2.1 Overview of the research area
The study was conducted at the Soil and Water Conservation Science and Technology Park in Wuhua County, Guangdong Province (Fig. 1). The park is located between 23°23′ and 24°12′ N latitude and 115°18′ and 116°02′ E longitude. Wuhua has a humid southern subtropical monsoon climate at the low-to-mid latitudes. The annual average temperature of the county between 1979 and 2000 was 21.2°C. The annual average precipitation, thunderstorm days, and frost-free period were 1519.7 mm, 77 days, and 330 days, respectively. Wuhua has a complex geological composition, primarily consisting of five major types of rocks: intrusive rocks, extrusive rocks, sandy rocks, limestone, and granite. These geological formations give rise to three major landforms: mountains, hills, and basins.
2.2 Preparation of test materials
The field test and soil sample measurement were conducted at the Soil and Water Conservation Technology Demonstration Park in Wuhua County, Guangdong Province and the Guangdong Environmental Science and Technology Public Laboratory, respectively. The study equipment comprised a portable rainfall system (Fig. 2), a laser disdrometer, a digital camera (ORDRO), and consumables such as buckets, sample boxes, sample bags, and testing reagents. The study also used P. vulgaris of the same year (planted 15 cm between plants × 20 cm between rows) and the redworm Eisenia foetida (5 g/pc; 400 g/m2 density). The red soil is predominant in South China; soil properties are presented in Table 1.
2.3 Experimental sample site establishment and survey
One-year-old P. vulgaris, a perennial herbaceous plant of the Lamiaceae family, was planted in the experimental plots. Standard planting practices were followed for propagation, seedling establishment, water and fertiliser management, cultivation, weeding, harvesting, and processing of the plant. The earthworms selected for the study were the “Taiping II” redworm—E. foetida. These earthworms were chosen based on specific criteria, including ≥ 90 mm body length, ≥ 3 mm body width, and ≥ 5 g weight (200 g/m2 density). We performed a sample plot survey before and after planting P. vulgaris to determine the soil’s physicochemical and biological properties. All the indicators were determined according to the standard practices. The simulated rainfall experimental sample site had dimensions of 1.2 × 1.0 m. Four erosion collectors were installed along contour lines in all four directions, each measuring 5 × 5 × 20 cm. The investigation plots had dimensions of 5.0 × 1.2 m to assess the soil properties and characteristics of P. vulgaris. The sample plots were ploughed and transplanted with P. vulgaris, and earthworms were introduced without further cultivation.
2.4 Design of field simulated rainfall experiments
Five slope angles (2°, 5°, 10°, 15°, and 20°) were chosen for the experiment, with five plots of earthworm treatment and bare ground plots for each slope angle. One rainfall test at five rain intensities (0.7, 1.0, 1.6, 2.2, and 2.6 mm/min) was conducted on the 10° slope with earthworm treatment and the bare ground sample plots. Furthermore, one rainfall test was conducted at 1.6 mm/min on the other four slopes (2°, 5°, 15°, and 20°), with 36 rainfall tests in total. Experimental treatments and sessions are presented in Table 2.
2.5 Measurement of soil and grass cover indicators
Soil bulk density was determined using the ring knife sampling method. Soil density was determined using the specific gravity determination method. Additionally, soil organic matter content was determined using the potassium dichromate-sulfuric acid digestion method. Soil mechanical composition was determined using a laser particle size analyser—MasterSizer 2000 (Malvern, UK). Soil porosity was determined via soil bulk density and density calculations. Plant root system characteristics (including root length, surface area, diameter, and volume) were determined using the WinRHIZO root system analyser. The dry weight of the root was obtained by rinsing the live roots, drying them in an oven at 80°C to a constant mass, and weighing them. Before the rainfall test, a digital camera was used to take portrait photos of the plot, which were subsequently analysed and processed in the laboratory. The method was to select the vegetation part of the photos as the area of interest (AOI) in the imageJ environment. Unsupervised classification was conducted using the minimum distance method. The classification results were visually interpreted and grouped into two major categories: vegetation cover and no vegetation cover. The number of image elements was counted to accurately determine the vegetation cover.
2.6 Field simulated rainfall experiments
Field simulated rainfall was conducted after the growth of P. vulgaris was stabilised. Before rainfall, the rain intensity was adjusted according to the experimental protocol. The sampling process commenced immediately with the onset of rainfall and continued at 3-min intervals until the rainfall ended, which lasted 40 min. Before sampling, the flow velocity within the velocity measurement zone (2 cm from the bottom and 2 cm from the edges of the erosion monitoring area) was measured using a dye tracer method. Measurements were taken on both sides of the velocity measurement zone, covering 50 cm. The average value represented the stream velocity for the sampling period. A thermometer was used to determine the temperature of the muddy water during rainfall, and a laser rain spectrometer was used to measure the size and terminal velocity of raindrops. A runoff bucket was set at the bottom of each sample plot to receive water and sediment samples. Four erosion collectors placed along the contour lines around the plot, with two in each direction, were flushed with 10 mL of water during sampling. After drying the sediment from the sheet and splash erosion, a small bucket was used to collect all the water and sediment from the designated time. The total volume of the muddy water was measured using a graduated cylinder. The collected water was then allowed to settle, and the clear water was separated and discarded. The remaining sediment was dried and weighed.
Similarly, water and sand samples were collected during rainfall to determine their physicochemical properties. The observed rainfall parameters include flow rate, sediment content, flow velocity, muddy water temperature, splash detachment (spattering), and raindrop size composition. The other research parameters were derived from calculations based on the observed data.
2.7 Estimating indicators
Based on relevant literature, hydraulic, rainfall, and basic erosion parameters were calculated(Dongdong et al.,2018;Feng et al.,2019). The regulatory effect was estimated by dividing the difference between the runoff separation and sediment transport rates with the bare soil and planting method by the rates with the bare soil, expressed as a percentage. Soil migration indicators were calculated using the sediment transport and erosion balance principles and raindrop runoff separation and sediment transport balance. The splash detachment rate was the total amount of the upward, downward, left, and right splash per unit time and area of the splash collection area (Eq. 1). The raindrop sediment transport rate was the difference in sediment separation between the upward and downward directions per unit time and area of the splash collection area in the upward and downward directions (Eq. 2). The runoff sediment transport rate (runoff sediment transport capacity) was equal to the difference between the erosion and the raindrop sediment transport rates (Eq. 3). The runoff separation rate was the difference between the erosion rate and the splash detachment capacity (Eq. 4). The formula for calculating the runoff sediment transport rate remained unchanged.
D1 = (Supwards + Sdownwards + Sleft + Sright) / (AS × t) (1)
T1 = (Sdownwards - Supwards) / (Ao × t) (2)
T2 = SE - T1 (3)
D2 = SE - D1 (4)
Where D1, T1, SE, T2, and D2 represent the splash detachment rate, raindrop sediment transport rate, erosion rate, runoff sediment transport rate, and runoff separation rate, respectively, with units in kg m− 2 s− 1. Supwards, Sdownwards, Sleft, and Sright represent the upward, downward, left, and right splash volumes, respectively, with units in kg. AS and AO represent the projected splash collection areas in the upward and downward directions, respectively, with units in m2. Finally, t represents the rainfall duration in s (seconds).
2.8 Data processing
The data analysis and charting software used were SPSS 19.0, EXCEL 2003, and Auto CAD 2014. The data analysis methods include correlation analysis, least significant difference (LSD) multiple comparisons, analysis of variance (ANOVA), regression analysis, time series analysis, and interval estimation F-test. The optimal multivariate equation was established by incorporating the independent factors with the dependent variables, and the multiple regression squared was determined using the contribution formula (Eq. 5).
Where Oi represents the contribution of the ith factor, R2 represents the coefficient of determination, bi represents the regression coefficient of the ith factor, σi represents the mean square deviation of the ith factor, and σy represents the mean square deviation of the dependent variable.