The technology of slope vegetation system stability enhancement is an important part of the comprehensive ecological security improvement in small watersheds area of the Loess Plateau. The results of the comprehensive soil erosion improvement in the sub-basin are used to give an evaluation of the effectiveness of the gully slope vegetation restoration project. Soil erosion quantification distribution in the Sheep Sap Gully sub-basin of the Loess Plateau hilly gully area was simulated and explored by combining the modified universal soil loss equation RUSLE model with GIS and RS spatial information technology. The quantitative values of LS factor were extracted using DEM data, the R factor of rainfall erosion force was calculated using meteorological monitoring station data around this region, the K factor of erosion resistance was obtained based on the soil survey database, and the C factor and P factor of soil and water conservation measures were obtained by combining MODIS image data and previous research experience. The study concluded that (1) The erosion area ratio within the study area is 36.33%; (2) The quantitative grading standard of the pattern, the size of the erosion distribution area at all levels is ranked as light>moderate>very strong>strong>intense; (3) Under the conditions of land use and vegetation cover, strong erosion is mostly found in farming areas with sparse vegetation, while weak erosion is found in areas with lush vegetation such as forests and grasslands; (4) In terms of spatial distribution, erosion is greater in the south-western part of the basin than in the north-eastern part, and there is also strong erosion in the south-western part. The results of the study provide a reference for research into integration and synthesis of ecological security technologies for gully and slope management projects. The research content provides the basis and support for watershed governance and soil and water resource management and conservation.

  • Quantification of soil erosion on the Loess Plateau for the regulation of hydrological and water resources.

  • Quantitative means of each model's key factors.

  • RUSLE model is a new technology that combines RS and GIS.

  • Data support of quantitative results for integrated soil and water management in small watersheds.

  • For soil and water conservation research support in the Loess Plateau region.

Graphical Abstract

Graphical Abstract
Graphical Abstract

One of the modern measures to combat soil erosion, one of the biggest ecological problems in the world, is the implementation of reforestation and grass restoration projects (Yue et al. 2015). By changing the pattern of the ecosystem in the area, the project has a positive effect on the restoration of vegetation, improving soil erosion and controlling soil erosion in the area. Since the project was implemented in August 1999, significant results have been achieved. In the 21st century, soil erosion research in China is facing new opportunities and challenges (Gong et al. 2022). Therefore, how to accurately reveal the principles, processes and laws of soil erosion, investigate the ways in which natural and human factors act on soil erosion, establish soil erosion models, make objective and systematic evaluations of the environmental benefits of soil erosion on a regional and global (Liu et al. 2021), propose targeted strategic solutions and technical approaches to prevent and control soil erosion in conjunction with the use of water and soil resources, and create a benign ecological environment are of far-reaching significance to achieve national ecological security and healthy economic development, and to promote scientific and technological research and development in the field of soil and water conservation (Wang et al. 2017).

In this paper, with the help of the RUSLE model, the factors and calculation results and raster data results related to soil erosion such as R rainfall erosion force, K soil erosion resistance, LS terrain factor, C vegetation cover factor were obtained respectively, and the calculation of soil erosion A was constructed with the help of ArcGIS10.2 platform using Model Builder modelling work The model is used to discuss the use of the RUSLE model and the application of GIS Geographical statistics for this study, which gives a reference for the research on soil erosion in small watersheds of Loess Plateau and has far-reaching significance to promote soil and water conservation disciplines (Wang et al. 2022).

In the first section of the article, the background of the study and the technical methods of the study and the significance of the study are introduced; the second section introduces the study area and the basic data of the study, and lists the calculation methods and formulas one by one; the third section discusses and analyzes the quantitative results of each factor, and estimates the amount of erosion in the study area and the overall analysis; and the final section summarizes the research methods and the results of the study in the whole article.

Study area scope

The areas studied in this paper are in the Sheep Circle Gully sub-basin, which is part of the central area of the Loess Plateau region, located from 109 °31′∼109 °71′ and 36 °42′∼36 °82′ north of Liqu Town, Baota District, Yan'an City, northern Shaanxi Province. This small watershed is a secondary tributary of the Yan River Basin, a first-order tributary of the Yellow River, and is a typical model of small watershed synthesis with a watershed area of 2.02 km2 (Figure 1).
Figure 1

Study area location diagram.

Figure 1

Study area location diagram.

Close modal

The study area is in a geomorphic gully interspersed with loess beams and loess ditches, with a density of 2.74 kg/km2 and an altitude of about 1,050–1,295 m. It has a semi-arid continental monsoon climate and an average annual precipitation of 535 mm, and the rainy season is mostly concentrated in July-September each year, accounting for about 79% of the annual precipitation, with a large interannual variability of precipitation. The annual sunshine hours are 2,528.8 h, the annual average temperature is 9.40 °C, and the annual difference in temperature is 29.40 °C (Zhao et al. 2016). The soils are mainly yellow cotton soil with loose structure, poor erosion resistance, and severe soil erosion (Azadeh et al. 2017). The vegetation of the watershed belongs to the forest-steppe transition zone in terms of zoning. The natural vegetation in the watershed has been destroyed due to the disturbance of human activities, and nowadays it is mostly planted with artificial secondary vegetation, and the artificial species of plant species mainly include acacia, oleander, sea buckthorn, lemon grass, alfalfa, etc. (Zhang et al. 2015). The average multi-year rainfall erosion force was 1,755.83 (MJ·mm)/(hm2·h) before reforestation (before 1999) and 2,155.71(MJ·mm)/(hm2·h) after reforestation. At present, the land types in the watershed are mainly sloping farmland, terraced farmland, forest, scrub, grassland, residential land, and water.

Experimental data

Soil erosion conditions in small watersheds were simulated with the modified universal soil loss equation (RUSLE), involving arithmetic parameters including precipitation, soil, topography, vegetation, soil and water conservation measures and other factors in the study target area. Daily rainfall data from 1997–2016 from the Yan'an Climate Observatory from the China Meteorological Data Network (http://data.cma.cn/) were used in this study. Soil texture data were provided through the Chinese soil texture data base from the Resource and Environment Science Data Center of the Chinese Academy of Sciences (http://www.resdc.cn), which has small variability due to the characteristics of loess and is based on the current year in terms of temporal resolution. Topographic data were generated using DEM data with 5 m spatial resolution by using interpolation of actual elevation points in the watershed. Vegetation cover data were obtained from MODIS vegetation index data through the Geospatial Data Cloud (http://www.gscloud.cn/). The vegetation cover and soil conservation data were taken depending on the empirical values of the researchers.

Methodology

In this study, the study area was quantitatively calculated using the RUSLE model, which was modified on the USLE model to establish a model for calculating soil erosion with a broader range of applicability. The calculation model consists of five input parameters: rainfall, soil erodibility, slope break length, vegetation coverage, and anthropogenic measures, and its quantitative calculation equation is:
(1)
where each parameter is: A denotes the average annual soil erosion in the study area; R denotes the rainfall erosivity factor; K denotes the soil erodibility factor; LS is the dimensionless slope length-slope factor; and where L is the slope length factor, defined as a power function of slope length, S is the slope factor, so LS represents the ratio of soil loss on a slope for a given slope length and slope to soil loss on a typical slope in a standard runoff plot, all else being equal; C represents the dimensionless vegetation cover, which refers to the ratio of soil loss under a particular crop or vegetation cover, all other factors being equal, to that of continuous recreational land after cultivation; P refers to the dimensionless ratio of soil loss after soil and water conservation measures to soil loss from downhill planting.

R factor

This indicator of the ability of rainfall to strip and transport soil and is usually estimated as a model input parameter in terms of rainfall. The rainfall erosion force is calculated as
(2)
and according to the formula for calculating the average annual rainfall in the study by Zhang Wenbo (Zhang et al. 2014) et al.
(3)

In Equation (2), F is the average annual rainfall (mm) and R the average annual rainfall erosion force; α and β are the model inputs; and in Equation (3) P is the total annual rainfall (mm) and ri is the monthly rainfall (mm) in that year. Based on the results of literature studies, the values of parameters α and β in the rainfall model of this study area were taken as α = 0.184 and β = 1.96, respectively, and the average rainfall erosion force of the past 20 years in the Sheep Shed Gully sub-basin was derived from the daily rainfall data and equations from 1997 to 2016.

K factor

The calculation of soil erodibility factor K is based on the participation of soil particle composition and organic matter parameters in soil texture properties, including content of sand, clay grain, chalk and organic matter. The factor K in the RUSLE model is a value obtained through experiments. It is quantitative and the K-factor is usually obtained by direct measurement of soil loss per unit of precipitation erosion force in standard natural plots, but it is difficult to deploy natural plots in this watershed on a large scale, and in this paper we use Williams' calculation of K-value in the EPIC model (Williams 1990).

(4)
where: 0.1317 is the conversion factor between American and international units; SAN (0.1∼2 mm) is the sand content (%); SIL (0.002∼0.1 mm) is the powder content (%); CLA (<0.002 mm) is the clay content (%); and C is the organic carbon content (%). Using the raster computer in ArcGIS10.2 for the above parameters combined with Equation (4), the results were calculated between 0.022 and 0.040 t·hm2·h·MJ−1·mm−1·hm−2, which is close to the results of Li Kui (Li 2014) and others, with a K value of 0.047 t·hm2·h·MJ−1·mm−1·hm−2 for yellow cotton soil.

LS factor

Slope and slope length, as a basic geomorphological indicator, affect the surface soil structure, surface runoff and soil erosion by influencing gravity, which in turn affects the deployment of soil and water conservation measures, the layout of agriculture, forestry and livestock production, and land resource development (Chen et al. 2018). From this paper, the LS factor proposed by Wischmeier and Smith was calculated using the formula.
(5)
(6)
where λ is the horizontal slope length, α is the slope length index, 22.13 is the slope length of the standard plot(m), and θ is the slope extracted using DEM data. S-factor is calculated using the formula of McCool, and the results are more satisfactory.
(7)
and using the steep slope formula of Liu Baoyuan.
(8)

C-factor

The vegetation cover and management C-factor is highly sensitive parameters to soil erosion changes, and it has this good correlation with vegetation cover (Kong 2013), which is calculated from the NDVI value calculated from remote sensing images as a dimensionless parameter. In this study, vegetation cover was obtained under the ENVI platform with the help of MYD13A1 synthetic vegetation index data from MODIS. Using the Tsai-Chung method to calculate C (Cai et al. 2000) as follows.
(9)
where VFC is the formula shows that the vegetation cover is greater than 0.783, the erosion of the surface is extremely weak and the erosion is negligible. And when the vegetation cover is less than 0.1%, its erosion reduction effect is basically not reflected. The distribution of its C value ranges from 0 to 1 (no vegetation cover area such as bare rock and bare soil).

P-factor

The P-factor is related to the type of land use in the subsurface, and Wischmier and Smith (Wischmeier 1978) showed that the P-value is between 0 and 1 and is a dimensionless number (Li et al. 2013). The soil conservation measure factor P is the most difficult parameter to obtain in the model. It can be measured by runoff plot tests, but this method is costly and time-consuming, and the measured results have regional limitations. In this study, we combined the research results of Deyan Zhong and Baoyuan Liu (Liu et al. 1999) on the Sheep Circle Ditch sub-basin with Landsat8 data to superimpose the results of land use classification decoding to get the P value, and other contained classifications as dams, roads, industrial and mining land, and oil fields.

Factor calculation results

K-factor calculation

The obtained soil texture data were assigned and rasterized separately to obtain the range values of soil sand, powder, clay, and organic matter (Qin 2017), where the soil sand was 0–12.6%, powder was 0–68.6%, clay was 0–18.7%, and organic carbon was 0.66–12.2%, (Figure 2). Then, combined with Equation (4), the K-factor's calculation results, which takes values between 0.022 and 0.04 (Figure 3).
Figure 2

Soil mass content distribution. (a) Quantification value of clay particles. (b) Quantification value of sand particles. (c) Quantification value of powder particles. (d) Quantification of organic carbon values.

Figure 2

Soil mass content distribution. (a) Quantification value of clay particles. (b) Quantification value of sand particles. (c) Quantification value of powder particles. (d) Quantification of organic carbon values.

Close modal
Figure 3

K factor content of soil erodibility.

Figure 3

K factor content of soil erodibility.

Close modal

R factor calculation

The Loess Plateau area's erosion type is water erosion, and rainfall is positively correlated with it, i.e., the greater the rainfall, the greater the intensity of soil erosion (Xie et al. 2001). In this study, the average annual rainfall data of the region for the past 10 years were obtained, and the rainfall from 2006 to 2015 was obtained after the removal of abnormal values (Figure 4). The total annual rainfall during this 10-year period was basically stable with small fluctuations, with a maximum peak of 959.1 mm in 2013, and the results were close to the R value of 2,155.71 (MJ·mm·hm-2·h−1·a−1) after reforestation according to Equations (2) and (3), (Figure 5).
Figure 4

Rainfall change from 2006 to 2015.

Figure 4

Rainfall change from 2006 to 2015.

Close modal
Figure 5

Rainfall erosivity R factor content.

Figure 5

Rainfall erosivity R factor content.

Close modal

LS factor calculation

For small watersheds and regions, to obtain the slope length L factor usually DEM is used to extract the watershed slope S and the D8 algorithm in the hydrological analysis module to obtain the flow direction, and after repeated filling of the DEM depressions to obtain the depression-free DEM, the sink accumulation is finally calculated (Qi 2017). The product of the sink accumulation and the raster image element is used as the valuation of λ, the quantified values of LS were calculated in Raster Calculator using Equations (5), (6) and (8), respectively (Figures 6 and 7).
Figure 6

S factor quantized value.

Figure 6

S factor quantized value.

Close modal
Figure 7

L factor quantized value.

Figure 7

L factor quantized value.

Close modal

C-factor calculation

This factor is a model parameter related to vegetation cover, and the calculation combines the land use data with the MODIS16 synthetic vegetation normalized index NDVI raster data for overlaying, and the quantitative values of C factor under different vegetation types are calculated by Equation (9) (Table 1 and Figure 8). The value is inversely proportional to the amount of soil erosion, i.e., the smaller the C-value, the lighter the soil erosion condition.
Table 1

Quantized value of factor C and P under different land use classification

Land use typeGrasslandForestCroplandOrchardTerraceswaterbare groundOtherResidential
0.23 0.03 0.35 0.05 0.31 1.0 0.9 
0.9 0.7 0.8 0.35 
Land use typeGrasslandForestCroplandOrchardTerraceswaterbare groundOtherResidential
0.23 0.03 0.35 0.05 0.31 1.0 0.9 
0.9 0.7 0.8 0.35 
Figure 8

C factor quantized value.

Figure 8

C factor quantized value.

Close modal

P-factor calculation

The P-factor in the general soil erosion equation has been a difficult problem to confirm, usually obtained by runoff plot test, and its value is from 0 to 1. The smaller the value, the lower the degree of soil erosion, representing perfect soil and water conservation measures deployed in the area, and vice versa, the higher the degree of soil erosion. This factor calculation is like the C-factor, with GIS platforms used to obtain from land use data, and the following P-factor quantification values are obtained according to Liu Baoyuan's research results (Table 1), and the raster quantification data are obtained by overlaying the land use data (Figure 9).
Figure 9

P factor quantized value.

Figure 9

P factor quantized value.

Close modal

Calculation results and analysis

Based on the raster data of each factor, the raster data were processed in the ARCGIS platform to construct a unified coordinate reference, and the raster image elements were converted into 5*5 size GRID format, and the Raster Calculator was used to bring in the linear calculation formula (1) for soil erosion to obtain the values of soil erosion and spatial distribution of the Sheep-Ring Ditch sub-basin, and the soil erosion was classified symbolically Based on soil erosion classification standards (SL196–2007). (SL196-2007) for symbolic classification (Table 2; Figure 10).
Table 2

Soil erosion grade classification table

Soil erosion amountTinyMildMediumStrongintenseSevere
Area (km271.75 125.75 43.86 18.74 6.26 9.07 0.59 
Proportion (%) 36.33 63.67 22.21 9.49 3.17 4.59 0.3 
Soil erosion amountTinyMildMediumStrongintenseSevere
Area (km271.75 125.75 43.86 18.74 6.26 9.07 0.59 
Proportion (%) 36.33 63.67 22.21 9.49 3.17 4.59 0.3 
Figure 10

P Soil erosion intensity classification mapping.

Figure 10

P Soil erosion intensity classification mapping.

Close modal

The calculated soil erosion in this study area is 3,175.16 t/(km2·a), which is close to 3,270.19 t/(km2·a) as calculated by Zhao Wenqi (Yi et al. 2015) for the 2014 Sheep Sap Ditch sub-basin, concluding that the Loess Plateau area has experienced many years of reforestation and comprehensive small watershed management projects have contributed to soil and water conservation, which is a non-intense erosion area. From Table 2, From this, it is clear that the soil erosion area of the whole watershed is 71.75 km2, accounting for 36.33% of the total area of the study area. The area and proportion of different levels of soil erosion are: 27.25 km2 for slight erosion, accounting for 63.67% of the area; 9.27 km2 for light erosion, area share of 22.21%; 19.53 km2 for moderate erosion, accounting for 9.49% of the area; 75.65 km2 for strong erosion, accounting for 3.17% of the total area of the study area; and 3.17% for very strong erosion. 3.17%; very strong erosion area 4.59 km2, accounting for 1.16% of this; severe erosion's area is2.71 km2, accounting for 0.3% of the total area of the study area. Slight and mild erosion is the mainly soil erosion, accounting for 85.88% of the occurring soil erosion area. The largest area of slight erosion and the smallest area of severe erosion, and the overall distribution is contiguous. The soil erosion intensity of each level is arranged in the order of area: mild > medium > intense > strong > severe.

  • (1)

    In the model calculation, the perennial rainfall in the Loess Plateau area is basically stable, and the changes of soil texture, which is mainly yellow sheep soil, are not significant, resulting in constant rainfall erosion force R and soil erodibility factor K. Only the topographic factor LS and vegetation index-related C and P factors become strongly correlated to the soil erosion amount. The steeper the slope and the longer the slope length, both will make the soil erosion intensity become larger and the erosion volume increase. Therefore, we can consider reducing the slope by cutting a steep slope into a gentle slope when carrying out comprehensive watershed management in land engineering, and truncating the length of slope gully management to reduce the runoff pooling time.

  • (2)

    From the perspective of vegetation cover, the planting of protective forests should be vigorously carried out within the red line of guaranteed arable land, and plants suitable for the survival of the Loess Plateau should be planted in the gully and slope to improve the capability of soil and water conservation and increase the area covered by various types of vegetation to decrease the occurrence of soil erosion.

  • (3)

    In this study, remote sensing data and spatial GIS data were innovatively used to quantify the factors for soil erosion calculation, in which the P-factor was quantified using the integrated management of small watersheds and the current land use status, so that the P-value could be more accurately derived in comparison with previous research results.

  • (4)

    The study has shown that the areas with higher soil erosion intensity have been mainly concentrated in the south-west channel, including slope land, steps of ditches, gully slope, and construction land in the north-east. Meanwhile, the soil erosion intensity on the arable land, channels with good vegetation, and gully bottom with gentle terrain have been lower. The soil erosion modulus of 3,175.16 t/(km2·a) (2018) calculated in this study was reduced from 3,270.19 t/(km2·a) derived from the study of Zhao et al. (2016). This also indicates that the adoption of integrated ecological management projects in the Sheep Sap Ditch sub-basin of the Loess Plateau has produced some improvement in local soil erosion. The ecological restoration policy and afforestation projects in recent years have also transformed the local land use from arable land to woodland and grassland, which has played a certain role in preventing soil erosion.

GIS and RS were used to obtain quantitative data of each parameter of the modified soil erosion equation (RUSLE), and the quantitative calculation of the soil erosion evaluation was carried out in the Sheep Sap Ditch sub-basin of Loess Plateau by combining the modeling method, and the influence of the differentiation of different factors on the change of soil erosion intensity in the sub-basin was analyzed. The soil erosion area is 71.75 km2, accounting for 36.33% of the total area of the study area. The study area has the largest area of slight erosion and the smallest area of intense erosion, and the soil erosion intensity of each class in the order of area is light>moderate>strong>very strong>intense. In terms of spatial distribution, the watershed is larger to the south-west than to the north-east and is accompanied by highly erosive conditions, so appropriate protection measures and comprehensive watershed management are needed to stabilize the ecological environment. The erosion covers the whole small watershed, and the soil erosion varies significantly under different land use types and vegetation cover, mostly occurring in arable land, terraces, and orchards, where the erosion condition is more severe; woodland, grassland, and shrubs are mainly slightly and lightly eroded. Another influencing factor of PTU is that the soil erosion intensity also varies greatly under different slope topography, where the erosion area accounts for the largest 15%∼25% slope zone, and the erosion intensity is mainly slight and mild.

This research was fund by Technology Innovation Center for Land Engineering and Human Settlements, Shaanxi Land Engineering Construction Group Co., Ltd and Xi'an Jiaotong University (2021WHZ0088) and (2021WHZ0091). and Scientific Research Item of Shaanxi Provincial Land Engineering Construction Group (DJNY2022-33).

All relevant data are included in the paper or its Supplementary Information.

Azadeh
K.
,
Abdulvahed
K. D.
&
Seyed
J. A.
2017
Changeability of soil erosion variables in small field plots from different rainfall durations with constant intensity
.
Journal of African Earth Sciences
129
,
751
758
.
Cai
C.
,
Ding
S.
,
Shi
Z.
,
Huang
L.
&
Zhang
G.
2000
Application of USLE model and GIS IDRISI to predict soil erosion in small watersheds
.
Journal of Soil and Water Conservation
2
,
19
24
.
Chen
F.
,
Li
H.-B.
&
Liu
Y.-J.
2018
Spatio-temporal differentiation and influencing factors of vegetation net primary productivity using GIS and CASA:a case study in Yuanyang County, Yunnan
.
Chinese Journal of Ecology
37
(
7
),
2148
2158
.
Kong
R. R.
2013
Application progress of remote sensing approaches in soil erosion investigation
.
Journal of Anhui Agricultural Sciences
. Doctoral dissertation.
Li
K.
2014
Research on Soil Erosion in Huangtuwa sub-Basin Based on Polyquad Sediment Deposition Analysis and RUSLE Simulation
.
Shaanxi Normal University
. Doctoral dissertation.
Li
X. X.
,
Guo
Q. X.
,
Yang
X. F.
,
Li
J. C.
,
Yan
R.
,
Yan
S. J.
,
Zhao
F. C.
&
Lu
Q. M.
2013
A quantitative study of soil erosion in a small watershed based on GIS and RUSLE model fork
.
Journal of Shanxi Agricultural University: Natural Science Edition
5
,
413
419
.
Liu
B.
,
Keli
Z.
&
Juying
J.
1999
Soil erodibility and its application in erosion forecasting
.
Journal of Natural Resources
4
,
345
350
.
Liu
X.
,
Liangjie
X.
&
Yahan
L.
2021
National scale assessment of the soil erosion and conservation function of terraces in China
.
Ecological Indicators
.
doi:10.1016/J.ECOLIND.2021.107940
.
Qi
J. Y.
2017
Research on Soil Erosion and its Influencing Factors in the Weixiang River Basin Based on GIS and RS
.
Gansu Agricultural University
.
Qin
Z.
2017
A Comparative Study of Soil Erosion in Guizhou Province Based on Remote Sensing Survey and RUSLE Model
.
Guizhou Normal University
.
Wang
X. F.
,
Xiao
F. Y.
,
Yi
L. C.
&
Tan
C.
2017
Evaluation of soil erosion at different watershed scales in the middle reaches of the Yellow River
.
Journal of Anhui Agricultural University
44
(
6
),
8
.
Wang
J.
,
Lu
P.
,
Valente
D.
,
Petrosillo
I.
,
Babu
S.
,
Xu
S.
,
Li
C.
,
Huang
D.
&
Liu
M.
2022
Analysis of soil erosion characteristics in small watershed of the loess tableland plateau of China
.
Ecological Indicators
137
,
108765
.
Wischmeier
W. H.
1978
Predicting rainfall erosion losses-a guide to conservation planning
.
Agriculture Handbook
537 (1978), 285–291.
Xie Yun, Zhang Wenbo, Liu Baoyuan 2001
Calculation of rainfall erosion force by daily rainfall and rain intensity[J]
.
Soil and Water Conservation Bulletin
2001 (6), 53–56.
Yi
K.
,
Wang
S.-Y.
&
Wang
X.
2015
The characteristics of spatial-temporal differentiation of soil erosion based on RUSLE model: a case study of Chaoyang City, Liaoning Province
.
Scientia Geographica Sinica
35
(
3
),
365
372
.
Yue
D.
,
Li
K.
,
Liu
P.
,
Yan
Y.
&
Li
W.
2015
Study on soil erosion and soil conservation in a small watershed of loess pits based on RUSLE
.
Journal of Shaanxi Normal University: Natural Science Edition
43
(
2
),
7
.
Zhang
Q.-H.
,
Wang
H.-Y.
&
Xu
L.
2014
Characteristics of sediment from Huangdong Reservoir in Huizhou, Guangdong Province and their implications on soil erosion during the past 50 years
.
Geographical Research
33
(
4
),
643
653
.
Zhang
L. W.
,
Fu
B. J.
,
Lu
Y. H.
&
Zeng
Y.
2015
Balancing multiple ecosystem services in conservation priority setting
.
Landscape Ecology
30
(
3
),
535
546
.
Zhao
W. Q.
,
Liu
Y.
,
Luo
M. L.
,
Wang
Y. F.
&
Lv
Y. H.
2016
Soil erosion effect assessment of vegetation restoration in small watersheds on Loess Plateau
. Journal of Soil and Water Conservation 5, 89–94.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).