1 Introduction

Soil organic carbon (SOC) pool is an important part of terrestrial carbon pool. Small changes in SOC pool may considerably destroy the global carbon balance [1] and decrease soil quality because SOC promotes soil aggregation, forms soil structure, and affects soil properties [2]. Water erosion widely and considerably affects SOC pool worldwide because of sediment loss that usually accompanies SOC transport [3]. Furthermore, given that raindrops and runoff preferentially detach and transport light particles, SOC can be enriched in sediments because of light density and easy bonding with clay-size mineral particles of SOC. Hence, further investigation of SOC loss under water erosion will help estimate SOC changes. However, soil erosion and SOC loss regulations in soils suffering from severe water erosion (e.g., red soil, black soil, and loess) vary. Thus, understanding the SOC loss for more than two serious eroded soil types will provide necessary knowledge for the SOC stock changes and soil quality degradation.

In Northern China, black soil suffers from severe soil erosion [4]. Freezing and thawing decrease the rainfall stability of soils given the cold and wet climate characteristics [5], thereby aggravating water erosion. The high annual precipitation and frequent rainfall provide sufficient water for the red soil in the hilly region of South China to be transported. The unique soil properties of loess on the Loess Plateau considerably contribute to the severe regional soil erosion. For example, factors, such as porosity, carbonate content, metal oxides, hydroxides, and soluble salts, affect loess collapsibility [6]. The erosive rainfall concentrated in summer promotes the abnormally significant soil erosive intensity. Moreover, the decreasing vegetation coverage is another primary reason for the soil erosion in China [7]. Slope is the basic unit of soil erosion. Many researchers have focused on the soil erosion on slopes for the three seriously eroded soils. Among the soil erosion factors, soil loss and runoff are positively correlated with rainfall intensity [8,9,10,11]. Soil loss is positively correlated or initially increases and then decreases with the increase in slope gradient [8, 12,13,14]. In addition, soil properties, such as soil bulk density and clay mineralogy, affect the slope soil erosion process [9, 10, 15]. However, studies concerning the difference of SOC loss among the three soils are few.

The sediment-bound organic carbon (OC) loss contributes remarkably to the total SOC loss under water erosion [16, 17]. Thus, many researchers have focused on SOC loss accompanying sediments, which is positively correlated with soil loss [15, 18,19,20]. However, light particles within SOC are usually preferentially transported [21]. SOC is enriched in eroded sediment, and the enrichment ratio of SOC (ERoc) changes with the relationship between soil loss and SOC loss. The transport of clay or fine particles is one of the causes of ERoc; other mechanisms include particulate OCs and aggregate fragments [16, 19, 22, 23]. For red soil, black soil, and loess with different soil textures and OC concentrations, SOC transportation patterns vary, affecting ERoc. However, the comparison of the ERoc of the three soils has not been studied. ERoc value is usually negatively correlated with rainfall intensity, rainfall duration, slope sediment loss, or runoff volume [23, 24], but detailed effects of factors on SOC enrichment of the three soils are different, and also require further investigation.

Previous studies mainly focused on the SOC loss pattern in a particular soil type. However, few studies focused on comparing the SOC loss patterns of black soil, red soil, and loess. Therefore, the objectives of this study are as follows: (1) to compare the soil loss of black soil, red soil, and loess; (2) to investigate the differences in SOC transport and enrichment in sediments for the three soils; (3) to clarify the differences in the SOC loss associated with soil loss mechanisms among the three soils. For these purposes, the study is structured as follows: the research methods used are shown in Sect. 2. The results of the study are described in Sect. 3, starting with a comparison of runoff and sediment loss for red soil, loess, and black soil in Sect. 3.1. Subsequently, the results of the SOC loss comparison are reported in Sect. 3.2, and the discussion is presented in Sect. 4. Finally, the conclusions are summarized in Sect. 5. This study can provide important theoretical bases for SOC evaluation and related modeling for soil under water erosion on a large scale.

2 Materials and methods

2.1 Study areas and soil properties

Black soil was collected from the plow layer (20 cm) of arable fields of the typical black soil region, i.e., Jilin (44° 43′ N, 126° 12′ E) and Heilongjiang (45° 43′ N, 126° 36′ E) Provinces in Northeast China (Fig. 1). The annual average temperature in the black soil regions is 3.7 °C. The mean annual precipitation in the black soil regions is 545 mm, and the rainfall from July to September accounts for 70% of the annual average precipitation. The soil textures of the black soil are silty clay loam or silty loam. The SOC contents of the black soils in Jilin and Heilongjiang Provinces are 13.61 and 14.21 g kg−1, respectively (Table 1). The slope gradient in the black soil area is generally 1°–8° (mostly less than 10°). The red soil was collected from the hilly red soil area, i.e., Hunan (27° 03′ N, 111° 22′ E; 28° 30′ N, 112° 54′ E) and Guangdong (24° 05′ N, 115° 37′ E) Provinces in the southeast of China (Fig. 1). The sites of the red soil are located in a subtropical humid monsoon climate zone. The annual average temperature in the red soil region is 19.0 °C. The annual average precipitation in the red soil region is 1743.5 mm. The soil from Hunan Province was developed from quaternary red soil. The red soil from Guangdong Province is located in the hilly weathered granite red soil area. The soil textures of the red soil are silty clay, loamy clay, and clay. The loess was collected from Shaanxi Province, which is located in the Loess Plateau. Four typical loess soils from Yangling (34° 16′ N, 108° 04′ E), Changwu (35° 12′ N, 107° 47′ E), Ansai (36° 58′ N, 109° 20′ E), and Suide (37° 31′ N, 110° 16′ E) were selected (Fig. 1). Yangling and Changwu have a subhumid continental monsoon climate, whereas Ansai and Suide have a semiarid continental monsoon climate. The mean annual temperature in the loess region is 9.5 °C. The mean annual rainfall in the loess region is 565 mm, 60% of which falls between June and September. The soil textures of the loess are silty clay loam and silty loam. The clay contents (< 0.002 mm) of the four typical loess soils are 24.90%, 20.63%, 15.60%, and 12.10%, showing the characteristics of the zonal distribution. The other properties of the three soils are shown in Table 1. All soils were collected from a depth of 0–20 cm and from the farmland. Outdoor experiments were conducted by establishing runoff plots directly on cultivated land.

Fig. 1
figure 1

Locations of the sampling sites

Full size image
Table 1 Selected properties of the original soils of black soil, red soil and loess
Full size table

2.2 Experimental design

According to the Chinese Soil and Water Conservation Act [25] and the classification of farmland slopes on the Loess Plateau [26], the 25° slope is the maximum slope for cultivated land and the highest floor level of slope. Thus, the selected slopes for loess soils were 5°, 10°, and 15°. The selected slopes for black soil were 5° and 10° given that the natural black soil area has a gentle slope between 3° and 10° [27]. The selected slopes for red soil and loess were 5°, 10°, and 15°. Most erosive rainfall in northeast China is short and heavy that lasts for less than 1 h [28]. Moreover, in accordance with the standard of erosive rainfall for agricultural land in northeast black soil region [29] and the rainfall intensity under which SOC easily enriched, rainfall intensities of 45 and 90 mm h−1 were used in experiments on black soil. Rainfall intensities of 45 and 90 mm h−1, representing the low- and high- intensity storms of this region, respectively, were also used in experiments on red soil, considering the change in rainfall intensity in the red soil area. Rainfall intensities of 90 and 120 mm h−1, which represented typical rainfall intensities of strong storms in subhumid climatic regions of China [25, 30], were used in the experiments for loess.

For laboratory experiments, plant residuals and gravels were removed, and all soil samples were passed through a 10 mm sieve and mixed thoroughly prior to air drying to the desired moisture content. Prior to packing the soil, a 5 or 10 cm thick layer of coarse sand was added to the bottom of the soil pan to maintain permeable conditions. Then, the experimental soil layer was placed over the coarse sand layer in 5 cm increments. Each soil layer was raked lightly before the next layer was packed; a total of 30 cm thick soil was packed. The length and width of experimental soil pans were 8 × 1.5 m, 5 × 2 m, and 5 × 1.5 m for black soil, red soil, and loess, respectively. The cultivated soil slopes were bare and without vegetation. For field experiments, the frame of each plot was surrounded with hard impermeable material, e.g., metal or PVC, driven into the ground to shield them from runoff from adjacent plot areas to ensure minimal soil disturbance. A V-shaped runoff funnel was positioned at the end of each plot to collect runoff samples. The nozzles of the sprinkler rainfall simulator system were placed 18 and 4 m above the ground for laboratory and field experiments, respectively, to ensure that the raindrops created in the experiments were similar to natural raindrops. For all experimental treatments, pre-rain with low rainfall intensity was performed 12 or 24 h prior to each experiment. The surface soil moisture contents for black soil, red soil, and loess were 21.7–29.7%, 8–39%, and approximately 10%, respectively. Prior to each simulated event, simulated rainfall devices were tested to ensure the rainfall intensity and the evenness of the raindrop spatial distribution (uniformity > 80%). After runoff initiation, the samples of runoff and sediment were collected between 3 min intervals. After each experiment, all runoff suspensions in buckets were immediately weighed to obtain the total weights of runoff and sediment. Sediment was dried in a forced-air oven until a constant mass was achieved and weighed to determine the sediment loss parameters. All SOC contents in the original soil were determined using the dichromate oxidation method. All experimental treatments were repeated twice or three.

2.3 Data processing

Data in all studies on the SOC loss process for loess, black soil, and red soil were extracted by searching key words “water erosion” and “SOC transport” among others. Original data were used to generate the regressed process curve for data reproduction and to obtain the laws of data effectively. For the soil and SOC loss process data, a 3 min interval was used to retrieve data from the regression curve. The average retrieved data were used for analysis. Rainfall process data were calculated as runoff rate (Tr, L m−2 h−1), soil loss rate (Ts, kg m−2 h−1), SOC loss rate (TSOC, g m−2 h−1), and ERoc (TERoc) to avoid the effect of different sizes of soil pan. The average rainfall data were calculated as the average total runoff (Sr, L m−2), total soil loss (Ss, kg m−2), sediment concentration (Ssc, g L−1 m−2), SOC loss rate (SSOC, g m−2 h−1), and ERoc (SERoc). Sediment concentration is the dry sediment mass per unit runoff volume per unit area. The process sediment or SOC loss data were determined as the sediment or SOC loss rate per unit area, respectively. All total sediment or SOC loss indices were calculated as the total sediment or SOC loss per unit area, respectively. Important indices were calculated as follows:

$$T_{r} = \frac{R}{S \times IT} \times 60,$$
(1)
$$T_{s} = \frac{SL}{{S \times IT}} \times 60,$$
(2)
$$T_{SOC} = \frac{{C_{SOC} \times SL}}{S}/IT \times 60,$$
(3)
$$T_{ERoc} = \frac{{C_{SOC} }}{{C_{SOC0} }},$$
(4)

where R is the runoff over time intervals (L), S is the plot area (m2), IT is the interval time (min), SL is the soil loss over time intervals (kg), CSOC is the SOC concentration in sediment over time intervals (g kg−1), and CSOC0 is the SOC concentration in the original soil (g kg−1).

The change in sediment concentration (∆SCα) when the rainfall intensity increases by 10 mm h−1 or when the slope gradient increases by 1° (g L−1 m−2 h−1), the change in SOC loss rate (∆SOCα) when the rainfall intensity increases by 10 mm h−1 or the slope gradient increases by 1° (g m−2 h−1), and the change in ERoc (∆ERocα) when the rainfall intensity increases by 50 mm h−1 or the slope gradient increases by 1° were used to normalize the effects of rainfall intensity and slope gradient on the average rainfall data, including sediment concentration, SOC loss rate, and ERoc for black soil, red soil, and loess. These indices could be calculated as follows:

$$\Delta SC_{\alpha } = \sum \left( {\frac{{SC_{max} - SC_{min} }}{{\alpha_{max} - \alpha_{min} }} \times \beta } \right)/n,$$
(5)
$$\Delta SOC_{\alpha } = \sum \left( {\frac{{SOC_{max} - SOC_{min} }}{{\alpha_{max} - \alpha_{min} }} \times \beta } \right)/n,$$
(6)
$$\Delta ERoc_{\alpha } = \sum \left( {\frac{{ERoc_{max} - ERoc_{min} }}{{\alpha_{max} - \alpha_{min} }} \times \gamma } \right)/n,$$
(7)

where α is the rainfall intensity (∆SCRI, ∆SOCRI, and ∆ERocRI) or slope gradient (∆SCSG, ∆SOCSG, and ∆ERocSG), SCmax/SCmin is the sediment concentration under the highest/lowest rainfall intensity or the maximum/minimum slope gradient (g L−1 m−2 h−1), αmax/αmin is the highest/lowest rainfall intensity (mm h−1) or the maximum/minimum slope gradient (°), β is 10 (α is rainfall intensity) or 1 (α is slope gradient), n is the number of groups of the same slope length and slope gradient (α is rainfall intensity) or groups of the same slope length and rainfall intensity (α is slope gradient), SOCmax/SOCmin is the SOC loss rate under the highest/lowest rainfall intensity or the maximum/minimum slope gradient (g m−2 h−1), ERocmax/ERocmin is the ERoc under the highest/lowest rainfall intensity or the maximum/minimum slope gradient, and γ is 50 (α is rainfall intensity) or 10 (α is slope gradient).

2.4 Data analyses

Nonlinear regression analysis was used to investigate the relationship of the total soil loss with total runoff volume. Exponential and linear regression analysis was used to investigate the relationship of the total soil loss with ERoc and the relationship of the clay content in original soils with the coefficient of regression equation, respectively. Pearson correlation analyses were used to study the correlations of the rainfall intensity, slope gradient, sediment concentration, SOC loss rate, and ERoc. All statistical analyses were conducted with IBM SPSS Statistics 26.0 software and Origin 2019b software.

3 Results

3.1 Soil loss characteristics for black soil, red soil, and loess

The runoff rate of black soil increased slowly within 20 min and stabilized below 100 mm h−1 (Fig. 2). The runoff rate of red soil increased within 15 min and stabilized at 30–100 mm h−1. The runoff rate of silty clay loam of loess increased rapidly within 5 min, and the increased range was approximately 40 mm h−1 (Fig. 3). The runoff rate of silty loam of loess increased slowly within 30 min. The stable value of the runoff rate for loess was 50–120 mm h−1, which is mostly greater than 80 mm h−1. In summary, the time to enter a stable period increased in the following order: loess < red soil < black soil. The stable value increased in the following order: black soil < red soil < loess.

Fig. 2
figure 2

Temporal variations of runoff rate and sediment concentration for black soil and red soil

Full size image
Fig. 3
figure 3

Temporal variations of runoff rate and soil loss rate for loess

Full size image

The sediment concentration for black soil fluctuated approximately in the first 15 min, and then reached a stable stage (Fig. 3). The stable sediment concentration for black soil was less than 18 g L−1. The sediment concentration for red soil varied approximately in the first 20 min. Most sediment concentrations for red soil fluctuated and then entered a stable stage with a sediment concentration of less than 70 g L−1. The soil loss rate of loess increased and then decreased in the first 10 min. The stable soil loss rate of loess was 1–30 kg m−2 h−1. In addition, the soil loss rate for loess increased in the late rainfall period and then decreased. The time to enter a stable period increased in the following order: loess < black soil < red soil. The stable value increased in the following order: black soil < red soil < loess.

∆SCRI and ∆SCSG were investigated to compare the effect of rainfall intensity and slope gradient on soil loss of the three soils (Table 2). Rainfall intensity and slope gradient were significantly correlated with sediment concentration (Table 3; P < 0.01). The ∆SCRI of black soil, red soil, and loess was 0.01, 0.25, and 0.06 g L−1 m−2 h−1, respectively. The ∆SCSG of black soil, red soil, and loess was 0.01, 0.09, and 0.70 g L−1 m−2 h−1, respectively. By contrast, the variation of rainfall intensity had the greatest effect on red soil and the least effect on black soil. The variation of slope gradient had the greatest effect on loess and the least effect on black soil.

Table 2 Effects of rainfall intensity and slope gradient on sediment concentration (g L−1 m−2) and SOC loss rate (g m−2 h−1) for black soil, red soil and loess
Full size table
Table 3 Correlation coefficients of the rainfall characteristics versus the sediment and SOC loss for three soils
Full size table

3.2 SOC loss features for black soil, red soil, and loess

The SOC loss rate of black soil increased with increasing rainfall intensity (Fig. 4). The SOC loss rate of black soil increased initially and then decreased with the increase in slope gradient, similar to the rule of soil loss rate. The SOC loss rate of red soil varied approximately in the first 20 min (Fig. 5). The SOC loss rate of red soil initially increased and then decreased, finally reaching stability. The stable value of the SOC loss rate of red soil was lower than 40 g m−2 h−1, especially for treatments with small slope gradient and low rainfall intensity. The SOC loss rate of loess decreased in the first 10 min (Fig. 6). The soils with high clay content had a considerable decline. The SOC loss rate of loess was lower than 30 g m−2 h−1 after entering the stable stage. Moreover, the SOC loss rate of loess increased in the late rainfall period and then decreased, and the peak value of each rainfall event varied considerably. By contrast, the time for the red soil to enter the stable stage was approximately twice that of the loess. The SOC loss rate of red soil in stable period was mostly lower than that of loess. In addition, the loess formed peaks in the late rainfall, which were not evident in red soil. The ∆SOCRI of black soil was 1.13 g m−2 h−1, nearly 50% of that of red soil (Table 2). The ∆SOCRI of loess was 3.36 g m−2 h−1. The ∆SOCSG of black soil was 0.10 g m−2 h−1, which was 20% of loess. However, the ∆SOCSG of red soil was negative, which indicated that the SOC loss rate for red soil decreased with the increase of in slope gradient except when the rainfall intensity was 90 mm h−1 (Fig. 5). The variation of rainfall intensity and slope gradient had the least effect on black soil but greatest effect on loess.

Fig. 4
figure 4

SOC loss rate and ERoc for black soil under different rainfall intensity and slope

Full size image
Fig. 5
figure 5

SOC loss rate and ERoc for red soil under different rainfall intensity and slope

Full size image
Fig. 6
figure 6

SOC loss rate and ERoc for loess under different rainfall intensity and slope

Full size image

The ERoc of black soil decreased and tended to 1 with the increase in rainfall intensity and slope gradient (Fig. 4). The average ERoc of black soil ranged from 1.0 to 1.5. During erosion process, the initial value of red soil ERoc was between 1.5 and 4 and then tended to 1 with time (Fig. 5). The average ERoc of red soil ranged from 0.31 to 2.21. The ERoc of loess with high rainfall intensity (120 mm h−1) tended to 1 (Fig. 6). However, the stable value of loess ERoc with 90 mm h−1 rainfall intensity was greater than 1 despite a decline period in the beginning. The initial value of loess ERoc ranged from 1 to 2.29. The average ERoc of loess ranged from 0.92 to 1.63. The ∆ERocRI exhibited the following order: black soil < red soil < loess (Table 4). The variation range of ERoc with rainfall intensity for loess was greatest, followed by red soil and black soil. The ∆ERocSG of red soil was greatest, followed by black soil and loess. Thus, slope gradient had a greater effect on the ERoc of red soil. All ERocs decreased with increasing rainfall intensity and slope gradient. The total soil loss and ERoc showed a negative exponential function relationship in black soil, red soil, and loess (P < 0.001) (Fig. 7), but it showed a negative logarithmic function relationship in loess with clay content of 24.9%. ERoc variation in red soil and loess with clay content of 24.9% significantly decreased with increasing soil loss rate. In addition, the clay content and coefficient of regression equation (“a” in y = abx) showed a positive linear correlation in black soil, red soil, and loess (R2 = 0.51, P = 0.176) (Fig. 8).

Table 4 Effects of rainfall intensity and slope gradient on ERoc for black soil, red soil and loess
Full size table
Fig. 7
figure 7

Relationships between total soil loss and ERocs

Full size image
Fig. 8
figure 8

Relationship between clay content and coefficient of regression equation

Full size image

4 Discussion

4.1 Comparisons of soil loss under water erosion for black soil, loess, and red soil

Loess has higher bulk density and smaller porosity than black soil, and appropriate capillary pores cause the infiltration water to become fully retained rather than continuously infiltrated. Loess showed the characteristics of infiltration excess runoff. Thus, entering the stable period of runoff and reaching a high stable runoff required a short time for loess, whereas black soil showed the opposite and tended to saturation excess runoff. Therefore, the runoff for soils with fine texture [15, 19], high bulk density, and low OC content increases rapidly and remains relatively high. The rainfall runoff erosivity, which is positively correlated with precipitation and runoff amount [31], causes the sediment yield of soils to show an order similar to runoff characteristics. For loess, clay mineral composition and carbonate [32] and soil microstructure [33] lead to easy collapse and mass production of eroded materials. Soluble carbonate cemented aggregates can be broken easily when suffering from water wetting [34]. This condition causes loess to enter the stable stage of soil loss in a short period of time and obtain the highest soil loss rate. By contrast, for red soil, aggregates cemented by kaolinite and Fe–Al oxides have good water stability [32]. Furthermore, the dispersion of soil aggregates during rapid wetting is reduced [35]. Thus, relatively more slaking-resistant macroaggregate contents improve the water erosion resistance of red soil. In addition, high aggregate stability and low soil erodibility are related to high SOC content [36, 37], as reflected in black soil. Thus, investigations about the internal mechanisms of the SOC adsorption by clay under water erosion will be beneficial to further understand aggregate stability and soil loss mechanisms.

For the effect factors on sediment loss, rainfall intensity and slope had the least effect on black soil, but rainfall intensity had the highest effect on red soil. This finding may be because the aggregated protection mechanism of clay or Fe–Al oxides in red soil [38] is more sensitive to the change in rainfall intensity than slope. The increase in rainfall intensity is accompanied by increases in raindrop splash intensity [39] and runoff coefficient [18]. The protection mechanism of large amount of SOC on aggregates for black soil is more stable than that of large amount of clay or Fe–Al oxides in red soil. Slope had the highest effect on loess soil because the increase in slope can change the contact angle between raindrops and slope surface and the acting force of raindrops and hydrodynamic characteristics of runoff erosion [40] on the eroded soil surface, to which loess erosion is sensitive. Therefore, the differences in the effects of other factors, such as mineral composition, soil microstructure, SOC features on soil aggregate stability, and soil erosion resistance for the three typical soils, should be investigated further.

4.2 Comparisons of SOC loss and enrichment for black soil, loess, and red soil

The SOC loss rate increases with increasing rainfall intensity, initially increases and then decreases with slope [16, 22], but the increase in SOC loss rate differs in different soil types. The change in SOC loss rate of loess was more sensitive to the changes in rainfall intensity and slope than those of black and red soils. The runoff infiltration and the weak cementation effect of soil minerals on aggregates for loess contribute to its SOC loss and enrichment. Thus, SOC loss rates in different soils are closely associated with their SOC form [41], e.g., SOC easily bonding with clay or large-sized humus wrapped by aggregate.

The SOC in black soil with low clay and high macroaggregate contents [42] is difficult to transport preferentially [43]. By contrast, for zonal acidic soils, e.g., red soil, SOC that easily combines with clay or Fe–Al oxides [44, 45] is easily preferentially transported, leading to a relatively evident ERocs in sediments. For loess, macroaggregates that usually combine with inorganic calcium compounds are easily broken by raindrop impact and form a large amount of clay and silt particles [46] associated with SOC. Thus, different from that of black soil, the SOC of red soil and loess can be evidently enriched in sediments when the soil loss rate is small, and ERocs decrease with increasing soil loss rate [18, 47]. Therefore, the SOC enrichment in sediments easily occurs in the soil containing abundant clay minerals under low rain-induced erosive power. However, for soil with low clay content and high SOC content, SOC cannot be easily enriched in sediments. Soil texture and mineral composition are the key factors for SOC loss and enrichment. Our study suggests that the variations in soil texture and mineral composition between soil types should be considered for the SOC enrichment in SOC dynamic prediction model in future studies. In addition, among loess soils with different soil texture, in soils with high aggregate and clay contents, e.g., loess with 24.90% clay content, the SOC was most evidently enriched in sediments when the soil loss rate was small. With increased soil loss, the major sediment particle composition changes from clay and microaggregates to sand and macroaggregates after rill erosion, and the degree of SOC enrichment rapidly decreases [48]. Thus, the logarithmic function may be more suitable than the exponential function for determining the relationships between ERocs and sediment loss for soils with high aggregate and clay contents. Lastly, more field rainfall experiments about soil erosion and SOC loss should be conducted to verify the results obtained in this study.

5 Conclusion

Soil erosion and SOC loss and enrichment were compared for black soil, red soil, and loess. Results indicated that the stable soil loss rate followed the order: black soil < red soil < loess. Soil and SOC loss rate of loess were the highest and had a noticeable increase in late rainfall events. Moreover, the SOC loss rate of loess was more susceptible to the change in slope gradient than those of two other soils. The SOC loss rate of red soil was more susceptible to the change in rainfall intensity than those of two other soils. Sediment loss amount determines the SOC loss amount for the black soil, loess, and red soil. The ERocs of red soil and loess was higher than that of black soil under low rain-induced erosive power. In black soil, aggregates with high SOC and low clay contents may physically prevent the preferential transport of SOC bonded with clay. The SOC enrichment in sediments easily occur in soils containing abundant clay minerals and SOC but not in soil with high SOC content and low clay content. Relationships between ERocs and total soil loss could be regressed by a negative exponential function. However, for soils with high clay and aggregate, the logarithmic function was suitable for determining the relationships between ERocs and sediment loss. In general, SOC loss regulations in different soils are associated with soil texture, mineral composition, and soil microstructure features. In the future, the effects of combining soil minerals and SOC in soils on sediment and SOC loss should be studied to further understand soil and SOC changes under water erosion.