The study area is situated in Qiaojia County, Zhaotong City, Yunnan Province, China. Qiaojia County borders Dongchuan County in Kunming, Huize County in Qujing, Ludian County and Zhaoyang County in Zhaotong, Yunnan Province, and is separated by the Jinsha River from Huidong County in Liangshan Prefecture, Sichuan Province (Figures 1A, B). Qiaojia is situated within the Sichuan-Yunnan meridian-directional structural system and lies at the intersection of the eastern margin of the northern section of the Sichuan-Yunnan meridian-directional structural belt and the eastern Yunnan structural region. The primary tectonic feature in the area is the Yaoshan tectonic belt. The tectonic development in the study area is primarily characterized by faults (Zhao and Xu, 2013). The area predominantly features a tectonic erosion alpine canyon landform. The Jinsha River flows from the southwest to the northwest of the county, while its tributary, the Niulan River, runs through the northeast. To the west, Yaoshan Mountain, which reaches an elevation of 4,040 m, is situated at the peak of northeast Yunnan. Xiaohe Town is situated in the northwest of the main stream of the Niulan River, nestled in a valley between two mountains, with an elevation of less than 1,000 m (Gao et al., 2011). Qiaojia features a “two rivers sandwiching a mountain” landform, characterized by significant elevation differences and pronounced terrain relief. Xiaohe Town is one of the 17 towns in Qiaojia County and is situated in the northeastern part of the county. The highest point in Xiaohe Town is Mantianxing in Shanbao Village, with an altitude of 3,300 m. Xiaohe is located at the intersection of Lufang Valley and Yinchang Valley, which are part of the Niulan River system—a tributary of the Jinsha River that originates in Songming County, Kunming City. Due to intense erosion by the Niulan River, combined with the effects of fractures and folds, geological disasters frequently occur in Xiaohe Town, resulting in significant economic losses and casualties. The sampling location for this paper is Lanniqing Village, an administrative village within Xiaohe Town. At approximately 4:40 a.m. on 5 September 2019, the northeast slope of Lanniqing Village experienced a landslide due to continuous heavy rainfall and the long-term impact of driving loads on the Shan-ma line road at the trailing edge. Approximately 20,000 cubic meters of fractured mudstone mixed with gray, purple clay, and sub-clay rapidly slid down the slope. The significant height difference exacerbated the situation, leading to a geological disaster that resulted in nine people being buried, two families affected, and substantial destruction of houses, causing significant casualties and economic losses (Figure 1C) (Chen et al., 2024b).

Through a field study of the Lanniqing landslide and analysis of remote sensing images, we determined that the sliding source area of the landslide is approximately 2,000 square meters (Figure 2A). The transport-scraping area measures about 2,500 square meters (Figure 2B), while the deposit and disaster-causing area totals around 5,000 square meters. The sliding-zone soils from the source area, primarily surrounded by the Shan-ma line, the access route to Lanniqing Village, and its left and right boundaries, was studied to assess the impact of different pretreatment methods on the particle size. This involved identifying a suitable pretreatment method and analyzing how particle size affects the mechanical properties of the soil, as well as the role of sliding-zone soil in landslide occurrence. Sampling was conducted in the sliding source area (Figures 1C, 2A). We selected an area with minimal anthropogenic disturbance, removed surface piles and debris formed by rainfall runoff erosion, and collected approximately 2 kg of soil from the smooth surface of the sliding zone in two locations, labeled QJ1-1 and QJ2-1. Additionally, undisturbed subsoil samples from a depth of 100 cm were collected for comparative analysis, labeled QJ1-5 and QJ2-5. These samples were used to determine particle size characteristics. For soil shear strength determination, samples were also taken using 30 cm² × 2 cm ring cutters at the same locations (Figures 2C–F).

To investigate the impact of various pretreatment methods on both sliding-zone soil and undisturbed subsoil, the samples were air-dried and then sieved through a 2 mm mesh. Seven sets of experiments will be conducted for each soil sample, corresponding to seven pretreatment methods (denoted as A, B, C, D, E, F, and G). In total, 28 samples will be tested. To minimize measurement errors, each sample will undergo three repetitions for each experiment, and the final results will be averaged for analysis. Each soil sample tested weighed 0.2 g. The particle size analysis was conducted using the Mastersizer 2000 (M2000), which has a range of 0.02–2000 μm. The Mastersizer 2000 (M2000) features 52 detectors on a single lens, enabling it to measure a wide particle size range from 0.02 to 2000 μm without the need for initial sample fractionation (Callesen et al., 2018). The M2000 uses laser diffraction to measure particle size. The process begins with monochromatic light emitted from a laser, which is expanded into parallel light through a filtering device and directed onto the particles, creating a diffraction pattern. The diffracted light passes through a Fourier lens, focusing the scattered light in the same direction to form a diffraction ring. The intensity distribution of the scattered light is related to the size and quantity of the particles. This scattered light is focused onto detectors, which measure its angular distribution. The light intensity signal is then converted into an electrical signal, which is amplified and used to calculate the equivalent particle size (Ramaswamy and Rao, 2006; Jiang et al., 2015). The different pretreatment methods were applied as follows (Zha et al., 2022):

A: The soil samples were saturated with purified water and left in a beaker for 24 h before being tested without additional treatment.

In the pretreatment methods described, hydrochloric acid (HCl) is primarily used to remove the effects of carbonate cementation, while hydrogen peroxide (H₂O₂) is employed to eliminate the influence of organic matter agglomeration. Additionally, sodium hexametaphosphate ((NaPO₃)₆) is used to address issues related to particle dispersion. The results from various pretreatment methods were compared to assess the impact of different treatments and reagents on particle size characteristics. For instance, methods A and F were compared to evaluate the effect of oscillation, while methods B and F were used to determine the influence of (NaPO₃)₆. The comparison between methods A and G examined the effect of HCl in the presence of other reagents, and the comparison of methods C and E highlighted the effect of HCl when used with H₂O₂ and (NaPO₃)₆. It is necessary to note that mineral composition analysis was performed using X-ray diffraction (XRD), microstructural observations were made with a Scanning Electron Microscope (SEM), and shear testing was conducted using a direct shear test apparatus.

Here, the widely used graphical method (Folk and Ward, 1957) was employed to calculate, compare, and analyze the particle size test results of different pretreatment methods. This method is commonly utilized for grain size analysis of sediments. This method is preferred for its speed and accuracy, effectively minimizing the impact of outliers on the data. It is particularly useful for comparing grain size data across different processing results and can also be applied to study the particle size of sliding-zone soil. The parameters commonly used to describe the statistical characteristics of particle size in this method include: median particle size (Md), mean particle size (Mz), sorting coefficient (Sd), skewness (Sk) and kurtosis (Ku). These parameters offer various perspectives on particle size characteristics. The particle size distribution can be analyzed through the median and mean particle sizes, which reflect the central tendency of particle size. The sorting coefficient indicates the uniformity of the particle size distribution. Skewness measures the asymmetry of the particle size distribution, while kurtosis assesses the sharpness of the particle frequency curve. The cumulative curve (Figures 3B, D, F, H) was derived from the frequency distribution of particle size measurements (Figures 3A, C, E, G) obtained using the M2000. By selecting values from representative points and applying the calculation equations listed in Table 1, the parameters were quantified. In the table, the ϕ values are calculated using the Krumbein transform method: ϕ = −Log₂(D), where D represents the particle size in millimeters (mm). The values ϕ₅, ϕ₁₆, ϕ₂₅, ϕ₅₀, ϕ₇₅, ϕ₈₄, and ϕ₉₅₋ represent the particle sizes corresponding to 5%, 16%, 25%, 50%, 75%, 84%, and 95% on the cumulative curves, respectively.

Figure 3 displays the particle size frequency distribution curves for the sliding-zone soil samples QJ1-1 and QJ2-1, as well as the undisturbed subsoil samples QJ1-5 and QJ2-5. It is evident that the frequency distribution curves for the same sample vary significantly depending on the pretreatment method used. Firstly, the analysis of sample QJ1-1 revealed that the frequency curve for QJ1-1A, which underwent only purification with water, exhibited a single-peak shape. The peak particle size was approximately −1.6ϕ, indicating a coarse particle size. The oscillatory process using an ultrasonic cleaner (QJ1-1F) resulted in noticeable changes, producing a multi-peak morphology. The highest peak occurred around −0.46ϕ, and the overall curve shifted to the right, indicating that the particle size became finer. With the addition of (NaPO₃)₆ (QJ1-1B), the curve transitioned to a single-peak morphology. The peak particle size remained at −0.46ϕ, but the percentage increased from 3.42% to 4.42%, indicating a noticeable rise in the proportion of fine particles. The above analysis indicates that both the oscillatory process and the addition of (NaPO₃)₆ had a substantial dispersing effect on the particles, leading to a notably finer particle size. Additionally, the analysis of the effect of HCl on particle size showed that the curve for soil sample QJ1-1G, treated with HCl, was similar to that of QJ1-1A, with only a slight shift to the right. The peak particle size changed from −1.6 ϕ to −1.36 ϕ, suggesting a slight tendency toward coarser particle size after the addition of HCl, though the change was not pronounced. Comparing QJ1-1B and QJ1-1D shows that, despite the particles being well-dispersed due to the oscillatory process and (NaPO₃)₆, the addition of HCl induces significant changes in particle size. The curve for QJ1-1D shifts to the left, with the peak particle size changing from −0.46 ϕ to −1.48 ϕ. The morphology of the curve transitions from single-peak to multi-peak, indicating a noticeable coarsening of the particle size. This is because, when particles are finely dispersed, incomplete removal and cleaning with HCl can lead to the cohesion of clay particles, significantly reducing the content of fine particles. Additionally, during the removal and cleaning process with HCl, some finer particles may be lost. As a result, these factors contribute to the observed coarsening of the particle size. The results from QJ1-1G reveal that the coarsening effect of particle size caused by HCl is more pronounced compared to other methods. Comparing QJ1-1B and QJ1-1C shows that after adding H₂O₂, (NaPO₃)₆, and using the oscillatory process, the peak particle size remains at −0.46ϕ, and the curve shape is consistent. The addition of H₂O₂ resulted in only a slight increase in the content of fine particles, with the most notable change being a reduction in peak content from 4.42% to 3.83%, representing a decrease of −0.59%. Therefore, the impact of H₂O₂ on particle size in the pretreatment of sliding-zone soils with minimal organic matter is negligible. Additionally, it was found that the traditional E method for particle size testing pretreatment is not the most suitable approach.

In Figures 3E–H, the particle size curves for QJ2-1 show noticeable differences due to various pretreatment methods, with the particle sizes generally being coarse. For sample QJ2-1A, the peak particle size remains large, approximately −1.36ϕ, which is similar to the value observed in QJ1-1A. The curve for QJ1-1F shifts to the right overall, indicating a significant reduction in particle size. In contrast, the curve for QJ2-1F, processed using the same method, exhibits a pronounced multi-peak morphology. The peak particle size for QJ2-1F is about −0.46ϕ, with a percentage of 3.65%, which is comparable to that of QJ1-1F. The peak particle size of the QJ2-1B curve is also at −0.64ϕ, but it shows an increase in fine-grained particles. In contrast, the QJ2-1C curve, which was treated with H₂O₂, differs significantly from QJ1-1. It shifts to the right and displays a multi-peak morphology, with a peak particle size of −0.34ϕ and a percentage of 5.26%. The particle size of QJ2-1 is notably finer, likely because QJ2-1 was sampled from a slope where organic matter, carried by surface water, accumulates, resulting in a higher organic matter content compared to QJ1-1. The addition of H₂O₂, which removes organic matter, had a significant impact on the particle size results for QJ2-1. In other respects, the effects of different pretreatment methods on the particle size results for QJ2-1 are similar to those observed for QJ1-1, and thus will not be further discussed.

Further research is needed to quantitatively analyze the impacts of different pretreatment methods on particle size parameters. Following the particle size grading standards (Shaanxi North team of Chengdu University of Geology, 1978), soil samples are categorized into clay (<2 μm), silt (2–63 μm), and sand (>63 μm), and the percentage for each size interval is calculated (Figure 4). The particle size parameters for each soil sample were then computed using the graphical method, as detailed in Table 2. The impact of different pretreatment methods on the percentage of each particle size interval is also significant. For QJ1-1, the particles are primarily silt, followed by clay, with sand being the least prevalent. After adding purified water (QJ1-1A), the sample showed a lower clay content (17.46%) and a higher silt content (76.09%), with median and mean particle sizes of 6.33ϕ and 6.56ϕ, respectively. In contrast, QJ1-1G, treated with HCl, exhibited a decrease in clay content (16.74%) and an increase in sand content (82.09%), with median and mean particle sizes of 6.70ϕ and 6.86ϕ, respectively, indicating a coarsening of particle size. QJ1-1F, which underwent ultrasonic treatment, demonstrated a significant increase in clay content (40.31%) and a decrease in silt (58.60%), with median and mean particle sizes of 8.27ϕ and 8.06ϕ, respectively. This sample had the finest particle size among all pretreatment methods. Compared to QJ1-1F, the particle size interval percentages, median particle size (8.15ϕ), and mean particle size (8.03ϕ) of QJ1-1B were smaller, indicating that (NaPO₃)₆ had a dispersing effect on the particles. However, this effect was less pronounced than that achieved through ultrasonic oscillation. QJ1-1D, which was treated with HCl, showed a low clay content (21.57%) and median and mean particle sizes of 7.03ϕ and 7.05ϕ, respectively. This indicates that HCl significantly coarsened the particle size compared to samples treated with other chemical reagents. The HCl tends to increase the grain size, a phenomenon supported by existing literature. On one hand, the removal and cleaning effects of HCl can lead to the loss of finer particles, particularly when other chemical reagents are involved (Wang et al., 2018). Additionally, excess HCl can react with clay minerals (such as chlorite and kaolinite), causing particle coagulation and resulting in further coarsening of the particle size (Liu et al., 2012). In contrast, QJ2-1C, treated with H₂O₂, had the highest clay content (56.05%), with silt at 43.95% and no sand, resulting in a median particle size of 9.07φ. This suggests that the dispersing effect of H₂O₂ was more pronounced compared to the effect of (NaPO₃)₆, as seen in QJ2-1B.

Based on the above analysis of particle size curves and parameters, it is clear that both ultrasonic oscillation and (NaPO₃)₆ effectively dispersed the particles during pretreatment. However, the dispersing effect of ultrasonic oscillation was more pronounced than that of (NaPO₃)₆. The use of HCl resulted in a reduction of clay content, an increase in silt, and a general coarsening of the particle size. For sliding-zone soil with a high clay content, these changes can lead to biased particle size test results. Therefore, it is advisable not to use HCl in the particle size measurement of sliding-zone soil. Given that the organic matter in the sliding-zone soil is generally low, the impact of H₂O₂ during pretreatment is minimal. Therefore, this paper recommends discussing the particle size results obtained using pretreatment method C for analyzing landslide occurrences.