Distribution of soil nutrients and erodibility factor under different soil types in an erosion region of Southeast China

Study area

The study area is located in the red soil region of Southeast China, in which the Jiulongjiang River basin is one of the typical erosion regions. Thus, the Jiulongjiang River basin was selected as the sampling region. The Jiulongjiang River basin (24°18′−25°88′N, 116°78′−118°03′E) is located in Fujian Province, with a drainage area of 14,741 km² (Fig. 1). The average elevation is less than 200 m above sea level. The terrain of the basin decreases from north to south, transits from mountain to plain. The basin is controlled by the subtropical and tropical monsoon climate, annual precipitation is about 1,400∼1,800 mm and annual temperature is about 19.9∼21.1 °C (Zhang et al., 2012). The original forest accounts for over 60% of the basin area, other lands are mainly used for agriculture and residential area. This area distributes yellow soils, red soils, lateritic red soils, purple soils, and paddy soils (Fig. 1). Over 90% of the basin area is covered by red soils, lateritic red soils, and yellow soils. Moreover, most soils are developed from granites and clastic rocks (Quan et al., 2007). These soils can be classified into the Alfisols based on the soil taxonomy of the United States Department of Agriculture (USDA) (Soil Survey Staff, 2014).

Soil sampling

Soil samples were collected in the winter season of January 2018. The three soil sites were selected according to different soil types, including red soils, lateritic red soils, and yellow soils, respectively (Fig. 1). The soil profiles were located at the excavation ground, furthermore, the sampling work was obtained support and help from the person in charge of the construction company. Considering the strong spatial heterogeneity of soils, especially in the vertical direction, the averaged results from the parallel soil profiles (if the distance between them is extremely far) will lose, even mislead the information that individual soil profile carries. Thus, the distance between the three parallel soil profiles at the same site is less than one m. A total of 9 soil profiles with a thickness of three m were used to collect soil samples. Soil samples were orderly collected from the bottom to the top to avoid pollution between samples, with a five cm-interval. The three samples at the same depth derived from the three parallel soil profiles were mixed to be one sample. In total, there are 180 samples. The replication of sampling sites and soil samples were considered together in the present study. The thickness and visible characteristics of the three soil profiles are shown in Table 1.

Sample analysis

Soil samples were air-dried after removing gravel and fresh roots, and then were preserved after passing through a two mm sifter. Soil pH values were determined by a pH meter (Liu, Han & Li, 2021a). Soil particle distributions were determined on a laser particle size analyzer (Liu, Han & Zhang, 2020). Soil samples (<2 mm) were ground into powder (<75 µm), using for further analysis of the concentrations of soil major elements. The soil samples were digested with three mL HF, one mL HClO₄, and three mL HNO₃ at 120 °C for 3 days (Li et al., 2020). The concentrations of soil major elements were expressed as the concentrations of their oxides (i.e., Al₂O₃, CaO, Fe₂O₃, K₂O, Na₂O, MgO, and TiO₂, %) in the soils (Liu, Han & Li, 2021a). The SiO₂ concentrations were analyzed according to Du et al. (2014).

Carbonates in soil samples (<75 µm) were removed by treating with 0.5 mol L⁻¹ HCl for 24 h (Midwood & Boutton, 1998), and inorganic N were removed by treating with 2 mol L⁻¹ KCl for 24 h (Meng, Ding & Cai, 2005). The treated samples were washed with purified water until neutrality, then were dried at 55 °C until constant weight and ground into powder. The SOC and SON concentrations were measured by a multi-element analyzer. Actual SON and SOC concentrations in the original soils can be calibrated by multiplying of measured value by the ratio of the sample mass after treating to it before treating.

Calculation of the K factor on the EPIC model

Soil erodibility can be used to indicate the erosion risk of surface soils because of the significant quantitative relationship between the amount of soil loss and the K factor (Wischmeier & Smith, 1965). Many modified models have been proposed based on correlations between soil physicochemical properties and the K values (Wang et al., 2013). For example, the K factor of the Erosion Productivity Impact Calculator (EPIC) model (Sharpley & Williams, 1990) is calculated using the parameters, including SOC concentration and different-sized particle distribution. Soil structure affects soil erodibility by restricting the force of slope flow. Many researchers reported that water-stable macro-aggregates could decrease soil erodibility due to the effects on soil structure (Ding & Zhang, 2016; Liu & Han, 2020). However, the formation and stabilization of water-stable aggregates are closely associated with the rich soil organic matter (SOM) (Six & Paustian, 2014). Water-stable macro-aggregates were hardly formed in the SOM-poor red soils of Southeast China, based on field observation. Therefore, SOC concentration and different-sized particle distribution are the main factors that affect soil erodibility. Thus, the EPIC model is fit to estimate soil erodibility K factor in the study area, and the formula is shown as follow: ${\displaystyle \text{Kepic}=\left\{0.2+0.3exp\left[-0.0256\mathrm{Sa}\left(1-\frac{\mathrm{Si}}{100}\right)\right]\right\}\times {\left(\frac{\mathrm{Si}}{\mathrm{Si}+\mathrm{Cl}}\right)}^{0.3}\times \left[1.0-\frac{0.25\mathrm{C}}{\mathrm{C}+exp\left(3.72-2.95\mathrm{C}\right)}\right]\times \left[1.0-\frac{0.7\mathrm{SN}1}{\mathrm{SN}1+exp\left(-5.51+22.9\mathrm{SN}1\right)}\right]}$ where the Sa, Si, and Cl (%) are the proportions of sand-sized, silt-sized, and clay-sized particles, respectively; the SN1 (%) = 1–Sa/100; the C (%) is SOC concentration. The unit of the K factor is t acre h/100 acre/ft/tanf/in. For brevity, the unit of the K factor will not be noted.

Statistical analysis

Boxplot was used to show the ranges of the proportions of different-sized particles, soil pH values, the concentrations of SOC, SON, and soil major elements (expressed as Al₂O₃, CaO, Fe₂O₃, K₂O, Na₂O, MgO, TiO₂, and SiO₂), and the K factor values in the red soils, lateritic red soils, and yellow soils. One-way ANOVA with the least significant difference (LSD) test was performed to determine the differences in the proportions of different-sized particles, soil pH values, the concentrations of SOC, SON, and soil major elements, and the K factor values among the three soil types at the level of P < 0.05. All data were calculated via the Kolmogorov–Smirnov test, which is one of the non-parametric tests commonly applied to analyze the normal distribution of the sample data set (Zeng & Han, 2020). Spearman’s rank correlation coefficient determined the relationship between different soil properties. All statistical analyses were performed by the SPSS 18.0 software (SPSS Inc., Chicago, IL, USA) and all graphs were drawn by SigmaPlot 12.5 software (Systat Software GmbH, Erkrath, Germany).

Soil particle distribution and soil pH

The proportions of different-sized particles and soil pH in the three profiles are shown in Figs. 2 and 3, respectively. The clay proportions in the YS and RS profiles were almost constant with increasing soil depth, while silt proportions increased in the YS profile and decreased in the RS profile, and sand proportions decreased in the YS profile and increased in the RS profile (Fig. 2). The vertical distributions of the different-sized particles in the LRS profile showed intensive fluctuation, which was attributed to the non-uniform distribution of the large-sized unweathered quartz particles. On the whole, the clay proportions in the yellow soils (mean: 13%) were significantly lower than those in the lateritic red soils (mean: 15%) and were significantly larger than those in the red soils (mean: 11%) (Fig. 3). The silt proportions in the yellow soils (mean: 68%) were significantly larger than those of the lateritic red soils (mean: 64%) and red soils (mean: 64%). While the sand proportions in the yellow soils (mean: 19%) were significantly lower than those in the red soils (mean: 25%) and were similar to those in the lateritic red soils (mean: 20%). The texture of most soils in the three profiles belongs to silt loams (Soil Survey Staff, 2014).

The soil pH values in the YS profile decreased from 4.3 to 3.8 with increasing soil depth at the 0∼50 cm depth, then kept constant at the 50∼160 cm depth, subsequently increased to 4.7 at the bottom (Fig. 2). In the RS profile, the soil pH values increased from 4.2 to 4.6 with increasing soil depth. In the LRS profile, the soil pH values decreased from 4.9 to 4.3 at the 0∼90 cm depth, then increased to 5.1 at the 120 cm depth, and were constant (4.7) in the deeper soils. On the whole, the soil pH values in red soils (mean: 4.3) were significantly lower than those in the lateritic red soils (mean: 4.7) and were significantly larger than those in the yellow soils (mean: 4.1) (Fig. 3). All soils in the three profiles were strong acid.

SOC and SON concentration and C/N ratio

The concentrations of SOC and SON in the three profiles showed decreasing trends with increasing soil depth at the 0∼80 cm depth, while they were almost constant in the soils below the 80 cm depth (Fig. 4). In the soils at the 0∼80 cm depth, the SOC and SON concentrations in the YS profile decreased from 8 g/kg to 2 g/kg and from 0.9 g/kg to 0.4 g/kg with increasing soil depth, respectively. Similarly, the SOC concentrations in the LRS profile decreased from 11 g/kg to 2 g/kg, and the SON concentrations decreased from 0.8 g/kg to 0.2 g/kg. However, the SOC and SON concentrations in the RS profile showed abnormal increasing trends at the depth of 20∼70 cm, which was closely associated with the abundant plant fine root (Table 1). In the soils below the 80 cm depth, the SOC concentrations in the three profiles were not significantly different, while the SON concentrations in the YS profile were significantly lower than those in the LRS and RS profiles.

The C/N ratios in the YS and LRS profiles slowly decreased from 8.6 to 4.0 and from 13.4 to 11.6 with increasing soil depth, respectively. However, the C/N ratios in the RS profile intensively decreased at the 80 cm depth. For the soils below the 80 cm depth, the C/N ratios in the YS profile were significantly lower than those in the LRS and RS profiles.

Soil major elements

The vertical distributions of soil major elements, including Al₂O₃, MgO, TiO₂, Fe₂O₃, K₂O, Na₂O, CaO, and SiO₂ in the three soil profiles are shown in Fig. 5. In the YS profiles, the concentrations of Al₂O₃, MgO, and Fe₂O₃ in the surface soils were significantly lower than those in the deep soils. The K₂O and CaO concentrations slightly increased with increasing soil depth, while the concentrations of TiO₂, Na₂O, and SiO₂ were almost constant, except for the abnormally high SiO₂ concentrations that occurred in the layer of the 5∼35 cm depth. The distributions of these soil major elements were fluctuant in the LRS and RS profiles. Generally, the MgO and K₂O concentrations in the RS profile slightly increased with increasing soil depth, while Fe₂O₃ and CaO concentrations slightly decreased. In the LRS profile, the Al₂O₃, MgO, Fe₂O₃, and K₂O concentrations increased and SiO₂ decreased with increasing soil depth. Particularly, the Na₂O and CaO concentrations at the 70∼90 cm and 200∼295 cm depth occurred abnormal peaks. On the whole, the Al₂O₃ concentrations in the YS profile were significantly lower than those in the RS and LRS profiles, while the MgO, TiO₂, Fe₂O₃, K₂O, Na₂O, and CaO were significantly larger (Fig. 6). The MgO and TiO₂ concentrations in the LRS profile were significantly lower than those in the RS profile, and the Na₂O were significantly lower, while the Al₂O₃, Fe₂O₃, K₂O, CaO concentrations in the two profiles were not significantly different. The SiO₂ concentrations in the three soils were similar with a mean proportion of 68%. Compared to the background values of soil major elements in Fujian Province (Chen et al., 1992), except for TiO₂ and CaO in the YS profile, the major elements in the three profiles were almost lower than these background values.

Soil erodibility K factor

The K factors in the soils at the 0∼30 cm depth of the three profiles ranged from 0.348 to 0.414 (Fig. 7). The K factors in the red soils (mean: 0.393) were significantly lower than those in the yellow soils (mean: 0.407) and significantly higher than those in the lateritic red soils (mean: 0.369).

Soil erodibility K factor in different soils

Soil erodibility K factor on the EPIC model is affected by soil physicochemical properties, including SOC concentration and the proportions of different-sized particles (Sharpley & Williams, 1990). Soil structure significantly affects the erosive force of runoff (Liu & Han, 2020). In the study area, SOM abundance and the distribution of different-sized particles mainly determined soil structure. SOM plays an important role in affecting soil erodibility, due to the ability to absorb water and enhance soil infiltration (Zhu et al., 2010). The distribution of soil particles determines soil porosity, which is also key for soil infiltration. Although water-stable aggregates have been widely reported that can effectively enhance soil structure (Ding & Zhang, 2016; Liu & Han, 2020). However, the low SOC concentration in the yellow soils, lateritic red soils, and red soils restricted the formation of water-stable aggregates. Thus, the K factor estimated on the EPIC model can be sufficiently used to predict erosion risk in the red soil region. The K factor values (0.348∼0.414, Fig. 7) in this study were accordant with the range of 0.15∼0.45 in the surface soils (at the 0∼30 cm depth) of Southern China (Zhang, DeAngelis & Zhuang, 2011b). Undoubtedly, the exact values were determined by the SOC concentration and the proportions of different-sized particles in soils. The K factors in the yellow soils were significantly higher than those in the red soils, and those in the red soils significantly higher than those in the lateritic red soils (Fig. 7). This result showed that soil erodibility significantly decreased with the progress of soil evolution. Under the effects of weathering and leaching processes, the progress of soil evolution generally follows as yellow soils, red soils, and lateritic red soils in Southeast China (Quan et al., 2007). In the soil evolution process, soil texture trends more and more clayey (Quan et al., 2007). Soil crust is easily formed in the clayed soils to reduce the loss of soil particles (Wang et al., 2018). Thus, the soil erodibility K factors of the three soils were significantly different.

Soil organic carbon and nitrogen in different soils

The organic carbon level in soils depends on the balance between the plant input and output due to microbial decomposition and soil erosion (Han et al., 2020; Jobbagy & Jackson, 2000; Yu et al., 2020). Deciduous forests are widely distributed in the red soil hilly area of southern China, which provides sufficient organic matter input. Generally, the low SOC concentration in the red soil region is mainly attributed to the rapid SOC turnover (Chen, He & Huang, 2002). Firstly, the warm humid monsoon climate in this region is conducive to microbial activities, resulting in the rapid decomposition of organic matter (Goh, 2011; Xia et al., 2021). Secondly, organic matters are lost as dissolved humic acid under soil erosion. The accumulation of humic acid generally leads to the decreasing soil pH, thus it is commonly observed that soil pH increases with increasing soil depth. However, the decreasing trend of soil pH with increasing soil depth in the YS and LRS profiles at the 0∼80 cm depth (Fig. 2), which mainly is attributed to the loss of dissolved humic acid. Thirdly, lots of Ca²⁺ and Mg²⁺ are lost in the processes of red soil evolution and leaching, resulting in reduced SOC stabilization. Soil Ca²⁺ and Mg²⁺ play important roles in concatenating organic matters with clay minerals to form stable organic–inorganic complexes (Li et al., 2017). The shortage of Ca²⁺ and Mg²⁺ can cause the acceleration of SOC decomposition due to the decreased stability. Additionally, it has been widely reported that the loss of particle organic matter causes decreased SOC concentration in surface soils (Ge et al., 2007; Quinton et al., 2010; Zhang et al., 2018). Thus, the soil erosion and soil physicochemical properties (depended on soil evolution) affect SOC distribution.

The SON concentrations in the yellow soils below the 80 cm depth were significantly larger (twice) than those in the red soils and lateritic red soils (Fig. 4). The large SON concentrations in the deep layer of the yellow soils were mainly attributed to the re-assimilation of inorganic N through microbial immobilization. This conclusion can be inferred by the results of soil C/N ratios and δ¹⁵N values (Table S1). The C/N ratio of plant-derived organic matter generally decreases in the processes of microbial decomposition (Marshall, Brooks & Lajtha, 2007), but it is not lower than 5 of the mean C/N ratio of microbial biomass (Zhang et al., 2000). The C/N ratios in the yellow soils below the 80 cm depth were almost near 5, thus it can be confirmed that the most SON is derived microbial biomass or necromass. The mean δ¹⁵N value of SON in the deep soils was 3‰, which was ¹⁵N-depleted compared to the decomposed organic matter (5‰) in the upper layer. It is can be inferred that the re-assimilation of ¹⁵N-depleted inorganic N through microbial immobilization into microbial biomass (Boutton & Liao, 2010; Corre et al., 2007), resulting in the ¹⁵N-depleted SON in the deep soils. In agricultural ecosystems, excess inorganic N generally is assimilated by soil microbes in the deep soils (Baggs et al., 2003; Liu, Han & Li, 2021b).

Soil major elements in different soils

Compared to the background values of soil major elements in Fujian Province (Chen et al., 1992), the major elements in the three profiles were almost lower than the background values (Fig. 5). The dissolved metal ion is easily lost under the strong erosion environment. However, the concentrations of many major elements in the yellow soils with a high erodibility were significantly larger than those in the red soils and lateritic red soils (Fig. 6). The result indicated that soil evolution significantly affected the distribution of soil major elements under different soil types. In the region of Southeast China, soil Na⁺, K⁺, Ca²⁺, and Mg²⁺ are preferentially dissolved and lost under strong leaching conditions, while the iron and aluminum, these have a low mobilization, are gradually accumulated with soil evolution (Belton & Goh, 1992; Treder, 2005). Furthermore, soil pH generally decreases with increasing soil depth because H⁺ replaces the position of the positive charge. Thus, theoretically, soil pH positively correlates with K₂O, Na₂O, CaO, and MgO concentrations, while negatively correlates with Al₂O₃, TiO₂, and Fe₂O₃ concentrations. In the yellow soils, similar correlations between soil pH and the concentrations of major metal elements were observed below 80 cm depth (Table 2), indicating that the distribution of these elements in the deep soils is controlled by soil evolution. However, soil pH negatively correlated with the concentrations of all major metal elements and positively correlated with SiO₂ concentration at 0∼80 cm depth (Table 2). This result indicated that the distribution of the major elements was also controlled by other factors. Silt proportion positively correlated with the concentrations of all major metal elements and negatively correlated with SiO₂ concentration; whereas sand proportion negatively correlated with the concentrations of all major metal elements and positively correlated with SiO₂ concentration (Table 2). These results indicated that the distribution of soil major elements in the surface soils was closely associated with the distribution of soil particles. Soil erosion preferentially removes fine particles (Ostovari et al., 2018). Thus, the variation of soil major elements in the surface layer of the yellow soils was affected by both soil erosion and soil evolution. The concentrations of most major elements did not significantly correlate with soil pH or the proportion of different-sized particles in the red soils and lateritic red soils, likely associated with intricate factors, for examples, extremely low concentration of some elements, strong leaching, and differences in soil parent materials. More studies are needed to explore the reasons that cause the vertical variations of soil major elements in different soil types.