Soil carbon and nitrogen sources and redistribution as affected by erosion and deposition processes: A case study in a loess hilly-gully catchment, China

Agriculture, Ecosystems and Environment 253 (2018) 11–22

Contents lists available at ScienceDirect

Agriculture, Ecosystems and Environment journal homepage: www.elsevier.com/locate/agee

Research paper

Soil carbon and nitrogen sources and redistribution as affected by erosion and deposition processes: A case study in a loess hilly-gully catchment, China ⁎

MARK

⁎⁎

Chun Liua,b,c, Zhongwu Lia,b,c, , Xiaofeng Changb, Jijun Hed, , Xiaodong Niee, Lin Liub, Haibing Xiaob, Danyang Wanga,c, Hao Penga,c, Guangming Zenga,c a

College of Environmental Science and Engineering, Hunan University, Changsha, 410082, PR China State Key Laboratory of Soil Erosion and Dryland Farming on the Loess Plateau, Institute of Soil and Water Conservation, CAS and MWR, Yangling, Shaanxi Province 712100, PR China c Key Laboratory of Environmental Biology and Pollution Control (Hunan University), Ministry of Education, Changsha, 410082, PR China d Base of the State Laboratory of Urban Environmental Processes and Digital Modeling, Key Laboratory of 3D Information Acquisition and Application, Ministry of Education, Capital Normal University, Beijing, 100048, PR China e Guangdong Key Laboratory of Integrated Agro-environmental Pollution Control, PR China b

A R T I C L E I N F O

A B S T R A C T

Keywords: Soil erosion Carbon Nitrogen Sources Isotopes Loess plateau

Understanding how organic carbon (C) and nitrogen (N) move with soil along the fluvial system and where eroded organic C and N sources occur over catchment landscape is of significant importance for evaluating accurately carbon or nitrogen budget at the catchment scale and designing proper management practice in the fragile ecosystem. In this study, we selected a dam-controlled catchment in the southwestern hilly-gully region of the Chinese Loess Plateau with severe soil erosion, and explored the catchment scale carbon and nitrogen redistribution by erosion and deposition processes as well as organic carbon and nitrogen sources in sediments retained by check dam. The physio-chemical characteristics, stable isotopic signatures (13C, 15N), total organic C (TOC) and total N (TN) of soils and sediments in studied catchment were determined and an isotopic mixing model was employed to quantify the relative contributions of each source type (ie., forests, cropland, and gully) to eroded sediment C and N. The results showed that the check dam intercepted 309.6 Gg of eroded soil, 1405.1 Mg of TOC, and 153.5 Mg of TN. The sediment the eroded TOC was mainly sourced from cropland, accounting for 53.54%, followed by gully (29.28%) and then forests (17.18%), respectively. Eroded TN sources was similar to C sources, showing that cropland contributed 53.53%, with gully and forests contributing 30.86% and 15.61%, respectively. Moreover, the forests contributions to eroded C and N gradually decreased in the direction of the runoff pathway at the check dam, and the C and N contributions of cropland and gullies showed the orders of mid-check dam > post-check dam > pre-check dam and pre-check dam > post-check dam > mid-check dam, respectively. Soil erosion and deposition processes induced 1569.8 Mg TOC and 146.7 Mg TN losses, with an average soil C and N erosion rate of 0.051 Mg C ha−1 yr−1 and 0.005 Mg N ha−1 yr−1 during the period (i.e., from 2004 to 2016), accounting for approximately 52.8% and 48.87% of the total amount of eroded TOC and TN, respectively. The results indicated that although the check dam served as a carbon and nitrogen storage and sequestration structure in the loess hilly-gully region, erosion-induced carbon or nitrogen redistribution might still act as a major source for atmospheric CO2 or nitrogen oxide in our studied catchment.

1. Introduction Soil erosion has been intensively studied as regards the impact of biogeochemical cycling of essential elements, especially elemental carbon (C) and nitrogen (N) in terrestrial ecosystems (Lal 2003;

⁎

Quinton et al., 2010; Berhe et al., 2007, 2017). Accelerated soil erosion redistribute large quantities of sediment and associated soil C and N on the Earth's surface, thereby not only leading to loss of surface soil in uplands and resulting in soil quality degradation and agricultural production reduction (Lal., 2010; Kirkels et al., 2014; Li et al., 2016a,b),

Corresponding author at: College of Environmental Science and Engineering, Hunan University, Changsha 410082, PR China. Corresponding author. E-mail addresses: [email protected] (Z. Li), [email protected] (J. He).

⁎⁎

http://dx.doi.org/10.1016/j.agee.2017.10.028 Received 10 September 2017; Received in revised form 29 October 2017; Accepted 31 October 2017 0167-8809/ © 2017 Elsevier B.V. All rights reserved.

Agriculture, Ecosystems and Environment 253 (2018) 11–22

C. Liu et al.

et al., 2015; Garzon-Garcia et al., 2016, Lu et al., 2016). In addition, stable isotopes have also been used to trace the sources of eroding material occurring on catchment hillslopes (Meusburger et al., 2013). The successful employment of stable isotope compositions of carbon and nitrogen are attributed to the significant difference among ecosystem pools because of isotopic fractionation during the cycling of C and N caused by physical, chemical, and biological processes (Weihmann et al., 2007; McCorkle et al., 2016). Furthermore, other tracers, particularly elemental ratios (i.e., the C/N ratio), might also be used to complement isotopic approaches for determining organic matter sources (Kendall et al., 2001; Liu et al., 2017a). The loess hilly-gully regions in northwestern China are the most severe soil erosion region in the world because of its precipitation characteristic, high soil erodibility, and intensive human activities in this region (Fu et al., 2011; Yue et al., 2016). More than 60% of the land in the Loess Plateau has subjected to soil erosion, with an average annual soil loss of 20–25 Mg ha−1 (Shi and Shao, 2000 Li et al., 2017a). Furthermore, significant amounts of nutrients associated with surface soil are lost during soil erosion in the Chinese Loess Plateau (Cai, 2001; Li et al., 2015). Approximately 0.8–1.5 kg of ammonia, 1.5 kg of total phosphorus, and 20 kg of total potassium in each ton of eroded soil are lost as estimated by Cai (2001). Carbon and nitrogen losses in the soil ecosystem not only have seriously depleted land resources and degraded the ecosystem in the Loess Plateau, but also have impacted the biogeochemical cycles of carbon and nitrogen in the terrestrial and aquatic ecosystem (Liu and Xing, 2012). With the implementation of a series of soil and water conservation measures since the 1950s on the Loess Plateau, vast areas of cropland with a slope gradient that exceeded 25° in mountainous areas are converted to forestland or grassland in the gully and hilly zones and more than 90,000 check dams are constructed in gullies and streams (Wang et al., 2016; Zhao et al., 2016). Consequently, the intensity of soil erosion has been greatly mitigated and the sediment export to lower reaches of Yellow river has decreased significantly over the past six decades (Miao et al., 2010; Gao et al., 2010). As the most widespread and effective strategy to reduce soil and water loss, check dams not only trap all of the sediments that are derived from upstream soil erosion but also intercept massive amounts of SOC and TN in the alluvial wedges, which form an important C or N sink area (Boix-Fayos et al., 2009; Lü et al., 2012; Liu et al., 2017b). The organic matter content in sediment cores trapped by check dam can provide an important history information on soil erosion and land use change (Meyers, 2003; Chen et al., 2016). Although the information regarding the rate of sediment yield and source has been well-documented (Chen et al., 2016; Zhao et al., 2017), however, owing to the complexity and uncertainty as to the fate of the OC and TN that are exported out of eroding catchments during erosion, transport, and deposition processes, little information on soil C and N redistribution patterns and sources at the catchment scale is available on the Loess Plateau, especially as regards the N dynamic characteristics in soil erosion subsystem. Therefore, the main objectives of this study were to: 1) determine the potential sources of eroded OC and TN in sediment cores using SIAR model and isotope tracing, 2) analyze the spatial variation of eroded OC and TN sources, 3) quantify the soil OC and TN mobilized by erosion at the catchment scale. The results will provide scientific basis on the study of C and N budgets and implementation of valid soil conservation strategies at larger scale catchment.

but also contributing to the aggravation of global warming and water eutrophication (Lal., 2004; Liu et al., 2017a,b). As estimated globally, soil erosion redistributes on the order of 75 Gt (1Gt = 1015g) of soil and 1–5 Gt soil organic carbon (SOC) (Stallard, 1998; Berhe et al., 2007) and 23–42 Tg soil total nitrogen (TN) (1Tg = 1012g) from soil ecosystem annually (Quinton et al., 2010). However, until recently, with the removal of C or N-rich topsoil from eroding slope positions and burial at a depressional and/or protected site, the redistribution of SOC and TN induced by soil erosion still have been intensely debated as to the fate of the C or N that is exported out of eroding catchments (Kirkels et al., 2014; Doetterl et al., 2016). The lateral flows of topsoil by erosion can result in the depletion of labile C and N at eroding site and the concentration of both sediment and SOC and TN at depositional site downslope, thereby indicating a selective mobilization in the light fractions of SOC and TN at the slope scale (Kirkels et al., 2014; Berhe et al., 2017). However, the redistribution of C and N by lateral flows at the watershed scale seem to be much more complex due to the interference of other ecogeomorphological processes (Boix-Fayos et al., 2015, 2017). Erosion process, which redistributes sediment and organic matter dominated by tillage, rill, and gully erosion, will be non-selective on catchment scale (Van Oost et al., 2008). The sediment generated by non-selective erosion is considered to have identical characteristics as the soil from which it derived (Van Oost et al., 2008; Kuhn et al., 2009). At the catchment scale, the bulk of the OC and TN fractions are transported with the sediment mobilization, which lead to the average C and N enrichment ratios (including both suspended and bedload sediment) lower than 1 as reported by previous studies (Haregeweyn et al., 2008; Wang et al., 2010; Liu et al., 2017b). Hence, the fate of eroded SOC and TN still remain considerable uncertainty owing to the multiple control factors, such as depth-related soil conditions (temperature, moisture and aeration, etc.) (Kuhn et al., 2009; Quinton et al., 2010), the stability of soil organic matter (SOM) and associated stabilisation mechanisms, as well as the possible differing nature of SOM in sediments as derived from different delivering locations within a catchment (Nadeu et al., 2012, 2014; Kirkels et al., 2014). Source pools of TOC and TN from different landform positions can affect the persistence of eroded C and N in deposition sites. For instance, soil material eroded from the surficial soils by sheet erosion might have a large proportion of relatively undecomposed OM (i.e., active OM pool which is easy to decompose and mineralize during transport or post-deposition downstream) with a high C and N contents (Wang et al., 2010; McCorkle et al., 2016). Gullies can produce a large amount of sediments, but soil materials by rill or gully erosion may contain very little C and N contents with a high proportion of relatively decomposed OM (i.e., passive OM pool which is difficult to decompose and mineralize by associating with soil minerals through aggregation or sorptive interactions during transport or after deposition downstream) (McCorkle et al., 2016). As a result, the soil material eroded from deeper soil layers should be stabilized compared to soil material eroded from surficial soil horizons during erosion, transport, and post-deposition processes (McCorkle et al., 2016; Li et al., 2017a,b; Nie et al., 2018). Moreover, although sediment OC and TN can be delivered from the terrestrial watershed (allochthonous input), the net primary production of plants on site is also likely to have a high proportion contribution to C and N pools in eroded sediments at deposition site (autochthonous input) (Liu et al., 2017a). Therefore, quantification of the fluxes of C or N between the uplands and alluvial plain induced by soil erosion, as well as identification of source materials contributing to sediment C and N within a catchment, are essential to determine C and N budgets, develop prediction models and implement soil conservation strategies at catchment scale. Carbon and nitrogen stable isotope ratios (δ13C, δ15N) and elemental compositions are a potentially powerful tool to trace the contributions of catchment sources to sediment and to in-stream particulate organic matter export out of the catchment (Liu and Xing, 2012; Laceby

2. Materials and methods 2.1. Study area The Luoyugou watershed (34°34′–34°34′N, 105°30′–105°45′E) is located in the northern suburbs of Tianshui city in the Gansu Province of China, and covers an area of 72.3 km2. This watershed is situated in the southwestern hilly-gully region of the Chinese Loess Plateau 12

Agriculture, Ecosystems and Environment 253 (2018) 11–22

C. Liu et al.

Fig. 1. Location of studied catchment.

out since the 1950s. The control practices mainly included grazing exclusion, constructing check dams, terraces building, and afforestation, as shown in Fig. 1. In addition, a road was constructed in the middle part of catchment in order to connect nearby villages. A dry main stream and four tributaries have been formed in gully and the flow direction is from north to south. Moreover, a sediment-trapping reservoir (ie., check dam) was constructed in 2004 at the outlet of this catchment, and large amounts of sediments were retained with the depth of 1 m. The dominate land use types are forests and terraced cropland, followed by sparse pastures and fallows. The forests is distributed on the gully heads and ridges of the catchment, whereas the terraced cropland is situated at the lower latitude in this catchment (Figs. 1 and 2).

(Fig. 1). This region is characterized as a semi-humid continental monsoon climate, with an average annual precipitation of 533.7 mm (Xin et al., 2016). More than 80% of precipitation is mainly concentrated on the rainy season (i.e., from May to October) in the form of intense, short duration rainstorms, leading to severe soil erosion and high sediment yield (Li et al., 2017a; Liu et al., 2017b). Historically, the annual sediment yield can reach 4851.6 Mg km−2 yr−1 in the Luoyugou watershed. In this area, the average annual temperature is approximately 10.7 °C, with a minimum mean temperature of −2.3 °C in January and a maximum of 22.6 °C in July. The main soil type is dominated by Calcaric Cambisols, occupying more than 90% of the watershed area, and Calcaric Fluvisols cover 3% of the watershed area according to the recent soil classification system of FAO (2014). The soils in this region have primarily developed from loess parent materials and have a silty loam texture, which are very susceptible to water and tillage erosion (Xin et al., 2016; Li et al., 2017a,b). As a well-known typical watershed of the Loess Plateau, since the 1950s a series of soil and water conservation measures have been implemented in the Luoyugou watershed, consisting of terracing, several phases of reforestation, grassland, and 24 check-dams in order to control soil and water loss from watershed. The forests are mainly dominated by Robinia pseudoacacia and Pinus tabulaeform, which are situated on hill tops and gully heads. The abandoned grassland may be found in mountainous areas with steep terrain in the form of level benches, level terraced fields, and fish-scale pits. The terraced cropland is the dominant landscape, which is divided by gullies covered with ecological forest and widely distributed at gentle plateau surface covered by loess soil throughout the Luoyugou watershed. In addition to cereals such as wheat, corn, or sorghum, etc., the orchards (i.e., mainly cherries (Cerasus pseudocerasus), apples (Malus pumila Mill.), peaches (Pyrus sorotina), and apricots (Prunus armeniaca L.)) are planted in the terraced cropland. Moreover, the check dams are constructed on the gullies for flood control and sediment retention. In this study, the Xijiazhai catchment (34°38′N, 105° 43′ E), delimited by check dam (3.1 km2) and located in the upstream within the Luoyugou watershed, was selected to estimate the soil C and N loss and sources based on the sediment trapped by the dams (Fig. 1). In this catchment, several soil and water conservation measures were carried

2.2. Field sampling Prior to sampling, the land use and check dam construction history were investigated by consulting local villager and government manager. Field sampling in the Xijiazhai catchment was undertaken in August 2016. In this catchment, natural fallow land hardly contributes sediments (Chen et al., 2016). The source materials that might be potentially mobilized by rainfall events and transported downstream were sampled in this catchment. A total of 54 sampling locations of soil source materials were established from eroding areas representative of cropland (18), forests (18), and gullies (18). Replicate and triple samples of forests and cropland were collected at 0-0.1 m depth at each of three landform positions along the transects: upper, middle, lower. At each sampling site, at least five samples were taken and combined in the field to form a representative composite sample by using a 5 m2 plot at random (Fig. 2). The gullies from stream bank soils and stream bed materials were also collected to evaluate these locations as possible sources of eroded material along the main stream. At each location, stream bank samples were collected on an approximately level surface within 0.1 m of the exposed bank face of the stream to maintain soil profile continuity at all augered point. Stream bed materials were also collected within a 0.1 × 0.1 m area to a depth of 0.1 m using a flat trowel in the middle of the stream path in this study. In addition, 11 terrestrial C3 plants (i.e., mainly tree leaves including pine, rice, black 13

Agriculture, Ecosystems and Environment 253 (2018) 11–22

C. Liu et al.

Fig. 2. Land use and sampling points of the study catchment.

(Chen et al., 2016). Moreover, the width, length, and thickness of the sediment wedge behind the check dam were measured in the field and undisturbed bulk samples were also taken from the retained sediments to estimate the average BD and SMC.

locust, cherry, and gully and non-gullied grasses) and 8 terrestrial C4 plants (mainly including maize and sorghum) were also collected during sampling period. Moreover, the undisturbed samples with three replicates for estimating bulk density (BD) and soil moisture content (SMC) of soils in rings of 100 cm3 were taken. All sampling sites were georeferenced in situ with a GPS Trimble GeoXM with differential correction. The sediment sampling was performed from head to tail along the sediment wedge silted by check dam. A total of 90 sediment samples (10 depths, 3 profiles, and 3 replications) were collected using a 7.0 cmdiameter soil sampler for the physio-chemical properties analysis. Three entire soil profiles perpendicular to the channel were dug and sampled to a depth of 1 m at 0.1 m depth intervals, on the apex (post-check dam: site A3), center (mid-check dam: site A2), and close (pre-check dam: site A1) to the check-dam (Fig. 2). in each transect three replicate samples were taken, on the extremes of the transect and in its center

2.3. Laboratory analysis Prior to laboratory analysis, visible root residues and stones were removed manually from the soil and sediment samples. All soil and sediment samples were air-dried, then crushed and sieved through a 2mm mesh sieve to remove coarse fragments. Afterwards, the samples were then passed through a 0.15-mm mesh to determine TOC and TN contents as well as their isotopic signatures (Liu et al., 2017c). The soil pH was determined using HI 3221 pH metre with a soil–water ratio of 1:2.5 (Hanna Instruments Inc., USA). The BD and SMC were determined gravimetrically by oven drying the whole bulk sample at 105 °C for 14

Agriculture, Ecosystems and Environment 253 (2018) 11–22

C. Liu et al.

Table 1 Characteristics of soils and sediments sampled in August 2016 at the studied catchment. Source types

Landscape position

Bulk density (Mg m−3)

Soil moisture content (g 100 g−1)

pH

Forests

Upper Middle Lower Upper Middle Lower Plain

1.33 1.37 1.36 1.13 1.14 1.24 1.35

10.40 ( ± 2.11) 8.67 ( ± 2.40) 10.14 ( ± 4.17) 9.12 ( ± 2.32) 7.91 ( ± 2.28) 8.66 ( ± 1.47) 22.56 ( ± 9.51)

7.89 7.91 7.90 8.14 8.13 8.00 8.32

Cropland

Gullies

( ± 0.19) ( ± 0.14) ( ± 0.12) ( ± 0.15) ( ± 0.08) ( ± 0.08) ( ± 0.13)

( ± 0.42) ( ± 0.42) ( ± 0.39) ( ± 0.03) ( ± 0.18) ( ± 0.05) ( ± 0.18)

Clay (g 100 g−1)

Silt (g 100 g−1)

Sand (g 100 g−1)

17.27 16.15 19.44 22.11 24.18 21.46 26.83

35.75 33.44 36.83 42.50 41.85 41.48 39.22

46.98 50.41 43.74 35.40 33.97 37.07 33.95

( ± 2.41) ( ± 4.94) ( ± 7.07) ( ± 2.49) ( ± 1.93) ( ± 2.45) ( ± 7.56)

( ± 4.75) ( ± 8.69) ( ± 3.56) ( ± 1.41) ( ± 1.70) ( ± 0.19) ( ± 4.43)

( ± 6.06) ( ± 13.47) ( ± 4.24) ( ± 3.42) ( ± 1.92) ( ± 4.48) ( ± 9.24)

Note: All values are averaged with standard deviations in parentheses (n = 18).

out to understand the relationships between the basic attributes of the sediments (texture, BD, pH, SMC SOC, and TN). Significant differences were evaluated among potential sources and sediment at depths via using the one-way analysis of variance (ANOVA) and the least significant differences (LSD). In our study, in order to determine the relative contributions to sediment organic carbon and nitrogen in sediments trapped by check dam from potential sources (i.e., forests, cropland, and gullies), we graphed δ13C versus SOC and δ15N versus TN to identify potential sources of eroded organic carbon and nitrogen, and an isotope mixing model (SIAR model) was then used to quantify the different source contributions to sediment organic carbon and nitrogen using 13C, 15N, TOC and TN with the concentration dependency of 13C and 15N (Jackson et al., 2009; Parnell et al., 2010; Hopkins and Ferguson, 2012; Garzon-Garcia et al., 2016).

48 h. The sample grain size (clay, silt, and sand fractions) was measured using a laser particle size analyzer (Mastersizer 2000, Marlven, Ltd. UK). Briefly, 20 g dry mineral soil/sediment (< 2 mm) were added with 10% H2O2 to eliminate the organic matter, and dispersed by ultrasound shaking overnight in a 0.1 L solution of sodium hexametaphosphate (5 g L−1). The measurements spanned sizes from 0.02 to 2000 μm. Moreover, the soil and sediment samples were treated repeatedly with 1 M HCl solution to remove carbonates, then rinsed with de-ionized water until the filtrate pH was neutral. Following the HCl treatment, samples were oven dried at 60 °C for 48 h for isotope analysis. The total organic carbon and total nitrogen were determined by the elemental analyzer (FLASH EA1112, Thermo); the C and N isotope ratios were measured with an isotope ratio mass spectrometer. The δ13C and δ15N values were reported relative to the Vienna PeeDee Belemnite standard (V-PDB) and atmospheric N2, respectively (Lu et al., 2016). The average analytical error was typically 0.01‰ for δ13C and 0.01‰ for δ15N (Liu et al., 2017a, c). All sample analysis was conducted at the State Key Laboratory of Soil Erosion and Dryland Farming on the Loess Plateau, Institute of Soil and Water Conservation, Chinese Academy of Sciences, China.

3. Results 3.1. Physio-chemical characteristics of soils and sediments The physio-chemical properties of soils at different potential sources and landscape positions are shown in Table 1. The soil BD (0–0.1 m depth) differed significantly among forests, cropland and gullies, with the highest value of BD in forests (1.35 ± 0.15 Mg m−3) and the lowest value in cropland (1.16 ± 0.11 Mg m−3), respectively. However, no significant difference was found at all landform positions (Table 1). SMC in gullies was higher compared with other land uses. Meanwhile, no regular values were found at the landform positions. The soils were slightly alkaline, with gullies being the highest values (pH = 8.32) and forests being the lowest values (7.90). Regarding the soil particles in different source types and landform positions, a predominance of silt (on average, 41.08 g 100 g−1 of the total), followed by sand (34.86 g 100 g−1) and clay (24.06 g 100 g−1) was observed, with the sand proportion being much higher in forests than cropland or gullies (Table 1). The BD in sediments trapped by check dam declined as depth increased with a slight fluctuation at the three part of wedge (pre-check dam (site A1), mid-check dam (site A2), and post-check dam (site A3)) (Fig. 3). The average BD reduced gradually from post-check dam (site A3) to pre-check dam (site A1) along the flow pathway (1.43 Mg m−3, 1.38 Mg m−3, 1.36 Mg m−3), which were higher than those of soils in cropland and were similar to those of soils in forests and gullies. Combined with soils in source materials, the SMC in sediments was higher at all depths (from 16.01 to 40.05 g 100 g−1) of different parts of wedge, which also showed a tendency in increase with depth. The pH values displayed little variation trends as depth increased at different parts of wedge, with varying from 8.06 to 8.22 at pre-check dam (site A1), from 8.09 to 8.47 at mid-check dam (site A2), and from 8.14 to 8.38 at post-check dam (site A3) (Fig. 3). The pH values in sediments were closer to those of cropland and higher than those of forests. A higher proportion of silt (41.1 g 100 g−1) was also found in sediments deposited by check dam, followed by sand (30.9 g 100 g−1) and then clay (27.9 g 100 g−1). The particle size distribution in sediments were

2.4. Organic carbon and total nitrogen mobilization by soil erosion The estimation of the eroded C and N were based on the sediment sampling and the estimated sediment trapped by check dam. In this study, the volume of sediment was estimated by assuming that the alluvial wedge had the form of a prismatic channel with a rectangular section (Castillo et al., 2007). The mass of sediment (Mg) retained by check dam was calculated based on the volume of sediments (m3) trapped and the average of bulk density (Mg m−3) of the sediment at the alluvial wedge. The specific sediment yield data (Mg ha−1 yr−1) was assessed by the total mass of sediments retained by each checkdam, the trap efficiency of check dam, the drainage area and time of check dam. The trap efficiency (TE) of check dam was calculated by a curve that related TE to the structure capacity–watershed area ratio in the context of background information on check dam obtained from local government, referring to the estimation of the Mid- and long term trap efficiency in the small reservoir when no inflow data area was available (Brown 1943; Boix-Fayos et al., 2009). The eroded C (Mg C ha−1 yr−1) and N (Mg N ha−1 yr−1) were calculated as the product of the mass of eroded sediment (kg) and the C and N contents (g kg−1) measured in the sediment, related to the drainage area and the year of check dam construction (Boix-Fayos et al., 2009). 2.5. Statistical analyses and source apportionment modelling All statistical tests were carried out using SPSS version 18.0 (SPSSInc., Chicago, IL, USA). The mean ± standard deviation of physio-chemical properties (BD, SMC, pH, and soil grain size), organic carbon and nitrogen and their isotopic signatures were conducted (or occasionally, ± 95% Confidence Interval). Pearson liner analysis was also used to perform correlation among the soil physio-chemical variables, eroded C and N. Principal Component Analysis (PCA) was carried 15

Agriculture, Ecosystems and Environment 253 (2018) 11–22

C. Liu et al.

Fig. 3. The profile distribution of bulk density, soil moisture content, and pH in sediment cores (A1: Pre-check dam; A2: Mid-check dam; A3: Post-check dam).

Fig. 4. The profile distribution of particle size in sediment cores (A1: Pre-check dam; A2: Mid-check dam; A3: Post-check dam).

soils (i.e., cropland and gully) were isotopically more similar to sediments in superficial layer (0–0.2 m depth) (Fig. 7). The δ13C values showed enrichment with fluctuation as depth increased on different transects in the wedge of check dam. Values of δ15N varied greater than δ13C value with depth at different sediment transects (Fig. 8). Some depth profiles even showed 15N depletion with depth (for example, the locations of mid-check dam and post-check dam). On different transects the 13C values in sediments were highest at pre-check dam, followed by mid-check dam and then post-check dam at each sediment layer, except for in the 0.6–1.0 m layer, which showed more enrichment at the post-check dam than at the mid-check dam (Fig. 8a). No regular variation was obviously found on three transects at the sediment wedge in the values of 15N. However, at deeper sediment depth (0.8–1.0 m depth) the variation of 15N values on three transects were similar to the values of 13C, showing the decrease in the following order of pre-check dam > post-check dam > mid-check dam (Fig. 8b).

similar to those of cropland (Fig. 3 and Table 1). In general, the sediments in the middle part of the wedge (23.2 g 100 g−1) had a lower proportion of fine material than those from the upper (29.7 g 100 g−1) or lower part (31.0 g 100 g−1). No significant variation at profile depths at the pre-check dam (CV = 15%) and mid-check dam (10%) except for post- check dam (21%) were found (Fig. 4). 3.2. Elemental compositions and isotopic signatures of soils and sediments The TOC and TN contents for the 0–0.1 m depth in soils and sediments are presented in Fig. 5a and b. Contents of TOC and TN significantly differ with source materials and sediment. The C and N contents in soils were the highest in forests and the lowest in sediments in the deposition zone. Moreover, the contents of C and N in soils of gullies were similar to those of sediments in deposition (Figs. 5a and b). However, little difference in atomic carbon to nitrogen ratio (C:N ratio) was found among source materials and sediment (Fig. 5c). The contents of carbon and nitrogen in sediments trapped by check dam at the three parts of wedge are presented in Fig. 6. There was significant linear correlation relationship between C and N content (r2 = 0.90). The average C and N contents at profile depths in sediments increased along the flow pathway (i.e., average TOC from 3.59 g kg−1 for post-check dam to 5.03 g kg−1 for pre-check dam and average TN from 0.41 g kg−1 for post-check dam to 0.55 g kg−1 for precheck dam, respectively). No obvious regular variations in both C and N contents were found as depth increased, varying from 2.47 to 7.07 g kg−1 for TOC and from 0.31 to 0.72 g kg−1 for TN at profile depths (Fig. 6). The isotopic compositions of carbon and nitrogen in SOM of soils and sediments were distinct from each other (Fig. 7). The carbon and nitrogen isotopic compositions of soils (e.g., cropland and gully) had higher values by 1 to 2‰ than that of soils which came from forests. The mean of δ13C and δ15N in sources (-26.41 ± 0.34‰, 2.86 ± 0.49‰) were statistically more depleted than those of sediments at different depths, which ranged from −26.24 to −24.84‰, with an average δ13C of −25.54 ± 0.69‰ for δ13C and varied from 3.38 to 3.85‰, with an average δ15N of 3.62 ± 0.42‰. The mineral

3.3. Apportionment of eroded organic carbon and nitrogen sources In order to explore the sources of organic C and N in sediments eroded by soil erosion from the potential sources (i.e., forests, cropland, and gully), the source discrimination potential of the measured properties was tested prior to modelling. δ13C, δ15N, TOC, and TN discriminated between all the source materials (Table 2). The significant differences in δ13C, δ15N, TOC, and TN among source materials were found in our study, except for between cropland and gully for δ15N. Therefore, these four sediment properties in combination were employed to model the source contributions quantitatively owing to the complete discrimination among all the sources. Moreover, the source materials were plotted with sediment samples in the dual relationship between δ13C and TOC, δ15N and TN, as well as between δ13C and δ15N in Figs. 7 and 9 . The results indicated cropland and gully were the primary potential TOC and TN sources in the studied region. The mixing model results supported the findings derived from the qualitative descriptions above. The substantial variability in the C and N contributions from the three sources to the individual sediment samples are shown in Fig. 10a and b. The mean relative contribution 16

Agriculture, Ecosystems and Environment 253 (2018) 11–22

C. Liu et al.

Fig. 5. Soil organic carbon and nitrogen contents as well as C:N ratio in the top 0.1 m of mineral soils and sediments from all sampling locations.

source type were also found, especially at pre-check dam (site A1) and post-check dam (site A3), as shown in Fig. 10a and b, thereby indicating the variations in hydrodynamic conditions, soil erosion, and land use in this region.

calculated of each source type to the sediment C indicated that cropland was the dominate source type, which contributed 53.54%, followed by gully (29.28%) and then forests (17.18%), respectively (Fig. 10a). The result of sediment N contributions was similar to C contributions, thus showing that cropland contributed 53.53%, with gully and forests contributing 30.86% and 15.61%, respectively (Fig. 10b). The relative C and N contributions of each source type varied from one location to the next near the check dam, and both the mean C and N contributions were similar in the direction of the runoff pathway (Fig. 10a and b). The forests contributions gradually decreased from post-check dam (site A3) to pre-check dam (site A1). In mid-check dam (site A2), the C and N contributions of cropland were higher than those of pre-check dam (site A1) and post-check dam portions (site A3), whereas the gullies contribution was highest in pre-check dam (site A1), followed by postcheck dam (site A3) and then mid-check dam (site A2). The vertical variations obviously of sediment C and N contributions from each

3.4. Soil carbon and nitrogen mobilized during erosion and deposition processes The TOC and TN mobilized by water erosion at the studied subcatchment were estimated compared with the total erosion in the surface soil layer (0-0.1 m depth). The total amount of the sediment was estimated to be approximately 30.96 × 104 Mg trapped by the check dam according the total silted volume and average bulk density, suggesting that the average annual sediment yield was about 9.99 Mg ha−1 yr−1 in the catchment (Table 3). A higher sediment yield generally led to greater TOC and TN in the

Fig. 6. The profile distribution of organic carbon and nitrogen in the sediment cores (A1: pre-check dam; A2: mid-check dam; A3: post-check dam).

17

Agriculture, Ecosystems and Environment 253 (2018) 11–22

C. Liu et al.

significantly and negatively correlated with pH (p < 0.05), thus indicating sedimentary condition was also major factor influencing soil organic carbon and nitrogen pool in sediments in the deposition zone.

4. Discussion 4.1. Organic carbon and nitrogen mobilized by erosion at the catchment scale The TOC and TN contents in the sediments (4.54 ± 1.06 g kg−1, 0.50 ± 0.10 g kg−1) were, in general, much lower than those of the forest soils (21.21 ± 1.14 g kg−1, 1.97 ± 0.09 g kg−1) and cropland soils (8.77 ± 0.54 g kg−1, 0.93 ± 0.09 g kg−1) of the catchment and were closer to the average OC and TN contents of the gully soils (4.51 ± 1.01 g kg−1, 0.51 ± 0.09 g kg−1). The eroded sediments retained by check dam were impoverished in OC and TN compared with the corresponding source soils in the loess hilly-gully catchment. The results were in consistent with previous studies at the catchment scale in Mediterranean ecosystems, in which the enrichment ratios for TOC and TN contents were lower than 1, although the exported sediments were slightly enriched in clay and sand particles (Boix-Fayos et al., 2009, 2017). However, these enrichment ratios were lower, when compared to the values reported in the literature referring to recently exported sediments due to water erosion in plot scale or at smaller catchment (Owens et al., 2002; Quinton et al., 2006; Rhoton et al., 2006; Li et al., 2017a). The difference can be attributed to the different erosion processes mobilizing sediments at different spatial scales (BoixFayos et al., 2017). On the one hand, the C and N oxidation in eroded sediment are highly correlated with transport distance and geomorphological characterisitic, thus indicating that longer transport distance and more complex geomorphology will result in higher rate of mineralization during sediment mobilization and transport within the catchment compared with the erosion plots (Kirkels et al., 2014). On the other hand, the type of erosion affects the sources of sediments and how sediments are connected to the fluvial channel and transported, thereby determining the enrichment or depletion of OC and TN in sediments with respect to the reference soils (Nadeu et al., 2012; BoixFayos et al., 2015). For instance, at slope landscape scale, sheet and interrill erosion by overland flow mobilizes surface mineral soils with relatively small mass but high C and N contents. In addition to sheet and interrill erosion, gully erosion will produce a large amount of sediment with poor C and N contents from deep soil layer at the catchment scale (McCorkle et al., 2016; Liu et al., 2017a). Moreover, other types of erosion, gravitational erosion (resulting in landslides), can mobilize soil materials from the entire profile, along with C and N depleted saprolite and bedrock in the hill catchment (Hilton et al.,

Fig. 7. Stable isotope (δ13C and δ15N) graphs of different potential source types and collected sediments at different depths (0–1.0 m).

sediments. As estimated by the mass eroded sediment and sediment C or N content, 0.14 × 104 Mg C and 0.02 × 104 Mg N were retained by check dam, about 0.45% and 0.05% of the total amount of sediment, respectively, suggesting an average soil C and N storage rates of 0.045 Mg C ha−1 yr−1and 0.005 Mg N ha−1 yr−1 for the studied subcatchment (Table 3). A total of 1569.8 Mg TOC and 146.7 Mg TN were lost during the soil erosion processes at an average soil C and N erosion at a rate of 0.051 Mg C ha−1 yr−1 and 0.005 Mg N ha−1 yr−1 during the period (i.e., from 2004 to 2016), accounting for approximately 52.8% and 48.87% of the total amount of eroded TOC and TN, respectively (Table 3), whereby indicating that although a large amount of C and N were buried in the sediments of wedge trapped by check dam in the studied catchment, a relative proportion of C and N might be exported during erosion and deposition processes. In order to further explore the associations between the organic carbon and nitrogen pools and other sediment characteristics, principal component analysis was conducted in this study (Fig. 11). Accounting for 84.9% of the variance from both major components, the first component separated sediment wedges with high TOC, TN, silt, and SMC contents from those rich in BD. The second component separated the sediment wedges rich in clay and pH contents from those and rich in sand. From pearson analysis, TOC and TN pools in sediments showed highly significant positive correlation with SMC as well as silt and negative correlation with sand (p < 0.01), and the TOC and TN pools

Fig. 8. Depth profiles of δ13C and δ15N of sediments at different locations (A1: Pre-check dam; A2: Mid-check dam; A3: Post-check dam) from wedge trapped by check dam.

18

Agriculture, Ecosystems and Environment 253 (2018) 11–22

C. Liu et al.

Table 2 Carbon and nitrogen isotopic and elemental compositions of potential sources (i.e., forests, cropland, and gully) in the studied watershed. Sources

TOC (g kg−1)

TN (g kg−1)

δ13C (‰)

δ15N (‰)

C/N

Forests Cropland Gully

21.21 ( ± 1.14)a 8.77 ( ± 0.54)b 4.51 ( ± 1.01)c

2.01 ( ± 0.09)a 0.93 ( ± 0.09)b 0.51 ( ± 0.09)c

−27.24 ( ± 0.17)c −25.61 ( ± 0.30)a −26.39 ( ± 0.76)b

1.85 ( ± 0.24)c 3.67 ( ± 0.31)a 3.06 ( ± 0.45)ab

10.84 ( ± 0.11)a 9.38 ( ± 0.27)bc 8.84 ( ± 0.96)c

Note: TOC, total organic carbon; TN, total nitrogen; C/N, total organic carbon to total nitrogen ratio. All values are averaged with standard deviations in parentheses. Different letters (a, b, and c) indicate significant different among potential sources at p < 0.05 (n = 18 for forests, 18 for cropland, and 18 for gullies, respectively).

landscapes likely functioned as a carbon source, a considerable proportion of OC was buried in a variety of depositional basins (Berhe et al., 2007; Ran et al., 2014; Liu et al., 2017a,b). In this studied catchment, a total of 1405.1 Mg eroded TOC was retained by check dam at the storage rate of 0.045 Mg C ha−1 yr−1, accounted for approximately 47.2% of the TOC mobilized by soil erosion. This finding was consistence with the observation of previous studies, which the important carbon sink function of check dams (Boix-Fayos et al., 2009; Wang et al., 2011; Liu et al., 2017a). The main reason was attributed to the favourable condition to favour carbon sequestration in check dams through burial, and new and reconfigured associations of the eroded OM with soil minerals (Berhe et al., 2007; Berhe and Kleber, 2013).

2008). In all, in our hill-gully catchment with complex topography and rainfall properties, the possibilities of various types of erosion led to the depletion of C and N in eroded sediment relative to catchment source materials. The average TOC erosion rate in our studied catchment (0.096 Mg C ha−1 yr−1) was close to those reported for subhumid Mediterranean catchments − such as 0.07 Mg ha−1 yr−1 (Smith et al., 2005) and 0.06 Mg ha−1 yr−1 (Boix-Fayos et al., 2009) − but was lower in comparison with other studies which reported the average rate of TOC erosion in humid ecosystems (0.113 Mg ha−1 yr−1 (Izaurralde et al., 2007)) and in subarid region of Chinese loess plateau (0.36 ± 0.16 Mg ha−1 yr−1 (Li et al., 2015)). The average TN erosion rate (0.010 Mg N ha−1 yr−1) in our study was sightly higher than the value reported by Berhe et al., (2017), which indicated that erosion redistributed 0.0025 to 0.0047 Mg N ha−1 yr−1 from the eroding landform positions in hillslope environments. Considerable TOC amounts on the Loess Plateau are redistributed and affected by soil erosion each year, which have significant effect on the local ecological security and biogeochemical cycling of elements (Cai, 2001; Wang et al., 2011, 2014). In this catchment, a total of 1569.8 Mg TOC was lost during the soil erosion processes at an average soil C and N erosion at a rate of 0.051 Mg C ha−1 yr−1, accounting for approximately 52.8% of the total amount of TOC from the eroding areas. We obtained this result based on the minor changes in the TOC concentration observed between present cropland surface soil data and historical data collected thirty years ago. This finding validated the commonly held assumption that 20–40% at the displaced OC would be lost in agricultural system because of selective erosion of organic C and subsequent mineralization during transport and landscape deposition, although a part of SOC is protected by burial, estimated at 0.4–0.6 Gt C yr−1 (Lal, 2003, 2004). The light OM fractions (i.e., labile organic carbon) from the O horizon and surface mineral soil was preferentially removed by surface runoff over slope scale, which will increase the mineralization of soil organic matter during soil OC transportation and deposition processes (Lal, 2003; Li et al., 2017b). Although erosion-induced OC transportation at agricultural

4.2. Controls on sources contribution at the catchment scale In order to quantify the flux of C and N among different C and N pools and determine the fate of eroded C and N at the catchment scale accurately, identifying the sources of organic C and N in eroded sediments by deposition is of great importance (McCorkle et al., 2016; Berhe et al., 2017). In our study, three potential sources (i.e., forests, cropland and gullies) were successfully quantified based on an isotope mixing model with δ13C, δ15N, and TN contents. The cropland was the main source contribution to sediment C and N, followed by gully and forest (Fig. 11a and b). Our results were consistent with previous studies reported by Wang et al. (2017) recently, indicating that the majority of C mainly derived from cropland, even though gullies system and steep slopes dominated the catchment topography in an intensive agricultural catchment. This phenomenon might be explained by certain governmental policies. For example, the Household Responsibility System was established in 1984 as the basic institution in rural China. The land was assigned to the farmer under this system for at least 15 year with no reassignment during the time of the contract (Krusekopf, 2002). As a result, farmers were highly motivated to manage land, which would lead to surface crushing and aggravate soil erosion. When intensive heavy rainfall events occurred, the cropland would generate high sediment contributions, which was the major

Fig. 9. Mean TOC and δ13C along with TN and δ15N for sediments at different depths (0–1.0 m) and potential sources for the studied catchment. Horizontal and vertical bars denote ± 95% confidence intervals.

19

Agriculture, Ecosystems and Environment 253 (2018) 11–22

C. Liu et al.

Fig. 10. Mean percent contributions to total organic carbon (a) and total nitrogen (b) in eroded sediments for each profile from forests, cropland and gully.

contributor to sediment OC. Moreover, erosion in the gullies typically decreased when the in-stream sediment loads were large (Juracek and Ziegler, 2009); therefore, erosion decreased in the gully system because the surface runoff carried substantial amounts of sediments derived from cropland areas. Moreover, the source of organic matter in eroded sediment might vary with erosion types. For instance, gullies can produce a large amount of sediment but the eroded material is likely to contain very little C (deep soil and associated OM) (McCorkle et al., 2016). The constructions of road and terraces would destroy and loosen surface soils in cropland further and then aggravate soil erosion when extreme events happen, thereby resulting in more sediment and associated OM pouring to the check dams downstream (Wang et al., 2014). The relative C and N contents of each source type varied from one location to the next near the check dam, and both the C and N contributions were similar in the direction of the runoff pathway. The reason might be attributed to variations in hydrodynamic conditions (Wang et al., 2014). In our study, there were large variations in physiochemical properties in the different sediment cores along the river valley (Figs. 3 and 4). The selectivity mobilization of sediment by water force along the flow pathway will result in the preferential deposition of coarse particles materials at the post-check dam and the finer particles materials moving to further distance with flow (Rhoton et al., 2006; Doetterl et al., 2016). The precipitation characteristic in studied area with relatively light rainfall events would contribute to the relatively fine sediments, and fine sediments are usually richer in SOC than are coarse sediments (Chaplot and Poesen, 2012; McCorkle et al., 2016). Moreover, the anthropogenic feature will have obvious influence on the sediment and associated sediment C source (Wang et al., 2016). For example, the construction of check dam in gullies resulted in the sediment C and N sources from gully were larger than those of other

Fig. 11. Bi-plot of the two first components from a PCA including bulk density (BD), soil moisture content (SMC), pH, textural variables, soil organic carbon (TOC), and total nitrogen (TN) pools in the sediments retained by check dam. Both components explained 84.9% of the variance.

source materials at pre-check dam. The vertical variation of sediment C and N contributions from each source type were also found, thereby indicating the temporal heterogeneity of soil erosion and land use (Chen et al., 2016; Li et al., 2017a), which led to the difference of contribution proportions to eroded C and N in the sediment profile of this catchment.

Table 3 Total erosion, mobilized total organic carbon (TOC) and total nitrogen (TN) as well as their pools at the studied catchment. Specific sediment yield (Mg ha−1 yr−1)

TOC (TN) mobilized by water erosion (×104 Mg)

TOC(TN) buried in sediments ( × 104 Mg)

TOC (TN) erosion rate (Mg ha−1 yr−1)

TOC (TN) storage rate (Mg ha−1 yr−1)

Ratio TOC (TN) erosion/total soil erosion

TOC (TN) loss rate (Mg ha−1 yr−1)

TOC (TN) loss percent (%)

9.99

0.30 (0.03)

0.14 (0.02)

0.096 (0.010)

0.045 (0.005)

0.010 (0.001)

0.051 (0.005)

52.77 (48.87)

Note: TOC, total organic carbon; TN, total nitrogen. The calculated values of TN are shown in parentheses.

20

Agriculture, Ecosystems and Environment 253 (2018) 11–22

C. Liu et al.

retained 1405.1 Mg eroded TOC and 153.5 Mg eroded TN at the storage rate of 0.045 Mg C ha−1 yr−1 and 0.005 Mg C ha−1 yr−1, accounted for approximately 47.2% of the TOC and 51.1% of the TN mobilized by soil erosion, indicating that although a large amount of C and N were buried in sediments trapped by check dam, a considerable proportion of C and N might be exported during erosion, transport and post-deposition processes. The highly correlation relationships among TOC and TN and soil variables in sediments indicated that sedimentary condition was also major factor affecting soil carbon and nitrogen pool in the deposition zone. Further study of soil C and N redistribution and sources at the catchment scale should take some uncertainties into consideration comprehensively in the process of field sampling, indoor analyzing, and model employing.

4.3. Uncertainties analysis: eroded C and N estimation and sources appointment As illustrated in Fig. 1, we obtained the sediment profiles at sites A1 and A3 along the flow pathway in the dam-controlled watersheds. In our study, the estimation of the eroded C and N were based on the sediment sampling and the estimated sediment trapped by check dam. The volume of sediment was estimated assuming that the alluvial wedge had the form of a prismatic channel with a rectangular section (Castillo et al., 2007). It was difficult to measure the width, length, and thickness of the sediment accurately behind the check dams in the field due to the complex geomorphological feature of the gully in the hillygully region (Zhao et al., 2017). This could result in the most important uncertainty source for the estimation of sediment and C or N yield. To improve the estimation of the sediment volume, more sediment cores should be drilled within the catchment, which may be better to capture the geomorphology of the deposit. However, this method would be time consuming and expensive (D’Haen et al. 2013). Until recently, the combination of sediment coring and paleolimnological techniques to obtain accurate information on sediment volume were proposed by Alatorre et al. (2012). Using stable isotope signatures to identify the source of eroded SOM in sediments should consider several assumption that can not be met in other studies. On the one hand, the 13C and 15N compositions among the selected potential sources should show significant differences within the catchment and the 13C and 15N isotopes signatures of sediment should be within the scope of variation of different potential sources (Doetterl et al., 2016). On the other hand, the application of natural stable isotopes signatures should also consider litter kinetic isotopic fractionation effects during soil erosion, sediment transport and post-depositional processes (Lu et al., 2016; Doetterl et al., 2016). However, it was difficult to assess the extent to which erosion and postdepositional processes might have altered the tracers. In order to reduce the possible alterations of the biochemical properties (including δ13C, δ15N, TOC and TN) of source materials mainly caused by some biological processing (e.g., mineralization), source samples should processed for elemental and isotopic analysis in a manner designed to mimic stream transport (Garzon-Garcia et al., 2016). In addition, the sampling processes may also produce uncertainties, not only because of the distinct sediment sources and geomorphology of valley, but also due to the sorting of the different particles from the gully head to the dam and the frequency of intensive agricultural activities (Zhao et al., 2017; Li et al., 2017a). Therefore, further investigation should take this uncertainty mentioned above into consideration in the future in order to study C and N redistribution and sources and estimate C and N budgets more accurately at similar or larger catchment scale.

Acknowledgments Financial support for this study came from the “Hundred-talent Project” of the Chinese Academy of Sciences (A315021407) and the National Natural Science Foundation of China (41271294). The authors are very grateful to anonymous reviewers and responsible editors of this journal for valuable comments and suggestions to improve this manuscript. References Alatorre, L.C., Begueria, S., Lana-Renault, N., Navas, A., Garcia-Ruiz, J.M., 2012. Soil erosion and sediment delivery in a mountain catchment under scenarios of land use change using a spatially distributed numerical model. Hydrol. Earth Syst. Sci. 16, 1321–1334. Berhe, A.A., Kleber, M., 2013. Erosion, deposition, and the persistence of soil organic matter: mechanistic considerations and problems with terminology. Earth Surf. Process. Landf. 38, 908–912. Berhe, A.A., Harte, J., Harden, J.W., Torn, M.S., 2007. The significance of the erosioninduced terrestrial carbon sink. Bioscience 57, 337–346. Berhe, A.A., Torn, M.S., Harden, J.W., 2017. Soil nitrogen storage and stabilization in eroding landscapes. Biogeochemistry 132, 37–54. Boix-Fayos, C., de Vente, J., Albaladejo, J., Martínez-Mena, M., 2009. Soil carbon erosion and stock as affected by land use changes at the catchment scale in Mediterranean ecosystems. Agric. Ecosyst. Environ. 133, 75–85. Boix-Fayos, C., Nadeu, E., Quiñonero, J.M., Martínez-Mena, M., Almagro, M., De Vente, J., 2015. Sediment flow paths and associated organic carbon dynamics across a Mediterranean catchment. Hydrol. Earth Syst. Sci. 19, 1209–1223. Boix-Fayos, C., Martínez-Mena, M., Cutillas, P.P., et al., 2017. Carbon redistribution by erosion processes in an intensively disturbed catchment. Catena 149, 799–809. Brown, C.B., 1943. Discussion of sedimentation in reservoirs, by J. Witzig. Trans. Am. Soc. Civ. Eng. 69, 1493–1500. Cai, Q.G., 2001. Soil erosion and management on the Loess Plateau. J. Geogr. Sci. 11, 53–70. Castillo, V.M., Mosch, W.M., Garcia, C.C., Barbera, G.G., Cano, J.A.N., Lopez-Bermudez, F., 2007. Effectiveness and geomorphological impacts of check dams for soil erosion control in a semiarid Mediterranean catchment: el Carcavo (Murcia, Spain). Catena 91, 103–123. Chaplot, V., Poesen, J., 2012. Sediment, soil organic carbon and runoff delivery at various spatial scales. Catena 88, 46–56. Chen, F.X., Fang, N.F., Shi, Z.H., 2016. Using biomarkers as fingerprint properties to identify sediment sources in a small catchment. Sci. Total Environ. 557 (-558), 123–133. D'Haen, K., Verstraeten, G., Dusar, B., Degryse, P., Haex, J., Waelkens, M., 2013. Unravelling changing sediment sources in a Mediterranean mountain catchment: a Bayesian fingerprinting approach. Hydrol. Process. 27, 896–910. Doetterl, S., Berhe, A.A., Nadeu, E., Wang, Z.G., Sommer, M., Fiener, P., 2016. Erosion, deposition and soil carbon: a review of process-level controls, experimental tools and models to address C cycling in dynamic landscapes. Earth Sci. Rev. 154, 102–122. FAO, 2014. World Reference Base for Soil Resources 2014: International Soil Classification System for Naming Soils and Creating Legends for Soil Maps. World Soil Resources Reports 106. FAO, Rome, Italy. Fu, B.J., Liu, Y., Lu, Y.H., He, C.S., Zeng, Y., Wu, B.F., 2011. Assessing the soil erosion control service of ecosystems change in the Loess Plateau of China. Ecol. Complex. 8, 284–293. Gao, P., Zhang, X.C., Mu, X.M., Wang, F., Li, R., Zhang, X.P., 2010. Trend and changepoint analyses of streamflow and sediment discharge in the Yellow River during 1950–2005. Hydrol. Sci. J. 55, 275–285. Garzon-Garcia, A., Laceby, J.P., Olley, J.M., Bunn, S.E., 2016. Differentiating the sources of fine sediment, organic matter and nitrogen in a subtropical Australian catchment. Sci. Total Environ. 575, 1384–1394. Haregeweyn, N., Poesen, J., Deckers, J., Nyssen, J., Hayle, M., Govers, G., Verstraeten, G., Moeyersons, J., 2008. Sediment-bound nutrient export from micro-dam catchments in Northern Ethiopia. Land Degrad. Dev. 19, 136–152.

5. Conclusions In this study, the redistribution patterns of TOC and TN by lateral flows and their sources at sediment cores in a loess hill-gully catchment were explored based on field sampling and tracing methods. The 13C and 15N isotopes signatures as well as their element compositions (TOC and TN) have been successfully employed to identify the eroded sediment C and N qualitatively and quantitively in a loess hilly-gully region. The sedimentary TOC and TN were mostly derived from cropland surface soil, followed by gullies and forests. The phenomenon of the eroded C and N sources showing significant variations at the horizontal and vertical patterns can be attributed to the variations in hydrodynamic conditions, soil erosion, and land use in this region. In this catchment, a total of 1569.8 Mg TOC and 146.7 Mg TN were lost during the soil erosion processes at an average soil C and N erosion at a rate of 0.051 Mg C ha−1 yr−1 and 0.005 Mg N ha−1 yr−1, accounting for approximately 52.8% and 48.9% of the total amount of TOC and TN mobilized from the eroding areas. However, check dam 21

Agriculture, Ecosystems and Environment 253 (2018) 11–22

C. Liu et al.

Meusburger, K., Mabit, L., Park, J.H., Sandor, T., Alewell, C., 2013. Combined use of stable isotopes and fallout radionuclides as soil erosion indicators in a forested mountain site, South Korea. Biogeosciences 10, 5627–5638. Meyers, P.A., 2003. Applications of organic geochemistry to paleolimnological reconstructions: a summary of examples from the Laurentian Great Lakes. Org. Geochem. 34, 261–289. Miao, C.Y., Ni, J.R., Borthwick, A.G.L., 2010. Recent changes of water discharge and sediment load in the Yellow River basin. China. Prog. Phys. Geogr. 34, 541–561. Nadeu, E., Berhe, A.A., de Vente, J., Boix-Fayos, C., 2012. Erosion, deposition and replacement of soil organic carbon in Mediterranean catchments: a geomorphological, isotopic and land use change approach. Biogeosciences 9, 1099–1111. Nadeu, E., Van Oost, K., Boix-Fayos, C., de Vente, J., 2014. Importance of land use patterns for erosion-induced carbon fluxes in a Mediterranean catchment. Agric. Ecosyst. Environ. 189, 181–189. Nie, X.D., Li, Z.W., Huang, J.Q., Liu, L., Xiao, H.B., Liu, C., Zeng, G.M., 2018. Thermal stability of organic carbon in soil aggregates as affected by soil erosion and deposition. Soil Till. Res. 175, 82–90. Owens, L.B., Malone, R.W., Hothem, D.L., Starr, G.C., Lal, R., 2002. Sediment carbon concentration and transport from small watersheds under various conservation tillage practices. Soil Till. Res. 67, 65–73. Parnell, A.C., Inger, R., Bearhop, S., Jackson, A.L., 2010. Source partitioning using stable isotopes: coping with too much variation. PloS One 5, e9672. Quinton, J.N., Catt, J.A., Wood, G.A., Steer, J., 2006. Soil carbon losses by water erosion: experimentation and modeling at field and national scales in the UK. Agric. Ecosyst. Environ. 112, 87–102. Quinton, J.N., Govers, G., Van Oost, K., Bardgett, R.D., 2010. The impact of agricultural soil erosion on biogeochemical cycling. Nat. Geosci. 3, 311–314. Ran, L., Lu, X.X., Xin, Z., 2014. Erosion-induced massive organic carbon burial and carbon emission in the Yellow River basin, China. Biogeosciences 11, 945–959. Rhoton, F.E., Emmerich, W.E., Goodrich, D.C., Miller, S.N., McChesney, D.S., 2006. Soil geomorphological characteristics of a semiarid watershed: influence on carbon distribution and transport. Soil Sci. Soc. Am. J. 70, 1532–1540. Shi, H., Shao, M.G., 2000. Soil and water loss from the Loess Plateau in China. J. Arid Environ. 45, 9–20. Smith, S.V., Sleezer, R.O., Renwick, W.H., Buddemeier, R., 2005. Fates of eroded soil organic carbon: Mississippi basin case study. Ecol. Appl. 15, 1929–1940. Stallard, R.F., 1998. Terrestrial sedimentation and the carbon cycle: coupling weathering and erosion to carbon burial. Glob. Biogeochem. Cycles 12, 231–257. Van Oost, K., Six, J., Govers, G., Quine, T.A., Gryze, S., 2008. Response to soil erosion: a carbon sink or source? Science 319, 1042. Wang, Z., Govers, G., Steegen, A., Clymans, W., Van den Putte, A., Langhans, C., Merckx, R., Van Oost, K., 2010. Catchment-scale carbon redistribution and delivery by water erosion in an intensively cultivated area. Geomorphology 124, 65–74. Wang, Y.F., Fu, B.J., Chen, L.D., Lü, Y.H., Gao, Y., 2011. Check dam in the Loess Plateau of China: engineering for environmental services and food security. Environ. Sci. Technol. 45, 10298–10299. Wang, Y.F., Chen, L.D., Fu, B.J., Lü, Y.H., 2014. Check dam sediments: an important indicator of the effects of environmental changes on soil erosion in the Loess Plateau in China. Environ. Monit. Assess. 186, 4275–4287. Wang, S., Fu, B.J., Piao, S.L., Lv, Y.H., Ciais, P., Feng, X.M., Wang, Y.F., 2016. Reduced sediment transport in the Yellow River due to anthropogenic changes. Nat. Geosci. 9, 38–3841. Wang, Y.X., Fang, N.F., Tong, L.S., Shi, Z.H., 2017. Source identification and budget evaluation of eroded organic carbon in an intensive agricultural catchment. Agric. Ecosyst. Environ. 247, 290–297. Weihmann, J., Mansfeldt, T., Schulte, U., 2007. Stable carbon (12/13C) and nitrogen(14/ 15N) isotopes as a tool for identifying the sources of cyanide in wastes and contaminated soils e a method development. Anal. Chim. Acta. 582, 375–381. Xin, Z.B., Qin, Y.B., Yu, X.X., 2016. Spatial variability in soil organic carbon and its influencing factors in a hilly watershed of the Loess Plateau, China. Catena 137, 660–669. Yue, Y., Ni, J.R., Ciais, P., et al., 2016. Lateral transport of soil carbon and land-atmosphere CO2 flux induced by water erosion in China. PNAS 113, 6617–6622. Zhao, G.J., Kondolf, G.M., Mu, X.M., Han, M.W., He, Z., Rubin, Z., Wang, F., Gao, P., Sun, W.Y., 2016. Sediment yield reduction associated with land use changes and check dams in a catchment of the Loess Plateau, China. Catena 148, 126–137. Zhao, G.J., Mu, X.M., Han, M.W., An, Z.F., Gao, P., Sun, W.Y., Xu, W.L., 2017. Sediment yield and sources in dam-controlled watersheds on the northern Loess Plateau. Catena 149, 110–119.

Hilton, R.G., Galy, A., Hovius, N., 2008. Riverine particulate organic carbon from an active mountain belt: importance of landslides. Glob. Biogeochem. Cycles 22, 1–12. Hopkins III, J.B., Ferguson, J.M., 2012. Estimating the diets of animals using stable isotopes and a comprehensive Bayesian mixing model. PLoS One 7, e28478. Izaurralde, R.C., Williams, J.R., Post, W.M., Thomson, A.M., McGill, W.B., Owens, L.B., Lal, R., 2007. Long-term modeling of soil C erosion and sequestration at the small watershed scale. Clim. Change 80, 73–90. Jackson, A.L., Inger, R., Bearhop, S., Parnell, A., 2009. Erroneous behaviour of MixSIR, a recently published Bayesian isotope mixing model: a discussion of Moore & Semmens (2008). Ecol. Lett. 12, E1–E5. Juracek, K.E., Ziegler, A.C., 2009. Estimation of sediment sources using selected chemical tracers in the Perry lake basin, Kansas. U.S.A. Int. J. Sediment. Res. 24, 108–125. Kendall, C., Silva, S.R., Kelly, V.J., 2001. Carbon and nitrogen isotopic compositions of particulate organic matter in four large river systems across the United States. Hydrol. Process 15, 1301–1346. Kirkels, F.M.S.A., Cammeraat, L.H., Kuhn, N.J., 2014. The fate of soil organic carbon upon erosion, transport and deposition in agricultural landscapes-A review of different concepts. Geomorphology 226, 94–105. Krusekopf, C.C., 2002. Diversity in land-tenure arrangements under the household responsibility system in China. China Econ. Rev. 13, 297–312. Kuhn, N.J., Hoffmann, T., Schwanghart, W., Dotterweich, M., 2009. Agricultural soil erosion and global carbon cycle: controversy over? Earth Surf. Process. Landf. 34, 1033–1038. Lü, Y.H., Sun, R.H., Fu, B.J., Wang, Y.F., 2012. Carbon retention by check dams: regional scale estimation. Ecol. Eng. 44, 139–146. Laceby, J.P., Olley, J., Pietsch, T.J., Sheldon, F., Bunn, S.E., 2015. Identifying subsoil sediment sources with carbon and nitrogen stable isotope ratios. Hydrol. Process. 29, 1956–1971. Lal, R., 2003. Soil erosion and the global carbon budget. Environ. Int. 29, 437–450. Lal, R., 2004. Soil carbon sequestration to mitigate climate change. Geoderma 123, 1–22. Lal, R., 2010. Managing soils and ecosystems for mitigating anthropogenic carbon emissions and advancing global food security. Bioscience 60, 708–721. Li, Y., Quine, T.A., Yu, H.Q., Govers, G., Six, J., Gong, D.Z., Wang, Z., Zhang, Y.Z., VanOost, K., 2015. Sustained high magnitude erosional forcing generates an organic carbon sink Test and implications in the Loess Plateau, China. Earth Planetary Sci. Lett. 411, 281–289. Li, Z.W., Liu, L., Nie, X.D., Chang, X.F., Liu, C., Xiao, H.B., 2016a. Modeling soil. organic carbon loss in relation to flow velocity and slope on the Loess Plateau of China. Soil Sci. Soc. Am. J. 80, 1341–1351. Li, Z.W., Lu, Y.M., Nie, X.D., Huang, B., Ma, W.M., Liu, C., Xiao, H.B., 2016b. Variability of beryllium-7 and its potential for documenting soil and soil organic carbon redistribution under event-based short-term water erosion. Soil Sci. Soc. Am. J. 80, 693–703. Li, Z.W., Liu, C., Dong, Y.T., Chang, X.F., Nie, X.D., Liu, L., Xiao, H.B., Lu, Y.M., Zeng, G.M., 2017a. Response of soil organic carbon and nitrogen stocks to soil erosion and land use types in the Loess hilly–gully region of China. Soil Till. Res. 166, 1–9. Li, Z.W., Nie, X.D., He, J.J., Chang, X.F., Liu, C., Lin Liu Sun, L.Y., 2017b. Zonal characteristics of sediment-bound organic carbon loss during water erosion: a case study of four typical loess soils in Shaanxi Province. Catena 156, 393–400. Liu, W.G., Xing, M., 2012. Isotopic indicators of carbon and nitrogen cycles in river catchments during soil erosion in the arid Loess Plateau of China. Chem. Geol. 296 (-297), 66–72. Liu, C., Dong, Y.T., Li, Z.W., Chang, X.F., Nie, X.D., Liu, L., Xiao, H.B., Bashir, H., 2017a. Tracing the source of sedimentary organic carbon in the Loess Plateau of China: an integrated elemental ratio, stable carbon signatures, and radioactive isotopes approach. J. Environ. Radioact. 167, 201–210. Liu, C., Li, Z.W., Dong, Y.T., Chang, X.F., Nie, X.D., Liu, L., Xiao, H.B., Zeng, G.M., 2017b. Do check dam construction and land use change affect a real estimate of soil carbon and nitrogen stocks on the Loess Plateau of China? Ecol. Eng. 101, 220–226. Liu, C., Li, Z.W., Dong, Y.T., Chang, X.F., Nie, X.D., Liu, L., Xiao, H.B., Wang, D.Y., Peng, H., 2017c. Response of sedimentary organic matter source to rainfall events using stable carbon and nitrogen isotopes in a typical loess hilly-gully catchment of China. J. Hydrol. 552, 376–386. Lu, L., Cheng, H.G., Pu, X., Wang, J.T., Cheng, Q.D., Liu, X.L., 2016. Identifying organic matter sources using isotopic ratios in a watershed impacted by intensive agricultural activities in Northeast China. Agric. Ecosyst. Environ. 222, 48–59. McCorkle, E.P., Berhe, A.A., Hunsaker, C.T., et al., 2016. Tracing the source of soil organic matter eroded from temperate forest catchments using carbon and nitrogen isotopes. Chem. Geol. 445, 172–184.

22