Assessment of Soil Degradation by Erosion in a Small Catchment in the Black Soil Region of Northeast China

Abstract

Soil erosion and deposition processes act as key drivers of soil resources distribution across landscapes, affecting soil quality and functionality. However, the impacts of long-term soil erosion on soil quality and degradation in the black soil region remain unclear. Here, we assessed soil quality and degradation as a consequence of historical erosion and soil redistribution in an agricultural catchment in Northeast China. Soil quality indices (SQI) were calculated using both linear and non-linear scoring function methods, along with soil indicator selection approaches, including Total Data Set (TDS) and Minimum Data Set (MDS). Soil degradation indices (SDI), resistance indices (SRI), and the change of SQI (CSQI) were computed and compared. The mean SDI for bulk density (BD) and sand was greater than 0. When BD and sand were excluded, the mean SDI and SRI for the 0–10 cm and 10–20 cm soil layers were −29.8% and −21.9%, and 0.57 and 0.65, respectively. Surface soil (0–10 cm) organic matter (SOM), available potassium (AK), structure stability index (SSI), and total nitrogen (TN) in eroding sites, as well as AK, SSI, SOM, TN, and available phosphorus (AP) in depositional sites, are particularly sensitive to long-term erosion. Field capacity, sand, AK, and SSI were selected to develop the SQI, with the non-linear method utilizing MDS outperforming other SQIs. Most SQIs in eroding sites were lower than those in depositional sites and increased with higher soil redistribution rates. The assessment of soil degradation using SDI, SRI, and CSQI revealed that long-term erosion markedly diminished soil quality, although deposition somewhat alleviated this impact. The lower SQI in the 10–20 cm compared to the 0–10 cm soil layer was primarily attributed to decreased FC, while long-term erosion degraded soil quality by negatively affecting AK and sand content. These findings enhance our comprehension of soil degradation caused by erosion in the Mollisol region of Northeast China.

1. Introduction

Soil is a non-renewable natural asset that is indispensable for food production, ecosystem services, environmental health, and socioeconomic growth. However, over one-third of soil resources worldwide are experiencing degradation due to unsustainable land management practices [1]. These practices deplete soil nutrients and increase soil erodibility, further accelerating soil degradation. Soil degradation negatively impacts soil functions, potentially hindering crop production, food security, and ecosystem services [2]. Consequently, more than 1500 million people globally are at risk of hunger owing to poor soil quality [3]. Therefore, maintaining and enhancing soil quality holds great significance for agricultural production and human well-being.

Soil quality denotes the capability of soil to operate within the boundaries of ecological systems, with the aim of sustaining biological productivity, safeguarding environmental quality, and enhancing plant and animal health [4]. Estimating soil quality is quite difficult because soil is more complex than air and water, consisting of solid, liquid, and gaseous phases. Soil quality is not measurable directly, but can be assessed through quantitative evaluations of soil physical, chemical, and biological properties [5]. Currently, various methods for soil quality assessment have been developed, including soil quality and test kits, fuzzy methods, grey correlation method, soil management assessment frameworks, and soil quality indices [6,7,8,9]. Among these, soil quality indices are the most commonly used way due to their flexibility and simplicity [10]. Generally, the assessment of soil quality index typically includes three key procedures: indicator selection, scoring transformation, and index integration. Common approaches for determining the minimum data set (MDS) include regression analysis, factor analysis, principal component analysis (PCA), and discriminant analysis, as well as various scoring functions. To date, soil quality evaluation frameworks have been widely and successfully applied across diverse ecosystems, including forestlands, coastal zones, wetlands, grasslands, and croplands.

Land use type and agricultural management regimes are closely associated with soil quality, since different management strategies can modify soil physical, chemical, and biological characteristics, potentially leading to soil degradation. Previous studies have indicated that conventional tillage often results in lower soil quality compared with reduced tillage or undisturbed soil conditions. Rainfed croplands generally exhibit inferior soil conditions relative to irrigated systems. Importantly, once natural soils are converted to cultivated land, they become highly vulnerable to degradation across multiple properties, often leading to irreversible reductions in soil quality. Intensive agricultural management has triggered severe soil erosion in many regions [11]. Moreover, erosion processes tend to be more intense in already degraded soils, creating a positive feedback loop that further reduces soil quality over time. For instance, soil physical properties including bulk density, porosity, and aggregate stability are commonly degraded by erosional processes [12]. As erosion intensifies and progressively impairs soil physical structure, the severity and rate of erosion tend to increase further [13]. Soil erosion is accompanied by the selective loss of clay particles and soil organic matter. Previous studies have documented that soil erosion in agricultural landscapes redistributes roughly 42 teragrams (Tg) of nitrogen annually, accompanied by annual fluxes of 2.1–3.9 Tg of organic phosphorus and 12.5–22.5 Tg of inorganic phosphorus [14]. Moreover, approximately 70–90% of eroded topsoil tends to accumulate within the same or neighboring topographical zones rather than being exported outside the source watershed [15]. During sediment deposition, soil organic carbon, total nitrogen, and total phosphorus are generally transported with sediment movement. Such lateral translocation of topsoil can alter the spatial redistribution of soil nutrients. Therefore, the interaction between erosion and deposition gives rise to marked spatial reallocation of soil components at the landscape scale, ultimately influencing soil quality and functional capacity [16,17,18].

The black soil region of Northeast China is one of major grain production base in our country. Unfortunately, the area has experienced significant soil erosion over the past few decades, particularly in agricultural lands, due to long-term intensive agricultural practices, improper management, and the combined effects of multiple erosive forces (wind, water, freezing, and thawing) [19,20]. Consequently, severe erosion not only resulted in soil redistribution but also contributed to soil degradation [21,22,23]. Several studies in this region have explored the impact of different land use, land use conversion, and cultivation on soil quality [24,25,26]. Most recently, serval studies have focused on the effect of erosion and deposition on soil characteristics, bacterial community dynamics, and soil quality on the slope [22,27,28]. However, few studies have examined the variation of soil quality in relation to soil erosion and deposition within the catchment or watershed.

Accordingly, the objectives of this study were to (i) develop soil quality indices (SQI) for assessing soil quality using two methods of indicator selection (TDS and MDS) and two scoring methods (linear and non-linear); (ii) evaluate the effects of erosion and deposition on the degradation of soil quality and soil quality indicators; and (iii) elucidate the influencing paths of erosion and deposition on soil quality.

2. Materials and Methods

2.1. Study Catchment

The Xianhehu catchment is located in Nenjiang Region of Heilongjiang Province, northeast China (125°11′38″~125°12′5″ E, 45°56′22″~45°56′45″ N; Figure 1a) and covers an area of 28.5 ha. The region exhibits a continental semihumid climate, characterized by a prolonged and cold winter. Average annual precipitation is 534 mm, while the temperatures in January and July stand at −20 °C and 21 °C, respectively [29]. The rainy season spans from May to October, while snowfall predominantly occurs between November and April. Catchment elevation varies from 320 to 360 m a.s.l. The catchment has a mean slope of 4.2%, with values varying between 0.4% and 8.4%.

Parent materials are quaternary lacustrine and fluvial sand beds or loess sediments, and soil is Luvic Phaeozem [30]. Historically, most areas of the study catchment were covered with scrubland, which has been converted to farmland since the 1950s. Currently, the dominant land use type is agricultural lands, with mainly maize (Zea mays L.) and soybean (Glycine max (Linn.) Merr.). Downslope tillage represents the primary tillage practice in the study area, which to some extent exacerbates water erosion processes.

2.2. Field Soil Collection and Analysis

Soil sampling was conducted in 2022, when maize and soybean were growing. According to the nuclide tracing results [31], disturbed and undisturbed soil samples were gathered at 0–10 cm and 10–20 cm soil depths across 34 sites where erosion and deposition were more intense (Figure 1c). Soil samples were also collected from forest sites with no erosion or deposition, which had remained undisturbed since at least 1950. Soil properties include soil pH, bulk density (BD), field capacity (FC), total soil porosity (TPO), soil texture, soil organic matter (SOM), total nitrogen (TN), available phosphorus (AP), and available potassium (AK). Soil properties were analyzed using standard laboratory methods (Table S1).

The soil structure stability index (SSI) also serves as an indicator of soil’s resistance to external disruptive forces, with its percentage value calculated as follows [32]:

$$SSI={\displaystyle \frac{SOM}{\left(SIL+CLA\right)}}\times 100$$

(1)

where SIL, CLA and SOM are the content of sixlt, clay, and soil organic matter, respectively.

2.3. Soil Erosion and Deposition Assessment

We used the soil redistribution rate (SRR) to characterize soil erosion and deposition patterns. The SRR was determined using the ¹³⁷Cs tracer method. It has the characteristics of decades-scale retrospective monitoring and no need for long-term field observation [31]. The SRR data are obtained from Fang et al. [31] and applied the Mass balance model 2 to transform ¹³⁷Cs activity measurements into soil redistribution rates (t·ha⁻¹·yr⁻¹) [33].

For a site experiencing soil degradation, the Mass balance model 2 can be formulated as follows:

$${\displaystyle \frac{dA\left(t\right)}{dt}}=\left(1-\Gamma \right)I\left(t\right)-\left(\lambda +P{\displaystyle \frac{R}{d}}\right)A\left(t\right)$$

(2)

where A(t) represents the total ¹³⁷Cs inventory present per unit area (Bq·m⁻²), d indicates the accumulated mass depth that aligns with the typical plowing depth (kg·m⁻²), The variable R denotes the soil erosion rate (kg·m⁻²·yr⁻¹), Γ represents the proportion of freshly deposited ¹³⁷Cs that gets washed away by erosion before it can mix into the topsoil layer, while I(t) stands for the annual deposition rate of ¹³⁷Cs at any given time t (Bq·m⁻²·yr⁻¹), measured in becquerels per square meter per year., λ stands for the yearly depositional flux of ¹³⁷Cs at time t (yr⁻¹), and P serves as a correction factor accounting for variations in particle size.

With respect to depositional zones, the corresponding soil deposition rate is formulated as

$$R^{\prime }={\displaystyle \frac{A_{ex}}{{\int }_{t_0}^t{C}_d\left(t^{\prime }\right)e^{-\lambda \left(t-t^{\prime }\right)}d{t}^{\prime }}}$$

(3)

A_ex denotes the surplus amount of ¹³⁷Cs inventory (Bq·m⁻²), which is determined by subtracting the local direct fallout input Aref from the total measured inventory A(t), while C_d(t′) (Bq·kg⁻¹) indicates the concentration of ¹³⁷Cs found within the sediment deposits. Sites with SRR value “>0” were identified as depositional site, and SRR value “<0” were identified as eroding site. The SRR were classified into six erosion classes based on their magnitudes. The soil redistribution rate (SRR) was categorized into six grades with corresponding names: severe erosion (SRR < −20 t·ha⁻¹·yr⁻¹), moderate erosion (−20~−10 t·ha⁻¹·yr⁻¹), mild erosion (−10~0 t·ha⁻¹·yr⁻¹), mild deposition (0~10 t·ha⁻¹·yr⁻¹), moderate deposition (10~20 t·ha⁻¹·yr⁻¹), and severe deposition (SRR > 20 t·ha⁻¹·yr⁻¹).

2.4. Soil Degradation and Resistance Indices

We used soil degradation index (SDI) and resistance indices (SRI) to quantify the magnitude of degradation of soil indicators. SDI was obtained through the equation listed below [34]:

$$SDI=\left({\displaystyle \frac{X_a-X_b}{X_b}}\right)\times 100\%$$

(4)

where X_a and X_b represent the measured average values of individual soil properties in disturbed agricultural land and undisturbed forestland, respectively. Mean positive and negative SDI values, categorized by soil property, served as land degradation indices in response to the impacts of land use change. In the calculation of SDI, we compared the average value of each soil property in disturbed cultivated lands with that of the same soil property in undisturbed forests.

SRI represents the response of soil properties to the disturbance. SRI was computed using the following equation [35] (Equation (5)):

$$SRI=1-{\displaystyle \frac{2\left|D_0\right|}{(X_b+|D_0\left|\right)}}$$

(5)

D₀ represents the difference between disturbed soil (X_a) following the completion of disturbance and the undisturbed reference soil (X_b). The Soil Resilience Index (SRI) ranges from −1 to +1, with lower values indicating more severe soil degradation.

2.5. Soil Quality Indices (SQI) Assessment

Drawing upon prior research [36,37], SQI was derived through a three-phase methodology [10,26,38]: (a) pinpointing appropriate metrics for the minimum data set (MDS); (b) scoring MDS indicators according to their response characteristics; and (c) synthesizing the individual metric scores to formulate a comprehensive soil quality index. A total of 13 soil indicators were adopted for the TDS in this study, and the MDS was further established via principal component analysis (PCA) and Pearson’s correlation analysis (PCC). The selected soil indicators were then normalized and scored using both linear and non-linear scoring functions.

We applied linear scoring functions (Equations (6) and (7)) to quantify “more is better” and “less is better” indicators, respectively, as follows:

$$LS={\displaystyle \frac{X}{X_{max}}}$$

(6)

$$LS={\displaystyle \frac{X_{min}}{X}}$$

(7)

where X represents a soil indicator’s measured value, and X_max and X_min stand for the maximum and minimum values.

A sigmoidal function (Equation (8)) was employed to carry out the non-linear scoring process, with X representing the measured indicator, X_m denoting the mean, and b standing in for the slope—set at −2.5 for scenarios where “more is better” and +2.5 for cases where “less is better”.

$$NLS={\displaystyle \frac{1}{1+{\left(\frac{X}{X_m}\right)}^b}}$$

(8)

Upon score determination, SQI were calculated through TDS and MDS approaches per Equation (9) [39]:

$$SQI=\sum _{i=1}^n{W}_i\times S_i$$

(9)

where S_i represents the indicator score derived from linear or non-linear curves; n represents the count of soil indicators in the TDS and MDS; and W_i is the weight of soil indicators obtained via PCA. Higher SQI values signify better soil quality and more outstanding performance of soil functions.

The sensitivity index (SI) was utilized to estimate the performance of SQI, as shown in Equation (10):

$$SI={\displaystyle \frac{SQ{I}_{max}}{SQ{I}_{min}}}$$

(10)

where SI represents the sensitivity index for SQI, and SQI_max and SQI_min are the maximum and minimum values of each individual SQI scenario, respectively.

We used an index termed the change of SQI (CSQI) [34] to estimate the degree of soil quality degradation, which is computed via Equation (11):

$$CSQI={\displaystyle \frac{SQ{I}_a-SQ{I}_b}{SQ{I}_a}}\times 100\%$$

(11)

where SQI_a and SQI_b denote the measured average values of individual soil quality indices (SQI) in disturbed croplands and undisturbed forest sites, respectively.

2.6. Data Analysis

Statistical analyses encompassed one-way analysis, Pearson’s correlation analysis, and principal component analysis, all implemented with SPSS 19 (version 25.0 SPSS Inc: Armonk, NY, USA) [40]. The structural equation model was developed using “lavvan” package in R4.4.1 [41]. The fit of SEM was evaluated using the Chi-square test (χ2), p value, goodness of fit index (GFI), and comparative fit index (CFI).

3. Results

3.1. Effect of Erosion and Deposition on Soil Indicators

Soil erosion significantly reduced the contents of soil nutrients (e.g., SOM, TN, and AK), TPO, CP, Silt, and SSI. This demonstrated that soil erosion contributed to lower soil nutrients and soil physical properties in Mollisols, although deposition slightly alleviated such negative effects (

Table 1. Effects of erosion and deposition on soil properties.

Variable	Forest Sites	Eroding Sites			Depositional Sites			Mean
		<−20 t·ha−1·yr−1	−20~−10 t·ha−1·yr−1	−10~0 t·ha−1·yr−1	0~10 t·ha−1·yr−1	10~20 t·ha−1·yr−1	>20 t·ha−1·yr−1
0–10 cm
BD g·cm−3	1.36 ± 0.00 a	1.39 ± 0.14 a	1.42 ± 0.10 a	1.37 ± 0.10 a	1.37 ± 0.19 a	1.27 ± 0.07 a	1.36 ± 0.14 a	1.36 ± 0.05
FC %	34.41 ± 4.97 a	26.53 ± 7.95 a	28.04 ± 5.42 a	28.32 ± 5.48 a	29.89 ± 10.09 a	35.01 ± 6.18 a	31.12 ± 6.66 a	29.82 ± 3.00
TPO %	51.22 ± 3.33 a	39.96 ± 7.92 b	43.27 ± 4.61 ab	41.61 ± 5.78 ab	45.37 ± 8.24 ab	48.95 ± 3.95 ab	45.53 ± 4.79 ab	44.12 ± ± 3.20
CP %	48.97 ± 2.71 a	35.94 ± 7.59 b	39.40 ± 5.17 ab	38.42 ± 5.49 b	40.00 ± 7.55 ab	42.89 ± 5.09 ab	41.83 ± 4.08 ab	39.75 ± 2.48
Silt %	49.13 ± 1.12 a	38.17 ± 4.55 b	39.06 ± 5.17 b	39.23 ± 4.59 b	44.09 ± 1.41 ab	41.76 ± +0.66 b	43.76 ± 7.54 ab	41.01 ± 2.56
Sand %	10.40 ± 1.24 b	27.76 ± 7.04 a	27.15 ± 7.84 a	25.30 ± 7.43 a	18.35 ± 5.00 ab	20.58 ± 2.67 ab	21.57 ± 7.36 ab	23.37 ± 3.72
Clay %	40.47 ± 1.74 a	34.57 ± 2.64 ab	33.79 ± 4.36 b	35.48 ± 3.57 ab	37.56 ± 3.62 ab	37.65 ± 2.17 ab	34.67 ± 1.72 ab	35.62 ± 1.63
pH	5.51 ± 0.21 a	4.89 ± 0.19 b	4.88 ± 0.21 b	4.85 ± 0.12 b	4.96 ± 0.06 b	4.86 ± 0.17 b	5.05 ± 0.22 b	4.92 ± 0.08
SOM g·kg−1	74.35 ± 5.12 a	39.98 ± 3.69 b	42.01 ± 7.09 b	46.74 ± 10.1 b	43.88 ± 5.53 b	43.93 ± 8.70 b	35.41 ± 2.79 b	41.99 ± 3.93
TN g·kg−1	3.49 ± 0.16 a	2.11 ± 0.27 b	2.11 ± 0.34 b	2.15 ± 0.38 b	1.90 ± 0.22 b	2.10 ± 0.36 b	1.99 ± 0.11 b	2.06 ± 0.10
AP mg·kg−1	64.25 ± 16.21 a	30.03 ± 12.94 ab	62.31 ± 42.09 a	49.22 ± 40.60 a	19.93 ± 16.73 b	37.78 ± 16.72 ab	40.38 ± 7.38 ab	39.96 ± 14.76
AK mg·kg−1	321.89 ± 9.63 a	100.08 ± 28.92 b	137.05 ± 61.25 b	105.44 ± 37.06 b	128.25 ± 70.94 b	100.50 ± 19.09 b	109.33 ± 1.04 b	113.44 ± 15.51
SSI %	12.51 ± 1.20 a	5.51 ± 0.37 bc	5.77 ± 0.75 bc	6.31 ± 1.45 b	5.41 ± 1.08 bc	5.56 ± 1.30 bc	4.54 ± 0.48 c	5.52 ± 0.58
10–20 cm
BD g·cm−3	1.30 ± 0.00 b	1.62 ± 0.15 a	1.57 ± 0.14 ab	1.60 ± 0.07 a	1.52 ± 0.06 ab	1.49 ± 0.13 ab	1.50 ± 0.23 ab	1.56 ± 0.06
FC %	34.41 ± 0.85 a	19.80 ± 6.07 b	22.15 ± 5.11 b	20.86 ± 3.97 b	23.87 ± 4.98 ab	26.91 ± 3.66 ab	25.93 ± 10.42 ab	23.13 ± 2.47
TPO %	46.79 ± 2.01 a	34.14 ± 7.86 b	37.87 ± 5.72 ab	35.52 ± 5.94 b	39.35 ± 6.43 ab	42.36 ± 2.08 ab	40.02 ± 11.40 ab	38.18 ± 3.04
CP %	45.25 ± 1.86 a	31.36 ± 6.99 b	34.26 ± 5.47 ab	32.99 ± 5.16 ab	36.14 ± 6.41 ab	39.82 ± 1.82 ab	37.25 ± 11.00 ab	35.14 ± 2.98
Silt %	45.98 ± 3.01 a	39.08 ± 1.54 b	38.30 ± 4.08 b	39.86 ± 5.63 b	44.20 ± 2.04 ab	46.53 ± 4.57 a	42.53 ± 2.76 ab	41.24 ± 1.58
Sand %	11.57 ± 1.15 b	25.78 ± 4.82 a	26.13 ± 7.49 a	24.31 ± 8.52 ab	15.18 ± 3.95 ab	18.46 ± 3.49 ab	20.11 ± 4.88 ab	22.96 ± 2.02
Clay %	43.04 ± 3.18 a	35.14 ± 3.43 b	35.57 ± 4.18 b	35.82 ± 4.01 b	40.62 ± 1.93 ab	35.11 ± 4.72 b	37.36 ± 3.69 b	35.80 ± 0.96
pH	5.39 ± 0.31 a	5.11 ± 0.21 ab	5.17 ± 0.27 ab	5.19 ± 0.23 ab	5.13 ± 0.09 ab	4.92 ± 0.04 b	5.31 ± 0.25 ab	5.14 ± 0.09
SOM g·kg−1	58.14 ± 2.62 a	41.75 ± 6.14 b	40.57 ± 7.19 b	42.73 ± 5.88 b	41.11 ± 8.57 b	44.67 ± 10.64 b	45.95 ± 3.91 ab	43.13 ± 2.33
TN g·kg−1	2.63 ± 0.25 a	2.10 ± 0.27 ab	2.21 ± 0.39 ab	2.05 ± 0.32 b	1.96 ± 0.35 b	2.17 ± 0.58 ab	2.39 ± 0.23 ab	2.18 ± 0.12
AP mg·kg−1	34.84 ± 11.16 a	25.31 ± 12.39 a	28.97 ± 20.89 a	28.69 ± 16.50 a	20.56 ± 8.37 a	20.96 ± 8.37 a	18.58 ± 8.56 a	24.50 ± 5.25
AK mg·kg−1	140.78 ± 60.74 ab	106.17 ± 41.74 ab	121.00 ± 46.07 ab	98.31 ± 29.21 b	101.63 ± 36.67 ab	102.83 ± 9.22 ab	171.33 ± 90.08 a	119.93 ± 26.81
SSI %	10.23 ± 0.74 a	5.62 ± 0.65 b	5.51 ± 0.94 b	5.72 ± 1.02 b	4.88 ± 1.22 b	5.48 ±1.30 b	5.78 ± 0.81 b	5.62 ± 0.25

Note: BD, bulk density; FC, field capacity; TPO, total porosity; CP, capillary porosity; SOM, soil organic matter; TN, total nitrogen; AP, available phosphorous; AK, available potassium; SSI: soil structure stability index. Different lowercase letters following the values indicate significant differences among forest sites and different soil redistribution rate (SRR) classes (p < 0.05).

). However, the impacts of erosion and deposition on soil BD and FC were negligible in the 0–10 cm layer. Soil sand contents across the 0–20 cm soil depth exhibited a trend opposite to that of other soil properties (e.g., SOM, TN, AK, Silt, Clay, and pH). As much, with the increase in soil redistribution rate (SRR), FC, TPO, CP, TN, Silt, and Clay content showed an upward trend in the 0–10 cm soil layer, whereas no obvious trend were observed for SOM, AP, AK, and SSI. These findings indicated that long-term soil erosion markedly influenced soil physical and chemical properties as well as their spatial variability.

3.2. SDI and SRI as Influenced by Soil Properties Under Erosion and Deposition

Compared with the SDI values of other soil properties, sand shows the strongest mobility, followed by BD. The SDIs for the BD and sand were both greater than 0 although the SDI values for BD were slightly lower than 0 under 10~20 t·ha⁻¹·yr⁻¹ and >20 t·ha⁻¹·yr⁻¹ (Figure 2a). Excluding BD and sand, SDIs in the 0–10 cm soil layer varied from −68.91% to −3.02% in eroding sites, while the SDI values ranged from −68.98% to 1.74% in depositional sites. A similar trend was observed in the 10–20 cm soil layer (Figure 2b). For 0–10 cm soil layer, the SDIs were less than the mean SDI value for AK, SSI, SOM, and TN in the eroding site and AK, SSI, AP, SOM, and TN in the depositional sites (Figure 2c,d). Notably, SDIs of sand in the 0–10 cm soil layer ranged from 76.45% to 107.41% in eroding sites and from 143.23% to 162.03% in depositional sites.

Likewise, the SRIs in the 0–10 cm soil layer ranged from 0.18 to 0.96 and from 0.31 to 0.98 in the 10–20 cm soil layer when BD and sand were excluded (Figure 3a,b). The lowest SRI was observed for AK, SSI, SOM, and TN in eroding sites, whereas the lowest SRI in depositional sites was AK, SSI, AP, SOM, and TN in the 0–10 cm soil layer (Figure 3c). For 10–20 cm soil layer, the lowest SRI was SSI, FC, SOM, CP in eroding sites and SDI, AP, SOM and FC in depositional sites (Figure 3d). The SRI was significantly correlated with the SDI in the 0–10 cm (R_adj² = 0.99, p < 0.001) and 10–20 cm (R_adj² = 0.94, p < 0.001) soil layer, suggesting that the tools could equally be employed for measuring the degradation of soil quality (Figure 4).

3.3. Soil Quality Assessment by PCA

PCA was performed to select the minimum data set (MDS). According to

Table 2. Results of principal component analysis for soil variables.

Soil Variable	PC1	PC2	PC3	PC4	Communality
BD	−0.894	0.132	0.207	−0.034	0.860
FC	0.972	−0.116	−0.170	−0.055	0.991
TPO	0.960	−0.109	−0.141	−0.064	0.958
CP	0.954	−0.050	−0.095	−0.095	0.932
Silt	0.006	0.690	−0.540	0.209	0.811
Sand	−0.021	−0.798	0.568	−0.183	0.994
Clay	0.033	0.728	−0.459	0.106	0.753
SOM	0.236	0.731	0.514	−0.168	0.882
TN	0.162	0.459	0.717	0.031	0.752
AP	0.187	−0.596	0.148	0.609	0.782
AK	0.210	0.251	0.505	0.693	0.842
SSI	0.320	0.473	0.740	−0.176	0.904
Eigenvalue	3.842	3.075	2.528	1.017
% of Variance	32.013	25.624	21.067	8.472
Cumulative variance (%)	32.013	57.637	78.705	87.176

Note: boldface factor loadings are considered highly weighted; bold-underlined factor loadings correspond to the indicators included in the MDS.

, four principal components with eigenvalue > 1 explained about 87.176% of soil variability. The first principal component (PC1) explained 32.013% of the total variance, with the remaining three components contributing 25.624%, 21.067%, and 8.472% sequentially. BD, FC, TPO and CP possessed relatively higher loading values in PC1, but only FC was retained according to the significant correlations between these four indicators (Figure 5). Similarly, SSI were selected as the key indicator to represent PC3 based on significant correlations. Given the significant correlations between sand and clay content, and SOM and SSI, only sand content was selected as the indicator for PC2. Meanwhile, AK possessed the highest loading value and thus was retained in PC4 (Figure 5). Overall, FC, sand, SSI, and AK were selected as the key indicators for the MDS, which is used to characterize soil quality changes in the study catchment, and their weight values were ranked as Sand (0.266) and FC (0.266) (equal weight) > SSI (0.242) > AK (0.226).

All soil indicators were transformed using the linear and non-linear scoring functions (

Table 3. Type of scoring curves, the attributes of linear and non-linear equations, and calculated weights for the indictors in the TDS and MDS.

Soil Variable	Scoring Curve	Non-Linear		Linear		Weight
		Mean	Slope	Xmax	Xmin	TDS	MDS
BD	Less is better	1.47	2.5		1.10	0.082
FC	More is better	25.92	−2.5	47.69		0.095	0.266
TPO	More is better	40.62	−2.5	54.01		0.092
CP	More is better	37.08	−2.5	48.85		0.089
Silt	More is better	40.59	−2.5	48.71		0.078
Sand	Less is better	23.51	2.5		10.40	0.095	0.266
Clay	More is better	35.90	−2.5	41.54		0.072
SOM	More is better	43.06	−2.5	64.26		0.084
TN	More is better	2.14	−2.5	2.96		0.072
AP	More is better	34.23	−2.5	131.19		0.075
AK	More is better	118.32	−2.5	356		0.080	0.226
SSI	More is better	5.77	−2.5	8.84		0.086	0.242

). For the TDS, SQI values ranged from 0.496 up to 0.685 for SQI-L, and from 0.381 to 0.572 for SQI-NLT. When employing the MDS method, the SQI values were in the range of 0.369–0.545 for SQI-LM, and 0.368–0.579 for SQI-NLM.

The results demonstrated that accurate quantification of SQI in the Mollisol region is achievable through the most interpretive indicator of this key soil indicator combination (MDS), explaining approximately 64~76% of the overall variation (Figure 6). Regarding four SQI scenarios, the levels of the sensitivity index (SI) associated with SQI-NLM (1.572 in 0–10 cm and 2.014 in 10–20 cm) were comparatively higher than those of other SQIs (Figure 7). Non-linear SQIs, particularly SQI-NLM, were more sensitive to environmental and management conditions than linear SQIs.

3.4. Impact of Erosion and Deposition on Soil Quality

The four SQIs measured in intact forest areas were notably higher than those in erosional sites and deposition zones across the research catchment (

Table 4. Soil quality indices under different erosion rate.

Variable	Forest Sites	Eroding Sites (t·ha−1·yr−1)			Depositional Sites (t·ha−1·yr−1)
		<−20	−20~−10	−10~0	0~10	10~20	>20
0–10 cm
SQI-LT	0.84 ± 0.02 a	0.57 ± 0.04 c	0.61 ± 0.05 bc	0.59 ± 0.05 bc	0.62 ± 0.02 bc	0.65 ± 0.03 b	0.60 ± 0.02 bc
SQI-LM	0.90 ± 0.06 a	0.43 ± 0.03 c	0.47 ± 0.05 bc	0.44 ± 0.08 bc	0.52 ± 0.02 b	0.51 ± 0.02 b	0.47 ± 0.01 bc
SQI-NLT	0.73 ± 0.01 a	0.46 ± 0.03 c	0.51 ± 0.06 bc	0.48 ± 0.05 bc	0.51 ± 0.01 bc	0.54 ± 0.03 b	0.50 ± 0.01 bc
SQI-NLM	0.83 ± 0.03 a	0.44 ± 0.04 c	0.49 ± 0.06 bc	0.46 ± 0.10 bc	0.54 ± 0.03 b	0.53 ± 0.00 b	0.48 ± 0.02 bc
10–20 cm
SQI-LT	0.71 ± 0.03 a	0.53 ± 0.05 c	0.56 ± 0.06 bc	0.55 ± 0.03 bc	0.59 ± 0.02 bc	0.60 ± 0.03 bc	0.61 ± 0.08 b
SQI-LM	0.72 ± 0.05 a	0.40 ± 0.05 c	0.42 ± 0.07 bc	0.41 ± 0.02 bc	0.48 ± 0.02 bc	0.47 ± 0.02 bc	0.51 ± 0.10 b
SQI-NLT	0.66 ± 0.04 a	0.42 ± 0.05 b	0.45 ± 0.07 b	0.44 ± 0.03 b	0.47 ± 0.03 b	0.49 ± 0.05 b	0.50 ± 0.09 b
SQI-NLM	0.77 ± 0.06 a	0.41 ± 0.08 c	0.44 ± 0.10 bc	0.43 ± 0.03 bc	0.49 ± 0.05 bc	0.50 ± 0.04 bc	0.54 ± 0.11 b

Note: Different lowercase letters following the values indicate significant differences among forest sites and different soil redistribution rate (SRR) classes (p < 0.05).

). Within the 0–10 cm soil layer, the four SQIs across the study catchment exhibited a pattern of initial increase followed by a decrease as SRR gradually increased. Notably, soil quality indices (SQIs) in the range of −10~0 t·ha⁻¹·yr⁻¹ were slightly lower than those at −20~−10 t·ha⁻¹·yr⁻¹. A similar pattern was also detected between the 10~20 t·ha⁻¹·yr⁻¹ and >20 t·ha⁻¹·yr⁻¹ intervals. In the 10–20 cm soil layer, SQIs showed an increasing trend with rising soil redistribution rate (SRR). In addition, the composite soil quality index (CSQI) exhibited a distinct increasing trend along the SRR gradient, with the only exception being the 0–10 cm layer at SRR > 20 t·ha⁻¹·yr⁻¹ (Figure 8).

A structural equation model (SEM) was adopted to clarify the mechanisms by which erosional and depositional processes influenced soil quality. The influence of erosion and deposition on SQI was primarily mediated by indirect effects (Figure 9a). Both the total and indirect effects of soil depth on SQI were negative, with standardized path coefficients (SPCs) of −0.17 and −0.22, respectively (Figure 9b). Notably, erosion and deposition exerted a negative direct effect (SPC = −0.08), while their total effect was positive (SPC = 0.28)—a result largely attributed to the indirect effect (SPC = 0.36). Specifically, soil depth indirectly influenced SQI by negatively affecting field capacity (FC; SPC = −0.50), whereas erosion and deposition impacted SQI indirectly through their negative effect on sand content (SPC = −0.24). Among the key indicators, FC, soil structure stability index (SSI), and available potassium (AK) exhibited positive effects on SQI, while sand content had a negative impact with an SPC of −0.54 (Figure 9c).

4. Discussion

4.1. Degradation of Soil Properties Induced by Soil Erosion

In the study catchment, most of soil nutrients levels and SOM were found to be the highest in the forest sites and lowest in the eroding sites (Table 1). A similar trend was observed for FC, TPO, clay and silt content, whereas relatively higher BD and higher sand content were observed in eroding and depositional sites. Notably, we observed that mean SDIs for BD and sand were higher than 0, whereas the SDI for other soil properties were lower than 0 (Figure 2). The higher sand content may be attributed to sediment selectivity. Soil erosion caused light and fine soil particles with low density to migrate from the eroding sites, while coarse particles are usually retained in situ [42]. Related studies have confirmed that soil clay particles and reactive organic carbon are preferentially transported via hydraulic erosion [43]. In this study, erosion induced losses of light and fine particles and SOM can be partly responsible for high BD in the eroding sites [44].

For the 0–10 cm soil layer, AK, SSI, SOM, and TN in eroding sites and AK, SSI, AP, SOM, and TN in depositional sites stated the lowest (i.e., more degraded) SDI values below the mean, suggesting these soil indicators were the most deteriorated properties (Figure 2). Likewise, the SRI values for these soil indicators had the lowest values (Figure 3). Zahedifar (2023) reported negative SDI values for cropland soils, with SDI of SSI and SOM was below the mean, supporting our results [8]. These results indicated that long-term erosion has significantly contributed to topsoil soil degradation. Previous studies have demonstrated that soil erosion strips away vital nutrients through direct transportation, and simultaneously depleting SOM through both physical removal and the accelerated breakdown of remaining organic materials [13,25,45,46,47,48]. In addition, soil acidification caused by long-term erosion and cultivation can result in the leaching of base caution and decreased levels of soil nutrients [49]. Contrary to our expectation, lower SDIs of SSI, SOM, TN and AP in depositional sites than those in eroding sites in the 0–10 cm soil layer, indicating these soil indicators were more degraded in depositional sites (Figure 2). This phenomenon can be attributed to lower TN and AP at soil redistribution rates of 0~10 t·ha⁻¹·yr⁻¹ and SOM at SRR of 20 t·ha⁻¹·yr⁻¹. It could also be partially explained by field survey, indicating that some eroded soil particles flowed out of the catchment, despite a portion being deposited at the catchment outlet [31].

4.2. Effect of Erosion and Deposition and Soil Quality Degradation

Four indicators were selected for the MDS to assess soil quality and degradation in the black soil region of Northeast China (Table 3). This indicated that these indicators played an important role in assessing soil quality in the study area. Previous studies have also identified these indicators as suitable for capturing the dynamics of soil quality [5,24,38,50,51,52]. SSI is an important indicator of soil quality due to the role of soil structure and stability in maintaining soil fertility, water infiltration and resistance to erosion. SOM, a key indicator of soil fertility and crop productivity, was widely recognized and frequently selected in the MDS for assessing soil quality [53].

In the present study, given that SSI calculated from SOM and soil texture (clay and silt), SOM was not maintained in MDS due to its significant correlation with SSI (Figure 5). Soil sand content directly affects soil fertility and soil quality [54,55]. Here, SDIs for sand were greater than 0, indicating severe soil degradation in the study catchment. The result aligns with Liu et al. (2024), who identified sand as one of the most representative indicators of soil quality in the black soil region [24]. Other studies have also suggested that soil sand content is a crucial physical indicator for assessing soil quality [27,56].

The soil in forest sites exhibited greater water retention at the field capacity (FC). Higher BD and sand content and lower SOM could be factors in lower FC availability in eroding and depositional sites [38]. On the other hand, erosion and deposition induced soil loss in eroding sites and soil accumulation in depositional sites, altering the distribution of soil depth and soil particle composition in deposition area. This could alter soil pore characteristics and field capacity. Potassium is the third significant essential macronutrient for plant productivity [57]. Our studies confirmed the results of Zhu et al. (2024) in the hilly region of southern Jiangsu, which emphasized soil AK was a critical factor to consider in the evaluation of soil quality [52].

Soil erosion is a major threat to soil quality degradation [58]. Several studies have documented the effects of soil erosion on soil quality and its degradation, but limited research has focused specifically on how erosion and deposition affect soil quality in the catchment. Our study revealed that long-term erosion has led to serious soil quality deterioration in agricultural land, while deposition somewhat alleviated these negative impacts (Table 4 and Figure 8). Specifically, the lowest CSQI in the 0–10 cm soil layer under SRR less than −20 t·ha⁻¹·yr⁻¹, whereas the highest CSQI were obtained in the 10–20 cm soil layer under SRR greater than 20 t·ha⁻¹·yr⁻¹. Notably, lower CSQI in the 0–10 cm was observed under SRR greater than 20 t·ha⁻¹·yr⁻¹ compared to those under SRR of 0~10 and 10~20 t·ha⁻¹·yr⁻¹. As a result, the correlation between SRR and SQI was weak, despite the mean CSQI in depositional sites being higher than in eroding sites. Fang et al. (2024) stated that slope soil quality was significantly affected by soil erosion in the black soil region of Northeast China [27]. The discrepancy may be partly attributed to variations in slope length, shape, and gradient. In addition, field surveys indicated that while sediment deposition generally occurs at the catchment outlet, some eroded sediments flowed out of the catchment (Figure 1), further complicating the correlation between SRR and SQI in the area.

Furthermore, the study catchment experienced soil quality degradation characterized by reductions in FC, SSI, and AK, coupled with an increase in sand content. The SEM further additionally indicated that the degradation of soil quality in Mollisols was primarily driven by long-term erosion, which indirectly influenced key soil quality indicators such as FC, SSI, AK, and sand content. As previously discussed, erosion and deposition induced soil loss in eroding areas soil accumulation in depositional areas. This dynamic altered the distribution of soil depth, soil particle size, and the contents of SOM and nutrients in both eroding and deposition areas, ultimately impacting soil quality (Figure 9). Interestingly, FC, SSI and sand are soil physical properties, indicating that these indicators are particularly sensitive to the influence of agricultural soil erosion under long-term cultivation. Unfortunately, the degradation of soil physical properties is relatively difficult to restore. Similar results haves been reported in Hailun city, Northeast China. Consequently, if current practices remain unchanged, soil quality is likely to continue to decline over time.

4.3. Implications and Limitations

Agricultural land is the predominant land use in black soil region of Northeast China. In this study, the SDI and SRI were applied to quantify the degradation of soil quality indicators, while also employing a composite SQI and CSQI to measure overall soil quality and its decline. Furthermore, this study quantified the responses of soil quality and its degradation to erosion and deposition processes. Long-term soil erosion has a negative impact on key soil quality indicators (FC, sand, AK, and SSI), leading to overall soil quality degradation. In contrast, deposition can alleviate such negative effects to a certain extent. Moreover, AK, SSI and sand in the 0–10 cm layer and FC, SSI, as well as sand in the 10–20 cm soil layer are more degraded. These findings significantly advance understanding of Mollisol region soil degradation, offering critical scientific guidance, emphasizing the need for more attention to controlling the degradation of soil physical indicators induced by soil erosion under long-term cultivation.

Previous studies have identified that soil physical, chemical, and biological properties were key determinants of soil quality [59]. In this study, the lack of biological indicators incorporated into the SQI restricted our understanding of the pathways driving soil quality degradation. Further research should incorporate a broader range of biological indicators to enhance the applicability of the results. The relatively small size of the catchment led to the deposition of some eroded sediments within the catchment, while a portion has been transported away. This phenomenon indicated that soil quality in depositional sites could be underestimated when evaluated at large regional scale. Within large regions, variables such as topography, land use type and management practices can vary greatly across the black soil region [24,60]. Furthermore, gully erosion has escalated dramatically, increasing from approximately 6853 and 92,899 hm² from 1950 to 2013, thereby reducing arable land and posing a threat to food security. To better understand SQI variability in the black soil region, future studies should consider the impacts of land use, topography, shelterbelts, and gully erosion, as well as their interactive effects. This approach will not only illuminate the challenges posed by climate change but also contribute to the development of effective management strategies.

5. Conclusions

Soil quality was assessed using principal component analysis together with a scoring method in an agricultural catchment in the black soil region of Northeast China. The results indicated that sand content, SSI, AK, and SOM could be more sensitive to soil quality degradation in this region according to the SRI and SDI. Mean SDI values greater than 0 for sand and BD, alongside mean SDI values less than 0 for other soil indicators, highlighted a greater vulnerability to long-term erosion under cultivation. FC, sand content, AK, and SSI were selected to develop an integrated SQI based on PCA and PCC. The SQI derived from the non-linear method showed greater max/min ratios than the SQI calculated using linear method. Consequently, the SQI values obtained from the non-linear scoring method are more sensitive and effective. Meanwhile, SQI based on MDS can serve as a useful tool for the comprehensive evaluation of SQI under erosion conditions. The SQIs in depositional sites were obviously higher than those in eroding sites within this catchment. SEM revealed that erosion and deposition markedly affected the degradation of soil quality by increasing the proportion of sand and decreasing AK and FC within the catchment. This study enhances our understanding of soil quality degradation, which is crucial for maintaining agricultural sustainability in the Mollisol region of Northeast China.