Geoderma 436 (2023) 116539

Contents lists available at ScienceDirect

Geoderma
journal homepage: www.elsevier.com/locate/geoderma

Multi-temporal spatial modelling to assess runoff and sediment dynamics

under different microtopographic patterns
Jian Luo a, b, Zicheng Zheng b, *, Shuqin He c, d, Wenfeng Ding d
a
Inner Mongolia Key Laboratory of River and Lake Ecology, School of Ecology and Environment, Inner Mongolia University, Hohhot 010021, China
b
College of Resources Science, Sichuan Agricultural University, Chengdu 611130, China
c
College of Forestry, Sichuan Agricultural University, Chengdu 611130, China
d
Key Laboratory of River and Lake Control and Flood Control of the Middle and Lower Reaches of the Yangtze River, Ministry of Water Resources, Wuhan 430010,
China

A R T I C L E I N F O A B S T R A C T

Handling Editor: Morgan Cristine L.S. Soil erosion is a multi-scale geographical interface process with high spatiotemporal variability. This study aimed
to assess runoff and sediment dynamics under different microtopographic patterns based on wavelet analysis and
Keywords: SIMulated Water Erosion model (SIMWE). Three rainfall intensities of 1.0 mm min− 1, 1.5 mm min− 1, and 2.0
Microtopography mm min− 1, combined with three slope gradients of 10◦ , 15◦ , and 20◦ , were utilized to conduct rainfall simu­
Rill erosion
lations on three soil boxes measuring 4 m in length and 0.8 m in width. The soil boxes exhibited distinct
Wavelet analysis
microtopographic patterns, including smooth slope (CK), artificial digging (AD), and ridge tillage (RT). The
SIMWE model
Simulated rainfall results showed that the runoff and sediment reduction benefits of different tilled slopes decreased with an in­
crease in both slope and rainfall intensity. The AD and RT slopes reduced runoff yields by 8.4% − 50.2% and
17.3% − 75.0% respectively compared to CK slope. Although sediment yields from AD and RT slopes were
reduced by 4.2% − 51.6% and 1.8% − 53.0%, respectively, the sediment control benefits of AD and RT slopes
were basically lost when the slope gradient increased to 20◦ . The rill erosion was the main source of soil erosion
on different tilled slopes, accounting for 72% − 87% of the total soil loss. And the rates of both soil erosion and
rill erosion exhibited a power function relationship with an increase in rainfall intensity and slope. The impact of
slope gradient on soil erosion on the RT slope was greater than those on the CK and AD slopes. During the
development of rill erosion, non-stationary and multi-scale periodic phenomena exist in the time series of runoff
and sediment on different microreliefs. As surface microrelief increased, the fluctuations in runoff and sediment
on tilled slopes intensified, resulting in a gradual shortening of the variation period. And this study verified the
feasibility of the SIMWE model to simulate the runoff and sediment dynamics on microtopographic scale under a
single rainfall event. The results can provide a theoretical basis for the scientific layout of soil and water con­
servation practices in sloping farmland and the construction of soil erosion model in purple soil area.

1. Introduction microtopography is regarded as an effective measure to control soil and

water loss in sloping farmland (da Rocha Junior et al., 2016). Tillage
Soil surface microtopography is the result of the joint action of practices cause soil redistribution, resulting in rapid changes in soil
erosion behavior (wind erosion or water erosion) and farming activities profile characteristics and microtopography, which in turn affect hy­
(Zobeck and Popham, 1997; Zheng et al., 2014; Bauer et al., 2015), drological and erosive processes. Many studies have indicated that
which has a vital impact on surface runoff and the erosion process. At tillage-induced microtopography affects runoff-erosion processes by
the same time, the runoff accumulation in depressions, sediment changing the microtopographic patterns and the near-surface soil physic
transportation, and deposition affect the spatial distribution of micro­ properties (Sun et al., 2021; Zhao et al., 2021; Luo et al., 2022). The
topography (Bahddou et al., 2023; Luo et al., 2023). The interaction influence of tillage-induced microtopography on soil loss remains
between microtopography and the erosion process alters with the controversial. Some authors suggested that tillage-induced micro­
change in microtopographic pattern (Zhao et al., 2016). Tillage-induced topography can increase the depression storage capacity and soil

* Corresponding author.
E-mail address: [email protected] (Z. Zheng).

https://doi.org/10.1016/j.geoderma.2023.116539
Received 27 February 2023; Received in revised form 4 May 2023; Accepted 25 May 2023
Available online 31 May 2023
0016-7061/© 2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
J. Luo et al. Geoderma 436 (2023) 116539

infiltration, intercept runoff and weaken runoff energy, thus decreasing used the spatial distributions of water depth, discharge, and sediment
soil loss (Zheng et al., 2014; Zheng et al., 2021). Some authors indicated flux simulated by the SIMWE model to identify the damaged positions of
that tillage-induced microtopography can also increase soil loss by terraces. Pijl et al. (2020) found that the SIMWE model had a good
concentration overland flow (Liu et al., 2014a; Luo et al., 2018). performance in evaluating the erodible areas under different soil and
Rill erosion is the dominant sediment source of slope soil erosion and water conservation measures through water depth distribution. In this
a transport channel of eroded products. The contribution of rill erosion paper, the SIMWE model was used to simulate the overland flow
to total soil loss on sloping farmland can reach more than 70% (Kimaro movement and erosion pattern, identify the erodible areas on different
et al., 2008; Auerswald et al., 2009). Rill erosion is mainly caused by the tilled slopes and then clarify the essential differences among the runoff
surface runoff generated by rainfall (Govers et al., 2007). Under the and sediment reduction benefits on different tilled slopes. Therefore,
splashing, scouring, and transportation of rainfall and runoff, the rill this study aims to 1) elucidate the effect of microtopographic patterns on
networks are formed on the slope soil surface (Berger et al., 2010). On rill erosion; 2) investigate the relationships among soil erosion, slope
the one hand, the rill networks would enhance sediment and hydro­ and rainfall intensity on different tilled slopes; 3) analyze the temporal
logical connectivity and reduce the water storage capacity of the soil variation characteristics of runoff and sediment yields using wavelet
surface (Zhang et al., 2019). On the other hand, the formation of rills analysis; and 4) simulate overland flow-sediment fluxes on different
enhances the sediment transport capacity of overland flow, which tilled slopes based on SIMWE model.
further aggravates soil erosion (Wirtz et al., 2012). Overland flow is the
main driving force of rill development on the slope. The rill development 2. Materials and methods
with time and space affects the hydraulic characteristics of overland
flow (Sun et al., 2021). Meanwhile, the dynamic change of overland flow 2.1. Experimental material
will inevitably lead to the further development of rill morphology
(Zhang et al., 2016). This phenomenon is mainly related to the inter­ The purple soil used in the experiment was collected from farmland
active relationship between microtopography change and rill develop­ with 20 cm topsoil, which located in Ziyang key field scientific obser­
ment. And studies on this topic have rarely been reported. vation station of agricultural resources and ecological environment in
Slope gradient and rainfall intensity are important factors affecting the upper reaches of Yangtze River, Ministry of Agriculture and Rural
soil erosion in sloping farmland. Shen et al. (2016) found that the rill Affairs, China. The test soil is classified as Entisol (Soil Survey Staff,
erosion rate increased by a power function as rainfall intensity (ranging 2010) (as detailed in Table 1). This study employed three soil boxes,
from 50 to 100 mm h− 1) and slope (ranging from 10◦ to 20◦ ) increased each with dimensions of 4.0 m in length, 0.8 m in width, and 0.5 m in
on a smooth slope of loessial soil, and the rainfall intensity had a greater depth. The soil utilized in the experiment underwent air drying until it
impact on rill erosion compared to slope gradient. However, Jiang et al. reached a 5% moisture content and was subsequently sifted through a
(2018) showed that the influence of slope on rill erosion was greater 10 mm sieve. To facilitate free drainage, 10 cm deep sand was placed at
than that of rainfall intensity (ranging from 100 to 140 mm h− 1) on steep the bottom of each soil box. Each box contained 40-cm deep air-dried
colluvial deposits (ranging from 20◦ to 40◦ ). The relationships among soil and the soil was deposited in successive 10-cm thick layers (An
rill erosion, slope gradient and rainfall intensity on different tilled slopes et al., 2012). After each soil box was filled, three tillage practices were
are still unclear. applied to simulate microreliefs. These tillage practices, namely smooth
The formation and evolution of soil erosion in sloping farmland slope (CK), artificial digging (AD), and ridge tillage (RT), are widely
affect the runoff pattern (Luo et al., 2017), resulting in periodic and utilized for agricultural production in sloping farmland located in the
transient phenomena in the hydrological process on the slope. The hilly region of the Sichuan basin. CK creates a smooth surface, AD
runoff-sediment process has complex nonlinear and multiscale temporal produces mounds and depressions, with a depression depth of 7 cm and
variations (Luo et al., 2020). Linear analysis of essentially nonlinear a 25 cm distance between depressions, while RT creates a ridge-furrow
runoff-sediment changes is inherently limited (Zhang et al., 2017). At system with a 15 cm high ridge and an 80 cm between-ridge distance
present, when studying soil erosion on slopes, researchers tend to focus (Fig. 1). CK and AD were categorized as random roughness, while RT
solely on the spatial variability of runoff and sediment, while dis­ was identified as oriented roughness.
regarding their temporal variability. Several studies have demonstrated
that wavelet analysis can accurately detect the main periods hidden in 2.2. Simulated rainfall experiment
time series (Liu et al., 2020; Rezaei and Gurdak, 2020). The temporal
variation of runoff and sediment serves not only as an indicator of the A side-sprinkle rainfall simulator equipped with two V-80100 noz­
erosive behave against soil surface roughness (SSR), but also as a po­ zles was used in this study. The height of the rainfall simulator was 6 m,
tential metric to differentiate between various microrelief features on and it had an effective rainfall area of 48 m2. The simulator exhibited a
the soil surface (Luo et al., 2020). Compared with the data obtained rainfall uniformity of approximately 85%. Three typical slope gradients
through physical experiments, the model simulation method can more (10◦ , 15◦ , and 20◦ ) and three representative erosive rainfall intensities
quickly and effectively clarify the characteristics of overland flow and (1.0 mm min− 1, 1.5 mm min− 1, and 2.0 mm min− 1) were applied in this
sediment movement on the slope. The empirical soil erosion prediction study. The rainfall kinetic energy values for intensities of 1.0 mm min− 1,
model commonly used at present, its disadvantage is that it usually only 1.5 mm min− 1, and 2.0 mm min− 1 were 327 J m− 2 h− 1, 518 J m− 2 h− 1,
considers the total runoff and sediment yields at the outlet of the slope and 835 J m− 2 h− 1, respectively. To equalize soil moisture levels for all
and regards the soil erosion process as a “black box” (Shen et al., 2016; treatments, a 10-min prewetting rain event at an intensity of 0.25 mm
Xu et al., 2017), neglects or simplifies the dynamic change of hydro­ min− 1 was carried out 24 h prior to the start of the rainfall experiments.
logical processes and the spatial distribution of sediment in the slope- A total of 27 treatments were conducted in the rainfall experiments, and
gully system, and it is incapable of simulating the complex erosion each treatment was repeated 3 times. For each run, soil boxes were
patterns resulting from the transport and conservation of surface fluxes prepared with fresh soils. The rainfall event for each treatment was
across time and space (Pijl et al., 2020). Early models, such as SEMMED performed with 90-minute. Runoff were sampled at the outlet of the soil
(De Jong et al., 1999), ANSWERS (Beasley et al., 1980), and LISEM (De box at the intervals of 3 min. The measurement of runoff and sediment
Roo et al., 1994), all simplify the description of the soil erosion process, yields has been described by Luo et al. (2022).
so they cannot depict the spatial pattern of soil erosion in high detail.
SIMWE model is developed based on the theoretical framework of the 2.3. The effect of tillage-induced microtopography on soil and water loss
WEPP model and it can accurately simulate the flow movement and
erosion pattern at a microtopographic scale. Fernandes et al. (2017) Runoff and sediment control benefits of tillage practices can be

2
J. Luo et al. Geoderma 436 (2023) 116539

Table 1
Physical-chemical properties of experimental soil.
Soil type pH Bulk density CEC SOC CaCO3 Sand Silt Clay Texture
− 3 − 1 − 1
(g cm ) (cmol kg ) (g kg ) (%) (%) (%) (%)

Purple soil 7.5 1.2 21 13 12 49 29 22 Clay loam

Fig. 1. Photographs of three boxes with different tillage-induced microreliefs. CK: smooth slope; AD: artificial digging; RT: ridge tillage.

expressed as follows (Zhao et al., 2019):

∫∞
( ) 1 t− b
Rs − Rt Wf (a, b) = √̅̅̅ f (t)ϕ* ( )dt (4)
RCB = × 100% (1) a a
Rs − ∞

( )
Ss − St where Wf (a, b) is the wavelet coefficient, a is the scale factor, b is the
SCB = × 100% (2) time factor, i is the imaginary number, c is a constant, and * is the
Ss
conjugate.
where RCB and SCB are runoff and sediment control benefits, respec­ After the completion of the wavelet transform, the wavelet co­
tively; Rs and Rt are runoff loss from CK slope and tilled slope, respec­ efficients stemming from the extended data located at both ends were
tively; Ss and St are soil loss from CK slope and tilled slope, respectively. removed, while the coefficients derived from the original time series
data for runoff and sediment were retained.
2.4. Wavelet analysis The wavelet variance is defined as follows:
∫∞
Wavelet analysis is a powerful time-frequency analysis technique
Var(a) = |Wf (a, b)|2 db (5)
that can efficiently detect periodic variations in a time series and predict
the evolutionary patterns of a system across various time scales. The
− ∞

selection of a suitable wavelet type is crucial for wavelet analysis. The The wavelet variance map provides insight into the periodic distri­
Morlet wavelet is a periodic function enveloped by a Gaussian function, bution of wave energy at different scales. The main period in a time
which is extensively employed in the recognition of periodic oscillations series can be identified through the wavelet variance map.
in hydrological processes (Labat, 2005).
Wavelet functions are a class of functions with oscillatory charac­
2.5. SIMWE model
∫∞
teristics that decay rapidly to zero, i.e., ∅(t)dt = 0. The Morlet
The SIMulated Water Erosion model (SIMWE) was developed by
Mitas and Mitasova (1998) according to the basic principle of the Water
− ∞
wavelet is a type of complex wavelet that exhibits a difference of π/2
Erosion Prediction model (WEPP). It is a physical process model for
between its real and imaginary parts, eliminating the oscillations of the
estimating soil erosion starting from a single rainfall event. The SIMWE
modulus of the real wavelet coefficients during the wavelet trans­
model mainly includes two modules. The first module is based on the
formation process.
bivariate form of Saint-Venant equations that combines the continuity
The Morlet wavelet is defined as follows:
relation of the water flow with the momentum conservation equation to
ϕ(t) = eict e− t2 /2
(3) describe the overland flow movement. The second module describes the
process of sediment transport according to the continuity relation of
As the runoff and sediment data collected for this study was treated sediment mass. These two modules are respectively integrated in the r.
as finite time series, there was a possibility of boundary effects arising at sim.water and r.sim.sediment of the GRASS GIS software.
the beginning and end of the time series. To mitigate these effects, the To simulate the overland flow movement on different microtopo­
runoff and sediment signals underwent an extension process at both graphic patterns, the input parameters mainly included high-resolution
ends of the time series, utilizing the signal extension function within DEM and its first-order derivatives along the × and y directions, rainfall
Matlab software. This enabled successful anomaly treatment of the intensity, infiltration rate, and Manning’s roughness coefficient. Point
runoff and sediment data, paving the way for a continuous Morlet clouds with spatial resolution of 1 mm × 1 mm was acquired by a laser
wavelet transform to be performed. scanner (MAPTEK I-Site 8820) to generate high-resolution DEM. The
For any time series f(t) ∈ L2 (R), its continuous wavelet transform last two parameters were directly measured in the experiment. In this
is defined as: part, the treatment of 1.5 mm min− 1 with 90 min rainfall duration, 15◦
was selected to obtain high-resolution DEM. The measured infiltration

3
J. Luo et al. Geoderma 436 (2023) 116539

rates of CK, AD, and RT slopes were 19 mm h− 1, 24 mm h− 1, and 28 mm

h− 1, respectively. And the measured Manning’s roughness coefficients of
CK, AD, and RT slopes were 0.06, 0.16, and 0.27, respectively. At the
end of the simulation, the spatial distributions of water depth and water
discharge on different tillage slopes were obtained. The input parame­
ters for simulating the sediment transport process on different tillage
slopes mainly included the water depth, soil detachment capacity co­
efficient, transport capacity coefficient, and critical shear stress. The last
three parameters were calculated directly by in-situ observation. The
soil detachment capacity coefficient of different tillage slopes was set
equal to 0.003 s m− 1 according to the soil texture of purple soil and the
calculation formula of soil detachment capacity in the WEPP Model
(Flanagan and Nearing, 1995). The transport capacity coefficient was set
equal to 0.008 s based on the relationship between soil particles of
purple soil and sediment transport capacity by overland flow (Zhang
et al., 2011). The critical shear stress of the CK, AD, and RT slopes were
set equal to 1.214 Pa, 2.505 Pa, and 3.323 Pa, respectively. At the end of
the simulation, the spatial distributions of sediment flux on different
tillage slopes were obtained.
Fig. 3. Change characteristics of sediment yield on tilled slopes under different
3. Results slope gradients and rainfall intensities. CK: smooth slope; AD: artificial digging;
RT: ridge tillage.

3.1. Runoff and sediment yield characteristics on different tilled slopes

and RT slopes respectively increased by 36.9% and decreased by 1.8%.
According to Fig. 2, the runoff on the CK slope was between 119.6
and 480.5 L, with an average value of 299.5 L under different treat­ 3.2. Slope gradient and rainfall intensity impact on soil erosion
ments. The runoff on the AD slope was between 59.6 and 440.1 L, with
an average value of 202.5 L. For the RT slope, the runoff yield was be­ As the rainfall intensity and slope increased, the runoff, soil loss, and
tween 29.9 and 397.5 L, with an average value of 168.2 L. In general, the rill erosion rates on various tilled slopes increased (Table 2). On the CK
runoff yields of different tillage practices were CK > AD > RT, and the slope, the average soil loss rate was 10.9 kg m− 2 h− 1 and the rill erosion
runoff yield on each tilled slope increased as rainfall intensity and slope rate was 7.5 kg m− 2 h− 1. On the AD slope, the average soil loss rate was
increased. As shown in Fig. 3, the sediment yield on the CK slope was 11.0 kg m− 2 h− 1 and the rill erosion rate was 8.6 kg m− 2 h− 1. The
between 15.7 and 101.9 kg, with an average value of 50.1 kg under average soil loss rate and rill erosion rate were 10.8 kg m− 2 h− 1 and 9.0
different treatments. The sediment yield of the AD slope was between kg m− 2 h− 1, respectively, on the RT slope. Rill erosion accounted for
9.7 and 106.9 kg, with an average value of 44.5 kg. For the RT slope, the 58% − 74% of the soil loss rate on the CK slope, and for the AD and RT
sediment yield was between 8.1 and 104.2 kg, with an average value of slopes, rill erosion rates accounted for 72% − 84% and 78% − 87% of
40.4 kg. The variations of sediment yield on tilled slopes were compli­ the total loss rates, respectively.
cated under different conditions. For example, the sediment yields were Note. CK: smooth slope; AD: artificial digging; RT: ridge tillage. The
shown as CK > AD > RT in the 1.0 mm min− 1, 10◦ treatment. The data are the average of three replicates ± standard deviation.
sediment yields of AD and RT slopes were reduced by 38.2% and 48.4% In order to further quantify the rainfall intensity and topography
respectively compared to CK slope. However, the sediment yields of factors impacts on soil erosion on different tilled slopes, the Curve
different tillage practices were AD > CK > RT in the 2.0 mm min− 1, 15◦ Fitting Tool in MATLAB R2018a was used to fit the soil erosion equation
treatment. Compared with the CK slope, the sediment yields of the AD (Fig. 4). The optimal fitting formulas are as follows:
( )
CK : SL = 0.2553R1.073 S1.221 R2 = 0.954, n = 27 (6)
( )
AD : SL = 0.1908R1.668 S1.231 R2 = 0.933, n = 27 (7)
( )
RT : SL = 0.1546R1.425 S1.337 R2 = 0.970, n = 27 (8)

where SL is soil loss rate (kg m− 2 h− 1); R is rainfall intensity (mm

min− 1); S is slope gradient (◦ ).
Soil loss rates on tilled slopes were found to increase in a power
function as both rainfall intensity and slope increased, as indicated by
Eqs. (6), (7), and (8). Comparing the exponents of rainfall intensity in
these equations, it is apparent that the effect of rainfall intensity on soil
loss rates was greater for AD and RT slopes than for CK slope. Addi­
tionally, the impact of slope gradient on soil loss rates was greater for RT
slope compared to CK and AD slopes. And the pattern of rill erosion was
similar to that of soil erosion, as shown in Fig. 5. The relationship be­
tween rill erosion and rainfall intensity and topography factors can be
expressed as follows:
( )
CK : RE = 0.1260R1.286 S1.306 R2 = 0.960, n = 27 (9)
Fig. 2. Change characteristics of runoff yield on tilled slopes under different
( )
slope gradients and rainfall intensities. CK: smooth slope; AD: artificial digging; AD : RE = 0.1114R1.853 S1.302 R2 = 0.952, n = 27 (10)
RT: ridge tillage.

4
J. Luo et al. Geoderma 436 (2023) 116539

Table 2 Fig. 6(A), the AD slope had a strong runoff control benefit (RCB) under
Runoff, soil loss and rill erosion rates on tilled slopes under different slope different conditions. Specifically, compared to the CK slope, the runoff
gradients and rainfall intensities. was reduced by 8.4% − 50.2%. As shown in Fig. 6(B), in the 1.0 mm
Tillage Slope Rainfall Runoff Soil loss Rill erosion min− 1 treatment, the SCB of AD slope decreased by 4.1% and 23.7%
practices gradients intensities rate (L rate (kg rate (kg respectively when the slope gradient increased from 10◦ to 15◦ and 20◦ .
(◦ ) (mm min− 1) m− 2 h− 1) m− 2 m− 2 h− 1) In the 1.5 mm min− 1 treatment, the SCB of AD slope decreased by 49.6%
h− 1)
and 91.9% respectively when the slope gradient increased from 10◦ to
CK 10 1.0 28.8 ± 3.8 ± 2.2 ± 0.1 15◦ and 20◦ . It is noteworthy that when the rainfall intensity increased
3.4 0.3
to 2.0 mm min− 1, the SCB of the AD slope changed from sediment
1.5 57.2 ± 6.8 ± 4.6 ± 0.2
6.9 0.8 control benefit to sediment increase effect as slope increased. Under the
2.0 94.4 ± 10.9 ± 7.6 ± 0.6 same condition of slope gradient, the SCB of the AD slope decreased
12.3 1.5 gradually when rainfall intensity was increased.
15 1.0 35.8 ± 6.1 ± 3.7 ± 0.2 Based on Fig. 7(A), the RT slope had a strong runoff control benefit
4.3 0.7
1.5 65.3 ± 9.6 ± 6.2 ± 0.6
(RCB) under different conditions. Compared with the CK slope, the
9.1 1.3 runoff was reduced by 17.3% − 75.0%. As shown in Fig. 7(B), in the 1.0
2.0 96.1 ± 13.0 ± 9.5 ± 1.0 mm min− 1 treatment, the SCB of RT slope decreased by 16.3% and
13.9 1.8 28.3% respectively when slope was increased from 10◦ to 15◦ and 20◦ .
20 1.0 42.7 ± 10.9 ± 7.2 ± 0.6
In the 1.5 mm min− 1 treatment, the SCB of the RT slope decreased by
5.8 1.3
1.5 67.8 ± 15.8 ± 10.7 ± 0.9 24.7% and 87.6% respectively when slope was increased from 10◦ to 15◦
8.1 2.4 and 20◦ . In the 2.0 mm min− 1 treatment, the SCB of the RT slope
2.0 100.1 ± 21.2 ± 15.7 ± 1.7 decreased by 94.3% when slope was increased from 10◦ to 15◦ . And the
17.9 3.2 SCB of the RT slope was lost when slope was reached to 20◦ .

AD 10 1.0 23.3 ± 3.0 ± 2.2 ± 0.1 3.4. Multi-temporal variations of runoff and sediment yield
2.6 0.2
1.5 41.0 ± 4.3 ± 3.3 ± 0.2
5.3 0.3 By calculating the wavelet coefficient variance, the main periods of
2.0 63.6 ± 10.0 ± 7.9 ± 1.2 runoff and sediment dynamics on tilled slopes can be verified. Fig. 8(A)
8.9 1.1 shows the wavelet variance maps of runoff on tilled slopes under
15 1.0 30.2 ± 4.7 ± 3.4 ± 0.3 different treatments. Taking the slope of 10◦ and rainfall intensity of 1.0
3.3 0.4
mm min− 1 as an example, there were two peaks in the time series of
1.5 46.0 ± 9.1 ± 6.6 ± 0.7
5.5 1.0 surface runoff on the CK slope, corresponding to 14 min and 32 min,
2.0 85.1 ± 20.6 ± 16.1 ± 2.1 with the largest peak at a timescale of 32 min, indicating that the surface
14.5 2.5 runoff changed most drastically at this timescale, which was the first
20 1.0 35.2 ± 9.2 ± 6.8 ± 0.6
main period of runoff changes under this condition, with 14 min as the
4.9 1.1
1.5 49.7 ± 15.7 ± 12.1 ± 1.5 second main period. There were two peaks in the time series of surface
6.5 1.7 runoff on the AD slope, corresponding to 23 min and 32 min, with the
2.0 91.7 ± 22.3 ± 18.7 ± 2.6 largest peak at a timescale of 32 min, which was the first main period of
11.9 3.1 runoff changes for the AD slope, with 23 min as the second main period.
Also there were two peaks in the time series of surface runoff on the RT
RT 10 1.0 17.0 ± 3.6 ± 2.8 ± 0.1 slope, corresponding to 7 min and 22 min, with the largest peak at a
2.6 0.2 timescale of 22 min, which was the first main period of runoff changes
1.5 28.2 ± 4.4 ± 3.6 ± 0.2
3.1 0.3
for the RT slope, with 7 min as the second main period. Fig. 8(B) shows
2.0 62.2 ± 9.5 ± 8.0 ± 0.9 the wavelet variance maps of sediment on tilled slopes. Taking the slope
7.5 1.2 of 10◦ and rainfall intensity of 1.0 mm min− 1 as an example, the main
15 1.0 21.1 ± 7.7 ± 6.4 ± 0.7 period of sediment changes on the CK slope was 13 min, and for the AD
3.4 0.7
slope was also 13 min. There were two peaks in the time series of
1.5 41.1 ± 9.4 ± 7.8 ± 0.9
5.7 0.9 sediment on the RT slope, corresponding to 6 min and 32 min, with the
2.0 80.1 ± 16.0 ± 13.1 ± 1.4 first main period at 32 min and the second main period at 6 min.
9.8 1.8
20 1.0 24.6 ± 7.5 ± 6.3 ± 0.6
3.5. Spatial simulation of overland flow and sediment fluxes
3.9 0.6
1.5 54.7 ± 15.9 ± 13.2 ± 1.5
7.7 2.1 Fig. 9(A) shows the spatial distribution of water depths computed for
2.0 85.7 ± 22.4 ± 19.5 ± 3.1 SIMWE simulations on different microtopographic patterns. For the CK
11.4 3.4 slope, the water depth gradually increased from upslope to downslope.
The water depth of less than 0.5 cm was mainly distributed in the up­
( ) slope area, accounting for 71.22% of the total grid of water depth on the
RT : RE = 0.1038R1.482 S1.406 R2 = 0.973, n = 27 (11) slope. The water depth on the downslope increased as upslope drainage
where RE is rill erosion rate (kg m − 2 − 1
h ); R is rainfall intensity (mm area increased. The water depth of less than 1.5 cm was mainly
distributed in rills at the middle and lower sections of the slope. The
min− 1); S is slope gradient (◦ ).
water depths in the rill area were significantly larger than those in the
interrill area, and the water depths in the rills were more than 2 cm. For
3.3. Soil erosion control of different tilled slopes the AD slope, the runoff was collected at the bottom of the depression
and the water depth was relatively large. The water depth of less than 1
Fig. 6 shows the soil and water control benefits of AD slope under cm accounted for 90.75% of the total grid of water depth on the slope.
different treatments. In general, soil and water control benefits The water depths greater than 4 cm were recorded in the rills that were
decreased as both rainfall intensity and slope increased. As shown in formed on the middle and lower sections of the slope. And for the RT

5
J. Luo et al. Geoderma 436 (2023) 116539

Fig. 4. Relationships of soil loss rate with rainfall intensity and slope gradient on different tilled slopes. CK: smooth slope; AD: artificial digging; RT: ridge tillage.

Fig. 5. Relationships of rill erosion rate with rainfall intensity and slope gradient on different tilled slopes. CK: smooth slope; AD: artificial digging; RT: ridge tillage.

Fig. 6. Runoff control benefit (RCB) (A) and sediment control benefit (SCB) (B) on AD slope under different slope gradients and rainfall intensities.

Fig. 7. Runoff control benefit (RCB) (A) and sediment control benefit (SCB) (B) on RT slope under different slope gradients and rainfall intensities.

6
J. Luo et al. Geoderma 436 (2023) 116539

Fig. 8. Wavelet variance diagrams of runoffs (A) and sediment yields (B) on tilled slopes under different slope gradients and rainfall intensities. CK: smooth slope;
AD: artificial digging; RT: ridge tillage.

slope, the water depth of less than 1.5 cm accounted for 97.80% of the 0.0003 m3 s− 1 accounted for 92.09% of the total grid of water discharge
total grid of water depth on the slope, which was mainly distributed on on the slope. The maximum water discharge at the position of ridge
the ridge sideslope, while the water depth could reach 5.7 cm at the failure was 0.0014 m3 s− 1.
position of ridge collapse. Compared with the CK slope, the percentage When assessing the soil and water loss on the slope, only the runoff
of water depth grids greater than 1 cm on AD and RT slopes increased by and sediment yields at the outlet are usually considered, while the dy­
12.3% and 25.5%, respectively. namic migration process of surface fluxes is often simplified or ignored.
Fig. 9(B) shows the spatial distribution of water discharge computed Fig. 9(C) shows the spatial distribution of sediment flux on different
for SIMWE simulations on different microtopographic patterns. For the microreliefs based on the SIMWE model. The sediment fluxes in the rill
CK slope, the water discharge increased gradually as slope length area on different microtopographic patterns are significantly larger than
increased. The water discharges in the rill area were obviously larger those in the interrill area. For the CK slope, the average value of sedi­
than those in the interrill area. The water discharge of less than 0.00015 ment flux was 0.0025 kg m− 2 s− 1, and the maximum value was 0.0036
m3 s− 1 was mainly distributed in the upper and middle parts of the slope, kg m− 2 s− 1. On the AD slope, the average value of sediment flux was
accounting for 81.19% of the total grid of water discharge on the slope. 0.0023 kg m− 2 s− 1, and the maximum value was 0.0037 kg m− 2 s− 1. The
The maximum water discharge on the slope was 0.00079 m3 s− 1. For the average value of sediment flux on the RT slope was 0.0026 kg m− 2 s− 1,
AD slope, the flow paths connection between depressions made the and the maximum value is 0.0079 kg m− 2 s− 1, which occurred at the
water discharge relatively large in rills, and the maximum water ridge collapse on the downslope. According to Fig. 10, the predicted
discharge was 0.00092 m3 s− 1. For the RT slope, the water discharges on values of sediment fluxes based on SIMWE simulation on different tilled
the ridge sideslope were relatively small, the water discharge of less than slopes were less than the measured values.

7
J. Luo et al. Geoderma 436 (2023) 116539

Fig. 9. Spatial distributions of water depths (A), water discharges (B) and sediment fluxes (C) computed for SIMWE simulations on different tilled slopes. CK: smooth
slope; AD: artificial digging; RT: ridge tillage.

4. Discussion

The influence of soil surface microtopography on soil erosion is still

unclear. Some researchers suggest that soil surface microtopography can
increase the dispersion resistance of soil particles caused by raindrop
splashing, increase the water storage capacity on the surface and reduce
the flow velocity, leading to decrease soil loss (Darboux et al., 2002;
Planchon et al., 2002). However, some authors claim that a rougher
microtopography can concentrate flow and have potential scouring ac­
tion, leading to the increase of rills and aggravating soil erosion (Gómez
and Nearing, 2005; Le Bissonnais et al., 2005). Our study found that the
impact of microtopography on soil erosion was closely related to rainfall
intensity and slope factors, which is in line with the results reported by
Zhao et al. (2021). As the rainfall intensity and slope increased, the RCB
and SCB of various tilled slopes showed a gradual decrease. The AD and
RT slopes had strong runoff control benefits under different treatments.
The maximum RCB of the AD and RT slopes reduced by 50.2% and
75.0% respectively compared to CK slope. Similarly, the maximum SCB
of the AD and RT slopes reduced by 51.6% and 53.0% respectively
Fig. 10. Comparison of measured and simulated values of sediment flux. compared to CK slope. The RCB on the AD slope in the 15◦ , 2.0 mm
min− 1 treatment changed from sediment control benefit to sediment
increase effect, and the maximum soil loss on the slope increased by
36.9% compared to CK slope. And the maximum soil loss on the RT slope

8
J. Luo et al. Geoderma 436 (2023) 116539

in the 20◦ , 2.0 mm min− 1 treatment increased by 2.3%. The results wavelet spectrum was controlled by the time scale distribution of runoff
indicated that there were critical conditions for the soil and water and sediment, which in turn reflects the evolution characteristics of
conservation effect on tilled slopes. Due to the influence of tillage runoff and sediment time series and its response mechanism to surface
practice, concentrated flow was formed on tilled slopes whereas sheet microtopography. Wavelet analysis results showed that there were pe­
flow was generated more on the smooth slope. The detachment and riodic variations with 3–32 min in runoff and sediment time series on
transportation capacities of concentrated flow were higher than those of different tilled slopes, and as the surface microrelief increased, the
sheet flow, resulting in an increase in soil loss on the tilled slopes, which fluctuation of runoff and sediment gradually enhanced, and the varia­
is magnified with an increase in rainfall intensity and slope gradient tion period gradually shortened. In particular, for the RT slope, the
(Zhao et al., 2021). And the response of surface microtopography to soil strong signal oscillation displayed in the time–frequency domain of the
erosion varies with changes in hydrological-sediment connectivity. Luo real part of the wavelet transform was closely related to the ridge
et al. (2023) indicated the formation of rills resulted in increased collapse.
hydrological-sediment connectivity on the tilled slopes, therefore, flow The spatial distribution of water depth on different microreliefs was
velocity and sediment transport increased with the evolution of water closely related to the microtopographic structure. For the CK slope, the
erosion. These could be the reason for the larger soil loss from AD and RT water depth increased gradually as slope length increased. The distri­
slopes as compared to CK slope. We also found that the rill erosion rate bution of water depth of less than 0.5 cm was the largest on the slope,
on different tilled slopes accounted for 72% − 87% of the total soil loss and the water depth in the rill area was significantly larger than that of
rate, which was similar to the results reported by Mutchler and Young the interrill area. The water depths in rills were more than 2 cm. For the
(1975). The result proved that rill erosion was the most severe form of AD slope, the rainwater was collected in the depressions. The distribu­
water erosion on hillslopes. Conducting research on rill erosion is of tion of water depth of less than 1 cm was the largest on the slope. The
great significance for the effective prevention and control of soil erosion water depths greater than 4 cm were recorded in the rills that were
on hillslopes (Shen et al., 2016). formed on the middle and lower sections of the slope. For the RT slope,
The rainfall intensity and slope gradient played important roles in the distribution of water depth less than 1.5 cm was the largest on the
soil erosion (Ziadat and Taimeh, 2013; Xu et al., 2017). Based on slope. And the water depth can reach up to 5.7 cm at the position of ridge
Table 2, both the soil loss rate and the rill erosion rate on different failure. Compared with the CK slope, the percentage of grids with water
microreliefs increased with an increase in both rainfall intensity and depth greater than 1 cm on the AD and RT slopes increased by 12.3%
slope, but the increments of the soil loss rate and the rill erosion rate and 25.5%, respectively, which indicated that the soil surface micro­
under different treatments were quite different. The reason for this topography played a role in harvesting rainwater and promoting the
phenomenon was that the critical values of soil erosion and rill erosion flow convergence and connectivity. The sediment flux in rills on
on different tilled slopes were different. Eqs. (6), (7) and (8) indicate that different tilled slopes was significantly greater than that in the interrill
rainfall intensity had greater impact on soil erosion on AD and RT slopes area. The mean values of sediment fluxes were shown as RT > CK > AD.
than that on CK slope, which was mainly related to the change of runoff The measured sediment flux values were higher than the predicted
generation patterns. The runoff generation patterns directly affect the values obtained through SIMWE simulation on different tilled slopes.
soil erosion process. The impact of slope gradient on soil erosion on the This result occurred because the existence of seepage on the slope
RT slope was greater than those on CK and AD slopes. This was mainly altered the rainfall-runoff relationships. Many studies have demon­
due to ridge tillage being horizontal tillage in the direction that was strated that the presence of water seepage flow increased the runoff and
perpendicular to the slope. The higher the slope gradient is, on the one soil moisture content, decreased the soil aggregate cohesion and sta­
hand, the water storage in the furrows decreases; on the other hand, the bility, and thus aggravated soil erosion (Vandamme and Zou, 2013;
component force parallel to the slope increases, the weaker the stability Estabragh et al., 2016; Masoodi et al., 2017). Purple soil formed by the
of the soil mass, and the more likely to cause ridge failure driven by series of Triassic-Cretaceous purple rock has the characteristic of rock-
runoff (Liu et al., 2014b; Luo et al., 2020). This finding expanded our soil structure (Zhong et al., 2019), contributing to the infiltrated water
understanding of the dominant control factors of rill erosion on slope accumulates and seepage develops actively (Tao et al., 2022). In future
farmland of purple soil. In the low rainfall intensity and slope treatment, studies on soil erosion in sloping farmland of purple soil, the influence of
at the beginning of rainfall, the soil infiltration rate was higher than the seepage on the runoff-sediment process should be taken into consider­
rainfall intensity and almost all the rainwater infiltrated into soil. Soil ation. Additionally, this study was conducted under laboratory condi­
infiltration rate decreased as cumulative rainfall increased. When the tions, and the edge effect of the soil box may have affected the results of
rainfall intensity was higher than the soil infiltration capacity, the excess the experiment. Further research should be conducted under field con­
rainwater was stored in the depressions. Compared with the CK slope, ditions to validate the findings.
the AD and RT slopes have stronger water storage capacity, and signif­
icantly reduced soil and water loss in low rainfall intensity and slope 5. Conclusions
treatment. However, the SCB of the AD and RT slopes changed from
sediment control benefit to sediment increase effect with an increase in This study explored the effect of microtopographic patterns on hill­
rainfall intensity and slope. This is mainly due to the increase in rainfall slope soil erosion. The spatial distribution of microtopography patterns
intensity and slope gradient promoting the evolution of water erosion. A differed under different tillage practices, which differently affected the
large amount of sediment was transported with the enhancement of flow soil and water control benefits of tillage practices. In general, the runoff
connectivity on the slope, and even a higher degree of microrelief acted and sediment reduction benefits on different microreliefs gradually
as a source of soil erosion (Luo et al., 2018). In this study, the SCB of decreased with an increase in both rainfall intensity and slope, and the
different tilled slopes was basically lost under the condition of 20◦ . sediment control benefits of tilled slopes were basically lost when the
Research has shown that the generation and evolution of runoff and slope gradient increased to 20◦ . The rill erosion exhibited significant
sediment under continuous rainfall conditions is a nonlinear and com­ contributions to soil erosion, accounting for 72%-87% of hillslope soil
plex process that is controlled by multiple factors and exhibits multiple loss. A power function relationship was observed in the increase of both
time scale variations (Tian et al., 2019). Due to the complex composition soil erosion and rill erosion rates as the rainfall intensity and slope
of materials, poor homogeneity, and interference from tillage practices, increased. Non-stationary and multiscale periodic phenomena was
the spatiotemporal variability of runoff and sediment processes on tilled observed in runoff and sediment time series on tilled slopes. As surface
slope is even more complex. Wavelet analysis is a multiscale analysis microrelief increased, the fluctuations of runoff and sediment on tilled
method that can reveal the time-varying characteristics of time series slopes intensified, resulting in a gradual shortening of the variation
from both the time and frequency domains (Labat, 2005). The pattern of period. A physically-based model (SIMWE) was used in this study to

9
J. Luo et al. Geoderma 436 (2023) 116539

assess spatial patterns of runoff and soil erosion at microtopographic Liu, H., Yu, Y., Zhao, W., Guo, L., Liu, J., Yang, Q., 2020. Inferring subsurface
preferential flow features from a wavelet analysis of hydrological signals in the Shale
scale. Our results also indicated that the existence of seepage on the
Hills catchment. Water Resour. Res. 56 (11). e2019WR026668.
slope altered the rainfall-runoff relationships. The influence of seepage Liu, Q.J., Shi, Z.H., Yu, X.X., Zhang, H.Y., 2014a. Influence of microtopography, ridge
on soil and water loss on sloping farmland of purple soil cannot be geometry and rainfall intensity on soil erosion induced by contouring failure. Soil
ignored. Till. Res. 136, 1–8.
Liu, Q.J., Zhang, H.Y., An, J., Wu, Y.Z., 2014b. Soil erosion processes on row sideslopes
within contour ridging systems. Catena 115, 11–18.
Luo, J., Zheng, Z., Li, T., He, S., 2017. Spatial heterogeneity of microtopography and its
Declaration of Competing Interest influence on the flow convergence of slopes under different rainfall patterns.
J. Hydrol. 545, 88–99.
The authors declare that they have no known competing financial Luo, J., Zheng, Z., Li, T., He, S., 2018. Assessing the impacts of microtopography on soil
erosion under simulated rainfall, using a multifractal approach. Hydrol. Process. 32
interests or personal relationships that could have appeared to influence (16), 2543–2556.
the work reported in this paper. Luo, J., Zheng, Z., Li, T., He, S., 2020. Temporal variations in runoff and sediment yield
associated with soil surface roughness under different rainfall patterns.
Geomorphology 349, 106915.
Data availability
Luo, J., Wang, N., Zheng, Z., Li, T., He, S., Tarolli, P., 2022. Tillage-induced
microtopography alters time-dependent intrinsic correlation of runoff and sediment
Data will be made available on request. yield. Soil Till. Res. 221, 105423.
Luo, J., Zheng, Z., Li, T., He, S., Tarolli, P., 2023. Impact of tillage-induced
microtopography on hydrological-sediment connectivity and its hydrodynamic
Acknowledgments understanding. Catena 228, 107168.
Masoodi, A., Majdzadeh Tabatabai, M.R., Noorzad, A., Samadi, A., 2017. Effects of soil
physico-chemical properties on stream bank erosion induced by seepage in
This study was funded by the CRSRI Open Research Program northeastern Iran. Hydrolog. Sci. J. 62 (16), 2597–2613.
(CKWV2021870/KY), the National Natural Science Foundation of China Mitas, L., Mitasova, H., 1998. Distributed soil erosion simulation for effective erosion
(42177316), and the Start-up funding from Inner Mongolia University prevention. Water Resour. Res. 34 (3), 505–516.
Mutchler, C.K., Young, R.A., 1975. Soil detachment by raindrops, present and
(21800-5223728).
prospective technology for predicting sediment yield and sources. USDA Annu. Rep.
5–40.
References Pijl, A., Reuter, L.E.H., Quarella, E., Vogel, T.A., Tarolli, P., 2020. GIS-based soil erosion
modelling under various steep-slope vineyard practices. Catena 193, 104604.
Planchon, O., Esteves, M., Silvera, N., Lapetite, J.M., 2002. Microrelief induced by
An, J., Zheng, F., Lu, J., Li, G., 2012. Investigating the role of raindrop impact on
tillage: measurement and modelling of surface storage capacity. Catena 46 (2–3),
hydrodynamic mechanism of soil erosion under simulated rainfall conditions. Soil
141–157.
Sci. 177 (8), 517–526.
Rezaei, A., Gurdak, J.J., 2020. Large-scale climate variability controls on climate,
Auerswald, K., Fiener, P., Dikau, R., 2009. Rates of sheet and rill erosion in Germany-A
vegetation coverage, lake and groundwater storage in the Lake Urmia watershed
meta-analysis. Geomorphology 111 (3–4), 182–193.
using SSA and wavelet analysis. Sci. Total Environ. 724, 138273.
Bahddou, S., Otten, W., Whalley, W.R., Shin, H.-C., El Gharous, M., Rickson, R.J., 2023.
Shen, H., Zheng, F., Wen, L., Han, Y., Hu, W., 2016. Impacts of rainfall intensity and
Changes in soil surface properties under simulated rainfall and the effect of surface
slope gradient on rill erosion processes at loessial hillslope. Soil Till. Res. 155,
roughness on runoff, infiltration and soil loss. Geoderma 431, 116341.
429–436.
Bauer, T., Strauss, P., Grims, M., Kamptner, E., Mansberger, R., Spiegel, H., 2015. Long-
Soil Survey Staff, 2010. Keys to Soil Taxonomy, 11th ed. USDA/NRCS, Washington, D.C.
term agricultural management effects on surface roughness and consolidation of
Sun, J., Zhang, N., Shi, M., Zhai, Y., Wu, F., 2021. The effects of tillage induced surface
soils. Soil Till. Res. 151, 28–38.
roughness, slope and discharge rate on soil detachment by concentrated flow: An
Beasley, D.B., Huggins, L.F., Monke, A., 1980. ANSWERS: A model for watershed
experimental study. Hydrol. Process. 35 (6), e14261.
planning. Trans. Asae 23 (4), 938–0944.
Tao, T., Han, Z., Li, Y., Gu, X., Chen, X., 2022. Effect of subsurface water flow depth on
Berger, C., Schulze, M., Rieke-Zapp, D., Schlunegger, F., 2010. Rill development and soil
the rill erosion process on purple soil slopes. Catena 214, 106297.
erosion: a laboratory study of slope and rainfall intensity. Earth Surf. Proc. Land. 35
Tian, S., Xu, M., Jiang, E., Wang, G., Hu, H., Liu, X., 2019. Temporal variations of runoff
(12), 1456–1467.
and sediment load in the upper Yellow River. China. J. Hydrol. 568, 46–56.
da Rocha Junior, P.R., Bhattarai, R., Alves Fernandes, R.B., Kalita, P.K., Vaz Andrade, F.,
Vandamme, J., Zou, Q., 2013. Investigation of slope instability induced by seepage and
2016. Soil surface roughness under tillage practices and its consequences for water
erosion by a particle method. Comput. Geotech. 48, 9–20.
and sediment losses. J. Soil Sci. Plant Nut. 16 (4), 1065–1074.
Wirtz, S., Seeger, M., Ries, J.B., 2012. Field experiments for understanding and
Darboux, F., Davy, P., Gascuel-Odoux, C., Huang, C., 2002. Evolution of soil surface
quantification of rill erosion processes. Catena 91, 21–34.
roughness and flowpath connectivity in overland flow experiments. Catena 46 (2–3),
Xu, X., Zheng, F., Wilson, G.V., Wu, M., 2017. Upslope inflow, hillslope gradient and
125–139.
rainfall intensity impacts on ephemeral gully erosion. Land Degrad. Dev. 28 (8),
De Jong, S.M., Paracchini, M.L., Bertolo, F., Folving, S., Megier, J., De Roo, A.P.J., 1999.
2623–2635.
Regional assessment of soil erosion using the distributed model SEMMED and
Zhang, P., Tang, H., Yao, W., Zhang, N., Xizhi, L.V., 2016. Experimental investigation of
remotely sensed data. Catena 37 (3–4), 291–308.
morphological characteristics of rill evolution on loess slope. Catena 137, 536–544.
De Roo, A.P.J., Wesseling, C.G., Cremers, N.H.D.T., Offermans, R.J.E., Ritsema, C.J., Van
Zhang, G.H., Wang, L.L., Li, G., Tang, K.M., Cao, Y., 2011. Relationship between
Oostindie, K., 1994. LISEM: a new physically-based hydrological and soil erosion
sediment particle size and transport coefficient on steep slopes. Trans. Asabe. 54 (3),
model in a GIS-environment, theory and implementation. IAHS Publications-Series
869–874.
of Proceedings and Reports-Intern Assoc Hydrological Sciences 224, 439–448.
Zhang, H., Zhang, H., Dong, Y., Zhang, Q., 2017. Characterization of runoff and sediment
Estabragh, A.R., Soltani, A., Javadi, A.A., 2016. Models for predicting the seepage
yield in farmlands on loess slopes based on R/S and wavelet analysis. Acta
velocity and seepage force in a fiber reinforced silty soil. Comput. Geotech. 75,
Pedologica Sinica 54 (6), 1345–1356.
174–181.
Zhang, X., Zhou, Y., Dong, Y., Zhang, H., Zhang, C., Zhang, Q., 2019. Measurement of the
Fernandes, J., Bateira, C., Soares, L., Faria, A., Oliveira, A., Hermenegildo, C., Moura, R.,
connectivity of runoff source areas on a bare soil surface in a laboratory experiment.
Gonçalves, J., 2017. SIMWE model application on susceptibility analysis to bank
J. Soil Water Conserv. 74 (5), 513–519.
gully erosion in Alto Douro Wine Region agricultural terraces. Catena 153, 39–49.
Zhao, L., Huang, C., Wu, F., 2016. Effect of microrelief on water erosion and their
Flanagan, D.C., Nearing, M.A., 1995. USDA-Water Erosion Prediction Project: Hillslope
changes during rainfall. Earth Surf. Proc. Land. 41 (5), 579–586.
profile and watershed model documentation. Nserl. Rep. 10, 1–123.
Zhao, L., Hou, R., Wu, F., 2019. Rainwater harvesting capacity of soils subjected to
Gómez, J.A., Nearing, M.A., 2005. Runoff and sediment losses from rough and smooth
reservoir tillage during rainfall on the Loess Plateau of China. Agr. Water Manage.
soil surfaces in a laboratory experiment. Catena 59 (3), 253–266.
217, 193–200.
Govers, G., Giménez, R., Van Oost, K., 2007. Rill erosion: exploring the relationship
Zhao, L., Fang, Q., Hou, R., Wu, F., 2021. Effect of rainfall intensity and duration on soil
between experiments, modelling and field observations. Earth-Sci. Rev. 84 (3–4),
erosion on slopes with different microrelief patterns. Geoderma 396, 115085.
87–102.
Zheng, Z.C., He, S.Q., Wu, F.Q., 2014. Changes of soil surface roughness under water
Jiang, F., Zhan, Z., Chen, J., Lin, J., Wang, M.K., Ge, H., Huang, Y., 2018. Rill erosion
erosion process. Hydrol. Process. 28 (12), 3919–3929.
processes on a steep colluvial deposit slope under heavy rainfall in flume
Zheng, H., Nie, X., Liu, Z., Mo, M., Song, Y., 2021. Identifying optimal ridge practices
experiments with artificial rain. Catena 169, 46–58.
under different rainfall types on runoff and soil loss from sloping farmland in a
Kimaro, D.N., Poesen, J., Msanya, B.M., Deckers, J.A., 2008. Magnitude of soil erosion on
humid subtropical region of Southern China. Agr. Water Manage. 255, 107043.
the northern slope of the Uluguru Mountains, Tanzania: Interrill and rill erosion.
Catena 75 (1), 38–44.
Labat, D., 2005. Recent advances in wavelet analyses: Part 1. A review of concepts. J.
Hydrol. 314 (1–4), 275–288.
Le Bissonnais, Y., Cerdan, O., Lecomte, V., Benkhadra, H., Souchère, V., Martin, P., 2005.
Variability of soil surface characteristics influencing runoff and interrill erosion.
Catena 62 (2–3), 111–124.

10
J. Luo et al. Geoderma 436 (2023) 116539

Zhong, S., Han, Z., Du, J., Ci, E., Ni, J., Xie, D., Wei, C., 2019. Relationships between the Ziadat, F.M., Taimeh, A.Y., 2013. Effect of rainfall intensity, slope, land use and
lithology of purple rocks and the pedogenesis of purple soils in the Sichuan Basin. antecedent soil moisture on soil erosion in an arid environment. Land Degrad. Dev.
China. Sci. Rep. 9 (1), 13272. 24 (6), 582–590.
Zobeck, T.M., Popham, T.W., 1997. Modification of the wind erosion roughness index by
rainfall. Soil Till. Res. 42 (1–2), 47–61.

11