Effects of Land Use Change on Rainfall Erosion in Luojiang River Basin, China

Abstract

This paper, based on daily rainfall erosivity model, ArcGIS, trend analysis and Kriging interpolation method, analyzed the spatial and temporal distribution characteristics of rainfall erosivity in the Luojiang River Basin of China, and then explored the influence relationship between land use change types and rainfall erosivity potential. The results showed the following: (1) from 1980 to 2019, the distribution range of multi-annual rainfall erosivity in the Luojiang River Basin was 14,674–15,227 MJ·mm/ (hm²·h), with an average value of 14,102 MJ·mm/(hm²·h), showing an overall increasing trend; (2) the spatial distribution of rainfall erosivity value tends to be consistent with the multi-year average rainfall, showing a decreasing trend from the middle to the periphery of the basin; (3) land use change is an important factor affecting the spatial and temporal distribution characteristic of rainfall erosivity value in the basin. The increase in rainfall erosivity will undoubtedly increase the potential of soil erosion. This study can provide theoretical reference for future basin land use planning and put forward preventive suggestions according to the distribution characteristics of rainfall erosivity.

1. Introduction

Rainfall is the power source of regional soil and water loss. Rainfall erosivity reflects the potential ability of rainfall to soil erosion, and is the main dynamic factor causing soil erosion. It is also one of the main parameters of soil erosion, sediment yield and water environment modeling [1,2]. In 1958, Wischmeier [3] and Smith et al. [4] proposed the concept of rainfall erosivity, firstly, which was further developed by Hudson [5], Wischmeier and Smith [6,7,8] and used the product of the kinetic energy of subrainfall (E) and the maximum 30 min rainfall intensity (I30) as the measurement index [9]. However, due to the difficulty in obtaining the data required for calculation and the complexity and time-consuming nature of calculation steps [10,11], a simple algorithm for rainfall erosivity came into being [12]. Richardson et al. [13] established a rainfall erosivity calculation model based on daily rainfall scale, which is cited by many scholars. Yu et al. [4,14] introduced cosine function on the basis of daily rainfall model and analyzed rainfall in New South Wales, Australia and obtained good verification results. Duulatov et al. [8] predicted rainfall erosivity in Central Asia based on four global climate models. The research on rainfall erosivity in China began in the 1980s. Jia Zhijun [15], Wang Wanzhong [16], Huang Yanhe [17], Zhang Xiankui [18] and other scholars summarized the calculation model of rainfall erosivity suitable for different regions in China [11]. Zhang Wenbo et al. [19,20] studied and established a simple model for daily rainfall estimation of rainfall erosivity, which has been well-used in many drainage basins in southern China. Gao Ge et al. [21] discussed the temporal and spatial variation of rainfall erosivity in the past and future of the Yellow River Basin based on the multi-year daily precipitation and future medium discharge scenarios. Wang Yousheng et al. [22] introduced seasonal variation into the Richardson model, estimated rainfall erosivity of Huangshan City, Anhui Province, China, and tested the seasonal performance of the Richardson model.

The soil loss equation RUSLE (Revised Universal Soil Loss Equation) [23] model considers soil erosion as a function of climate erosion (which is affected by rainfall and intensity), topography, soil erodibility and vegetation cover [23,24], and studies the variation law of soil erosion from rainfall, soil properties, topography, land use and vegetation cover [25]. Because soil and topography have a large but relatively stable influence on water erosion, scholars attribute the increase in soil erosion to heavy rain, inappropriate land use and degraded vegetation [26,27]. Vegetation controls soil erosion through its canopy, root and litter components, but soil erosion also affects the composition, structure and growth pattern of plant communities [28,29]. However, due to their interaction, the relationship between precipitation, vegetation and erosion is complex and uncertain. Therefore, the positive or negative correlation between precipitation and erosion mainly depends on land use and vegetation [30].

Land use mainly affects soil erosion characteristics by influencing soil properties and vegetation coverage [7,31,32]. Land use changes include the quantity and quality of land resources and the change of land use structure over time, as well as the change in the spatial structure and the combination of land use types [33]. However, with the increasing intensity of human activities, unreasonable land use changes the regional topographic conditions, deteriorates soil characteristics and destroys vegetation resources, which aggravates soil erosion, limits the way and structure of land use, leads to land productivity degradation, and further intensifies the contradiction between human and land. A large number of studies have shown that the impact of rainfall on the characteristics of slope erosion is significantly different under different land use types [34]. Slopes with good vegetation coverage can effectively resist rainstorm erosion [35,36]. The degree of soil erosion in forest, shrub and dense grassland is lower than that in woodlands and sparse grasslands [37]. The greater the vegetation density, the more significant the interception effect of raindrops, preventing them from directly hitting the soil surface [38] and thus reducing the erosivity of rainfall. In addition, dense vegetation also slows down the formation of surface runoff, the root system of trees and shrubs increases soil porosity and organic matter content, thereby increasing the infiltration rates and reducing erosion. Sparse grassland and cultivated land are more likely to form runoff, so soil particles are directly exposed to the driving effect of surface runoff, which is prone to rainfall erosion [39,40].

At present, most scholars have studied the spatial and temporal characteristics of rainfall erosivity in different regions from different perspectives. However, due to the influence of vegetation coverage, the potential of rainfall erosivity under different land use conditions is also very different. The upper reaches of the Luojiang River Basin are a hilly region with multi-weathered sandy loam, loose soil, single land use type and serious soil and water loss. The middle and lower reaches are gentle slopes and plains, with red loam and yellow loam; the soil fertility is good. However, due to the serious influence of human activities, vegetation cover varies greatly, land use types are complex and changeable, and soil erosion occurs easily. In view of this, on the basis of previous studies, this paper still takes the Luojiang River Basin as an example to calculate the value and spatial–temporal variation characteristics of rainfall erosivity, and then it explores the influence relationship of rainfall erosion potential under land use change in the basin. The results not only can provide important reference for agricultural management in the basin, but they also have practical prospects for rational development and utilization of water and soil resources and soil erosion prevention.

2. Materials and Methods

2.1. Study Area

The Luojiang River is located in the low latitude area south of the North Return Line (110°10′ E~110°50′ E, 21°30′ N~22°30′ N), most of which are located in Huazhou City, Maoming City, Guangdong Province, China. It is the largest tributary of the Jianjiang River (Figure 1). Luojiang River originates from Wangjiang, Xinyi City, with a total length of 143 km, a basin area of 2055.067 km² and an average slope of 0.64‰. This region belongs to the south subtropical monsoon climate. The annual average temperature is 23.1 °C, and the rainfall is abundant. The annual average rainfall is 1890 mm, but the distribution of rainfall in space and time is extremely uneven. The rainfall in the middle of a year is mainly concentrated in May to September, accounting for 75% of the total rainfall of the year [41,42].

2.2. Data Source

The data required in this paper include land use data (2010, 2015 and 2018) and daily rainfall data of the Hejiang, Jialong and Baowei meteorological station from 1980 to 2019. The description, pretreatment and sources of each data are shown in

Table 1. Data description and Preprocessing.

Data Type	Data Description	Data Source	Data Processing
Land Use	Land use type maps of 2010, 2015 and 2018 with a resolution of 1 km × 1 km	Data Center for Resources and Environmental Sciences, Chinese Academy of Sciences http://www.resdc.cn (accessed on 30 March 2021)	Projection conversion, splicing, reclassification, calibration with actual data
Meteorological data	Daily precipitation at Jialong, Baowei and Hejiang stations from 1980 to 2019	Maoming Hydrology Sub-bureau of Guangdong Hydrology Bureau	Data extraction, missing data filling, correlation analysis

. Land use data of the three calculated years are shown in Figure 2.

2.3. Calculation Model of Rainfall Erosivity

The rainfall erosivity factor (R) is a basic factor in the soil loss equation, which is mainly estimated by establishing the power function structure or cosine function structure of rainfall and rainfall intensity. This paper adopts the simplified model of calculating rainfall erosivity by daily rainfall proposed by Zhang Wenbo et al. [19]. The model combines rainfall data in half a month period, uses the product of daily rainfall and daily rainfall intensity combination as the index to measure rainfall erosivity. In addition, the regional model parameters are used to modify the model, which can be well applied to the red soil region with abundant rainfall in southern China.

$$\mathrm{R}=\mathsf{\alpha }{\displaystyle \sum _{\mathrm{k}=1}^{\mathrm{n}}{\mathrm{P}}_{\mathrm{k}}^{\mathsf{\beta }}}$$

(1)

$$\mathsf{\alpha }=21.586{\mathsf{\beta }}^{-7.1891}$$

(2)

$$\mathsf{\beta }=0.8363+\frac{18.144}{{\mathrm{P}}_{\mathrm{d}12}}+\frac{24.455}{{\mathrm{P}}_{\mathrm{y}12}}$$

(3)

where R is rainfall erosivity within half a month, MJ·mm /(hm²·h); K is erosive rainfall days within half a month; ${\mathrm{P}}_{\mathrm{K}}$ is the daily rainfall, which greater than or equal to 12 mm on the kth day in half a month, mm; ${\mathrm{P}}_{\mathrm{d}12}$ refers to the average daily rainfall greater than or equal to 12 mm, the unit is mm; ${\mathrm{P}}_{\mathrm{y}12}$ is the annual average rainfall with daily rainfall greater than or equal to 12 mm, the unit is mm; $\mathsf{\alpha }$ and $\mathsf{\beta }$ are model parameters, the calculation formula is shown in Formulas (2) and (3).

2.4. Spatio-Temporal Analysis Method

2.4.1. Mann–Kendall Trend Analysis

Mann–Kendall (MK) is a nonparametric test method recommended by the World Meteorological Organization and has been widely used. This method judges the trend change of the sequence to be tested by the positive and negative statistics Z, which is suitable for the trend analysis of meteorological and hydrological sequences [43]. In this paper, MK trend analysis is used to analyze the change trend of multi-year rainfall erosivity in the basin from the perspective of time, and the confidence level of explicitness test is set to be 95 % (|Z| > 1.96).

2.4.2. Kriging Interpolation Method

Kriging estimation method [44] based on geostatistics is a commonly used spatial interpolation method. It can better fit the spatial pattern of rainfall erosivity by establishing the covariance function or variation function of variables and interpolating the estimated points according to the spatial autocorrelation of variables. In this paper, ArcGIS software is used to realize Kriging interpolation to analyze the spatial distribution of multi-year average rainfall erosivity in the basin.

2.5. Land Use Transfer Matrix

Land use state transition matrix is a quantitative description between system state and state transition in system analysis, which can describe the structural characteristics of regional land use change, reflect the direction of land use change, and better reveal the evolution process of land use pattern [45]. Its mathematical expression is as follows:

$${\mathrm{S}}_{\mathrm{i}\mathrm{j}}=\left[\begin{array}{cccc}{\mathrm{S}}_{11}& {\mathrm{S}}_{12}& \cdots & {\mathrm{S}}_{1\mathrm{n}}\\ {\mathrm{S}}_{12}& {\mathrm{S}}_{22}& \cdots & {\mathrm{S}}_{2\mathrm{n}}\\ \cdots & \cdots & \cdots & \cdots \\ {\mathrm{S}}_{\mathrm{n}1}& {\mathrm{S}}_{\mathrm{n}2}& \cdots & {\mathrm{S}}_{\mathrm{n}\mathrm{n}}\end{array}\right]$$

(4)

where S is the area, n is the number of land use types before and after the transfer, i and j are the land use types at the beginning and end of the research period, respectively, and S_ij is the area of land use conversion from land type i at the beginning of the research period to land type j at the end of the research period.

2.6. Sub-Basins Division

The division of sub-basins is an indispensable step in constructing hydrological models. The SWAT (Soil and Water Assessment Tool) model is a semi-distributed hydrological model constructed and developed by DILE et al. [46] based on ArcGIS software, which can be based on digital elevation model by setting the catchment area threshold, and the subwatershed in the simulated study area can be divided.

3. Results

3.1. Spatial and Temporal Distribution Characteristics of Rainfall Erosivity

Based on the daily observed rainfall data from 1980 to 2019 at meteorological stations in the Hejiang, Jialong and Baowei, the rainfall erosivity of each station is calculated by using Equations (1)–(3). The monthly rainfall erosivity, annual rainfall erosivity and annual average rainfall erosivity can be obtained through summary

3.1.1. Temporal Variation Characteristics of Rainfall Erosivity

The trend analysis of rainfall erosivity in the Luojiang Basin from 1980 to 2019 was conducted by MK test method, as shown in Figure 3.

It can be seen that from 1980 to 2019, the temporal pattern and trend of annual rainfall erosivity and annual rainfall basically tend to be consistent, showing an obvious trend of increase, with a growth rate of 29.048 MJ·mm/ (hm²·h·a). After the 1990s, the rainfall erosivity value showed an obvious upward trend, and the inter-annual fluctuation amplitude was significantly different. The maximum annual rainfall erosivity was 24,429.774 MJ·mm/(hm²·h) in 1993, and the minimum was 7171.761 MJ·mm/(hm²·h) in 1998, with a difference of 3.4 times.

3.1.2. Spatial Distribution Characteristics of Rainfall Erosivity

The ArcGIS software was used to interpolate the whole basin by Kriging interpolation method combined with the calculation results of rainfall erosivity at each station, and the spatial distribution map of rainfall erosivity in the Luojiang River Basin was obtained (Figure 4a). The rainfall erosivity of the Luojiang Basin ranges from 14,674 to 15,227 MJ·mm/(hm²·h), with an average of 14,102.67 MJ·mm/(hm²·h), and a standard deviation of 4325.206 MJ·mm/(hm²·h), indicating a large spatial difference. The rainfall erosivity showed a decreasing trend from the middle to the surrounding areas. The rainfall erosivity was the largest in the middle and most parts of the north-central basin, with the average annual rainfall erosivity exceeding 15,146 MJ·mm/(hm²·h). The rainfall erosivity value is the smallest in most areas of southwest of the basin, and the average rainfall erosivity is concentrated in 14,674–15,002 MJ·mm/(hm²·h). It can be seen from Figure 4b that the value of multi-year average rainfall erosivity is relatively consistent with the spatial pattern of multi-year average rainfall. The linear regression method is adopted to establish the correlation between the two (Figure 5); the results show that the correlation coefficient R of the multi-year average rainfall erosivity and the multi-year average rainfall is 0.974, and they are extremely significantly correlated.

3.2. Analysis of Land Use Change

IDRISI 17.0 Selva and ArcGIS 10.2 software developed by relevant departments in the United States were used, the land use data of the basin in 2010, 2015 and 2018 were statistically analyzed, such as

Table 2. Statistical table of land-use types in different years.

Land Type	2010		2015		2018
	Area/km2	Percentage/%	Area/km2	Percentage/%	Area/km2	Percentage/%
Arable land	272.195	13.40	272.212	13.40	278.821	13.77
Garden land	1643.112	80.88	1627.748	80.12	1622.603	80.10
Forest land	53.965	2.66	62.342	3.07	53.144	2.62
Grassland	23.213	1.14	23.213	1.14	17.069	0.84
Commercial land	39.126	1.93	46.097	2.27	55.060	2.72

; at the same time, the land use area transfer matrix from 2010 to 2015 and from 2015 to 2018 was produced, as shown in

Table 3. 2010–2015 Land use transfer matrix.

Land Transfer		2015
		Grassland	Arable Land	Forest Land	Commercial Land	Garden Land	Sum /km2
2010	Grassland	23.213					23.213
	Arable land		272.186	0.001		0.008	272.195
	Forest land			53.965			53.965
	Commercial land				39.126		39.126
	Garden land		0.026	8.375	6.971	1627.74	1643.112
	Sum/km2	23.213	272.212	62.342	46.097	1627.748	2031.611

and

Table 4. 2015–2018 Land use transfer matrix.

Land Transfer		2018
		Grassland	Arable Land	Forest Land	Commercial Land	Garden Land	Sum /km2
2015	Grassland	2.97	4.372			15.857	23.199
	Arable land	3.474	58.42	5.341	15.495	189.129	271.859
	Forest land	0.306	6.19	13.397	1.092	40.508	61.493
	Commercial land	1.17	14.691	0.094	3.918	26.091	45.964
	Garden land	9.149	195.148	34.312	34.555	1349.44	1622.603
	Sum/km2	17.069	278.821	53.144	55.060	1621.025	2025.119

. It can be seen from Table 2 that in the three calculation years, the land use types are mainly garden land, accounting for more than 80% of the total basin area, while the grassland use type is relatively small, accounting for less than 1.5% of the total basin area. The order of land use types by area is Garden land > Cultivated land > Forest land > Commercial land > Grassland. As can be seen from Table 3, the grassland area remained unchanged and the garden land area decreased slightly in 2010 to 2015, mainly converted to forest land, commercial land and arable land, with forest land being the largest. As can be seen from Table 4, more complex two-way conversions occurred in all land use types from 2015 to 2018, with the area of grassland, forest land and garden land decreased by 6.144 km², 9.198 km² and 5.145 km², respectively; the reduced area is mainly converted into cultivated land and commercial land.

3.3. Impact of Land Use Change on Rainfall Erosivity

The SWAT (Soil and Water Assessment Tool) plug-in in ArcGIS software was used to divide the basin into 27 sub-basins based on the digital elevation model. Combined with the rainfall erosivity model, the rainfall erosivity values of each sub-basin under land use scenarios in 2010, 2015 and 2018 were calculated successively (Figure 6).

The comparison shows that under the land use scenario in 2018, the rainfall erosion potential of the whole basin is minimal, with the rainfall erosivity value of each sub-basin is between 11,773–12,973 MJ·mm/(hm²·h). Under the land use scenario in 2010, the rainfall erosivity of the whole basin was relatively the largest, with the rainfall erosivity value of each sub-basin was between 17,889–18,209 MJ·mm/(hm²·h). Under the three land use conditions, the rainfall erosivity potential of each sub-basin in the middle and south is greater than that in the north of the basin, which is mainly because: the cultivated land use types in the north and northeast of the basin occupy more areas, while the grassland is mainly concentrated in the middle and south of the basin. Grassland can effectively reduce the direct effect of raindrops on soil particles, and has a stronger retention effect on surface runoff, which reduces the fluidity of soil particles. Moreover, according to the actual cultivated land types in the basin, most of them are low vegetation cover crops such as rice, peanut and vegetables; their interception effect on raindrops is not as good as that of grassland, so the potential of forming rainfall erosivity is greater.

Based on the distribution maps of rainfall erosivity in each year, the distribution maps of rainfall erosivity change rate from 2010 to 2015 and from 2015 to 2018 can be calculated (Figure 7). Compared with the land use scenario in 2010, the potential of rainfall erosion in the basin showed a decreasing trend in 2015, but the decrease was not significant, and the change rate was between 0.031 and 0.034. Among them, the decline of No.4, No.11, No.12 in the northeast and No.5, No.13 in the west of the basin is relatively obvious, with a change rate more than 0.032. Compared with the land use scenario in 2015, the potential of rainfall erosion in each sub-basin also showed a downward trend in 2018, and the change rate increased from the northern part of the basin to the surrounding areas, with a relatively large change range between 0.091 and 0.176. Among them, the southern and southwestern sub-basins of No. 17, No. 21, No. 26, have the most significant decline, with a change rate of more than 0.150.

The sub-basins No. 3, No. 17, No. 21 and No. 26, with the most significant rainfall erosion rate during 2010–2015 and 2015–2018, were selected as the main objects for land use change analysis (

Table 5. Proportion of land use type area in sub-basins.

Proportion of Land Use Type (%)	Sub-Basin No. 13					Sub-Basin No. 17
	Arable Land	Garden Land	Forest Land	Grassland	Commercial Land	Arable Land	Garden Land	Forest Land	Grassland	Commercial Land
2010	0.29	3.69	0.01	0.05	0.00	0.67	4.03	0.43	0.23	0.00
2015	0.29	3.59	0.09	0.05	0.03	0.67	3.90	0.51	0.23	0.05
2018	0.51	3.33	0.16	0.00	0.05	0.89	3.94	0.20	0.10	0.19
2010—2015	0.00	−0.11	0.08	0.00	0.03	0.00	−0.13	0.08	0.00	0.05
2015—2018	0.22	−0.25	0.07	−0.05	0.02	0.22	0.03	−0.31	−0.13	0.14
Proportion of Land Use Type (%)	Sub-Basin No. 21					Sub-Basin No. 26
	Arable Land	Garden Land	Forest Land	Grassland	Commercial Land	Arable Land	Garden Land	Forest Land	Grassland	Commercial Land
2010	0.52	1.33	0.15	0.05	0.00	0.36	1.07	0.00	0.00	0.08
2015	0.52	1.33	0.15	0.05	0.00	0.36	1.07	0.00	0.00	0.08
2018	0.41	1.66	0.10	0.00	0.01	0.35	0.99	0.00	0.00	0.16
2010—2015	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00
2015—2018	−0.11	0.33	−0.06	−0.05	0.01	−0.01	−0.08	0.00	0.00	0.08

). According to the comparison of land use transfer matrix in the basin, compared with the land use conditions in 2010, the area of garden land in the basin decreased in 2015, while the area of forest land and commercial land increased, which led to a downward trend of rainfall erosivity in 2015. In 2018, compared with 2015, the area of grassland, forest and garden land decreased and was mainly converted to commercial land, followed by cultivated land. Although the increase in cultivated land area will promote the increase of rainfall erosivity to a certain extent, the large increase in commercial land area is the main reason for the decrease in rainfall erosivity potential under the land use scenario in 2018.

Through the analysis of sub-basins No. 13, No. 17, No. 21, No. 26, it is found that the conversion of garden land to commercial land makes the area of regional hardening increase greatly. Although the increase in commercial land area easily produces runoff and form surface runoff, the hardening layer protects the soil surface and avoids the direct exposure of soil particles to splash erosion of raindrops, thereby reducing the likelihood of rainfall erosion. The conversion of garden land to woodland land increases the vegetation coverage in the watershed, and the interception effect on rainfall is more significant, which weakens the kinetic energy of rainfall, thus reducing the erosion potential of rainfall and reducing the possibility of soil erosion. The conversion of grassland, forest land and garden land into cultivated land will reduce the vegetation coverage of the watershed, enhance the splash erosion of raindrops on soil, and increase the turbulence of surface runoff and rainfall erosion force. On the other hand, under the type of cultivated land use, the underground root amount and root distribution in the basin will also be reduced accordingly, the soil structure is relatively single, the porosity becomes smaller, and the infiltration performance is weakened. When encountering heavy rainfall, it is easy to form super-infiltrating runoff, resulting in serious soil erosion and water loss of cultivated land, and the potential of rainfall erosion also increases correspondingly.

Although the conversion area of the above land use types is small, it provides reference for the optimal allocation of land resources and the reduction of soil loss caused by land use conversion in the basin.

4. Discussion

4.1. Effects of Land Use Change on Rainfall Erosivity

The process of soil erosion includes the comprehensive effect of rainfall, vegetation, soil and human factors, among which rainfall and vegetation are the two main factors affecting the degree of soil erosion, and other factors can be reflected in vegetation, while rainfall erosivity is the main index reflecting erosive rainfall. Based on the meteorological data from 1980 to 2019 and the land use data of 2010, 2015 and 2018 in the Luojiang River Basin, combined with the rainfall erosivity model, this paper compared the response of rainfall erosivity in each sub-basin under different land use change conditions, and concluded that the main reason for the decreasing trend of rainfall erosion potential in 2010, 2015 and 2018 is the decrease of rainfall, but land use change is also an important influencing factor. As the annual rainfall decreases, the rainstorm intensity and surface runoff decrease correspondingly, so the value of rainfall erosivity decreases and the potential of rainfall erosion decreases. On the one hand, land use change can change the watershed vegetation coverage; on the other hand, it can affect the rate and magnitude of runoff production and confluence over the underlying surface, and then affect the magnitude of rainfall erosion potential.

However, this study only explores the relationship between land use change and rainfall erosivity in the Luojiang River Basin in general, and then it provides a theoretical basis for the management of soil erosion. The difference of rainfall erosivity caused by the change of vegetation type, vegetation coverage and farming pattern is not much involved, and more attention should be paid to this aspect in the future. In addition, it should be noted that the spatial distribution characteristics of rainfall and rainfall erosivity values in the paper were obtained by Kriging interpolation method, which may deviate from the measured values in different regions.

4.2. Recommendations

The potential of rainfall erosion is different with different land use patterns. The increase in rainfall erosion potential will undoubtedly lead to soil erosion and other problems. Only by optimizing the land use structure, selecting a reasonable way of land use, efficiently utilizing limited land resources and controlling the potential of rainfall erosion can we realize the sustainable development of agriculture. According to the research experience of land use and rainfall erosion at home and abroad, combined with the characteristics of land use change and the change trend of rainfall erosivity in the basin, it is suggested that the watershed should develop crops with high vegetation coverage in the central and northern regions where the potential of rainfall erosion is relatively large, and establish a multi-level rainfall erosion prediction model, carry out the quantitative evaluation of the relationship between land use dynamic change and rainfall erosion, and construct a distributed land use and rainfall erosion monitoring network based on modern space technology.

5. Conclusions

This paper is based on the simple model of daily rainfall erosivity, combined with Mann–Kendall trend analysis and Kriging interpolation method to analyze the temporal and spatial distribution characteristics of rainfall erosivity in the Luojiang Basin. On this basis, by comparing the response relationships of rainfall erosivity values of each sub-basin under different land use scenarios in 2010, 2015 and 2018, the following conclusions are drawn:

(1)

From 1980 to 2019, the distribution range of annual rainfall erosivity in the Luojiang River Basin was 14,674–15,227 MJ·mm/(hm²·h), and the time series showed an obvious increasing trend, with the growth rate of 29.048 MJ·mm/(hm²·h·a), which was significantly correlated with annual rainfall (correlation coefficient R = 0.974).

(2)

The average value of rainfall erosivity is 14,102 MJ·mm/(hm²·h), and the standard deviation is 4325.206 MJ·mm/(hm²·h). The spatial difference is large, and the overall trend is decreasing from the middle to the surroundings of the basin.

(3)

Land use change is an important factor affecting the spatial and temporal distribution characteristics of rainfall erosivity in the basin. It is suggested to improve the vegetation coverage in the middle and north of the basin and establish a multi-level rainfall erosion prediction model.