Soil physical and chemical parameters affect litter decomposition in an agriculturally dominated catchment

1 Introduction

The decomposition of plant residues and other organic matter is a major factor in ecosystem functioning, and crucial to carbon and nitrogen cycling processes. Litter decomposition rates are mainly influenced by natural factors such as climate, soil nutrient concentrations, plant traits, and the soil microbial and detritivore communities (1–3). These processes are further controlled by soil physical and chemical properties, which affect microbial activity, nutrient availability, and the stabilization of decomposed organic matter. At the same time, soil properties are formed by land use type and topographic position, while slope position affects soil moisture and soil redistribution. Different land use types alter soil conditions through vegetation composition and management practices. Anthropogenic factors further influences the variations of decomposition rates, especially soil management techniques at different land use types, e.g., irrigation, tillage, or fertilizer application. Therefore, understanding how soil properties vary across land-use types and slope positions is essential for explaining spatial variation in litter decomposition and organic matter stabilization, especially in environmentally sensitive ecosystems (4, 5).

Soils represent one of the largest terrestrial carbon reservoirs, and their capacity to act as a carbon sink or a net source of greenhouse gases is strongly influenced by litter decomposition. As the primary pathway through which plant biomass is transformed into soil organic matter, litter decomposition regulates carbon sequestration, nutrient cycling, and microbial activity. Climatic changes and increasing atmospheric carbon dioxide concentrations can modify litter quality, including nitrogen concentrations and C:N ratios (6), potentially altering decomposition dynamics and carbon turnover. Soil organic matter content, especially soil organic carbon (SOC), can be used as an indicator of soil health, as higher organic matter present in the soil can enhance microbial activities, further helping ecosystem functioning. Therefore, the organic matter input can greatly influence soil stabilization and decomposition processes (7). Generally, the more natural ecosystems rely primarily on internally generated organic matter inputs, whereas managed ecosystems often experience altered organic matter dynamics due to external inputs and disturbance. For example, forests, particularly deciduous forests, continuously replenish topsoil litter layers, promoting the formation and accumulation of organic-rich soils. Further, the continuous input of plant residues and the absence of intensive soil disturbance favor the development of stable soil aggregates (8), which can protect organic matter from microbial decomposition and contribute to SOC stabilization. Grasslands are similar in that they typically experience little to no soil management and minimal biomass removal, with disturbances primarily affecting aboveground biomass through grazing or mowing. The dense root systems of grassland vegetation provide continuous belowground carbon inputs, while permanent vegetation cover reduces soil erosion and helps retain organic matter within the soil profile (9). Vineyards may be subject to varying management intensities, ranging from mechanical soil disturbance, such as inter-row tillage, to minimal soil disturbance where permanent grass cover is maintained between rows (10). These contrasting practices can lead to differences in decomposition and stabilization processes. Inter-row tillage can disrupt soil aggregates, increase oxygen availability, and stimulate microbial activity, while grassed inter-row can reduce erosion or can help retain organic matter, similar to grassland. In contrast, under-vine management often involves mechanical weed control to improve soil moisture availability and support grape production (11). Among agricultural land-use types, croplands generally experience the greatest anthropogenic impacts, as they are commonly characterized by intensive tillage, mechanized sowing, fertilization, and crop protection practices. Frequent disturbance of soil structure can reduce aggregate protection of organic matter, increase decomposition rates, and enhance SOC losses through mineralization and erosion (12, 13). Overall, reliance on externally supplied organic matter tends to decrease with increasing ecosystem naturalness, reflecting the greater capacity of more natural systems to maintain internal organic matter inputs and nutrient cycling. Therefore, a better understanding of soil organic matter (SOM) and SOC degradation processes in agricultural systems is necessary to help mitigate the negative effects of soil disturbance and land-use change.

Along with soil chemistry, soil physical properties can strongly influence soil microbial processes. Soil texture, such as predominantly sandy or clayey soil provides different environmental conditions for microbial growth. Variations in soil texture affect oxygen availability, which is particularly important for aerobic decomposition processes in surface and upper soil layers (14). At the same time, finer-textured soils with smaller particle sizes provide greater surface area, promoting microbial colonization and enabling a larger proportion of carbon derived from plant residues to be incorporated into microbial biomass (15). Clay particles can also stabilize litter-derived organic matter by protecting it from microbial decomposition, thereby reducing its accessibility to decomposers and slowing further decomposition (16). In contrast, sandy soils generally provide less physical protection for organic matter, resulting in lower stabilization potential and faster carbon turnover (14). Besides parent material and vegetation, the geographical setting of agricultural land is also important for understanding soil nutrient dynamics (17). For example, vineyards are often located on sloping terrain, where soil redistribution processes can alter both soil chemical and physical properties, resulting in substantial soil chemical and physical differences between upper and lower slope positions (18, 19). Therefore, the interactions among land use, soil properties, vegetation, and hydrological processes are best investigated at a larger spatial scale, such as catchment or sub-catchment. Where possible, land-use types should be further subdivided into more spatially representative units to better capture within-site variability.

The method of tea bag index (TBI) has been widely used for ecological research, as it is easy to use, cost-effective, and a standardized method, mainly reactive to abiotic changes in the soil such as soil temperature and soil moisture content (20, 21). Several research use this method to gather information on decomposition rates (k) and stabilization factors (S) in a wide range of environmental conditions, including different soil types, or aquatic conditions (22–24). Although the basic method developed by Keuskamp et al. (21) is widely accepted, its development and improvements for different environmental conditions are still ongoing (25). While the TBI uses standardized teas to calculate the k and S factors of the given soil, it has also been proved that these data can be representative of native litter decomposition (26). Therefore, the TBI method continues to be an important tool in ecosystem research because it facilitates standardized assessments of litter decomposition, allowing researchers to compare the effects of land use, management practices, and environmental change across diverse ecosystems.

Land-use change in agriculturally dominated catchments is often driven by economic incentives, but it may also arise from efforts to restore or conserve ecologically sensitive areas. Catchment-scale studies therefore provide a critical framework for understanding spatial patterns in soil processes and for generating baseline information on soil health in both agricultural and conservation-oriented landscapes. In this context, the aim of the present study was to quantify litter decomposition and stabilization rates across contrasting land-use types, including forest, grassland, vineyard, and cropland, and to identify the environmental factors controlling their spatial variability. In addition, we examined the influence of slope position, and thereby soil redistribution processes, on S and k parameters. We hypothesized that S would vary systematically among land use types and slope positions due to differences in vegetation cover and soil properties associated with soil redistribution processes. Furthermore, we expected soil physical and chemical properties to be significant predictors of litter stabilization, reflecting spatial variability in soil health across the catchment.

2 Materials and methods

2.1 Site description, meteorological information

The present study was located in a small agricultural catchment, Upper Balaton, Hungary (46.929948N, 17.655593E). In the analysis, SOC degradation and stabilization were measured at different land use types of cropland, vineyard, grassland, and forest (Figure 1) during the vegetation period of 2020. These land use types represent over 85% of the entire catchment area. Topographic elevation at the site varies between 220 (cropland) and 284 m (forest) above sea level.

Continental climate is dominant in this region with Mediterranean and oceanic influences. Normally hot and dry in the summer season, cold in winter, and most precipitation occurs during the spring and fall months (27). The region is characterized by a moderate rainfall deficit, with a mean annual temperature of 10.9 °C and mean annual precipitation of 635 mm during 2016–2020. During 2021–2025, the mean annual temperature increased to 11.3 °C, while mean annual precipitation decreased to 432 mm, indicating a warming and drying trend in the regional climate. Annual sunlight is around 1970–2000 hours. To determine the severity of drought during the study period, the Selyaninov hydrothermal coefficient (HTC) was calculated based on Equation 1. Moisture conditions were classified from the HTC according to the thresholds reported in Bartold et al. (28), where HTC < 0.4 → very dry; 0.4–0.7 → dry; 0.7–1.0 → moderately dry; 1.0–1.5 → sufficient moisture; 1.5–2.0 → humid; 2.0 → very humid. The conditions along with meteorological data during the study period are presented in Table 1.

Table 1. Precipitation sums, air temperature values, and Selyaninov’s hydrothermal coefficients (HTC) during the general vegetation periods at the study catchment.

Minimum and maximum temperatures were based on daily average values.

(1) HTC= ∑i=1nP0.1×∑i=1nT

Where P is the total precipitation amount of the given time period n (mm), and T is the mean daily air temperature above 10°C.

The soils at the investigated sites are Calcaric Cambisol, Loamic, according to WRB (29). Further classification of the soils showed Epileptic Cambisol (Loamic) for the grassland and Endoleptic Cambisol (Anoloamic) for the forest (29). The soil type at the Vineyard_1G site is Cambic Calcisol (Anoclayic, Endoloamic). At the Vineyard_2G and Vineyard_2T (the two sites are located next to each other) sites, the soil type is Calcaric Regosol (Pantoloamic, Ochric), with relatively homogeneous loamy loess soil (29).

2.2 Land use types and slope inclines

Four land use types at six locations were studied. Cropland has inorganic fertilizer (carbamide of 100kg N/ha) application annually and crop rotation is being practiced. During the investigation period maize (Zea mays L.) was grown. At the forest site, a mix of sessile oak (Quercus petraea) and black locust (Robinia pseudoacacia) trees are present. The grassland site is known for its shallow topsoil due to long-time soil erosion (the soil layer is approximately 40-45cm deep at parts), and it is normally highly moisture deficient. Three vineyard sites were included in the study. All vineyard sites have the same fertilizer treatments and once a year a light tipping is performed. Weed control is being performed when necessary, at approximately the same time for each vineyard site. There is no irrigation being performed by the stakeholders, all farmlands are rain fed at the catchment.

As the studied catchment is predominantly covered by vineyards, forests, and grasslands on sloping terrain, we investigated SOC degradation at two positions along these slopes, hereafter referred to as the upper (U) and lower (L) slope positions. In vineyard slopes, an additional intermediate position (M) was selected at locations where the slope angle changed (either becoming gentler or steeper) to assess whether the associated changes in soil physical and chemical properties influence the soil microbial communities responsible for SOC degradation. Because vineyards are the most representative agricultural land use type at this catchment, we chose three distinct vineyard slopes (Vineyard_1G, Vineyard_2G, and Vineyard_2T). The differences between the vineyard sites are the following: Vineyard_1G has no-till managed inter-rows with long-term grass strips between rows. Vineyard_1G has a highly eroded slope with an appr. 14% incline. The second vineyard site (Vineyard_2G and Vineyard_2T) is located on a less steep slope (appr. 8%); however, it has different inter-row soil management (Figure 1). Vineyard_2G has no tilled inter-row with long-term grass strips between rows, while Vineyard_2T has tilled inter-row soil management with no vegetation present between rows. All three vineyard sites have white wine types of grape varieties, and all plant management methods are performed in the same manner (e.g., fertilizer additions, trimming events, harvest). As the cropland site is located in the catchment valley, no slopes were present (< 5%), hence only one location was studied.

2.3 Soil physical and chemical analyses

Soil samples for physical and chemical determinations were collected from the top 20 cm soil layer. Particle size distribution was determined using the sieve–pipette method. Briefly, a soil–water suspension was prepared in a sedimentation cylinder, and aliquots were collected with a pipette at predetermined times to obtain particle-size fractions based on their settling velocities.

Samples were homogenized, air-dried, sieved (<2 mm), and analyzed for total nitrogen (N_Tot), NH₄⁺-N, NO₃^--N, K₂O (ammonium lactate-Al soluble), P₂O₅ (Al soluble), soil organic carbon (SOC) content, electrical conductivity, and pH_H2O. The amount of N_Tot was determined using the modified Kjeldahl method (ISO 11261:1995) and the amount of SOC was measured by wet digestion using the Tyurin method (30). K₂O and P₂O₅ measurements were conducted using an inductively coupled plasma optical emission spectrometry (Quotation ICP-OES, Ultima 2, Thermo Fischer Scientific, Waltham, MA, USA) after ammonium lactate (Al) extraction. As the forest upper point had a high amount of plant tissues in the soil samples, the chemical parameters were measured but were excluded in several analyses due to data uncertainty and extremity. The soil pH was measured in 1:2.5 soil-to-water suspensions. Soil element concentrations are reported as mg kg^-1 dry-weight soil.

2.4 Tea bag index, tea bags used, and data analyses

The activities of soil microbiological communities were measured through decomposition and stabilization rates using a standardized method of the TBI (21). Green and red tea filters in non-woven polypropylene tea bags were purchased from the Netherlands (www.dutchsupermarket.com). Together 16 points were investigated in the catchment using a total of 320 teabags at the beginning of the study. Tea bags were buried at 8 cm below the soil surface. All bags were weighed and labeled before burial. Tea bags were left for the full vegetation growth period (from mid-May 2020 to early September 2020) as during the time of the study unexpectedly dry meteorological conditions were present (Table 1). 10 pieces of both types of tea filters were placed at each location approximately 1 m apart from each other to capture possible soil heterogeneities and to minimize any misleading or lost data (e.g., presence of mole’s burrow or ant colonies). All recovered tea bags were cleaned and air dried prior to measurements. All bags were opened, and the remaining tea materials were measured separately.

Calculations for k and S were based on Equations 2–4 after Keuskamp et al. (21):

(2) Wr(t)=are−kt+(1−ar)

(3) S=1−ag/Hg

(4) ar=Hr(1−S)

where Wr(t) is the substrate weight of rooibos after incubation time (t in days), a_r is the labile, and 1 − a_r is the recalcitrant fraction of the substrate described by the decomposition rate constant k. S is the stabilization factor, a_g is the decomposable fraction of green tea (calculated from the mass loss), and H_g is the hydrolyzable fraction of green tea. The decomposable fraction of rooibos tea is calculated using Equation 4 based on the hydrolyzable fraction (H_r) and the stabilization factor S. With W_r(t) and a_r known, k was calculated using Equation 2.

2.5 Tea bag burial

All six study sites received the tea bags in mid-May and were collected in early September 2020, a few days apart. For the cropland sites tea bags were placed in the soil approximately two months after tillage and when the maize plants were in the two-leaf growth stage, while tea bags were collected prior to harvest. At the vineyard sites, tea bags were placed in the soil in May 2020 prior to plant flowering stage, and the tea bags were recovered after harvest, while the plants were still active. At the grassland and forest soils, both upper and lower slope positions received tea bags with no intermediate study points. However, in the forest soils, some animal damage to the filters resulted in a second tea bag burial event, approximately two weeks after the first event. The days were adjusted for the S and k calculation accordingly.

2.6 Statistics

All statistical calculations were performed using the software package R (R Core Team, Version 4.0.2), with statistical significance determined at p < 0.05. The factors of stabilization and degradation rate constants at different land use types were analyzed using nonparametric statistical analyses. Data normality was assessed via the rstatix (v0.7.2) package using the Shapiro-Wilk test, which revealed that the analyzed variables did not follow a normal distribution. To evaluate significant differences among group medians, the Kruskal-Wallis test was selected, which was followed by Dunn’s post-hoc multiple comparison test using the rstatix (v0.7.2) and multcompView (v0.1.10) packages to identify specific pairwise differences. Pearson’s correlation coefficients (r) were computed using the Hmisc (v5.2.3) package to evaluate the linear correlation between the soils’ physical and chemical characteristics, S and k factors, and land use types or slope positions. The multivariate data analysis was performed by principal component analysis (PCA) using the vegan (v2.7.1) package to evaluate the relationship between the measured soil parameters and S and k factors. All figures and plots were generated using ggplot2 (v3.5.2) and ggpubr (v0.6.1).

3 Results

3.1 Soil physical and chemical characteristics of the land use sites and slope positions

The particle size distribution and the average soil water content at the time of soil sampling are presented in Table 2 for all 16 sampling points. From the physical data analyses, we excluded the forest upper points’ values due to some uncertainties resulting from high amount of plant residues in the soil samples. To compare land use types, values were averaged across all sampling points from the different slope positions. Based on these averages we found that cropland and Vineyard_1G had the lowest sand contents (11.5 and 14.8%, respectively) and the highest clay contents (42.2 and 43.7%, respectively). The highest sand contents were observed for Vineyard_2T and Vineyard_2G (25.2 and 27.7%, respectively) along with the lowest clay contents (17.1 and 16.7%, respectively). Soil textures within slopes also showed high variations (Table 2). For example, most upper points were classified as silty clay loam while the lower points had higher sand contents. Vineyard_1G samples showed increasing clay contents toward the bottom of the slope (Table 2), where the steepness of the slope was over 14%.

Table 2. Soil physical characteristics of the investigated sites (0–20 cm layer).

G indicates no-tilled grass inter-row, T indicates tilled inter-row soil. U, M, and L represent upper, middle, and lower points of the given slope, respectively. Values from land use marked with an asterisk (*) were measured in soil samples containing high amounts of plant litter residues, which may have affected the measurements. Volumetric water contents (VWC) and soil temperatures (Soil Temp) were measured at the time of filter collection from the upper soil layer. Different lowercase letters indicate significant differences between slope positions within each land use type at p < 0.05. n=3; ± SD.

The basic soil chemical characteristics of the investigated land use types are summarized in Table 3. Again, to compare the different land use types, values were averaged across all sampling points from the different slope positions. Forest and grassland had slightly acidic or close to neutral soils (pH = 6.24 and 6.95 respectively), while the cropland and all vineyard samples’ pH varied between 7.41 and 7.76. Grassland and forest soils also had the highest total N, NH₄⁺-N, SOC contents, C/N ratios, and the lowest P₂O₅ content among the sites (Table 3). Particularly high K₂O contents were observed in the Vineyard_1G samples, which receives biennial or annual fertilizer treatments.

Table 3. Chemical characteristics of the investigated land use sites (0–20 cm layer).

SOC represents soil organic carbon values.

G indicates no-tilled grass inter-row soil and T indicates tilled inter-row soil. U, M, and L represent upper, middle, and lower slope positions, respectively. Values from land use marked with an asterisk (*) were measured in soil samples containing high amounts of plant litter residues, which may have affected the measurements. Different lowercase letters indicate significant differences among slope positions within the same land-use type at p < 0.05. Data was collected in 2020. n=3; ± SD.

3.2 Decomposition rate constants and stabilization factors of the land use types and slope positions

A clear distinction of both decomposition rate constant and stabilization factor between the investigated land use types after the incubation period was observed (Figure 2). The measured k ranged from 0.0036 to 0.0451 d^–1, with an average of 0.014 ± 0.007 d^–1 for all sites and slope positions. The S ranged from 0.085 to 0.514 with an average of 0.320 ± 0.108 for all study points (Figure 2). We found a moderate, positive correlation between the k and S values (r = 0.50; p < 0.001).

The overall decomposition rates and stabilization factors for each land use type are shown in Figures 3a, b, respectively. The rate constants were similar between the land uses, except for the grassland where the values were significantly lower than most of the other land uses (Figure 3a). The highest rate constants were observed for Vineyard_2G (0.0159 d^-1).

For the stabilization factor, more distinguishable differences were observed (Figure 3b). Vineyard_2T had the highest, while the cropland had the lowest overall S values (0.400 and 0.161, respectively), with most cases significant differences between land use values (p < 0.05). Specifically, Vineyard_2G and Vineyard_2T, and the Vineyard_1G and grassland sites showed similar overall values (p > 0.05).

3.3 Principal component analysis and correlations

The PCA explained 67.2% of the total variance in soil properties and Tea Bag Index parameters, with PC1 and PC2 accounting for 46.8% and 20.4% of the variation, respectively (Figure 5). Land-use types showed clear separation along the principal component axes. PC1 primarily separated vineyard soils from grassland, forest, and cropland sites. Vineyard_2G and Vineyard_2T were associated with positive PC1 scores and were aligned with higher k, S, and greater sand and silt contents. In contrast, grassland and forest sites clustered at negative PC1 values and were associated with higher SOC and total N concentrations. Cropland occupied an intermediate position near the center of the ordination.

Moderate, positive correlations were noted between S and k factors (r = 0.50) when all collected data was used (n = 49). However, when the land-use types were analyzed separately, strong and significant positive correlations were found for all six land-use types: Vineyard_1G r = 0.76, Vineyard_2G r = 0.90, Vineyard_2T r = 0.73, and grassland r = 0.94 (Supplementary Figure 1). However, these correlations were derived from individual site data, resulting in relatively small sample sizes (n ~ 9), which may reduce the reliability of the correlation estimates. Soil physical properties differed among land use types and were strongly associated with the observed S values. Based on the full dataset, lower clay contents were associated with higher S values (r = -0.69; Figure 5). For the silt loam soils with less than 20% clay content, the S values exceeded 0.30. In contrast, sand content was positively correlated with S (r = 0.63), indicating that soils with higher sand content tended to exhibit greater stabilization factors. When analyzed by each land use type separately, weak to moderate correlations between S and sand content were observed (Supplementary Figure 1). Decomposition rate constants showed a weak relationship with the measured soil chemical properties. However, analyses conducted separately for each land use type revealed stronger relationships between soil particle-size distribution and k. Significant correlations with sand and clay contents were observed for grassland (r = 0.71 and r = -0.60, respectively; Supplementary Figure 1). However, due to the low sample sizes, these results should be interpreted with caution.

To study the relationships between soil chemical parameters and litter decompositions, we also analyzed each land use site separately. Our findings revealed stronger relationships between litter decomposition parameters and soil chemical properties than overall values. Cropland exhibited the strongest positive correlations between both S and k and several soil chemical variables, including total nitrogen (total N; r = 0.49 and 0.67, respectively), NH₄⁺–N (r = 0.93 and 0.90), and soil pH (r = 0.81 and 0.75). Similarly, grassland showed strong positive correlations between S and k and N_Tot, SOC, K₂O, and P₂O₅, while S was additionally positively correlated with NH₄⁺–N and NO₃^–N. In contrast, negative correlations were observed between S values and N_Tot, SOC, and K₂O in the Vineyard_2T (r < -0.55, p < 0.06). Positive, although non-significant, relationships between S, k, and soil pH were observed in Vineyard_1G (r > 0.55, p > 0.115).

4 Discussion

4.1 Influence of land use types on litter stabilization and decomposition rates

Plant litter decomposition is a vital process of the soil nutrient cycle in the ecosystems, where land use might greatly influence decomposition rates. While decomposition in the field mainly depends on the type of plant materials to be decomposed (31), the present experiment used the same types of substance (i.e., green and rooibos tea) to investigate the soils’ potentials to stabilization and decomposition rates. Therefore, in the present study the roles of the soil physical and chemical compositions were investigated in the overall decomposition processes. We found that land use types influenced both decomposition rates and stabilization factors of the soils. However, land use exerted stronger control on litter stabilization than on decomposition rates. Decomposition rates varied moderately while S differed substantially among land uses, indicating that land management and associated soil properties might exert a greater influence on the decomposed organic matter than on the decomposition process itself. Grassland had significantly lower decomposition rates than all other land use types except Vineyard_1G site. Stabilization factor was the lowest for the cropland and the highest for Vineyard_2G, Vineyard_2T, and forest, indicating that a larger proportion of the hydrolysable litter fraction became stabilized under the environmental conditions present at these sites. Fu et al. (32) investigated four land use types, and, contrary to our results, found that grassland had the highest S and k values, further highlighting that litter decomposition responses to land use are strongly site dependent. However, the values of S and k might differ under different precipitation regimes (33) or seasonality (20), indicating that our sites with high air temperatures and prolonged drought conditions might cause reduced decomposition values. Although k and S describe different components of the TBI framework, the positive association observed in our study is not unprecedented, similar relationships have been reported under varying ecosystems (34–36). However, k and S are not entirely independent because they are derived from the same mass reduction measurements, and biased positive relationships might occur (37). Therefore, the positive relationship between k and S observed in our study might resulted from the combined effect of environmental controls and the inherent properties of the TBI methodology. The relatively narrow range of k values across the catchment and the limited number of significant pairwise differences suggest that decomposition varied less among land-use types than litter stabilization. This interpretation is consistent with the PCA, in which k was located near the center of the ordination space and showed only weak associations with the measured soil physical and chemical properties. Similarly, Hayes et al. (15) found weak correlations between TBI and soil nutrient levels. The only land use type in which significantly lower decomposition rates were observed was grassland. This research site is characterized by shallow topsoil and persistent moisture deficiency resulting from long-term erosion. These conditions likely limit microbial activity, despite the soil’s relatively high organic carbon and nitrogen content (38). In water-deficient environments, microbial metabolism often depends more on moisture deficiency than on substrate deficiency, which reduces decomposition rates even in the presence of abundant organic resources (33). Since all agricultural systems within the catchment were rain-fed, spatial differences in soil water availability may have exerted a stronger influence on decomposition than soil nutrient status.

4.2 Soil texture, soil cover, and slope position affecting litter stabilization

Soil physical properties strongly influenced the S values, with lower clay contents and higher sand contents associated with higher S values. In contrast, no significant correlations were observed between k and soil particle size distribution. Soil texture, particularly clay content, can substantially affect soil moisture retention, porosity, and chemical properties, thereby influencing litter decomposition and organic matter stabilization processes. While our study showed strong negative relations among S and clay content, strong positive correlations between soil clay content and organic matter stabilization have been reported previously (39), while clay content might not be a factor in other cases (40, 41). The contrasting relationship observed in our study may reflect the fact that soil texture did not act independently but was closely associated with land use type, management intensity, soil moisture content, and aggregate stability. Previous studies at the Vineyard_2 research site found that higher moisture content was associated with lower clay content; however, the aggregate stability remained consistently higher in areas with a more sandy texture (10). This finding suggests that factors beyond clay content contribute to soil structural stability at our study sites. These findings further highlight the importance of local soil characteristics in evaluating litter decomposition. The higher stabilization factors observed in the no-till vineyard with permanent grass strips suggest that vegetation cover contributes to favorable conditions for retaining litter-derived carbon. Permanent cover vegetation may enhance aggregate formation, improve soil structure, reduce erosion, and promote microsites that facilitate carbon retention (42–44). Similar to Vineyard_2G, Vineyard_2T also showed relatively high stabilization values, indicating that soil parent materials, texture, and topographic controls may partially compensate for the soil disruption effects of tillage.

Variations in erosion intensity, sediment deposition, soil moisture distribution, and aggregate formation along the slopes can alter the transformation and retention of decomposed organic matter while not significantly affecting overall decomposition rates (45, 46). In our study, the absence of topographic effects on k indicates that decomposition processes were relatively resistant to spatial variation within individual vineyard slopes. In opposition, the significant variability of S suggests that stabilization processes are more sensitive to local changes in soil conditions associated with topography. Although the investigated hillslopes were relatively short (approximately 100 m in length), the upper and lower sampling points at the Vineyard_2 site were located approximately 500 m apart. Therefore, differences among sampling locations may not solely reflect local slope effects but also broader-scale soil redistribution processes.

4.3 Soil chemical parameters affecting decomposition rates

The different land uses are characterized by contrasting management practices. Fertilizer application and conventional tillage are regularly applied in croplands, whereas forests and grasslands remain largely undisturbed. Independent microbiological investigations conducted at the same experimental sites demonstrated that these land use specific management practices significantly altered soil bacterial community composition (38, 47). In 2020, 16S rRNA sequencing of soil samples collected from the investigated sites revealed that the bacterial communities were dominated by members of the phyla Pseudomonadota (formerly Proteobacteria), Acidobacteriota, Actinobacteriota and Bacteroidota. A similar community composition can be expected during the present experiment. These bacterial groups have contrasting ecological strategies: Pseudomonadota and Bacteroidota are generally regarded as copiotroph taxa that rapidly utilize fresh organic substrates (48), whereas Acidobacteriota are typically associated with oligotrophic conditions and slower carbon turnover (49). Actinobacteriota plays an important role in degrading more complex plant polymers such as cellulose and lignin and also contributes to the formation of microbial residues that may become stabilized in soil organic matter (50, 51). Although microbial communities were not analyzed in the present study, these previously reported differences provide a biologically plausible mechanism for variation in TBI parameters among land uses. therefore changing the decomposition rates of plant resides and other organic matter in soils (52). SOC was the lowest at the cropland with similar nitrogen contents as found in the vineyards. These sites receive external nitrogen and phosphorous inputs that may increase nutrient availability for decomposer organisms and thereby promote litter decomposition (53–55). The added inorganic fertilizer type (i.e., carbamid) to our cropland site is known to increase maize yield with conventional tillage method (56); however, it disrupts the microbial communities in the soils and can cause reduced microbial diversities (38, 57). The inorganic fertilizer addition can enhance litter decomposition rates in soils (52, 58), which can partly explain our highest k values observed for the cropland. This interpretation is consistent with previous microbiological studies from our research sites, which demonstrated that cropland soils supported bacterial communities distinct from those of grassland and forest (38, 47). Because copiotroph bacterial groups, particularly members of the Pseudomonadota and Bacteroidota, rapidly metabolize easily degradable plant-derived carbon (59), increase in their density might contribute to faster litter decomposition. While annual fertilizer addition is occurring in our cropland site, the nitrogen content was not significantly different from the other investigated sites’ value (0.17%), which is historically typical to our study site (60). The relatively high decomposition rates and low stabilization factors at the cropland site indicate the potential for low carbon accumulation in these ecosystems.

Our findings showed negative relationship between S and SOC, N_Tot, and NH₄⁺-N concentrations. Forest and grassland soils had the highest organic matter contents; however, these variables were not positively associated with litter stabilization. This observation might imply that factors or processes controlling accumulated SOC stocks are different from those governing the stabilization of recently introduced organic substrates. The negative association between S and clay content further supports this interpretation. While clay minerals are widely recognized as important agents of long-term SOC protection (61, 62), the stabilization factor derived from the TBI primarily represents the short-term transformation of labile litter into a relatively recalcitrant fraction. Therefore, the mechanisms responsible for litter stabilization during the incubation period differ from those controlling the persistence of mineral-associated organic carbon over decadal timescales.

5 Conclusions, limitations, and future work

Decomposition rates remained largely invariant across land uses and slope positions, whereas stabilization processes responded strongly to both land use and topographic position. Hence, erosion and soil physical properties could have a greater influence on carbon retention than on carbon turnover. The strong positive association between SOC and N_Tot highlights the close linkage between carbon and nitrogen cycling in soils at our study sites. Under these conditions, a greater fraction of tea litter may be decomposed and mineralized rather than incorporated into stabilized soil organic matter, resulting in lower short-term stabilization rates as measured by the TBI. Consequently, long-term SOC accumulation and short-term litter stabilization might be considered complementary, yet distinct, indicators of soil ecosystem functioning.

The results from our study help to improve our understanding of how different land use systems influence soil carbon dynamics, litter decomposition, and stabilization processes, thereby contributing to a more comprehensive assessment of soil ecosystem functioning in agricultural and semi-natural environments. However, some limitations should be acknowledged. Our measurements were based on a single sampling period, whereas decomposition and stabilization processes exhibit temporal variability driven by seasonal changes in climate and biological activity. The changes in climatic conditions already show a decreasing trend in precipitation amounts and increase in air temperature, which would make research findings from such studies as ours valid for a decade or two. Therefore, repetitions or long-term monitoring would enable us to see prolonged differences among the research sites’ litter decomposition processes. While the present study mainly focused on soil physical and chemical parameters affecting decomposition rates, biological, especially microbiological differences could help better explaining some of the differences among sites and measurement points, hence for the long-term studies, parameter expansion would enable more founded deductions. Some background microbiological data are available for the study sites; however, collecting soil samples from the same burial sites at the time of tea bag collection and analyzing them using the newest sequencing technologies (e.g., full-length sequencing, shotgun metagenomics) would provide a more robust basis for explaining the observed differences in stabilization factors and decomposition rates among land use types.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. The authors declare that generative AI tools were used solely for language editing and grammar improvement.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fsoil.2026.1908668/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Pu Shi, Jilin University, China

Reviewed by: Yangyang Jia, Xinjiang University, China

Wenju Chen, Northwest A and F University, China