Responses of soil organic carbon cycle to land degradation by isotopically tracing in a typical karst area, southwest China

Study area

The sampling area situates in Yinjiang county within the territory of Tongren city, the northeast of Guizhou Province (Fig. 1), which is characterized by karst landform with abundant mountains. Yinjiang county widely stretches across 700–800 m in elevation, with an elevation difference exceeding 2,100 m (Han & Xu, 2022). According to the statistical results from the Yinjiang government, the forest coverage in Yinjiang county reached 68.37%. Statistics from the bureau reported that the dominant economic industry in Yinjiang county was agriculture; the proportion of crops seeded reached 64,471 hectares and that of cereals covered 37,833 hectares in 2016. The soils are classified as calcareous soils by the United States Department of Agriculture (USDA) classification standard, which are mostly derived from carbonate rocks with sporadic coal seams and silicate rocks. The sub-tropical monsoon climate of Yinjiang county is notable for its significant seasonality and simultaneous variations in precipitation and temperature. Most rainfall occurs between April and October each year (>70%). The temperature varied from −2.2 °C to 38.6 °C, and the annual precipitation was 1281.4 mm in 2016.

Field sampling

Large amounts of cropland have been abandoned or converted back to natural land resulting from the ‘Grain-for-Green’ program in China in the last 20 years. The native vegetation and abandoned cropland are vastly distributed in Yinjiang county, which proved useful for examining how agriculture affects soils. Three land-use soil profiles were chosen to be sampled in Sept. 2016 (Fig. 1). The sum of 20 samples was obtained in the profile of secondary forest land (SF), 16 samples were collected in the profile of abandoned cropland (AC), and 11 samples were collected in the profile of shrubland (GS). The description of sampling methods and soil profile characteristics have been reported by Han, Zhang & Xu (2023) and Han & Xu (2022). The sampling intervals were defined as 5 cm in shallow soils (above 20 cm), and as 10 cm in soil depths below 20 cm. The descriptions of sampling sites are provided in Table 1.

Soil composition measurements

After plant residues and stones were removed from air-dried soil samples, two groups were divided. Following the passing of the ground soils through a two mm mesh sieve, one group of samples was used for the analysis of soil chemical properties. The soil and bedrock samples were digested with HNO₃–HF–HClO₄, and the Ca contents were determined by ICP–OES (Optima 5300DV; Perkin Elmer, Waltham, MA, US). The suspension of soil and water (2:5) was employed to determine soil pH by a glass electrode with a precision of ±0.05 (Han et al., 2020). Soil particles (0.002 mm for clay, 0.053 mm for silt, and 0.250 mm for sand) were classified by the USDA Soil Taxonomy, and separated in wet-sieving by a laser particle size analyzer (Mastersizer 2000; Malvern Panalytical, Malvern, UK) (Van Bavel, 1950). Residual soils were ground for particle size less than 150 µm, then digested for 24 h with 0.5 mol/L HCl and repeatedly washed with pure water (18.2 M Ω cm) until neutrality was reached to eliminate the disturbance of carbonates (Liu, Han & Zhang, 2020). Subsequently, samples were dried at 55 °C to measure SOC and δ¹³C_SOC. The SOC contents were obtained on an elemental analyzer (Vario TOC cube; Elementar, Langenselbold, Germany) with a precision of ±0.1%. In addition, the actual SOC contents were calibrated due to the loss of inorganic carbon: ${\displaystyle \mathrm{SOC}=SO{C}_M\times \frac{M_1}{M_2}\times 10}$

where the SOC_M (%) represents the measured values of the soil organic carbon (g/kg), the M₁ represents the mass of soils after removing carbonates, and M₂ represents the mass of soils before removing carbonates.

Isotopic measurements

The carbon isotopes were measured by a stable isotope mass spectrometer (Thermo, MAT-253, USA) with a precision better than 0.1‰. A duplicate measurement was performed on each sample to ensure accuracy. The δ¹³C_SOC (¹³C_SOC/¹²C_SOC) values were relative to Vienna Pee Dee Belemnite (V-PDB), which is expressed in the unit of‰: ${\displaystyle {\delta }^{13}{\mathrm{C}}_{\mathrm{SOC}}=\frac{{\delta }^{13}{\mathrm{C}}_{\mathrm{M}}-{\delta }^{13}{\mathrm{C}}_{\mathrm{V}}}{{\delta }^{13}{\mathrm{C}}_{\mathrm{V}}}\times 1000}$

where M means measured samples, and V means V-PDB standard. All measurements were accomplished in the Institute of Geographic Sciences and Natural Resources Research, CAS.

Calculation and data processing

Soil erodibility factor

Soil erodibility denotes the resistance to erosion that is induced by external influence, which can be evaluated by the soil erodibility K factor (Kiani & Ghezelseflo, 2016). Depending on the physical and chemical properties of soils, soil erodibility presents regional characteristics within a different area. Based on soil properties (e.g., soil structure, SOC contents, and soil texture), several experiential models were established (Favis-Mortlock & Boardman, 1995; Pal & Chakrabortty, 2019). The Erosion Productivity Impact Calculator (EPIC) model calculates K (t ha h(ha MJ/mm)) using soil texture and SOC contents, and equation (4) has been modified to accord for Chinese soils (Zhang et al., 2008):

$\mathrm{K}=0.5158\left\{0.2+0.3\times exp\left[-0.0256{\mathrm{S}}_{\mathrm{a}}\left(1-\frac{{\mathrm{S}}_{\mathrm{i}}}{100}\right)\right]\right\}\times {\left(\frac{{\mathrm{S}}_{\mathrm{i}}}{C_l}+S_i\right)}^{0.3}$

$\times \left[1-\frac{0.25\mathrm{C}}{{\mathrm{C}}_{\mathrm{SOC}}+exp\left(3.72-22.95{\mathrm{C}}_{\mathrm{SOC}}\right)}\right]\times \left[1-\frac{0.7{\mathrm{S}}_{\mathrm{A}}}{{\mathrm{S}}_{\mathrm{A}}+exp\left(-5.51+22.9{\mathrm{S}}_{\mathrm{A}}\right)}\right]-0.01383$

where the S_a, S_i, C_l represents the percent of sand, silt, and clay in the total soil particles; S_A represents the values of 1-S_a/100, and C_SOC represents the SOC contents (%). And the units of K factors were omitted for simplifying analyses. The higher values of K factors indicated stronger soil anti-erodibility.

Data processing

The relationships among SOC, δ¹³C_SOC and related parameters were analyzed by linear regression analysis, with the determination of the coefficient R and p-values by IBM SPSS Statistics 25, USA. All graphs were created by ArcMap 10.8 (Esri, Redlands, CA, USA) and Origin 2017 (OriginLab, Northampton, MA, USA).

Profile characteristics under different land uses

Soil pH, particle distribution and MWD values are listed in Table S1. The detail of soil pH in three profiles was reported by Han, Zhang & Xu (2023): ranged from 7.1 to 7.9, 4.8 to 5.2, and 6.3 to 7.0 in the SF, AC, and GS profiles, respectively. The descriptions primarily focus on the soils above 30 cm) because the cycling of SOC is basically active in the surface soil. The distributions of soil particles in the surface soil (0 to 30 cm) of the three profiles are presented in Fig. 2. Obviously, silt occupied the largest proportion of soil particles among the three profiles. The sand in the AC profile occupied a larger proportion than that in others and decreased with increasing depth in the surface soil. In addition, the soil textures were uniform in the SF and GS profiles compared to those in the AC profile (Table S1). The mean values of MWD were in the order of that in the AC profile (2.14 mm) >GS profile (1.31 mm) >SF profile (1.24 mm). The values of MWD decreased above 100 cm and increased below 100 cm with depth in the AC profile, but little difference was observed between the SF and GS profiles.

SOC content, δ¹³C_SOC value, and K factor

The vertical distributions of SOC contents, δ¹³C_SOC values, MWD values, and K factor are illustrated in Fig. 3. The SOC contents ranged from 3.84 g/kg to 42.16 g/kg, 5.11 g/kg to 17.67 g/kg, and 7.63 g/kg to 108.19 g/kg in the SF, AC, and GS profiles, respectively. At the same soil depth of over 30 cm, the order of SOC and calcium (Ca) contents in the profiles was GS profile >SF profile >AC profile. A significant reduction in SOC content was observed with increasing depth below 30 cm in the SF and GS profiles, suggesting that human disturbance had a limited impact at greater depths. The values of δ ¹³C_SOC ranged from −25.00‰ to −22.30‰ in the SF profile, −24.97‰ to −22.52‰ in the AC profile, and −26.49‰ to −23.82‰ in the GS profile. High increasing rates of δ¹³C_SOC values were found in the 0 to 30 cm soil layers of three profiles. The values of δ ¹³C_SOC roughly showed an upward trend in the upper soils (above 50 cm) of the SF profile (rising by 2.55 ‰), while showing a downward (decreasing by 2.69‰) towards the bedrock (δ¹³C_SOC = −26.73‰) at deeper depths. In contrast, the δ ¹³C_SOC values consistently increased with increasing depth in the GS profile (an increment of 2.68 ‰). Apart from the topsoil, the variation of δ ¹³C_SOC values was slight with the range of 1.53‰ in the AC profile. The δ¹³C_SOC values in the GS profile were obviously lighter than those in the other profiles at the same depth. The δ¹³C_SOC values in the soils over 70 cm of AC profile were similar to those in the SF profile at the same depth. The values of K factor were in the range of 0.0262 to 0.0266, 0.0243 to 0.0259, and 0.0262 to 0.0264 in the SF, AC, and GS profiles, respectively. The K factors exhibited an increasing trend with depth in the SL profile and a slight increase up to 100 cm followed by a decrease in the AC profile. Notably, a significant variation in K factors was observed between the soils above 30 cm (0.0263 to 0.0266) and those below 30 cm (0.0293 to 0.0306) in the GS profile.

Effects of land use on SOC contents

The dynamics of SOC are mainly controlled by covered vegetation, soil structure, and soil erosion, which may affect the input, decomposition, and loss of SOC, respectively (Han, Li & Tang, 2015; He et al., 2016). The impact of lithology on the spatial distribution of SOC contents within a region is expected to be limited, given the similarity in soil mechanical compositions (Doetterl et al., 2015). All soil profiles were selected at a small catchment with similar soil parent materials and climate. Therefore, the soil evolutions affected by land uses may be the key factor affecting SOC contents. The vertical variations of SOC and Ca in three profiles are presented in Fig. 3. Consistently, the contents of Ca followed the decreasing sequence of GS profile (mean values: 26.33 g/kg), SF profile (mean values: 13.05 g/kg), and AC profile (mean values: 0.47 g/kg). Generally, the stable formation of organic–inorganic complexes with Ca²⁺ in calcareous soil can strongly influence the stability of sequestrated SOC (Chen et al., 2012). The lower contents of Ca also may induce the loss of SOC. The possible explanation for the highest levels of SOC in the GS profile and the lowest levels in the AC profile, except for the SOC contents in 100 cm soil, may be attributed to the calcareous organic complexes.

Abundant microorganisms exist in the natural soils, especially in the surface soil, which can decompose SOC (Chen et al., 2002). The rapid decrease of SOC contents of the soil over 30 cm may result from the decomposition by abundant microorganisms (Han, Li & Tang, 2015). Plant residues provided continuous supplies of organic carbon to topsoil, and the SOC contents remained relatively constant due to the improvement of soil and water conservation by plant root (Sun, Zhu & Guo, 2015). The covered vegetation in the location of SF profile effectively improves the capacity of soil and water conservation to maintain SOC. Moreover, there was abundant plant residue to supply SOC to the soils in the SF profile. Although the SOC may be decomposed by soil microorganisms, some ions (e.g., calcium and magnesium) can be complexed in a more stable form afterward (Heckman et al., 2009). The study area is characterized by carbonate rocks, where the natural soils have abundant calcium and magnesium to be conducive to fixing SOC (Li et al., 2017).

Except for the topsoil, the SOC contents were basically identical in the whole AC profile (range of 3.22 g/kg). The higher level of SOC in the topsoil may be attributed to the covered ruderal. After consuming SOC for growth, cultivated crops will never return to the soil at maturity. Furthermore, the decomposition rate of SOC is possibly improved by microbial biomass (Geisseler, Linquist & Lazicki, 2017). Additionally, the exposure time of soil will increase by frequent repeat tilling, and further consume SOC by decomposition and soil mineralization (Schjønning & Thomsen, 2013; Wang et al., 2022c). A mass of acidic compounds (e.g., ammonium chloride and urea) will be added to the soil during the fertilizer application, which may decrease the soil pH (Čakmak et al., 2014). Ca and heavy metals are easily released into soil solutions in an acidic environment (Čakmak et al., 2014). Frequent irrigation can promote the loss of Ca, Mg, and heavy metals (Hall et al., 2001). Furthermore, the finer soil particles prefer to limit the external disturbance rather than accumulate the amount of SOC (Schjønning & Thomsen, 2013). Therefore, the SOC in the AC profile was difficult to maintain and store. The results indicated that nonnegligible SOC might be consumed in long-term cultivation, and difficulty recovering to the original level.

The location of the GS profile only experienced goats grazing in five years with no cultivated activities. Theoretically, the stock of C aboveground will decrease because the uptake of goats consumes part of returned SOC to the soil by the plant. However, SOC was concentrated in the surface soil (0–30 cm) of the GS profile. The SOC content was contained in the highest level of the topsoil (0–5 cm, 108.19 g/kg), which was 1.53 times greater than that in the 5 to 10 cm soil layer and 2.25 times greater than that in the 10 to 15 cm soil layer of the GS profile. Successive excretion of feces may increase biological productivity by abundantly supplying nitrogen, further increasing the uptake of CO₂ by plants, and stimulating root biomass growth (Francisco et al., 2022; He et al., 2015). And the SOC inputs from plant debris may increase. Moreover, the formation of biogenic aggregates will be improved, which have higher sequestrated contents of nitrogen and carbon than physicogenic aggregates, especially in the surface soil (He et al., 2015). The Ca contents showed the highest level in the GS profile, which may be complexed with SOC in a stable formation. It is notable that the fastest decrease of SOC contents with depth was exhibited in the GS profile among the three profiles. There is also evidence that the effect of SOC contents in soils and roots from solid feces is mainly reflected in the topsoil (Francisco et al., 2022; Loss et al., 2017). The mean value of soil pH was 7.0 in the GS profile, in a neutral environment. In general, the Actinomycetes Strain is suitable for reproduction in a neutral environment, further improving the decomposition of SOC (Heckman et al., 2009). Moreover, the soil microorganisms can decompose SOC more effectively after acclimatizing to the combined environment of feces and soil (Wang et al., 2021). Accordingly, the SOC contents extremely decrease in the shallow soil layer of GS profile.

Carbon cycle under diverse land uses by δ13C tracing

Previous studies found that vegetation type and moisture are the dominant factors of isotope fractionation during plants assimilation (Han, Li & Tang, 2015; Qu & Han, 2022; Yakir & Sternberg, 2000). The surficial debris contributes substantially to the carbon pool in soil (Yakir & Sternberg, 2000). Based on the pathways of carbon turnover during plant photosynthesis, the plants are mainly divided to C₃ plants (use the Benson-Bassham-Calvin Cycle, −32‰ < δ ¹³C_SOC < −22‰) and C₄ plants (use the Hatch-slack Cycle, −17‰ < δ ¹³C_SOC < −9‰) (Calvin & Benson, 1948; Johnson & Hatch, 1969; Kohn, 2010; Yakir & Sternberg, 2000). The C₃ plants were the dominant vegetation in the study area, and the δ¹³C_SOC values in the three profiles ranged from −26.49‰ to −22.30‰. It can be inferred that carbon fractionation in the study area may be controlled by C₃ plants, and the long-term cultivation of potatoes (C₄) may have slightly changed δ¹³C_SOC composition in the AC profile. Additionally, the distributions of δ¹³C_SOC also have diverse characteristics under various land uses in a region (Jou et al., 2021). The turnover rate of SOC and carbon isotope fractionation in the cropland are lower than those in the natural land with limited human activities may be attributed to the lower amount of microorganisms and soil enzyme activity (Ni et al., 2015). The variation range of δ¹³C_SOC values was minimum in the AC profile (varied within 2.44‰), which was slightly smaller than that in the SF (varied within 2.69‰) and GS (varied within 2.68‰) profiles. The slight discrepancy indicates that the terminated cultivation may narrow the variation gap of δ¹³C_SOC values between cropland and natural land. Moreover, the soil parent material also influences the δ¹³C_SOC values in regional soils (Liu, Han & Zhang, 2020). The δ¹³C_SOC values were −26.73‰ and −28.77‰ in the bedrock of the SF and GS profiles, respectively. The proximity of SF and AC profiles, located at a distance less than 1 km, suggested a homogeneity in their geological origins. Therefore, it can be inferred that the δ¹³C_SOC value of the bedrock in the AC profile can be approximated by that in the SF profile. The lower δ¹³C_SOC values in the GS profile relative to other soil profiles may be inherited from bedrock.

The δ¹³C_SOC values are strongly affected by the decomposition of fresh SOM in the surface soil, while the old carbon in deeper soils is stable without evident isotopic fractionation (An et al., 2010). Therefore, the discussion between SOC contents and δ¹³C_SOC values is concentrated on soils from 0 to 30 cm in this study. The SOC contents and δ¹³C_SOC values had a negative correlation in the soils over 30 cm in three profiles (Fig. 4), and only had a positive correlation in the deep soil horizon of the SF profile. Although the utilization of CO₂ during the respiration of soil organisms and plant roots may only trigger slight kinetic isotope fractionation, the selective decomposition of plant biochemicals can also affect the distribution of δ¹³C_SOC in soils (Breecker et al., 2015). Compared with the bulk SOC, the roots are generally enriched in ¹³C, whereas the ¹³C in roots may reduce in relation to that in leaves with plant growth (Philben et al., 2022). The decomposition of SOC with the enrichment of ¹³C cannot be entirely explained by roots decay in all profiles. In addition, it has been reported that the decrease in ¹³C levels was always accompanied by the increase of precipitation through reducing the decomposition rate of soil microorganism (Deng, Liu & Shangguan, 2014; Wang et al., 2022b). The soil microorganism activity may be restrained due to the decrease of the stomatal conductance of plants and dissolved oxygen under drought (annual precipitation lower than about 2000 mm) or excessively humid environments (annual precipitation higher than 3000 mm) (Xu et al., 2016). In contrast, a well-humid environment will increase SOC availability through the decomposition of plant debris and enrich ¹³C in soil by microorganism utilization (Deng, Liu & Shangguan, 2014; Peri et al., 2012). Since 1984, the annual precipitation has ranged from 800 mm to 1500 mm without extreme rainstorms or drought events in September (the sampling month) (Wu et al., 2017), which is suitable for the transport of SOC in the plant–soil–microorganism cycle to accumulate ¹³C. Moreover, ¹³C will continue to accumulate in the soil, along with the death of most soil microorganisms (Henn, Gleixner & Chapela, 2002). The additional input of soil organisms and fungi could make a substantial contribution to the enrichment of ¹³C, leading to the highest rate of increase in δ ¹³C_SOC (δ¹³C_SOC = 2.07‰) in the GS profile compared to the others, for depths over 30 cm. The accumulation of SOC accompanied by the enrichment of δ¹³C_SOC in deep soils of the SF profile may result from a mixing effect. The ¹³C dilution caused by bedrock weathering (δ¹³C_SOC value of −26.73‰) possibly prevents aboveground organisms from controlling SOC compositions.

Assessment of soil degradation in the karst soil

The adverse influence of soil degradation on agriculture has severely restricted the development of the agriculture-driven economy in Guizhou Province. Obvious soil acidification had occurred in the AC profile (soil pH at an average of 5.0). The process of soil acidification may lead to excessive release of Ca, Mg, and heavy metals into soil solution, thereby reducing the storage of SOC and posing a threat to the growth of soil organisms and plants (Čakmak et al., 2014). Mn, Mo, Ni, Fe, and Cr exhibited reduced concentrations in the abandoned cropland compared to those in the uncultivated land as previously reported in Yinjiang county (Han & Xu, 2022). The above-mentioned analyses also demonstrated that agricultural practices significantly impacted the soil ecological system. Therefore, it is imperative to artificially protect cropland after abandoned rather than solely rely on the self-restoration of ecosystems. To better evaluate soil erosion, the soil aggregate stability index of MWD in different depths of three profiles was determined. SOC storage is more affected by soil aggregates rather than soil organisms in deeper soil (An et al., 2010). The MWD and SOC at deep soils (>30 cm) showed a negative linear correlation in the SF and GS profiles, and a negative curve correlation in the AC profile (Fig. 5). Generally, the majority of SOC and decayed plant residues are generally protected by larger soil particles, while part of SOC is immobilized in small soil particles (Adamczyk, Heinonsalo & Simon, 2020). Positive correlations between SOC contents and MWD values have been found in several areas (Gupta Choudhury et al., 2014; Li et al., 2015; Sun, Zhu & Guo, 2015), whereas the study soils showed the contrary. The soil particles were mainly composed of clay and silt (the mean proportions of 98.82%, 90.62%, and 97.34% in the SF, AC, and GS profiles, respectively), resulting in low MWD values in soils. Abundant fine-sized particles improve the sequestration of SOC by promoting the more stable formation of soil aggregates and mineral complexes (Huang & Hartemink, 2020). However, the soil texture will be coarsened in the short term after cropland abandonment because of the decreased capacity of cation exchange and SOC sequestration (Huang & Hartemink, 2020). Obviously, the loss of clay-silt particles was influenced by the cropland abandonment in the AC profile, which has adverse effects on protecting SOC sequestration. Conversely, moderate grazing levels may be conducive to protecting ecological balance and soil fertility. Moderate grazing even may compact the sandy soils to improve the cohesion forces of soils (Wang et al., 2022a).

Besides SOC contents and soil acidification, soil erodibility is also an essential indicator of soil degradation (He et al., 2016). The K factor was determined for evaluating soil erodibility in the three profiles. The range of K factor values (0.0259 to 0.0266) in the study area was in accordance with that of surface soils in Guizhou Province (0.0230 to 0.0477) but lower than that of most surface soils in southern China (0.0635 to 0.2183) (Zhang, De Angelis & Zhuang, 2011). It may be attributed to the high levels of SOC and silt proportion in Guizhou Province. The similar values of K factors in the SF and GS profiles at the same depth with a narrow range, suggest that moderate grazing did not affect soil erodibility. However, the K factors in the AC profile were inferior to those in other profiles, which may result from long-term cultivation on sloping farmland. The plant roots can strongly consolidate soil to improve soil texture, and the canopy of vegetation and litter layer can further reduce the splashing effect by precipitation (Sun et al., 2016). It is indicated that the soils in Yinjiang county are more susceptible to erosion, especially soils that have experienced cultivation. The management of slope farmland possibly increases soil erosion by rainfall erosivity, especially in karst areas (Li et al., 2015). The vertical distributions of SOC and K factor levels were dissimilar in the three profiles (p > 0.05). In general, an increase in sand apportion may reduce soil erodibility by increasing infiltration (Panagos et al., 2014). The lower levels of K factor and SOC on abandoned cropland should be noted. It is necessary to convert the existing form of traditional slope farming in Guizhou Province (such as the terrace). And the soil erodibility of terraces under reasonable management may also be greater than that of woodland (Jiang et al., 2020).