Effects of Land Use on Soil Parameters and Carbon Dynamics in Surface Soil of Ecosystems of Rila Mountains, Bulgaria

Abstract

This study quantifies how different land-use types influence surface soil characteristics (0–5 cm) and the dynamics of soil organic carbon (SOC) and nitrogen in the mountainous ecosystems of the Rila Mountains. Across 54 forest and agricultural plots, pH, bulk density, coarse fraction, C:N ratio, SOC, total nitrogen (TN), and their respective stocks were assessed using standard analytical methods and statistical tests (Shapiro–Wilk, ANOVA, Kruskal–Wallis, correlation and regression analysis). Land use significantly affected all soil parameters except pH. Forest soil showed lower bulk density and lower SOC stocks compared with grasslands. Unmown meadows exhibited the highest SOC and TN concentrations and stocks, while potato fields recorded the highest bulk density and elevated TN stocks, reflecting intensive management impacts on surface soil properties. Forest soils displayed species-specific patterns, with Scots pine and Silver fir showing comparatively lower SOC and TN stocks attributable to historical degradation and site limitations. As the study focused on the uppermost soil layer (0–5 cm), the results should be interpreted more as indicators of surface soil dynamics rather than as estimates of total topsoil carbon and nutrient storage. Correlation analysis revealed strong positive relationships among SOC, TN, and the C:N ratio, and strong negative relationships between SOC and both bulk density and coarse fraction in managed agricultural lands. The findings demonstrate that minimizing soil disturbance and maintaining permanent vegetation cover—particularly through conservation of unmanaged grasslands—offer great capacity for enhancing the soil organic matter accumulation in mountainous ecosystems.

1. Introduction

Soil is a key natural resource supporting plant growth for food and raw materials, regulating water through filtration and storage, and playing an important role in the nutrient and carbon cycles [1,2]. Additionally, soil serves as a habitat for diverse organisms, provides structural support for infrastructure, and contributes to climate regulation and the mitigation of natural hazards such as flooding [1]. Soil’s properties and its ability to provide various ecosystem services [3] are of particular importance for sensitive ecosystems such as those in mountainous areas. Mountain soils are highly vulnerable due to steep slopes, shallow profiles, and intense climatic pressures that accelerate erosion, nutrient loss, and structural degradation [4,5,6].

In addition to the inherent natural pressures on soil systems related to the natural characteristics of the mountainous landscape, the anthropogenic impacts on soils are also considerable, particularly in the context of deforestation or unsustainable land use [7,8]. Land-use changes are a primary driver of soil physicochemical property alterations [9], often leading to increased erosion, sedimentation, and overall land degradation [10,11,12], thereby reducing land productivity [13].

Another implication of land-use changes and subsequent land management regimes is related to the organic matter dynamics and soil carbon sequestration potential [14,15]. This is of particular importance in the context of increasing atmospheric CO₂ emissions. Soils store approximately twice as much carbon as the atmosphere and nearly three times more than vegetation, making them a critical component of the global carbon cycle [16,17]. In a global meta-analysis study, Guo & Gifford (2002) [7] reported that soil carbon (C) stocks declined by an average of 9% due to land conversion across all land-use categories, although the responses varied widely among specific transitions. For example, C increased after conversions from forest to pasture (+8%), crop to plantation (+18%), crop to secondary forest (+53%), and crop to pasture (+19%). In contrast, C stocks declined following transitions from pasture to plantation (−10%) and forest to plantation (−13%), with the largest losses occurring when forest or pasture were converted to cropland (−42% and −59%, respectively). These trends are further confirmed by more recent global meta-analysis by Beillouin et al. (2023) [18], who reported the conversion of forest lands, grasslands, and wetlands to croplands consistently resulted in large SOC losses, with a mean change of −25% and SOC gains are significant when croplands are converted to forests (+57%) and grasslands (+26%). These outcomes highlight that land-use and -management decisions determine whether soils function as carbon sources or sinks [19], with direct implications for soil fertility, ecosystem functioning, and climate mitigation. However, the effect of land-use conversions on SOC is considerably determined by the local soil and climate conditions; thus, regional studies on how specific land-use practices modify fundamental soil parameters are essential.

This paper aims to study and assess the influence of land use and land conversion on soil properties and the carbon and nitrogen dynamics in surface soil as the most biologically active zone [20] and to determine the interrelationships between certain soil parameters. It should be acknowledged, however, that the focus on the uppermost soil layer, while appropriate for detecting recent management-induced changes [21], does not represent the full topsoil carbon and nutrient dynamics [22], and the findings should be interpreted within this constraint. Understanding these patterns is important for developing effective land management strategies that mitigate soil degradation and enhance ecosystem resilience in mountainous areas. This research is particularly relevant given that mountainous ecosystems are inherently fragile and highly susceptible to anthropogenic disturbances, making them critical areas for studying the long-term impacts of land-use change on ecosystem services such as soil conservation and carbon storage [23]. Furthermore, increasing soil organic carbon is important for carbon sequestration, addressing climate change, improving land productivity, enriching soil biological and chemical properties, and supporting microbial communities [24]. These benefits collectively contribute to ecological restoration efforts and the mitigation of global environmental challenges.

2. Materials and Methods

2.1. Study Area

The study area encompasses parts of Yundola’s village territory (around 5200 ha), spread around the eastern and southeastern ridges of the Rila Mountains, as well as the western and northwestern ridges of the Rhodope Mountains (Figure 1). An evident characteristic of the topography in the region is its elevated altitude and substantial topography complexity. The mean altitude of the territory is 1479 m, with a maximum elevation difference of 804 m.

The climate is categorized as mountainous, with cold winters and relatively cool summers. Thermal conditions are highly seasonal due to winter–summer radiation differences. The greatest air temperature fluctuations occur at the end of the growing season (Figure 1). On average, temperatures stay above 0 °C for 255 days, above 5 °C for 191 days, and above 10 °C for 122 days. The active temperature sum averages 2105° (above 5 °C) and 1595° (above 10 °C). The growing season lasts about 5.3–5.9 months at elevations of 1380–1600 m. The precipitation regime exhibits two peaks in May (89.3 mm) and November (58.6 mm), with lows in September (41.9 mm) and February (45.1 mm). At 1600 m above sea level, July and August are critical months for drought, whereas at 1380 m, the diagram does not show any dry periods (Figure 1). Average annual relative humidity is 78%, with about 12 days below 30% and 100 days above 80% each year.

The vegetation is indicative of the middle mountain belt, where beech and coniferous forests dominate at altitudes ranging from 700 to 2000 m above sea level. Most of the area (92.3%) is located within the middle mountain forests area according to the Bulgarian classification of forest vegetation zones, which are characterized by beech, fir, and spruce (1200–1700 m above sea level). The soils across the area are classified as Cambisols (WRB, 2022), consistent with the typical pedological conditions for the mid-montane elevation in Bulgaria [25,26], developed on metamorphic and magmatic rocks such as gneisses and granites.

A total of 54 sample plots (SPs) of varying land-use types have been established, including 27 located in woodlands (Pinus sylvestris L., Picea abies (L.) Karst., and Abies alba Mill.) and the remaining 27 situated on agricultural lands, encompassing potato fields, annually mown meadows, and unmown meadows. All 54 sample plots were distributed within an altitudinal range of 1300–1700 m a.s.l., where Dystric–Eutric Cambisols predominate [27].

2.2. Sampling Design and Laboratory Analysis

The study focuses on the uppermost soil layer (0–5 cm), which is the most biologically active soil layer, characterized by high microbial activity, rapid decomposition, and dynamic carbon cycling, making it especially responsive to management practices, land-use changes, and environmental fluctuations [28,29,30]. The sample plots have a size of 900 m². From each plot, a composite sample for the surface soil layer (0–5 cm) has been collected. In addition, a dedicated core sample in two replicates has been extracted as suggested by Cools and De Vos (2013) [31] to determine the bulk density for 0–5 cm following the Kachinsky method [32].

The soil samples were oven-dried at 105 °C (ISO 11465:1993) [33] and then processed for physicochemical analysis. The following properties were analyzed: pH in H₂O (ISO 10390:2005) [34], nitrogen content (Kjeldahl method), and carbon content (modified Turin method). Carbon and nitrogen stocks were derived using the IPCC GPG-LULUCF formula [35]—Equation (1), which has been adapted for nitrogen stocks by Ellert & Bettany (1995) [36].

$${SOC stock}_{ layer}={({SOC}_{layer}· Bulk Density · {Depth}_{layer}· (1-Frag)·10)}_{layer}$$

(1)

where

• SOC stock _layer—SOC stock in a layer, tons C∙ha⁻¹;
• SOC _layer—SOC content, gC∙kg⁻¹ soil;
• Bulk Density—g∙cm⁻³;
• Depth _layer—sampled depth, m;
• Frag—coarse fraction (particles with a diameter > 2 mm), %/100.

2.3. Statistical Analysis

The analysis started with deriving the basic statistical parameters: mean; median; minimum and maximum values; standard error; confidence interval ±95%; lower and upper quantile; coefficient of variance (CV). The data was checked for normal distribution using Shapiro–Wilk test. The analysis continued with parametric (one-way ANOVA) and non-parametric (Kruskal–Wallis) statistical tests to check if there is statistically significant difference between the means of the groups defined by the land-use types. Duncan’s test was used to identify the differences between groups after a positive one-way ANOVA. The analysis continued with exploring the relationship between the variables applying correlation analysis and linear regression. The processing of data and the statistical analysis were done in R programming.

3. Results

The analysis of the two main characteristics of the normal distribution, specifically the mean and variance, shows that the parameters such as pH and N content exhibited low CV, indicating relatively stable distributions and minimal dispersion around the mean. The pH of the analyzed soils was strongly acidic with mean values between 4.86 and 5.07 across all the land-use types (

Table 1. Descriptive statistics for soil parameters.

Parameter	Land Use	n	Mean	se	±CI95%	CV, %
pH	Norway spruce	9	4.86	0.12	0.27	7
	Scots pine	9	5.06	0.07	0.17	4
	Silver fir	9	4.94	0.06	0.14	4
	Unmown meadow	9	5.08	0.10	0.22	6
	Meadow	9	5.17	0.17	0.39	10
	Potato fields	9	5.04	0.19	0.43	11
Bulk Density, g∙cm−3	Norway spruce	9	0.83	0.13	0.30	47
	Scots pine	9	0.52	0.03	0.06	15
	Silver fir	9	0.65	0.04	0.08	17
	Unmown meadow	9	0.66	0.07	0.16	32
	Meadow	9	0.96	0.08	0.18	25
	Potato fields	9	1.18	0.03	0.08	8
Coarse fraction, %	Norway spruce	9	12.16	1.81	4.18	45
	Scots pine	9	17.22	0.51	1.18	9
	Silver fir	9	11.46	0.89	2.06	23
	Unmown meadow	9	2.67	0.98	2.27	111
	Meadow	9	5.24	1.49	3.44	85
	Potato fields	9	8.95	1.61	3.72	54
SOC, g∙kg−1	Norway spruce	9	46.68	3.93	9.06	25
	Scots pine	9	30.95	4.45	10.27	43
	Silver fir	9	37.14	4.78	11.02	39
	Unmown meadow	9	81.26	21.31	49.13	79
	Meadow	9	31.73	6.59	15.20	62
	Potato fields	9	27.69	2.52	5.82	27
TN, g∙kg−1	Norway spruce	9	11.42	0.54	1.24	14
	Scots pine	9	9.89	0.44	1.02	13
	Silver fir	9	10.74	0.54	1.23	15
	Unmown meadow	9	11.31	0.15	0.35	4
	Meadow	9	9.97	0.32	0.75	10
	Potato fields	9	10.05	0.16	0.36	5
C:N	Norway spruce	9	4.14	0.37	0.85	27
	Scots pine	9	3.22	0.54	1.24	50
	Silver fir	9	3.49	0.45	1.03	38
	Unmown meadow	9	7.03	1.71	3.95	73
	Meadow	9	3.06	0.53	1.22	52
	Potato fields	9	2.73	0.20	0.47	22
SOC stock, t C∙ha−1	Norway spruce	9	16.68	2.80	6.46	50
	Scots pine	9	6.50	0.75	1.73	35
	Silver fir	9	10.73	1.68	3.87	47
	Unmown meadow	9	21.44	1.84	4.25	26
	Meadow	9	12.80	1.52	3.51	36
	Potato fields	9	14.90	1.44	3.32	29
TN stock, t N∙ha−1	Norway spruce	9	4.15	0.68	1.56	49
	Scots pine	9	2.16	0.18	0.41	25
	Silver fir	9	3.08	0.22	0.51	21
	Unmown meadow	9	3.62	0.36	0.84	30
	Meadow	9	4.42	0.28	0.66	19
	Potato fields	9	5.39	0.13	0.31	7

). The mean total nitrogen content varied between 9.88 g∙kg⁻¹ and 11.42 g∙kg⁻¹ dry weight within the analyzed land uses. Both pH and total nitrogen content exhibited coefficients of variation below 15% (Table 1), indicating relatively low dispersion around the mean (Figure 2A and Figure 3B).

In contrast, the means of bulk density, coarse fraction, SOC content, SOC stock and C:N ratio exhibited large variation among the different land uses (Table 1), thereby signifying relative heterogeneity across sampling sites. These observations are also indicative of the natural landscape variability present within the region.

The bulk density is greatest in potato fields with a mean value of 1.18 g∙cm⁻³ followed by meadows with a mean of 0.96 g∙cm⁻³ (Table 1). Unmown meadows have a much lower average bulk density (0.66 g∙cm⁻³) than meadow sites, but their coefficient of variation is higher (CV 32% versus CV 25%). Forest soils are on average less compacted compared with the managed agricultural lands with mean values of bulk density of 0.52 g∙cm⁻³ in Scot pine sites, 0.63 g∙cm⁻³ in Silver fir sites and 0.83 g∙cm⁻³ in Norway spruce (Table 1). The increased CV observed in Norway spruce indicates greater spatial variability in bulk density across these sites.

The average content of the coarse fraction in the studied areas varies from 2.67% to 8.95% in managed and unmanaged agricultural lands, compared to 11.46% to 17.22% in forests (Table 1). The CVs in the agricultural lands are over 54%, reaching 111%, which reflects the high variability of this parameter, particularly in the complex and diverse landscapes of mountainous areas.

Carbon content in the surface soil layer (0–5 cm) differs across all studied land-use types. Managed agricultural lands (potato fields and meadows) show the lowest mean carbon concentration (27.69 g∙kg⁻¹ dry weight), while unmown meadows have the highest mean value (81.26 g∙kg⁻¹ dry weight). Among forest plots, the differences in average carbon values are smaller compared to the grassy lands. The highest mean carbon content is observed in Norway spruce plots (46.68 g∙kg⁻¹ dry weight), and in Scots pine the lowest (Table 1). Forest plots show relatively consistent carbon levels, with CVs ranging from 25% to 43%. In contrast, agricultural fields exhibit wider variation, with CVs between 62% and 79%, except potato fields where the CV is 27% (Table 1).

In terms of SOC stock in the surface soil layer, the highest mean (21.44 t C∙ha⁻¹) is still reported for the unmown meadows, while the lowest mean (6.50 t C∙ha⁻¹) is observed for the Scots pine sites. This observation is also valid for the total nitrogen stock (TN stock), where the variation is bigger than the N content and the CVs range between 7% for potato sites and 49% for Norway spruce. The highest mean nitrogen stock is estimated for potato sites (5.39 t N∙ha⁻¹), whereas the lowest is for the pine plots—2.16 (t N∙ha⁻¹).

The results for the C:N ratio indicate that potato fields exhibit the lowest average ratio (2.73), while unmown meadows demonstrate the highest average value (7.03). Regarding forest ecosystems, the average values of the ratios are between 3.22 and 4.14, with CVs spanning 27% to 50%. Grassland ecosystems (meadows and unmown meadows) show higher variability, with CVs between 52% and 73%, while potato fields present the lowest variation at 22% (Table 1).

The higher CVs for some of the soil parameters necessitated checking the normality of the distributions of the analyzed data. The results revealed no significant deviations from a normal distribution for SOC and C:N ratio within meadow sites. In contrast, the Shapiro–Wilk test indicates departures from normality in several land-use types: specifically, SOC for Norway spruce, bulk density and SOC stock for Silver fir, and pH values for Scots pine (

Table 2. Shapiro–Wilk normality test results.

Parameter	Norway Spruce	Scots Pine	Silver Fir	Unmown Meadow	Meadow	Potato Fields
pH	0.676	0.023 *	0.182	0.080	0.547	0.078
Bulk density	0.333	0.976	0.025 *	0.259	0.293	0.521
Coarse fraction	0.247	0.719	0.656	0.024 *	0.380	0.083
SOC	0.008 *	0.124	0.050	0.000 *	0.049 *	0.306
N	0.609	0.917	0.386	0.422	0.908	0.977
C:N	0.092	0.108	0.144	0.000 *	0.049 *	0.318
SOC stock	0.383	0.617	0.016 *	0.789	0.722	0.330
TN stock	0.387	0.906	0.728	0.036 *	0.653	0.631

(p-value—p = 0.05, * showing that the data is not normally distributed).

). For unmown meadow sites, the data exhibited non-normal distribution for multiple parameters, including coarse fraction, C:N ratio, total nitrogen stock (TN stock), SOC, and SOC stock.

Given the heterogeneity of the dataset in terms of its normal distribution, both ANOVA and Kruskal–Wallis tests were applied to assess the differences among groups (

Table 3. ANOVA single-factor (land use) and Kruskal–Wallis test.

Parameter	Df	ANOVA		Kruskal–Wallis
		F-Value	p-Value	Chi-sq	p-Value
pH	5	0.749	0.591	4.690	0.455
Bulk density	5	11.26	0.000 *	27.317	0.000 *
Coarse fraction	5	16.07	0.000 *	33.454	0.000 *
SOC	5	4.308	0.003 *	24.345	0.000 *
N	5	3.129	0.016 *	16.564	0.005 *
C:N	5	3.898	0.005 *	20.042	0.001 *
SOC stock	5	8.311	0.000 *	28.116	0.000 *
TN stock	5	9.826	0.000 *	28.929	0.000 *

(* p-value < 0.05).

).

As both methods produced consistent results, indicating that land use significantly influences all the soil properties examined with the exception of pH, Duncan’s test was then used for post hoc comparisons to identify specific group-level differences. These results are visually represented in Figure 2, Figure 3 and Figure 4, where boxplots show the distribution of each variable together with a grouping letter, defined by the post hoc Duncan test after a positive ANOVA result. The letters in the figures indicate which subsets are statistically distinct from one another, reflecting variations attributable to land-use categories.

Pearson’s correlation coefficient analysis revealed consistently strong positive relationships across land-use types (

Table 4. Correlation analysis.

	Parameter	pH	N	C:N	BD	CF	SOC Stock	TN Stock
All studied plots n = 54	SOC	−0.130	0.378 **	0.988 **	−0.381 **	−0.353 **	0.572 **	−0.248
	pH		−0.135	−0.113	0.175	−0.003	−0.062	0.122
	N			0.259	−0.123	−0.099	0.455 **	0.172
	C:N				−0.405 **	−0.349 **	0.540 **	−0.302 *
	BD					−0.049	0.302 *	0.940 **
	CF						−0.447 **	−0.209
	SOC stock							0.495 **
All except unmown meadows n = 45	SOC	−0.142	0.401 **	0.942 **	−0.271	−0.080	0.631 **	−0.107
	pH		−0.142	−0.091	0.185	0.028	−0.059	0.115
	N			0.100	−0.047	0.059	0.382 **	0.234
	C:N				−0.318 *	−0.081	0.509 **	−0.233
	BD					−0.211	0.539 **	0.947 **
	CF						−0.241	−0.305 *
	SOC stock							0.672 **
Forest plots n = 27	SOC	−0.055	0.211	0.906 **	0.155	−0.296	0.736 **	0.248
	pH		−0.083	0.049	−0.096	−0.106	−0.114	−0.068
	N			−0.191	0.291	−0.071	0.373	0.570 **
	C:N				0.000	−0.285	0.537 **	−0.012
	BD					−0.033	0.755 **	0.939 **
	CF						−0.244	−0.148
	SOC stock							0.799 **
Agricultural plots n = 27	SOC	−0.184	0.724 **	0.999 **	−0.750 **	−0.424 *	0.646 **	−0.732 **
	pH		−0.196	−0.184	0.213	0.319	−0.179	0.130
	N			0.724 **	−0.721 **	−0.483 *	0.888 **	−0.523 **
	C:N				−0.750 **	−0.424 *	0.646 **	−0.733 **
	BD					0.708 **	−0.467 *	0.946 **
	CF						−0.345	0.538 **
	SOC stock							−0.255
Managed lands n = 18 (potato fields and meadows)	SOC	−0.133	0.910 **	0.999 **	−0.627 **	−0.339	0.788 **	−0.387
	pH		−0.198	−0.134	0.276	0.374	−0.122	0.119
	N			0.910 **	−0.396	−0.250	0.909 **	−0.064
	C:N				−0.628 **	−0.340	0.788 **	−0.387
	BD					0.753 **	−0.072	0.908 **
	CF						0.021	0.536 *
	SOC stock							0.238
Unmown meadows n = 9	SOC	−0.484	0.825 **	0.999 **	−0.835 **	−0.387	0.449	−0.811 **
	pH		−0.554	−0.484	0.276	0.156	−0.517	0.240
	N			0.825 **	−0.765 *	−0.391	0.703 *	−0.671 *
	C:N				−0.835 **	−0.387	0.449	−0.811 **
	BD					0.207	−0.107	0.985 **
	CF						−0.481	0.086
	SOC stock							0.017
Unmown meadows (without outlier) n = 8	SOC	−0.430	0.965 **	0.999 **	−0.361	−0.274	0.826 *	−0.135
	pH		−0.412	−0.430	−0.100	0.021	−0.458	−0.175
	N			0.965 **	−0.487	−0.231	0.731 *	−0.261
	C:N				−0.361	−0.275	0.826 *	−0.134
	BD					−0.121	0.216	0.958 **
	CF						−0.426	−0.346
	SOC stock							0.444

(significance levels: ** p < 0.01; * p < 0.05).

), most notably between SOC and C:N (R = 0.91 ** ÷ 0.99 **) and between BD and TN stock (R = 0.91 ** ÷ 0.99 **). The analysis was conducted both collectively across all sample plots and within specific groups based on land-use and -management combinations to determine key relationships. For instance, forest plots (n = 27) and agricultural lands (n = 27) were assessed as distinct categories. Managed lands (n = 18), including potato fields and meadows, were also examined independently. Additionally, the influence of outliers was evaluated by analyzing all unmown meadows (n = 9) and subsequently excluding the outlier from this group (n = 8) for comparative purposes. Like this, the SOC was positively correlated with total N (R = 0.38 ** ÷ 0.97 **) and with SOC stock (R = 0.57 ** ÷ 0.83 *), while SOC stock also showed strong associations with C:N (R = 0.51 ** ÷ 0.83 *) and TN stock (R = 0.50 ** ÷ 0.89 **). Negative correlations were mainly observed in managed and unmanaged agricultural soils, where SOC, N, and C:N were strongly inversely related to BD, CF, and TN stock (R = −0.42 * to −0.75 **). Forest soils showed no significant negative relationships. In potato fields and meadows, SOC, N, and C:N remained tightly coupled (R = 0.91 ** ÷ 0.99 **), with additional positive correlations between BD and CF (R = 0.75 **) and between TN stock and BD (R = 0.91 **). Unmown meadows (n = 9) exhibited the strongest sensitivity to individual values but maintained the same overall structure: very strong SOC–C:N (R = 0.99 **) and SOC–N (R = 0.83 **) correlations, alongside pronounced significant correlation negative relationships between SOC, C:N and BD or TN stock (R = −0.81 ** to −0.84 **).

The results of the entire linear regression analysis are presented in Appendix A (see

Table A1. Regression equations derived from the whole dataset; n = 54 (all studied plots).

Y = a ∗ x + b	r2
SOC = 9.7334 ∗ N − 60.2467	0.14
SOC = 12.0854 ∗ C/N − 5.1129	0.98
SOC = −42.0013 ∗ BD + 76.1866	0.15
SOC = −1.9374 ∗ CF + 61.2019	0.12
SOC = 2.743 ∗ SOC stock + 4.6078	0.33
N = 0.0146 ∗ SOC + 9.9404	0.14
N = 0.0846 ∗ SOC stock + 9.3924	0.21
C/N = 0.0809 ∗ SOC + 0.5036	0.98
C/N = −3.6522 ∗ BD + 6.8686	0.16
C/N = −0.1565 ∗ CF + 5.4510	0.12
C/N = 0.2116 ∗ SOC stock + 1.0165	0.29
C/N = −0.5649 ∗ TN stock + 6.095	0.09
BD = −0.0035 ∗ SOC + 0.9477	0.15
BD = −0.045 ∗ C/N + 0.9779	0.16
BD = 0.0131 ∗ SOC stock + 0.6184	0.09
BD = 0.1955 ∗ TN stock + 0.0566	0.88
CF = −0.0644 ∗ SOC + 12.3587	0.12
CF = −0.7783 ∗ C/N + 12.6862	0.12
CF = −0.391 ∗ SOC stock + 15.0272	0.20
SOC stock = 2.4502 ∗ N − 12.0425	0.21
SOC stock = 1.3777 ∗ C/N + 8.4048	0.29
SOC stock = 6.9409 ∗ BD + 8.2861	0.09
SOC stock = −0.5119 ∗ CF + 18.7634	0.20
SOC stock = 2.362 ∗ TN stock + 4.8546	0.24
TN stock = −0.1612 ∗ C/N + 4.4407	0.09
TN stock = 4.5269 ∗ BD + 0.1817	0.88
TN stock = 0.1036 ∗ SOC stock + 2.371	0.24

,

Table A2. Regression equations derived from the whole dataset without the outlier; n = 53.

Y = a ∗ x + b	r2
SOC = 6.1763 ∗ N − 26.3554	0.21
SOC = 11.1905 ∗ C/N − 1.9766	0.92
SOC = −17.6612 ∗ BD + 53.0365	0.09
SOC = −0.9531 ∗ CF + 48.0422	0.11
SOC = 1.8834 ∗ SOC stock + 13.0553	0.54
N = 0.0335 ∗ SOC + 9.2366	0.21
N = 0.0817 ∗ SOC stock + 9.4208	0.19
C/N = 0.0825 ∗ SOC + 0.4415	0.92
C/N = −1.7218 ∗ BD + 5.0326	0.11
C/N = −0.0776 ∗ CF + 4.3958	0.10
C/N = 0.1422 ∗ SOC stock + 1.6991	0.42
BD = −0.005 ∗ SOC + 1.0053	0.09
BD = −0.0662 ∗ C/N + 1.0522	0.11
BD = 0.0169 ∗ SOC stock + 0.5818	0.15
BD = 0.1931 ∗ TN stock + 0.0674	0.88
CF = −0.1121 ∗ SOC + 14.1350	0.11
CF = −1.2379 ∗ C/N + 14.2971	0.10
CF = −0.3661 ∗ SOC stock + 14.7822	0.17
CF = −1.1626 ∗ TN stock + 14.2756	0.08
SOC stock = 0.2878 ∗ SOC + 2.4776	0.54
SOC stock = 2.3018 ∗ N − 10.6284	0.19
SOC stock = 2.9467 ∗ C/N + 2.9062	0.42
SOC stock = 9.0888 ∗ BD + 6.2432	0.15
SOC stock = −0.4758 ∗ CF + 18.2795	0.17
SOC stock = 2.8009 ∗ TN stock + 2.8280	0.34
TN stock = 4.5455 ∗ BD + 0.1640	0.88
TN stock = −0.066 ∗ CF + 4.4990	0.08
TN stock = 0.1223 ∗ SOC stock + 2.1867	0.34

,

Table A3. Regression equations derived from all studied plots except unmown meadows; n = 45.

Y = a ∗ x + b	r2
SOC = 4.4198 ∗ N − 11.1919	0.16
SOC = 10.5259 ∗ C/N − 0.1940	0.89
SOC = 1.518 ∗ SOC stock + 16.1335	0.40
N = 0.0364 ∗ SOC + 9.1446	0.16
N = 0.0835 ∗ SOC stock + 9.385	0.15
C/N = 0.0843 ∗ SOC + 0.3924	0.89
C/N = −1.3686 ∗ BD + 4.4605	0.10
C/N = 0.1097 ∗ SOC stock + 1.9768	0.26
BD = −0.0748 ∗ C/N + 1.0764	0.10
BD = 0.0271 ∗ SOC stock + 0.4936	0.29
BD = 0.1938 ∗ TN stock + 0.0829	0.90
CF = −1.1109 ∗ TN stock + 15.2723	0.09
SOC stock = 1.7479 ∗ N − 5.8824	0.15
SOC stock = 2.3635 ∗ C/N + 4.4549	0.26
SOC stock = 10.6800 ∗ BD + 3.4843	0.29
SOC stock = 2.7292 ∗ TN stock + 1.8382	0.45
TN stock = 4.6298 ∗ BD + 0.0103	0.90
TN stock = −0.0835 ∗ CF + 4.7597	0.09
TN stock = 0.1654 ∗ SOC stock + 1.8035	0.45

,

Table A4. Regression equations derived from forest plots only; n = 27.

Y = a ∗ x + b	r2
SOC = 9.4496 ∗ C/N + 4.0687	0.82
SOC = 1.5007 ∗ SOC stock + 21.2952	0.54
N = 0.6163 ∗ TN stock + 8.7562	0.32
C/N = 0.0869 ∗ SOC + 0.2930	0.82
C/N = 0.1052 ∗ SOC stock + 2.4288	0.29
BD = 0.0279 ∗ SOC stock + 0.3507	0.57
BD = 0.1656 ∗ TN stock + 0.1479	0.88
SOC stock = 2.7523 ∗ C/N + 1.3447	0.29
SOC stock = 20.3413 ∗ BD − 2.2522	0.57
SOC stock = 3.7975 ∗ TN stock − 0.5834	0.64
TN stock = 0.5268 ∗ N − 2.4992	0.32
TN stock = 5.3391 ∗ BD − 0.4278	0.88
TN stock = 0.1681 ∗ SOC stock + 1.2304	0.64

,

Table A5. Regression equations derived from agricultural plots only; n = 27.

Y = a ∗ x + b	r2
SOC = −118.0884 ∗ BD + 157.2154	0.56
SOC = −30.2093 ∗ TN stock + 182.2053	0.54
SOC = −3.9706 ∗ CF + 69.1938	0.18
SOC = 4.8514 ∗ SOC stock − 32.5740	0.42
SOC = 36.0066 ∗ N − 329.0916	0.52
SOC = 12.4486 ∗ C/N − 6.3107	0.99
N = −2.2775 ∗ BD + 12.5698	0.52
N = −0.4336 ∗ TN stock + 12.3841	0.27
N = −0.0909 ∗ CF + 10.9528	0.23
N = 0.0146 ∗ SOC + 9.7594	0.52
N = 0.1812 ∗ C/N + 9.6675	0.52
N = 0.1342 ∗ SOC stock + 8.2437	0.79
C/N = −9.4860 ∗ BD + 13.1361	0.56
C/N = −2.4267 ∗ TN stock + 15.1435	0.54
C/N = −0.3190 ∗ CF + 6.0653	0.18
C/N = 0.3897 ∗ SOC stock − 2.1097	0.42
C/N = 2.8924 ∗ N − 25.9290	0.52
C/N = 0.0803 ∗ SOC + 0.5069	0.99
BD = −0.0593 ∗ C/N + 1.1875	0.56
BD = −0.0048 ∗ SOC + 1.1575	0.56
BD = −0.2271 ∗ N + 3.3053	0.52
BD = −0.0222 ∗ SOC stock + 1.2977	0.22
BD = 0.0420 ∗ CF + 0.6983	0.50
BD = 0.2480 ∗ TN stock − 0.1764	0.90
CF = −2.5646 ∗ N + 32.3968	0.23
CF = −0.0453 ∗ SOC + 7.7403	0.18
CF = −0.5637 ∗ C/N + 8.0260	0.18
CF = 2.3694 ∗ TN stock − 4.9962	0.29
CF = 11.8839 ∗ BD − 5.4856	0.50
SOC stock = −9.7414 ∗ BD + 25.4808	0.22
SOC stock = 1.0690 ∗ C/N + 11.8112	0.42
SOC stock = 5.8754 ∗ N − 44.9711	0.79
TN stock = −0.2211 ∗ C/N + 5.4241	0.54
TN stock = −0.0178 ∗ SOC + 5.3121	0.54
TN stock = −0.6304 ∗ N + 11.0622	0.27
TN stock = 0.1222 ∗ CF + 3.7931	0.29
TN stock = 3.6161 ∗ BD + 1.1008	0.90

,

Table A6. Regression equations derived from arable plots only (potato and meadows); n = 18.

Y = a ∗ x + b	r2
SOC = −43.4513 ∗ BD + 76.1587	0.39
SOC = 2.6026 ∗ SOC stock − 6.3357	0.62
SOC = 18.008 ∗ N − 150.5166	0.83
SOC = 12.4486 ∗ C/N − 6.3107	0.99
N = 0.1518 ∗ SOC stock + 7.9056	0.83
N = 0.046 ∗ SOC + 8.6416	0.83
N = 0.5726 ∗ C/N + 8.3513	0.83
C/N = −3.4904 ∗ BD + 6.6248	0.39
C/N = 0.2091 ∗ SOC stock − 0.002	0.62
C/N = 1.4466 ∗ N − 11.5841	0.83
C/N = 0.0803 ∗ SOC + 0.5069	0.99
BD = −0.0091 ∗ SOC + 1.339	0.39
BD = −0.1131 ∗ C/N + 1.3964	0.39
BD = 0.0325 ∗ CF + 0.8382	0.57
BD = 0.2362 ∗ TN stock − 0.0904	0.83
CF = 3.2219 ∗ TN stock − 8.7195	0.29
CF = 17.3921 ∗ BD − 11.4994	0.57
SOC stock = 2.97 ∗ C/N + 5.2556	0.62
SOC stock = 5.4484 ∗ N − 40.6784	0.83
TN stock = 0.0891 ∗ CF + 4.2755	0.29
TN stock = 3.4932 ∗ BD + 1.1735	0.83

and

Table A7. Regression equations derived from unmown meadow plots only; n = 9.

Y = a ∗ x + b	r2
SOC = −253.0788 ∗ BD + 249.4135	0.70
SOC = −47.6001 ∗ TN stock + 253.6749	0.66
SOC = 116.9218 ∗ N − 1241.2588	0.68
SOC = 12.446 ∗ C/N − 6.2938	0.99
N = −1.6352 ∗ BD + 12.3976	0.58
N = −0.2777 ∗ TN stock + 12.317	0.45
N = 0.0574 ∗ SOC stock + 10.0807	0.49
N = 0.0058 ∗ SOC + 10.8381	0.68
N = 0.0725 ∗ C/N + 10.8014	0.68
C/N = −20.3334 ∗ BD + 20.5449	0.70
C/N = −3.8242 ∗ TN stock + 20.8864	0.66
C/N = 9.3954 ∗ N − 99.2376	0.68
C/N = 0.0803 ∗ SOC + 0.5057	0.99
BD = −0.0028 ∗ SOC + 0.8883	0.70
BD = −0.0343 ∗ C/N + 0.9056	0.70
BD = −0.3575 ∗ N + 4.7083	0.58
BD = 0.1909 ∗ TN stock − 0.0269	0.97
SOC stock = 8.6176 ∗ N − 76.0338	0.49
TN stock = −0.0138 ∗ SOC + 4.7448	0.66
TN stock = −0.1719 ∗ C/N + 4.8317	0.66
TN stock = −1.6188 ∗ N + 21.9323	0.45
TN stock = 5.0888 ∗ BD + 0.241	0.97

). The equations employed for the evaluation of the relationship of SOC stock in the uppermost layer using SOC concentration are presented in Figure 5A,B. The coefficient of determination (r²) is lowest in the linear function involving all sample plots (r² = 0.33, n = 54). This is followed by the regression involving agricultural areas (r² = 0.42, n = 27) and forest areas (r² = 0.54, n = 27). The coefficient of determination is comparatively low in the regression analysis of data for all areas except unmown meadows (r² = 0.40, n = 45). The most robust prediction is achievable in the function incorporating plots with potatoes and meadows (r² = 0.62, n = 18) (Figure 5B). Conversely, the correlation coefficient determining the relationship in unmown meadows is statistically insignificant (Table 4), which necessitates the exclusion of the outlier. The regression, which incorporates solely unmown meadows and excludes the outlier, is based on a relatively small sample (r² = 0.68, n = 8), which makes it less reliable.

4. Discussion

Our study confirmed that land-use change constitutes a significant factor for alteration among various soil parameters in the surface soil layer. From all of the analyzed parameters, only pH values showed no significant difference between the analyzed groups of land use. The study showed that the pH values vary slightly. This could be attributed to natural characteristics of soil acidity predefined by the parent rock. The bedrock in the study area is predominantly composed of silicate-rich metamorphic rocks, which are known to weather slowly, releasing base cations gradually [37]. The active soil acidity ranges from 4.34 to 6.17, which classifies the examined brown forest soils as highly acidic to acidic. This is in line with other studies from the region [38,39] and across the Rila Mountains [40] and Western Rhodopes [21]. In addition to the parental material, the acidity of the soils could also be attributed to litter quality and the legacy of land use in the region. Land-use changes and management practices are known to affect soil properties. However, this study found no significant differences in average pH across the land uses examined, although there is greater variation in pH in arable plots and meadows compared to forest and unmown meadow sites (Figure 2A). This might be due to lasting effects from the long-term legacy of land use—the area was managed as forest for many centuries. Coniferous forests typically create more acidic soil conditions than deciduous forests [41,42]. The higher concentrations of recalcitrant compounds like lignin in coniferous needles [43,44], along with the usually higher C:N ratio in litterfall compared to the deciduous leaf litter [45], favors a slow decay of the organic matter which releases organic acid during decomposition. For the Yundola region, this is confirmed by published data on the concentration of trace elements of the aboveground litterfall, which shows that the C:N ratio in needles and wood fraction is high—60 and 65, respectively [38].

Regarding physical characteristics, both bulk density and coarse fraction exhibited statistically significant differences in mean values across the studied land-use categories (Table 3). Bulk density is divided into five distinct subsets (Figure 2B), where arable fields represent the group with the highest bulk density values, ranging from 1.03 to 1.30 g cm⁻³. This range exceeds 75% of the observations recorded in meadows and spruce forests and surpasses all values recorded for the remaining land-use types. Spruce plantations constitute a separate subset, with values falling within the lower part of the range observed for meadows. Meadows themselves display considerable internal variability, with bulk density ranging from 0.57 to 1.25 g·cm⁻³. More than three-quarters of meadow measurements exceed the median value observed for spruce. At their upper end, the meadow values for bulk density closely approach the maximum recorded for unmown meadows and surpass the bulk density values associated with both Scots pine and Silver fir stands. The observed tendency of bulk density being higher in cropland > meadows > woodland is consistent with large-scale studies on bulk density across Europe [46]. The higher bulk density in managed agricultural lands is mainly due to soil compaction related to land-management practices, as well as low input of organic matter compared to forest stands and unmanaged grassland [47,48,49].

Regarding the content of the coarse fragments (>2.0 mm), the analysis identifies three distinct subsets (Figure 2C). Soils beneath Scots pine exhibit the highest skeleton content. A separate subset includes soil under Norway spruce and Silver fir trees, as well as those found in potato fields. Among forest land uses, fir stands demonstrate the lowest coarse fraction content. The third subset comprises meadows and unmown meadows, where the content of coarse fragments is particularly low. The relatively high percentage of the coarse fraction in the forest soils could be attributed to erosion of the fine soil particles on mountain slopes, as most sampling plots are located on inclines of 11–20°. Although potato fields belong to the middle subset, their coarse fragment content is relatively high compared to what is expected for cultivated lands. This is likely a result of deep tillage practices [50] that bring stones to the surface [51], combined with vulnerability to erosion during periods of bare soil exposure in spring and autumn. The low coarse fragment content in grassland soils—both meadows and unmown grasslands—reflects the stabilizing effect of permanent vegetation cover, which prevents fine-earth loss, and the absence of soil disturbance practices.

The SOC content across the studied land-use categories indicates potential losses of SOC attributable to decomposition and erosion processes associated with the prevailing land-use and -management practices in the Yundola area. The analysis of variance (ANOVA) confirmed that land-use type has a statistically significant effect on SOC content in the surface soil layer (0–5 cm). However, the post hoc Duncan test identified only two distinct subsets, separating only the unmown meadows in a single group with the highest SOC values (Figure 3A). All remaining land-use types were grouped into a single subset, with mean SOC values ranging from 27.69 g∙kg⁻¹ dry weight to 46.68 g∙kg⁻¹ dry weight. This observation suggests a relatively limited differentiation in SOC content among the remaining land categories despite differences in management intensity. The SOC content in forest soils ranges from below 20 g∙kg⁻¹ to 50 g∙kg⁻¹ dry weight, which is consistent with large-scale observations for forest soils across Europe, where organic carbon concentrations in the soil of mid-mountain ranges fall between 2% and 5% [52]. Potato fields and meadows exhibit comparable SOC concentrations, although both land-use types record generally lower values than forest plots. This pattern is probably indicative of the combined effect of reduced organic matter inputs and greater susceptibility to SOC losses through tillage and erosion in managed non-forest lands [53].

Considering the link between the C accumulation and soil nitrogen dynamics, the observation of similar patterns of both C and N content in soils is expected. In contrast to SOC content, the Duncan grouping for total nitrogen concentration distinguishes three distinct groups (Figure 3B). The highest nitrogen concentration is observed in unmown meadows and spruce stands. Regarding the unmown meadows, this could be partially explained by the presence of legume cover (Trifolium sp., Lotus sp., Vicia sp.) [54] supporting high biomass productivity and accumulation of organic matter [55] due to the absence of biomass removal. The lowest nitrogen concentration is observed in the agricultural lands—both potato fields and meadows, categorized together in another subset following Duncan’s test. This is attributed to the continuous nutrient export following potato and hay harvest and erosion processes associated with different management practices [56]. This is also confirmed by the results on the C:N ratio which provides insight into organic matter quality and decomposition. The lowest mean ratio (2.73) is observed in potato fields (Figure 3C), suggesting faster turnover and nitrogen mineralization due to application of nitrogen fertilizers [57,58] and tillage practices [59]. The relatively low concentration of nitrogen in meadow sites is associated with nitrogen depletion following frequent mowing [60] combined with the absence of inorganic fertilization. Typically, in Yundola’s region the meadows are not amended with inorganic fertilizers and are mowed once per year which is consistent with general practice across Europe [61].

The nitrogen concentration in soils among the three types of forest stands differ significantly, categorizing them in three distinct groups following Duncan’s test (Figure 3B). This result suggests that the patterns of nitrogen in soil are mainly influenced by species-specific litter input and decomposition processes instead of management practices [62], especially in the absence of pollution sources, as seen in Yundola. The lowest total nitrogen concentration among the forest stands is observed in Scots pine, followed by Spruce and Silver fir, which also reflects the observed nitrogen content in the litterfall of these species [63]. The C:N ratio in forest plots, however, is rather low—the mean is around 4 (Figure 3C)—compared to other studies on brown forest soils across Bulgaria [64,65]. This is mainly attributed to the lower SOC concentration observed in forest soils in the area indicating that the soil is far from its natural equilibrium. The reason could be related to long-lasting anthropogenic activity in the area: active forestry, long-term grazing legacy during the past centuries. Most of the Scots pine stands in the Yundola region have been reforested over the past century, aiming to restore the area’s natural forest cover after the unsustainable forest management in the period before the XX century [66,67].

The combined effects of the SOC and TN concentrations, together with the altered physical properties previously discussed, contribute to a more comprehensive assessment of carbon and nitrogen availability across the studied land-use types. An inherent limitation of the study is its focus on the shallowest soil layer only, which restricts the capacity for a comprehensive topsoil assessment. Referring to the SOC and TN stocks of the uppermost soil layer (0–5 cm), the analysis identified five distinct subsets (Figure 4), reflecting the strong land-use signal [68] which is mainly linked to alterations in the physical properties. The results are consistent with other observations reporting that woodland and grassland soils store more organic carbon compared to arable lands [69,70,71]. Unmown meadows constitute a distinct subset, characterized by the highest SOC stock values in the upper 0–5 cm layer (11.8–29.4 t∙ha⁻¹). This range exceeds more than 75% of the observed values in the subsequent subset, which pertains to Norway spruce stands (Figure 4A). The high SOC stock values recorded in unmown meadows are predominantly attributable to elevated SOC concentrations, given that both bulk density and coarse fragment content are notably low in these plots—factors that would otherwise lead to a dilution of organic carbon stocks. The distribution of SOC stocks in unmown meadows is characterized by the presence of a distinct outlier, reflecting considerable within-group variability (Figure 4A). This is consistent with the spatial heterogeneity inherent to unmanaged grassland ecosystems [72], where the absence of regular mowing and soil disturbance promotes uneven accumulation of organic residues [53]. The contributing factors likely include variability in the distribution of plant biomass, microtopographic differences that create locally distinct moisture and decomposition regimes, and the patchy establishment of leguminous and herbaceous vegetation [73]. Together, these processes generate spatially variable organic matter inputs that are not homogenized by management intervention, resulting in localized patches of particularly high SOC accumulation.

The spruce stands also demonstrate considerable variability in the observed SOC stock in the surface (0–5 cm) soil layer (7.6–33.5 t∙ha⁻¹) (Figure 4A). This is further substantiated by the extensive distribution of the observed coarse fraction and bulk density in the spruce forest plots (Figure 2B,C). The two managed agricultural land-use types (potato fields and meadows) form an intermediate subset, with potato fields showing slightly higher SOC stocks in the uppermost 0–5 cm soil layer (12.6–17.1 t∙ha⁻¹) compared to meadows (9.8–14.5 t∙ha⁻¹), mainly attributed to the higher bulk density (Figure 2B). Silver fir stands constitute a separate group characterized by predominantly low SOC stock levels, with more than three quarters of the values below 11 t ha⁻¹, except for a single plot—22.9 t∙ha⁻¹ (Figure 4A).

The lowest SOC stocks in the surface soil layer (0–5 cm) are observed under Scots pine stands (2.6–9.4 t∙ha⁻¹), forming a distinct subset and highlighting the limited carbon storage in the surface soil layer. The relatively low SOC stocks observed in certain forest plots—particularly those under Silver fir and Scots pine—reflect a combination of historical land degradation, incomplete ecosystem recovery, and inherent site limitations. Historical records from the Yundola region confirm that these areas were subject to intensive deforestation during the XVIII–XIX centuries, primarily to support grazing [66]. The period of unsustainable resource use resulted in severe soil degradation, the effects of which probably persist in the current SOC stocks despite approximately 100 years of forest recovery following the establishment of the forest plantations. It should be noted, however, that the present study is limited to the uppermost 0–5 cm soil layer, which is the primary link between the atmosphere, vegetation, and deeper soil horizons. While this layer is highly responsive to recent land-use and -management changes, it does not capture the full depth of the topsoil, and therefore the long-term legacy effects of historical deforestation and subsequent reforestation on SOC distribution throughout the entire topsoil profile cannot be fully assessed from these data alone. Deeper soil layers may retain distinct imprints of historical land use that are not reflected in surface measurements, and future studies incorporating full topsoil profile sampling would provide a more comprehensive picture of SOC recovery trajectories in the region. Site characteristics further constrain SOC accumulation under Scots pine, as this species is known to colonize and thrive on nutrient-poor, shallow soils with inherently low organic matter content [74]. The pine plantations in the study area received limited silvicultural interventions during their development, meaning stand density was not actively managed to optimize conditions for organic matter accumulation. The observed stand densities in the pine plantations of the sample plots range from 0.7 to 1.0. Furthermore, Scots pine maintains a more open canopy structure compared to Norway spruce and Silver fir, resulting in reduced litterfall and consequently lower annual carbon inputs to the soil surface [75].

The highest TN stocks in the uppermost 0–5 cm soil layer were observed in the cultivated potato fields, where three quarters of the values exceeded those of all other land-use types, except for one spruce plot (7.7 t∙ha⁻¹). The elevated soil nitrogen levels observed in potato fields are attributable to the routine application of inorganic nitrogen fertilizers. Potato is among the most nitrogen-demanding arable crops [76], requiring substantial nitrogen inputs to support optimal vegetative growth and yield [77,78]. However, nitrogen use efficiency in potatoes is known to be low [79,80]. Therefore, a significant fraction of the applied nitrogen is not recovered by the harvested biomass and remains in the soil [79], contributing to elevated total nitrogen concentrations in the surface soil layer. This residual nitrogen, accumulated over repeated growing seasons, is therefore reflected in the soil nitrogen measurements recorded for potato plots in the present study. In addition, the higher TN stock in the surface soil layer is also attributed to the alterations in physical properties due to land management and more specifically the increase in bulk density compared to the other land uses. The observation is also confirmed by the correlation analysis and the strong positive relationship between the bulk density and the TN stock (Table 4).

Meadows and unmown meadows constitute the subsequent two distinct groups in terms of TN stock in the uppermost soil layer. Meadows display broader variability around the median, with the interquartile range spanning 3.9–5.1 t∙ha⁻¹, while unmown meadows exhibit a more compact distribution varied between 3.7 and 4.2 t∙ha⁻¹, punctuated by two notably low outlying values of 1.3 and 2.5 t∙ha⁻¹ (Figure 4B), which likely reflect the spatial heterogeneity in organic matter accumulation discussed previously. Among the forest plots, the spruce stands cluster into a common subset due to similar TN levels to those of unmown meadow. Forest soils under Silver fir and Scots pine form two additional and separate subsets. Silver fir shows higher TN stocks (1.9–4.0 t∙ha⁻¹), with 75% of the values exceeding 75% of those in spruce, whereas Scots pine exhibits the lowest TN stocks (1.4–3.1 t∙ha⁻¹). The differentiation of the forest plots in three distinct groups, combined with the positive correlations between the SOC and TN stocks observed in forest soil (Table 4), underlines the role of the organic inputs—primarily leaf litter and woody debris—in supporting nutrient retention and SOC accumulation.

The observed consistently strong positive relationships between SOC and C:N ratio across land-use types (Table 4) show that the increase in SOC is associated with a change in litter quality and organic matter composition [22,81,82,83]. The content and stability of organic carbon in soils, especially in mountainous areas, is significantly influenced by temperature conditions and the duration of snow cover [84]. Soil moisture and temperature also determine the rate at which organic matter decomposes, although higher temperatures do not always lead to accelerated decomposition [85]. The observed correlation between BD and TN stocks indicates that nitrogen storage is strongly driven by soil compaction and consequent changes in soil porosity. BD is considered an indicator of soil health and structural condition, with lower BD values generally associated with healthier soils that have higher SOC and TN stocks [86]. BD tends to increase with soil depth and is negatively correlated with SOC and TN, suggesting that more compacted soils may have reduced nutrient availability [87,88]. However, our study has limitations in terms of the soil depth at which changes are detected (0–5 cm), which prevent us from confirming these observations within the scope of the current study; however, similar analyses in the region [39,40] and the country [64,89] suggest that this also applies to the soils in the Rila Mountains.

Positive associations of SOC stock with SOC, N, C:N and TN stock are detected within most of the specific groups analyzed (Table 4), indicating that the quantity and quality of organic matter play a key role in the retention of nutrients. Similar positive correlations have also been found by Lacatusu et al. (2024) [90] and Paltineanu et al. (2024) [91]. The full dataset (n = 54, Table 4) reveals that SOC and C:N have negative correlations with BD and CF. The physical structure of organic matter accounts for its influence on soil properties. The association between SOC and BD stems from the comparatively lower density of organic matter, which is considerably lighter than mineral particles such as sand, silt, and clay. As SOC increases, it effectively “dilutes” the mineral content, leading to a reduction in overall bulk density. Additionally, elevated SOC levels enhance soil porosity by promoting the formation of larger aggregates, resulting in a greater number of macropores and air spaces. These changes contribute to a looser, lighter soil texture and further decrease in BD.

The relationship between SOC and CF is logical since sites with high CF content have limited root volume and organic matter accumulation. Conversely, stonier soils tend to be better drained and aerated, accelerating the oxidation of carbon and maintaining low SOC levels [92,93,94]. This suggests that accumulation of organic matter is sensitive to physical disturbance and coarse-textured substrates which limit stabilization potential.

This study shows that SOC stock in the uppermost layer (0–5 cm) is shaped by SOC concentration, nitrogen availability, and soil physical properties, with land-use context determining the strength and direction of these relationships (Figure 5, Table A1, Table A2, Table A3, Table A4, Table A5, Table A6 and Table A7). Regression functions predicting SOC stock from SOC further emphasized these differences. The performance was weakest when all plots were pooled and improved when analyses were restricted to more homogeneous land-use groups, particularly potatoes and meadows (Figure 5B).

5. Conclusions

This study conducted in the Yundola region confirms that soil physical properties and carbon–nitrogen dynamics are strongly governed by the land-use pattern, although the analysis is confined to surface soil processes within the uppermost 0–5 cm layer and does not capture subsurface dynamics. This represents an inherent limitation of the study, as the sampled layer, while highly responsive to recent land-use and -management changes, does not reflect the full topsoil profile. Consequently, the observed patterns in SOC and TN stocks should be interpreted as indicative of surface-layer responses rather than as a comprehensive assessment of total topsoil carbon and nitrogen reserves. Future studies incorporating full topsoil profile sampling—typically to 30 cm depth—are necessary to confirm and extend the present observations, particularly with respect to the long-term legacy effects of historical land use on SOC recovery trajectories.

Soil pH was found to be mainly governed by parent material and long-term land-use legacy, showing limited sensitivity to current management practices. On the contrary, both bulk density and coarse fragment content exhibited strong responses to land-use type, with values reflecting the combined influence of soil disturbances, vegetation cover and erosion processes. Arable soils recorded the highest bulk density, consistent with compaction associated with tillage and lower organic matter inputs, while unmown meadows and forest soils generally displayed more favorable physical conditions. These structural differences influenced SOC and TN stocks, emphasizing the central role of soil physical properties in regulating surface nutrient storage.

Unmown meadows were identified as the most effective land use for SOC accumulation in the surface layer, combining elevated SOC concentrations with low bulk density and minimal coarse fragment content in the uppermost (0–5 cm) soil layer. Forest soils exhibited strong species-specific patterns, with Scots pine and Silver fir displaying lower SOC and TN stocks due to site limitations, historical degradation, and litter input differences. The extent to which these patterns extend beyond the 0–5 cm layer remains an open question that merits targeted investigations. The managed agricultural lands showed intermediate SOC stocks but elevated TN stocks, driven partially by organic and inorganic nitrogen fertilization and its associated residual soil nitrogen, with bulk density emerging as an important mediating factor.

The findings show that management practices which limit soil disturbance and maintain a permanent vegetation cover, like the conservation or restoration of unmanaged grasslands, offer favorable conditions for improving soil carbon storage in mountainous regions such as Yundola. In land-use contexts characterized by strong historical legacies and susceptibility to erosion, physical soil indicators should be systematically integrated into land-use planning, carbon accounting frameworks, and ecosystem restoration strategies. Nevertheless, the generation of robust policy recommendations and reliable carbon stock estimates will necessitate further research. Such studies should aim to resolve the full topsoil depth, thereby facilitating a more comprehensive understanding of how land use influences carbon and nitrogen dynamics across the entire rooting zone.