Herbaceous and Woody Root Biomass, Seasonal Changes in Root Turnover, and Arbuscular Mycorrhizal and Ectomycorrhizal Colonization during Primary Succession in Post-Mining Sites

Abstract

Seasonal changes in the biomass and length of fine roots and their growth into ingrowth cores were measured in a chronosequence of post-mining sites represented by 6-, 16-, 22-, and 45-year-old study sites, located on spoil heaps after brown coal mining in the Sokolov coal mining district. The depth distribution of roots differed between herbs and woody species and also with succession age. At the 22-year-old site, the greatest root biomass was found in the fermentation layer (248.9 ± 113.4 g m²) and decreased with depth. In the case of herbaceous root biomass, the greatest root biomass was found in the 16-year-old site (63.7 ± 15.2 g m²), again in the fermentation layer, which decreased with depth. Overall root biomass increased with succession age, reaching its highest value in the 45-year-old site. In younger sites, the root biomass was dominated by herbs and grasses, whereas woody roots dominated in older sites. After one year, the root biomass in ingrowth cores reached up to one quarter of in situ biomass, which would suggest a low turnover rate. However, the difference between the minimum and the maximum value during the course of one year represents more than half of the mean value. Analysis of the number of arbuscules on roots of Plantago lanceolata sown in soil from all succession stages revealed extensive colonization by arbuscular mycorrhizal fungi in early succession (14.2 ± 0.3 mm root⁻¹), decreasing with succession age, and reaching the lowest value in the 22-year-old site (2.4 ± 0.08 mm root⁻¹) before increasing in the oldest site. Colonization of roots by ectomycorrhizal fungi increased with succession age, reaching a maximum in the 16-year-old site. In comparison with the extent of ectomycorrhizal colonization in relation to root length, the greatest length of ectomycorrhiza-colonized roots was found in the 22-year-old site; hence, the pattern was the opposite of the one observed in arbuscular mycorrhiza-colonized roots.

1. Introduction

Open-cast coal mining causes massive destruction of ecosystems, which are either excavated or covered by overburden deposited in large heaps. These newly formed substrates differ remarkably from natural soil in that they have an extreme texture with either too low or too high a clay content, too low or too high pH, high salinity, or increased content of certain toxic compounds [1,2]. Restoration of soil, the soil biota, and the biological functions of the soil are important steps in the reconstruction of functional ecosystems in post-mining landscapes [1,2]. However, when substrate conditions are not extremely severe, natural succession processes typically result in the formation of a functional vegetation cover, and possibly even a forest, within a few decades [3,4,5,6,7,8]. Consequently, vegetation succession in post-mining sites has been the subject of many studies [9,10,11]. However, the vast majority of these studies only consider the aboveground part of plant communities, whereas studies describing root biomass development and its spatial distribution and mycorrhiza colonization are rare despite the large effect of mycorrhiza–plant interaction during succession and ecosystem recovery [12,13,14,15,16]. This is despite the fact that some studies have shown that belowground species traits change during ecological succession [17,18,19]. In particular, interaction between plant-roots-associated mycorrhizal fungi interaction [12,13,14,15,16], including ectomycorrhizal and arbuscular mycorrhiza fungi competition [20,21], plays a principal role in succession of plant and microbial communities. Symbiotic mycorrhizal fungi are the principal link between plants and soil; they help plants to obtain essential nutrients such as P and N [22] and increase plant resistance to water stress [23,24] and some fungal pathogens [25]. Arbuscular mycorrhiza (AM) and ectomycorrhiza (EcM) are the major types of mycorrhizae in temperate zones. Herbs and grasses mostly form AM with fungi from the phylum Glomeromycota, whereas EcM is characteristic for woody plant species and fungi from the phyla Ascomycota and Basidiomycota [22]. A wide range of plants have dual status that supports both AM and EcM fungi; this strategy lowers the cost of seed establishment and brings larger plasticity in a wide range of environmental conditions [26,27,28]. However, studies exploring both AM and EcM fungi are rare [20,29]. Various interactions of plants with soil biota are also likely to strongly modify plant succession in post-mining sites [5,6,7,14,15,16]. Additionally, plants also respond to the development of the soil profile during succession, which is associated with plant litter accumulation and bioturbation activity of the soil fauna [5,30].

Moreover, processes in the rhizosphere are important components of terrestrial carbon and nutrient cycles [31]. Turnover of fine tree roots is one of the major carbon (C) pathways in forest ecosystems [32,33,34], which is currently receiving considerable attention owing to the importance of this C pool for biogeochemical models of C dynamics [32,35]. The turnover rate of fine roots is dependent on the biomass and annual production of fine roots, but different methods of its calculation yield somewhat different results, e.g., [36,37,38,39]. This methodological complexity is one of the reasons why there is still a considerable degree of uncertainty regarding the biomass and turnover rates of fine roots. Despite recent developments of new techniques based on other principles [40,41], the more classical techniques based on evaluating seasonal dynamics of root biomass [34] or direct measurements of root growth into a mesh or root-free soil are still the most frequently used [42,43].

In this study, we used a well-established chronosequence of post-mining sites at the Sokolov study site, which is part of the worldwide network of long-term ecosystem research sites (LTER), where many aspects of plant-community aboveground biomass, nutrient turnover, and soil and soil biota development have been intensively studied. However, a study that would combine root biomass root turnover and mycorrhizal colonization has not yet been done. Together with information already available, such information would represent an important piece of the puzzle that would allow for better explanation of many ecosystem processes such as C dynamic or plant community development, as explained earlier. There are generally very few studies that consider the recovery of root biomass and root turnover during ecosystem development after major disturbances such as landslides, retreat of glaciers, or mining activities [5,6]. In particular, information about the distribution of roots of herbs and woody species during the course of succession, their turnover rate, depth distribution, and distribution of associated AM and EcM symbionts may be essential for better understanding herb–tree competition, which may be useful in the elucidation of mechanisms that determine whether a system is likely to develop toward a forest or grassland type of vegetation.

For addressing these questions in this study, we used two methods for measuring root turnover, namely, the maximum–minimum technique based on evaluation of seasonal dynamics of root biomass, together with direct measurement of root growth into root-free soil to observe if both give similar results.

The aim of this study was to examine the depth distribution of root biomass, root length, root turnover, and associated AM and EcM symbionts for sites of various ages in a well-characterized chronosequence of post-mining sites from which we can benefit from the availability of data on plant communities and soil development [5,20,30], and describe patterns within the chronosequence. These data show that mainly grasses and rarely herbs occur during initial succession, that intermediate succession stages are dominated by shrubs, and that late succession leads to an open forest with a dense herb and grass understory [5,20]. Based on previous research in the same area, we state following hypotheses (labeled as H1–H4). H1—Based on aboveground vegetation data [5,20,44], herbaceous roots dominate in the early stages of succession, intermediate stages of succession are dominated by roots of woody plants, and the later succession is characterized by the co-occurrence of both herbaceous and woody plant roots. H2—Based on the spatial distribution of EcM and AM at intermediate chronosequence sites conducted earlier [16], EcM dominates in the intermediate succession stages, with Salix caprea, AM, and herbs and grasses dominating during the early succession. H3—Based on previous data concerning soil profile development [44,45], the roots of dominant plants in less developed soils are located close to the soil surface, but roots in soil with a well-developed A horizon are distributed more evenly throughout the A horizon. H4—Root turnover decreases with increasing succession age as the proportion of plants with a longer lifespan increases during succession [9].

2. Material and Methods

2.1. Study Sites

The study considered a chronosequence of four 6–45-year-old post-mining plots covered by spontaneously emerged vegetation, located in the Sokolov coal-mining district, Czech Republic (Figure 1). The spoil dumps are composed of tertiary clay material with a pH 8–9 [46]. The prevailing minerals are kaolinite, illite, calcite, and quartz [46]. The surface, because it was not leveled after the heaping process, is characterized by longitudinal depressions and elevations. For this reason, two microhabitats, one in a depression and one elevated, were sampled at each site [45].

The age of stages responds to the time after the heaping process when spontaneous succession started. In the 6-year-old-site, vegetation of herbs and grasses was dominated by Tussilago farfara and Calamagrostis epigeios, respectively, occurring only rarely (see 5 and 9 for more details about the plant community). A dense herb cover with scattered shrubs dominated by Salix caprea occurred in the 16-year-old site. Shrubs with Salix caprea dominated in the 22-year-old site and shade nearly the entire soil surface, resulting in a weak herb and grass cover. When the shrubland eventually developed into a forest, a dense herb and grass cover appeared, mostly on top of the waves [5,7,45]. A tree cover with Populus tremuloides and Betula spp. developed and dominated in the 45-year-old site, where only weak understory grass and herb cover occurred, e.g., with Festuca ovina [44]. Vegetation was patchy in the sense that patches that were covered by woody vegetation had very little understory and vice versa, patched without woody vegetation was covered by herbs and grasses; however, except for the youngest site with its high proportion of bare ground, basically all the area was covered by some vegetation. Refer to [5,20] for a more detailed description of the vegetation development.

There were also apparent changes in soil development along the chronosequence [5,25], which can be summarized as follows. In the 6-year-old site, there was just bare overburden. In the 16-year-old site, by contrast, herb litter accumulated on the soil surface and became partly mixed into the soil by the litter-feeding macrofauna and cryoturbation. In the 22-year-old site, there were a few centimeters of a deep fermentation (Oe) layer on the soil surface, and in the 45-year-old site, the A horizon was developed due to the bioturbation activity of earthworms [5,30]. The site characteristics are summarized in

Table 1. Characteristics of the studied sites based on [5,13,25]. Plot age is expressed in years after the heaping process, soil parameters refer to topsoil 0–10 cm, pH refers to pH in H₂O, C% refer to percentage of soil organic carbon.

Plot Age (Years)	Vegetation	Dominated Species and Their Cover	Soil Development	pH	C%
6	rarely herbs and grasses	Tussilago farfara (20%), Calamagrostis epigeios (5%) seedlings of Salix caprea 10%.	bare overburden	8.5	5
16	dense herbs, scattered shrubs	Salix caprea (20%) Calamagrostis epigeios (15%)	herb litter	7.5	8
22	mostly shrubs	Salix caprea (50%)	thick fermentation; layer A horizon absent	6.8	10
45	mostly trees	Populus tremuloides (50%), Betula pendula (40%).	A horizon 5 to 10 cm	6.5	12

.

The study was conducted from March 2007 until April 2008. Unfortunately, owing to limited access to the 6-year-old site, root growth cores were not evaluated for this site.

2.2. Root Sampling and Ingrowth Cores

Root samples consisted of tree campaign sampling to determine the depth of root distribution in March 2007. The samples were used to determine the seasonal dynamic; thus, the root samples require one year from the end of March 2007 until the beginning of April 2008, which includes the start of the vegetation season (late March–early April). Peak or vegetation season included August 2007 through the end of the vegetation season, October 2007.

Roots were sampled by coring (cores 6 cm in diameter to the depth of 20 cm) to explore and assess the depth distribution of roots. Individual cores were separated into 5 cm layers in March 2007. If a fermentation layer was present on the soil surface, it was also separated as an extra layer and the depth of soil was measured from the bottom of the fermentation layer. The root samples were placed in coolers, transported to the laboratory, and stored at 4 °C until processing, which took place within one week.

To start ingrow cores, 24 holes 12 cm in diameter and 15 cm deep were drilled into the individual sites and filled with root-free soil from the same location in March 2007. Root-free soil was obtained by digging soil from depths of 0–15 cm, hand-sorting larger roots, and sifting soil through a sieve with 2 mm mesh openings. Twenty-four holes were placed to cover all microhabitats in the sites, including six holes on top of the land waves, six at the bottom of depressions, and twelve on the slopes of the waves (six for each aspect).

In August and October 2007 and in April 2008, soil from the central part of eight holes filled with root-free soil was sampled by drilling cores 6 cm in diameter to a depth of 10 cm. At the same sites, a soil sample (6 cm in diameter to the depth of 10 cm) was taken 10 cm from the edge of each root-free soil-filled sample hole. The root samples were placed in coolers, transported to the laboratory, and stored at 4 °C until processing, which took place within one week. These samples were used to assess seasonal dynamics of roots and their production; samples also used to install ingrow cores were used for this purpose, so samples for the seasonal dynamic of root biomass were taken in March 2007, August 2007, October 2007, and April 2008. They were located in two parallel rows perpendicular to terrain waves that occur on the sites. During each sampling occasion, eight samples were taken; these covered an equal number of samples from elevated and depression locations.

2.3. Root Processing

Roots were rinsed with cold running water on a 1 mm screen and divided into herb and woody plant roots. This was enabled by a simple and patchy plant community and a previous collection of all dominant plants with roots, which served as each category, was then further divided into fine (<1 mm) and coarse (>1 mm) roots. Owing to substrate constraints, most of the roots extracted from soil were very small fragments that do not allow morphological separation of absorptive and transport fine roots [47]. However, in some cases we obtained sufficiently larger fragments of the absorptive short-lived roots, which were mostly less than 1 mm. Therefore, this cut was used as a proxy for short-lived fine roots. The roots were stored in a refrigerator and their lengths were measured using a DELTA T Image Analysis System (England). After the measurement of length, the roots were dried for 48 h at 60 °C and weighed. Specific root length (SRL) was calculated for fine roots by dividing root length by root dry mass.

2.4. Mycorrhizal Colonization Potential

We conducted a pot experiment to evaluate the arbuscular mycorrhiza colonization potential [21,48] of individual types of soil along the succession gradient. Soil for the experiment was collected from each site in August 2007 and sifted through a sieve with 5 mm mesh openings. The soil was loosely packed into 140 cm³ plastic pots. Many studies that examine mycorrhiza inoculation potential use maize [21,48]; we wanted to use the same plant that could occur in all the sites. Plantago major, although never dominant, was present in all the sites where mycorrhiza inoculation potential was studied; we therefore knew it would grow in all the sites. Eight pots were prepared for each plot. Six seeds of Plantago major from a commercial seed provider (Planta Naturalis) were planted in each pot. These pots were incubated in a climatic chamber from 19 September 2007 to 29 February 2008 with a photoperiod of 14 h of light (500 W·m⁻² light intensity) at 20 °C and 10 h of the dark at 15 °C. After the experiment, seedlings of Plantago were harvested and their roots were cleaned with KOH and stained with trypan blue to visualize AM colonization. Three root sections from each of the nine plants from each site were then examined for the presence of mycorrhizal arbuscules. In each root, the number of arbuscules in a 1 cm section of the root was counted under a stereomicroscope.

The proportion of roots infected by ectomycorrhizal fungi (EcM) fungi was quantified in fine woody roots sampled in March 2007. The proportion of roots colonized by EcM fungi was determined by the intercept method. A set of lines was projected across fine woody roots and at each intersection of a root with this line, we recorded whether the root was or was not colonized by an EcM fungus; this was done until the 100th intercept for each sample was attained. The result is expressed as the percentage of root length colonized by EcM fungi. To obtain absolute numbers, we multiplied the percentage for each sample (divided by 100) by the length of fine woody roots in the given sample, which gave us the length of EcM-colonized roots.

2.5. Data Evaluation and Statistics

Root biomass and length were compared between individual sites for individual categories of roots using one-way ANOVA. If that returned significant results, Fisher’s post hoc test was applied. One-way ANOVA and Fisher’s post hoc test were also applied to test for differences between sites using arbuscular density, root length, the proportion of EcM fungi colonization, and the length of roots colonized by EcM fungi. Because of the large spatial variability of roots, the effect of season on root biomass was studied by two-way ANOVA, where microhabitat and sampling time were used as the dependent variables. Paired t-test was applied to compare biomass for individual months with biomass at the start of observation (in March 2007). In these t-tests, biomass ascertained from a sample taken at the very same spot was used for comparison. All statistical computations were carried out using Statistica 13.0.

Root production was estimated from the seasonal dynamics of roots using the maximum–minimum method; the maximum–minimum (MM) formula calculates the annual fine-root production (Pa) by subtracting the lowest biomass (Bmin) from the highest biomass value (Bmax), regardless of other biomass values recorded during a full year [49]. Turnover was estimated as production divided by mean biomass calculated as an average of all biomass values recorded during the year [32].

3. Results

3.1. Depth Distribution of Roots

Based on depth distribution sampling in March 2007, the distribution of roots with depth noticeably differed among the individual study sites (Figure 2). In the 6-year-old site, roots occurred only in the top 10 cm of soil; however, finer roots seemed to be equally distributed in the 0–5 and 5–10 cm layers. In the 16-year-old site, herbaceous roots dominated. They had the greatest biomass in the 0–5 cm layer and their biomass decreased with depth. Fine woody roots had much lower biomass and appeared to have the opposite pattern of depth distribution. In the 22-year-old site, fine woody roots dominated and had the greatest biomass in the fermentation layer, and their biomass decreased with increasing depth. Coarse woody roots reached a maximum in the 10–15 cm layer. In the 45-year-old site, there were more woody roots, both coarse and fine, than herbaceous roots. In this site, all root types were more or less equally distributed among the soil depths studied (Figure 2).

3.2. Root Biomass and Length Variation among Succession Age and Season

Based on four sampling occasions during the season, at depths from 0 to 20 cm, pooled fine herbaceous roots had significantly greater biomass and length in the 16-year-old site than in the 22- and 45-year-old sites (

Table 2. Mean root biomass and root length in 16-, 22-, and 45-year-old post-mining sites ± SEM, based on all sampling occasions (n = 24 for each site) in the whole soil profile measured. Root production was estimated by the maximum–minimum method, and root turnover was calculated as production divided by mean biomass. Statistically homogeneous groups of sites are marked by the same letter (one-way ANOVA, LSD post hoc test).

Root	Root Mass g m−2
Type	16-Year-Old Plot		22-Year-Old Plot		45-Year-Old Plot
Herb fine	28.1	±3.5 b	0.3	±3.6 a	6.2	±3.6 a
Herb coarse	5.7	±3.6	5.5	±3.7	7.1	±3.7
Woody fine	5.4	±22.6 a	160.3	±23.2 b	109.3	±23.2 b
Woody coarse	7.5	±16.7 a	38.6	1 ± 7.1 a	97.4	±17.1 b
	Root mass production (maximum minimum method) g m−2 y−1
Herb fine	22		n/a		5
Woody fine	14		97		47
	Root turnover rate y−1
Herb fine	2.59		n/a		0.80
Woody fine	0.78		0.60		0.43
	Root length m m−2
Herb fine	1473	±202 c	31	±207 a	785	±207 b
Herb coarse	18	±29	27	±29	61	±29
Woody fine	11	±35.6 a	2398	±364 c	1274	±364 b
Woody coarse	28	±19 a	53	±19 ab	102	±19 b
	Specific root length m g−1
Herb fine	52.4	±2.2	103.3	±22.3	126.6	±105.8
Woody fine	2.0	±1.4 a	15.0	±4.2 b	11.7	±3.8 b

). The biomass and length of coarse herbaceous roots did not differ significantly between the sites. Fine woody roots in the 22- and 45-year-old sites had significantly greater biomass than in the 16-year-old site (Table 2). The same applied to root length, but the 22-year-old site also had samples with significantly greater root length than those collected from the 45-year-old site (Table 2). Coarse woody root biomass increased with increasing plot age, and the 45-year-old site was significantly greater than in the other two sites (Table 2). The length of coarse roots also increased with succession age; the samples from the 45-year-old site had significantly greater root length than those collected from the 16-year-old one, but the 22-year-old site did not differ significantly from either the younger or the older site (Table 2). Specific root length was larger for herbs than for trees. In trees, specific root length was significantly smaller in the youngest site (Table 2).

Looking at seasonal dynamics (Figure 3), two-way ANOVA comparing months and microhabitats did not find any significant effect of the month on any of the sites. For the 22-year-old site, we found a significant effect of the microhabitat, with root biomass in depressions between the land waves being significantly greater than in all the other habitats (data not shown). However, using paired t-tests, some significant differences were found between March 2007 and some other months. In the 16-year-old site, we observed a significant (paired t-test) monotonous increase in herbaceous root biomass during the observation period (Figure 3a). The 22- and 45-year-old sites, which were dominated by fine woody roots, exhibited a bell-shaped seasonal dynamic of root biomass, with a minimum in spring and a maximum in summer or autumn (Figure 3b,c); however, only in the case of the 45-year-old site was there a difference between March 2007 and the maximum value in October 2007.

3.3. Ingrowth Cores

Root biomass in ingrowth cores was found/observed in all cases, with the only exception of herbaceous roots in the 45-year-old site, significantly lower than corresponding root biomass outside the core (paired t-test) (Figure 3). In the 45-year-old site, root herb biomass in the cores reached values found in the soil surrounding the cores. At all the other sites, the percentage of both herbaceous and woody roots found in ingrowth cores reached 10–20% of those in the surrounding soil; the maximum was 29% in the case of fine woody roots in the 22-year-old site in April 2007. In all cases, the proportion of roots in ingrowth cores to root biomass in the surrounding soil established in April 2007 apparently did not increase over time.

3.4. Mycorrhizal Colonization Potential

The density of AM arbuscules found in the roots of Plantago plants sown at the different sites was significantly greater in the 6-year-old site than at the all the other sites, and significantly lower in the 22-year-old site than in the all the other sites, with the 16- and 45-year-old sites having an intermediate value (one-way ANOVA, LSD post hoc test) (Figure 4a). The proportion of fine woody roots infected by EcM fungi was the lowest in the youngest, 6-year-old site; in all the older sites the proportion of EcM colonization was significantly greater with no significant difference between the older sites (one-way ANOVA, LSD post hoc test) (Figure 4b). However, when the data on the proportion of infected roots were combined with root length data (Figure 4c), the length of roots infected by EcM fungi (Figure 4d) showed an inverse pattern to that exhibited by the occurrence of AM fungi (Figure 4a) among the sites. The length of roots infected by EcM fungi in the 6- and 16-year-old sites was significantly lower than in the 22- and 45-year-old sites (one-way ANOVA, LSD post hoc test).

4. Discussion

The data show apparent vicariant distribution between woody and herbaceous roots and their mycorrhizal symbionts during succession. AM are suppressed particularly in intermediate stages of succession, which correspond with a high density of fine woody roots and EcM (Figure 4). This can be driven by the development of plant communities, as taller woody vegetation may, during succession, become more successful in competition for light and outcompete herb vegetation, which as a consequence may result in replacement of AM by EcM. However, previous studies conducted at the same sites suggest that competition of woody roots and associated EcM against herb roots and associated AM may substantially contribute to this vicariant distribution. Namely, the results of [21], who observed an increase in herb growth and AM inoculation potential after woody roots in these sites were isolated by trenching [9,16], also observed an apparent absence of herbaceous roots and AM in patches dominated by woody roots and EcM [16]. A negative relationship between AM and EcM has also been observed in other ecosystems. For example, some studies [50,51] found that EcM willows were negatively influenced by the AM colonization of understory plants via their EcM fungal partners and leaf litter deposition in a meadow–forest transition zone of an alpine ecosystem. In dual hosts using EcM, plants may be benefitted more when major nutrients occur preferentially in the organic form [26]. This can be expected in intermediate succession stages in our sites where most nutrients are partly decomposed litter on the soil surface (Oe horizon). On the other hand, it has been shown that the presence of AM understory shrubs significantly decreased the productivity of EcM trees [51,52] and colonization by EcM fungi [7,53]. These interactions between EcM and AM fungi are more likely to cause this effect rather than changes in soil chemistry, as this effect occurred in 22-year-old sites densely covered by Salix caprea but not on younger or older sites, while chemistry on the 22-year-old site was intermediate between older and younger sites (Table 1). However, data available in this study do not really allow us to judge to what extent changes in root and mycorrhiza colonization just follow succession changes in plant community and to what extent they drive these changes. Moreover, both explanations are supportive and not exclusive as plants which are stronger aboveground can better support root growth and mycorrhiza symbionts, and plants which monopolize more belowground resources may better compete aboveground. In addition to development of plant communities, soil development may also play role as soils in our sites are characterized by high pH (8.5) and a low amount or recent organic matter. Previous studies [5] show that during soil development soil pH decreased and the content of soil organic matter increased (Table 1). Data from the literature [54] suggest that ectomycorrhizal fungi are associated with sites with lower pH and higher soil organic matter content.

Not only root density and density of mycorrhizal symbionts, but also their depth distribution indicate competition between herbaceous and woody plants. The depth distribution of roots and the pattern of herbaceous root competition may be affected by the development of the soil profile. In soils where the A horizon is not developed, the prevailing type of roots grows close to the soil surface. Results reported in [55] describe differences in the vertical distribution of roots between different tree species, with early-succession species being more evenly distributed in the soil profile and late-succession species being concentrated more toward the surface. In our study, the woody species present were the same along the chronosequence [5,9], but we observed remarkable differences between the sites. In the 6-year-old site, fine woody roots were distributed evenly with depth, apparently due to the fact that the substrate created by heaping was well-mixed and rather homogeneous vertically [56]. Similarly, a homogeneous distribution of roots with depth was found in the oldest site. We assume that this is because of earthworm bioturbation, which mixed the soil and homogenized the soil properties through the depth profile of the topsoil [30]. Rather surprising is the absence of fine herd roots in the 6-year-old site. This may be connected with fact that samples were taken in March, which is when Tusilago just overwinters using thick roots but had no leaves. Probably most of the fine roots die during winter and new ones begin to be produced just later in the season when leaves appear. In the 22-year-old site, most roots were concentrated in the fermentation layer, which concentrated most nutrients [42], and where roots can benefit most from the ability of EcM to obtain nutrients from decomposing litter [57]. In the 16-year-old site, the biomass of woody roots increased with depth, most likely in response to herbaceous root competition. This suggests, in agreement with [58], that local site conditions, namely the distribution of organic matter and nutrients in the soil profile and macropore availability [59], may have a stronger influence on the vertical distribution of roots than the tree species alone.

Total root biomass and root length were found to increase over the course of succession on sand dunes in northwestern China and in Brazil’s Atlantic Forest [19,60]. In our study, we found a gradual increase in overall root biomass, although the length of roots reached a maximum in the intermediate succession stages and later decreased slightly. This corroborates the results of [61], who found that root length peaked in the forest stage of intermediate age.

About 20–30% of the biomass found in the surrounding soil occurred in ingrowth cores, which would suggest a much lower turnover rate than estimates based on sequential sampling. This discrepancy between the two methods used to estimate turnover rates in this study might be caused by the alternation of the substrate during the preparation of the root-free soil. In our study sites, the overburden is formed of clay mudstones, which in the first stage of weathering disintegrate into sheetlike lamellae, giving them the appearance of a book or office file consisting of many pages. Roots then penetrate the soil through larger macrospores in gaps between leaves. By sieving the soil, we made the environment more isotropic, which may slow root penetration.

Looking at the data obtained by our sequential sampling, root turnover values are in a range given by worldwide review for temperate ecosystems [34]. The turnover of herbaceous roots seems faster than that of woody roots. This may appear surprising following a global review of root turnover [34]. However, our case does not consider that some of the herbs and grasses are annuals, which renew their roots every year. This is particularly true in early succession stages with an ephemeral plant. Additionally, root turnover seems to decrease with succession age. This is possibly connected with the fact that in early succession there is much space for roots to colonize. Moreover, the rate of root turnover in fact includes not only the renewal of existing root networks but also the growth and colonization of new space, which does not happen in more mature sites.

Specific root length in woody species was in the range of data provided by [62], which showed lower values than those found in the youngest site. Here, very few woody roots were extracted, possibly affecting the results. Furthermore, the site is alkaline, which may play a role as SRL was shown to decrease with increasing pH [62].

5. Conclusions

In conclusion, we observed an increase in root biomass with succession age. Herbaceous roots dominated in early succession whereas woody roots formed most root biomass in later succession stages. Consequently, the mycorrhizal potential for colonization by AM and EcM fungi shows a vicariant pattern. AM colonization potential was the highest in the young site and the lowest in the 22-year-old site, whereas EcM colonization reached a maximum in the 22-year-old site. This indicates that this age can be very suitable for introducing EcM trees into succession sites.