dimensions of plants and growth [56]     the likelihood of rooting success [62; 63]. However, it appears that the effect of carbohydrate reserves on rooting ability depends on the site-specific environmental conditions and on the tree species [58; 64]. The difference in root/shoot ratio has a significant effect on growth because of its implications for water uptake and water loss in trees [65]. The survival and growth of cuttings and whips were as high as bareroot seedlings despite the fact that they needed to develop new roots to access soil resources. Bareroot plants already have a root system at planting and therefore have rapid access to water and nutrients. However, bareroot plant roots could have poor contact with the soil (air pockets between root and soil) [65]. Moreover, bareroot plants have to support large aboveground parts and must regenerate new roots to facilitate water and nutrient uptake. The buried portion of the stem of poplar cuttings and whips can develop roots from their stems when conditions are appropriate [66][67][68]. Other studies have reported varying performances among planting material types [61,[69][70][71]. Additionally, Duddles and Owston [72] indicated that the site conditions could have more impact on growth and survival than the planting material type. The rate of survival and height growth of the tested hybrid poplar clone (26-143 cm. yr −1 ) were similar or superior to those observed for the same poplar clone in plantations on other sites with similar boreal conditions (40 cm. yr −1 [73,74]. It appears that, with the use of a 50-cm-thick layer of overburden soil, hybrid poplar planting could be appropriate for the revegetation of waste rock slopes under both short-and medium-term conditions. Soil quality directly influences tree development and survival. Notably, the presence of organic matter influences many functions in the soil [75; 76]. Moreover, prior studies have shown a positive effect of soil thickness on tree growth and survival [4,49]. Contrary to the authors' second hypothesis, there were no significant differences in the growth and survival of trees when the topsoil (and the organic matter) quantity increased. The development of the hybrid poplar tested in this study was equivalent in 50 cm of topsoil as in a combination of 40 cm of mineral soil and 10 cm of topsoil. This could be explained by the improvement of the quality of the mineral soil after the four years of planting. In particular, the establishment of herbaceous plants can constitute a source of organic matter by providing litter [77; 78].
The anchorage of a tree is controlled by the architecture of its root system, as well as its interactions with the soil. Root architecture has an important influence on the tensile forces mobilised in the roots [79]. In the present study, neither the planting material diameter nor the soil quality influenced the root architecture of the tested hybrid poplar. Both rooted (bareroots) and unrooted plants (cuttings and whips) developed coarse root system with similar architectures after four growing seasons. In particular, there was no difference observed between treatments in terms of the quantitative variables used to evaluate and describe the root system structures. Thiffault [53] measured stability (resistance to winching) of large containerised and bareroot black spruce (Picea mariana) seedlings in the seventh growing season and characterised their root architecture. Their results showed no significant effects due to planting material on the trees' stabilities or root architectures. In the light of the results of the present study, which showed that there was no significant difference between the six treatments in terms of the maximum resistance to the uprooting, the second and third hypotheses were dismissed. Indeed, it is especially the lateral roots located in the first horizons of the soil that intervene in the anchoring [80; 81]. A homogeneous distribution of these roots is essential to form the soil root plate around the stem to ensure anchorage. After four growing seasons, however, the studied poplars were too young to observe a well-developed root-plate. Moreover, during a mechanical stress, if the 360° distribution of the roots is homogeneous, tension will be applied in a homogeneous way, which will increase the tree's resistance to an uprooting force [52].
The establishment of the root architecture during the development of the plant is a complex process [82]. The physical properties of the soil (structure and texture) play a major role in the anchorage resistance of the plant. Depending on the nature and texture of the soil, the anchorage of the plant can be modified [83; 84]. Root development could be affected by constraints encountered in certain soils, such as rocks [84,85]. Research has shown that root systems are more mechanically resistant in clays than in sandy soils [83]. Prior works also show that soil moisture conditions can affect plant stability. However, the effect of the degree of saturation of soil on plant stability depends on the type of soil and remains poorly understood [86]. Contrary to the authors' fourth hypothesis, the results of this study showed a similar effect of both tested soils in terms of tree stability. For the studied hybrid poplar, there were no significant differences between the two soil qualities in terms of the number of roots, diameter of the roots, maximum rooting depth, the number of sinker roots (in-depth). In the present study, it appears that the thickness of the soil above waste rocks, rather than its quality, could have mostly controlled the root development and, therefore, the anchorage of the trees. Indeed, maximum rooting depths were lower than the soil thickness of 50 cm for all treatments (no coarse roots observed in the underlying waste rocks).
In line with the results of the maximum resistance to the uprooting, root architecture, shoot/root ratio, and plant size measurements showed no significant differences due to either planting material or soil quality. This was reflected in the two calculated root anchorage indexes. The two root anchorage indexes are strongly correlated to the maximum resistance to the uprooting. The use of these indexes appears to be an interesting option for evaluating the root anchorage of trees without using lateral uprooting tests. The first index (I1) integrates the 180º distribution of the root system and could allow evaluating the stability of older trees, as well as for species with a branched root system for which anchorage is achieved mainly by the distribution of the roots [87]. The second index (I2) also integrates the vertical distribution of roots and would be more interesting to use for species growing in deeper soils for which anchorage depends mainly on taproots [88,89].

Conclusion
This study demonstrated that the root anchorage of a four-year-old plantation of hybrid poplar was not affected by soil quality (variable topsoil thickness) or by planting material (variable shoot/root ratio) when 50 cm of overburden soil was used to cover waste rock slopes. Results indicated that there was no significant difference between treatments in terms of: the maximum resistance to uprooting, number of lateral roots, mean root diameter, root biomass, aboveground biomass, shoot/root ratio, maximum height, and basal diameter. This finding is interesting as it demonstrates that the performance of cuttings (in terms of stability, root development, aboveground growth, and survival) is as high as that of whips and bareroot plants, while the costs related to their production are lower. Moreover, in the short term, there was no significant difference between hybrid poplars planted on a 50 cm topsoil layer or on a soil layer comprised of 10 cm of topsoil and 40 cm of mineral soil in terms of survival, growth, root development, and tree stability. Therefore, lower quantities of topsoil could be used, thus reducing the challenge that mining companies face in finding sufficient quantities of soil for the revegetation of sites.
In the future, it is believed that these results will be maintained. Four years after planting, the root architecture of hybrid poplars is already defined and, without change due to external environmental factors, the root distribution will only become more complex over time. However, it could be interesting to evaluate the effect of soil quality and planting material on native species and species with different root systems, or to examine greater soil thicknesses over waste rocks. [90][91][92][93][94], [95]; [96]; [97]; [98]; [99]; [100,101]; [102]; [103,104]; Note 1. The length was considered constant with L: 50 cm. According to field observations, the diameter of the main structural roots decreases rapidly after 50 cm, thus it is chosen to consider a maximum length of 50 cm and constant in the calculations. It is considered that the maximum root reinforcement occurs at this level, close to the trunk.