Next Article in Journal
Spatial Correlation between Urban Planning Patterns and Vulnerability to Flooding Risk: A Case Study in Murcia (Spain)
Previous Article in Journal
Examining the Spatial Effect of “Smartness” on the Relationship between Agriculture and Regional Development: The Case of Greece
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Land
Volume 12
Issue 3
10.3390/land12030542
land-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Using Source Fingerprinting Techniques to Investigate Sediment Sources during Snowmelt and Rainfall Erosion Events in a Small Catchment in the Black Soil Region of Northeast China

by
Pengfei Du
1,*,
Donghao Huang
2,
Bing Liu
1 and
Wei Qin
1
1
State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin, China Institute of Water Resources and Hydropower Research, Beijing 100048, China
2
College of Water Conservancy, Shenyang Agricultural University, Shenyang 110866, China
*
Author to whom correspondence should be addressed.
Land 2023, 12(3), 542; https://doi.org/10.3390/land12030542
Submission received: 23 December 2022 / Revised: 17 February 2023 / Accepted: 20 February 2023 / Published: 23 February 2023
Browse Figures
Versions Notes

Abstract

:
Concern for the offsite impact of eroded sediment and the need to develop effective catchment sediment management strategies has directed attention to the need for an improved understanding of the primary sediment sources within catchments and the potential of sediment source fingerprinting techniques to provide such information. The study reported here was undertaken in the black soil region of Northeast China, where soil erosion is seen as a serious threat to the sustainable use of soil resources and offsite impacts of eroded soil are also concerned. The study applies source fingerprinting techniques to the evaluation of sediment sources in a small (3.46 km2) agricultural catchment. Sediment sources from five snowmelt and five rainfall events of varying magnitude were contrasted. Three key potential sediment sources were identified within the study catchment: gullies, cultivated topsoil and uncultivated topsoil. Geochemical properties of the source materials (Ti, Ga, Br and Ba) were used as composite fingerprints capable of discriminating between the three potential sources. A mixing model, optimized using a genetic algorithm and coupled with a Monte Carlo procedure to quantify the uncertainty associated with the resulting estimates, was used for source apportionment. The results indicated mean source contributions for the set of 10 events for cultivated topsoil, uncultivated topsoil and the gullies of ~30%, ~10% and ~60%, respectively. In general, snowmelt and rainfall events were characterized by increased contributions from gully sources and cultivated topsoil, respectively. The study is seen as demonstrating the potential for using source tracing techniques to investigate sediment sources in environments where strong seasonal contrasts between snowmelt and rainfall events exist.

1. Introduction

The black soil region of Northeast China has a history of cultivation extending back less than 100 years [1], and has already experienced severe soil loss [2,3]. This makes it one of the regions with the greatest potential soil erosion risk in China [4,5]. In recent years, research in the region has focused on developing an improved understanding of individual erosion processes, including splash erosion [6], particularly surface erosion [7], gully erosion [8] and ephemeral gully erosion [9]. A few studies have linked these erosion processes to catchment sediment yields and catchment sediment budgets [10,11]. As the role of fine sediment in degrading freshwater ecosystems and habitats becomes increasingly recognized, there is a need to also direct attention to the off-site impact of the eroded sediment and to measures for reducing this impact. Information on sediment sources can be very important for designing sediment control strategies and optimizing the return on investment in such strategies. Within the black soil region, to arrange reasonable control measures, it is important to know the main sediment source (i.e., surface and gully erosion) and the contrasts between their contributions during rainfall and snow-melt events.
After more than 40 years of development, sediment source fingerprinting techniques have become capable of identifying the source of sediment delivered to the outlet of a catchment [12,13,14,15]. Its background, development, application in various research have been well elaborated [16]. Compared to the previous methods, it overcomes the various shortcomings and establishes the relative contribution of different sediment sources [17], although its application is based on some assumptions. Thus far, it has been successfully applied in many different areas of the world and different environments [18,19,20,21,22,23], and has helped solve problems, such as water storage reservoirs [24], geomorphic coupling [25], river and sea connections [26], etc., hence playing an important role both in scientific and practical significance. Tang et al. [27] provided a useful review of the application of sediment source fingerprinting techniques in China, where the use of the approach is still in its early stages. Most Chinese studies have focused on the Loess Plateau [28,29,30] and the upper reaches of the Yangtze River [31,32]. Several recent papers have examined the black soil region of northeast China [11,33,34,35,36]; this region is characterized by snow cover for approximately 5–8 months of the year. Most areas experience a continuous snow cover for at least two months [37,38]. Based on runoff sediment information from 27 typical watershed hydrological stations, snowfall accounts for ~7 to 25% of annual precipitation, and snowmelt runoff accounts for ~6 to 28% of the annual sediment yield [37]. However, the sediment sources during snowmelt events remain unclear, not to mention how they differ from the sources during rainfall events. This study, following previous studies developed in the black soil region that tried to investigate the sources contributions to the annual sediment yields in approximately 40 years chronology established by excess lead-210 [33], that aimed to determine the response of soil erosion to land management changes through estimating source contributions in the sediment deposited in the reservoir [34], that designed to assess fingerprinting models performance, including multivariate mixing models, Bayesian model and DFA model in sediment source contribution studies [39], that proposed to explore scale effects in source contributions deposited in the bottom of gullies [35], that purposed to know the general sources contribution in deposited sediment caused by snowmelt during a period [36]. This study aims to further explore the potential for using source fingerprinting techniques to establish the relative importance of the sediment sources contributing to the sediment yield during individual snowmelt and rainfall erosion events.

2. The Study Catchment

The study was undertaken in a 3.46-km2 sub-catchment of the Hebei catchment (27.6 km2) located in the rolling-hill region marking the transition between the Daxing’anling mountains and the Songnen plain, Nenjiang County, Heilongjiang Province, China (Figure 1). The study catchment lies at an elevation of 321–376 m. It is characterized by low topographic relief and long, gentle slopes with angles of 1–6°. Lacustrine and fluvial sand beds are the parent materials.
A monitoring station with a water level recorder and a compound sharp-crested weir was constructed at the outlet of this catchment in 2004 to monitor the runoff and sediment output (see Figure 1). Depth- and width-integrated suspended sediment samples have been collected manually from the weir crest. A tipping bucket rain gauge was installed near the monitoring station in 2006 to provide detailed precipitation records.
The region is characterized by a cold, temperate, continental, semihumid climate, with a long, cold winter and a short growing season. The average temperatures in January and July are −21 °C and 21 °C, respectively. Data from the local meteorological station indicate that the mean annual total sum precipitation for the period 1976–2015 was 534 mm. The snow thickness between November and February varies greatly from year to year, and averages 10–20 cm. This snow cover melts rapidly from mid-March to early April, resulting in snowmelt runoff and associated erosion. The frost-free period extends to 115–120 days, and the soil can be frozen to a depth of up to 2 m during the winter season [34].
The black soil region in northeast China is a key area for agricultural production in China, and the dominant land use is cultivated land, which accounts for approximately 92.8% of the study catchment. The primary crops are soybean (Glycine max (L.) Merr.) and corn (Zea mays L.). Other land-use types within the study catchment include forests (4.3%), permanent grassland (1.7%) and residential areas (1.2%). The study catchment is underlain by Quaternary lacustrine clay deposits. Based on the USDA Soil Taxonomy, the soil is classified as a Udic Argiboroll [40], which is characterized by a thickness of 30–50 cm, a bulk density of approximately 1.1 g cm−3 and a humus content of 4–6%.
Three main types of soil erosion occur in the study area: wind erosion, snowmelt erosion and erosion driven by rainfall. Wind erosion occurs during the dry winter and spring seasons, but rates of soil loss associated with wind erosion are low. Annual soil loss due to wind erosion of 0.58 t ha−1 year−1 has been reported by Jiang [41]. The sediment output from the study catchment is dominated by erosion associated with rainfall and snowmelt. Significant rainfall does not occur during the spring-thaw period [37]. Erosional events associated with snowmelt and rainfall are therefore temporally distinct. Linear gullies are a key feature of the catchment, as shown in Figure 1, and these gullies function during erosion events associated with both snowmelt and rainfall. These features are commonly in the order of 300–400 m in length, 1–2 m wide and approximately 1 m deep and they are actively developing. They are frequently extended by ephemeral gullies developed during major erosional events. Sheet and rill erosions are widespread across the catchment during erosional events. On the uniform slopes, the most serious rill erosion occurs on the lower slopes, although it can affect 85% of the slope length. Where midslope gradients are variable, rill erosion is concentrated in the low-lying areas, typically occurring over ~65% of the entire slope. Figure 2A presents an image of a typical gully in the study catchment. Examples of typical sheet and rill erosion during snowmelt and during a major rainfall event are shown in Figure 2B,C, respectively.
Table 1 presents existing information on the magnitude of the annual suspended sediment yield from the study catchment and the relative contributions of snowmelt and rainfall erosion to the annual sediment yields for the period 2011–2015, when an intensive program of sediment monitoring was in operation. These data were made available by the faculty of geographical science of Beijing Normal University, who operate the monitoring station. The mean annual suspended sediment yield for the catchment for the years 2011–2015 is ~2.8 t ha−1 year−1, although the annual suspended sediment yields show considerable interannual variation. Major rates of annual soil loss generated by rainfall and snowmelt erosion events are likely to be considerably greater, due to the appreciable conveyance losses and associated deposition within the catchment resulting from the relatively gentle slopes. Connectivity between the slopes and the main gully network is limited. Dong et al. estimated that the sediment delivery ratio for the Hebei catchment was in the order of 10% [10]. Table 1 indicates that the relative contributions of snowmelt and rainfall erosion to the annual sediment yield during the period 2011–2015 varied markedly from year to year. During the 5-year period, snowmelt erosion and erosion driven by rainfall contributed between zero and ~80% and between ~20% and ~95%, respectively, of the annual sediment yield. Over the 5-year period, snowmelt erosion contributed ~23% and rainfall erosion ~77% of the total sediment yield.

3. Methods

The statistical methods used in this study involve the dual-range bracket test, the nonparametric Kruskal–Wallis test, Monte Carlo procedure test, goodness-of-fit test, linear regression analysis, etc. The details are described as the following.

3.1. Collection of Source Material and Target Sediment Samples

3.1.1. Collection of Source Material Samples

Field observations in the study catchment prior to source material sampling indicated the existence of three main potential sediment sources. These represented, firstly, the surface soil of areas under cultivation, secondly, the surface soil of uncultivated areas (pasture and woodland) and, thirdly, the sides and floors of the gullies. By using the uniform grids function in ArcGIS 10.3, the area of the study catchment was overlaid with a grid, the intersections of which were used as a basis for identifying the locations of the sampling points to be used to characterize the first two potential sources. Since cultivated land represented the main land-use and covered the majority of the study catchment, a relatively small sampling grid size (i.e., a 400 m × 250 m grid) was utilized to define the sampling points. In contrast, since forest land and grassland occupied a much smaller area, an increased sampling grid size (i.e., an approximately 80 m × 25 m grid) was employed to define sampling points. Sampling of the network of gullies involved the collection of representative composite samples of gully-side material every 40 m along selected gully thalwegs. These included some ephemeral gullies. Using this approach, a total of 71 samples of potential source material were collected, representing 25 collected from cultivated land, 19 collected from uncultivated land and 27 collected from the gullies.
The samples of potential source material were collected in June 2016. The surface samples (0–5 cm) were collected using a sampling ring (depth = 5 cm, diameter = 7.5 cm). In order to take account of local scale spatial variability of source material properties, three samples were collected within a 4-m diameter circle at each sampling point and these were combined to provide a single composite sample. The gully samples were collected using a trowel to provide a composite sample representative of the sides and floor of the gully at the sampled cross section.

3.1.2. Collection of Target Suspended Sediment Samples

The sediment output from the study sub-catchment has been monitored since the establishment of a monitoring station at the outlet in 2004. During the period 2011–2015, an intensive program of suspended sediment sampling was undertaken. Attention was focused on this period since many of the samples of suspended sediment had been retained and could be made available for geochemical analysis. A total of 123 significant runoff events occurred in the study catchment during this period (Figure 3). Sampling was undertaken manually across the weir crest using 1-liter plastic bottles multiple times during an event. This provided depth- and width-integrated samples. The sediment was recovered by settling and subsequent evaporation. Most samples were small, therefore, this study necessarily focused on samples collected from rainfall and snowmelt events with high sediment concentrations. These samples were collected at or near the peaks of runoff events. They represent the period of maximum sediment flux and a substantial proportion of the total suspended sediment load associated with individual snowmelt and rainfall events. Suitable sediment samples from 10 individual events were selected. These comprised five rainfall runoff events and 5 snowmelt runoff events (Figure 3; Table 2). We assumed that source material samples collected in 2016 could be compared with target sediment samples collected during the period 2011–2015. There had been no substantive changes in land use or in cropping or tillage practices during the period 2011–2016.

3.2. Characterizing Source Material and Sediment Samples

All source material samples were air-dried, and fragments of organic material and larger particles were manually removed prior to manual disaggregation and sieving to <0.063 mm. The dried suspended sediment samples were similarly disaggregated and sieved to <0.063 mm. Aliquots of both the source material and the suspended sediment samples were subsequently analyzed for a range of geochemical properties by X-ray fluorescence spectrometry (XRF). The 29 geochemical properties, to be potentially used in establishing source fingerprints, represented the suite of properties that could be measured by the available XRF equipment. This comprised a wide range of source material properties, including nutrients (i.e., total-P), metals (i.e., Mn, Co, Cr, Cu, Ni, Pb, Zn, Ti, V, Ga, Rb, Sr, Zr, Nb and Ba), metalloid elements (i.e., As), a halogen element (i.e., Br), rare earth elements (i.e., La, Ce, Y and Nd), and metallic oxides (i.e., SiO2, Al2O3, Fe2O3, MgO, CaO, Na2O, K2O). It is important to note that, with the possible exception of total-P, these geochemical properties were not expected to be susceptible to seasonal variation in response to the effects of freeze-thaw conditions and were, in that context, judged to be conservative. When comparing the geochemical properties of source material and target samples in fingerprinting investigations, it is important to ensure that differences in particle size composition between source and target samples do not compromise comparisons. All source material and sediment samples were, as indicated above, sieved to <0.063 mm. However, as emphasized by Smith and Blake [42], contrasts between source material and target sediment samples in the particle size distribution of the sediment within the <0.063 mm fraction samples could still compromise the direct comparison of their fingerprint properties. The particle size distribution (PSD) of the <0.063 mm fraction of the source material was measured by laser granulometry after pretreatment with hydrogen peroxide to remove the organic fraction and chemical and ultrasonic dispersion. It was not possible to obtain sufficient amounts of samples to undertake equivalent measurements on the target suspended sediment samples.
It was, however, possible to access surrogate information relating to the PSD of the suspended sediment samples. This surrogate information was provided by the sediment deposited in a small reservoir at the outlet of the study catchment that operated between 1974 and 2004. The dam was removed in 2004 in order to construct the gaging station installed at the outlet of the catchment. The trap efficiency of the reservoir was high, since there was no spillway and the dam was never overtopped. Huang et al. reconstructed the record of annual sediment yield from the study catchment for the period 1974–2004 [33]. A sediment core collected from the deposits indicated a constant PSD over its depth, representing the period 1974–2004 [33] (Figure 4). Focusing on the silt and clay content of the sediment (i.e., the <0.063 mm fraction), the mean percentage of clay (i.e., <2 µm) in the sediment associated with the 42 depth-incremental (5 cm) core sections was 10.0%. The comparison of this value with the mean clay contents of the source material samples representing cultivated topsoil, uncultivated topsoil and gullies, which were 10.7%, 12.0% and 9.7%, respectively, provided confirmation that the sediment source and target samples were characterized by similar PSDs. This similarity suggests that changes in the relative contributions of different sources over the period concerned would not be associated with substantive changes in the PSD of the sediment load of the stream.
To identify the optimum composite fingerprint for discriminating the three potential sources, the dual-range bracket test was used as an initial step to eliminate nonconservative properties, i.e., fingerprint properties where target sample concentrations exceeded the maximum values or were lower than the minimum values associated with the potential sources. Subsequently, the nonparametric Kruskal–Wallis (K–W) test was used to assess the contrasts between the property concentration values associated with the individual sources and thereby identify potential fingerprint properties. Those properties showing statistically significant differences between the sources were then used in a stepwise discriminant function analysis to select the optimal combination of source material properties for discriminating the three potential sources.

3.3. The Source Apportionment Model

A mixing model has been developed to determine the contribution of various sources in sediment samples by minimizing errors [43] and has been successfully used in different places across the world [19,33,44,45]. The optimized mixing model proposed by Walling et al. [23] and Collins et al. [46] was used in the study for source apportionment and therefore for estimating the relative contribution of the three potential sources to the target (suspended sediment) samples. This model has been widely used by other researchers, and Haddadchi et al. [47] indicate that it provides reliable estimates of source contributions. Furthermore, a study involving artificial mixtures of material collected from the study catchment and representing the same potential sources, reported by Du et al. [39], again confirmed the reliability of this model. It generated better results, in terms of reproducing the experimentally known source contributions, than a Bayesian model and a model based on discriminant function analysis.
The model minimizes the sum of the squared relative errors (E) for a multivariate linear mixing model, viz:
E = ∑ i = 1 m C i − ∑ s = i n P s   S s i ] / C i } 2  
It operates with the constraints that all sediment sources are represented, that contributions from individual sources cannot be negative and that the relative contributions from all sources must sum to 1, i.e.,
0 ≤ P s ≤ 1
∑ s = 1 n P s = 1  
where n is the number of sources, Ci is the concentration of the ith fingerprint property, m is the number of fingerprint properties, Ps is the sediment contribution from the sth sediment source and Ssi is the mean concentration of the ith fingerprint property in the sth sediment source. No corrections for differences in grain size composition between the source material and the target samples were applied. In order to take into account the uncertainty associated with the fingerprint properties of individual sources, due to the spatial variability of those sources, the model was coupled with a Monte Carlo routine based on the standard deviation of the fingerprint properties associated with individual sources. This involved determining source contributions through 2500 iterations of a Latin hypercube sampling procedure. A genetic algorithm was used to determine the global optimal solution for each iteration. A goodness-of-fit test was used to assess the uncertainty of the estimates of source contributions. This was based on the mean absolute fit (MAF) [48], which was calculated as follows:
M A F = 1 − ∑ i = 1 m C i − ∑ s = i n P s S s i / C i m
Generally, the study flowchart is shown in Figure 5.

4. Results and Discussion

4.1. Selecting the Optimal Composite Fingerprint for Source Apportionment

The mean and coefficient of variation (CV) for all fingerprint properties of sources and sediment samples are shown in Table 3. The mean CV values ranged from 0.12 to 0.14 for source fingerprint properties, while the mean sediment fingerprint property CV was 0.10. Eight fingerprint properties successfully passed the dual-range bracket test, namely, P, Ti, V, Co, Ga, Br, Ba and Pb. Only Co did not pass the (K–W) test due to the lack of a significant difference between concentrations associated with the three sediment sources. The other seven geochemical properties showed statistically significant differences between the sources (p < 0.05) (Table 4). Stepwise discriminant function analysis indicated that the optimal combination of source material properties for discriminating the three potential sources comprised Ti, Ga, Br and Ba. Figure 4 further confirms that this set of fingerprint properties provides highly effective discrimination between the samples representing the three sources. It also demonstrates that, despite variations in the size and location of the sampled gullies within the study catchment and inclusion of some ephemeral gullies, within the set of samples, the fingerprint properties for the gullies show limited variation and effectively represent a single population. The scatter is less than for the other two sources. The cumulative sediment source discrimination success rates for this combination of fingerprint properties indicated by the stepwise discriminant function analysis were 63.5%, 90.1%, 97.5% and 100%.

4.2. The Relative Source Contributions Associated with Snowmelt Erosion and Rainfall Erosion Events

As a measure of the goodness-of-fit of the multiple solutions of the mixing model for each event provided by the Monte Carlo procedure, mean values of MAF are provided in Table 5. These values are consistently >0.96 and confirm that the mixing model solutions are characterized by very close agreement between the measured values of the sediment properties comprising the sediment fingerprints of the target samples associated with the 10 individual events and the values generated by the optimized mixing model. The source apportionment results summarized in Table 5 present the mean, the standard error of the mean (SE) and the 95% confidence limits of the range of estimates of source contributions provided by the Monte Carlo procedure. The confidence limits are indicated by the lower limit (LL) and upper limit (UL). Taken together, these values confirm the limited uncertainty associated with the estimated source contributions and therefore the precision of those results. The uncertainty increases for the estimates of the relative source contributions associated with uncultivated topsoil because of the low values involved. The uncertainty at the 95% level of confidence for the estimated contributions from this source is typically ~±10.5%. The uncertainty decreases for the estimated source contributions from cultivated topsoil and gully sources, which are of a greater magnitude, and the equivalent values for these two sources are typically ~±5.0% and ±4.3%, respectively. The values for source contributions reported subsequently in the text represent the mean values provided by the Monte Carlo procedure, but the above confidence limits should be recognized.
The results of the source apportionment for the five snowmelt erosion events and the five rainfall erosion events are presented in Table 5 and Figure 6. It is important to recognize that these results are based on a single sample collected near each event maximum. The use of additional samples from each event to reflect possible variations in source contributions during the events could result in different estimates of the relative contributions of the three sources, as during the process of flow events, sediment sources may vary [49]. However, since the single samples used were representative of flows and suspended sediment concentrations close to their peaks, and therefore peak sediment fluxes, they are likely to provide a meaningful estimate of source contributions to the sediment load associated with a given event. As expected, the contributions from cultivated topsoil are substantially greater than those from uncultivated topsoil because the latter occupies a relatively small proportion of the total catchment surface and erosion rates associated with uncultivated soils are likely to be lower than those for cultivated soils. However, since uncultivated areas represent only approximately 6% of the catchment surface, the magnitude of their contribution (mean = 9.6% across all events) could be seen as substantially greater than might be expected. This apparent anomaly is highlighted further when considering solely the contribution of topsoil sources, since, across all events, uncultivated topsoil contributed ~24% of the sediment mobilized from topsoil sources. The uncultivated areas are in the upper part of the valley bottom and above the gully head; particularly, the pasture areas are close to the channel (gully) network, and therefore, their runoff and sediment loads have greater connectivity. This means that a substantial proportion of the sediment mobilized from these areas reaches the channel network and thence the catchment outlet, whereas much of the sediment mobilized from the upper slopes of the catchment, occupied by the cultivated areas, is deposited before reaching the channel network. As indicated above, the sediment delivery ratio for the catchment is estimated to be relatively low at approximately 10%.
Table 2 indicates that the ten events that provide the basis of the study span a substantial range of event magnitudes, as represented by peak discharge, suspended sediment concentration at the time the sample was collected and event sediment yield. Despite this range, there is a reasonable degree of consistency among the results for the individual events. The relative contribution from the uncultivated areas is consistently the smallest; gully sources are the dominant source for seven out of ten of the events and across the ten events there is no major deviation from the mean contributions, namely ~60% from gullies, ~30% from cultivated topsoil and ~10% from uncultivated topsoil. This result is similar to the work by Lizaga et al., who found the channel bank and agricultural lands are the main sediment sources in a Mediterranean agroforestry catchment, accounting for more than 70% of the total source [50]. In another study conducted by Haddadchi et al., the dominant source from the channel with an average proportion of 71.5% was revealed in a mountainous agricultural catchment of western Iran, where cultivated land is major land use (61.5%) [51]. Notwithstanding this general consistency of source contributions, Table 5 indicates that for the snowmelt events, the mean contribution from gully sources (~65%) is appreciably greater than for the rainfall events (~54%), whereas for rainfall events, the mean contribution from cultivated topsoil (~35%) is appreciably greater than for snowmelt events (~25%). The increased importance of gully sources during snowmelt events shown by these results concurs with the findings of Lamba et al. [52]. These researchers reported that the relative contribution of channel sources to the sediment output from a 50 km2 agricultural catchment in south-central Wisconsin, USA, increased during a period when snowmelt events occurred. Additionally, the findings that gully erosion in snowmelt is greater than that in rainfall and that rill erosion in the catchment in snowmelt is less than that in rainfall were reported by Rodzik et al. [53] in the Kolonia Celejow catchment in the period 2003–2006. Gellis and Noe [54] similarly reported that the relative contribution of channel sources to event sediment yields from a 147 km2 agricultural catchment in Maryland, USA, increased in the winter when freeze-thaw processes were active. Some caution is, however, necessary when viewing the above generalizations as providing a definitive assessment of sediment source contributions in the study catchment since they are based on only 10 events, although, as indicated above, those events span a range of event magnitudes. It is also important to recognize that a simple mean of the source proportions associated with the individual events may not provide a good indication of the source of the total mass of sediment transported to the catchment outlet by the 10 events or by the five snowmelt and five rainfall events. This is because the source proportions associated with the events with the largest sediment loads (t) will exert a dominant influence on the source of the total load. Table 5 therefore also presents the load-weighted mean contributions associated with the total mass of sediment transported by the 5 snowmelt events, the 5 rainfall events and the 10 events, in combination. These were calculated using the information on the sediment loads associated with the sampled events provided in Table 2. The lack of major differences between the simple mean values and the load-weighted means further emphasizes the general consistency of the source contributions across the range of events documented. However, it is notable that for the total mass of sediment transported by the 10 events, the estimates of the gully contribution provided by the load-weighted mean (~52%) are appreciably less than those indicated by the simple mean (~60%). In contrast, the estimate of the contribution from cultivated topsoil provided by the load-weighted mean (~38%) increases appreciably relative to that indicated by the simple mean (~30%). These trends suggest that large events in the study catchment with high sediment loads are characterized by reduced gully contributions and increased contributions from cultivated topsoil. This result differs from the findings of Rodzik et al. [53] and Kociuba et al. [55]; their studies demonstrate that large events result in increased gully contribution. However, there is also a need to recognize another limitation of the results presented in Table 5 for the 10 events. These cannot provide a definitive estimate of the relative contribution of the three sources to the annual sediment export from the study catchment since these values do not take into account the relative contributions of snowmelt and rainfall erosion events to the annual sediment load and their interannual variations highlighted in Table 1.
Figure 6 indicates that although, as suggested above, source contributions show a general consistency across the range of events included in the sample of 10 events, there is, nevertheless, evidence of some variation in the source contributions associated with individual events, for both the snowmelt erosion and rainfall erosion events. For example, in the case of the five snowmelt erosion events, the contribution of sediment mobilized from the gullies ranged from a minimum of ~44% to a maximum of ~73%. Equally, for the rainfall erosion events, the contribution of sediment eroded from the gullies ranged between ~42% and ~67%. Since, as indicated above, the contribution of uncultivated topsoil shows little variation between individual events, this variation in gully contributions is essentially mirrored by variations of a similar magnitude in the contributions of cultivated topsoil. The relatively small number of events associated with each dataset precludes detailed analysis of the key controls on this variation since they are likely to include parameters representing both the characteristics of the events, including event magnitude and also antecedent conditions. However, Figure 7 attempts to explore the relationships between the relative contributions of sediment mobilized from the gullies and from cultivated and uncultivated topsoil, both discharges at the time of sampling and the total suspended sediment load associated with an event. The latter two independent variables are seen as providing measures of event magnitude. Snowmelt and rainfall events are considered separately due to differences in the erosion processes involved.
Figure 7A indicates that the relationships between source contributions and water discharge at the time of sampling are not well defined for snowmelt events. However, the plot suggests that the gully contribution increases as discharge increases (r = 0.45) and that the contribution from cultivated topsoil decreases (r = −0.41) as discharge increases. As noted previously, the contribution from uncultivated topsoil remains essentially constant across the range of discharges involved. The changes in the gully and cultivated topsoil contributions as discharge increases suggest that, as water discharge through the gully network increases, sediment mobilization by gully erosion accelerates and its relative contribution increases, resulting in a parallel decrease in the contribution from cultivated topsoil. The equivalent relationship between rainfall events presented in Figure 7B demonstrates much clearer positive (r = 0.96) and negative (r = −0.95) relationships between the gully and cultivated topsoil contributions, respectively, and water discharge at the time of sampling. As with the snowmelt events, the contribution from uncultivated topsoil shows little variation across the range of water discharge values involved. The close positive relationship between the gully contribution and water discharge is seen as reflecting the close link between sediment mobilization from the gully system and water discharge for rainfall events and the connectivity between sediment sources within the gully network and the catchment outlet. The parallel negative relationship between the contribution of cultivated topsoil and water discharge is seen as being likely to reflect the greater increase in the gully erosion contribution, relative to that of the cultivated topsoil, as discharge increases, rather than a reduction in the absolute magnitude of the latter. The less well-defined relationships between the magnitude of the relative source contributions and discharge shown by snowmelt events could reflect the complicating influence of freeze-thaw processes on sediment mobilization.
The relationships between the relative source contributions and the total sediment load of the event for snowmelt and rainfall events are presented in Figure 7C,D, respectively. Again, there is little variation in the contribution of uncultivated topsoil across the range of event sediment yields involved. However, there is some evidence of meaningful relationships between the relative contributions of gully and cultivated topsoil sources and event sediment yield for both sets of events. However, the r values for these relationships are greater for snowmelt events than for rainfall events. For the snowmelt events shown in Figure 7C, the gully contribution shows a positive relationship with event sediment yield (r = 0.70), whereas the contribution of cultivated topsoil shows a negative relationship (r = −0.74). These trends are seen as reflecting the increasing dominance of gully sources during snowmelt events with high sediment loads. This could again be related to the increased connectivity between the gully system and the catchment outlet, compared to that between the slopes and the catchment outlet. The relationships between source contributions and event sediment yield shown by the rainfall events (Figure 7D) are less well defined than for snowmelt events and the trends of the relationships are reversed. The relationship for gully sources is negative (r = −0.50), whereas that for cultivated topsoil sources is positive (r = 0.47). Here, the gully and cultivated topsoil contributions decrease and increase, respectively, as the event sediment yield increases. These contrasting trends for snowmelt and rainfall events may reflect the influence of splash detachment associated with rainfall events but absent from snowmelt events.

4.3. Comparison of Sediment Sources for Snowmelt and Rainfall Erosion Events

As indicated above, the two dominant sediment sources for both snowmelt and rainfall erosion events were the gullies and cultivated topsoil. Figure 6 indicates that, for four of the five snowmelt events represented, gully erosion contributes a substantially greater proportion of the sediment yield than cultivated topsoil and is thus the dominant source. For the remaining events, the gully contribution is similar to that from the cultivated topsoil. Considering all five events, the mean contribution from gully sources is 65.6%, whereas the mean contribution from cultivated topsoil is 25.5%. This situation contrasts with that for the rainfall events, where the contribution from cultivated topsoil is of increased importance (35.4%) and the mean contribution from gullies reduces to 54.4%. For two of the rainfall events represented, the contribution from cultivated topsoil exceeds that from the gullies. This finding is consistent with Tiecher‘s study, which demonstrated the precipitation events are related to the sediment increase from topsoil in two paired agricultural catchments in southern Brazil [22].
The importance of gully erosion as a sediment source in the study catchment, where the gully contribution dominates for seven out of the ten documented events and the mean gully contribution for the ten events was 60.0%, can be primarily attributed to the properties of the black soils and the local topography. However, the important contrasts between snowmelt and summer rainfall events in terms of their source contributions, with the relative contribution of gully sources dominating in snowmelt events and the contribution of cultivated topsoil sources increasing for rainfall events, can be accounted for by three key contrasts between the erosion processes operating during snowmelt and during rainfall events. The first is the time of flow, unlike rainfall runoff, snowmelt runoff occurs only where there is snow accumulation, and the runoff quickly joins the gully, where it flows for a much longer period of time than on the slope. The second is the instability of the gully system during the spring thaw, which increases its importance as a sediment source. The absence of protective vegetation increases the susceptibility of gully systems to erosion. This is further increased by the loosening and collapse or slumping of gully sides, as a result of freeze-thaw activity, and the occurrence of saturated conditions in the upper horizons of the soil as it thaws, which in turn is likely to decrease gully stability and promote gully extension and the development of ephemeral gullies. Due to freeze-thaw activity, substantial amounts of loose sediment can accumulate within the gullies and provide a source of readily mobilizable sediment when flow occurs during discharge events generated by snowmelt. The third is the importance of both rainfall and surface runoff in mobilizing sediment during summer rainfall events. The kinetic energy of falling rain promotes the detachment of soil particles by splash detachment and this sediment is readily transported by surface runoff during storm events, along with further sediment mobilized by the runoff. These erosion processes drive interrill and rill erosion, which mobilize sediment from the catchment surface. As a result, the catchment surface is a more important sediment source during the summer than during the winter when, in the absence of splash detachment, surface erosion is driven only by runoff detachment, which is considered less erosive with lower kinetic energy compared with rainfall [56] and is therefore less effective.

5. Conclusions

The study reported has demonstrated that sediment source tracing techniques can be successfully used to investigate contrasts in sediment sources within a catchment between snowmelt events and rainfall events. The Monte Carlo procedure indicated that the final estimates of source contributions involved limited uncertainty. Estimates of the precision at the 95% level of confidence associated with these estimates indicated a precision of ~±4.3% and 5.0% for the source contributions from gullies and cultivated topsoil, respectively, and ~±10.5% for the uncultivated topsoil.
Generally, for the five snowmelt events, gully erosion contributed more sediment yield (65.6%) than cultivated topsoil (25.5). For the five rainfall events, the gully and cultivated topsoil contributions changed to 54.4% and 35.4%, respectively. Taking the 10 events together, the results indicate mean contributions of sediment from cultivated topsoil, uncultivated topsoil and gullies of ~30%, ~10% and ~60%, respectively. The mean contribution values for three sources indicate that while the contributions from uncultivated topsoil remain similar for snowmelt and rainfall events, gully contributions either dominate or are very similar to the contributions from cultivated topsoil for both sets of events, but the contributions from cultivated topsoil are generally higher for rainfall events than for snowmelt events.
For both snowmelt and rainfall events, with the increase in discharge at the time of sampling, the contributions from gully and cultivated topsoil were increased and decreased, respectively. However, with the increase in event sediment load, the contributions from gullies increased for snowmelt events but decreased for rainfall events, and the trends of the relationships are reversed for contributions from the cultivated topsoil. These trends may relate to the increased connectivity between the gully system and the catchment outlet and may reflect the influence of splash detachment associated with rainfall events but absent from snowmelt events. These relationships also demonstrated contrasts in both strength and trend between the snowmelt and rainfall events, which were tentatively described as differences between the processes operating. Further work, involving samples from additional events, is required to explore these and other relationships further in order to develop a better understanding of the factors controlling variations in source contributions between events and the key contrasts between snowmelt and rainfall events.

Author Contributions

Conceptualization, P.D. and D.H.; methodology, D.H. and B.L.; software, D.H.; Validation: B.L.; formal analysis, P.D.; investigation, P.D. and W.Q.; resources, W.Q.; writing—original draft preparation, P.D. and D.H.; writing—review and editing, P.D. and B.L.; funding acquisition, P.D., D.H. and B.L. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the special fund of the National Natural Science Foundation of China [grant number 42007050, 41501299], the scientific Program of State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin [grant number SKL2022TS10], and the Jiangxi Province technology innovation introductory program [grant number 20212AEI91011]. Grateful thanks are extended to the Jiusan Soil Conservation Experimental Station supported by the State Key Laboratory of Earth Surface Processes and the Resource Ecology at Beijing Normal University, for permitting geochemical analysis of the 10 suspended sediment samples collected during 2011–2015, that were used in the study and for providing background data for the sampled events and sediment yields from the study catchment.

Data Availability Statement

Not applicable.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Wu, Y.; Zheng, Q.; Zhang, Y.; Liu, B.; Cheng, H.; Wang, Y. Development of gullies and sediment production in the black soil region of northeastern China. Geomorphology 2008, 101, 683–691. [Google Scholar] [CrossRef]
  2. Yan, B.; Yang, Y.; Liu, X.; Zhang, S.; Liu, B.; Shen, B.; Wang, Y.; Zheng, G. Present Status of Soil Erosion and Evolution Tendency of Black Soil Region of Northeast. Soil Water Conserv. China 2008, 12, 26–30. (In Chinese) [Google Scholar] [CrossRef]
  3. Zhang, T.; Liu, G.; Duan, X.; Wilson, G.V. Spatial distribution and morphologic characteristics of gullies in the Black Soil Region of Northeast China: Hebei watershed. Phys. Geogr. 2016, 37, 228–250. [Google Scholar] [CrossRef]
  4. Fan, H.; Cai, Q.; Wang, H. Condition of Soil Erosion in Phaeozem Region of Northeast China. J. Soil Water Conserv. 2004, 18, 66–70. (In Chinese) [Google Scholar] [CrossRef]
  5. Wang, H.; Yang, S.; Wang, Y.; Gu, Z.; Xiong, S.; Huang, X.; Sun, M.; Zhang, S.; Guo, L.; Cui, J.; et al. Rates and causes of black soil erosion in Northeast China. Catena 2022, 214, 1–8. [Google Scholar] [CrossRef]
  6. Hu, W.; Zheng, F.; Bian, F. The Directional Components of Splash Erosion at Different Raindrop Kinetic Energy in the Chinese Mollisol Region. Soil Sci. Soc. Am. J. 2016, 80, 1329–1340. [Google Scholar] [CrossRef] [Green Version]
  7. Fan, H.; Cai, Q.; Cui, M. Soil erosion developed with the vertical belts in the gentle hilly black soil regions in Northeast China. Trans. Chin. Soc. Agric. Eng. 2005, 21, 8–11. (In Chinese) [Google Scholar] [CrossRef]
  8. Wang, W.; Deng, R.; Zhang, S. Effect of land use change on gully erosion in Nemoer River basin of typical black soil region of Northeast China. J. Nat. Resour. 2016, 31, 833–844. (In Chinese) [Google Scholar] [CrossRef]
  9. Zhang, Y.; Wu, Y.; Wang, Y.; Liu, B. Seasonal discrepancy of ephemeral gully erosion in Northeast China with black soils. Geogr. Res. 2008, 27, 145–154. Available online: http://www.dlyj.ac.cn/EN/Y2008/V27/I1/145 (accessed on 22 December 2022). (In Chinese).
  10. Dong, Y.; Wu, Y.; Zhang, T.; Yang, W.; Liu, B. The sediment delivery ratio in a small catchment in the black soil region of Northeast China. Int. J. Sediment Res. 2013, 28, 111–117. [Google Scholar] [CrossRef]
  11. Fan, H. Temporal variations of sediment source from a reservoir catchment in the black soil region, Northeast China. Soil Tillage Res. 2015, 153, 59–65. [Google Scholar] [CrossRef]
  12. Collins, A.L.; Pulley, S.; Foster, I.D.L.; Gellis, A.; Porto, P.; Horowitz, A.J. Sediment source fingerprinting as an aid to catchment management: A review of the current state of knowledge and a methodological decision tree for end-users. J. Environ. Manag. 2017, 194, 86–108. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  13. Walling, D.E. Tracing suspended sediment sources in catchments and river systems. Sci. Total Environ. 2005, 344, 159–184. [Google Scholar] [CrossRef] [PubMed]
  14. Walling, D.E. The evolution of sediment source fingerprinting investigations in fluvial systems. J. Soils Sediments 2013, 13, 1658–1675. [Google Scholar] [CrossRef]
  15. Li, Z.; Xu, X.; Zhang, Y.; Wang, K. Fingerprinting sediment sources in a typical karst catchment of southwest China. Int. Soil Water Conserv. 2020, 8, 277–285. [Google Scholar] [CrossRef]
  16. Foster, I.D.L. Tracers in Geomorphology; John Wiley & Sons Ltd.: Chichester, UK, 2000. [Google Scholar]
  17. Collins, A.L.; Walling, D.E. Documenting catchment suspended sediment sources:problems, approaches and prospects. Prog. Phys. Geogr. Earth Environ. 2004, 28, 159–196. [Google Scholar] [CrossRef]
  18. Collins, A.L.; Walling, D.E.; Sichingabula, H.M.; Leeks, G.J.L. Suspended sediment source fingerprinting in a small tropicasl catchment and some management implications. Appl. Geogr. 2001, 21, 387–412. [Google Scholar] [CrossRef]
  19. Devereux, O.H.; Prestegaard, K.L.; Needelman, B.A.; Gellis, A.C. Suspended-sediment sources in an urban watershed, Northeast Branch Anacostia River, Maryland. Hydrol. Process. 2010, 24, 1391–1403. [Google Scholar] [CrossRef]
  20. Schuller, P.; Walling, D.E.; Iroume, A.; Quilodran, C.; Castillo, A.; Navas, A. Using 137Cs and 210Pbex and other sediment source fingerprints to document suspended sediment sources in small forested catchments in south-central Chile. J. Environ. Radioact. 2013, 124, 147–159. [Google Scholar] [CrossRef]
  21. Smith, H.G.; Sheridan, G.J.; Lane, P.N.J.; Noske, P.J.; Heijnis, H. Changes to sediment sources following wildfire in a forested upland catchment, southeastern Australia. Hydrol. Process. 2011, 25, 2878–2889. [Google Scholar] [CrossRef]
  22. Tiecher, T.; Caner, L.; Minella, J.P.G.; Pellegrini, A.; Capoane, V.; Rasche, J.W.A.; Schaefer, G.L.; Rheinheimer, D.d.S. Tracing sediment sources in two paired agricultural catchments with different riparian forest and wetland proportion in southern Brazil. Geoderma 2017, 285, 225–239. [Google Scholar] [CrossRef]
  23. Walling, D.E.; Woodward, J.C.; Nicholas, A.P. A multi-parameter approach to fingerprinting suspended-sediment sources. In Tracers in Hydrology; Peters, N.E., Hoehn, E., Leibundgut, C., Tase, N., Walling, D., Eds.; IAHS Press: Wallingford, UK, 1993; pp. 329–338. [Google Scholar]
  24. Verstraeten, G.; Poesen, J. Estimating trap efficiency of small reservoirs and ponds: Methods and implications for the assessment of sediment yield. Prog. Phys. Geogr. 2000, 24, 219–251. [Google Scholar] [CrossRef]
  25. D’Haen, K.; Dusar, B.; Verstraeten, G.; Degryse, P.; De Brue, H. A sediment fingerprinting approach to understand the geomorphic coupling in an eastern Mediterranean mountainous river catchment. Geomorphology 2013, 197, 64–75. [Google Scholar] [CrossRef]
  26. Gu, J.; Chen, J.; Sun, Q.; Wang, Z.; Wei, Z.; Chen, Z. China’s Yangtze delta: Geochemical fingerprints reflecting river connection to the sea. Geomorphology 2014, 227, 166–173. [Google Scholar] [CrossRef]
  27. Tang, Q.; Fu, B.; Wen, A.; Zhang, X.; He, X.; Collins, A. Fingerprinting the sources of water-mobilized sediment threatening agricultural and water resource sustainability: Progress, challenges and prospects in China. Sci. China Earth Sci. 2019, 62, 2017–2030. [Google Scholar] [CrossRef] [Green Version]
  28. Jia, X.; Wang, H.; Wan, H. Sources and trace element geochemical characteristics of the coarse sediment in the Ningxia-Inner Mongolia reaches of the Yellow River. Geosci. J. 2013, 18, 181–192. [Google Scholar] [CrossRef]
  29. Wen, A.B.; Zhang, X.B.; Walling, D.E. A study on soil erosion rates and sediment sources using caesium-137 technique in a small drainage of the loess hills. Acta Geogr. Sin. 1998, 53, 124–133. (In Chinese) [Google Scholar] [CrossRef]
  30. Yang, M.; Xu, L. Fingerprinting suspended sediment sources in a small catchment on the loess plateau. J. Soil Water Conserv. 2010, 24, 30–34. Available online: https://www.cnki.net/kcms/doi/10.13870/j.cnki.stbcxb.2010.02.018.html (accessed on 22 December 2022). (In Chinese).
  31. Guo, J.; Wen, A.; Yan, D.; Shi, Z. Quantifying catchment scale sediment source using composite fingerprinting technique. Trans. Chin. Soc. Agric. Eng. 2014, 30, 94–104. Available online: https://en.cnki.com.cn/Article_en/CJFDTOTAL-NYGU201402013.htm (accessed on 22 December 2022). (In Chinese).
  32. Zhang, X.; He, X.; Wen, A.; Walling, D.E.; Feng, M.; Zou, X. Combing Cs-137 and Pb-210 to study sediment source on small catchment in Sichuan Hilly Basin. Chin. Sci. Bull. 2004, 49, 1537–1541. (In Chinese) [Google Scholar] [CrossRef]
  33. Huang, D.; Du, P.; Walling, D.E.; Ning, D.; Wei, X.; Liu, B.; Wang, J. Using reservoir deposits to reconstruct the impact of recent changes in land management on sediment yield and sediment sources for a small catchment in the Black Soil region of Northeast China. Geoderma 2019, 343, 139–154. [Google Scholar] [CrossRef]
  34. Huang, D.; Du, P.; Wang, J.; Wei, X.; Liu, B.; Xu, J. Using reservoir deposits to quantify the source contributions to the sediment yield in the Black Soil Region, Northeast China, based on the fingerprinting technique. Geomorphology 2019, 339, 1–18. [Google Scholar] [CrossRef]
  35. Huang, D.; Pei, M.; Zhou, L.; Fan, H.; Jia, Y. Identification of sediment sources and exploration of scale effects in the black soil region of Northeast China. Catena 2020, 195, 104848. [Google Scholar] [CrossRef]
  36. Huang, D.; Su, L.; Zhou, L.; Fan, H. Gully is the dominant sediment source of snowmelt erosion in the black soil region—A case study. Soil Tillage Res. 2022, 215, 105232. [Google Scholar] [CrossRef]
  37. Jiao, J.; Xie, Y.; Lin, Y.; Zhao, D. Study on snowmelt runoff and sediment yields in northeast China. Geogr. Res. 2009, 28, 333–344. (In Chinese) [Google Scholar] [CrossRef]
  38. Zhang, R.; Wang, X.; Fan, H.; Zhou, L.; Wu, M.; Liu, Y. Study on the regionalization of freeze-thaw zones in China and the erosion characteristics. Sci. Soil Water Conserv. 2009, 7, 24–28. (In Chinese) [Google Scholar] [CrossRef]
  39. Du, P.; Huang, D.; Ning, D.; Chen, Y.; Liu, B.; Wang, J. Application of Bayesian models and discriminant function analysis to the estimation of sediment source contributions. Int. J. Sediment Res. 2019, 34, 577–590. [Google Scholar] [CrossRef]
  40. Sun, J.M.; Liu, T.S. Desertification in the Northeastern China. Quat. Sci. 2001, 1, 72–78. (In Chinese) [Google Scholar] [CrossRef]
  41. Jiang, H. Research on the Wind Erosion in the Black Soil Region and Inner Mongolia Semi-Arid Grassland Region. Ph.D. Thesis, Beijing Normal University, Beijing, China, 2013. (In Chinese). [Google Scholar]
  42. Smith, H.G.; Blake, W.H. Sediment fingerprinting in agricultural catchments: A critical re-examination of source discrimination and data corrections. Geomorphology 2014, 204, 177–197. [Google Scholar] [CrossRef]
  43. Haddadchi, A.; Ryder, D.S.; Evrard, O.; Olley, J. Sediment fingerprinting in fluvial systems: Review of tracers, sediment sources and mixing models. Int. J. Sediment Res. 2013, 28, 560–578. [Google Scholar] [CrossRef] [Green Version]
  44. Motha, J.A.; Wallbrink, P.J.; Hairsine, P.B.; Grayson, R.B. Determining the sources of suspended sediment in a forested catchment in southeastern Australia. Water Resour. Res. 2003, 39, 1056. [Google Scholar] [CrossRef]
  45. Hughes, A.O.; Olley, J.M.; Croke, J.C.; McKergow, L.A. Sediment source changes over the last 250 years in a dry-tropical catchment, central Queensland, Australia. Geomorphology 2009, 104, 262–275. [Google Scholar] [CrossRef]
  46. Collins, A.L.; Walling, D.E.; Leeks, G.J.L. Source type ascription for fluvial suspended sediment based on a quantitative composite fingerprinting technique. Catena 1997, 29, 1–27. [Google Scholar] [CrossRef]
  47. Haddadchi, A.; Olley, J.; Laceby, P. Accuracy of mixing models in predicting sediment source contributions. Sci. Total Environ. 2014, 497–498, 139–152. [Google Scholar] [CrossRef] [PubMed]
  48. Zhang, X.C.; Liu, B.L. Using multiple composite fingerprints to quantify fine se-diment source contributions: A new direction. Geoderma 2016, 268, 108–118. [Google Scholar] [CrossRef]
  49. Gonzales-Inca, C.; Valkama, P.; Lill, J.-G.; Slotte, J.; Hietaharju, E.; Uusitalo, R. Spatial modeling of sediment transfer and identification of sediment sources during snowmelt in an agricultural watershed in boreal climate. Sci. Total Environ. 2018, 612, 303–312. [Google Scholar] [CrossRef] [PubMed]
  50. Lizaga, I.; Gaspar, L.; Latorre, B.; Navas, A. Variations in transport of suspended sediment and associated elements induced by rainfall and agricultural cycle in a Mediterranean agroforestry catchment. J. Environ. Manag. 2020, 272, 111020. [Google Scholar] [CrossRef]
  51. Haddadchi, A.; Nosrati, K.; Ahmadi, F. Differences between the source contribution of bed material and suspended sediments in a mountainous agricultural catchment of western Iran. Catena 2014, 116, 105–113. [Google Scholar] [CrossRef]
  52. Lamda, J.; Karthikayan, K.G.; Thompson, A.M. Apportionment of suspended sediment sources in an agricultural watershed using sediment fingerprinting. Geoderma 2017, 239–240, 25–33. [Google Scholar] [CrossRef]
  53. Rodzik, J.; Furtak, T.; Zgłobicki, W. The impact of snowmelt and heavy rainfall runoff on erosion rates in a gully system, Lublin Upland, Poland. Earth Surf. Process. Landf. 2009, 34, 1938–1950. [Google Scholar] [CrossRef]
  54. Gellis, A.C.; Noe, G.B. Sediment source analysis in the Lingamore Creek watershed, Maryland, USA, using the sediment fingerprinting approach: 2008–2010. J. Soils Sediments 2013, 13, 1735–1753. [Google Scholar] [CrossRef]
  55. Kociuba, W.; Janicki, G.; Rodzik, J.; Stępniewski, K. Comparison of volumetric and remote sensing methods for assessing the development of a permanent forested loess gully. Nat. Hazard 2015, 77, 139–158. [Google Scholar] [CrossRef]
  56. Jensen, T.; Tiessen, K.; Salvano, E.; Kalischuk, A.; Flaten, D.N. Spring snowmelt impact on phosphorus addition to surface runoff in the Northern Great Plains. Better Crops 2011, 95, 28–31. [Google Scholar] [CrossRef] [Green Version]
Land 12 00542 g001 550
Figure 1. The location of Nenjiang County in Northeast China (A), the location of Hebei catchment in Nenjiang County (B), the location of study sub-catchment in Hebei catchment and the fluvial network (C), and the topography and landuse and the distribution of sampling points of the study sub-catchment (D).
Figure 1. The location of Nenjiang County in Northeast China (A), the location of Hebei catchment in Nenjiang County (B), the location of study sub-catchment in Hebei catchment and the fluvial network (C), and the topography and landuse and the distribution of sampling points of the study sub-catchment (D).
Land 12 00542 g001
Land 12 00542 g002 550
Figure 2. A typical gully in the study catchment (A), examples of sheet and rill erosions associated with a snowmelt erosion event (B) and a rainfall erosion event (C).
Figure 2. A typical gully in the study catchment (A), examples of sheet and rill erosions associated with a snowmelt erosion event (B) and a rainfall erosion event (C).
Land 12 00542 g002
Land 12 00542 g003 550
Figure 3. Maximum discharges associated with the 123 rainfall and snowmelt events sampled at the outlet of the study catchment during the period 2011–2015. The events sampled are identified, and further details of these events are provided in Table 2.
Figure 3. Maximum discharges associated with the 123 rainfall and snowmelt events sampled at the outlet of the study catchment during the period 2011–2015. The events sampled are identified, and further details of these events are provided in Table 2.
Land 12 00542 g003
Land 12 00542 g004 550
Figure 4. Scatter plot of the discriminant functions representing the samples collected from the three potential sediment sources.
Figure 4. Scatter plot of the discriminant functions representing the samples collected from the three potential sediment sources.
Land 12 00542 g004
Land 12 00542 g005 550
Figure 5. The flowchart shows the whole study process.
Figure 5. The flowchart shows the whole study process.
Land 12 00542 g005
Land 12 00542 g006 550
Figure 6. The relative contributions of the different potential sources to the target samples representing the five snowmelt erosion events (left) and the five rainfall erosion events (right).
Figure 6. The relative contributions of the different potential sources to the target samples representing the five snowmelt erosion events (left) and the five rainfall erosion events (right).
Land 12 00542 g006
Land 12 00542 g007 550
Figure 7. Relationships between source contributions and water discharge at the time of sampling and event sediment yield for snowmelt events (A,C) and rainfall events (B,D). Statistically significant fitted regression lines are shown as solid lines. An asterisk (*) is used to denote r values associated with relationships that are significant at a >95% level of confidence.
Figure 7. Relationships between source contributions and water discharge at the time of sampling and event sediment yield for snowmelt events (A,C) and rainfall events (B,D). Statistically significant fitted regression lines are shown as solid lines. An asterisk (*) is used to denote r values associated with relationships that are significant at a >95% level of confidence.
Land 12 00542 g007
Table
Table 1. Annual suspended sediment yields from the study catchment during the period 2011–2015 and the relative contributions of snowmelt and rainfall erosion to those sediment yields.
Table 1. Annual suspended sediment yields from the study catchment during the period 2011–2015 and the relative contributions of snowmelt and rainfall erosion to those sediment yields.
YearSediment Yield Contributed by
Snowmelt Erosion Events		Sediment Yield Contributed by
Rainfall Erosion Events		Total Sediment Yield
(t ha−1 Year−1)%(t ha−1 Year−1)%(t ha−1 Year−1)
20110.6079.50.1620.50.76
201200.000.0021000.002
20132.1726.75.9773.38.14
20140.234.64.7795.45.00
20150.1860.60.1239.40.30
Mean0.6422. 52.2077.52.84
Table
Table 2. Details relating to the 10 suspended sediment samples selected for geochemical analysis.
Table 2. Details relating to the 10 suspended sediment samples selected for geochemical analysis.
Date of Sample CollectionType of EventPeak Discharge
(m3 s−1)		Sediment
Concentration
(g l−1)		Event Sediment Load
(t)
7 April 2011Snowmelt0.04109.55.97
12 April 2011Snowmelt0.010916.72.51
29 April 2013Snowmelt0.003716.10.69
7 June 2013Rainfall0.258439.149.5
13 June 2013Rainfall0.969939.6112.9
19 June 2013Rainfall0.005342.80.08
18 May 2014Rainfall0.507438.90.70
14 July 2014Rainfall0.939628.00.84
31 March 2015Snowmelt0.011611.610.21
15 April 2015Snowmelt0.010710.716.08
Table
Table 3. Results of the mean and coefficient of variation (CV) of the fingerprint properties for all samples.
Table 3. Results of the mean and coefficient of variation (CV) of the fingerprint properties for all samples.
MaterialSediment Fingerprint PropertiesSample Size
PTiVCrMnCoNiCuZnGaAsBrRbSrYZrNbBaLaCeNdPbSiO2Al2O3Fe2O3MgOCaONa2OK2O
μg/gμg/gμg/gμg/gμg/gμg/gμg/gμg/gμg/gμg/gμg/gμg/gμg/gμg/gμg/gμg/gμg/gμg/gμg/gμg/gμg/gμg/g%%%%%%%
Cultivated topsoilMean998.0 4382.0 126.7 90.5 1051.0 18.2 41.4 14.1 47.9 25.0 16.2 5.3 93.5 131.1 20.8 210.4 12.7 696.9 40.0 94.7 32.5 25.2 60.0 13.7 4.5 1.1 1.0 1.3 2.2 25
CV0.10 0.07 0.14 0.15 0.10 0.12 0.30 0.23 0.11 0.14 0.20 0.26 0.09 0.09 0.13 0.12 0.12 0.04 0.15 0.11 0.15 0.21 0.04 0.03 0.09 0.11 0.09 0.10 0.04
Uncultivated topsoilMean1149.3 3968.2 62.7 40.9 931.7 19.7 16.4 13.1 51.1 9.4 7.3 6.1 88.8 129.3 18.4 184.6 11.4 955.5 36.8 95.1 31.7 10.9 51.2 13.1 4.6 1.2 1.2 1.5 2.2 19
CV0.17 0.05 0.08 0.13 0.08 0.14 0.24 0.22 0.10 0.15 0.09 0.36 0.07 0.11 0.11 0.20 0.11 0.03 0.10 0.10 0.10 0.23 0.07 0.03 0.08 0.10 0.14 0.10 0.08
GullyMean663.5 4748.5 102.5 75.1 768.8 20.9 31.9 19.1 59.3 18.6 10.8 1.8 105.7 141.1 23.9 231.3 14.6 644.5 41.8 87.7 34.1 19.0 61.9 15.8 5.4 1.4 0.9 1.2 2.4 27
CV0.27 0.05 0.16 0.18 0.38 0.27 0.23 0.16 0.14 0.09 0.20 0.50 0.07 0.07 0.10 0.11 0.09 0.03 0.14 0.09 0.10 0.13 0.05 0.04 0.10 0.15 0.08 0.07 0.03
Sediment samplesMean842.5 4379.6 92.5 75.5 565.4 17.6 32.1 24.1 75.3 16.8 12.1 3.0 116.2 132.4 22.7 182.3 15.0 656.9 49.9 89.4 39.5 19.2 60.5 13.4 5.9 1.4 1.1 1.2 2.4 10
CV0.08 0.06 0.10 0.15 0.06 0.10 0.17 0.18 0.11 0.09 0.08 0.21 0.09 0.12 0.08 0.20 0.04 0.04 0.10 0.10 0.10 0.07 0.03 0.06 0.14 0.15 0.03 0.21 0.02
Table
Table 4. Results of the nonparametric Kruskal–Wallis test.
Table 4. Results of the nonparametric Kruskal–Wallis test.
FingerprintH-Valuep-Value
P37.6670.000
Ti31.7940.000
V43.3050.000
Co5.7450.057 *
Ga42.0910.000
Br42.5040.000
Ba20.0780.000
Pb15.0420.001
* Differences were not statistically significant at the p < 0.05 level.
Table
Table 5. Estimates of relative source contributions to the target samples representing the snowmelt erosion, rainfall erosion events and all events.
Table 5. Estimates of relative source contributions to the target samples representing the snowmelt erosion, rainfall erosion events and all events.
TypeDateMAFSediment Contribution (%)
Cultivated TopsoilUncultivated TopsoilGully
MeanSE95% LL95% ULMeanSE95% LL95% ULMeanSE95% LL95% UL
Snowmelt erosion events7 April 20110.9722.60.8720.824.37.60.506.78.669.81.0567.771.9
12 April 20110.9724.10.8822.425.87.70.506.78.768.21.0666.170.3
29 April 20130.9745.60.9843.747.510.90.599.712.043.51.1641.245.8
31 March 20150.9617.70.8516.119.49.20.538.210.273.11.0571.075.1
15 April 20150.9617.70.8316.119.49.10.548.010.173.21.0371.275.2
Mean0.9725.50.8823.827.38.90.537.99.965.61.0763.567.7
Load-weighted mean/19.50.8517.921.28.80.537.89.871.71.0469.673.7
Rainfall erosion events7 June 20130.9731.60.9529.733.510.60.569.511.757.81.1655.560.0
13 June 20130.9746.00.9944.047.911.40.6010.212.642.61.1340.444.8
19 June 20130.9747.70.9645.849.510.50.579.411.641.81.1339.644.0
18 May 20140.9727.60.9025.829.49.70.538.710.862.71.1060.564.8
14 July 20140.9724.10.8622.425.88.80.537.89.867.11.0765.069.2
Mean0.9735.40.9333.637.210.20.569.111.354.41.1252.256.6
Load-weighted mean/41.40.9839.543.311.20.5910.012.347.41.1445.249.6
All eventsMean0.9730.50.9128.732.29.60.548.510.660.01.0957.862.1
Load-weighted mean/37.5 0.95 35.7 39.4 10.7 0.58 9.6 11.9 51.7 1.12 49.5 53.9
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Du, P.; Huang, D.; Liu, B.; Qin, W. Using Source Fingerprinting Techniques to Investigate Sediment Sources during Snowmelt and Rainfall Erosion Events in a Small Catchment in the Black Soil Region of Northeast China. Land 2023, 12, 542. https://doi.org/10.3390/land12030542

AMA Style

Du P, Huang D, Liu B, Qin W. Using Source Fingerprinting Techniques to Investigate Sediment Sources during Snowmelt and Rainfall Erosion Events in a Small Catchment in the Black Soil Region of Northeast China. Land. 2023; 12(3):542. https://doi.org/10.3390/land12030542

Chicago/Turabian Style

Du, Pengfei, Donghao Huang, Bing Liu, and Wei Qin. 2023. "Using Source Fingerprinting Techniques to Investigate Sediment Sources during Snowmelt and Rainfall Erosion Events in a Small Catchment in the Black Soil Region of Northeast China" Land 12, no. 3: 542. https://doi.org/10.3390/land12030542

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop