, , , ,

1. College of Forestry, Henan University of Science and Technology, Luoyang 471003, China; 2. State Forestry Administration Key Laboratory of Forest Ecology and Environment, Research Institute of Forest Ecology, Environment and Protection, Chinese Academy of Forestry, Beijing 100091, China; 3. College of Forestry, Henan Agricultural University, Zhengzhou 451002, China

1 Introduction

Water loss and soil erosion not only brings about the decline of soil quality, but also results in siltation of ditches and ponds. Water loss and soil erosion is the result of multi-scale comprehensive action of vegetation, land use, rainfall, topography and soil environmental factors. For a certain scale or area, there are factors controlling the runoff and sediment yield[1-2]. Wang Yan[3], Tian Dalun[4], Gu Yushu[5], Kang Bingetal.[6]studied the effects of environmental factors on the runoff and sediment yield from the soil environment, the geography and climate, the characteristics of vegetation and other factors. In the research methods, Huang Junetal.[7]analyzed the influencing factors of runoff and sediment yield based on the path analysis and gray relational theory. Ai Ningetal.[8]studied the influencing factors of runoff and sediment yield. In the Three Gorges Reservoir Region, some natural secondary forests and artificial shelter forests have simple structure, the function of water conservation and soil conservation is relatively low, these have brought about a series of ecological and environmental problems. During the 13 year from 2003 to 2016, the suspended sediment of the Three Gorges Reservoir reached 2 158.51 million t[9], threatening the normal operation of the dam and the survival and development of human beings in the region. In the late construction of the ecological barrier of the Three Gorges Dam, it is badly necessary to assess the existing forest runoff and sediment yield. The objectives of this study are: (i) to reveal the correlation between runoff and sediment yield and environmental factors in the Three Gorges Reservoir Region; (ii) to make a quantitative analysis on the direct and indirect effects of environmental factors on runoff and sediment yield using path analysis, and determine the main factor influencing runoff and sediment yield.

2 Materialsandmethods

2.1OverviewofthestudyareaThe study area is located in the south bank of the Three Gorges Reservoir Region, Lanlingxi and Shanmuxi valleys of the central and west regions of Maoping Town, Zigui County of Hubei Province, the subtropical continental monsoon region, with annual average temperature of 18.0℃ and annual average rainfall of 1 439 mm. Soil is dominated by yellow soil and bedrock is granite. The study area is a typical agroforestry basin in the reservoir region. The forest land is mainly dominated by mixed forest ofPinusmassonianaLamb.,Cunninghamialanceolata,Quercusacutissima, andLiquidambarformosana. The stand density is not uniform and herbaceous cover is not even. Since 2008, the Forest Ecological Research Station of Three Gorges Reservoir Region (Zigui) has established 15 fixed runoff plots (10 m × 25 m each) in Shanmuxi and Lanlingxi valleys on the basis of topography, soil and tree species (group), canopy closure, and tree density. The runoff plots are surrounded by a cement board and the wall is 30 cm above the ground level. Under ground, it extends to the parent rock layer, to prevent lateral infiltration of the external water flow and the internal water flow. A confluence ditch is provided at the foot of each runoff plot, the runoff pool has size of 2.65 m×2 m×1.6 m (L×W×H). Rainfall devices are installed in the runoff plot. When rainfall generates runoff, the runoff pool collects the runoff and sediment, the runoff and sediment yield can be calculated and converted.

2.2SelectionandsurveyofenvironmentalfactorsCombined with the references[2-6], we selected 15 environmental factors: altitude (HB), slope (PD), aspect (PX), tree species (SZ), plant density (MD), canopy closure (YB), average breast diameter (XJ), average tree height (SG), stand age (NL), leaf area index (YZ), canopy cover (GG), herbaceous cover (CG), litter thickness (DH), status index (SI), and soil bulk density (RZ). The 15 fixed runoff plots established by the Forest Ecological Research Station of Three Gorges Reservoir Region (Zigui) are listed in Table 1.

Table1Basicinformationof15runoffplotsinthestudyarea

RunoffplotSZYBHB∥mPD∥°PXNL∥yDH∥cmYZMD/(plant/ha)XJ∥cmSG∥mGG∥%CG∥%1P．massoniana0．6973710．81．3622．49．53．037279．014．862．455．32P．massoniana0．4351421．00．5034．86．92．71107215．413．165．851．03P．massoniana0．5879912．31．2747．47．82．96100713．213．623．00．04P．massoniana0．6257421．00．9625．08．52．8098011．511．923．80．15P．massoniana0．5754715．00．6856．68．02．43114515．918．426．015．66C．lanceolata0．4043127．30．6460．06．22．01134017．215．368．147．87C．lanceolata0．5265520．20．7949．07．12．78106419．519．536．956．58C．lanceolata0．5074715．00．9845．27．32．77106414．518．231．040．29C．lanceolata0．5251216．00．5655．07．53．00103419．517．531．040．210C．lanceolata0．4648918．70．6642．87．72．48121712．312．162．455．311Mixedforest0．4456220．21．0360．06．72．81115517．915．742．563．512Mixedforest0．4848925．00．5145．06．72．49118015．518．731．040．213Mixedforest0．5545222．00．5660．07．62．26109515．416．344．064．414Mixedforest0．5064018．50．6351．06．62．31124016．615．648．065．315Mixedforest0．4455725．30．4351．05．52．41118712．316．852．464．0

2.3StatisticalanalysisBased on the IBM SPSS Statistics 19 analysis module, we analyzed the correlation between the runoff, sediment yield and environmental factors. Using the correlation analysis, we tested the relationship between the runoff, sediment, and environmental factors; using stepwise multiple regression analysis, we determined the regression equation between runoff, sediment yield and environmental factors. For the significance influencing factors in the regression analysis, we carried out the path analysis and determined the degree of influence and relative importance of environmental factors on the runoff and sediment yield.

3 Resultsandanalyses

3.1Relationshipbetweenrainfall,runoffandsedimentyield

The average rainfall of the runoff plot in the study area was 1 210.6 mm with a runoff of 220.7 mm and a sediment loss of 22.5 t/ha (Table 2). There was a significant correlation between rainfall and runoff (R=0.836,P<0.01) (Table 3). Runoff and sediment loss were significant (R=0.970,P<0.01). The variation coefficient of the rainfall, runoff, and sediment yield was 5.8%, 17.4% and 33.9%, respectively, showing a large difference. The variation coefficient of runoff was 3.0 times of that of rainfall, and the variation coefficient of sediment yield was 1.95 times of that of runoff. Under the influence of the environmental factors, these three indicators were relatively independent.

Table2Changesinrunoff,sedimentyield,andrainfallofrunoffplotsinthestudyarea

3.2Correlationandregressionrelationshipbetweenrunoff,sedimentyield,andenvironmentalfactorsThe correlation between runoff, sediment yield, and environmental factors is shown in Table 3. At 0.01 level, the runoff was correlated with environmental factors such as soil bulk density (R=0.863), herbaceous cover (R=-0.830), slope (R=0.814), canopy cover (R=-0.714), aspect (R=0.778), and litter thickness (R=-0.716); at 0.01 level, the runoff was significantly correlated with the average tree age (R=0.593). At 0.01 level, the sediment yield was correlated with environment factors including litter thickness (R=-0.908), herbaceous cover (R=-0.899), slope (R=0.892), canopy cover (R=-0.835), altitude (R=0.825), bulk density (R=0.832), and aspect (R=0.789); at 0.05 level, it was significantly correlated with the average tree age (R=0.519). The correlation of runoff and sediment yield was extremely significant with the rainfall (P<0.01).

Table3Correlationcoefficientofrunoff,sedimentyield,andenvironmentalfactors

PDPXHBMDYBXJSGNLYZGGCGDHSIRZJYJLNSPD1PX0．822??1HB0．917??0．645??1MD-0．444-0．461-0．4111YB-0．873??-0．662??-0．874??0．5061XJ0．1230．0120．1750．075-0．3461SG-0．004-0．0830．021-0．396-0．1560．551?1NL0．4250．2590．478-0．310-0．567?0．712??0．4911YZ-0．106-0．3450．0720．299-0．0440．3760．084-0．0841GG-0．651??-0．516?-0．579??0．4360．806??-0．356-0．247-0．298-0．2671CG-0．879??-0．766??-0．750??0．547?0．839??-0．255-0．132-0．536?0．1020．794??1DH-0．821??-0．773??-0．680??0．518?0．860??-0．346-0．179-0．451-0．0350．849??0．923??1SI-0．0670．210-0．018-0．366-0．104-0．1660．159-0．240-0．1430．0290．162-0．0531RZ0．813??0．746??0．713??-0．393-0．789??0．215-0．1070．321-0．029-0．613??-0．790??-0．836??0．1451JY0．792??0．728??0．848??-0．497-0．829??0．4180．1380．700??-0．009-0．546??-0．762??-0．739??0．0650．686??1JL0．814??0．778??0．415-0．460-0．732??0．4380．1530．593?0．045-0．714??-0．830??-0．716??-0．0340．863??0．836??1NS0．892??0．789??0．825??-0．511-0．721??0．3940．1430．519?0．045-0．835??-0．899??-0．908??0．0260．832??0．844??0．970??1

Note:*and**denote significant difference at 0.01 and 0.05 levels.

Table4Stepwiseregressionequationofrunoff,sedimentyieldandenvironmentalfactors

ItemRegressionequationR2AdjustedR2PRunoff y=376．069+0．064HB-99．473YB-21．319DH-4．474RZ0．8680．8550．000Sedimentyield y=47．893+0．354PD-39．426YB-0．115CG-1．845DH0．8160．7830．000

3.3DirectandindirecteffectsofenvironmentalfactorsontherunoffThe path analysis results of runoff regression equation showed that the direct path coefficient of the above four environmental factors was: canopy closure (-0.628) > litter thickness (-0.547) > soil bulk density (0.509) > altitude (0.289) (Table 5). Therefore, the canopy closure, litter thickness and soil bulk density affected the runoff mainly through the direct action. The canopy closure had the strongest direct effect on the runoff, manifested as the negative effect, while the litter thickness had the stronger direct effect on the runoff, manifested as the negative effect. The soil bulk density also had a great effect on the runoff, manifested as the positive effect. The direct path coefficient of altitude was 0.289, which had a low positive direct effect on the runoff.

In the runoff regression equation, the order of indirect path coefficient was: soil bulk density (0.354) > litter thickness (-0.169) > altitude (0.126) > canopy closure (-0.104). The soil bulk density (0.354) was the largest in indirect path coefficient. Indirect effect of soil bulk density on the runoff was the largest. Soil bulk density exerted an indirect negative effect on the runoff mainly through canopy closure and litter thickness. The canopy closure (YB) exerted an indirect positive effect on the runoff mainly through the soil bulk density and altitude. The litter thickness (-0.169) exerted an indirect positive effect on the runoff mainly through the canopy closure and altitude. The altitude had the smallest indirect effect on the runoff.

Table5Decompositionofcorrelationcoefficientofrunoffandenvironmentalfactors

EnvironmentalfactorsCorrelationcoefficientDirectpathcoefficientIndirectpathcoefficientAltitudeCanopyclosureLitterthicknessSoilbulkdensitySubtotalAltitude0．4150．289-0．1520．037-0．0630．126Canopyclosure-0．732-0．6280．239--0．123-0．012-0．104Litterthickness-0．716-0．5470．167-0．194--0．142-0．169Soilbulkdensity0．8630．509-0．0030．475-0．118-0．354

3.4DirectandindirecteffectsofenvironmentalfactorsonthesedimentyieldThe path analysis results (Table 6) indicated that in the sediment yield regression equation, the direct path coefficient of environmental factors was herbaceous cover (-0.815) > litter thickness (-0.777) > canopy closure (-0.624) > slope (0.620). Thus, the herbaceous cover, litter thickness, canopy closure, and slope affected the sediment yield mainly through the direct action. The direct path coefficient of herbaceous cover was the largest and herbaceous cover had the largest negative effect on the sediment yield. The canopy closure and litter thickness also had strong direct negative effects on the sediment yield, and the slope (PD) had high correlation with the sediment (R=0.892), but its direct path coefficient was only 0.620, and its direct effects on the sediment yield was lower than other factors.

In the sediment yield regression equation, the indirect path coefficient was ranked as slope (0.272) > litter thickness (-0.131) > canopy closure (-0.097) > herbaceous cover (-0.084). The slope had the largest indirect path coefficient, and it exerted the indirect positive effect on the sediment yield of canopy closure and herbaceous cover. Then, it was the litter thickness, which exerted the indirect negative effect on the sediment yield mainly through the canopy closure and slope. The herbaceous cover had the lowest indirect effect on the sediment yield. The path analysis of sediment yield and environmental factors further indicated that there was the interaction between environmental factors and the sediment yield, and such interaction affected the output of sediment yield.

Table6Decompositionofcorrelationcoefficientofsedimentyieldandenvironmentalfactors

EnvironmentalfactorsCorrelationcoefficientDirectpathcoefficientIndirectpathcoefficientSlopeCanopyclosureHerbaceouscoverLitterthicknessSubtotalSlope0．8920．620-0．2340．251-0．2130．272Canopyclosure-0．721-0．6240．232--0．123-0．012-0．097Herbaceouscover-0．899-0．8150．253-0．182--0．156-0．084Litterthickness-0．908-0．777-0．3130．438-0．256 --0．131

In the runoff equation of Table 4, the correlation coefficient between the soil bulk density and runoff was the highest (R=0.863), the correlation coefficient between the canopy closure and runoff was -0.732, but its direct path coefficient was the largest (-0.628). In the sediment yield equation of Table 4, the correlation coefficient between the litter thickness and sediment yield was the largest (R=-0.908), and the direct path coefficient was -0.777 (Table 6); the correlation coefficient was -0.899, but the direct path coefficient was the largest (-0.815); the correlation coefficient between canopy closure and sediment yield was the smallest (-0.721), and the direct path coefficient (-0.624) was larger than that of the slope (0.620). This indicated that the size of correlation coefficient between environmental factors and sediment yield was not consistent with the direct path coefficient, there was obvious interaction between environmental factors, and such interaction jointly affected the sediment yield.

4 Discussions

4.1EffectsofenvironmentalfactorsontherunoffandsedimentyieldThere have been reports about the relationship between the runoff or sediment yield and stand structure[10-13], canopy closure[12], and vegetation cover[14-15]. Through correlation analysis, we concluded that the environmental factors highly correlated with the runoff are ranked as the soil bulk density, herbaceous cover, slope, canopy closure, aspect, litter thickness and so on; the environmental factors highly correlated with sediment yield are ranked as litter thickness, herbaceous cover, slope, canopy cover, and canopy closure, and this conclusion is consistent with the relevant literature. For example, Wang Xiaoyan[16]stated that the rainfall and canopy closure have the largest effect on the slope runoff. The soil bulk density has a greater effect on the runoff, when the soil bulk density increases, the runoff will increase, which is consistent with findings of Wang Huietal.[17], Lu Dianqingetal.[18]and Shen Yitong[19]in the relationship between the soil bulk density and runoff. Based on multivariate correlation analysis on the relationship between surface runoff and underforest vegetation and soil characteristics, Liu Simianetal.[20]believed that the amount of surface runoff has a significantly negative correlation with the litter thickness, canopy cover, soil non-capillary porosity and water-stable soil aggregates, and has a significant positive correlation with the soil bulk density. Chen Qiangetal.[21]considered that when the soil bulk density decreased by 0.15 g/cm3, the initial soil infiltration rate could be increased by more than 30%, and the water and soil loss could be reduced by 20% and 40% respectively. The average water holding capacity of forest litters was 309.5% of the total dry weight. The maximum water holding capacity of forest litter was 4.18 mm[22]. The litter had the function of blocking runoff, intercepting sediment yield and balancing soil water heat conditions and other effects, the reduction and inhibition of runoff is also particularly significant[23-24]. The type of forest vegetation selected in this study has better site conditions, litter distribution is relatively uniform, the level is obvious, and the litter has a strong negative effect on the runoff and sediment yield.

4.2PathanalysisofsignificantenvironmentalfactorsThrough the stepwise regression equation, we obtained the direct path coefficient of runoff as follows: canopy closure > litter thickness > soil bulk density > altitude. The indirect effect was ranked as the soil bulk density > litter thickness > altitude> canopy closure. From the perspective of direct path coefficient, the canopy closure had the greatest effect on the runoff, and it also inhibited the runoff together with other factors. Through the path analysis of the significant factors of the stepwise multivariate regression equation, we obtained the direct and indirect effects of herbaceous cover, canopy closure, litter thickness and slope on the sediment yield. From the perspective of direct path coefficient, the herbaceous cover > litter thickness> canopy closure > slope; from the indirect path coefficient, the slope > litter thickness > canopy closure > herbaceous cover. The herbaceous cover, litter thickness and canopy closure play a decisive role in inhibiting the sediment yield, while slope has a positive effect on sediment yield. From the indirect path coefficient analysis, the slope promotes the sediment yield through other factors such as the herbaceous cover and canopy closure. Based on the correlation and multiple regression analysis, we decomposed the correlation coefficient between environmental factors and the runoff and sediment yield into direct and indirect path coefficients. Through this, we can not only evaluate the direct effect of environmental factors on the runoff and sediment yield, but also analyze the indirect effects of environmental factors on the runoff and sediment yield through other related factors. Path analysis can reveal the causal relationship between environmental factors and the runoff and sediment yield through simple correlations, assess the relative importance of environmental factors to the runoff and sediment output, and reveal the role of actual and other un-included environmental factors to the runoff and sediment yield, and more objectively reflect the relationship between environmental factors.

4.3MaincontrollingfactorsoftherunoffandsedimentyieldWater loss and soil erosion is the result of multi-scale comprehensive action of vegetation, land use, rainfall, topography and soil environmental factors. For a certain scale or area, there are factors controlling the runoff and sediment yield[1-2]. In related researches, the correlation between soil erosion and environmental factors is mostly based on correlation and multiple regression analysis methods[25-26]. Multivariate regression analysis can only assess the direct effect of environmental variables, but can not assess the interaction and transmission relationship between environmental variables, so the analysis results are relatively one-sided[8]. On the basis of correlation and regression analysis, Zheng Jiangkunetal.[25]held that the slope position and soil quality are the dominant environmental factors of the runoff and sediment yield in the loess hilly and gully regions. Zhao Mingsong[27]believed that the slope and elevation in the Dabie Mountains in western Anhui Province dominate the spatial distribution of soil erosion intensity in the region and the slope is the dominant factor of soil erosion intensity. Li Yang[28]stated that the slope and elevation of the typical watershed in Yimeng Mountain closely affect the soil erosion.

In this study, through separate analysis of factors influencing the runoff and sediment yield, we concluded that main controlling factors of the runoff are the canopy closure and litter thickness, and the soil erosion is mainly influenced by the herbaceous cover and litter thickness. The determination coefficients of error term of the runoff and sediment yield are both significant at 0.01 level, but the residual path coefficient is larger, indicating that the we have not covered the environmental factors that enter the stepwise regression equation and affect runoff and sediment yield. In view of the main controlling environmental factors of the runoff and sediment yield in the Three Gorges Reservoir Region, it is required to consider the initial planting density of stands in the subsequent construction, cultivate reasonable canopy and herbaceous cover. This is of great significance for the ecological construction of the Three Gorges Reservoir Region.

[1] QIU Y, FU BJ, WANG Y. Spatiotemporal variation in soil erosion and its relation to environmental factors[J]. Journal of Soil and Water Conservation, 2002, 16(1): 108-111. (in Chinese).

[2] QIU Y, FU BJ, WANG J,etal. Spatiotemporal variability of the soil erosion and its relations to the influencing factors on the Loess Plateau, China[J]. Acta Ecologica Sinica, 2004, 24(9): 1871-1877. (in Chinese).

[3] WANG Y, WANG B, ZHAO GD,etal. Soil moisture physical characteristics of three forest types in Dagangshan Mountain in Jiangxi Province[J]. Journal of Soil and Water Conservation, 2008, 22(14): 151-153, 173. (in Chinese).

[4] TIAN DL, CHEN SJ. Analysis of soil hydrology-physic properties in the artificial forest ofCinnamomumcamphora(L.) Presl[J]. Journal of Central South Forestry University, 2005, 25(2): 1-6. (in Chinese).

[5] GU YS, XING ZK, HAN YZ,etal. Water-holding characteristics of four typical forests litter in upstream of Hunhe river[J]. Journal of Nanjing Forestry University (Natural Sciences), 2013, 37(1): 31-36. (in Chinese).

[6] KANG B, LIU SR, CAI DX,etal. Effects ofPinusmassonianaplantation stand density on understory vegetation and soil properties[J]. Chinese Journal of Applied Ecology, 2009, 20(10): 2323-2331. (in Chinese).

[7] HUANG J, ZHAO XN, WU PT. Factors analysis of slope runoff and sediment based on path analysis and grey relational analysis[J]. Journal of Sichuan University (Engineering Science Edition), 2012, 44(5): 64-70. (in Chinese).

[8] AI N, WEI TX, ZHU QK,etal. Factors affecting slope runoff and sediment yield in northern Shaanxi Province based on path analysis[J]. Journal of Beijing Forestry University, 2015，37(6)：77-84. (in Chinese).

[9] Changjiang Water Resources Commission, The Ministry of Water Resources of the People’s Republic of China. Changjiang Sediment Bulletin[M]. Changjiang Publishing House, 2016. (in Chinese).

[10] LI DN, WANG B, CAI TJ,etal. Rainfall redistribution traits of three main forest types in Dagangshan Mountains of Jiangxi Province, China[J]. Chinese Journal of Applied Ecology, 2014, 25(8): 2193-2200. (in Chinese).

[11] GONG GT, CHEN JH, LI YQ,etal. A comparative analysis of the hydrological effects of the four cypress stand types in Sichuan Basin[J]. Acta Ecologica Sinica, 2011, 31(10): 2716-2726. (in Chinese).

[12] GONG GT, LI YQ, ZHU ZF,etal. The suitable stand structure and hydrological effects of the cypress protection forests in the central Sichuan hilly region[J]. Acta Ecologica Sinica, 2012, 32(3): 923-930. (in Chinese).

[13] WANG L, SUN CZ, ZHOU B. Effects of different structure Beijing Jiulongshan on of platycladusorientalis plantation in redistribution of precipitation[J]. Forest Research, 2016, 29(5): 752-758. (in Chinese).

[14] JIANG GJ, ZHENG XX, NING YC. Relationship between forest stand structure and function of water conservation——A case study of Badaling forest farm[J]. Journal of Northwest Forestry University, 2012, 27(2): 175-179. (in Chinese).

[15] HUO XP, LI XW, ZHANG J,etal. Experimental on soil storage capacity and infiltration capability under different vegetation types in the subalpine of Western Sichuan[J]．Science of Soil and Water Conservation, 2009, 7(6): 74-79. (in Chinese).

[16] WANG XY, BI HX, GAO LB,etal. Discrimination of factors influencing the runoffs of different spatial scales on loess region in Western Shanxi[J]. Journal of Northwest Sci-Tech University of Agriculture and Forestry(Natural Science Edition), 2014, 42(1): 159-166. (in Chinese).

[17] WANG H, WANG QJ, SHAO MA. Effect of soil bulk density on soil nutrient in runoff from loess slope[J].Journal of Soil and Water Conservation, 2007, 21(3): 10-13. (in Chinese).

[18] LU DQ, WANG H, PAN Y,etal. Effect of bulk density changes on soil solute transport characteristics[J]. Journal of Natural Science of Hunan Normal University, 2010, 33(1): 75-79. (in Chinese).

[19] SHEN YT, GUO CJ, JIN S,etal. Effect of soil bulk density on nutrient loss of black soil slope[J]. Journal of Soil and Water Conservation, 2016, 30(1): 26-30. (in Chinese).

[20] LIU SM, TIAN DL, XIANG WH,etal. The impacts of thinning intensity on overland flow in a Chinese fir plantation[J]. Acta Ecologica Sinica, 2015, 35(17): 5769-5775. (in Chinese).

[21] CHEN Q, YURIY SK, CHEN Y,etal. Seasonal variations of soil structures and hydraulic conductivities and their effects on soil and water conservation under no-tillage and reduced tillage[J]. Acta Pedologica Sinica, 2014, 51(1): 11-21. (in Chinese).

[22] WANG PC, XIAO WF, ZHANG SG,etal. Amount and water holding capacity of forest litter in Three Gorges Reservoir area[J]. Science of Soil and Water Conservation, 2008, 6(4): 41-47. (in Chinese).

[23] LU HJ, LIU WJ, LUO QJ. Eco-hydrological effects of litter layer in a mountainous rubber plantation in Xishuangbanna, Southwest China[J]. Chinese Journal of Ecology, 2011, 30(10): 2129-2136. (in Chinese).

[24] WANG SY, YU XX, JIA GD,etal. Hydrological effects of forest litters of different forests in Beijing mountainous area[J]. Science of Soil and Water Conservation, 2011, 9(5): 42-47. (in Chinese).

[25] ZHENG JK, WEI TX, ZHAO J,etal. Environmental interpretation on slope runoff-erosion-sediment process in loess hilly and gully region[J]. Scientia Geographica Sinica, 2015, 35(7): 905-911. (in Chinese).

[26] TIAN YW, HUANG ZL, XIAO WF,etal. Changes of runoff and sediment yield in Three Gorges Reservoir’s ecological barrier range based on AnnAGNPS model[J]. Journal of Northwest Forestry University, 2014, 29(6): 24-28. (in Chinese).

[27] ZHAO MS, LI DC, CHENG XF,etal. Spatial characteristics of soil erosion and its relationship to topography in Dabie Mountains of West Anhui[J]. Journal of Anhui Normal University(Natural Science Edition), 2017, 40(3): 265-270. (in Chinese).

[28] LI Y, LIU X, YAO XY,etal. Coupling relationship between soil erosion and environmental factors in typical watershed of Yimeng Mountion[J]. Research of Soil and Water Conservation, 2011, 18(2): 66-70. (in Chinese).

AboutKIT

The Royal Tropical Institute (KIT) in Amsterdam is an independent centre of knowledge and expertise in the areas of international and intercultural cooperation, operating at the interface between theory and practice and between policy and implementation. The Institute contributes to sustainable development, poverty alleviation and cultural preservation and exchange.