Introduction

Karst areas are unique ecosystems that differ from non-karst areas in terrain and landform, hydrothermal characteristics, and site-specific conditions that influence vegetation and soil development. Because of their special geological and climatic conditions, karst areas exhibit low environmental capacity, weak anti-interference, low stability, and poor self-regulation (Zheng et al. 2016; Chen et al. 2014; Zhou et al. 2010). In karst regions, the soil environment is coupled with bedrock outcrops and small soil stocks with a scattered and discontinuous distribution, and the micro-topography is complex and diverse, leading to the development of a discontinuous shallow soil of varying thickness. These factors complicate karst soil organic carbon density calculations (Ying et al. 2012). Due to the unique characteristics of the karst region, the method for estimating carbon density in non-karst areas is not suitable for these special areas. Many scholars have studied the karst ecosystem characteristics that influence soil organic carbon density heterogeneity. Currently, some indexes, such as the rate of rock coverage and soil thickness, have been considered in the estimation of SOC storage in karst areas (Koulouri and Giourga 2007; Schlesinger 1982), but other indexes have not. By investigating the spatial variability of related indexes, including soil distribution and acreage, rate of rock coverage, soil thickness, bulk density, and SOC content, in the plateau-type karst of small watersheds and revising the SOC density and storage relationship, a method of calculation based on soil type has been proposed to improve the reliability of SOC storage estimates for these areas. However, the publication of such research is rare.

Soil connects the atmosphere, hydrosphere, biosphere, and lithosphere. Soil carbon is an important part of the terrestrial carbon pool as well as an important basis for soil fertility (Weissert et al. 2016; Akpa et al. 2016; Bohn 1982). Soil carbon consists of organic carbon and inorganic carbon. The content of inorganic carbon is stable, exerting little influence on the carbon cycle (Fox and Bryan 1999; Kinsey-Henderson and Wilkinson 2013). Therefore, soil organic carbon (SOC) is the main factor affecting the carbon balance of terrestrial ecosystems (Pouyat et al. 2002). The transformation of SOC has a direct impact on the global carbon cycle, soil fertility and plant growth, and an indirect impact on terrestrial carbon stores, which makes SOC both an important indicator of soil quality and a comprehensive reflection of terrestrial productivity and environmental health (Niu et al. 2015). Serious human disturbance and fragility in karst ecosystems have caused severe land degradation with heavy losses of SOC that further decrease the sustainability of the ecosystem. Therefore, understanding the changes in the SOC of karst ecosystems not only contributes to an underlying theory for their ecological restoration and reconstruction but also becomes an important element for understanding the carbon cycle (Lal 2002).

The primary methods for calculating SOC include the soil-type method, life zone method, and model method (Pouyat et al. 2002). Due to the density of sampling points and sampling methods, both the geostatistical and life zone methods generate large errors, because they estimate the SOC of an entire study area based on a limited number of sampling points (Juan et al. 2013). However, the soil-type method, which is used more widely, employs a soil attribute database to estimate the SOC storage in a small or medium-sized area by summing all estimated SOC stocks for each soil type and its corresponding acreage (Grüneberg et al. 2013). Because the factors influencing SOC storage are poorly understood, the estimates by different people can be quite different, and it is difficult to determine, whose calculation is more accurate. Therefore, all factors that affect SOC content should be considered in the calculation.

The primary objective of this research was to perform a comprehensive analysis of the factors impacting the soil organic carbon in the Houzhai River Basin. Furthermore, we explored how the accuracy of the estimation methods varied with different soil types, soil depths, and SOC contents. This in-depth study of a karst area and the accurate estimation of its carbon pool provide a reference to inform regional sustainable development and the accurate estimation of the global soil carbon pool.

Materials and methods

Study region

The study region (105°40′43″–105°48′2″E, 26°12′29″–26°17′15″N) includes the towns of Chengguan, Maguan and Baiyan in Puding County of Guizhou Province, and it covers an area of 72 km2. The elevation is between 1223.4 and 1567.4 m above sea level, and the air pressure is between 806.1 and 883.8 hpa. There are three major soil categories: limestone soil, paddy soil, and yellow soil. Vegetation includes cedarwood (Cupressus funebris Endl.), Chinese aspen (Populus adenopoda Maxim), Chinese mahogany [Toona sinensis (A. Juss.) Roem.], and Chinese pear [Pyrus pyrifolia (Burm.) Nakai.]. The main crops are paddy rice (Oryza sativa and Oryza glaberrima), corn (Zea mays Linn. Sp.), soybean [Glycine max (Linn.) Merr], and sunflower (Helianthus annuus).

Soil sampling

The study design followed a grid-based sampling method, and a total of 3180 sampling grids (150 m × 150 m) were established. Sampling sites were defined at the center of each sampling grid (Fig. 1). From March 2013 to January 2015, 2755 soil profiles consisting of 22,057 soil samples were determined according to designed sampling grids. Of the sampling sites, 425 were located, where sampling could not be performed, such as on a roadway or tractor road or in residential housing, industrial parks, or streams. Each profile was divided into 12 soil horizons (0–5, 5–10, 10–15, 15–20, 20–30, 30–40, 40–50, 50–60, 60–70, 70–80, 80–90, and 90–100 cm) if the soil thickness was equal to or greater than 95 cm. Otherwise, sampling was carried out to the actual soil depth. For instance, if a soil profile was 26 cm in depth, 5 soil samples were taken (0–5, 5–10, 10–15, 15–20, and 20–26 cm), and if a soil profile was 33 cm thick, 5 soil samples would also be taken (0–5, 5–10, 10–15, 15–20, and 20–30 cm).

Fig. 1
figure 1

Location of Houzhai River Basin and the distribution of sample sites

Full size image

Local information about each sampling point including the soil bulk density, soil thickness, rock coverage, and other indexes was measured and recorded at the time of sampling. Soil samples were air dried, ground, and prepared as required for laboratory analysis, and the SOC content was then tested and analyzed via the potassium dichromate method. Soil acreage was calculated using GIS technology and field surveys soil bulk density was measured layer by layer from the top to the bottom of the soil profile via the cutting-ring method (Nelson and Sommers 1982; Wang et al. 2010). Soil thickness was recorded in accordance with the type of ecological niche with a metal stick that was 60 or 120 cm long depending on the soil mass at different depths. Bare rock rate was surveyed using the line-transect method. Due to the complex karst landscape, this measurement would have been more accurate but less operable with a longer line transect. Therefore, the length of the transect was set at 10 m, and the grid cells with rock coverage were surveyed with a tape measure.

Soil bulk density (SBD) was determined at the time of sampling (cylindrical core method). For each layer of all soil profiles, 181.58 cm3 of soil was sampled with a cutting ring (r = 3.4 cm, h = 5 cm), and the fresh weight was measured with a portable balance. Approximately 5 g of soil from each layer were collected into an aluminum cup, whose weight had been previously determined. Three milliliters of alcohol (95%) were added to the subsample and ignited (this was repeated three times), and the weight was determined pre- and post-calcination. The SBD was calculated using

$$ {\text{SBD}} = \frac{{(W_{{{\text{cr}} + {\text{s}}}} - W_{\text{cr}} ) \times (W_{\text{post}} - W_{\text{cup}} )}}{{(W_{\text{pre}} - W_{\text{cup}} ) \times 181.58}}, $$
(1)

where SBD is the soil bulk density (g/cm3), \( W_{\text{cr}} \) is the weight of the cutting ring (g), \( W_{{{\text{cr}} + {\text{s}}}} \) is the weight of the cutting ring with fresh soil (g), \( W_{\text{cup}} \) is the weight of the aluminum cup (g), \( W_{\text{pre}} \), and \( W_{\text{post}} \) are the pre- and post-calcination weights of the aluminum cup with soil (g) and 181.58 is the volume of the cutting ring (cm3).

Data analysis

Data from 2755 sampling points surveyed during the early stage of this program were analyzed. The soils could be divided into nine types on the basis of soil type as follows: 457 samples of yellow clay, 613 samples of rendzina, 397 samples of yellow limestone soil, 129 samples of large loam, 439 samples of small clay, 125 samples of white large loam, 106 samples of white sand, 185 samples of large mud field loam, and 304 samples of yellow clayey soil. Data analysis was conducted in Excel 2013 and ArcGIS 10.3 software.

Calculations and statistical analysis

To minimize the estimation error of the SOC stock, the error due to rock coverage in the karst area can be reduced by revising its bare rock rate (Zhang et al. 2017):

$$ {\text{SOCD}}_{i,j} = C_{{{\text{soc}}_{i,j} }} \times \rho_{i,j} \times T_{i,j} \times 10^{ - 2} $$
(2)

where \( {\text{SOCD}}_{i,j} \) is the SOC density in the ith layer of soil type j (kg/m2), \( C_{{{\text{soc}}_{i,j} }} \) is the SOC content in the ith layer of soil type j (g/kg), \( \rho_{i,j} \) is the soil bulk density of the ith layer of soil type j (g/cm3), \( T_{i,j} \) is the soil thickness of the ith layer of soil type j (cm), and 10−2 is the conversion coefficient:

$$ {\text{SOCS}} = \sum\limits_{j = 1}^{m} {\sum\limits_{i = 1}^{n} {{\text{SOCD}}_{i,j} \times S_{j} \times (1 - \delta_{j} ) \times (1 - G_{j} ) \times 10^{3} } } , $$
(3)

where SOCS is the total stock of SOC in the study area (kg C), \( S_{j} \) is the soil acreage of soil type j (km2), 103 is the unit conversion factor, \( \delta_{j} \) is the bare rock rate in the sampling area of soil type j (%), and Gj is the fraction of the gravel volume that is greater than 2 mm in soil type j. The other indexes are the same, as in Eq. (2).

Results

Soil thickness and rock exposure in the study region

As mentioned above, 425 of the sampling sites were designated in places, where sampling could not be performed. One was situated on a railway, 120 on roadways, 10 on tractor roads, 36 in streams, 124 in residential housing, 100 in industrial parks, and 34 in factories.

Detailed information on the soil thickness and rock exposure of the Houzhai River Basin is listed in Tables 1 and 2, and spatial information related to these factors is shown in Fig. 2. First, it is obvious that the thicknesses of soils on mountains are generally lower than those of soils in flat areas or on small hills, and rock exposure in mountainous areas is much greater than that in flat areas. Second, the mean soil thickness and rock exposure values varied significantly with the different soil types.

Table 1 Statistics of soil thickness a of different soil types in study region
Full size table
Table 2 Statistics of rock exposure of different soil types in study region
Full size table
Fig. 2
figure 2

Spatial information of soils thickness and rock exposure in the study region

Full size image

The limestone soil, paddy soil, and yellow soil are interwoven throughout the study area, so the soils in this watershed are highly heterogeneous. The limestone soil areas suffered from severe stony desertification and are scattered with rock exposure. With rock coverage being considered, soil coverage in the area is overestimated. Therefore, the soil acreage value should be revised by considering rock acreage. In addition, the rate of rock coverage is very different in different soil types. The mean rock coverage rate in the rendzina area was approximately 43.34%, which is the highest value, while it was 29.22%, the lowest value, in the large loam of tilled soil. There was little rock exposure in the three major tillage areas as well as the yellow clay, large mud field loam, and yellow clayey soils, so the rate of rock acreage in these areas is very low.

Soil bulk densities and SOC concentrations of different soil types in the study area

Soil bulk density is of great importance in the assessment of SOC contents. As shown in Table 3, the differences in soil bulk densities among the different soil types were large. The average soil bulk density of large mud field loam, which ranged from 1.25 to 1.52 g/cm3, was significantly higher than that of other soil types. Among the different soil layers, the soil bulk density of rendzina gradually increased with soil depth, while that of large loam and large mud field loam first increased and then tended to stabilize. In contrast, the soil bulk density of yellow clay, yellow clayey soil, yellow lime soil, small clay, white sand, and white large loam first increased and then decreased as soil depth increased. The largest soil bulk density values of rendzina and large loam occurred at the bottom of the profile, while the largest values for the remaining soil types were observed in the transition zone between the A and B layers.

Table 3 Profile characteristics of soil bulk density of different soil types (g cm−3)
Full size table

The average SOC concentrations of different soil types are shown in Fig. 3. Vertically, the distribution of the SOC content in the profile was the most obvious characteristic. With increasing soil depth, the organic carbon content gradually decreased. In the 0–50-cm soil horizon, the content of SOC decreased significantly, but at depths greater than 50 cm, SOC content decreased slowly. This finding suggests that 50 cm is the turning point in the graph of the SOC profile. The discrepancies in the average SOC contents of the different soil types mainly existed in the top 50 cm of the soil layer. There were no significant differences among the average SOC contents of the different soil types in the 50–100-cm soil layer. The SOC content of the surface soil (0–10 cm) was the highest for all soil types, and it was ranked as follows: rendzina > yellow lime soil > white sand > large mud field loam > white large loam > small clay > yellow clayey soil > large loam > yellow clay. However, the distribution of the SOC concentration throughout the profile of the different soil types (0–100-cm soil horizon) conforms to a different model.

Fig. 3
figure 3

Profile characteristics of soil organic carbon of different soil types in study area

Full size image

Soil organic carbon density in the Houzhai River Basin

In the present study, the SOC density at the 0–20 cm and 0–100 cm soil thickness horizons were calculated for each designated sampling grid, and this spatial information for the study region is presented in Fig. 4. Obviously, SOC density exhibits high inhomogeneity in response to soil thickness and rock exposure. In the Houzhai River Basin, the SOC density for the 0–20 cm and the 0–100 cm soil horizons ranged from 0.09 to 19.54 kg/m2 and from 0.09 to 44.66 kg/m2 with mean values of 4.28 and 8.70 kg/m2, respectively. Detailed information about the SOC density of different soil types is listed in Table 4. The mean SOC density values of the 0 to 20-cm soil horizon of the different soil types descended in the following sequence: large mud field loam > yellow clayey soil > rendzina > large loam > yellow lime soil > small clay > white sand > white large loam > yellow clay. Conversely, the mean SOC density values of the 0–100-cm soil horizon of the different soil types descended in the following sequence: white large loam > large loam > large mud field loam > small clay > yellow clay > yellow lime soil > yellow clayey soil > white sand > rendzina. The SOC density of the 10-cm-thick soil layer gradually decreased as soil depth increased before gradually stabilizing in the deeper soil horizon. The mean SOC density for the 0–100-cm soil horizon in the Houzhai River Basin was approximately 8.70 kg/m2, but the average density of soil organic carbon in China at 100 cm is 10.53 kg/m2. Due to the large spatial variation in soil thickness, the 10-cm thick karst SOC density and the mean SOC density for China were compared. The average SOC density in the top 10 cm of the soil in the Houzhai River Basin was 2.09 kg/m2, but the average 10-cm SOC density in China is 1.21 kg/m2. The SOC density in the upper 10 cm of the soils in the study region is approximately 1.7 times the mean for China.

Fig. 4
figure 4

Spatial information of SOC density in study region (a for 0–20 cm soil horizon, b for, 0–100 soil horizon) (unit: kg/m2)

Full size image
Table 4 Statistics of SOC density of different Soil types
Full size table

Soil organic carbon storage in the Houzhai River Basin

The total SOC storage in the 0–100-cm soil horizon of the Houzhai River Basin was approximately 5.39 × 108 kg, and the SOC content in each sampling grid is shown in Fig. 5. The SOC for the different grids ranged from 2112 to 1,004,938 kg. The stored SOC in the 0–10-, 0–20-, and 0–30-cm soil horizons of the Houzhai River Basin were 1.48 × 108, 2.65 × 108, and 3.44 × 108 kg, respectively.

Fig. 5
figure 5

Spatial information of SSOC stores in each designed grid in Houzhai River Basin

Full size image

The differences in the stored SOC among soil types were mainly caused by the differences in the SOC density of the different soil types. As shown in Fig. 6, the SOC storage in topsoil (0–20 cm) of the different soil types was ranked as follows: large mud field loam > yellow clayey soil > large loam > white large loam > rendzina > yellow lime soil > small clay > white sand > yellow clay (Fig. 6a). The SOC storage of the whole profile (0–100 cm) was ranked as follows: large mud field loam > white large loam > yellow clayey soil > large loam > rendzina > small clay > yellow clay > yellow lime soil > white sand (Fig. 6b). At the same time, the soil-forming environment and the soil-forming characteristics of the different soil types were different. During the field surveys, most of the large mud field loam was flooded, and the soil layer under the water and wet vegetation was in a state of gas reduction for a long time. This state is conducive to the accumulation of organic matter but not its decomposition. Therefore, the SOC density of the large mud field loam was very high, and the SOC storage was also relatively large. Most of the white sand and yellow clay were distributed in steeply sloping areas, and the corresponding vegetation coverage was very low. Species of small shrubs and bushes were the dominant vegetation, and the decay of plant residues was slow, while the mineralization of organic matter was fast. We believe that this is the main reason why the organic carbon accumulation speed in these soils was very weak.

Fig. 6
figure 6

Percentage of SSOC in different Soil types in Houzhai River Basin (a for 0–20 cm soil horizon, b for 0–100 cm soil horizon)

Full size image

Discussion

The study region is part of a karst landform and is characterized by high topographical diversity and soil nutrient spatial in homogeneity. The factors affecting SOC concentration and storage in this region are extremely complex. Therefore, SOC concentration and density in this region are highly spatially variable.

Comparison of organic carbon content in karst and non-karst areas

The content and dynamic equilibrium of soil organic carbon are vital to the assessment of soil quality and soil health, and they directly affect soil fertility and crop production. Some studies indicate that soil organic carbon has great influence on soil structure formation and stability, water maintenance, bioavailability of plant nutrients, soil buffering properties, and soil biodiversity (Adak et al. 2014). Therefore, it plays an important role in the mitigation and regulation of a series of soil processes related to soil degradation and soil productivity. The results of the present study suggest that the soil organic carbon contents in the top soil layers (0–20 cm) in the Houzhai River Basin range from 1.61 to 119.11 g/kg with a mean value of 25.07 g/kg. In comparison with the previous studies, the mean value is obviously lower than the soil organic carbon contents in the non-karst area (Parrasalcántara et al. 2015), which likely indicates that the soil in the karst area is infertile and thus the weakness of the karst ecological system (Wu et al. 2015).

The total amount of soil and soil organic carbon in the karst area was very low. The soil organic carbon density in the upper soil layers (0–20 cm) of the sloping karst lands was approximately 4.28 kg m−2, which was significantly lower than that in non-karst areas. The unique karst geological environment has a profound influence on soil organic carbon density that results in bare rock, soil erosion, discontinuous soil cover, habitat complexity, and widely exposed bedrock, and a large number of weathering residues accumulate in the rock fractures and funnels. Plant root development with discontinuous shallow soils scatters the distribution of soils. The thickness of the karst soil varies greatly with most of the soil being only 4–9 cm deep. Therefore, the total amount of soil in the karst area is significantly lower than in non-karst areas. In addition, the decline of the fragile karst ecosystem is a complex process in which human disturbance is a strong driving force, vegetation reduction is an important incentive, land productivity degradation is an innate character, and increasing desertification landscapes are a symbol.

Effects of different soil and soil rock types on SOC concentrations

Some research has shown that the soil parent material has a remarkable impact on the spatial distribution of SOC (Fang et al. 2015). In karst regions, soluble rock was the base development material, and the insoluble residue from the corrosion of carbonate rock was the main source of the soil (Fu et al. 2014). The tropical peak cluster and needle karst landforms were largely developed under humid and rainy climate conditions, leading to the wide distribution of limestone soil throughout the peak forest plain or the arc peak plain (Xu et al. 2013). Therefore, the rock types in the study area were varied, and the spatial distribution patterns of the soils were different. In the studied basin, the rock types are mainly dolomite, limestone, marl, sand–shale, and quaternary yellow clay (Gong et al. 2012). The dolomite and limestone were mainly distributed in the peak upper reaches and the peak cluster depression area, and their distribution area was very large.

However, the distribution area of the dolomite and limestone in the karst and butte areas was relatively low. The peak, peak cluster, and regional soil thicknesses were relatively shallow, but the amount of soil in the depression region was profound. The sandstone area was marl that was primarily distributed in the upstream slope land and paddy fields, whose distribution area was small (Fontaine et al. 2007). Other regions also had a small number of areas with sporadically distributed soil that was relatively shallow. The quaternary yellow clay was mainly distributed in the middle and lower reaches of the hillock; their distribution area was large, and the amount of soil was profound. Based on the results of the analysis, the impact of soil type and lithology on the spatial variation in SOC was remarkable. The soil types reflected differences of soil-forming conditions and soil-forming processes. Because the soil-forming conditions and soil-forming processes of the three types of soil in the study area (lime soil, paddy soil, and yellow soil) were obviously different, significant differences in soil organic carbon storage resulted (Table 5).

Table 5 Lithologic distribution characteristics of the Houzhai River Basin
Full size table

Effect of slope gradient and slope position on the SOC stock of the study region

Based on geologic records, the soil slope position was divided into the bottom, the lower part, the upper middle part, the upper part, and the top part of the mountains in the Houzhai River Basin. As shown in Table 6, rocks were exposed along the slope positions, affecting the soil depth, the SOC contents, and the SOC density. The soil depth was closely related to the slope position, and declined from 61.81 to 35.73 cm. The rate of exposed rock gradually increased from 16 to 27% and then stabilized. The density of SOC (0–10 cm) increased from 1.96 to 2.16 kg/m2 and then declined to 2.11 kg/m2. Because of the influence of soil depth, rock contents and SOC from the bottom of the slope to the top part of the mountains. The most of central part, the upper part, and the top part of the mountains were woodland and shrub land that were rich in organic matter and characterized by vegetation in good condition. In addition, the higher terrain and reduced human influence resulted in a higher SOC content. In contrast, most of the bottom part of the slope and the lower mountains were agricultural land, which was poor in vegetation and soil input. The SOC content was low in this area. In general, the sunny aspect was dry, so the SOC decomposed faster, and its content was lower. The content of SOC was higher on the shady aspect; thus, the content and density of SOC were higher than on the sunny aspect.

Table 6 Effect of slope position on rock exposure, soil thickness, SBD, SOC concentration, and SOC density
Full size table

Effects of other environmental factors on the SOC stock of the study region

The correlation analyses between the SOC density and the environmental factors in the Houzhai River Basin is shown in Table 7. The density of SOC exhibited a remarkable correlation with elevation, soil depth, gravel content, slope gradient, and rate of exposed rock (p < 0.01). However, multiple environmental factors were correlated with each other, so the contribution of each environmental factor on density of SOC could not be determined. Therefore, a principal component analysis (PCA) was needed to extract the major factors influencing the density of SOC. Slope position, elevation, slope gradient, and slope aspect were regarded as the main indicators of the first principal component, which consisted of the main topographic parameters driving SOC in the Houzhai River Basin. The soil bulk density, soil type, rock content, and soil depth were regarded as the main indicators of the second principal component, which was treated as a factor that correlated with soil formation. The vegetation was regarded as the main indicator of the third principal component, which was considered the vegetation factor. To reveal the contribution of each environmental factor to SOC density, a stepwise regression analysis (SR) was used to select the principal component index, which included the gravel content, elevation, slope position, soil depth, slope gradient and the rate of exposed rock, and the following multiple linear regression equation describing the relationship between SOC density (dependent variable) and each environmental factor (independent variables) was established:

$$ {\text{SOCD}} = { 1}. 8 6D + { 8}. 2 1L + \, 0. 1 5A + { 3}. 6 5 4B \, {-} 10 8 9. 9 4 6 { }(R^{ 2} = \, 0. 5 4 , { }p\; < \;0.00 1), $$

where SOCD is the SOC density (kg m−2), D is the soil depth, L is the rate of exposed rock (%), A is the elevation (m), and B is the gravel content (%). Soil depth was most closely correlated with SOC density (β = 0.67) followed by the rate of exposed rock (β = 0.23), the gravel content (β = 0.15) and elevation (β = 0.11). Overall, SOC density was comprehensively affected by different environmental factors in the Houzhai River Basin.

Table 7 Correlation between the density of SOC and different environmental factors
Full size table

Conclusion

Using a grid-based sampling method, 2755 soil profiles consisting of 22,057 soil samples were sampled and analyzed. We believe that the SOC density in the 0–20- and 0–100-cm soil horizons in the Houzhai River Basin in the Guizhou Province of southwest China ranged from 0.09 to 19.54 kg C m−2 and 0.09 to 44.66 kg/m2 with mean values of 4.28 and 8.70 kg/m2, respectively. The SOC stored in the 0–20 and 0–100 cm soil horizons of the Houzhai River Basin was 2.65 × 108 and 5.39 × 108 kg, respectively.

In the Houzhai River Basin, soil type is a critical factor that is closely associated with soil bulk density and SOC concentration. In addition, the correlation analyses of SOC density and different environmental factors in the Houzhai River Basin were inconclusive. Therefore, principal component analysis (PCA) was needed to extract the primary factors influencing the density of SOC. The slope position, elevation, slope gradient and slope aspect, which were main factors of SOC in the Houzhai River Basin, were determined to be the main indicators of the first principal component and were treated as topographic parameters. The soil bulk density, soil type, gravel content, and soil depth were regarded as the main indicators of the second principal component and were treated as soil formation correlation factors. Human activities and geographic characteristics are believed to be the primary factors that lead to high in homogeneity of SOC in mountainous karst areas.