Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

Variations in Soil Erosion Resistance of Gully Head Along a 25-Year Revegetation Age on the Loess Plateau

Variations in Soil Erosion Resistance of Gully Head Along a 25-Year Revegetation Age on the Loess... water Article Variations in Soil Erosion Resistance of Gully Head Along a 25-Year Revegetation Age on the Loess Plateau 1 1 , 2 1 , Zhuoxin Chen , Mingming Guo and Wenlong Wang * State Key Laboratory of Soil Erosion and Dryland Farming on the Loess Plateau, Institute of Water and Soil Conservation, Northwest A&F University, Yangling 712100, China; xiyu.zxchen@foxmail.com (Z.C.); guomingming@iga.ac.cn (M.G.) Key Laboratory of Mollisols Agroecology, Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Harbin 150081, China * Correspondence: wlwang@nwafu.edu.cn Received: 10 October 2020; Accepted: 19 November 2020; Published: 24 November 2020 Abstract: The e ects of vegetation restoration on soil erosion resistance of gully head, along a revegetation age gradient, remain poorly understood. Hence, we collected undisturbed soil samples from a slope farmland and four grasslands with di erent revegetation ages (3, 10, 18, 25 years) along gully heads. Then, these samples were used to obtain soil detachment rate of gully heads by the hydraulic flume experiment under five unit width flow discharges (2–6 m h). The results revealed that soil properties were significantly ameliorated and root density obviously increased in response to restoration age. Compared with farmland, soil detachment rate of revegetated gully heads decreased 35.5% to 66.5%, and the sensitivity of soil erosion of the gully heads to concentrated flow decreased with revegetation age. The soil detachment rate of gully heads was significantly related to the soil bulk density, soil disintegration rate, capillary porosity, saturated soil hydraulic conductivity, organic matter content and water stable aggregate. The roots of 0–0.5 and 0.5–1.0 mm had the highest benefit in reducing soil loss of gully head. After revegetation, soil erodibility of gully heads decreased 31.0% to 78.6%, and critical shear stress was improved by 1.2 to 4.0 times. The soil erodibility and critical shear stress would reach a stable state after an 18-years revegetation age. These results allow us to better evaluate soil vulnerability of gully heads to concentrated flow erosion and the eciency of revegetation. Keywords: soil erosion; gully erosion; vegetation restoration; soil erodibility; land use 1. Introduction Soil erosion is recognized as a global environmental problem, which severely damages infrastructure, causes land degradation and water pollution, and threatens the safety of human production and life [1–3]. In the past few decades, many scholars have made many e orts to study the process and mechanism of soil erosion, establish many soil erosion prediction models and try to control soil erosion [4–7]. At present, a set of soil erosion control measures system integrating engineering measures, agricultural measures, and biological measures has been formed [8–10], especially vegetation measures play an extremely important role in soil erosion control [11,12]. Previous studies have shown that revegetation can e ectively reduce soil erosion. For example, Wang et al. [13] found that soil detachment capacity of abandoned farmland was 1.02 to 2.29 times greater than four restored lands. Li et al. [14] reported that the ratios of the soil detachment capacity of cropland to those of orchard, shrubland, woodland, grassland, and wasteland were 7.14, 12.29, 25.78, 28.45, and 46.43, respectively. The improvement of soil erosion resistance by revegetation is mainly controlled Water 2020, 12, 3301; doi:10.3390/w12123301 www.mdpi.com/journal/water Water 2020, 12, 3301 2 of 16 by the combination of soil properties and root traits [15–18]. In terms of soil properties, many studies have verified that the revegetation significantly a ects soil erosion by changing the soil bulk density, organic matter content, and water-stable aggregate [19–21]. Furthermore, the vegetation root zone is the dynamic interface of soil–plant–atmosphere continuum in partitioning rain and irrigation water into evaporation, transpiration, runo , and deep drainage [22,23], but is also the home of “green water” which is the source of plant nutrition [24]. Especially, the vegetation root systems also play a great role in protecting soil against flow scouring by a ecting soil water movement [25,26]. Root-permeated soils exhibited lower erosion rates primarily through increasing the required shear stress before detachment [27]. Moreover, root growth can bind and bond soil particles and aggregate, thus, enhances soil resistance to erosion [28]. Some root parameters, for example root biomass, length density, and surface area density, were used to estimate the e ect of root on soil detachment [21,28–31]. De Baets and Poesen [25] found that soil detachment rate reduced exponentially with increasing root biomass. Some studies also showed that soil detachment was related to root architecture and fibrous root was more e ective than tap root in reducing soil loss [24,31]. However, the most of previous studies only focus on the impact of revegetation on soil erosion resistance of hillslopes. In the watershed dominated by gully erosion, the gully head is the main source of soil erosion, but the e ect of revegetation on soil erosion resistance of gully heads remains unclear. Therefore, there is a strong need to understand the e ect of revegetation on soil erosion of gully head by concentrated flow to develop a more reasonable vegetation model. Notably, in the gully region of the Loess Plateau, about 63% of total runo is generated from the loess tableland with a gentle slope of 1–5 , which can initiate gully headcut erosion and contribute 86.3% of total sediment [32]. The gully headcut erosion by concentrated flow became the main sediment resource. At present, the gully headcut erosion was controlled e ectively due to the implementation of a series of control measures (e.g., the “Three Protection Belts” and the “Green for Grain” project), which, to some extent, was attributed to the fact that the revegetation improves the soil resistance of gully heads to concentrated flow [21,33]. Since some ecological restoration projects were conducted, land use has changed dramatically in the Loess Plateau [34]. Hence, the land use has changed, and the natural succession of vegetation was promoted [35]. With progression in natural restoration of grassland, soil physical and chemical properties and vegetation characteristics (e.g., coverage, community structure, species composition and diversity, and root diameter, density, and diameter distribution) varied greatly [36,37]. These changes would result in dynamic variations in soil erosion resistance. However, the response of soil erosion resistance to vegetation succession process mainly focused on the hillslope in the hilly-gully region of Loess Plateau [15,16], and few studies were conducted to explore the response of soil resistance of gully heads by concentrated flow to vegetation succession process. Therefore, to evaluate the e ect of revegetation process on soil erosion resistance of gully heads and optimize revegetation measures for controlling gully headcut erosion in the gully region of the Loess Plateau, we selected four grasslands with di erent revegetation ages (3, 10, 18, 25 years) along gully heads with the slope farmland as the control. This study aimed to (1) quantify the e ect of revegetation age on soil detachment by concentrated flow, (2) clear the relationships between soil detachment rate and soil and root properties, and (3) confirm the dynamic variation in soil erosion resistance of gully head with revegetation age. 2. Material and Method 2.1. Study Area The study was conducted in the Nanxiaohegou watershed in the Xifeng Research Station of Soil 0  0  0  0 2 and Water Conservation (35 41 –35 44 N, 107 30 –107 37 E). The watershed has an area of 36.3 km and altitudes ranging from 1050 to 1423 m above mean sea level (Figure 1) in the typical gully region of the Loess Plateau. The climate is temperate continental semiarid. The mean temperature is 10 C, Water 2020, 12, x FOR PEER REVIEW 3 of 15 Water 2020, 12, 3301 3 of 16 the characteristic of annual variation and uneven distribution during the year. In the form of short and the frost-free period is 160–180 days. Annual precipitation is approximately 523 mm, which has heavy storms, 58.8% of the rainfall occurs from July to September. The soil type is yellow loamy soil. the characteristic of annual variation and uneven distribution during the year. In the form of short The original vegetation has disappeared due to human activities. Gully headcut erosion is the main heavy storms, 58.8% of the rainfall occurs from July to September. The soil type is yellow loamy resource of sediment yield in the watershed. Since the 1970s, some soil and water conservation soil. The original vegetation has disappeared due to human activities. Gully headcut erosion is the projects, for example the “three protection belts” project and the “Green for Grain” project and so on, main resource of sediment yield in the watershed. Since the 1970s, some soil and water conservation were implemented to control soil and water loss, and the vegetation cover of the Loess Plateau projects, for example the “three protection belts” project and the “Green for Grain” project and so on, were implemented to control soil and water loss, and the vegetation cover of the Loess Plateau increased increased to 59.6% in 2013. Additionally, the land use has undergone tremendous changes [38]. These to 59.6% in 2013. Additionally, the land use has undergone tremendous changes [38]. These e orts also efforts also effectively stabilized the gully heads and thus contained the gully headcut erosion [33]. e ectively stabilized the gully heads and thus contained the gully headcut erosion [33]. At present, −2 −1 At present, the annual soil erosion module is effectively controlled at the level of 2440 t km a in 2 1 the annual soil erosion module is e ectively controlled at the level of 2440 t km a in the study area, the study area, and the vegetation communities comprise mainly planted forests and shrubs and and the vegetation communities comprise mainly planted forests and shrubs and native secondary native secondary herbaceous plants [21]. herbaceous plants [21]. Figure 1. Location of the study area on the Loess Plateau and the location of sampling sites in the Figure 1. Location of the study area on the Loess Plateau and the location of sampling sites in the Nanxiaohegou watershed. Nanxiaohegou watershed. 2.2. Sampling Sites Selection 2.2. Sampling Sites Selection During our investigation of gully heads, some cracks developed near gully heads. Kompani-Zare et al. [39] and Guo et al. [21] stated that soil samples from 0 to 30 cm depth near the gully heads During our investigation of gully heads, some cracks developed near gully heads. Kompani- (the distance was less than 5 m) can represent soil properties of the gully heads. Therefore, in consideration Zare et al. [39] and Guo et al. [21] stated that soil samples from 0 to 30 cm depth near the gully heads of collapsibility and vertical joints development of loess, the sampling plots were established about 1.0 m (the distance was less than 5 m) can represent soil properties of the gully heads. Therefore, in meters from gully heads to ensure safety. As a result, four natural restoration grasslands with different consideration of collapsibility and vertical joints development of loess, the sampling plots were ages (3, 10, 18, 25 years) were selected (Figure 1). The natural restoration age was confirmed by consulting the village elders and scientists at the scientific experimental station. The slope aspect and gradient, established about 1.0 m meters from gully heads to ensure safety. As a result, four natural restoration elevation, soil type, and previous farming practices of the selected sites were similar to minimize the grasslands with different ages (3, 10, 18, 25 years) were selected (Figure 1). The natural restoration effects of these factors on the experimental results. For comparison, one corn-planted farmland site, with a age was confirmed by consulting the village elders and scientists at the scientific experimental station. topography similar to that of the grasslands, was selected as a control. The basic information of the five The slope aspect and gradient, elevation, soil type, and previous farming practices of the selected selected sites is listed in Table 1. sites were similar to minimize the effects of these factors on the experimental results. For comparison, Table 1. Basic information of the selected five sampling sites. one corn-planted farmland site, with a topography similar to that of the grasslands, was selected as Restoration Dominant a control. The basic information of the five selected sites is listed in Table 1. Site Code Slope ( ) Coverage (%) Altitude (m) Main Companion Species Age (yr) Communities SF 0 2.4 — 1420 Zea mays Setaria viridis NR3 3 2.9 72.3 1405 Artemisia capillaris Artemisia sacrorum Table 1. Basic information of the selected five sampling sites. NR10 10 2.4 80.8 1401 Artemisia sacrorum Artemisia capillaris NR18 18 3.2 93.4 1390 Artemisia sacrorum Artemisia capillaris Artemisia sacrorum + Site Restoration Slope Coverage Altitude Dominant Main Companion NR25 25 3.1 91.2 1380 Bothriochloa ischaemum Lespedeza daurica Code Age (yr) (° ) (%) (m) Communities Species SF 0 2.4 — 1420 Zea mays Setaria viridis Artemisia NR3 3 2.9 72.3 1405 Artemisia sacrorum capillaris Artemisia NR10 10 2.4 80.8 1401 Artemisia capillaris sacrorum Artemisia NR18 18 3.2 93.4 1390 Artemisia capillaris sacrorum Bothriochloa Artemisia sacrorum NR25 25 3.1 91.2 1380 ischaemum + Lespedeza daurica Water 2020, 12, 3301 4 of 16 2.3. Sampling and Measurement of Soil and Root In this study, seven soil and root property parameters including soil bulk density, capillary porosity, soil disintegration rate, soil water-stable aggregate, soil saturated hydraulic conductivity, organic matter content, and root mass density were measured. Firstly, three repeated sampling plots (5 m 5 m) were established in each of gully head sampling sites with the litter layer removed, and topsoil samples (0–30 cm) were collected. Then, three cutting rings (200 cm ) were used to randomly collect soil samples in each plot, and a total of nine samples were oven-dried at 105 C for 24 h to determine the soil bulk density of each gully head site. Similarly, the other 9 soil samples were also collected by cutting rings of 200 cm to determine the soil saturated hydraulic conductivity by applying the constant water head test method. Three cutting rings (100 cm ) were used to collect soil samples for the measurement of soil capillary porosity [33]. Three man-made steel cubical boxes (5 cm in length) were used to collect soil samples for measuring soil disintegration rate by using a disintegration box [14,40]. Lastly, the other three samples were randomly collected in each plot to form a mixed sample. A total of 45 mixed samples were obtained and used for laboratory analyses of organic matter content and water-stable aggregate and its stability. These mixed soil samples were air-dried at room temperature, with large roots and organic residues manually removed. Sieves with apertures (0.25, 0.5, 1.0, 2.5, and 5.0 mm) were used to test the water-stable aggregate. The potassium dichromate external heating method was used to measure the soil organic matter content. 2.4. Hydraulic Flume Experiments A hydraulic flume experiment was conducted to determine the soil resistance to concentrated flow upstream gully heads (Figure 2). The size of the flume was 2.0 m long and 0.15 m wide similar to the one used by De Baets et al. [28,31], which was enough to make water flow along the slope soil. An opening (0.5 m length and 0.1 m wide) was set at the bottom of the flume, and a metal sample box with the same size was used to collect undisturbed soil samples so that the surface of the soil sample was at the same level of the flume surface. The space between sampling box and flume edge sealed with painter ’s mastic to prevent boundary e ects. According to the study of Guo et al. [40], the flume experiment was carried out under five di erent unit width flow discharges of 2, 3, 4, 5, 6 3 1 m s , and thus, a total of 100 samples (5 sites 5 flow discharges 4 replications) were collected to measure soil resistance of gully heads. To simulate real flow generation conditions, the soil should be saturated by using a watering pot before experiment. During the experiment, a portable flow meter instrument (LS300-A) with 1.5% accuracy was used to measure flow velocity which was regarded as the flow velocity scoring soil area. Runo and sediment samples were collected with sampling tanks, and the sampling time was recorded. The measured flow velocity was modified according to flow regime [41]. Sampled sediment was oven-dried at 105 C for 24 h to determine the soil loss amount 2 1 (SLA, kg). Thus, the soil detachment rate (D , kg m s ) could be calculated as follows: please check font size, please check all reference citation SLA D = (1) AT where SLA is the oven-dry mass of every sediment sample (kg), T is the experimental period (s) and A is the soil sample area (m ). In addition, the relative soil detachment rate (RD ) was calculated as the ratio between D for the root-permeated soil samples and that for the farmland topsoil samples, tested at the same condition [28]. Water 2020, 12, x FOR PEER REVIEW 5 of 15 3 −1 where h is flow depth (m), q is the flow discharge (m s ), w is the width of the flume (m), v is the −1 −3 −2 mean flow velocity (m s ).  is the water mass density (kg m ), g is the gravity constant (m s ), and −1 S is the slope steepness (m m ). After each scouring test, a steel cubical box (10 cm in length) was used to take soil sample in the center of soil area of scouring flume, and the sample was soaked in tap water for about one hour to increase the dispersion of soil and then were placed on a 0.25 mm sieve and washed with tap water using low-pressure head. The living roots, plant debris and some pebbles were left on the sieve. Only the living roots were picked out carefully using tweezers one by one [15]. The washed roots were classified into 4 levels (0–0.5, 0.5–1.0, 1.0–2.0, and >2.0 mm) by vernier caliper and then were oven- Water 2020, 12, 3301 5 of 16 −3 dried for 24 h at 65 °C and weighed to calculate root mass density (RBD, kg m ). Figure 2. Skitch of scouring flume for determining soil erosion resistance of gully heads. Figure 2. Skitch of scouring flume for determining soil erosion resistance of gully heads. In addition, the flow depth (h, m) and shear stress (, Pa) were also calculated as follows: 2.5. Parameter Calculation after the Experiments h = (2) Soil particle is detached when flow shear vw stress exceeds the critical shear stress [6]. Soil erodibility parameter (Kr) and critical shear stress (c) were estimated for every natural restoration = ghS (3) stage as the slope coefficient and intercept on the abscissa axis of the regression line between soil 3 1 where h is flow depth (m), q is the flow discharge (m s ), w is the width of the flume (m), v is the detachment rate and shear stress as described in the WEPP model as follow: 1 3 2 mean flow velocity (m s ).  is the water mass density (kg m ), g is the gravity constant (m s ), and S is the slope steepness (m m ). DK = ()   (4) r e c After each scouring test, a steel cubical box (10 cm in length) was used to take soil sample in the center of soil area of scouring flume, and the sample was soaked in tap water for about one hour to Generally, soil detachment rate can be considered as zero when root reached the infinity. To increase the dispersion of soil and then were placed on a 0.25 mm sieve and washed with tap water quantify the relationship between detachment rate and root mass density, the Hill curve was selected using low-pressure head. The living roots, plant debris and some pebbles were left on the sieve. to simulate the relationship between them [21,31,42]. The Hill curve is expressed as follows [43]: Only the living roots were picked out carefully using tweezers one by one [15]. The washed roots were classified into 4 levels (0–0.5, 0.5–1.0, 1.0–2.0, and >2.0 mm) by vernier caliper and then were KX RD = , (K 0, a 0, b 0) oven-dried for 24 h at 65 C and weighed to calculate root mass density (RBD, kg m ). (5) Xb  2.5. Parameter Calculation after the Experiments where RDr is relative soil detachment rate; Xr is root mass density; K, a and b are constants. The Soil particle is detached when flow shear stress exceeds the critical shear stress [6]. Soil erodibility parameter a determines the shape of the curve, b determines the steepness of the curve and K is the parameter (K ) and critical shear stress ( ) were estimated for every natural restoration stage as the r c asymptote of Dr for infinitesimal Xr values. Additionally, the Hill curve can be used to evaluate the slope coecient and intercept on the abscissa axis of the regression line between soil detachment rate ability of roots to increase soil resistance against concentrated flow erosion. According to Li et al. [44], and shear stress as described in the WEPP model as follow: b^(1/a) is plant specific and can be used as an index to compare the effectiveness of different plant D = K (  ) (4) r r c roots in reducing soil erosion rates: the lower b^(1/a), the more effective the plant root. When the value of Xr is b^(1/a), the soil detachment rates is reduced by 50%. Generally, soil detachment rate can be considered as zero when root reached the infinity. To quantify the relationship between detachment rate and root mass density, the Hill curve was selected to simulate 2.6. Statistical the relationship Analysisbetween and Plothem tting[ 21,31,42]. The Hill curve is expressed as follows [43]: The analysis of one-way ANOVA follo KXwed by multiple comparisons with LSD was applied to RD = , (K > 0, a < 0, b > 0) (5) X + b assess the differences of soil properties (Soil bulk density, soil capillary porosity, soil disintegration rate, saturated hydraulic conductivity, organic matter content, and water-stable aggregate) and root where RD is relative soil detachment rate; X is root mass density; K, a and b are constants. The parameter r r a determines the shape of the curve, b determines the steepness of the curve and K is the asymptote of mass density among the five revegetation ages. All soil and root variables of each revegetation ages D for infinitesimal X values. Additionally, the Hill curve can be used to evaluate the ability of roots r r were tested whether the data exhibited a normal distribution and variance homogeneity by Shapiro- to increase soil resistance against concentrated flow erosion. According to Li et al. [44], bˆ(1/a) is plant Water 2020, 12, 3301 6 of 16 specific and can be used as an index to compare the e ectiveness of di erent plant roots in reducing soil erosion rates: the lower bˆ(1/a), the more e ective the plant root. When the value of X is bˆ(1/a), the soil detachment rates is reduced by 50%. 2.6. Statistical Analysis and Plotting The analysis of one-way ANOVA followed by multiple comparisons with LSD was applied to assess the di erences of soil properties (Soil bulk density, soil capillary porosity, soil disintegration rate, saturated hydraulic conductivity, organic matter content, and water-stable aggregate) and root mass density among the five revegetation ages. All soil and root variables of each revegetation ages were tested whether the data exhibited a normal distribution and variance homogeneity by Shapiro-Wilk test and Levene test, respectively. If the data failed to meet the two conditions, the Kruskal–Wallis test was performed for the above analysis. The interaction e ect of flow discharge and revegetation age was detected using a two-way ANOVA. Pearson’s correlation analysis was used to determine linkages among soil properties, root mass density, and soil detachment rate. Relationships among soil detachment rate, soil properties, flow shear stress and restoration age were analyzed by the regression method. The data analyses were conducted in SPSS v. 16.0 statistical software (IBM Corp., Armonk, NY, USA). The figure plotting was conducted by Origin v. 2020 (OriginLab Corp., Northampton, MA, USA). 3. Results and Discussion 3.1. E ect of Revegetation on Soil and Root Properties of Gully Heads Figure 3 illustrates that the six soil properties of gully heads exhibited a significant increase or decrease with revegetation age. Compared with slope farmland, the soil bulk density (SBD) and soil disintegration rate (SDR) of revegetated gully heads significantly decreased by 5.7–18.6% and 28.8–80.5%, respectively (p < 0.05, Figure 3a,c), while the soil capillary porosity (SCP), saturated soil hydraulic conductivity (SHC), organic matter content (OMC) and water-stable aggregate (WSA) significantly increased by 3.9–13.8%, 17.4–236.2%, 34.2–221.8%, and 27.7–64.4%, respectively (p < 0.05, Figure 3b,d–f). Figure 4 illustrates that the roots of 1–2 mm in slope farmland had the relatively higher root mass density (RMD, 0.20 kg m ) and accounted for 39% of total RMD. Notably, after revegetation, the RMD of >2.0 mm was significantly greater than those of the other three root diameters, and it can account for 40–61% of total RMD. When revegetation age was greater than 3 years, there was a significant di erence in RMD among four root diameter levels (p < 0.05). In addition, we found that the RMD of four root diameters (except for >2.0 mm) showed a non-significant increase in the first three-years and then significantly increased. These results were similar to previous findings regarding the e ects of revegetation on soil properties [45–47]. In fact, the improvement of soil properties of gully heads with revegetation age can be attributed to the accumulation of fresh plant residues in surface soil as well as roots and decomposed root residues in subsurface soil [48]. These materials can be directly transformed into soil organic matter and thus provide energy/carbon sources and nutrients for soil microorganisms [49,50], further promoting the development of soil aggregation and enhancing the cohesion of soil particles [51]. Hence, vegetation restoration would induce the formation of macroaggregates and increase the water stability of aggregates [19,52]. Water 2020, 12, x FOR PEER REVIEW 6 of 15 Wilk test and Levene test, respectively. If the data failed to meet the two conditions, the Kruskal– Wallis test was performed for the above analysis. The interaction effect of flow discharge and revegetation age was detected using a two-way ANOVA. Pearson’s correlation analysis was used to determine linkages among soil properties, root mass density, and soil detachment rate. Relationships among soil detachment rate, soil properties, flow shear stress and restoration age were analyzed by the regression method. The data analyses were conducted in SPSS v. 16.0 statistical software (IBM Corp., Armonk, NY, USA). The figure plotting was conducted by Origin v. 2020 (OriginLab Corp., Northampton, MA, USA). 3. Results and Discussion 3.1. Effect of Revegetation on Soil and Root Properties of Gully Heads Figure 3 illustrates that the six soil properties of gully heads exhibited a significant increase or decrease with revegetation age. Compared with slope farmland, the soil bulk density (SBD) and soil disintegration rate (SDR) of revegetated gully heads significantly decreased by 5.7–18.6% and 28.8– 80.5%, respectively (p < 0.05, Figure 3a,c), while the soil capillary porosity (SCP), saturated soil hydraulic conductivity (SHC), organic matter content (OMC) and water-stable aggregate (WSA) significantly increased by 3.9–13.8%, 17.4–236.2%, 34.2–221.8%, and 27.7–64.4%, respectively (p < 0.05, Figure 3b,d–f). Figure 4 illustrates that the roots of 1–2 mm in slope farmland had the relatively higher −3 root mass density (RMD, 0.20 kg m ) and accounted for 39% of total RMD. Notably, after revegetation, the RMD of >2.0 mm was significantly greater than those of the other three root diameters, and it can account for 40–61% of total RMD. When revegetation age was greater than 3 years, there was a significant difference in RMD among four root diameter levels (p < 0.05). In addition, we found that the RMD of four root diameters (except for >2.0 mm) showed a non- Water 2020, 12, 3301 7 of 16 significant increase in the first three-years and then significantly increased. Figure Figure 3. 3. V Var ariation iation inin soil sopr il operties properties with wr it evegetation h revegetati age. on a Note: ge. No Bar te: means Bar m the eans 95% the confidence 95% confi interval dence interval (95% CI). (95 Di %er ent CI).lower Different case letters lower indicate case letters significant indicate di er sience gnificant among dif di ference erent r am evegetation ong differen ages t revegetati (p < 0.05). on (a) a Soil ges bulk (p <density 0.05). ((SBD); a) Soil (b bulk ) Soil dcapillary ensity (SBD); porosity (b) (SCP); Soil capill (c) Soil ary disintegration porosity (SCP) rate ; (c(SDR); ) Soil Water 2020, 12, x FOR PEER REVIEW 7 of 15 (d) Saturated soil hydraulic conductivity (SHC); (e) Organic matter content (OMC); (f) Water-stable disintegration rate (SDR); (d) Saturated soil hydraulic conductivity (SHC); (e) Organic matter content aggregate (WSA). (OMC); (f) Water-stable aggregate (WSA). Figure 4. Changes in root mass density (a) and its proportion (b) of di erent root diameters with Figure 4. Changes in root mass density (a) and its proportion (b) of different root diameters with revegetation age. Note: Bar means the 95% confidence interval (95% CI). SF refers to the slope farmland. revegetation age. Note: Bar means the 95% confidence interval (95% CI). SF refers to the slope NR3, NR10, NR18, NR25 represents the 3, 10, 18, and 25 years of natural restoration time, respectively. farmland. NR3, NR10, NR18, NR25 represents the 3, 10, 18, and 25 years of natural restoration time, Di erent capital letters for the same restoration age indicate a significant di erence among di erent respectively. Different capital letters for the same restoration age indicate a significant difference root diameters (p < 0.05), and di erent lowercase letters for the same root diameter level indicate a among different root diameters (p < 0.05), and different lowercase letters for the same root diameter significant di erence among di erent revegetation ages (p < 0.05). (a) Root mass density; (b) Mass level indicate a significant difference among different revegetation ages (p < 0.05). (a) Root mass accumulated percent. density; (b) Mass accumulated percent. These results were similar to previous findings regarding the effects of revegetation on soil properties [45–47]. In fact, the improvement of soil properties of gully heads with revegetation age can be attributed to the accumulation of fresh plant residues in surface soil as well as roots and decomposed root residues in subsurface soil [48]. These materials can be directly transformed into soil organic matter and thus provide energy/carbon sources and nutrients for soil microorganisms [49,50], further promoting the development of soil aggregation and enhancing the cohesion of soil particles [51]. Hence, vegetation restoration would induce the formation of macroaggregates and increase the water stability of aggregates [19,52]. 3.2. Effect of Revegetation Age on Soil Detachment of Gully Heads As illustrated in Figure 5a, the Dr of gully heads showed a significant decrease during the 25- year revegetation. This result was not agreed with the conclusion of Wang et al. [16] who stated that soil detachment capacity of sloped lands fluctuated with abandonment time, and the soil detachment capacity of the slope farmland was significantly greater than those of the abandoned farmlands. The difference was mainly attributed to the great difference in erosion environment (e.g., plant type, geomorphological feature, climate) significantly affecting the succession process [36,47]. The mean Dr of slope farmland was 1.6 to 3.0 times greater than those of revegetated gully heads, which indicated that the revegetation played a role in enhancing the soil resistance of gully heads to concentrated flow erosion. Figure 5b shows the Dr of gully heads of slope farmland and four restored grasslands varied with flow discharge. The optimal relationships between Dr and flow discharge were fitted, which can reflect the response of Dr of gully heads to concentrated flow induced by rainstorms of different recurrence intervals. It was found that the response of Dr of slope farmland to flow discharge could be expressed as a power function (y = m × x ), and the n-value was greater than 1, indicating the soil loss of gully heads increases at an increased speed with increasing flow. However, for the restored gully heads, the optimal relationships between Dr and flow discharge could be described by a series Water 2020, 12, 3301 8 of 16 3.2. E ect of Revegetation Age on Soil Detachment of Gully Heads As illustrated in Figure 5a, the D of gully heads showed a significant decrease during the 25-year revegetation. This result was not agreed with the conclusion of Wang et al. [16] who stated that soil detachment capacity of sloped lands fluctuated with abandonment time, and the soil detachment capacity of the slope farmland was significantly greater than those of the abandoned farmlands. Water 2020, 12, x FOR PEER REVIEW 8 of 15 The di erence was mainly attributed to the great di erence in erosion environment (e.g., plant type, geomorphological feature, climate) significantly a ecting the succession process [36,47]. The mean D of linear functions (y = p × x + q), and the p-value decreased with revegetation age, indicating that the of slope farmland was 1.6 to 3.0 times greater than those of revegetated gully heads, which indicated sensitivity of Dr of the gully heads to concentrated flow erosion gradually decreased with increasing that the revegetation played a role in enhancing the soil resistance of gully heads to concentrated restoration age. Besides, the interacted effect of revegetation age and unit width discharge flow erosion. significantly affected Dr (p < 0.001) (Table 2). Figure 5. Change in soil detachment rate of gully heads with restoration age (a), and its relationships Figure 5. Change in soil detachment rate of gully heads with restoration age (a), and its relationships with flow discharge (b). Note: Bar means the 95% confidence interval (95% CI). SF refers to the slope with flow discharge (b). Note: Bar means the 95% confidence interval (95% CI). SF refers to the slope farmland. NR3, NR10, NR18, NR25 represents the 3, 10, 18 and 25 years of revegetation age, farmland. NR3, NR10, NR18, NR25 represents the 3, 10, 18 and 25 years of revegetation age, respectively. The respec diti v er el ent y. lower The di case fferent letters lowercas indicate e le a tte significant rs indicate di er a ence signifi among cant di di ff erence erent revegetation among different ages (revegetati p < 0.05). on a (a) Site ges code; (p < 0.(05) b) Unit . (a) Si width te code dischar ; (b) Unit ge. width discharge. Figure 5b shows the D of gully heads of slope farmland and four restored grasslands varied with r Table 2. Summary of two-way ANOVAs tests. flow discharge. The optimal relationships between D and flow discharge were fitted, which can reflect p- the response of D of gully heads to concentrated flow induced by rainstorms of di erent recurrence Sourc r e SS Df MS F Value intervals. It was found that the response of D of slope farmland to flow discharge could be expressed −4 −5 Revegetation age 1.08 × 10 4 2.69 × 10 151.44 <0.001 as a power function (y = m x ), and the n-value was greater than 1, indicating the soil loss of gully −4 −5 Unit width discharge 1.75 × 10 4 4.38 × 10 246.48 <0.001 heads increases at an increased speed with increasing flow. However, for the restored gully heads, Revegetation age × Unit width −5 −6 the optimal relationships between D and flow discharge could be described by a series of linear r 2.29 × 10 16 1.43 × 10 8.07 <0.001 discharge functions (y = p x + q), and the p-value decreased with revegetation age, indicating that the sensitivity −5 −7 Error 1.33 × 10 75 1.78 × 10 of D of the gully heads to concentrated flow erosion gradually decreased with increasing restoration Total 0.0011 100 age. Besides, the interacted e ect of revegetation age and unit width discharge significantly a ected D (p < 0.001) (Table 2). 3.3. Response of Soil Detachment to Soil Properties Table 2. Summary of two-way ANOVAs tests. Figure 6 showed that Dr was positively correlated with soil bulk density and soil disintegration rate (p < 0.01), but negatively correlated with capillary porosity, saturated hydraulic conductivity, Source SS Df MS F p-Value organic matter and water-stable aggregate of >0.25 mm (p < 0.01). Regression analysis showed that Dr 4 5 Revegetation age 1.08 10 4 2.69 10 151.44 <0.001 increased with soil bulk density as a power function (Figure 7a), which showed an opposite trend 4 5 Unit width discharge 4 246.48 <0.001 1.75 10 4.38 10 with the Wang et al. [13] and Yu et al. [15]. Lower soil bulk density was caused by greater root Revegetation age Unit 5 6 2.29 10 16 1.43 10 8.07 <0.001 physical and soil organisms’ activities, and thus a soil with lower bulk density was harder to be width discharge 5 7 Error 1.33 10 75 1.78 10 detached. Additionally, Dr decreased with capillary porosity as a logarithmic function (Figure 7b), Total 0.0011 100 which was caused by physically binding and chemically bonding effect of root improving soil structure and porosity and hence increasing soil resistance to erosion [28,53]. Soil disintegration rate referred to the dispersion speed of soil contacting with water, which is an important factor determining soil resistance to erosion [14]. In this study, the soil disintegration rate decreased with the revegetation time (Figure 3c) and Dr increased linearly with an increase in soil disintegration rate (Figure 7c). This is attributed to root wedging mechanism preventing soil from detaching that roots can bind soil and tie surface soil layer into strong and stable subsurface soil layer [14,54]. Saturated soil hydraulic conductivity is an integrating parameter for several physical characteristics such as bulk density, porosity, and mechanical composition. The conclusion that the Dr decreased with increasing soil hydraulic conductivity by a power function is reasonable (Figure 7d) because this Water 2020, 12, 3301 9 of 16 3.3. Response of Soil Detachment to Soil Properties Figure 6 showed that D was positively correlated with soil bulk density and soil disintegration rate (p < 0.01), but negatively correlated with capillary porosity, saturated hydraulic conductivity, organic matter and water-stable aggregate of >0.25 mm (p < 0.01). Regression analysis showed that D increased with soil bulk density as a power function (Figure 7a), which showed an opposite trend with the Wang et al. [13] and Yu et al. [15]. Lower soil bulk density was caused by greater root physical and soil organisms’ activities, and thus a soil with lower bulk density was harder to be detached. Additionally, D decreased with capillary porosity as a logarithmic function (Figure 7b), which was caused by physically binding and chemically bonding e ect of root improving soil structure and porosity and hence increasing soil resistance to erosion [28,53]. Soil disintegration rate referred to the dispersion speed of soil contacting with water, which is an important factor determining soil resistance to erosion [14]. In this study, the soil disintegration rate decreased with the revegetation time (Figure 3c) and D increased linearly with an increase in soil disintegration rate (Figure 7c). This is attributed to root wedging mechanism preventing soil from detaching that roots can bind soil and tie surface soil layer into strong and stable subsurface soil layer [14,54]. Saturated soil hydraulic conductivity is an integrating parameter for several physical characteristics such as bulk density, Water 2020, 12, x FOR PEER REVIEW 9 of 15 porosity, and mechanical composition. The conclusion that the D decreased with increasing soil hydraulic conductivity by a power function is reasonable (Figure 7d) because this study and previous study and previous research findings have also indicated that changes in soil bulk density and research findings have also indicated that changes in soil bulk density and porosity influenced soil porosity influenced soil detachment and also were affected by revegetation (e.g., Neves et al. [55]; detachment and also were a ected by revegetation (e.g., Neves et al. [55]; Zhang et al. [56]). A negative Zhang et al. [56]). A negative power function was found between Dr and soil organic matter content power function was found between D and soil organic matter content (Figure 7e). The accumulation (Figure 7e). The accumulation of soil organic matter in soil could promote the formation of aggregate of soil organic matter in soil could promote the formation of aggregate and enhance the cohesion of soil and enhance the cohesion of soil particles [57]. Hence, water-stable aggregate also was an indicator particles [57]. Hence, water-stable aggregate also was an indicator determining soil resistance to flow determining soil resistance to flow erosion [19]. The Dr decreased as a linear function of water-stable erosion [19]. The D decreased as a linear function of water-stable aggregate of >0.25 mm (Figure 7f). aggregate of >0.25 mm (Figure 7f). The results were in agreement with the findings of Li et al. [14]. The results were in agreement with the findings of Li et al. [14]. However, in Wang et al. [13,16] studies, However, in Wang et al. [13,16] studies, no significant relationships were found between Dr and no significant relationships were found between D and organic matter and water-stable aggregate of organic matter and water-stable aggregate of >0.25 mm, probably caused by small variations of the >0.25 mm, probably caused by small variations of the two factors in their studies and di erence in two factors in their studies and difference in land use between their studies and this study land use between their studies and this study (Podwojewski et al. [58]). (Podwojewski et al. [58]). Figure 6. Correlation matrix among soil detachment rate, soil properties, and root mass density. Note: Figure 6. Correlation matrix among soil detachment rate, soil properties, and root mass density. Dr, SBD, SCP, SDR, SHC, OMC, WAS, RMD < 0.5, RMD 0.5–1.0, RMD 1.0–2.0, and RMD > 2.0 refers Note: Dr, SBD, SCP, SDR, SHC, OMC, WAS, RMD < 0.5, RMD 0.5–1.0, RMD 1.0–2.0, and RMD > 2.0 to the soil detachment rate, soil bulk density, soil capillary porosity, soil disintegration rate, saturated refers to the soil detachment rate, soil bulk density, soil capillary porosity, soil disintegration rate, soil hydraulic conductivity, organic matter content, water-stable aggregate, root mass density of <0.5 saturated soil hydraulic conductivity, organic matter content, water-stable aggregate, root mass density mm, root mass density of 0.5–1.0 mm, root mass density of 1.0–2.0 mm, and root mass density of >2.0 of <0.5 mm, root mass density of 0.5–1.0 mm, root mass density of 1.0–2.0 mm, and root mass density mm, respectively. of >2.0 mm, respectively. Figure 7. Relationships between soil detachment rate and soil properties. (a) Soil bulk density (SBD); (b) Soil capillary porosity (SCP); (c) Soil disintegration rate (SDR); (d) Saturated soil hydraulic conductivity (SHC); (e) Organic matter content (OMC); (f) Water-stable aggregate (WSA). Water 2020, 12, x FOR PEER REVIEW 9 of 15 study and previous research findings have also indicated that changes in soil bulk density and porosity influenced soil detachment and also were affected by revegetation (e.g., Neves et al. [55]; Zhang et al. [56]). A negative power function was found between Dr and soil organic matter content (Figure 7e). The accumulation of soil organic matter in soil could promote the formation of aggregate and enhance the cohesion of soil particles [57]. Hence, water-stable aggregate also was an indicator determining soil resistance to flow erosion [19]. The Dr decreased as a linear function of water-stable aggregate of >0.25 mm (Figure 7f). The results were in agreement with the findings of Li et al. [14]. However, in Wang et al. [13,16] studies, no significant relationships were found between Dr and organic matter and water-stable aggregate of >0.25 mm, probably caused by small variations of the two factors in their studies and difference in land use between their studies and this study (Podwojewski et al. [58]). Figure 6. Correlation matrix among soil detachment rate, soil properties, and root mass density. Note: Dr, SBD, SCP, SDR, SHC, OMC, WAS, RMD < 0.5, RMD 0.5–1.0, RMD 1.0–2.0, and RMD > 2.0 refers to the soil detachment rate, soil bulk density, soil capillary porosity, soil disintegration rate, saturated soil hydraulic conductivity, organic matter content, water-stable aggregate, root mass density of <0.5 mm, root mass density of 0.5–1.0 mm, root mass density of 1.0–2.0 mm, and root mass density of >2.0 Water 2020, 12, 3301 10 of 16 mm, respectively. Figure 7. Relationships between soil detachment rate and soil properties. (a) Soil bulk density (SBD); Figure 7. Relationships between soil detachment rate and soil properties. (a) Soil bulk density (b) Soil capillary porosity (SCP); (c) Soil disintegration rate (SDR); (d) Saturated soil hydraulic (SBD); (b) Soil capillary porosity (SCP); (c) Soil disintegration rate (SDR); (d) Saturated soil hydraulic conductivity conductivity (S (SHC); HC); ( (e e) ) Or Org ganic anic matter matter conte content nt (O (OMC); MC); ((ff)) Wa Water ter -stable -stableaggr aggreg egate ate (WSA). (WSA). 3.4. Response of Soil Loss of Gully-Head to Root Traits Significant correlation was found between D and RMD of 0–0.5 mm, 0.5–1.0 mm, 1.0–2.0 mm and >2.0 mm (p < 0.01, Figure 6), of which the RMD of 0–0.5 mm had the highest correlation with D , indicating that roots of each diameter level had the significant impact on soil erosion of gully heads, especially the fibrous root of 0–0.5 mm. Furthermore, the Hill curve could well simulate the relationships between D and RMD of di erent root diameters with R varying from 0.42 to 0.57 (Figure 8). As illustrated in Figure 8, the D showed a rapid decrease when RMD of 0–0.5, 0.5–1.0, 3 3 3 3 1.0–2.0 and >2.0 mm ranged from 0 to 0.25 kg m , 0 to 0.3 kg m , 0 to 0.5 kg m , and 0 to 1.0 kg m , respectively, implying that soil erosion of gully heads could be controlled once vegetation restoration or root growth in soil. Although the roots were limited in density and flexible in early revegetation stage, whereas, roots can contribute to soil cohesion and additional strength, and be crucial in reduction of soil erosion [28,59]. Additionally, root system can bind soil and tie surface soil layer into strong and stable subsurface soil layer [54]. Well-developed root system had great physical binding and chemical bonding e ect that could well bind soil particles and soil aggregates together and enhance soil resistance to erosion [16,28,42]. In addition, judged by fitted eciency (R ), the optimal results were found in RMD of 0–0.5 mm (Figure 8a), indicating that fibrous root of 0–0.5 mm is the optimal root system reducing soil loss of gully heads. However, Li et al. [44] reported that the ability of plant roots to decrease soil erosion mainly depended on the number of fibrous roots <1.0 mm. Shangguan et al. [53] also found a similar result but recommended root surface area density as the root variable. The reason may be that plant species with contrasting root architectures have a di erent erosion reduction e ect [25]. Additionally, Amezketa [60] and Gyssels et al. [61] reported that monocotyledonous plants are superior to dicotyledonous plants and grasses are better than cereals in stabilizing soil aggregates. According to Li et al. [44], bˆ(1/a) can be used as an indicator to compare the e ectiveness of di erent diameter roots in reducing soil erosion. The lower bˆ(1/a), the more e ective the diameter root. The relatively lower bˆ(1/a) values (0.132 and 0.131 kg m ) were found in the roots of 0–0.5 mm and 0.5–1.0 mm than 1.0–2.0 mm and >2.0 mm (Figure 8), indicating that the 0–0.5 mm and 0.5–1.0 mm are the most e ective roots in reducing soil erosion of gully heads. However, De Baets et al. [25] study the e ect of the mixed community of four grasses [Lolium perenne (variety: tove), L. perenne (variety: starlet), Festuca rubra (variety: echo), and F. arundinacea (variety: starlet)] on SDR and found the bˆ(1/a) Water 2020, 12, x FOR PEER REVIEW 10 of 15 3.4. Response of Soil Loss of Gully-Head to Root Traits Significant correlation was found between Dr and RMD of 0–0.5 mm, 0.5–1.0 mm, 1.0–2.0 mm and >2.0 mm (p < 0.01, Figure 6), of which the RMD of 0–0.5 mm had the highest correlation with Dr, indicating that roots of each diameter level had the significant impact on soil erosion of gully heads, especially the fibrous root of 0–0.5 mm. Furthermore, the Hill curve could well simulate the relationships between Dr and RMD of different root diameters with R varying from 0.42 to 0.57 (Figure 8). As illustrated in Figure 8, the Dr showed a rapid decrease when RMD of 0–0.5, 0.5–1.0, 1.0– −3 −3 −3 −3 2.0 and >2.0 mm ranged from 0 to 0.25 kg m , 0 to 0.3 kg m , 0 to 0.5 kg m , and 0 to 1.0 kg m , respectively, implying that soil erosion of gully heads could be controlled once vegetation restoration Water 2020, 12, 3301 11 of 16 or root growth in soil. Although the roots were limited in density and flexible in early revegetation stage, whereas, roots can contribute to soil cohesion and additional strength, and be crucial in value is 0.79 kg m that is greater than that of our study. The result fully indicated that the di erent reduction of soil erosion [28,59]. Additionally, root system can bind soil and tie surface soil layer into plant communities had the markedly di erent influences on reducing soil erosion. The result also strong and stable subsurface soil layer [54]. Well-developed root system had great physical binding suggested us reasonably choosing plant species with di erent root architectures and root diameters for and chemical bonding effect that could well bind soil particles and soil aggregates together and revegetation at gully heads. enhance soil resistance to erosion [16,28,42]. Figure 8. Relationships between relative soil detachment rate and root mass density of different root Figure 8. Relationships between relative soil detachment rate and root mass density of di erent root diameters. (a) Root of 0–0.5 mm; (b) Root of 0.5–1.0 mm; (c) Root of 1.0–2.0 mm; (d) Root > 2.0 mm. diameters. (a) Root of 0–0.5 mm; (b) Root of 0.5–1.0 mm; (c) Root of 1.0–2.0 mm; (d) Root > 2.0 mm. 3.5. E ect of Revegetation Age on Soil Erosion Resistance of Gully Head In addition, judged by fitted efficiency (R ), the optimal results were found in RMD of 0–0.5 mm (Figure 8a), indicating that fibrous root of 0–0.5 mm is the optimal root system reducing soil loss of Rill soil erodibility parameter (K ) and critical shear stress ( ) were employed to characterize the r c gully heads. However, Li et al. [44] reported that the ability of plant roots to decrease soil erosion soil erosion resistance of gully heads [13,21], and were determined by the WEPP model (Equation (4)). mainly depended on the number of fibrous roots <1.0 mm. Shangguan et al. [53] also found a similar The fitted linear function between D and shear stress was illustrated in Figure 9. The slope of the fitted result but recommended root surface area density as the root variable. The reason may be that plant line is equal to the K , and the K of the restored grasslands were 31% to 78.6% less than that of slope r r species with contrasting root architectures have a different erosion reduction effect [25]. Additionally, farmland. In addition, we found that the soil erodibility of 3-year restored grassland rapidly declined Amezketa [60] and Gyssels et al. [61] reported that monocotyledonous plants are superior to by 31% compared with the slope farmland, indicating the short-term revegetation can rapidly reduce dicotyledonous plants and grasses are better than cereals in stabilizing soil aggregates. 1 soil erodibility of gully heads. The K of grasslands in this study ranged from 0.0009 to 0.0029 s m , According to Li et al. [44], b^(1/a) can be used as an indicator to compare the effectiveness of which were less than those reported by Li et al. [14]. Wang et al. [13] found that averaged K of restored different diameter roots in reducing soil eros1ion. The lower b^(1/a), the more effective the diameter lands of abandoned farmland was 0.0024 s m that was close to those of this study. The di erence was mainly caused by di erences in land use, plant species and restoration time. The soil samples were taken from di erent vegetation restoration models (korshinsk peashrub, black locust, Chinese pine and mixed forest of amorpha and Chinese pine) in the study of Wang et al. [13], and the restoration age (37 years) was greater than that of this study (3 to 25 years). Regression analysis found that the K decreased with restoration time in an exponential function and showed a slight change when restored age was greater than 18 years (Figure 10). Water 2020, 12, x FOR PEER REVIEW 11 of 15 −3 root. The relatively lower b^(1/a) values (0.132 and 0.131 kg m ) were found in the roots of 0–0.5 mm and 0.5–1.0 mm than 1.0–2.0 mm and >2.0 mm (Figure 8), indicating that the 0–0.5 mm and 0.5–1.0 mm are the most effective roots in reducing soil erosion of gully heads. However, De Baets et al. [25] study the effect of the mixed community of four grasses [Lolium perenne (variety: tove), L. perenne (variety: starlet), Festuca rubra (variety: echo), and F. arundinacea (variety: starlet)] on SDR and found −3 the b^(1/a) value is 0.79 kg m that is greater than that of our study. The result fully indicated that the different plant communities had the markedly different influences on reducing soil erosion. The result also suggested us reasonably choosing plant species with different root architectures and root diameters for revegetation at gully heads. 3.5. Effect of Revegetation Age on Soil Erosion Resistance of Gully Head Rill soil erodibility parameter (Kr) and critical shear stress (c) were employed to characterize the soil erosion resistance of gully heads [13,21], and were determined by the WEPP model (Equation (4)). The fitted linear function between Dr and shear stress was illustrated in Figure 9. The slope of the fitted line is equal to the Kr, and the Kr of the restored grasslands were 31% to 78.6% less than that of slope farmland. In addition, we found that the soil erodibility of 3-year restored grassland rapidly declined by 31% compared with the slope farmland, indicating the short-term revegetation can rapidly reduce soil erodibility of gully heads. The Kr of grasslands in this study ranged from 0.0009 −1 to 0.0029 s m , which were less than those reported by Li et al. [14]. Wang et al. [13] found that −1 averaged Kr of restored lands of abandoned farmland was 0.0024 s m that was close to those of this study. The difference was mainly caused by differences in land use, plant species and restoration time. The soil samples were taken from different vegetation restoration models (korshinsk peashrub, black locust, Chinese pine and mixed forest of amorpha and Chinese pine) in the study of Wang et al. [13], and the restoration age (37 years) was greater than that of this study (3 to 25 years). Regression analysis found that the Kr decreased with restoration time in an exponential function and showed a Water 2020, 12, 3301 12 of 16 slight change when restored age was greater than 18 years (Figure 10). Figure 9. Relationship between soil detachment rate and shear stress under di erent revegetation ages. (a) SF; (b) NR3; (c) NR10; (d) NR18; (e) NR25. Water 2020, 12, x FOR PEER REVIEW 12 of 15 Figure 9. Relationship between soil detachment rate and shear stress under different revegetation ages. (a) SF; (b) NR3; (c) NR10; (d) NR18; (e) NR25. Figure 10. Relationships between soil erodibility and critical shear stress and revegetation age. Figure 10. Relationships between soil erodibility and critical shear stress and revegetation age. In addition, the c increased with restoration age by a power function (Figure 10). However, the In addition, the  increased with restoration age by a power function (Figure 10). result was inconsistent with the finding of Wang et al. [16] that critical shear stress varied with However, the result was inconsistent with the finding of Wang et al. [16] that critical shear stress restoration age in a nonlinear pattern, reaching the minimum at the restored age of 18. The difference varied with restoration age in a nonlinear pattern, reaching the minimum at the restored age of 18. in the temporal change of critical shear stress between Wang et al. [13] and this study was caused The di erence in the temporal change of critical shear stress between Wang et al. [13] and this study was probably by differences in soil properties and vegetation characteristics of the sampling sites. caused probably by di erences in soil properties and vegetation characteristics of the sampling sites. Compared with slope farmland, the c of the grasslands was improved by 1.2 to 4.0 times, while c of Compared with slope farmland, the  of the grasslands was improved by 1.2 to 4.0 times, while  of c c restored land had a little decrease when restored time was more than 18 years (Figure 10). The result restored land had a little decrease when restored time was more than 18 years (Figure 10). The result further indicated that revegetation can effectively improve the soil erosion resistance of gully head further indicated that revegetation can e ectively improve the soil erosion resistance of gully head to to concentrated flow, and the critical shear stress would reach a stable state after a 18-year concentrated flow, and the critical shear stress would reach a stable state after a 18-year revegetation. revegetation. 4. Conclusions This study was carried out to explore the effect of revegetation age on soil erosion resistance of gully heads in the gully region of the Loess Plateau. The results showed that revegetation significantly improved soil properties and promoted root accumulation of gully heads. The mean Dr of slope farmland was 1.6 to 3.0 times greater than those of revegetated gully heads. The revegetation can effectively weaken the sensitivity of soil erosion of the gully heads to concentrated flow. The Dr of gully heads was positively related to bulk density and disintegration rate and negatively related to soil capillary porosity, saturated soil hydraulic conductivity, organic matter content, and water- stable aggregate. Roots of 0–0.5 mm and 0.5–1.0 mm were the most effective roots in reducing soil erosion of gully head, and the native plant species with rich root of 0.5–1.0 mm and 0–0.5 mm were recommended as the first choice for revegetation to restrain gully headcut erosion. Revegetation can reduce soil erodibility of gully heads by 31% to 78.6% and improve the critical shear stress by 1.2 to 4.0 times. This study allows us to better evaluate soil vulnerability of gully head to concentrated flow erosion during revegetation. Further studies are needed to quantify the effect of the different combinations of vegetation types with different root architecture types on soil erosion resistance of gully heads. Author Contributions: Conceptualization, Z.C. and M.G.; data curation, Z.C.; formal analysis, Z.C. and M.G.; funding acquisition, W.W.; investigation Z.C. and M.G.; methodology, Z.C.; project administration, W.W.; resources, W.W.; software, Z.C.; supervision, W.W.; validation, W.W.; writing—original draft, Z.C.; writing— review and editing, M.G. and W.W. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported by the National Natural Science Foundation of China (42077079; 41571275). Water 2020, 12, 3301 13 of 16 4. Conclusions This study was carried out to explore the e ect of revegetation age on soil erosion resistance of gully heads in the gully region of the Loess Plateau. The results showed that revegetation significantly improved soil properties and promoted root accumulation of gully heads. The mean D of slope farmland was 1.6 to 3.0 times greater than those of revegetated gully heads. The revegetation can e ectively weaken the sensitivity of soil erosion of the gully heads to concentrated flow. The D of gully heads was positively related to bulk density and disintegration rate and negatively related to soil capillary porosity, saturated soil hydraulic conductivity, organic matter content, and water-stable aggregate. Roots of 0–0.5 mm and 0.5–1.0 mm were the most e ective roots in reducing soil erosion of gully head, and the native plant species with rich root of 0.5–1.0 mm and 0–0.5 mm were recommended as the first choice for revegetation to restrain gully headcut erosion. Revegetation can reduce soil erodibility of gully heads by 31% to 78.6% and improve the critical shear stress by 1.2 to 4.0 times. This study allows us to better evaluate soil vulnerability of gully head to concentrated flow erosion during revegetation. Further studies are needed to quantify the e ect of the di erent combinations of vegetation types with di erent root architecture types on soil erosion resistance of gully heads. Author Contributions: Conceptualization, Z.C. and M.G.; data curation, Z.C.; formal analysis, Z.C. and M.G.; funding acquisition, W.W.; investigation Z.C. and M.G.; methodology, Z.C.; project administration, W.W.; resources, W.W.; software, Z.C.; supervision, W.W.; validation, W.W.; writing—original draft, Z.C.; writing—review and editing, M.G. and W.W. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported by the National Natural Science Foundation of China (42077079; 41571275). Acknowledgments: All authors thank anonymous reviewers for insightful comments on the original manuscript. Conflicts of Interest: The authors declare no conflict of interest. References 1. Mohammed, S.; Al-Ebraheem, A.; Holb, I.J.; Alsafadi, K.; Dikkeh, M.; Pham, Q.B.; Linh, N.T.T.; Szabo, S. Soil management e ects on soil water erosion and runo in central Syria—A comparative evaluation of general linear model and random forest regression. Water 2020, 12, 2529. [CrossRef] 2. Hao, H.X.; Cheng, L.; Guo, Z.L.; Wang, L.; Shi, Z.H. Plant community characteristics and functional traits as drivers of soil erodibility mitigation along a land degradation gradient. Land Degrad. Dev. 2020, 31, 1851–1863. [CrossRef] 3. Zarris, D.; Lykoudi, E.; Panagoulia, D. Assessing the impacts of sediment yield on the sustainability of major hydraulic systems. In Proceedings of the International Conference “Protection and Restoration of the Environment VIII”, Chania, Greece, 3–7 July 2006. 4. Brazier, R.E.; Beven, K.J.; Freer, J.; Rowan, J.S. Equifinality and uncertainty in physically based soil erosion models: Application of the GLUE methodology to WEPP–the water erosion prediction project–for sites in the UK and USA. Earth Surf. Process. Landf. 2000, 25, 825–845. [CrossRef] 5. Kinnell, P.; Risse, L.M. USLE-M: Empirical modeling rainfall erosion through runo and sediment concentration. Soil Sci. Soc. Am. J. 1998, 62, 1667–1672. [CrossRef] 6. Nearing, M.A.; Foster, G.R.; Lane, L.J.; Finkner, S.C. A process-based soil erosion model for USDA-water erosion prediction project technology. Trans. ASAE 1989, 32, 1587–1593. [CrossRef] 7. Efthimiou, N.; Lykoudi, E.; Panagoulia, D.; Karavitis, C. Assessment of soil susceptibility to erosion using the EPM and RUSLE models: The case of Venetikos river catchment. Glob. Nest. J. 2016, 18, 164–179. 8. Guo, W.Z.; Kang, H.L.; Wang, W.L.; Guo, M.M.; Chen, Z.X. Erosion-reducing e ects of revegetation and fish-scale pits on steep spoil heaps under concentrated runo on the Chinese Loess Plateau. Land Degrad Dev. 2020. [CrossRef] 9. Gramlich, A.; Stoll, S.; Stamm, C.; Walter, T.; Prasuhn, V. E ects of artificial land drainage on hydrology, nutrient and pesticide fluxes from agricultural fields—A review. Agric. Ecosyst. Environ. 2018, 266, 84–99. [CrossRef] 10. Kervroëdan, L.; Armand, R.; Saunier, M.; Ouvry, J.; Faucon, M. Plant functional trait e ects on runo to design herbaceous hedges for soil erosion control. Ecol. Eng. 2018, 118, 143–151. [CrossRef] Water 2020, 12, 3301 14 of 16 11. Sun, L.; Zhang, G.H.; Luan, L.L.; Liu, F. Temporal variation in soil resistance to flowing water erosion for soil incorporated with plant litters in the Loess Plateau of China. Catena 2016, 145, 239–245. [CrossRef] 12. Chang, E.; Li, P.; Li, Z.; Su, Y.; Zhang, Y.; Zhang, J.; Liu, Z.; Li, Z. The impact of vegetation successional status on slope runo erosion in the Loess Plateau of China. Water 2019, 11, 2614. [CrossRef] 13. Wang, B.; Zhang, G.H.; Shi, Y.Y.; Zhang, X.C. Soil detachment by overland flow under di erent vegetation restoration models in the Loess Plateau of China. Catena 2014, 116, 51–59. [CrossRef] 14. Li, Z.W.; Zhang, G.H.; Geng, R.; Wang, H.; Zhang, X.C. Land use impacts on soil detachment capacity by overland flow in the Loess Plateau, China. Catena 2015, 124, 9–17. [CrossRef] 15. Yu, Y.C.; Zhang, G.H.; Geng, R.; Sun, L. Temporal variation in soil detachment capacity by overland flow under four typical crops in the Loess Plateau of China. Biosyst. Eng. 2014, 122, 139–148. [CrossRef] 16. Wang, B.; Zhang, G.H.; Shi, Y.Y.; Zhang, X.C.; Ren, Z.P.; Zhu, L.J. E ect of natural restoration time of abandoned farmland on soil detachment by overland flow in the Loess Plateau of China. Earth Surf. Process. Landf. 2013, 38, 1725–1734. [CrossRef] 17. Scherer, U.; Zehe, E.; Tr Bing, K.; Kai, G. Prediction of soil detachment in agricultural loess catchments: Model development and parameterisation. Catena 2012, 90, 63–75. [CrossRef] 18. Zhang, G.H.; Tang, M.K.; Zhang, X.C. Temporal variation in soil detachment under di erent land uses in the Loess Plateau of China. Earth Surf. Process. Landf. 2010, 34, 1302–1309. [CrossRef] 19. Bernard, B.; Roose, E. Aggregate stability as an indicator of soil susceptibility to runo and erosion; validation at several levels. Catena 2002, 47, 133–149. 20. Knapen, A.; Poesen, J.; Govers, G.; Baets, S.D. The e ect of conservation tillage on runo erosivity and soil erodibility during concentrated flow. Hydrol. Process. 2008, 22, 1497–1508. [CrossRef] 21. Guo, M.M.; Wang, W.L.; Wang, T.C.; Wang, W.X.; Kang, H.L. Impacts of di erent vegetation restoration options on gully head soil resistance and soil erosion in loess tablelands. Earth Surf. Process. Landf. 2020, 45, 1038–1050. [CrossRef] 22. Lazarovitch, N.; Vanderborght, J.; Jin, Y.; van Genuchten, M.T. The root zone: Soil physics and beyond. Vadose Zone J. 2018, 17, 1–6. [CrossRef] 23. Panagoulia, D. Hydrological modelling of a medium-size mountainous catchment from incomplete meteorological data. J. Hydrol. 1992, 137, 279–310. [CrossRef] 24. Falkenmark, M. Land and water integration and river basin management. FAO Land Water Bull. 1995, 1, 15–16. 25. Baets, S.D.; Poesen, J. Empirical models for predicting the erosion-reducing e ects of plant roots during concentrated flow erosion. Geomorphology 2010, 118, 425–432. [CrossRef] 26. Guo, M.M.; Wang, W.L.; Shi, Q.H.; Chen, T.D.; Kang, H.L.; Li, J.M. An experimental study on the e ects of grass root density on gully headcut erosion in the gully region of China’s Loess Plateau. Land Degrad. Dev. 2019, 30, 2107–2125. [CrossRef] 27. Khanal, A.; Fox, G.A. Detachment characteristics of root-permeated soils from laboratory jet erosion tests. Ecol. Eng. 2017, 100, 335–343. [CrossRef] 28. De Baets, S.; Poesen, J.; Gyssels, G.; Knapen, A.; Adili, A.A. E ects of grass roots on the erodibility of topsoils during concentrated flow. Geomorphology 2006, 76, 54–67. [CrossRef] 29. Mamo, M.; Bubenzer, G.D. Detachment rate, soil erodibility, and soil strength as influenced by living plant roots part II: Field study. Trans. ASAE 2001, 5, 1175–1181. [CrossRef] 30. Mamo, M.; Bubenzer, G.D. Detachment rate, soil erodibility, and soil strength as influenced by living plant roots part I: Laboratory study. Trans. ASAE 2001, 5, 1167–1174. [CrossRef] 31. Baets, S.D.; Poesen, J.; Knapen, A.; Galindo, P. Impact of root architecture on the erosion-reducing potential of roots during concentrated flow. Earth Surf. Process. Landf. 2007, 32, 1323–1345. [CrossRef] 32. Shi, Q.; Wang, W.; Guo, M.; Chen, Z.; Feng, L.; Zhao, M.; Xiao, H. The impact of flow discharge on the hydraulic characteristics of headcut erosion processes in the gully region of the Loess Plateau. Hydrol. Process. 2019, 34, 718–729. [CrossRef] 33. Guo, M.M.; Wang, W.L.; Kang, H.L.; Yang, B. Changes in soil properties and erodibility of gully heads induced by vegetation restoration on the Loess Plateau, China. J. Arid Land 2018, 10, 712–725. [CrossRef] 34. Fu, B.; Chen, L.; Ma, K.; Zhou, H.; Wang, J. The relationships between land use and soil conditions in the hilly area of the loess plateau in northern Shaanxi, China. Catena 2000, 39, 69–78. [CrossRef] Water 2020, 12, 3301 15 of 16 35. Xiong, Y.; Zhou, J.Z.; Chen, L.; Jia, B.J.; Sun, N.; Tian, M.Q.; Hu, G.H. Land use pattern and vegetation cover dynamics in the three Gorges Reservoir (TGR) intervening Basin. Water 2020, 12, 2036. [CrossRef] 36. Jiao, J.Y.; Tzanopoulos, J.; Xofis, P.; Bai, W.J.; Ma, X.H.; Mitchley, J. Can the study of natural vegetation succession assist in the control of soil erosion on abandoned croplands on the Loess Plateau, China? Restor. Ecol. 2007, 15, 391–399. [CrossRef] 37. Wang, B.; Liu, G.B.; Xue, S.; Zhu, B.B. Changes in soil physico-chemical and microbiological properties during natural succession on abandoned farmland in the Loess Plateau. Environ. Earth Sci. 2011, 62, 915–925. [CrossRef] 38. Chen, Y.P.; Wang, K.B.; Lin, Y.S.; Shi, W.Y.; Song, Y.; He, X.H. Balancing green and grain trade. Nat. Geosci. 2015, 8, 739–741. [CrossRef] 39. Kompani-Zare, M.; Soufi, M.; Hamzehzarghani, H.; Dehghani, M. The e ect of some watershed, soil characteristics and morphometric factors on the relationship between the gully volume and length in Fars Province, Iran. Catena 2011, 86, 150–159. [CrossRef] 40. Guo, M.M.; Wang, W.L.; Kang, H.L.; Yang, B.; Li, J.M. Changes in soil properties and resistance to concentrated flow across a 25-year passive restoration chronosequence of grasslands on the Chinese Loess Plateau. Restor. Ecol. 2020, 28, 104–114. [CrossRef] 41. Luk, S.H.; Merz, W. Use of the salt tracing technique to determine the velocity of overland flow. Sci. Total Environ. 1992, 5, 289–301. 42. Li, Y.; Xu, X.Q.; Zhu, X.M.; Tian, J.Y. E ectiveness of plant roots on increasing the soil permeability on the Loess Plateau. Chin. Sci. Bull. 1992, 37, 1735–1738. 43. Hill, A.V. The possible e ects of the aggregation of the molecules of haemoglobin on its dissociation curves. J. Physiol. 1910, 40, 4–7. 44. Li, Y.; Zhu, X.M.; Tian, J.Y.; Al-Ebraheem, A. E ectiveness of plant roots to increase the anti-scouribility of soil on the Loess Plateau. Chin. Sci. Bull. 1991, 36, 2077–2082. 45. Gros, R.; Jocteur Monrozier, L.; Bartoli, F.; Chotte, J.L.; Faivre, P. Relationships between soil physico-chemical properties and microbial activity along a restoration chronosequence of alpine grasslands following ski run construction. Appl. Soil Ecol. 2004, 27, 7–22. [CrossRef] 46. Li, Y.Y.; Shao, M.A. Change of soil physical properties under long-term natural vegetation restoration in the Loess Plateau of China. J. Arid. Environ. 2006, 64, 77–96. [CrossRef] 47. Jiao, F.; Wen, Z.M.; An, S.S. Changes in soil properties across a chronosequence of vegetation restoration on the Loess Plateau of China. Catena 2011, 86, 110–116. [CrossRef] 48. Zhao, Y.G.; Wu, P.T.; Zhao, S.W.; Feng, H. Variation of soil infiltrability across a 79-year chronosequence of naturally restored grassland on the Loess Plateau, China. J. Hydrol. 2013, 504, 94–103. [CrossRef] 49. Six, J.; Elliott, E.T. Aggregation and soil organic matter accumulation in cultivated and native grassland soils. Soil Sci. Soc. Am. J. 1998, 62, 1367–1377. [CrossRef] 50. Six, J.; Bossuyt, H.; Degryze, S.; Denef, K. A history of research on the link between (micro)aggregates, soil biota, and soil organic matter dynamics. Soil Tillage Res. 2004, 79, 7–31. [CrossRef] 51. Six, J.; Paustian, K.; Elliott, E.T.; Combrink, C. Soil structure and organic matter I. distribution of aggregate-size classes and aggregate-associated carbon. Soil Sci. Soc. Am. J. 2000, 64, 681–689. [CrossRef] 52. Kong, A.Y.Y.; Six, J.; Bryant, D.C.; Denison, R.F.; Kessel, C.V. The relationship between carbon input, aggregation, and soil organic carbon stabilization in sustainable cropping systems. Soil Sci. Soc. Am. J. 2005, 69. [CrossRef] 53. Shangguan, Z.P.; Zheng, C.Z. Soil anti-scouribility enhanced by plant roots. J. Integr. Plant Biol. 2005, 47, 676–682. 54. Adili, A.A.; Azzam, R.; GiovanniSpagnoli, S.J. Strength of soil reinforced with fiber materials (Papyrus). Soil Mech. Found. Eng. 2012, 48, 241–247. [CrossRef] 55. Neves, C.S.V.J.; Feller, C.; Guimarães, M.F.; Medina, C.C.; Tavares Filho, J.; Fortier, M. Soil bulk density and porosity of homogeneous morphological units identified by the cropping profile method in clayey oxisols in Brazil. Soil Tillage Res. 2003, 71, 109–119. [CrossRef] 56. Zhang, B.J.; Zhang, G.H.; Yang, H.Y.; Wang, H. Soil resistance to flowing water erosion of seven typical plant communities on steep gully slopes on the Loess Plateau of China. Catena 2019, 173, 375–383. [CrossRef] Water 2020, 12, 3301 16 of 16 57. Fattet, M.; Fu, Y.; Ghestem, M.; Ma, W.; Foulonneau, M.; Nespoulous, J.; Le Bissonnais, Y.; Stokes, A. E ects of vegetation type on soil resistance to erosion: Relationship between aggregate stability and shear strength. Catena 2011, 87, 60–69. [CrossRef] 58. Podwojewski, P.; Orange, D.; Jouquet, P.; Valentin, C.; Nguyen, V.T.; Janeau, J.L.; Tran, D.T. Land-use impacts on surface runo and soil detachment within agricultural sloping lands in Northern Vietnam. Catena 2008, 74, 109–118. [CrossRef] 59. Gyssels, G.; Poesen, J. The importance of plant root characteristics in controlling concentrated flow erosion rates. Earth Surf. Process. Landf. 2003, 28, 371–384. [CrossRef] 60. Amézketa, E. Soil aggregate stability: A review. J. Integr. Plant Biol. 1999, 14, 83–151. [CrossRef] 61. Gyssels, G.; Poesen, J.; Bochet, E.; Li, Y. Impact of plant roots on the resistance of soils to erosion by water: A review. Prog. Phys. Geogr. 2005, 29, 189–217. [CrossRef] Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional aliations. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Water Multidisciplinary Digital Publishing Institute

Variations in Soil Erosion Resistance of Gully Head Along a 25-Year Revegetation Age on the Loess Plateau

Water , Volume 12 (12) – Nov 24, 2020

Loading next page...
 
/lp/multidisciplinary-digital-publishing-institute/variations-in-soil-erosion-resistance-of-gully-head-along-a-25-year-pANzKkQKMp

References (60)

Publisher
Multidisciplinary Digital Publishing Institute
Copyright
© 1996-2020 MDPI (Basel, Switzerland) unless otherwise stated Disclaimer The statements, opinions and data contained in the journals are solely those of the individual authors and contributors and not of the publisher and the editor(s). MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Terms and Conditions Privacy Policy
ISSN
2073-4441
DOI
10.3390/w12123301
Publisher site
See Article on Publisher Site

Abstract

water Article Variations in Soil Erosion Resistance of Gully Head Along a 25-Year Revegetation Age on the Loess Plateau 1 1 , 2 1 , Zhuoxin Chen , Mingming Guo and Wenlong Wang * State Key Laboratory of Soil Erosion and Dryland Farming on the Loess Plateau, Institute of Water and Soil Conservation, Northwest A&F University, Yangling 712100, China; xiyu.zxchen@foxmail.com (Z.C.); guomingming@iga.ac.cn (M.G.) Key Laboratory of Mollisols Agroecology, Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Harbin 150081, China * Correspondence: wlwang@nwafu.edu.cn Received: 10 October 2020; Accepted: 19 November 2020; Published: 24 November 2020 Abstract: The e ects of vegetation restoration on soil erosion resistance of gully head, along a revegetation age gradient, remain poorly understood. Hence, we collected undisturbed soil samples from a slope farmland and four grasslands with di erent revegetation ages (3, 10, 18, 25 years) along gully heads. Then, these samples were used to obtain soil detachment rate of gully heads by the hydraulic flume experiment under five unit width flow discharges (2–6 m h). The results revealed that soil properties were significantly ameliorated and root density obviously increased in response to restoration age. Compared with farmland, soil detachment rate of revegetated gully heads decreased 35.5% to 66.5%, and the sensitivity of soil erosion of the gully heads to concentrated flow decreased with revegetation age. The soil detachment rate of gully heads was significantly related to the soil bulk density, soil disintegration rate, capillary porosity, saturated soil hydraulic conductivity, organic matter content and water stable aggregate. The roots of 0–0.5 and 0.5–1.0 mm had the highest benefit in reducing soil loss of gully head. After revegetation, soil erodibility of gully heads decreased 31.0% to 78.6%, and critical shear stress was improved by 1.2 to 4.0 times. The soil erodibility and critical shear stress would reach a stable state after an 18-years revegetation age. These results allow us to better evaluate soil vulnerability of gully heads to concentrated flow erosion and the eciency of revegetation. Keywords: soil erosion; gully erosion; vegetation restoration; soil erodibility; land use 1. Introduction Soil erosion is recognized as a global environmental problem, which severely damages infrastructure, causes land degradation and water pollution, and threatens the safety of human production and life [1–3]. In the past few decades, many scholars have made many e orts to study the process and mechanism of soil erosion, establish many soil erosion prediction models and try to control soil erosion [4–7]. At present, a set of soil erosion control measures system integrating engineering measures, agricultural measures, and biological measures has been formed [8–10], especially vegetation measures play an extremely important role in soil erosion control [11,12]. Previous studies have shown that revegetation can e ectively reduce soil erosion. For example, Wang et al. [13] found that soil detachment capacity of abandoned farmland was 1.02 to 2.29 times greater than four restored lands. Li et al. [14] reported that the ratios of the soil detachment capacity of cropland to those of orchard, shrubland, woodland, grassland, and wasteland were 7.14, 12.29, 25.78, 28.45, and 46.43, respectively. The improvement of soil erosion resistance by revegetation is mainly controlled Water 2020, 12, 3301; doi:10.3390/w12123301 www.mdpi.com/journal/water Water 2020, 12, 3301 2 of 16 by the combination of soil properties and root traits [15–18]. In terms of soil properties, many studies have verified that the revegetation significantly a ects soil erosion by changing the soil bulk density, organic matter content, and water-stable aggregate [19–21]. Furthermore, the vegetation root zone is the dynamic interface of soil–plant–atmosphere continuum in partitioning rain and irrigation water into evaporation, transpiration, runo , and deep drainage [22,23], but is also the home of “green water” which is the source of plant nutrition [24]. Especially, the vegetation root systems also play a great role in protecting soil against flow scouring by a ecting soil water movement [25,26]. Root-permeated soils exhibited lower erosion rates primarily through increasing the required shear stress before detachment [27]. Moreover, root growth can bind and bond soil particles and aggregate, thus, enhances soil resistance to erosion [28]. Some root parameters, for example root biomass, length density, and surface area density, were used to estimate the e ect of root on soil detachment [21,28–31]. De Baets and Poesen [25] found that soil detachment rate reduced exponentially with increasing root biomass. Some studies also showed that soil detachment was related to root architecture and fibrous root was more e ective than tap root in reducing soil loss [24,31]. However, the most of previous studies only focus on the impact of revegetation on soil erosion resistance of hillslopes. In the watershed dominated by gully erosion, the gully head is the main source of soil erosion, but the e ect of revegetation on soil erosion resistance of gully heads remains unclear. Therefore, there is a strong need to understand the e ect of revegetation on soil erosion of gully head by concentrated flow to develop a more reasonable vegetation model. Notably, in the gully region of the Loess Plateau, about 63% of total runo is generated from the loess tableland with a gentle slope of 1–5 , which can initiate gully headcut erosion and contribute 86.3% of total sediment [32]. The gully headcut erosion by concentrated flow became the main sediment resource. At present, the gully headcut erosion was controlled e ectively due to the implementation of a series of control measures (e.g., the “Three Protection Belts” and the “Green for Grain” project), which, to some extent, was attributed to the fact that the revegetation improves the soil resistance of gully heads to concentrated flow [21,33]. Since some ecological restoration projects were conducted, land use has changed dramatically in the Loess Plateau [34]. Hence, the land use has changed, and the natural succession of vegetation was promoted [35]. With progression in natural restoration of grassland, soil physical and chemical properties and vegetation characteristics (e.g., coverage, community structure, species composition and diversity, and root diameter, density, and diameter distribution) varied greatly [36,37]. These changes would result in dynamic variations in soil erosion resistance. However, the response of soil erosion resistance to vegetation succession process mainly focused on the hillslope in the hilly-gully region of Loess Plateau [15,16], and few studies were conducted to explore the response of soil resistance of gully heads by concentrated flow to vegetation succession process. Therefore, to evaluate the e ect of revegetation process on soil erosion resistance of gully heads and optimize revegetation measures for controlling gully headcut erosion in the gully region of the Loess Plateau, we selected four grasslands with di erent revegetation ages (3, 10, 18, 25 years) along gully heads with the slope farmland as the control. This study aimed to (1) quantify the e ect of revegetation age on soil detachment by concentrated flow, (2) clear the relationships between soil detachment rate and soil and root properties, and (3) confirm the dynamic variation in soil erosion resistance of gully head with revegetation age. 2. Material and Method 2.1. Study Area The study was conducted in the Nanxiaohegou watershed in the Xifeng Research Station of Soil 0  0  0  0 2 and Water Conservation (35 41 –35 44 N, 107 30 –107 37 E). The watershed has an area of 36.3 km and altitudes ranging from 1050 to 1423 m above mean sea level (Figure 1) in the typical gully region of the Loess Plateau. The climate is temperate continental semiarid. The mean temperature is 10 C, Water 2020, 12, x FOR PEER REVIEW 3 of 15 Water 2020, 12, 3301 3 of 16 the characteristic of annual variation and uneven distribution during the year. In the form of short and the frost-free period is 160–180 days. Annual precipitation is approximately 523 mm, which has heavy storms, 58.8% of the rainfall occurs from July to September. The soil type is yellow loamy soil. the characteristic of annual variation and uneven distribution during the year. In the form of short The original vegetation has disappeared due to human activities. Gully headcut erosion is the main heavy storms, 58.8% of the rainfall occurs from July to September. The soil type is yellow loamy resource of sediment yield in the watershed. Since the 1970s, some soil and water conservation soil. The original vegetation has disappeared due to human activities. Gully headcut erosion is the projects, for example the “three protection belts” project and the “Green for Grain” project and so on, main resource of sediment yield in the watershed. Since the 1970s, some soil and water conservation were implemented to control soil and water loss, and the vegetation cover of the Loess Plateau projects, for example the “three protection belts” project and the “Green for Grain” project and so on, were implemented to control soil and water loss, and the vegetation cover of the Loess Plateau increased increased to 59.6% in 2013. Additionally, the land use has undergone tremendous changes [38]. These to 59.6% in 2013. Additionally, the land use has undergone tremendous changes [38]. These e orts also efforts also effectively stabilized the gully heads and thus contained the gully headcut erosion [33]. e ectively stabilized the gully heads and thus contained the gully headcut erosion [33]. At present, −2 −1 At present, the annual soil erosion module is effectively controlled at the level of 2440 t km a in 2 1 the annual soil erosion module is e ectively controlled at the level of 2440 t km a in the study area, the study area, and the vegetation communities comprise mainly planted forests and shrubs and and the vegetation communities comprise mainly planted forests and shrubs and native secondary native secondary herbaceous plants [21]. herbaceous plants [21]. Figure 1. Location of the study area on the Loess Plateau and the location of sampling sites in the Figure 1. Location of the study area on the Loess Plateau and the location of sampling sites in the Nanxiaohegou watershed. Nanxiaohegou watershed. 2.2. Sampling Sites Selection 2.2. Sampling Sites Selection During our investigation of gully heads, some cracks developed near gully heads. Kompani-Zare et al. [39] and Guo et al. [21] stated that soil samples from 0 to 30 cm depth near the gully heads During our investigation of gully heads, some cracks developed near gully heads. Kompani- (the distance was less than 5 m) can represent soil properties of the gully heads. Therefore, in consideration Zare et al. [39] and Guo et al. [21] stated that soil samples from 0 to 30 cm depth near the gully heads of collapsibility and vertical joints development of loess, the sampling plots were established about 1.0 m (the distance was less than 5 m) can represent soil properties of the gully heads. Therefore, in meters from gully heads to ensure safety. As a result, four natural restoration grasslands with different consideration of collapsibility and vertical joints development of loess, the sampling plots were ages (3, 10, 18, 25 years) were selected (Figure 1). The natural restoration age was confirmed by consulting the village elders and scientists at the scientific experimental station. The slope aspect and gradient, established about 1.0 m meters from gully heads to ensure safety. As a result, four natural restoration elevation, soil type, and previous farming practices of the selected sites were similar to minimize the grasslands with different ages (3, 10, 18, 25 years) were selected (Figure 1). The natural restoration effects of these factors on the experimental results. For comparison, one corn-planted farmland site, with a age was confirmed by consulting the village elders and scientists at the scientific experimental station. topography similar to that of the grasslands, was selected as a control. The basic information of the five The slope aspect and gradient, elevation, soil type, and previous farming practices of the selected selected sites is listed in Table 1. sites were similar to minimize the effects of these factors on the experimental results. For comparison, Table 1. Basic information of the selected five sampling sites. one corn-planted farmland site, with a topography similar to that of the grasslands, was selected as Restoration Dominant a control. The basic information of the five selected sites is listed in Table 1. Site Code Slope ( ) Coverage (%) Altitude (m) Main Companion Species Age (yr) Communities SF 0 2.4 — 1420 Zea mays Setaria viridis NR3 3 2.9 72.3 1405 Artemisia capillaris Artemisia sacrorum Table 1. Basic information of the selected five sampling sites. NR10 10 2.4 80.8 1401 Artemisia sacrorum Artemisia capillaris NR18 18 3.2 93.4 1390 Artemisia sacrorum Artemisia capillaris Artemisia sacrorum + Site Restoration Slope Coverage Altitude Dominant Main Companion NR25 25 3.1 91.2 1380 Bothriochloa ischaemum Lespedeza daurica Code Age (yr) (° ) (%) (m) Communities Species SF 0 2.4 — 1420 Zea mays Setaria viridis Artemisia NR3 3 2.9 72.3 1405 Artemisia sacrorum capillaris Artemisia NR10 10 2.4 80.8 1401 Artemisia capillaris sacrorum Artemisia NR18 18 3.2 93.4 1390 Artemisia capillaris sacrorum Bothriochloa Artemisia sacrorum NR25 25 3.1 91.2 1380 ischaemum + Lespedeza daurica Water 2020, 12, 3301 4 of 16 2.3. Sampling and Measurement of Soil and Root In this study, seven soil and root property parameters including soil bulk density, capillary porosity, soil disintegration rate, soil water-stable aggregate, soil saturated hydraulic conductivity, organic matter content, and root mass density were measured. Firstly, three repeated sampling plots (5 m 5 m) were established in each of gully head sampling sites with the litter layer removed, and topsoil samples (0–30 cm) were collected. Then, three cutting rings (200 cm ) were used to randomly collect soil samples in each plot, and a total of nine samples were oven-dried at 105 C for 24 h to determine the soil bulk density of each gully head site. Similarly, the other 9 soil samples were also collected by cutting rings of 200 cm to determine the soil saturated hydraulic conductivity by applying the constant water head test method. Three cutting rings (100 cm ) were used to collect soil samples for the measurement of soil capillary porosity [33]. Three man-made steel cubical boxes (5 cm in length) were used to collect soil samples for measuring soil disintegration rate by using a disintegration box [14,40]. Lastly, the other three samples were randomly collected in each plot to form a mixed sample. A total of 45 mixed samples were obtained and used for laboratory analyses of organic matter content and water-stable aggregate and its stability. These mixed soil samples were air-dried at room temperature, with large roots and organic residues manually removed. Sieves with apertures (0.25, 0.5, 1.0, 2.5, and 5.0 mm) were used to test the water-stable aggregate. The potassium dichromate external heating method was used to measure the soil organic matter content. 2.4. Hydraulic Flume Experiments A hydraulic flume experiment was conducted to determine the soil resistance to concentrated flow upstream gully heads (Figure 2). The size of the flume was 2.0 m long and 0.15 m wide similar to the one used by De Baets et al. [28,31], which was enough to make water flow along the slope soil. An opening (0.5 m length and 0.1 m wide) was set at the bottom of the flume, and a metal sample box with the same size was used to collect undisturbed soil samples so that the surface of the soil sample was at the same level of the flume surface. The space between sampling box and flume edge sealed with painter ’s mastic to prevent boundary e ects. According to the study of Guo et al. [40], the flume experiment was carried out under five di erent unit width flow discharges of 2, 3, 4, 5, 6 3 1 m s , and thus, a total of 100 samples (5 sites 5 flow discharges 4 replications) were collected to measure soil resistance of gully heads. To simulate real flow generation conditions, the soil should be saturated by using a watering pot before experiment. During the experiment, a portable flow meter instrument (LS300-A) with 1.5% accuracy was used to measure flow velocity which was regarded as the flow velocity scoring soil area. Runo and sediment samples were collected with sampling tanks, and the sampling time was recorded. The measured flow velocity was modified according to flow regime [41]. Sampled sediment was oven-dried at 105 C for 24 h to determine the soil loss amount 2 1 (SLA, kg). Thus, the soil detachment rate (D , kg m s ) could be calculated as follows: please check font size, please check all reference citation SLA D = (1) AT where SLA is the oven-dry mass of every sediment sample (kg), T is the experimental period (s) and A is the soil sample area (m ). In addition, the relative soil detachment rate (RD ) was calculated as the ratio between D for the root-permeated soil samples and that for the farmland topsoil samples, tested at the same condition [28]. Water 2020, 12, x FOR PEER REVIEW 5 of 15 3 −1 where h is flow depth (m), q is the flow discharge (m s ), w is the width of the flume (m), v is the −1 −3 −2 mean flow velocity (m s ).  is the water mass density (kg m ), g is the gravity constant (m s ), and −1 S is the slope steepness (m m ). After each scouring test, a steel cubical box (10 cm in length) was used to take soil sample in the center of soil area of scouring flume, and the sample was soaked in tap water for about one hour to increase the dispersion of soil and then were placed on a 0.25 mm sieve and washed with tap water using low-pressure head. The living roots, plant debris and some pebbles were left on the sieve. Only the living roots were picked out carefully using tweezers one by one [15]. The washed roots were classified into 4 levels (0–0.5, 0.5–1.0, 1.0–2.0, and >2.0 mm) by vernier caliper and then were oven- Water 2020, 12, 3301 5 of 16 −3 dried for 24 h at 65 °C and weighed to calculate root mass density (RBD, kg m ). Figure 2. Skitch of scouring flume for determining soil erosion resistance of gully heads. Figure 2. Skitch of scouring flume for determining soil erosion resistance of gully heads. In addition, the flow depth (h, m) and shear stress (, Pa) were also calculated as follows: 2.5. Parameter Calculation after the Experiments h = (2) Soil particle is detached when flow shear vw stress exceeds the critical shear stress [6]. Soil erodibility parameter (Kr) and critical shear stress (c) were estimated for every natural restoration = ghS (3) stage as the slope coefficient and intercept on the abscissa axis of the regression line between soil 3 1 where h is flow depth (m), q is the flow discharge (m s ), w is the width of the flume (m), v is the detachment rate and shear stress as described in the WEPP model as follow: 1 3 2 mean flow velocity (m s ).  is the water mass density (kg m ), g is the gravity constant (m s ), and S is the slope steepness (m m ). DK = ()   (4) r e c After each scouring test, a steel cubical box (10 cm in length) was used to take soil sample in the center of soil area of scouring flume, and the sample was soaked in tap water for about one hour to Generally, soil detachment rate can be considered as zero when root reached the infinity. To increase the dispersion of soil and then were placed on a 0.25 mm sieve and washed with tap water quantify the relationship between detachment rate and root mass density, the Hill curve was selected using low-pressure head. The living roots, plant debris and some pebbles were left on the sieve. to simulate the relationship between them [21,31,42]. The Hill curve is expressed as follows [43]: Only the living roots were picked out carefully using tweezers one by one [15]. The washed roots were classified into 4 levels (0–0.5, 0.5–1.0, 1.0–2.0, and >2.0 mm) by vernier caliper and then were KX RD = , (K 0, a 0, b 0) oven-dried for 24 h at 65 C and weighed to calculate root mass density (RBD, kg m ). (5) Xb  2.5. Parameter Calculation after the Experiments where RDr is relative soil detachment rate; Xr is root mass density; K, a and b are constants. The Soil particle is detached when flow shear stress exceeds the critical shear stress [6]. Soil erodibility parameter a determines the shape of the curve, b determines the steepness of the curve and K is the parameter (K ) and critical shear stress ( ) were estimated for every natural restoration stage as the r c asymptote of Dr for infinitesimal Xr values. Additionally, the Hill curve can be used to evaluate the slope coecient and intercept on the abscissa axis of the regression line between soil detachment rate ability of roots to increase soil resistance against concentrated flow erosion. According to Li et al. [44], and shear stress as described in the WEPP model as follow: b^(1/a) is plant specific and can be used as an index to compare the effectiveness of different plant D = K (  ) (4) r r c roots in reducing soil erosion rates: the lower b^(1/a), the more effective the plant root. When the value of Xr is b^(1/a), the soil detachment rates is reduced by 50%. Generally, soil detachment rate can be considered as zero when root reached the infinity. To quantify the relationship between detachment rate and root mass density, the Hill curve was selected to simulate 2.6. Statistical the relationship Analysisbetween and Plothem tting[ 21,31,42]. The Hill curve is expressed as follows [43]: The analysis of one-way ANOVA follo KXwed by multiple comparisons with LSD was applied to RD = , (K > 0, a < 0, b > 0) (5) X + b assess the differences of soil properties (Soil bulk density, soil capillary porosity, soil disintegration rate, saturated hydraulic conductivity, organic matter content, and water-stable aggregate) and root where RD is relative soil detachment rate; X is root mass density; K, a and b are constants. The parameter r r a determines the shape of the curve, b determines the steepness of the curve and K is the asymptote of mass density among the five revegetation ages. All soil and root variables of each revegetation ages D for infinitesimal X values. Additionally, the Hill curve can be used to evaluate the ability of roots r r were tested whether the data exhibited a normal distribution and variance homogeneity by Shapiro- to increase soil resistance against concentrated flow erosion. According to Li et al. [44], bˆ(1/a) is plant Water 2020, 12, 3301 6 of 16 specific and can be used as an index to compare the e ectiveness of di erent plant roots in reducing soil erosion rates: the lower bˆ(1/a), the more e ective the plant root. When the value of X is bˆ(1/a), the soil detachment rates is reduced by 50%. 2.6. Statistical Analysis and Plotting The analysis of one-way ANOVA followed by multiple comparisons with LSD was applied to assess the di erences of soil properties (Soil bulk density, soil capillary porosity, soil disintegration rate, saturated hydraulic conductivity, organic matter content, and water-stable aggregate) and root mass density among the five revegetation ages. All soil and root variables of each revegetation ages were tested whether the data exhibited a normal distribution and variance homogeneity by Shapiro-Wilk test and Levene test, respectively. If the data failed to meet the two conditions, the Kruskal–Wallis test was performed for the above analysis. The interaction e ect of flow discharge and revegetation age was detected using a two-way ANOVA. Pearson’s correlation analysis was used to determine linkages among soil properties, root mass density, and soil detachment rate. Relationships among soil detachment rate, soil properties, flow shear stress and restoration age were analyzed by the regression method. The data analyses were conducted in SPSS v. 16.0 statistical software (IBM Corp., Armonk, NY, USA). The figure plotting was conducted by Origin v. 2020 (OriginLab Corp., Northampton, MA, USA). 3. Results and Discussion 3.1. E ect of Revegetation on Soil and Root Properties of Gully Heads Figure 3 illustrates that the six soil properties of gully heads exhibited a significant increase or decrease with revegetation age. Compared with slope farmland, the soil bulk density (SBD) and soil disintegration rate (SDR) of revegetated gully heads significantly decreased by 5.7–18.6% and 28.8–80.5%, respectively (p < 0.05, Figure 3a,c), while the soil capillary porosity (SCP), saturated soil hydraulic conductivity (SHC), organic matter content (OMC) and water-stable aggregate (WSA) significantly increased by 3.9–13.8%, 17.4–236.2%, 34.2–221.8%, and 27.7–64.4%, respectively (p < 0.05, Figure 3b,d–f). Figure 4 illustrates that the roots of 1–2 mm in slope farmland had the relatively higher root mass density (RMD, 0.20 kg m ) and accounted for 39% of total RMD. Notably, after revegetation, the RMD of >2.0 mm was significantly greater than those of the other three root diameters, and it can account for 40–61% of total RMD. When revegetation age was greater than 3 years, there was a significant di erence in RMD among four root diameter levels (p < 0.05). In addition, we found that the RMD of four root diameters (except for >2.0 mm) showed a non-significant increase in the first three-years and then significantly increased. These results were similar to previous findings regarding the e ects of revegetation on soil properties [45–47]. In fact, the improvement of soil properties of gully heads with revegetation age can be attributed to the accumulation of fresh plant residues in surface soil as well as roots and decomposed root residues in subsurface soil [48]. These materials can be directly transformed into soil organic matter and thus provide energy/carbon sources and nutrients for soil microorganisms [49,50], further promoting the development of soil aggregation and enhancing the cohesion of soil particles [51]. Hence, vegetation restoration would induce the formation of macroaggregates and increase the water stability of aggregates [19,52]. Water 2020, 12, x FOR PEER REVIEW 6 of 15 Wilk test and Levene test, respectively. If the data failed to meet the two conditions, the Kruskal– Wallis test was performed for the above analysis. The interaction effect of flow discharge and revegetation age was detected using a two-way ANOVA. Pearson’s correlation analysis was used to determine linkages among soil properties, root mass density, and soil detachment rate. Relationships among soil detachment rate, soil properties, flow shear stress and restoration age were analyzed by the regression method. The data analyses were conducted in SPSS v. 16.0 statistical software (IBM Corp., Armonk, NY, USA). The figure plotting was conducted by Origin v. 2020 (OriginLab Corp., Northampton, MA, USA). 3. Results and Discussion 3.1. Effect of Revegetation on Soil and Root Properties of Gully Heads Figure 3 illustrates that the six soil properties of gully heads exhibited a significant increase or decrease with revegetation age. Compared with slope farmland, the soil bulk density (SBD) and soil disintegration rate (SDR) of revegetated gully heads significantly decreased by 5.7–18.6% and 28.8– 80.5%, respectively (p < 0.05, Figure 3a,c), while the soil capillary porosity (SCP), saturated soil hydraulic conductivity (SHC), organic matter content (OMC) and water-stable aggregate (WSA) significantly increased by 3.9–13.8%, 17.4–236.2%, 34.2–221.8%, and 27.7–64.4%, respectively (p < 0.05, Figure 3b,d–f). Figure 4 illustrates that the roots of 1–2 mm in slope farmland had the relatively higher −3 root mass density (RMD, 0.20 kg m ) and accounted for 39% of total RMD. Notably, after revegetation, the RMD of >2.0 mm was significantly greater than those of the other three root diameters, and it can account for 40–61% of total RMD. When revegetation age was greater than 3 years, there was a significant difference in RMD among four root diameter levels (p < 0.05). In addition, we found that the RMD of four root diameters (except for >2.0 mm) showed a non- Water 2020, 12, 3301 7 of 16 significant increase in the first three-years and then significantly increased. Figure Figure 3. 3. V Var ariation iation inin soil sopr il operties properties with wr it evegetation h revegetati age. on a Note: ge. No Bar te: means Bar m the eans 95% the confidence 95% confi interval dence interval (95% CI). (95 Di %er ent CI).lower Different case letters lower indicate case letters significant indicate di er sience gnificant among dif di ference erent r am evegetation ong differen ages t revegetati (p < 0.05). on (a) a Soil ges bulk (p <density 0.05). ((SBD); a) Soil (b bulk ) Soil dcapillary ensity (SBD); porosity (b) (SCP); Soil capill (c) Soil ary disintegration porosity (SCP) rate ; (c(SDR); ) Soil Water 2020, 12, x FOR PEER REVIEW 7 of 15 (d) Saturated soil hydraulic conductivity (SHC); (e) Organic matter content (OMC); (f) Water-stable disintegration rate (SDR); (d) Saturated soil hydraulic conductivity (SHC); (e) Organic matter content aggregate (WSA). (OMC); (f) Water-stable aggregate (WSA). Figure 4. Changes in root mass density (a) and its proportion (b) of di erent root diameters with Figure 4. Changes in root mass density (a) and its proportion (b) of different root diameters with revegetation age. Note: Bar means the 95% confidence interval (95% CI). SF refers to the slope farmland. revegetation age. Note: Bar means the 95% confidence interval (95% CI). SF refers to the slope NR3, NR10, NR18, NR25 represents the 3, 10, 18, and 25 years of natural restoration time, respectively. farmland. NR3, NR10, NR18, NR25 represents the 3, 10, 18, and 25 years of natural restoration time, Di erent capital letters for the same restoration age indicate a significant di erence among di erent respectively. Different capital letters for the same restoration age indicate a significant difference root diameters (p < 0.05), and di erent lowercase letters for the same root diameter level indicate a among different root diameters (p < 0.05), and different lowercase letters for the same root diameter significant di erence among di erent revegetation ages (p < 0.05). (a) Root mass density; (b) Mass level indicate a significant difference among different revegetation ages (p < 0.05). (a) Root mass accumulated percent. density; (b) Mass accumulated percent. These results were similar to previous findings regarding the effects of revegetation on soil properties [45–47]. In fact, the improvement of soil properties of gully heads with revegetation age can be attributed to the accumulation of fresh plant residues in surface soil as well as roots and decomposed root residues in subsurface soil [48]. These materials can be directly transformed into soil organic matter and thus provide energy/carbon sources and nutrients for soil microorganisms [49,50], further promoting the development of soil aggregation and enhancing the cohesion of soil particles [51]. Hence, vegetation restoration would induce the formation of macroaggregates and increase the water stability of aggregates [19,52]. 3.2. Effect of Revegetation Age on Soil Detachment of Gully Heads As illustrated in Figure 5a, the Dr of gully heads showed a significant decrease during the 25- year revegetation. This result was not agreed with the conclusion of Wang et al. [16] who stated that soil detachment capacity of sloped lands fluctuated with abandonment time, and the soil detachment capacity of the slope farmland was significantly greater than those of the abandoned farmlands. The difference was mainly attributed to the great difference in erosion environment (e.g., plant type, geomorphological feature, climate) significantly affecting the succession process [36,47]. The mean Dr of slope farmland was 1.6 to 3.0 times greater than those of revegetated gully heads, which indicated that the revegetation played a role in enhancing the soil resistance of gully heads to concentrated flow erosion. Figure 5b shows the Dr of gully heads of slope farmland and four restored grasslands varied with flow discharge. The optimal relationships between Dr and flow discharge were fitted, which can reflect the response of Dr of gully heads to concentrated flow induced by rainstorms of different recurrence intervals. It was found that the response of Dr of slope farmland to flow discharge could be expressed as a power function (y = m × x ), and the n-value was greater than 1, indicating the soil loss of gully heads increases at an increased speed with increasing flow. However, for the restored gully heads, the optimal relationships between Dr and flow discharge could be described by a series Water 2020, 12, 3301 8 of 16 3.2. E ect of Revegetation Age on Soil Detachment of Gully Heads As illustrated in Figure 5a, the D of gully heads showed a significant decrease during the 25-year revegetation. This result was not agreed with the conclusion of Wang et al. [16] who stated that soil detachment capacity of sloped lands fluctuated with abandonment time, and the soil detachment capacity of the slope farmland was significantly greater than those of the abandoned farmlands. Water 2020, 12, x FOR PEER REVIEW 8 of 15 The di erence was mainly attributed to the great di erence in erosion environment (e.g., plant type, geomorphological feature, climate) significantly a ecting the succession process [36,47]. The mean D of linear functions (y = p × x + q), and the p-value decreased with revegetation age, indicating that the of slope farmland was 1.6 to 3.0 times greater than those of revegetated gully heads, which indicated sensitivity of Dr of the gully heads to concentrated flow erosion gradually decreased with increasing that the revegetation played a role in enhancing the soil resistance of gully heads to concentrated restoration age. Besides, the interacted effect of revegetation age and unit width discharge flow erosion. significantly affected Dr (p < 0.001) (Table 2). Figure 5. Change in soil detachment rate of gully heads with restoration age (a), and its relationships Figure 5. Change in soil detachment rate of gully heads with restoration age (a), and its relationships with flow discharge (b). Note: Bar means the 95% confidence interval (95% CI). SF refers to the slope with flow discharge (b). Note: Bar means the 95% confidence interval (95% CI). SF refers to the slope farmland. NR3, NR10, NR18, NR25 represents the 3, 10, 18 and 25 years of revegetation age, farmland. NR3, NR10, NR18, NR25 represents the 3, 10, 18 and 25 years of revegetation age, respectively. The respec diti v er el ent y. lower The di case fferent letters lowercas indicate e le a tte significant rs indicate di er a ence signifi among cant di di ff erence erent revegetation among different ages (revegetati p < 0.05). on a (a) Site ges code; (p < 0.(05) b) Unit . (a) Si width te code dischar ; (b) Unit ge. width discharge. Figure 5b shows the D of gully heads of slope farmland and four restored grasslands varied with r Table 2. Summary of two-way ANOVAs tests. flow discharge. The optimal relationships between D and flow discharge were fitted, which can reflect p- the response of D of gully heads to concentrated flow induced by rainstorms of di erent recurrence Sourc r e SS Df MS F Value intervals. It was found that the response of D of slope farmland to flow discharge could be expressed −4 −5 Revegetation age 1.08 × 10 4 2.69 × 10 151.44 <0.001 as a power function (y = m x ), and the n-value was greater than 1, indicating the soil loss of gully −4 −5 Unit width discharge 1.75 × 10 4 4.38 × 10 246.48 <0.001 heads increases at an increased speed with increasing flow. However, for the restored gully heads, Revegetation age × Unit width −5 −6 the optimal relationships between D and flow discharge could be described by a series of linear r 2.29 × 10 16 1.43 × 10 8.07 <0.001 discharge functions (y = p x + q), and the p-value decreased with revegetation age, indicating that the sensitivity −5 −7 Error 1.33 × 10 75 1.78 × 10 of D of the gully heads to concentrated flow erosion gradually decreased with increasing restoration Total 0.0011 100 age. Besides, the interacted e ect of revegetation age and unit width discharge significantly a ected D (p < 0.001) (Table 2). 3.3. Response of Soil Detachment to Soil Properties Table 2. Summary of two-way ANOVAs tests. Figure 6 showed that Dr was positively correlated with soil bulk density and soil disintegration rate (p < 0.01), but negatively correlated with capillary porosity, saturated hydraulic conductivity, Source SS Df MS F p-Value organic matter and water-stable aggregate of >0.25 mm (p < 0.01). Regression analysis showed that Dr 4 5 Revegetation age 1.08 10 4 2.69 10 151.44 <0.001 increased with soil bulk density as a power function (Figure 7a), which showed an opposite trend 4 5 Unit width discharge 4 246.48 <0.001 1.75 10 4.38 10 with the Wang et al. [13] and Yu et al. [15]. Lower soil bulk density was caused by greater root Revegetation age Unit 5 6 2.29 10 16 1.43 10 8.07 <0.001 physical and soil organisms’ activities, and thus a soil with lower bulk density was harder to be width discharge 5 7 Error 1.33 10 75 1.78 10 detached. Additionally, Dr decreased with capillary porosity as a logarithmic function (Figure 7b), Total 0.0011 100 which was caused by physically binding and chemically bonding effect of root improving soil structure and porosity and hence increasing soil resistance to erosion [28,53]. Soil disintegration rate referred to the dispersion speed of soil contacting with water, which is an important factor determining soil resistance to erosion [14]. In this study, the soil disintegration rate decreased with the revegetation time (Figure 3c) and Dr increased linearly with an increase in soil disintegration rate (Figure 7c). This is attributed to root wedging mechanism preventing soil from detaching that roots can bind soil and tie surface soil layer into strong and stable subsurface soil layer [14,54]. Saturated soil hydraulic conductivity is an integrating parameter for several physical characteristics such as bulk density, porosity, and mechanical composition. The conclusion that the Dr decreased with increasing soil hydraulic conductivity by a power function is reasonable (Figure 7d) because this Water 2020, 12, 3301 9 of 16 3.3. Response of Soil Detachment to Soil Properties Figure 6 showed that D was positively correlated with soil bulk density and soil disintegration rate (p < 0.01), but negatively correlated with capillary porosity, saturated hydraulic conductivity, organic matter and water-stable aggregate of >0.25 mm (p < 0.01). Regression analysis showed that D increased with soil bulk density as a power function (Figure 7a), which showed an opposite trend with the Wang et al. [13] and Yu et al. [15]. Lower soil bulk density was caused by greater root physical and soil organisms’ activities, and thus a soil with lower bulk density was harder to be detached. Additionally, D decreased with capillary porosity as a logarithmic function (Figure 7b), which was caused by physically binding and chemically bonding e ect of root improving soil structure and porosity and hence increasing soil resistance to erosion [28,53]. Soil disintegration rate referred to the dispersion speed of soil contacting with water, which is an important factor determining soil resistance to erosion [14]. In this study, the soil disintegration rate decreased with the revegetation time (Figure 3c) and D increased linearly with an increase in soil disintegration rate (Figure 7c). This is attributed to root wedging mechanism preventing soil from detaching that roots can bind soil and tie surface soil layer into strong and stable subsurface soil layer [14,54]. Saturated soil hydraulic conductivity is an integrating parameter for several physical characteristics such as bulk density, Water 2020, 12, x FOR PEER REVIEW 9 of 15 porosity, and mechanical composition. The conclusion that the D decreased with increasing soil hydraulic conductivity by a power function is reasonable (Figure 7d) because this study and previous study and previous research findings have also indicated that changes in soil bulk density and research findings have also indicated that changes in soil bulk density and porosity influenced soil porosity influenced soil detachment and also were affected by revegetation (e.g., Neves et al. [55]; detachment and also were a ected by revegetation (e.g., Neves et al. [55]; Zhang et al. [56]). A negative Zhang et al. [56]). A negative power function was found between Dr and soil organic matter content power function was found between D and soil organic matter content (Figure 7e). The accumulation (Figure 7e). The accumulation of soil organic matter in soil could promote the formation of aggregate of soil organic matter in soil could promote the formation of aggregate and enhance the cohesion of soil and enhance the cohesion of soil particles [57]. Hence, water-stable aggregate also was an indicator particles [57]. Hence, water-stable aggregate also was an indicator determining soil resistance to flow determining soil resistance to flow erosion [19]. The Dr decreased as a linear function of water-stable erosion [19]. The D decreased as a linear function of water-stable aggregate of >0.25 mm (Figure 7f). aggregate of >0.25 mm (Figure 7f). The results were in agreement with the findings of Li et al. [14]. The results were in agreement with the findings of Li et al. [14]. However, in Wang et al. [13,16] studies, However, in Wang et al. [13,16] studies, no significant relationships were found between Dr and no significant relationships were found between D and organic matter and water-stable aggregate of organic matter and water-stable aggregate of >0.25 mm, probably caused by small variations of the >0.25 mm, probably caused by small variations of the two factors in their studies and di erence in two factors in their studies and difference in land use between their studies and this study land use between their studies and this study (Podwojewski et al. [58]). (Podwojewski et al. [58]). Figure 6. Correlation matrix among soil detachment rate, soil properties, and root mass density. Note: Figure 6. Correlation matrix among soil detachment rate, soil properties, and root mass density. Dr, SBD, SCP, SDR, SHC, OMC, WAS, RMD < 0.5, RMD 0.5–1.0, RMD 1.0–2.0, and RMD > 2.0 refers Note: Dr, SBD, SCP, SDR, SHC, OMC, WAS, RMD < 0.5, RMD 0.5–1.0, RMD 1.0–2.0, and RMD > 2.0 to the soil detachment rate, soil bulk density, soil capillary porosity, soil disintegration rate, saturated refers to the soil detachment rate, soil bulk density, soil capillary porosity, soil disintegration rate, soil hydraulic conductivity, organic matter content, water-stable aggregate, root mass density of <0.5 saturated soil hydraulic conductivity, organic matter content, water-stable aggregate, root mass density mm, root mass density of 0.5–1.0 mm, root mass density of 1.0–2.0 mm, and root mass density of >2.0 of <0.5 mm, root mass density of 0.5–1.0 mm, root mass density of 1.0–2.0 mm, and root mass density mm, respectively. of >2.0 mm, respectively. Figure 7. Relationships between soil detachment rate and soil properties. (a) Soil bulk density (SBD); (b) Soil capillary porosity (SCP); (c) Soil disintegration rate (SDR); (d) Saturated soil hydraulic conductivity (SHC); (e) Organic matter content (OMC); (f) Water-stable aggregate (WSA). Water 2020, 12, x FOR PEER REVIEW 9 of 15 study and previous research findings have also indicated that changes in soil bulk density and porosity influenced soil detachment and also were affected by revegetation (e.g., Neves et al. [55]; Zhang et al. [56]). A negative power function was found between Dr and soil organic matter content (Figure 7e). The accumulation of soil organic matter in soil could promote the formation of aggregate and enhance the cohesion of soil particles [57]. Hence, water-stable aggregate also was an indicator determining soil resistance to flow erosion [19]. The Dr decreased as a linear function of water-stable aggregate of >0.25 mm (Figure 7f). The results were in agreement with the findings of Li et al. [14]. However, in Wang et al. [13,16] studies, no significant relationships were found between Dr and organic matter and water-stable aggregate of >0.25 mm, probably caused by small variations of the two factors in their studies and difference in land use between their studies and this study (Podwojewski et al. [58]). Figure 6. Correlation matrix among soil detachment rate, soil properties, and root mass density. Note: Dr, SBD, SCP, SDR, SHC, OMC, WAS, RMD < 0.5, RMD 0.5–1.0, RMD 1.0–2.0, and RMD > 2.0 refers to the soil detachment rate, soil bulk density, soil capillary porosity, soil disintegration rate, saturated soil hydraulic conductivity, organic matter content, water-stable aggregate, root mass density of <0.5 mm, root mass density of 0.5–1.0 mm, root mass density of 1.0–2.0 mm, and root mass density of >2.0 Water 2020, 12, 3301 10 of 16 mm, respectively. Figure 7. Relationships between soil detachment rate and soil properties. (a) Soil bulk density (SBD); Figure 7. Relationships between soil detachment rate and soil properties. (a) Soil bulk density (b) Soil capillary porosity (SCP); (c) Soil disintegration rate (SDR); (d) Saturated soil hydraulic (SBD); (b) Soil capillary porosity (SCP); (c) Soil disintegration rate (SDR); (d) Saturated soil hydraulic conductivity conductivity (S (SHC); HC); ( (e e) ) Or Org ganic anic matter matter conte content nt (O (OMC); MC); ((ff)) Wa Water ter -stable -stableaggr aggreg egate ate (WSA). (WSA). 3.4. Response of Soil Loss of Gully-Head to Root Traits Significant correlation was found between D and RMD of 0–0.5 mm, 0.5–1.0 mm, 1.0–2.0 mm and >2.0 mm (p < 0.01, Figure 6), of which the RMD of 0–0.5 mm had the highest correlation with D , indicating that roots of each diameter level had the significant impact on soil erosion of gully heads, especially the fibrous root of 0–0.5 mm. Furthermore, the Hill curve could well simulate the relationships between D and RMD of di erent root diameters with R varying from 0.42 to 0.57 (Figure 8). As illustrated in Figure 8, the D showed a rapid decrease when RMD of 0–0.5, 0.5–1.0, 3 3 3 3 1.0–2.0 and >2.0 mm ranged from 0 to 0.25 kg m , 0 to 0.3 kg m , 0 to 0.5 kg m , and 0 to 1.0 kg m , respectively, implying that soil erosion of gully heads could be controlled once vegetation restoration or root growth in soil. Although the roots were limited in density and flexible in early revegetation stage, whereas, roots can contribute to soil cohesion and additional strength, and be crucial in reduction of soil erosion [28,59]. Additionally, root system can bind soil and tie surface soil layer into strong and stable subsurface soil layer [54]. Well-developed root system had great physical binding and chemical bonding e ect that could well bind soil particles and soil aggregates together and enhance soil resistance to erosion [16,28,42]. In addition, judged by fitted eciency (R ), the optimal results were found in RMD of 0–0.5 mm (Figure 8a), indicating that fibrous root of 0–0.5 mm is the optimal root system reducing soil loss of gully heads. However, Li et al. [44] reported that the ability of plant roots to decrease soil erosion mainly depended on the number of fibrous roots <1.0 mm. Shangguan et al. [53] also found a similar result but recommended root surface area density as the root variable. The reason may be that plant species with contrasting root architectures have a di erent erosion reduction e ect [25]. Additionally, Amezketa [60] and Gyssels et al. [61] reported that monocotyledonous plants are superior to dicotyledonous plants and grasses are better than cereals in stabilizing soil aggregates. According to Li et al. [44], bˆ(1/a) can be used as an indicator to compare the e ectiveness of di erent diameter roots in reducing soil erosion. The lower bˆ(1/a), the more e ective the diameter root. The relatively lower bˆ(1/a) values (0.132 and 0.131 kg m ) were found in the roots of 0–0.5 mm and 0.5–1.0 mm than 1.0–2.0 mm and >2.0 mm (Figure 8), indicating that the 0–0.5 mm and 0.5–1.0 mm are the most e ective roots in reducing soil erosion of gully heads. However, De Baets et al. [25] study the e ect of the mixed community of four grasses [Lolium perenne (variety: tove), L. perenne (variety: starlet), Festuca rubra (variety: echo), and F. arundinacea (variety: starlet)] on SDR and found the bˆ(1/a) Water 2020, 12, x FOR PEER REVIEW 10 of 15 3.4. Response of Soil Loss of Gully-Head to Root Traits Significant correlation was found between Dr and RMD of 0–0.5 mm, 0.5–1.0 mm, 1.0–2.0 mm and >2.0 mm (p < 0.01, Figure 6), of which the RMD of 0–0.5 mm had the highest correlation with Dr, indicating that roots of each diameter level had the significant impact on soil erosion of gully heads, especially the fibrous root of 0–0.5 mm. Furthermore, the Hill curve could well simulate the relationships between Dr and RMD of different root diameters with R varying from 0.42 to 0.57 (Figure 8). As illustrated in Figure 8, the Dr showed a rapid decrease when RMD of 0–0.5, 0.5–1.0, 1.0– −3 −3 −3 −3 2.0 and >2.0 mm ranged from 0 to 0.25 kg m , 0 to 0.3 kg m , 0 to 0.5 kg m , and 0 to 1.0 kg m , respectively, implying that soil erosion of gully heads could be controlled once vegetation restoration Water 2020, 12, 3301 11 of 16 or root growth in soil. Although the roots were limited in density and flexible in early revegetation stage, whereas, roots can contribute to soil cohesion and additional strength, and be crucial in value is 0.79 kg m that is greater than that of our study. The result fully indicated that the di erent reduction of soil erosion [28,59]. Additionally, root system can bind soil and tie surface soil layer into plant communities had the markedly di erent influences on reducing soil erosion. The result also strong and stable subsurface soil layer [54]. Well-developed root system had great physical binding suggested us reasonably choosing plant species with di erent root architectures and root diameters for and chemical bonding effect that could well bind soil particles and soil aggregates together and revegetation at gully heads. enhance soil resistance to erosion [16,28,42]. Figure 8. Relationships between relative soil detachment rate and root mass density of different root Figure 8. Relationships between relative soil detachment rate and root mass density of di erent root diameters. (a) Root of 0–0.5 mm; (b) Root of 0.5–1.0 mm; (c) Root of 1.0–2.0 mm; (d) Root > 2.0 mm. diameters. (a) Root of 0–0.5 mm; (b) Root of 0.5–1.0 mm; (c) Root of 1.0–2.0 mm; (d) Root > 2.0 mm. 3.5. E ect of Revegetation Age on Soil Erosion Resistance of Gully Head In addition, judged by fitted efficiency (R ), the optimal results were found in RMD of 0–0.5 mm (Figure 8a), indicating that fibrous root of 0–0.5 mm is the optimal root system reducing soil loss of Rill soil erodibility parameter (K ) and critical shear stress ( ) were employed to characterize the r c gully heads. However, Li et al. [44] reported that the ability of plant roots to decrease soil erosion soil erosion resistance of gully heads [13,21], and were determined by the WEPP model (Equation (4)). mainly depended on the number of fibrous roots <1.0 mm. Shangguan et al. [53] also found a similar The fitted linear function between D and shear stress was illustrated in Figure 9. The slope of the fitted result but recommended root surface area density as the root variable. The reason may be that plant line is equal to the K , and the K of the restored grasslands were 31% to 78.6% less than that of slope r r species with contrasting root architectures have a different erosion reduction effect [25]. Additionally, farmland. In addition, we found that the soil erodibility of 3-year restored grassland rapidly declined Amezketa [60] and Gyssels et al. [61] reported that monocotyledonous plants are superior to by 31% compared with the slope farmland, indicating the short-term revegetation can rapidly reduce dicotyledonous plants and grasses are better than cereals in stabilizing soil aggregates. 1 soil erodibility of gully heads. The K of grasslands in this study ranged from 0.0009 to 0.0029 s m , According to Li et al. [44], b^(1/a) can be used as an indicator to compare the effectiveness of which were less than those reported by Li et al. [14]. Wang et al. [13] found that averaged K of restored different diameter roots in reducing soil eros1ion. The lower b^(1/a), the more effective the diameter lands of abandoned farmland was 0.0024 s m that was close to those of this study. The di erence was mainly caused by di erences in land use, plant species and restoration time. The soil samples were taken from di erent vegetation restoration models (korshinsk peashrub, black locust, Chinese pine and mixed forest of amorpha and Chinese pine) in the study of Wang et al. [13], and the restoration age (37 years) was greater than that of this study (3 to 25 years). Regression analysis found that the K decreased with restoration time in an exponential function and showed a slight change when restored age was greater than 18 years (Figure 10). Water 2020, 12, x FOR PEER REVIEW 11 of 15 −3 root. The relatively lower b^(1/a) values (0.132 and 0.131 kg m ) were found in the roots of 0–0.5 mm and 0.5–1.0 mm than 1.0–2.0 mm and >2.0 mm (Figure 8), indicating that the 0–0.5 mm and 0.5–1.0 mm are the most effective roots in reducing soil erosion of gully heads. However, De Baets et al. [25] study the effect of the mixed community of four grasses [Lolium perenne (variety: tove), L. perenne (variety: starlet), Festuca rubra (variety: echo), and F. arundinacea (variety: starlet)] on SDR and found −3 the b^(1/a) value is 0.79 kg m that is greater than that of our study. The result fully indicated that the different plant communities had the markedly different influences on reducing soil erosion. The result also suggested us reasonably choosing plant species with different root architectures and root diameters for revegetation at gully heads. 3.5. Effect of Revegetation Age on Soil Erosion Resistance of Gully Head Rill soil erodibility parameter (Kr) and critical shear stress (c) were employed to characterize the soil erosion resistance of gully heads [13,21], and were determined by the WEPP model (Equation (4)). The fitted linear function between Dr and shear stress was illustrated in Figure 9. The slope of the fitted line is equal to the Kr, and the Kr of the restored grasslands were 31% to 78.6% less than that of slope farmland. In addition, we found that the soil erodibility of 3-year restored grassland rapidly declined by 31% compared with the slope farmland, indicating the short-term revegetation can rapidly reduce soil erodibility of gully heads. The Kr of grasslands in this study ranged from 0.0009 −1 to 0.0029 s m , which were less than those reported by Li et al. [14]. Wang et al. [13] found that −1 averaged Kr of restored lands of abandoned farmland was 0.0024 s m that was close to those of this study. The difference was mainly caused by differences in land use, plant species and restoration time. The soil samples were taken from different vegetation restoration models (korshinsk peashrub, black locust, Chinese pine and mixed forest of amorpha and Chinese pine) in the study of Wang et al. [13], and the restoration age (37 years) was greater than that of this study (3 to 25 years). Regression analysis found that the Kr decreased with restoration time in an exponential function and showed a Water 2020, 12, 3301 12 of 16 slight change when restored age was greater than 18 years (Figure 10). Figure 9. Relationship between soil detachment rate and shear stress under di erent revegetation ages. (a) SF; (b) NR3; (c) NR10; (d) NR18; (e) NR25. Water 2020, 12, x FOR PEER REVIEW 12 of 15 Figure 9. Relationship between soil detachment rate and shear stress under different revegetation ages. (a) SF; (b) NR3; (c) NR10; (d) NR18; (e) NR25. Figure 10. Relationships between soil erodibility and critical shear stress and revegetation age. Figure 10. Relationships between soil erodibility and critical shear stress and revegetation age. In addition, the c increased with restoration age by a power function (Figure 10). However, the In addition, the  increased with restoration age by a power function (Figure 10). result was inconsistent with the finding of Wang et al. [16] that critical shear stress varied with However, the result was inconsistent with the finding of Wang et al. [16] that critical shear stress restoration age in a nonlinear pattern, reaching the minimum at the restored age of 18. The difference varied with restoration age in a nonlinear pattern, reaching the minimum at the restored age of 18. in the temporal change of critical shear stress between Wang et al. [13] and this study was caused The di erence in the temporal change of critical shear stress between Wang et al. [13] and this study was probably by differences in soil properties and vegetation characteristics of the sampling sites. caused probably by di erences in soil properties and vegetation characteristics of the sampling sites. Compared with slope farmland, the c of the grasslands was improved by 1.2 to 4.0 times, while c of Compared with slope farmland, the  of the grasslands was improved by 1.2 to 4.0 times, while  of c c restored land had a little decrease when restored time was more than 18 years (Figure 10). The result restored land had a little decrease when restored time was more than 18 years (Figure 10). The result further indicated that revegetation can effectively improve the soil erosion resistance of gully head further indicated that revegetation can e ectively improve the soil erosion resistance of gully head to to concentrated flow, and the critical shear stress would reach a stable state after a 18-year concentrated flow, and the critical shear stress would reach a stable state after a 18-year revegetation. revegetation. 4. Conclusions This study was carried out to explore the effect of revegetation age on soil erosion resistance of gully heads in the gully region of the Loess Plateau. The results showed that revegetation significantly improved soil properties and promoted root accumulation of gully heads. The mean Dr of slope farmland was 1.6 to 3.0 times greater than those of revegetated gully heads. The revegetation can effectively weaken the sensitivity of soil erosion of the gully heads to concentrated flow. The Dr of gully heads was positively related to bulk density and disintegration rate and negatively related to soil capillary porosity, saturated soil hydraulic conductivity, organic matter content, and water- stable aggregate. Roots of 0–0.5 mm and 0.5–1.0 mm were the most effective roots in reducing soil erosion of gully head, and the native plant species with rich root of 0.5–1.0 mm and 0–0.5 mm were recommended as the first choice for revegetation to restrain gully headcut erosion. Revegetation can reduce soil erodibility of gully heads by 31% to 78.6% and improve the critical shear stress by 1.2 to 4.0 times. This study allows us to better evaluate soil vulnerability of gully head to concentrated flow erosion during revegetation. Further studies are needed to quantify the effect of the different combinations of vegetation types with different root architecture types on soil erosion resistance of gully heads. Author Contributions: Conceptualization, Z.C. and M.G.; data curation, Z.C.; formal analysis, Z.C. and M.G.; funding acquisition, W.W.; investigation Z.C. and M.G.; methodology, Z.C.; project administration, W.W.; resources, W.W.; software, Z.C.; supervision, W.W.; validation, W.W.; writing—original draft, Z.C.; writing— review and editing, M.G. and W.W. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported by the National Natural Science Foundation of China (42077079; 41571275). Water 2020, 12, 3301 13 of 16 4. Conclusions This study was carried out to explore the e ect of revegetation age on soil erosion resistance of gully heads in the gully region of the Loess Plateau. The results showed that revegetation significantly improved soil properties and promoted root accumulation of gully heads. The mean D of slope farmland was 1.6 to 3.0 times greater than those of revegetated gully heads. The revegetation can e ectively weaken the sensitivity of soil erosion of the gully heads to concentrated flow. The D of gully heads was positively related to bulk density and disintegration rate and negatively related to soil capillary porosity, saturated soil hydraulic conductivity, organic matter content, and water-stable aggregate. Roots of 0–0.5 mm and 0.5–1.0 mm were the most e ective roots in reducing soil erosion of gully head, and the native plant species with rich root of 0.5–1.0 mm and 0–0.5 mm were recommended as the first choice for revegetation to restrain gully headcut erosion. Revegetation can reduce soil erodibility of gully heads by 31% to 78.6% and improve the critical shear stress by 1.2 to 4.0 times. This study allows us to better evaluate soil vulnerability of gully head to concentrated flow erosion during revegetation. Further studies are needed to quantify the e ect of the di erent combinations of vegetation types with di erent root architecture types on soil erosion resistance of gully heads. Author Contributions: Conceptualization, Z.C. and M.G.; data curation, Z.C.; formal analysis, Z.C. and M.G.; funding acquisition, W.W.; investigation Z.C. and M.G.; methodology, Z.C.; project administration, W.W.; resources, W.W.; software, Z.C.; supervision, W.W.; validation, W.W.; writing—original draft, Z.C.; writing—review and editing, M.G. and W.W. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported by the National Natural Science Foundation of China (42077079; 41571275). Acknowledgments: All authors thank anonymous reviewers for insightful comments on the original manuscript. Conflicts of Interest: The authors declare no conflict of interest. References 1. Mohammed, S.; Al-Ebraheem, A.; Holb, I.J.; Alsafadi, K.; Dikkeh, M.; Pham, Q.B.; Linh, N.T.T.; Szabo, S. Soil management e ects on soil water erosion and runo in central Syria—A comparative evaluation of general linear model and random forest regression. Water 2020, 12, 2529. [CrossRef] 2. Hao, H.X.; Cheng, L.; Guo, Z.L.; Wang, L.; Shi, Z.H. Plant community characteristics and functional traits as drivers of soil erodibility mitigation along a land degradation gradient. Land Degrad. Dev. 2020, 31, 1851–1863. [CrossRef] 3. Zarris, D.; Lykoudi, E.; Panagoulia, D. Assessing the impacts of sediment yield on the sustainability of major hydraulic systems. In Proceedings of the International Conference “Protection and Restoration of the Environment VIII”, Chania, Greece, 3–7 July 2006. 4. Brazier, R.E.; Beven, K.J.; Freer, J.; Rowan, J.S. Equifinality and uncertainty in physically based soil erosion models: Application of the GLUE methodology to WEPP–the water erosion prediction project–for sites in the UK and USA. Earth Surf. Process. Landf. 2000, 25, 825–845. [CrossRef] 5. Kinnell, P.; Risse, L.M. USLE-M: Empirical modeling rainfall erosion through runo and sediment concentration. Soil Sci. Soc. Am. J. 1998, 62, 1667–1672. [CrossRef] 6. Nearing, M.A.; Foster, G.R.; Lane, L.J.; Finkner, S.C. A process-based soil erosion model for USDA-water erosion prediction project technology. Trans. ASAE 1989, 32, 1587–1593. [CrossRef] 7. Efthimiou, N.; Lykoudi, E.; Panagoulia, D.; Karavitis, C. Assessment of soil susceptibility to erosion using the EPM and RUSLE models: The case of Venetikos river catchment. Glob. Nest. J. 2016, 18, 164–179. 8. Guo, W.Z.; Kang, H.L.; Wang, W.L.; Guo, M.M.; Chen, Z.X. Erosion-reducing e ects of revegetation and fish-scale pits on steep spoil heaps under concentrated runo on the Chinese Loess Plateau. Land Degrad Dev. 2020. [CrossRef] 9. Gramlich, A.; Stoll, S.; Stamm, C.; Walter, T.; Prasuhn, V. E ects of artificial land drainage on hydrology, nutrient and pesticide fluxes from agricultural fields—A review. Agric. Ecosyst. Environ. 2018, 266, 84–99. [CrossRef] 10. Kervroëdan, L.; Armand, R.; Saunier, M.; Ouvry, J.; Faucon, M. Plant functional trait e ects on runo to design herbaceous hedges for soil erosion control. Ecol. Eng. 2018, 118, 143–151. [CrossRef] Water 2020, 12, 3301 14 of 16 11. Sun, L.; Zhang, G.H.; Luan, L.L.; Liu, F. Temporal variation in soil resistance to flowing water erosion for soil incorporated with plant litters in the Loess Plateau of China. Catena 2016, 145, 239–245. [CrossRef] 12. Chang, E.; Li, P.; Li, Z.; Su, Y.; Zhang, Y.; Zhang, J.; Liu, Z.; Li, Z. The impact of vegetation successional status on slope runo erosion in the Loess Plateau of China. Water 2019, 11, 2614. [CrossRef] 13. Wang, B.; Zhang, G.H.; Shi, Y.Y.; Zhang, X.C. Soil detachment by overland flow under di erent vegetation restoration models in the Loess Plateau of China. Catena 2014, 116, 51–59. [CrossRef] 14. Li, Z.W.; Zhang, G.H.; Geng, R.; Wang, H.; Zhang, X.C. Land use impacts on soil detachment capacity by overland flow in the Loess Plateau, China. Catena 2015, 124, 9–17. [CrossRef] 15. Yu, Y.C.; Zhang, G.H.; Geng, R.; Sun, L. Temporal variation in soil detachment capacity by overland flow under four typical crops in the Loess Plateau of China. Biosyst. Eng. 2014, 122, 139–148. [CrossRef] 16. Wang, B.; Zhang, G.H.; Shi, Y.Y.; Zhang, X.C.; Ren, Z.P.; Zhu, L.J. E ect of natural restoration time of abandoned farmland on soil detachment by overland flow in the Loess Plateau of China. Earth Surf. Process. Landf. 2013, 38, 1725–1734. [CrossRef] 17. Scherer, U.; Zehe, E.; Tr Bing, K.; Kai, G. Prediction of soil detachment in agricultural loess catchments: Model development and parameterisation. Catena 2012, 90, 63–75. [CrossRef] 18. Zhang, G.H.; Tang, M.K.; Zhang, X.C. Temporal variation in soil detachment under di erent land uses in the Loess Plateau of China. Earth Surf. Process. Landf. 2010, 34, 1302–1309. [CrossRef] 19. Bernard, B.; Roose, E. Aggregate stability as an indicator of soil susceptibility to runo and erosion; validation at several levels. Catena 2002, 47, 133–149. 20. Knapen, A.; Poesen, J.; Govers, G.; Baets, S.D. The e ect of conservation tillage on runo erosivity and soil erodibility during concentrated flow. Hydrol. Process. 2008, 22, 1497–1508. [CrossRef] 21. Guo, M.M.; Wang, W.L.; Wang, T.C.; Wang, W.X.; Kang, H.L. Impacts of di erent vegetation restoration options on gully head soil resistance and soil erosion in loess tablelands. Earth Surf. Process. Landf. 2020, 45, 1038–1050. [CrossRef] 22. Lazarovitch, N.; Vanderborght, J.; Jin, Y.; van Genuchten, M.T. The root zone: Soil physics and beyond. Vadose Zone J. 2018, 17, 1–6. [CrossRef] 23. Panagoulia, D. Hydrological modelling of a medium-size mountainous catchment from incomplete meteorological data. J. Hydrol. 1992, 137, 279–310. [CrossRef] 24. Falkenmark, M. Land and water integration and river basin management. FAO Land Water Bull. 1995, 1, 15–16. 25. Baets, S.D.; Poesen, J. Empirical models for predicting the erosion-reducing e ects of plant roots during concentrated flow erosion. Geomorphology 2010, 118, 425–432. [CrossRef] 26. Guo, M.M.; Wang, W.L.; Shi, Q.H.; Chen, T.D.; Kang, H.L.; Li, J.M. An experimental study on the e ects of grass root density on gully headcut erosion in the gully region of China’s Loess Plateau. Land Degrad. Dev. 2019, 30, 2107–2125. [CrossRef] 27. Khanal, A.; Fox, G.A. Detachment characteristics of root-permeated soils from laboratory jet erosion tests. Ecol. Eng. 2017, 100, 335–343. [CrossRef] 28. De Baets, S.; Poesen, J.; Gyssels, G.; Knapen, A.; Adili, A.A. E ects of grass roots on the erodibility of topsoils during concentrated flow. Geomorphology 2006, 76, 54–67. [CrossRef] 29. Mamo, M.; Bubenzer, G.D. Detachment rate, soil erodibility, and soil strength as influenced by living plant roots part II: Field study. Trans. ASAE 2001, 5, 1175–1181. [CrossRef] 30. Mamo, M.; Bubenzer, G.D. Detachment rate, soil erodibility, and soil strength as influenced by living plant roots part I: Laboratory study. Trans. ASAE 2001, 5, 1167–1174. [CrossRef] 31. Baets, S.D.; Poesen, J.; Knapen, A.; Galindo, P. Impact of root architecture on the erosion-reducing potential of roots during concentrated flow. Earth Surf. Process. Landf. 2007, 32, 1323–1345. [CrossRef] 32. Shi, Q.; Wang, W.; Guo, M.; Chen, Z.; Feng, L.; Zhao, M.; Xiao, H. The impact of flow discharge on the hydraulic characteristics of headcut erosion processes in the gully region of the Loess Plateau. Hydrol. Process. 2019, 34, 718–729. [CrossRef] 33. Guo, M.M.; Wang, W.L.; Kang, H.L.; Yang, B. Changes in soil properties and erodibility of gully heads induced by vegetation restoration on the Loess Plateau, China. J. Arid Land 2018, 10, 712–725. [CrossRef] 34. Fu, B.; Chen, L.; Ma, K.; Zhou, H.; Wang, J. The relationships between land use and soil conditions in the hilly area of the loess plateau in northern Shaanxi, China. Catena 2000, 39, 69–78. [CrossRef] Water 2020, 12, 3301 15 of 16 35. Xiong, Y.; Zhou, J.Z.; Chen, L.; Jia, B.J.; Sun, N.; Tian, M.Q.; Hu, G.H. Land use pattern and vegetation cover dynamics in the three Gorges Reservoir (TGR) intervening Basin. Water 2020, 12, 2036. [CrossRef] 36. Jiao, J.Y.; Tzanopoulos, J.; Xofis, P.; Bai, W.J.; Ma, X.H.; Mitchley, J. Can the study of natural vegetation succession assist in the control of soil erosion on abandoned croplands on the Loess Plateau, China? Restor. Ecol. 2007, 15, 391–399. [CrossRef] 37. Wang, B.; Liu, G.B.; Xue, S.; Zhu, B.B. Changes in soil physico-chemical and microbiological properties during natural succession on abandoned farmland in the Loess Plateau. Environ. Earth Sci. 2011, 62, 915–925. [CrossRef] 38. Chen, Y.P.; Wang, K.B.; Lin, Y.S.; Shi, W.Y.; Song, Y.; He, X.H. Balancing green and grain trade. Nat. Geosci. 2015, 8, 739–741. [CrossRef] 39. Kompani-Zare, M.; Soufi, M.; Hamzehzarghani, H.; Dehghani, M. The e ect of some watershed, soil characteristics and morphometric factors on the relationship between the gully volume and length in Fars Province, Iran. Catena 2011, 86, 150–159. [CrossRef] 40. Guo, M.M.; Wang, W.L.; Kang, H.L.; Yang, B.; Li, J.M. Changes in soil properties and resistance to concentrated flow across a 25-year passive restoration chronosequence of grasslands on the Chinese Loess Plateau. Restor. Ecol. 2020, 28, 104–114. [CrossRef] 41. Luk, S.H.; Merz, W. Use of the salt tracing technique to determine the velocity of overland flow. Sci. Total Environ. 1992, 5, 289–301. 42. Li, Y.; Xu, X.Q.; Zhu, X.M.; Tian, J.Y. E ectiveness of plant roots on increasing the soil permeability on the Loess Plateau. Chin. Sci. Bull. 1992, 37, 1735–1738. 43. Hill, A.V. The possible e ects of the aggregation of the molecules of haemoglobin on its dissociation curves. J. Physiol. 1910, 40, 4–7. 44. Li, Y.; Zhu, X.M.; Tian, J.Y.; Al-Ebraheem, A. E ectiveness of plant roots to increase the anti-scouribility of soil on the Loess Plateau. Chin. Sci. Bull. 1991, 36, 2077–2082. 45. Gros, R.; Jocteur Monrozier, L.; Bartoli, F.; Chotte, J.L.; Faivre, P. Relationships between soil physico-chemical properties and microbial activity along a restoration chronosequence of alpine grasslands following ski run construction. Appl. Soil Ecol. 2004, 27, 7–22. [CrossRef] 46. Li, Y.Y.; Shao, M.A. Change of soil physical properties under long-term natural vegetation restoration in the Loess Plateau of China. J. Arid. Environ. 2006, 64, 77–96. [CrossRef] 47. Jiao, F.; Wen, Z.M.; An, S.S. Changes in soil properties across a chronosequence of vegetation restoration on the Loess Plateau of China. Catena 2011, 86, 110–116. [CrossRef] 48. Zhao, Y.G.; Wu, P.T.; Zhao, S.W.; Feng, H. Variation of soil infiltrability across a 79-year chronosequence of naturally restored grassland on the Loess Plateau, China. J. Hydrol. 2013, 504, 94–103. [CrossRef] 49. Six, J.; Elliott, E.T. Aggregation and soil organic matter accumulation in cultivated and native grassland soils. Soil Sci. Soc. Am. J. 1998, 62, 1367–1377. [CrossRef] 50. Six, J.; Bossuyt, H.; Degryze, S.; Denef, K. A history of research on the link between (micro)aggregates, soil biota, and soil organic matter dynamics. Soil Tillage Res. 2004, 79, 7–31. [CrossRef] 51. Six, J.; Paustian, K.; Elliott, E.T.; Combrink, C. Soil structure and organic matter I. distribution of aggregate-size classes and aggregate-associated carbon. Soil Sci. Soc. Am. J. 2000, 64, 681–689. [CrossRef] 52. Kong, A.Y.Y.; Six, J.; Bryant, D.C.; Denison, R.F.; Kessel, C.V. The relationship between carbon input, aggregation, and soil organic carbon stabilization in sustainable cropping systems. Soil Sci. Soc. Am. J. 2005, 69. [CrossRef] 53. Shangguan, Z.P.; Zheng, C.Z. Soil anti-scouribility enhanced by plant roots. J. Integr. Plant Biol. 2005, 47, 676–682. 54. Adili, A.A.; Azzam, R.; GiovanniSpagnoli, S.J. Strength of soil reinforced with fiber materials (Papyrus). Soil Mech. Found. Eng. 2012, 48, 241–247. [CrossRef] 55. Neves, C.S.V.J.; Feller, C.; Guimarães, M.F.; Medina, C.C.; Tavares Filho, J.; Fortier, M. Soil bulk density and porosity of homogeneous morphological units identified by the cropping profile method in clayey oxisols in Brazil. Soil Tillage Res. 2003, 71, 109–119. [CrossRef] 56. Zhang, B.J.; Zhang, G.H.; Yang, H.Y.; Wang, H. Soil resistance to flowing water erosion of seven typical plant communities on steep gully slopes on the Loess Plateau of China. Catena 2019, 173, 375–383. [CrossRef] Water 2020, 12, 3301 16 of 16 57. Fattet, M.; Fu, Y.; Ghestem, M.; Ma, W.; Foulonneau, M.; Nespoulous, J.; Le Bissonnais, Y.; Stokes, A. E ects of vegetation type on soil resistance to erosion: Relationship between aggregate stability and shear strength. Catena 2011, 87, 60–69. [CrossRef] 58. Podwojewski, P.; Orange, D.; Jouquet, P.; Valentin, C.; Nguyen, V.T.; Janeau, J.L.; Tran, D.T. Land-use impacts on surface runo and soil detachment within agricultural sloping lands in Northern Vietnam. Catena 2008, 74, 109–118. [CrossRef] 59. Gyssels, G.; Poesen, J. The importance of plant root characteristics in controlling concentrated flow erosion rates. Earth Surf. Process. Landf. 2003, 28, 371–384. [CrossRef] 60. Amézketa, E. Soil aggregate stability: A review. J. Integr. Plant Biol. 1999, 14, 83–151. [CrossRef] 61. Gyssels, G.; Poesen, J.; Bochet, E.; Li, Y. Impact of plant roots on the resistance of soils to erosion by water: A review. Prog. Phys. Geogr. 2005, 29, 189–217. [CrossRef] Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional aliations. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Journal

WaterMultidisciplinary Digital Publishing Institute

Published: Nov 24, 2020

There are no references for this article.