Introduction

The growing costs of production, the decreasing quality and productivity of soils due to intensive agricultural use, and/or their contamination with various substances constitutes major reasons for the abandonment of agricultural areas in many regions, including Europe (Kurganova et al. 2014). Afforestation has commonly been used to protect abandoned soils against further negative changes (Kilian 1998). Moreover, the practice can also be applied for carbon sequestration (Paul et al. 2002; Miripanah et al. 2019) and to increase the potential to provide various ecosystem services (Pèrez-Silos et al. 2021). The spontaneous development of vegetation can also be observed on non-afforested areas. This is a multi-stage process that follows a typical pattern of species replacement. Annual plants are typical of early colonisation, and these being replaced by perennial species in the next stages (Corbet 1995; Wilcox 1998). The process of succession is controlled by various ecological mechanisms that are involved in three alterative models––facilitation, tolerance and inhibition (Connell and Slatyer 1977). The species composition of vegetation during this process depends on a complex of environmental factors, particularly the soil nutrient status and pH, the water regime, climate and the relief (Kilian 1998). Some species, such as silver birch, grey alder and common aspen, are characterised by a high succession potential due to their tolerance of a broad array of environmental conditions, their few site requirements, rapid growth and large seed fall (Špulák et al. 2010).

Significant changes have been recorded in soils during ecological succession (Mielnik et al. 2021). Root growth typically increases soil porosity, water filtration and water holding capacity, whereas the bulk density decreases (Osman 2013; Yan et al. 2016). Following this, the topsoil becomes better aerated (Wei et al. 2013). During the succession, there is usually an increase in the total organic carbon (TOC) content and sometimes also certain biogenic elements (Matamala et al. 2008; Kalinina et al. 2011; Spohn et al. 2016). However, the effect on the soil nutrients, soil organic matter (SOM) and indicators of the soil’s ecochemical state (pH, sorption and buffering properties) strongly depends on the plant species (Augusto et al. 2002). The uptake of elements by roots, their accumulation in the biomass and their return to the soil via throughfall, stemflow and litterfall are considered to be major mechanisms that influence the soil cover during the succession process (Nilsson et al. 1999; Parzych et al. 2010; Jonczak and Mackiewicz 2012).

Silver birch (Betula pendula Roth) is widely distributed in forests of the Northern Hemisphere and has a large succession potential (Oikonomakis et al. 2000). It is often the main tree species on abandoned land undergoing natural regeneration (Špulák et al. 2010; Zasada et al. 2014; Franiel and Kompała-Bąba 2021). Therefore, it has a large potential to influence the soil. Jonczak et al. (2020) reviewed the current knowledge on the effects of birches on certain soil components, characteristics and processes. The authors concluded that several aspects related to soil–birch interactions are still poorly understood and require further exploration. Generally, aspects related to forest management and biomass productivity are more frequently represented in the literature than those related to soil–birch ecological feedback processes. The effect of birch on SOM quality and quantity is among the least-explored topics. Several studies have focussed on SOM stocks and their major characteristics. Rosenvald et al. (2011) reported increasing SOM contents in soils with stand age. This observation confirmed the findings of an earlier study performed in Latvia by Daugaviete et al. (2003). An increase in SOM stocks after the introduction of birch has also been reported by Holubík et al. (2014) for the Czech Republic. There are also comparative studies available that cover birch and other tree species stands. For example, Fernández-Núñez et al. (2010) reported a greater SOM accumulation in the soil under birch than under pine. Conversely, Kirby and Potvin (2007) and Ma et al. (2015) reported lower SOM contents under birch stands than under conifers. Jonczak (2013) found lower SOM stocks in post-arable soils afforested with silver birch when compared to arable field, meadow and fallow soils and black alder stands. Data on the impact of birch on SOM quality are scarce. However, some positive effects have been reported by Gawęda et al. (2018). Smolander and Kitunen (2002) recorded higher dissolved organic carbon (DOC) contents in the soils of birch stands when compared to other tree species. This observation can be explained by the positive effect that birches have on soil microorganisms (Priha and Smolander 1997, 1999; Saetre et al. 1999), and the higher degradability of DOC in the litter layer of birch stands compared to other tree species (Kiikkilä et al. 2006). The fractional composition of SOM and the internal structure of humic substances have not been studied in any detail in birch stands. To summarise, the available knowledge on the effects of birch on SOM quantity and quality is insufficient and inconsistent.

Considering the importance of the issue, we initialised a broad study focussed on silver birch–soil interactions in post-arable stands. The study covered a wide spectrum of issues, with a particular emphasis on SOM. In this study, we evaluated the quantitative and qualitative changes of SOM in the chronosequence of a silver birch succession on poor quality, and sandy (Brunic Arenosol) soils that had been abandoned after agricultural production. The study was based on five plots, constituting a typical sequence of spontaneous vegetation development, including arable field and fallow land and three birch stands aged 8, 33 and 40 years.

Materials and methods

Study area

The study was conducted in central Poland (51.872 N, 20.297 E), representing a cold climatic zone, with a warm summer, humid and continental climate, according to the Köppen–Geiger classification (Peel et al. 2007). The climate is characterised by favourable solar conditions against the background territory of Poland. The mean annual temperature was 8.3 °C, and the mean annual sum of precipitation was 538.3 mm for Skierniewice station during the period 1951–2019. The warmest month was July, with a mean monthly air temperature of 18.6 °C, whilst the coldest was January, with a mean air temperature of −2.3 °C. The annual distribution of precipitation was characterised by a large bias between the warm (347.8 mm) and cold (190.5 mm) half-year. July was the month with the highest sum of precipitation, (81.7 mm) and January had the lowest (26.3 mm).

The study covered a sequence of five stands, located in the same complex of Brunic Arenosols developed from fluvioglacial sands, including arable field (AF), fallow (FA) and 8 year-old (YS), 33 year-old (MS) and 40-year-old (OS) successions of silver birch. Tree age was taken from Forest Database developed by National Forests. The birch stands differed in terms of their basic characteristics (Table 1). The highest density and lowest diameters (i.e. diameter at breast height and DBH) were typical of the YS. Tree density decreased, and DBH increased with stand age. The admixtures of black cherry (Prunus serotina Ehrh.), common aspen (Populus tremula L.) and grey alder (Alnus incana (L.) Moench) that were observed are typical of the spontaneous development of vegetation.

Table 1 Basic characteristics of the stands
Full size table

Soil sampling and analysis

The study was performed in 2019. The soils were sampled from 20 × 25 m plots. One soil profile was described from each plot to provide a basic characterisation of the soils. The soils were described using the Food and Agriculture Organisation (FAO) of the United Nations criteria (FAO 2006) and classified according to the World Reference Base (WRB) system of the International Union of Soil Sciences Working Group (WRB 2015). Then, one disturbed and two undisturbed (100 cm3) samples were taken from horizons. The bulk density and total porosity were determined from undisturbed samples using the gravimetric method. The disturbed samples were air dried and sieved through a 2.0 mm sieve to remove the coarse fraction. Measurements taken on the earth fraction included the particle size distribution by mixed sieve and pipette methods, the pH, evaluated potentiometrically in a suspension of water and 1 mol dm−3 potassium chloride (KCl), and the TOC content and total nitrogen (N) content by dry combustion (Vario MacroCube, Elementar, Germany).

In addition, the O and A horizons were sampled from 20 regularly distributed sub-plots per plot. The O horizon samples were taken as 20 cm diameter cores, dried at 65 °C, weighed and milled into powder. One disturbed and one undisturbed sample of the A horizon were taken from each sub-plot and prepared for analysis using the same procedures as for the samples taken from the soil profiles. The thickness of the A horizon was measured in each location. Laboratory analyses included measuring the bulk density, total porosity, pH, TOC and N contents using the same methods as described above. The SOM content was determined by loss on ignition at 550 °C. The SOM fractional composition was analysed using the Schnitzer procedure (Gonet 1999). Sequential extraction included the determination of fractions extracted during sample decalcification (Cdec), fulvic acids (FAs), humic acids (HAs) and humins (HUs). The contents of all the fractions are given as the percentage of TOC (%TOC) in this paper. The HAs were extracted using the Schnitzer procedure (Gonet 1999), purified using a soft hydrofluoric (HF) and hydrochloric (HCl) acid solution (6 ml 40% HF + 5 ml 38% HCl + 990 ml deionised water), washed with deionised water and freeze-dried. Their irradiation absorbances at wavelengths of 465 and 665 nm were determined in 0.01% alkaline (0.1 mol dm−3 sodium hydroxide [NaOH]) solutions using a Jenway 6105 UV–VIS spectrophotometer, and the E4/6 quotients were calculated based on these data. The elemental composition (C, N and hydrogen [H]) of the freeze-dried HAs was determined by dry combustion (Vario MacroCube, Elementar, Germany). The oxygen (O) content was calculated as 100%–%C–%N–%H. The contents of all the elements are given in atomic percentages relative to the ash-free mass. The ratio of internal oxidation (ω) of the HAs was calculated based on their elemental compositions, following Żdanow (1965). A 13C nuclear magnetic resonance (NMR) analysis, with cross-polarisation and magic-angle spinning (CPMAS), was performed using a Bruker Avance III 400 MHz spectrometer. The chemical shift regions were referred to alkyl C (0–45 ppm), O-alkyl C (45–110 ppm), aryl C (110–160 ppm) and carboxylic C (160–220 ppm), following Wilson (1987).

Statistical analysis

The statistical analyses included the calculation of significance of differences between the means using the one-way analysis of variance (ANOVA), and homogeneous groups were distinguished with the use of Tukey test for significance level equal 0.05. Moreover, principal component analysis (PCA) and cluster analysis were used to investigate the relationship between the variables characterising soils and the multidimensional differentiation of the studied stands. Ward's method and the square of the Euclidean distance were used in the cluster analysis. All analyses were conducted using Statistica 13 software (Dell 2022).

Results

Basic characteristics of the soils

The soils of the studied plots showed features typical of Brunic Arenosols, including the sequence of horizons (A–Bw–C) and the texture––loamy sand or sandy loam in the A horizons (Fig. 1) and sand or loamy sand in the deeper horizons. The thickness of the A horizons varied from 22.8 ± 3.3 cm in the OS to 30.5 ± 2.2 cm in the FA soils (Table 2). The bulk density ranged from 1.26 ± 0.11 g cm−3 in the FA to 1.46 ± 0.08 g cm−3 in the YS soils, whereas total porosity ranged from 43.4% in the YS to 51.1% in the FA soils. The pH was comparable in all the A horizons, changing from 4.4 ± 0.1 to 4.8 ± 0.5 for pH-H2O and from 3.8 ± 0.2 to 4.0 ± 0.1 for pH-KCl. Higher pH values were noted in the O horizons (5.3 ± 0.2–5.8 ± 0.3 for pH-H2O and 5.2 ± 0.2–5.7 ± 0.4 for pH-KCl). The soils were generally poor in TOC (7.3 ± 1.2–11.9 ± 2.0 g kg−1) and N (0.57 ± 0.07–0.97 ± 0.13 g kg−1) in the A horizons. The TOC: N ratio ranged from 11.6 ± 0.6:1 in the AF soils to 15.6 ± 1.6:1 in the MS soils. In the O horizons, it ranged from 31.5 ± 6.3:1 to 35.3 ± 4.6:1. The differences between the plots of TOC and N contents and the TOC:N ratios were statistically significant in the majority of cases (Table 2).

Fig. 1
figure 1

Texture in the A horizons of the studied soils

Full size image
Table 2 Basic characteristics of the O and A horizons of the studied soils (mean ± SD, n = 20, the same letters in the rows represent homogeneous groups according to Tukey's procedure at the significance level of 0.05)
Full size table

SOM content and fractional composition

The SOM content in A horizons ranged from 15.68 ± 1.92 to 22.90 ± 2.47 g kg−1, being the lowest in the YS and the highest in the AF soils (Table 3). Generally, the SOM content was comparable in the AF, FA and OS plots. In the birch stands, there was a positive correlation with stand age. When compared to the SOM stock, the highest values were noted in the AF and FA soils, whereas the lowest were in the birch stands. Differences between the SOM in the birch stand soils were low and usually statistically insignificant, with 89.4–91.3% of the SOM in the forest soils being allocated to the A horizons and 8.7–10.6% occurring in the litter layers.

Table 3 Mean ± SD SOM stocks and major SOM fraction contents (as %TOC) in the O and A horizons of the studied soils (n = 20, the same letters in the rows represent homogeneous groups according to Tukey's procedure at the significance level of 0.05)
Full size table

The humin fraction (HUs) predominated amongst the SOM components in all the studied soils (Table 3). In the O horizons, it constituted 76.9 ± 1.8–79.6 ± 2.2% of the TOC, whereas in the A horizons, and this was 63.1 ± 3.5–76.4 ± 2.9% of the TOC. The HUs contents in the A horizons increased from the AF to the MS and then, decreased with succession age. Differences between the stands were statistically significant in majority of cases. The FAs occurred in amounts of 9.3 ± 1.2% TOC to 10.6 ± 1.5% TOC in the O horizons and 13.7 ± 1.8% TOC to 26.0 ± 4.0% TOC in the A horizons. In the A horizons, the FAs contributed most to the OS and AF soils and much less to the soils of the remaining stands. As for the previous fraction, differences between the stands in terms of FAs content were usually statistically significant. The HAs constituted 8.7 ± 1.2–11.8 ± 1.1% TOC of the O horizons and 7.7 ± 1.2–13.5 ± 1.3% TOC of the A horizons. Significantly, lower HA contents were noted in the forest soils than the AF and FA soils. The Cdec fraction occurred in its lowest amounts in both the O and A horizons, mostly constituting around 2% TOC. Only in the A horizons of the OS soils did it occur in higher amounts (Table 3).

Elemental composition of the HAs

The major component of the HAs was H followed by C (Table 4). The H content fluctuated around 40 atomic % and showed relatively low variability amongst the stands and soil horizons. The C showed higher variability, with its content in the O horizons ranging from 31.91 to 33.65 atomic %, whereas in the A horizons, and it was 28.60 to 33.64 atomic %. The lowest C contents were recorded in the AF and OS soils, whereas the highest were in the FA soils. Considerable differences in the N content were found between the O and A horizons, and these values being 1.25–1.37 and 2.06–2.46 atomic %, respectively. The HAs richest in N were from the A horizons of the YS soils, whereas the poorest were from the O horizons of the MS soils. The O content in the HAs from the litter layer increased with stand age. In the A horizons, O showed the general tendency AF > FA > YS < MS < OS with homogenous two groups FA, YS, MS and AF, FA, OS. The similar tendency was noted for the ω developed by Żdanov (1965), where two homogenous groups were determined AF, FA, YS, MS and AF, FA, OS (Table 4).

Table 4 Mean ± SD of elemental content, H:C ratios and HA internal oxidation values in the O and A horizons of the studied soils (n = 5, the same letters in the rows represent homogeneous groups according to Tukey's procedure at the significance level of 0.05)
Full size table

HAs UV–visible spectroscopy and 13C-NMR

The E4/6 quotient is commonly used as an indicator of soil humus quality and maturity. In this studied soils, it varied significantly in the O (7.1 ± 0.3–8.0 ± 0.3) and A (4.5 ± 0.2–4.9 ± 0.3) horizons (Table 5). The tendency observed is typical. The 13C-NMR analysis indicated a predominance of aliphatic (alkyl, O-alkyl) over aromatic and carbonyl structures in all the studied soils (Table 6). The aliphatic structures comprised 57.1–67.9%, aromatic 21.3–26.3% and carbonyl 8.4–18.4% C. Some differences in the internal structure of the HAs were noted amongst the stands. Typically, the most clearest differences were recorded between the O and A horizons. The greatest variability was noted in the spectral regions 0–45, 90–110, 160–190 and 190–220 ppm. The HAs extracted from the O horizons were generally richer in aliphatic and poorer in carboxylic structures than those from the A horizons. There were also some clear tendencies when comparing the HAs from the A horizons amongst the stands. The transformation from AF to FA and then, YS to MS soils showed increasing tendencies in the 0–45 and 190–200 ppm spectral regions and decreasing tendencies in the 45–60, 60–90 and 90–110 ppm regions.

Table 5 Mean ± SD of E4/6 ratios in humic acids extracted from the O and A horizons of the studied soils (n = 20, the same letters in the rows represent homogeneous groups according to Tukey's procedure at the significance level of 0.05)
Full size table
Table 6 Results of the HA 13C-NMR analysis
Full size table

The characteristics of the studied stands based on selected variables

Fig. 2 shows the results of PCA analysis. PCA was carried out for all studied stands in the O and A horizons for the following variables: fractional composition (contents of Cdec, HAs, FAs and HUs expressed as a % of TOC), the E4/6 quotient of HAs, elemental (C, H, N and O) composition of HAs and the content of basic structures (Calip, Carom and Ccarb) based on 13C-NMR analysis. The first two components provide 72.92% information about the variability of the data set under consideration. The differences between stands in relation to the horizontal axis (PC1) are the most important. This means that the O and A horizons introduced the greatest differentiation of stands. Calip, E4/6 and HUs were strongly positively correlated, and all the O horizons (YS–O, MS–O and OS–O) had high values of these variables. They were strongly negatively correlated with variables FA, HAs component O and N and Ccarb which were very high in the OS–A stand as well as in the rest of A horizons of stands. The HAs component C was strongly negatively correlated with Cdec and FAs, which were very high in the OS–A stand. The FA–A stand had a very high value of HAs, and low values of HAs component H, thus, was significantly different from other A horizons.

Fig. 2
figure 2

Results of PCA carried out for all studied stands in the O and A horizons for the following variables: fractional composition Cdec, HAs, FAs and HUs expressed as a % of TOC, the E4/6 quotient, HAs components, i.e. C, H, N, O and NMR parameters Calip, Carom, Ccarb

Full size image

Based on the same variables for the stands in the O and A horizons, a cluster analysis was performed (Fig. 3). Three stand groups were distinguished as a result of the analysis. The stands in the O horizons were grouped showing similarities between them, as were the stands in the A horizons. In the O horizons, YS–O and MS–O were the most similar. In the A horizons, AF–A and OS–A were the most similar as were YS–A and MS–A.

Fig. 3
figure 3

Cluster analysis (Ward method, Euclidean distance) carried out for all studied stands in the O and A horizons for the following variables: fractional composition Cdec, HAs, FAs and HUs expressed as a % of TOC, the E4/6 quotient, HAs components, i.e. C, H, N, O and contents of Calip., Carom., Ccarb. based on 13C-NMR analysis

Full size image

Discussion

The changes recorded in the soils during succession must be considered in the context of a broad understanding of the initial state of the soils, their ability to buffer external factors, the species composition of invading vegetation and the quantitative and qualitative transformation of the soil microbiome. These factors strongly influence soil processes and the cycling of matter and energy and are also reflected in the characteristics of SOM. Abandoned arable soils are often poor in SOM. Several studies have shown that conventional agriculture causes a strong depletion of SOM stocks and a deterioration of its quality (Compton and Bonne 2000; Murty et al. 2002), followed by negative changes in soil aggregation (Polláková et al. 2018; Šimanský et al. 2019) and other soil characteristics. Finally, the soil fertility and productivity decrease below the levels profitable for agricultural production. This problem concerns mainly sandy soils, which are naturally poor in nutrients, usually acidified and characterised by low cation-exchange, water holding and buffering capacities (Okołowicz et al. 2003; Jonczak and Sztabkowski 2021). Negative changes in arable soils can be reduced by applying suitable fertilisation, in particular organic (Abiven et al. 2009; Šimanský et al. 2019), and using certain agrotechnical practices, such as no-till, strip-till, ridge-till and mulch-till systems. The positive effects of reduced tillage on soil quality have been reported by several authors (e.g. Gadermaier et al. 2012; Devine et al. 2014; Blanco-Canqui and Ruis 2018; Kobierski et al. 2020).

The rehabilitation of degraded soils is a priority of sustainable management policies. This process is partially controlled in afforested areas. Soil amelioration can essentially contribute to maintaining the sustainability of forest ecosystems (Kilian 1998). However, many abandoned soils undergo natural regeneration. The development of annual vegetation, including grasses and herbs, usually constitutes the first step in that process. The positive effects of such vegetation on SOM have been well reported in the literature. Leinweber et al. (1993) observed a rapidly growing SOM content and changes in its structure, in the initial stages of succession. Reuter (1991) found that the decomposition of grass residues led to a significant accumulation of SOM, and the steepest increase being between the second and seventh years. Positive effects have also been observed where arable fields have been transformed into pastures (Römkens et al. 1999; Pulleman et al. 2000; Guo and Gifford 2002). Grass vegetation effectively contributes to the development of a stable soil structure (Pulleman et al. 2005) and the physical protection of SOM in aggregates (Six et al. 1998; Chevallier et al. 2004). Moreover, it effectively prevents SOM mineralisation by limiting soil heating (Li et al. 2016). The most common positive effects of grass vegetation on SOM were partially confirmed in the present study. The SOM and TOC contents in the AF and FA soils did not differ statistically (Tables 2, 3). The same tendency was observed in the SOM stocks (Table 3). Some differences were observed in the SOM fractional compositions. The FA soils were characterised by a higher contribution of HAs and HUs and a lower contribution of FAs (Table 3). This observation is in line with the findings of Ukalska-Jaruga et al. (2019). Based on our results, it can be stated that grass vegetation contributed to the greater stability of humic substances. This was also confirmed by the results of the elemental analysis of the HAs; a decrease in H content with increasing C and N content was noted (Table 4). The H:C atomic ratios in the HAs, which fluctuated around 1.2–1.4, correspond to aromatic structures coupled to an aliphatic chain containing up to 10 C atoms (Gonet 1989). The 13C-NMR analysis did not reveal any clear differences between the AF and FA soils (Table 6). Generally, in both these soils, aliphatic predominated over aromatic structures, which is typical of poorly advanced humification (Dębska 2004). However, it should be noted that the CPMAS 13C-NMR spectra were semiquantitative (Kinchesh et al. 1995; Mathers et al. 2000), and the C detected using this technique represents only part of the total pool of C in the sample (Mao et al. 2000).

The transformation of SOM during the latter stages of succession strongly depends on the species composition of the invading trees (Marcos et al. 2010). There are several mechanisms involved in causing the effects of forest vegetation on soils, as reviewed by Augusto et al. (2002). Litterfall production and its microbiochemical transformation are one of the most important amongst these, especially in the context of SOM. Litterfall is a key link in the biogeochemical cycling of matter and energy (Nordén 1994; Astel et al. 2009; Krishna and Mohan 2017), providing an important pool of soil nutrients (Ukonmaanaho et al. 2008; Straková et al. 2010) and the precursors to humic substances (Taylor et al. 1989; Stevenson 1994; Cotrufo et al. 2015). Generally, trees with higher litter quality exert a more positive influence on soil biodiversity (van Calster et al. 2008) and the intensity of biochemical processes (Kooijman and Martinez-Hernandez 2009). The issue of litterfall production and its chemistry and decomposition in birch stands has not been sufficiently explored until now, however. Studies by Aussenac et al. (1972) and Tripathi et al. (2006) have shown that litterfall production in birch stands is no different from that in forests dominated by many deciduous tree species (Shen et al. 2019). Generally, silver birch litterfall is moderately abundant in nutrients (Berg and Staaf 1987; Perala and Alm 1990; Johansson 1995; Brandtberg et al. 2004; Carnol and Bazgir 2013), with leaves, as the major component, decomposing rapidly (in 1–2 years) under temperate climatic conditions (Mikola 1985; Tripathi et al. 2006; Hordecki and Jagodziński 2019) and more slowly under colder climatic conditions (Huttunen et al. 2009). Hence, mull-type humus is typical of birch stands. This was confirmed in the studied chronosequence. The litter horizons were poorly developed and relatively rich in N, which was also reflected in the low C:N ratios (Table 2). This relatively high N abundance can be explained by the post-arable nature of the stand. Pools of SOM in the O horizon were lower than in the A horizon (Table 3). The qualitative characteristics of the SOM were typical of litter layers, including the low contribution of humified components with high residual (non-humified) fraction contents (Table 3). Moreover, the HAs were rich in H and C, but poor in N (Table 4). However, the 13C-NMR results showed an increasing aromaticity of the HAs with stand age (Table 6). Alkyl and O-alkyl C predominated in all the studied samples. This is typical of most forest soils, particularly in their litter horizons (Ussiri and Johnson 2003; Rumpel et al. 2005). The observed tendencies cannot be unambiguously explained by birch age due to the admixture of other species. In addition, the characteristics of the litter horizons need to be carefully interpreted, especially in mull-type humus, due to the single sampling. This approach does not reflect dynamic changes in the SOM that can be observed over time as an effect of decomposition and the seasonal variability of the weather conditions, amongst other factors.

The process of litterfall decomposition involves the following sequence of changes: (1) the preferential degradation of structures, such as cellulose and hemicellulose, by the microbial community, resulting in a decrease in the O-alkyl C content; (2) the decomposition of aromatic C and (3) the accumulation of alkyl C. Dissolved organic matter (DOM), as a product of that process, is transported towards deeper soil horizons by leaching. The intensity of its production and transport is strongly affected by a complex of abiotic and biotic factors, which have been the focus of numerous studies. For example, Aber et al. (1993) and Currie et al. (1996) both highlighted the importance of tree species composition, Clark et al. (2005) soil pH and Møller et al. (1999) soil microbial activity. In the mineral horizons, DOM is involved in various soil structural components and processes (Chen and Avnimelech 1986). The secondary synthesis of its low-molecular compounds is of key importance in humus formation (Stevenson 1994) and development of the A horizon. Due to its location at the top of the soil profile, the characteristics of that horizon cause DOM to be highly dynamic and susceptible to impacts from external factors, particularly land-use type and vegetation.

Previous studies have demonstrated the various effects of tree species composition on SOM content and quality. Most authors have reported increasing SOM contents and pools during the transition from arable field or grassland to forest land use (e.g. Gonet et al. 2009; Oktaba and Kusińska 2012; Rytter and Rytter 2020). This is logical because the SOM content in soils is directly affected by the input of plant-derived C (Wang et al. 2017), and this is usually greater in forest ecosystems. However, in young tree stands, litterfall production is often low. Moreover, accordingly to Oades (1988), there is a greater intensity of mineralisation in forests; thus, their SOM contents are lower than those of grassland soils. Li et al. (2016) reported that the effects of afforestation on soil C and N can be strongly modified by lithology. The impact of birch trees is not fully understood in this context, although previous studies have highlighted relatively positive effects (Daugaviete et al. 2003; Rosenvald et al. 2011; Holubík et al. 2014). Vladychenskii et al. (2007) also found much higher SOM contents under birches when compared to arable soils, although the pools not differ much. Moreover, the authors noted significant differences in the distribution of humus in the microprofiles under birch, oak and pine forests. The positive effects of birches on SOM reported by previous authors were not confirmed by the present study, with our findings clearly indicating a rapid loss of that component following silver birch succession (Table 3). This tendency can be explained by increased dehydrogenase activity and accelerated mineralisation (Gawęda et al. 2019) in the young birch stands compared to the older stands. This hypothesis is also supported by the fractional composition of the SOM. The soils under the young birches were characterised by a lower content of low-molecular humus fractions (Cdec, FAs) (Table 3), and these are more susceptible to microbial degradation. Moreover, in the YS, there was typically a lower internal oxidation of HAs, and the highest content of aliphatic structures with the lowest aromatic content (Table 6). The accelerated mineralisation of low-molecular SOM fractions may also be partially associated with the high nutritional demands of young birches (Miller 1984). Generally, it can be concluded that the transition from FA to a YS of silver birch contributes to negative changes in SOM characteristics.

During the latter stages of succession, some evidence of SOM regeneration was observed. The SOM content and stock gradually increased, until ultimately reaching values typical of AF and FA soils at the OS stage (Table 3). This is in line with the findings of previous studies by Smal and Olszewska (2008) and Tanner et al. (2014). However, considering the admixtures in the studied stands, it was not possible to precisely estimate the role of silver birch in that process. The occurrence of alder in the OS stand could be of particular importance. Podwika et al. (2018) reported the strongest positive effects of alder on soil quality compared to several other species. Alder litterfall is rich in nutrients (Jonczak et al. 2016) and rapidly decomposes (Jonczak et al. 2015). The increasing tendency in SOM content in the studied silver birch chronosequence was mainly associated with increasing contents of FAs, whereas the HAs remained relatively stable (Table 3). The E4/6 ratios of the HAs did not changed significantly (Table 5); however, a large increase in the internal oxidation of the HAs was noted (Table 4), accompanied by considerable changes in their molecular structures (Table 6).

Conclusions

Our findings confirm the clear inter-relationship between land-use type, vegetation and SOM characteristics. The spontaneous development of vegetation was accompanied by dynamic changes in the SOM stocks and quality in the abandoned sandy soils. There was little difference, in the transition from AF to FA soil, in the SOM stocks in the A horizons, although the SOM fractional composition and internal structure were modified. Increasing contents of HAs with decreasing contents of FAs might be considered to be an indicator of SOM stabilisation under FA vegetation. This was also supported by the results of the elemental and 13C-NMR analyses of the HAs. This observation highlights the positive effects of grass vegetation on SOM quality and supports the observations of several other authors. In its initial stage, the silver birch succession was accompanied by a rapid loss in SOM in the A horizon. This was probably because of accelerated mineralisation and the high nutritional demands of that tree species. This hypothesis was supported by the qualitative changes in, and fractional composition of, the SOM. The soils under young birches were characterised by a lower content of low-molecular humus fractions, which are more susceptible to microbial degradation, when compared to the FA soils. The birch succession contributed to a lower internal oxidation of HAs and an increase in the content of aliphatic structures in their molecules. Consequently, it can generally be stated that the initial stage of silver birch succession caused negative changes in the SOM characteristics. However, during the following stages of succession, some evidence of regeneration was observed. The contents of SOM gradually increased, mainly due to an increase in FAs. A considerable increase in the internal oxidation of the HAs, accompanied by structural changes in their molecules, as evidenced by 13C-NMR analysis, confirms the stabilisation of those substances. To summarise, our study has provided strong evidence for negative rather than positive changes in SOM caused by silver birch succession on sandy soils. More utilitarial conclusion of this study is that afforestation of post-agricultural sandy soils with that tree does not contribute to sustainable management of its resources. The results of the study provide grounds for considering the relevance of this practice. However, further studies are highly recommended to confirm or reject our hypothesis based on a wider number of cases, covering a broader spectrum of environmental/soil conditions.