Aggregate-Breaking Mechanism Response to Polyacrylamide Application of Purple Soils in Southwestern China Using Le Bissonnais Method

Abstract

Polyacrylamide (PAM) is a water-soluble polymer with strong cohesiveness and a strong water absorption capacity, and it has been widely used to modify soil structural stability. However, little information is available on the impact of PAM application on the aggregate-breaking process of purple soils in hilly areas of southwestern China. Therefore, the current study aimed to examine the influence of PAM application on the aggregate stability of purple soil in terms of different breakdown mechanisms at different hillslope locations. Three disruptive tests employing the Le Bissonnais method (FW, fast-wetting sieving; SW, slow-wetting sieving; and WS, wet-stirring sieving) were used to determine the mean weight diameter (MWD), geometric mean diameter (GMD), and mass fractal dimension (D) of the soil aggregates, and soil erodibility factor (K) was calculated as an index of soil anti-erodibility. Overall, the major aggregate-breaking mechanism for purple soils was the following: SW (differential swelling) > WS (mechanical breakdown) > FW (slaking). The content of water-stable aggregates (>0.25 mm) obviously rose after PAM application, with the most significant influences shown under FW. A significant difference in MWD was observed between PAM application and without polyacrylamide application (CK) under WS (p < 0.05). However, there was a significant difference in GMD between PAM and CK (p < 0.05) under FW and SW. In comparison with CK, D value in PAM under FW and SW was significantly reduced, mainly at the slope locations of 0 and 20 m. A descending order of FW, WS, and SW was found on the basis of K value at different slope locations. These findings contribute to improved understanding of proper application of soil amendments to control soil and water loss in purple soils.

1. Introduction

Purple soils, developing from rapid physical weathering of nutrient rich sedimentary rock and classified as Inceptisols or Entisols in United States Department of Agriculture (USDA) Taxonomy, are the most important agricultural soils in the Sichuan basin of southwestern China [1]. The soils derived from sedimentary rocks formed during Mesozoic Era (Periods of Triassic, Jurassic, Cretaceous) and Tertiary Period, are peculiar soil resources in China, which haven’t been extensively discovered in other countries. Owing to rapidity of soil formation, complexity of mineral composition, richness of nutrients in soils, loam in texture, good tilth and high productivity of soils, as well as high natural fertility and the suitability to different crops growth in purple soils. In other words, purple soils are valuable agricultural soil resources, meanwhile for purple soils distribute dominantly in southwestern regions of China. In the hilly area in the Sichuan basin, they are unique in China and mainly distributed with a land area of approximately 160,000 km² [1].

However, the thickness of the soil layer is commonly less than 50 cm in the upper slop locations for purple soil areas, which makes it easily eroded by runoff and tillage operation. Therefore, soil erosion of purple soils in hilly areas will have adverse effects on soil quality and form a coarse sand, which indicates desertification. Earlier studies have primarily focused on the agricultural and engineering measures available to prevent soil erosion of purple soil sloping fields [2]; whereas limited research has been conducted on the use of chemical amendments. Polyacrylamide (PAM) is a kind of high molecular weight polymer that can positively affect soil aggregates by bridging the gap between soil particles through its adsorption capacity [3]. PAM can not only enhance the soil’s capacity for water and nutrients conservation and improve its structure [3,4,5], it can also effectively reduce soil erosion and improve soil quality [6]. Therefore, PAM has been widely promoted and applied around the world as an emerging soil amendment technology to control soil and water loss [7,8].

Soil structure is a fundamental property of productive soils, and soil aggregates are the basic units of soil structure [9]. Their stability is closely related to soil erosion and surface runoff and represents one of the main indices for assessing anti-erodibility and soil quality [10,11]. Soil disintegration and mechanical destruction are the main mechanisms of soil aggregate breakdown caused by rainfall [12]. Currently, wet sieving and rainfall simulation are the main traditional methods for determining the stability of soil aggregates [13]. However, these methods combine external forces and the intrinsic properties of the soil and focus only on simulating soil aggregate disintegration caused by fast wetting without considering the impact of different breakdown processes on the stability of aggregates [14].

Three disruptive tests of the Le Bissonnais method (LB) propose that fast-wetting sieving (FW), slow-wetting sieving (SW), and wet-stirring sieving (WS) tests should be used to simulate the disintegration of soil aggregates caused by slaking, differential swelling, and mechanical breakdown [15]. Many studies have demonstrated that LB is more suitable for analysing the stability of soil aggregates under three breakdown mechanisms than the traditional methods (wet-sieving method, rainfall simulation method, etc.) [13,16]. Against this background, the objectives of this study were: (i) to determine the distribution of aggregate sizes by LB at different slope locations; (ii) to ascertain the influence of different breakdown mechanisms on the structure and anti-erodibility properties of the purple soils after PAM application.

2. Materials and Methods

2.1. Study Area

The study site was conducted at the Suining Soil and Water Conservation Experimental Station. The station is located in Suining (30°21′51″ N, 105°28′37″ E) in the middle reaches of the Fujiang River in Sichuan Province, China (Figure 1). The study area has a monsoon-influenced humid subtropical climate. The annual average temperature is 16.7–17.4 °C. The annual average rainfall is 933 mm; the number of frost-free days is approximately 300, and the average altitude ranges from 300 to 600 m. The parent material of the purple soils was the Jurassic Suining formation (J3s) of early Late Jurassic and four samples of red brown purple soils developed from those rocks, which has been eroded and cut and has accumulated water year-round, thus forming a hilly landform. In this area, the cultivated land is dominated by purple soils, which accounts for 62% of the total mainly sloping farmland with poor soil anti-erodibility. The main crops include wheat (Triticum aestivum L.), maize (Zea mays L.), rape (Brassica napus L.), and sweet potato (Ipomoea batatas (L.) Lam.). The station has been engaged in the collection of relevant hydrogeological data in this area as well as the detection of soil erosion in sloping farmland at different slope gradients and the same tillage operation since 1984; thus, it can provide the basic data for this study.

2.2. Soil Sampling

At the experimental station, two neighboring runoff plots with a slope length of 20 m, width of 5 m and mean slope gradient of approximately 15° were chosen as the application of PAM and CK (without application PAM, Figure 1c). The entire plot was divided into five slope locations: 0 (from 0 to 4 m), 5 (from 4 to 8 m), 10 (from 8 to 12 m), 15 (from 12 to 16 m), and 20 m (from 16 to 20 m) from the summit to bottom. Previous studies have suggested that PAM has a significant soil amelioration effect at application rates of 0.8 g m⁻² in purple soil areas [17], and the PAM application rate was 0.8 g m⁻² in this study. PAM was dissolved in deionized water and stirred by hand. Then, the prepared PAM solution was sprinkled evenly on the soil surface of the plot using a knapsack sprayer (capacity 3 L) at the rate of 8 kg ha⁻¹. When the PAM had completely penetrated the soil after three days, soil sampling for the determinations of aggregate size fraction, were carried out along the slope at different slope locations of the surface layer. More detailed procedures of the sampling can be found in the literature via Wang et al. [9]. Soil samples were taken along the two sampling lines of the slope between which there was a contour distance of 2 m for the PAM application and CK treatments.

These samples were then packed into plastic boxes and brought back to the laboratory. Furthermore, soil structure destruction was avoided as much as possible during the collection and transport process. The determination of organic matter content was carried out using wet oxidation with K₂Cr₂O₇, soil bulk density was determined with oven-dried weight and sample volume. A portable soil compaction meter (SL-TSA, Baziu Shiye, Shanghai, China) was used to measure soil compaction in different slope locations. Soil physical and chemical properties of the PAM and CK treatments are shown in

Table 1. The physical and chemical characteristics for the experimental sites (±Standard Deviation).

Item	Bulk Density (g cm−3)	Soil Compaction (kPa)	Soil Water Content (%)	Organic Matter (%)	Total Nitrogen (%)	pH	Particle-Size Fraction (%)
							2–0.02 mm	0.2–0.002 mm	<0.002 mm
PAM	1.43 ± 0.10	993.94 ± 116.27	14.19 ± 1.43	1.24 ± 0.15	0.10 ± 0.01	8.28 ± 0.05	39.80 ± 2.37	49.08 ± 1.50	11.13 ± 1.00
CK	1.41 ± 0.08	1007.51 ± 296.77	12.78 ± 1.95	1.15 ± 0.29	0.12 ± 0.02	8.16 ± 0.09	36.40 ± 3.74	53.43 ± 4.77	10.17 ± 1.18

, and a detailed test method used can be found in Liu [18]. The analysis of the paired-samples t-tests showed that there was no significant difference in the soil physical and chemical properties between PAM and CK treatments (p > 0.05).

Figure 1. Study area location and study site layout: (a) China; (b) Sichuan province; (c) Study site.

Figure 1. Study area location and study site layout: (a) China; (b) Sichuan province; (c) Study site.

2.3. Laboratory Methods

The aggregate samples were stripped into approximately 10 mm piles along their natural joints, and the plant roots and stones were removed. After air-drying, soil aggregates with a size ranging from 5 to 2 mm were selected via dry sieving. Different breakdown mechanisms of soil aggregates under rainfall conditions were simulated using three sieving tests of LB (FW, fast-wetting sieving; SW, slow-wetting sieving; and WS, wet-stirring sieving). The aggregates of 5–2 mm were roasted in an oven at 40 °C for 24 h. After the soil water content was equilibrated, the particle-size distribution characteristics of soil aggregates under three sieving tests of LB were performed as follows: (1) FW: 5 g of aggregates was immersed in deionized water for approximately 10 min and then the water was absorbed by pipettes; (2) SW: 5 g of aggregates was placed on filter paper with a tension of −0.3 kPa for approximately 30 min to make the aggregate completely wet; and (3) WS: 5 g of aggregates was immersed in ethanol 95% for 10 min and then the ethanol was sucked off with pipettes. Then, the aggregates were transferred to a conical bottle with 50 mL of deionized water, and 150 mL of deionized water was added. The bottle was sealed and turned upside down 20 times, and the coarse dispersion was precipitated by standing for 30 min. Finally, the rest of the water was absorbed by pipettes.

The samples of the above three sieving tests were transferred to a sieve with 0.05 mm openings and completely immersed in ethanol 95%. The samples were oscillated up and down 20 times with an amplitude of 2 cm. Then, the samples were transferred to a beaker and the ethanol was dried in an oven at 40 °C. The mass fractions of different particle sizes were obtained by passing through sieves with 5, 2, 1, 0.5, 0.25, and 0.053 mm openings, in sequence.

2.4. Calculations

The mean weight diameter (MWD) and geometric mean diameter (GMD) can reflect the distribution characteristics and stability of soil aggregates. The larger the GMD and MWD values were, the stronger the stability and anti-erodibility of soil aggregates would be [14]. MWD and GMD were calculated as follows:

$$\mathrm{MWD}=\frac{{\displaystyle \sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{W}}_{\mathrm{i}}{\mathrm{X}}_{\mathrm{i}}}}{{\displaystyle \sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{W}}_{\mathrm{i}}}}$$

(1)

$$\mathrm{GMD}=\exp \left\{\frac{{\displaystyle \sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{W}}_{\mathrm{i}}{\mathrm{lnX}}_{\mathrm{i}}}}{{\displaystyle \sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{W}}_{\mathrm{i}}}}\right\}$$

(2)

where W_i is the proportion of aggregates in the size class i; and X_i is the mean diameter of each size class (mm).

The soil fractal dimension (D) reflects the composition characteristics of soil aggregates, which was expressed based on the following expression [19]:

$$\lg \left[\frac{\mathrm{M}(\mathrm{r}<\mathrm{R})}{{\mathrm{M}}_{\mathrm{T}}}\right]=(3-\mathrm{D})\lg \left(\frac{\mathrm{R}}{{\mathrm{X}}_{\max }}\right)$$

(3)

where R is the average diameter of aggregate at a level, mm; M(r < R) is the mass of aggregate with particle size smaller than R; M_T is the total mass of aggregates lower than X_max; and X_max is the maximum diameter of the aggregates.

In the case of limited data on soil physical and chemical properties, soil erodibility factor (K) value can be calculated according to the GMD [20] as follows:

$$\mathrm{K}=7.954\left\{0.0017+0.0494\exp \left[-\frac{1}{2}{\left(\frac{\lg (\mathrm{GMD})+1.675}{0.6986}\right)}^2\right]\right\}$$

(4)

3. Results

3.1. Aggregates Size Distribution with Three Breakdown Mechanisms at Different Slope Locations

Under three breakdown mechanisms of LB, the particle sizes of the aggregates in PAM and CK were mainly concentrated in the range of 0.25–0.053 mm, and the mean contents of the aggregates separately accounted for 39.88% (29.94–57.08%) and 42.92% (29.90–53.43%) in these plots, respectively (Figure 2). However, the average content of aggregates with a particle size of 5–2 mm in PAM was obviously higher than that in CK (p < 0.01). PAM mainly changed the particle-size distribution of soil by reducing the content of aggregates with a particle size of 0.25–0.053 mm and increasing the content of aggregates with particles sized 5–2 mm. A higher content of soil aggregates with a particle size larger than 0.25 mm corresponded to a better soil structural stability. The contents of aggregates with a particle size larger than 0.25 mm showed a descending order when using SW (51.52%), WS (46.61%), and FW (40.39%) in PAM. Similarly, a descending order of SW (47.21%), WS (42.14%), and FW (34.35%) was found according to the contents in CK. These findings suggest that the content of water-stable aggregates (>0.25 mm) obviously rose after PAM application, with the most significant influences were shown under FW.

The content of aggregates with particle sizes of 5–2 mm in CK gradually increased along the downward direction of the slope, while they progressively decreased in PAM. The content of aggregates with particle sizes of 0.25–0.053 mm gradually reduced along the downward direction of the slope in CK, while that in PAM gradually increased. At the 0 m slope location, the contents of aggregates with particle sizes of 5–2 mm, 2–1 mm, and 1–0.5 mm in PAM application were significantly higher, yet those of aggregates with particle sizes of 0.25–0.053 mm and smaller than 0.053 mm were remarkably lower than those in CK (p < 0.05). The coefficients of variation (CVs) of soil aggregate contents larger than 0.25 mm at different slope locations in PAM application and CK were 24.50% and 19.19%, respectively, indicating that slope location exerted an obvious influence on the content of water-stable aggregates. Particularly, the average contents of soil aggregates larger than 0.25 mm in PAM application increased by 76.78%, 48.27%, and 49.86% at the 0 m location compared with those in CK under FW, WS, and SW, respectively.

3.2. The MWD and GMD of Soil Water-Stable Aggregates at Different Slope Locations

The MWD and GMD at different slope locations in PAM application and CK are presented in Figure 3. Three sieving tests of LB were ranked in a descending order as SW, WS, and FW according to the MWD and GMD at different slope locations, thus showing that the expansive effect of soil clay had the least destructive effect on soil aggregates in the study area. MWD and GMD decreased along the downward direction of the slope in the PAM application and CK, with higher values at the upper slope locations (0–4 m) compared with the other slope locations. At the 0 m location, The MWD in PAM application increased by 63.08%, 45.56%, and 73.19%, and GMD increased by 47.44%, 43.88%, and 75.29% under FW, WS, and SW in comparison with that in CK, respectively. The MWD and GMD under FW, WS, and SW significantly increased at the 0 and 20 m slope locations of PAM application (p < 0.05) compared with CK.

The results of paired samples t-tests of the MWD and GMD in PAM application and CK demonstrated that differences of MWD reached an extremely significant level under FW and SW (p < 0.01) (

Table 2. Paired sample t-test (PAM–CK) for MWD and GMD in three breakdown mechanisms.

Item	Breakdown Mechanism	Paired Difference (PAM–CK)					t	df	p Value
		Mean	Standard Deviation	Standard Error	95% Confidence Interval of the Difference
					Lower	Upper
MWD	FW	0.148	0.139	0.036	0.071	0.225	4.107	14	0.001
	WS	0.156	0.209	0.054	0.040	0.272	2.888	14	0.012
	SW	0.225	0.191	0.049	0.119	0.330	4.555	14	0.000
GMD	FW	0.039	0.055	0.014	0.008	0.070	2.746	14	0.016
	WS	0.042	0.091	0.023	–0.007	0.093	1.819	14	0.090
	SW	0.075	0.099	0.026	0.020	0.129	2.926	14	0.011

). A significant difference in MWD was observed between PAM application and CK under WS (p < 0.05). Furthermore, there was a significant difference in GMD between PAM application and CK (p < 0.05) under FW and SW, while the difference was statistically insignificant under WS (p > 0.05). MWD and GMD at different slope locations in PAM application were always higher than those in CK under three sieving tests. These results suggest that PAM application improved the structure of soil aggregates. The MWD and GMD at different slope locations in PAM and CK treatments are presented in Figure 3. Three sieving tests of LB were ranked, in descending order, as SW, WS, and FW according to the MWD and GMD at different slope locations, thus showing that the expansive effect of soil clay had the least destructive effect on soil aggregates in the study area. The MWD and GMD decreased along the downward direction of the slope in PAM application and CK, with higher values at the upper slope locations (0–4 m) compared with the other slope locations by three sieving tests of LB. At the 0 m slope location, MWD in PAM increased by 63.08%, 45.56%, and 73.19%, and GMD increased by 47.44%, 43.88%, and 75.29% under FW, WS, and SW in comparison with that in CK, respectively. MWD and GMD under FW, WS, and SW significantly increased at the 0 and 20 m slope locations of PAM application (p < 0.05) compared with CK.

The results of paired samples t-tests of MWD and GMD in PAM and CK treatments demonstrated that differences of MWD reached an extremely significant level under FW and SW (p < 0.01) (Table 2). A significant difference in MWD was observed between PAM application and CK under WS (p < 0.05). Furthermore, there was a significant difference in GMD between PAM and CK treatments (p < 0.05) under FW and SW, while the difference was statistically insignificant under WS (p > 0.05). MWD and GMD at different slope locations in PAM application were always higher than those in CK under three breakdown mechanisms. These results suggest that PAM application improved the structure of soil aggregates, which made the structure more stable.

3.3. Changes in the D Value of Soil Aggregates in Three Breakdown Mechanisms

The D value can reflect soil texture, water stability, uniformity degree, and fertility characteristics and reveal regular changes [21]. In comparison with CK, the D value in PAM application under FW and SW was significantly reduced, mainly at the slope locations of 0 and 20 m (

Table 3. Changes in the D value with three breakdown mechanisms at different slope locations.

Slope Location	FW		WS		SW
	PAM	CK	PAM	CK	PAM	CK
0	2.37 ± 0.13	2.54 ± 0.01	2.47 ± 0.03	2.47 ± 0.04	2.38 ± 0.13	2.51 ± 0.04
5	2.32 ± 0.07	2.37 ± 0.08	2.42 ± 0.05	2.40 ± 0.09	2.33 ± 0.07	2.39 ± 0.10
10	2.59 ± 0.05	2.46 ± 0.10	2.49 ± 0.03	2.55 ± 0.07	2.59 ± 0.05	2.48 ± 0.09
15	2.40 ± 0.12	2.46 ± 0.01	2.51 ± 0.02	2.38 ± 0.03	2.41 ± 0.12	2.40 ± 0.06
20	2.32 ± 0.05	2.47 ± 0.08	2.41 ± 0.10	2.48 ± 0.05	2.29 ± 0.06	2.47 ± 0.09
Mean	2.40 ± 0.11	2.46 ± 0.06	2.46 ± 0.04	2.46 ± 0.07	2.40 ± 0.12	2.45 ± 0.05
CV (%)	4.65	2.46	1.77	2.77	4.83	2.04

), suggesting that PAM application can increase the stability of soil aggregates. Specifically, it dramatically reduced the effects of dissipation of entrapped air in aggregates at the slope locations of 0, 15, and 20 m and fracturing caused by clay expansion after wetting the soil aggregates, thereby improving the stability of soil aggregates. The tests of between-subjects effects demonstrated that the D value was significantly affected by the slope locations and breakdown mechanisms (p < 0.05), although their interactions did not significantly influence the D value (p > 0.05;

Table 4. Test of within-subject effects on the D value in PAM application and CK.

Item	III Sum of Square		df	Mean Square		F Value		p Value
	PAM	CK		PAM	CK	PAM	CK	PAM	CK
Correction model	0.373 a	0.285 b	14	0.027	0.020	5.261	3.629	0.003	0.001
Intercept	275.816	281.915	1	275.816	281.915	54,427.680	50,221.861	0.000	0.000
Slope location	0.189	0.182	4	0.094	0.091	18.647	16.212	0.000	0.000
Sieving test	0.116	0.077	2	0.029	0.019	5.706	3.416	0.002	0.020
Slope location × sieving test	0.069	0.026	8	0.009	0.003	1.691	0.589	0.142	0.779
Error	0.152	0.168	30	0.005	0.006
Total	276.342	282.369	45
Total correction	0.525	0.454	44

Note: ^a R² = 0.605 (Adjust R² = 0.420); ^b R² = 0.372 (Adjust R² = 0.079).

). Based on F test statistics, the main effects (slope location and breakdown mechanism) had relatively little influence on the D value for both PAM and CK treatments.

3.4. Change Characteristics of Soil Erodibility under Three Breakdown Mechanisms with LB at Different Slope Locations

A smaller K value corresponds to stronger erosion resistance, whereas a larger K can lead to the weakening of anti-erodibility [20]. As shown in Figure 4, a descending order of FW, WS, and SW was found on the basis of K value at different slope locations, which meant that the K value was the highest under FW in this study area, followed by WS and then SW. It is worth noting that the most obvious changes in the K value were found at the 0 and 20 m slope locations for both PAM and CK treatments. In comparison with those in CK, the K value after PAM application under SW at the 0 and 20 m slope location was reduced by 39.39% and 28.19%, respectively. Therefore, PAM can most obviously improve the erosion resistance of soil at the 0 and 20 m locations of sloping land, which is consistent with MWD, GMD, and D results.

Except for the 10 m slope location, where the K value in PAM application under SW was higher than that in CK, the K value at the remaining locations in the PAM application was lower than those in CK under the three breakdown mechanisms. Through the paired samples of t-tests, the differences in the K value under the three breakdown mechanisms were analyzed as shown in

Table 5. Paired sample t-test (PAM-CK) for the K value in three breakdown mechanisms.

Breakdown Mechanism	Paired Difference (PAM–CK)					t	df	p Value
	Mean	Standard Deviation	Standard Error	95% Confidence Interval of the Difference
				Lower	Upper
FW	–0.027	0.035	0.009	–0.047	–0.008	–3.004	14	0.009
SW	–0.019	0.033	0.008	–0.037	–0.001	–2.220	14	0.043
WS	–0.022	0.032	0.008	–0.040	–0.005	–2.710	14	0.017

. The results demonstrated that the differences in the K value with FW between the PAM application and CK were extremely significant (p = 0.009). The average K values with SW in the PAM application and CK were 0.12 (0.10–0.14) and 0.14 (0.12–0.14), respectively. Furthermore, the average K value of the whole slope in the PAM application under WS was significantly higher than that in CK (p < 0.05), indicating that PAM can significantly alter the physical and chemical properties and enhance the erosion resistance of purple soil.

As shown in Figure 5, the content of 5–2 mm aggregates was significantly positively correlated with the MWD and GMD (p < 0.05), yet significantly negatively correlated with the K value (p < 0.05) for both PAM and CK treatments. Meanwhile, the <0.053 mm aggregates was significantly negatively correlated with the MWD and GMD (p < 0.05), but significantly positively correlated with the K value (p < 0.05). Moreover, the MWD and GMD weren’t significantly correlated with D (p > 0.05), whereas they were significantly negatively correlated with the K value (p < 0.05), indicating that soil anti-erodibility increased with the improving stability of the soil structure on the steep hillslope. The principal component analysis results of different aggregate sizes, MWD, GMD, D, and K values are presented in Figure 6. The first two principle components axe explained 80.3% of the total variability for PAM application and 84.9% for the CK treatment, in which the variability in soil anti-erodibility could be explained by the contents of different aggregate sizes (Figure 6). Correlation analysis and principal component analysis show that soil anti-erodibility was mainly affected by the contents of large aggregates and the disintegration of soil aggregates.

4. Discussion

Soil aggregates are an important component of soil structure, and their amount and composition affect the maintenance and supply of soil moisture and nutrients [22]. The results indicate that PAM is a linear water-soluble polymer, and its molecular chain is relatively wide; thus, PAM is often used as a thickener, binder, flocculant, and soil amendment [23]. In this study, the MWD and GMD in PAM were larger than in CK, while the D and K values in the PAM application were smaller than in CK. These results show that the destruction degree of soil aggregates in the purple hilly area was lower in the plot treated with the PAM application than in the CK in terms of dissipation, clay disintegration, and mechanical disturbance, which shows that PAM could increase the stability of soil aggregates and enhance soil anti-erodibility. The content of water-stable aggregates was obviously increased after PAM application. This result is also consistent with previous studies [8,24], which indicate that PAM improves soil aggregate formation via the cohesion of adjacent particles.

The fragmentation of soil aggregates was considered from different breakdown mechanisms via three sieving tests of LB. FW simulated the destruction of soil aggregates caused by dissipation during fast wetting in heavy rains, SW simulated the destructive effect caused by swelling of clay particles after wetting in light rain, and WS simulated the destructive effect of mechanical disturbances on soil aggregates in moderate rain [25]. The results showed that FW caused the most serious destruction to soil, and it was followed by WS and SW, which caused the least destruction to soil. These findings demonstrated that the main fragmentation mechanism of soil aggregates on sloping farmland in the experimental area was the dissipation of heavy rains. This result was similar to the study results of Loess Plateau by Zeng et al. [26]. When rainfall rapidly wets the soil aggregates, the air in the aggregates is quickly compressed and the aggregates are broken, which caused macro-aggregates to disintegrate into micro-aggregates. Heavy rain occurs frequently in this study area [27]; therefore, the effect of heavy rain on soil aggregates should be investigated to improve the soil quality.

When PAM meets water, the hydrogen group in the molecular chain adsorbs the soil clay particles, the molecules entangle with each other in the form of a chain bridge, and the dispersed soil particles in the series are interwoven with each other such that the soil particles gradually agglomerate into macro-aggregates, thus forming strong water-stable aggregates [28,29]. In addition, PAM can form a protective film on the surface of soil when dissolved in water, and this film can effectively reduce the impact of raindrop beating and lessen the destruction and dissipation of soil [5,30].

The slope locations had a significant effect on the soil structure and anti-erodibility of sloped farmland in both PAM and CK. The MWD, GMD, and K values were more obvious at the summit locations (0 m) and the lower slope locations (20 m). The results showed that the impact of the PAM application on the structure and erodibility of purple soils was primarily associated with the boundary effect of sloping farmlands. It has been proven that the soil’s physical and chemical properties and the structure of soil changed significantly at the upper slope and lower slope boundaries [9]. In addition, the thickness of the soil profiles tended to gradually increase from the upper to the lower slope locations because of erosion at the upper slope locations caused by long-term tillage and soil deposition at the lower slope locations in the hilly areas [9]. Thus, soil depth in hillslopes will most likely affect the structure and erodibility of the upslope and downslope boundaries of the field. In addition, organic matter in soil can significantly affect the adsorption of PAM on soil particles because the anion of PAM has a stronger adsorption capacity for soil with low organic matter content [31,32]. Long-term soil erosion in this region resulted in the gradual decrease in soil organic matter content at the upper slope locations [33,34]; therefore, the lower content of soil organic matter at the upper slope would be favorable for PAM adsorption to soil particles.

5. Conclusions

The objectives of this study were to test the impact of the PAM application on the structure and different breakdown mechanisms of purple soil using three sieving tests of LB in a hilly landscape. The soil aggregate stability responded to slope location and differently according to each breakdown mechanisms, including slaking due to fast wetting, differential swelling breakdown caused by slow wetting, and mechanical breakdown through wet stirring. In purple hilly area of southwestern China, the order of the major aggregate-breaking mechanism was as follows: SW > WS > FW. The content of water-stable aggregates (>0.25 mm) obviously increased after the PAM application, with the most significant influences were shown under FW. The dissipation of slaking due to fast wetting represented the most significant destruction to the structure of purple soils among the three aggregate-breaking mechanisms. Significant differences in soil aggregate and erodibility property after PAM application were found close to the upper and lower slope boundaries of the field, showing that the boundary effect was prominent on steeply sloping land. Our experimental results suggest that the PAM application yield a significant effect on aggregate-breaking mechanism, and effectively enhance the aggregate stability and soil anti-erodibility of purple soils in southwestern regions of China. Therefore, the PAM application can provide measures available to protect land resources of purple soils. For future work, we recommend that PAM concentration should be tested with different physical properties of the soils as well as other types of soils (i.e., black soil and loess soil). We also intend to explore the relationship between PAM adsorption by the soil and the impact of the plants, microorganisms, and animals.