Landslides

, Volume 13, Issue 5, pp 1231–1242 | Cite as

Mitigation planning based on the prediction of river blocking by a typical large-scale debris flow in the Wenchuan earthquake area

  • Jinfeng Liu
  • Yong You
  • Xiaoqing Chen
  • Xingzhang Chen
Original Paper

Abstract

Due to scale amplification resulting from the blocking-bursting process associated with landslide dams in debris flow catchments, many of large-scale debris flows that occurred as a result of the Wenchuan earthquake blocked the main rivers and resulted in catastrophic dam-breaking floods. To decrease the damage caused by dam-breaking floods, mitigation works should be applied in debris flow gullies. Currently, few studies focus on how to determine the key parameters concerning the scale of debris flows (i.e., peak discharge and total volume) that need to be controlled for mitigation planning considering downstream objects affected by the flow. The Xiaojia debris flow is described and analyzed as a typical large-scale case from the Yingxiu area. A back-calculation method and numerical simulation were proposed for mitigation planning in Xiaojia gully based on predictions of river blocking. First, the maximum peak flood discharge that did not endanger the town of Yingxiu was calculated. Then, the permissible blockage from a debris flow was determined based on the back-calculation of a dam-breaking flood. Finally, the scale of the largest permissible debris flow was obtained for use in mitigation plans based on numerical simulations. The calculations showed that the peak discharge of a flood that would not endanger Yingxiu should be <1496.48 m3/s. Accordingly, based on the 1/3 and 1/2 breach modes, the permissible blocking height of a debris flow barrier dam should not exceed 43.09 and 31.64 m, respectively. The total volume and peak discharge of a single debris flow event should be controlled to not exceed 70.59 × 104 m3 and 784.59 m3/s for the 1/3 breach mode and 45.21 × 104 m3 and 551.77 m3/s for the 1/2 breach mode. Based on these determinations of the key debris flow parameters, the simulation results indicate that debris flow damage can be decreased to an acceptable level to ensure the safety of Yingxiu downstream by the implementation of two check dams in the downstream channel and a deposition works on the debris fan.

Keywords

Debris flow Disaster characteristics Wenchuan earthquake River-blocking prediction Mitigation planning 

Introduction

In the Wenchuan earthquake area, during the rainy seasons from 2008 to 2013, the frequency and scale of debris flows greatly increased due to the huge increase in loose materials from the earthquake and extreme weather conditions. On 7–9 July 2013, a heavy rainstorm with maximum daily rainfall of 751.4 mm (Xingfu Village, Dujiangyan, the highest recorded in the past 60 years) swept Wenchuan County and caused significant damage (Liu et al. 2014). Many large-scale debris flows occurred in Wenchuan County, destroying the Dujiangyan-Wenchuan highway (newly constructed and in service for only 7 months), the old National Road G213, and many other rural roads. Among them, the debris flow that occurred on 11 July 2013 in the Qipan gully was the largest in scale and resulted in the most catastrophic damage in this area. Eight people were confirmed dead, and six were listed as missing. In all, 285 buildings, 4 km of drainage channel, three transformer substations, and the manufacturing complexes of seven companies downstream from the gully were completely destroyed by the debris flow.

Due to the scale of amplification resulting from the blocking-bursting process of landslide dams in debris flow catchments (Cui et al. 2011; Huang et al. 2012; Cui et al. 2013a), many of the large-scale debris flows in the Wenchuan earthquake area that blocked main river channels resulted in severe catastrophes (Tang et al. 2009; Huang and Li 2009; Tang et al. 2011; Chen et al. 2012). For example, on 14 August 2010, debris flows in the Hongchun and Shaofang gullies in Wenchuan County blocked the Min River and produced barrier dams that changed the river course and resulted in flooding of the newly reconstructed town of Yingxiu (Tang et al. 2011; Cui et al. 2013b). When a debris flow barrier dam breaks up, the dam-breaking flood can cause devastating effects on downstream facilities, lives, and property (Chang et al. 2011). To decrease the damage of dam-breaking floods, mitigation works can be applied in debris flow gullies. To undertake such debris flow mitigation planning, the downstream objects affected by the dam-breaking flood should be considered in particular. However, few studies have focused on how to determine the key parameters controlling the scale of debris flows (i.e., peak discharge and total volume) that need to be regulated to undertake mitigation planning that considers downstream objects affected by such flows. In this paper, the Xiaojia debris flow is described and analyzed as a representative large-scale case from the Yingxiu area. A back-calculation method and numerical simulation are proposed to assist with mitigation planning for the Xiaojia gully based on a river-blocking prediction.

Background

Formation conditions

Xiaojia gully is located in Yingxiu Town, Wenchuan County, Aba Prefecture, Sichuan Province, China. The coordinates of the gully mouth are 103° 26′ 20″ E, 31° 04′ 32″ N. The gully is a first-grade left branch of the Yuzixi River, an arterial branch of the upper Minjiang River (Fig. 1). The catchment covers an area of 7.19 km2, and the main channel has a length of 4.50 km and a longitudinal slope of 48.5 %. The height difference is 2043 m, with a maximum elevation of 3483 m and a minimum elevation of 1080 m (Fig. 1). The S303 provincial road which is a vital traffic line for connecting the towns of Yingxiu and Wolong passes through the alluvial fan of the gully.
Fig. 1

The location and geology geological conditions of the study area

The study area is located on the eastern edge of the Qinghai–Tibetan Plateau. The terrain is complex and characterized by considerable elevation differences and steep slopes (Yin et al. 2009). High mountains and deep-cut rivers are the main topographic characteristics of the area (Zhang et al. 2006). The geological conditions of the study area are very complex. The gully is very near the deep fracture zone of the Yingxiu–Beichuan reverse fault (part of the Longmen mountain fault zone), running southwest to northeast (Fig. 1). The Ms 8.0 “5.12” Wenchuan earthquake that occurred in this area in 2008 was due to movement on the Yingxiu–Beichuan reverse fault (Wang et al. 2009). The Wenchuan earthquake caused widespread earth surface destruction across the whole catchment of Xiaojia gully and formed a large number of landslips and landslides, providing rich source of loose solid materials for debris flows (Fig. 1). Based on the interpretation of 0.5-m resolution aerial photography from 7 January 2011 (Fig. 1), the total volume of loose solid material available for the formation of debris flows in the catchment was 532.09 × 104 m3 distributed over 0.99 km2 of the mountain surface as the result of earthquake destruction (Liu et al. 2014).

The strata exposed in the gully include diorite (δ23) in the upper reaches, plagioclase granite (γo24) in the middle reaches, and granodiorite (γδ23) in the lower reaches (Fig. 1). The rock mass of the study area is part of a suite of magmatic rocks formed in the Jinning–Chengjing period with structural characteristics including high strength, jointed fissures, and strongly weathered layers at the surface. Consequently, it is prone to instability, collapse, and the formation of landslides.

The study area lies within the subtropical humid monsoon climatic zone of the Sichuan Basin border, and it is in the rainstorm center of the western Sichuan region. According to data from the Yingxiu weather station (Fig. 1), the average annual rainfall is 1253.1 mm, and the maximum and minimum annual recorded rainfall is 1688 and 836.7 mm, occurring in 1964 and 1974, respectively. The average number of annual precipitation days is up to 202.7 days, with a maximum daily rainfall of 269.8 mm. The precipitation over the year is distributed unevenly, concentrated mainly from June to September, which accounts for 60 to 70 % of the annual rainfall (Liu et al. 2010). This period also usually has a high occurrence of collapses, landslides, debris flows, and other geological disasters.

Debris flow events and damage

The landforms, geology, and rainfall conditions in the study area strongly favored the formation of debris flows following the Wenchuan earthquake. Following the Wenchuan earthquake, the Xiaojia gully exhibited several large-scale debris flow torrents. On 14–15 August 2010, the area around Yingxiu was affected by a widespread rainstorm with total precipitation of 212.5 mm in 40 h and a maximum rainfall intensity of 22.1 mm/h (Tang et al. 2011). Many debris flows were triggered in the region, with the largest being a debris flow that occurred in Xiaojia gully. In this event, debris flow materials were transported to the alluvial area at the gully mouth where a large debris fan was formed (Figs. 2a, b and 3b). The debris flow reached the Yuzixi River and blocked it to form a 31-m-high barrier dam (Figs. 2a and 4). Consequently, a barrier lake 1.0 km in length was formed that submerged the upstream S303 road. At the same time, the downstream Nanhua tunnel (Fig. 1), which was constructed after the Wenchuan earthquake (Fig. 2c), was completely buried by debris flow deposits from the 14 August 2010 event (Fig. 2d). The tunnel was seen as one of the key projects of the reconstructed S303 road which was aided by the Hong Kong government after the Wenchuan earthquake. On 6 September 2010, another debris flow occurred and buried the S303 road again. The two debris flows caused traffic disruption to the S303 road for about 1 month in total. Then, during 3–4 July 2011, continuous heavy rainfall in the area triggered another debris flow in Xiaojia gully. About 30,000 m3 of debris flow materials were transported to the gully mouth, resulting in disruption to the S303 road once again. Furthermore, on 17 August 2012, a small-scale debris flow in the gully generated about 3000 m3 of solid materials and buried about 200 m of the S303 road.
Fig. 2

Debris flow deposition blocked the Yuzixi River and formed a barrier lake in the 14 August 2010 debris flow event (a and b); the Nanhua tunnel constructed after the Wenchuan earthquake (c) was buried by the debris flow deposition of the 14 August 2010 debris flow event (d)

Fig. 3

The spatial relation between the debris flow barrier dam of Xiaojia gully and the Yingxiu Town (a); the measurements of the debris flow fan, the permissible peak discharges (Qp and Qbp) of the dam-breaking flood (b) and the permissible peak debris flow discharge (Qcp) (c)

Fig. 4

The longitudinal section (sections c–d in Fig. 3) of debris fan and barrier dam formed in the 14 August 2010 debris flow event

The debris flows in the Xiaojia gully cause two kinds of damage. One is the direct burial of the S303 road. Currently, the S303 road passes through the middle part of the debris fan and is very close to the gully mouth (within about 80 m, measured in Fig. 3b). If debris flows occur again, the road on the alluvial fan can easily be buried, causing disruption to traffic. The other type of damage is a secondary disaster caused by blocking the river with the debris flow. If the Yuzixi River is blocked by a debris flow, a barrier lake can form. The backed-up water can then submerge the road upstream of the barrier dam. However, most critically, a sudden dam-breaking flood can cause damage to the newly constructed town of Yingxiu, which is just 5.4 km downstream of the gully (Fig. 3a). The Yingxiu Town is an important traffic hub in Aba Prefecture, lying at the intersection of provincial road S303, national road G213, and the Dujiangyan to Wenchuan highway (Fig. 1). During the Wenchuan earthquake, the town was completely destroyed and 6566 lives were lost. Following the earthquake, the town was relocated and rebuilt as a new home to about 5700 local residents (Liu et al. 2014). As shown in Fig. 5, the Yuzixi River passes through the middle of the reconstructed town. In this section, protective embankments on both banks of the Yuzixi River were constructed to protect Yingxiu from flood damage. Here, the width of the Yuzixi River is 45 m, and the height of the protection embankment is about 4.5 m (Fig. 5). According to the spatial positioning of the Xiaojia gully relative to the town (Fig. 3a), a dam-breaking flood can still reach Yingxiu if a debris flow in Xiaojia gully blocks the Yuzixi River. If the flood depth is greater than the height of protection embankments (4.5 m), a dam-breaking flood can overflow the banks of the Yuzixi River and inundate the Yingxiu Town.
Fig. 5

Relative locations of the Minjiang and Yuzixi rivers, the reconstructed town of Yingxiu, and the protective embankments

Method

Data collection and preparation

Field investigations and topography measurements were carried out in the Xiaojia gully to obtain basic debris flow data. A 1:50,000 topographic map and a 25-m digital elevation model (DEM), provided by Sichuan Center of Basic Geographic Information, were used to analyze the topographic features of the debris flows. Aerial photography with a scale of 1:10,000 was used to clarify the spatial relationships between the debris flow depositional fan and the main river (Fig. 6). Field investigations and measurements were made to map the debris flow gullies on a 1:50,000 scale topographic map and then digitize them into a geographic information system (GIS). Topography map (1:1000) of the depositional area and 1:200 cross section of the main channel were measured for the numerical simulation of the debris flow that could then be used in mitigation planning.
Fig. 6

The spatial relation between the debris flow deposition fan and the main river, the mitigation planning in the Xiaojia gully

River-blocking prediction method

This section introduces a back-calculation method to determine the permissible debris flow scale (peak discharge and total volume of a single debris flow event) in the Xiaojia gully that ensures the safety of the downstream Yingxiu Town. First, to protect the Yingxiu Town, the depth of the flood passing through the town should be <4.5 m (lower than the protection embankments, Fig. 5). Here, the average velocity (vP, Fig. 3c) and peak discharge (QP, Fig. 3c) of the permissible flood can be calculated using the following equation (You et al. 2012a, b):
$$ {v}_P=\frac{1}{n_P}{J_P}^{1/2}{R_P}^{2/3}, $$
(1)
$$ {R}_P=\frac{S_P}{P_P}=\frac{H_P{B}_P}{B_P+2{H}_P}, $$
(2)
$$ {Q}_P={S}_P{v}_P={H}_P{B}_P{v}_P, $$
(3)
where nP, the roughness coefficient of the Yuzixi River, equals 0.04 from the common assignment table (Zhang 2008); JP is the slope of the hydraulic grade line of the Yuzixi River near Yingxiu and measures 1.50 % in the field; RP is the hydraulic radius of the overflowing cross section; BP, the width of the Yuzixi River, equals 45 m (Fig. 5); HP is the permissible flood depth that is the depth of the protection embankments of the Yuzixi River and equals 4.5 m (Fig. 5). As shown in Figs. 3a and 5, at the confluence of the two rivers, the water depth of flood in the Minjiang River can be the impact of raising the water level of Yuzixi River. The maximum flood depth in Minjiang River is about 4.20 m (occurred in 1985, historical observation data for the Yingxiu hydrometric station). Combined with the slope of the Yuzixi River (JP = 1.50 %), the backward water length to the Yuzixi River under the impact of the flood in the Minjiang River is about 280 m and much smaller than the length of the protection embankment (about 1260 m, Fig. 3a). Therefore, the impact of raising the water level of the Yuzixi River by the flood in the Minjiang River was ignored in the calculation. SP is the flood overflowing area. As shown in Fig. 5, the artificial embankments were designed as perpendicular training walls on both banks of the Yuzixi River. Therefore, the flowing cross section is assumed as rectangular shape, and SP equals the product of HP and BP. PP is the wetted perimeter.
Focused on the Xiaojia gully, flood damage is the result of a dam-breaking flood that occurs when the Yuzixi River is blocked by a debris flow. After blocking the Yuzixi River, a debris flow barrier dam can be formed. Usually, the debris flow barrier dam is regarded as an earth-type dam that breaks step by step. Historical data from past events show that a partial breach (1/3 and 1/2 of the dam height) of a debris flow barrier dam occurs most often, whereas an entire breach is almost impossible (Zhuang et al. 2010; Du et al. 2014). In the case of Xiaojia gully, the 1/3 and 1/2 breach modes are used for the following calculations and analysis. Based on the calculation of the maximum permissible peak discharge of a flood through Yingxiu, the peak discharge from the dam-breaking flood at the dam site (Qbp, Fig. 3b) can be calculated as (Liu et al. 2009; You et al. 2012b)
$$ {Q}_{bp}=\frac{Q_PW}{W-\frac{Q_PL}{v_mk}}, $$
(4)
$$ W=\frac{{h_P}^2{B}_m}{2{J}_R}. $$
(5)
The permissible blocking height (hP, Figs. 4 and 7) of the debris flow barrier dam can then be calculated (Walder and O’Connor 1997):
$$ {Q}_{bp}=0.9\left(\frac{h_P-h}{h_P-0.827}\right){B}_m\sqrt{h_P}\left({h}_P-h\right), $$
(6)
where L is the distance between the debris flow dam site and the downstream town of Yingxiu, which is measured to be roughly 5.4 km (Fig. 3a, the distance between a and b along the Yuzixi River); W is the water volume of the barrier lake, which can be calculated by using the topography map (Fig. 7); vm is the average velocity of the Yuzixi River in the rainy season which equals 8 m/s based on historical data; k is an empirical coefficient (for a mountain stream, k = 1.5) (You et al. 2012b). JR is the longitudinal slope of the Yuzixi River at the outlet of Xiaojia gully and measured as 2.50 % in the field (Fig. 7); h is the residual dam height after the breach (Figs. 4 and 7), which equals hP/2 for 1/2 breach mode and 2hP/3 for 1/3 breach mode; Bm is the average width of the Yuzixi River at the gully mouth, which was measured as 70 m in field investigation (Fig. 4).
Fig. 7

The cross section (sections e–f in Fig. 3) of debris barrier dam formed in the 14 August 2010 debris flow event

Based on the determined maximum permissible blocking height (hP), the debris flow volume participating in river blocking (VP, Fig. 4) can be calculated by (Zhou et al. 1991):
$$ {V}_P=\left(\frac{1}{2tg\theta }+\frac{1}{2tg\delta}\right){B}_m{h_P}^2, $$
(7)
where θ is the initiating slope for forming water-stony debris flow and usually determined as 14°; δ is the internal friction of the debris flow material at the saturation state and usually determined as 25°.
After the 14 August 2010 large-scale debris flow event, a debris fan was formed at the gully mouth (Fig. 3b). For the next event, part of the debris flow material accumulated initially on the debris fan; the remainder is possible to reach the main river and participate in river blocking (Keefer 1999; Chen et al. 2012). The debris flow materials (volume of Vdp) flowing out of the gully mouth can be divided into two parts: fan deposition (Vf, Fig. 4) and river-blocking deposition (VP, Fig. 4). For a debris flow event in Xiaojia gully, the maximum deposition length (Ld, Figs. 3b and 4), maximum deposition width (Bd, Figs. 3b and 7), and average deposition depth (Hd, Fig. 4) can be calculated by the following formulae, based on field measurements at 23 debris flow gullies distributed along the Yuzixi River (Xiang et al. 2012):
$$ {H}_d=0.141{V_{dp}}^{1/3}+4.21, $$
(8)
$$ {L}_d=2.446{V_{dp}}^{1/3}+0.7487, $$
(9)
$$ {B}_d=2.748{V_{dp}}^{1/3}+0.3483. $$
(10)
As shown in Fig. 4, the line a-b is the boundary delimiting Vf and VP. The distance between the gully mouth and the left bank of Yuzixi River (Lf, Fig. 4) was measured as 150 m in the field. If Ld ≤ Lf (Fig. 4), it shows that the debris flow material is unable to reach the main river and accumulates only on the debris fan. Consequently, no blockage to the Yuzixi River is formed in this situation. Conversely, a debris flow can reach the main river and possibly block it if Ld > Lf (Fig. 4). Based on the calculated result above for the debris flow volume participating in river blockage (VP, Fig. 4), the total permissible volume of a single debris flow in Xiaojia gully (Vdp) under the condition Ld > Lf can be calculated (Liu et al. 2014):
$$ {V}_{dp}={V}_f+{V}_P=\frac{L_f{B}_d{H}_d}{2}+{V}_P. $$
(11)
Finally, the permissible peak discharge (Qcp, Fig. 3b) of debris flow in Xiaojia gully that ensures the safety of downstream Yingxiu can be calculated (Ou 1992; Xiang et al. 2012):
$$ {Q}_{cp}=0.0188{V_{dp}}^{0.79}. $$
(12)
If a debris flow occurs again in Xiaojia gully, its peak discharge (Qc) can be estimated assuming the same occurrence frequencies for the rainstorms, floods, and debris flows (Zhou et al. 1991). Here, three frequencies (1, 2, and 5 %) of debris flow were considered for the calculation. The possible debris flow peak discharge (Qc) under different return periods can be calculated (Liu et al. 2013, 2014):
$$ {Q}_B=0.278\phi \frac{S}{\tau^n}F, $$
(13)
$$ {Q}_c=\left(1+\frac{\gamma_c-{\gamma}_m}{\gamma_s-{\gamma}_c}\right){Q}_B{D}_U, $$
(14)
where γm is the density of water, usually determined to be 1.00 t/m3; γs is the density of the solid material, usually set as 2.65 t/m3; γc is the density of the debris flow, usually determined by a spot investigation method (the “Specification of geological investigation for debris flow stabilization,” the Chinese geological mineral industry standard, DZ/T0220-2006). DU is the blockage coefficient in the debris flow gully; its range is between 0 and 3 and determined based on actual field conditions. QB is the peak flood discharge in Xiaojia gully. F is the watershed area, equal to 7.19 km2; φ is the runoff coefficient for the flood peak; S is the rainfall intensity; τ is the runoff confluence time of the rainstorm; n is the attenuation index of the rainstorm. φ, S, τ, and n can be determined based on the “The Rainstorm and Flood Calculation Manual of Medium and Small Basins in Sichuan Province” (published in 2010, rainfall data from 1978 to 2004) (Liu et al. 2013). The calculating results of related parameters for the possible debris flow peak discharge under different return periods are listed in Table 1.
Table 1

The calculating results of related parameters for the possible debris flow peak discharge under different return periods

Parameters

Unit

Frequency

1 %

2 %

5 %

n

0.53

0.55

0.57

S

mm

79.36

70.68

58.90

τ

h

1.14

1.18

1.25

φ

0.91

0.91

0.89

QB

m3/s

135.51

117.03

92.42

DU

2.5

2.3

2.1

γc

t/m3

2.0

2.0

2.0

Qc

m3/s

790.50

628.04

452.86

Above all, the calculating procedure for the permissible and possible debris flow peak discharge is shown as Fig. 8. If the possible peak debris flow discharges (Qc) for different return periods are larger than the permissible peak discharge (Qcp), then mitigation measures should be applied (Fig. 9) to decrease Qc to an acceptable level for the safety of the downstream town.
Fig. 8

The calculating procedure for the permissible and possible debris flow peak discharge

Fig. 9

The interface of the Kanano simulator and the related settings

Numerical simulation method for mitigation planning

Based on the above analysis, a numerical simulation method was used to implement mitigation planning for Xiaojia gully for instances when the peak debris flow discharge (Qc) of different return periods is larger than the permissible peak discharge (Qcp). Here, a Kanako simulator was used for the numerical simulation. The simulator is a kind of numerical simulation tool, equipped with an efficient graphical user interface (GUI), that can be used to constrain the most efficient mitigation measures used in construction planning and implementation visualization (Nakatani et al. 2007, 2008). The basic equations of debris flow, for momentum, continuation, riverbed deformation, erosion/deposition, and riverbed shearing stress, are based on previous research (Takahashi and Nakagawa 1991, Takahashi et al. 2001). The effect of check dams constructed in debris flow channels can be simulated based on a model developed by Satofuka and Mizuyama (2005 and 2006). Before simulation, input parameters concerning the following need to be set: debris flow dynamic parameters, channel topographic conditions, and mitigation words (Table 2). Some of the parameters (such as the mass and density of bed material, mass and density of the fluid phase, concentration of the movable bed, coefficient of erosion rate and coefficient of accumulation rate) are empirical values used in most debris flow cases (Satofuka and Mizuyama 2005). The simulation channel was set as the downstream channel, where two check dams were planned (Fig. 6, section a–b, and Fig. 9). The input data of the channel (channel elevation and width) were obtained from the measured 1:1000 topographic map. The total channel length for the simulation was 870 m. The channel width varied from 7.38 to 53.1 m (Fig. 9) based on the measured topographic data. The movable bed layer was set at 10 m based on field measurements. Six observation points (obs. 1–6, Fig. 9) were set for recording the simulation results of the change in debris flow peak discharge.
Table 2

Input parameters for the Kanako simulator

Parameters/variables

Value

Unit

Simulation continuance time

3600

s

Time interval of calculation

0.01

s

Mass density of bed material

2650

kg/m3

Mass density of fluid phase (water and mud, silt)

1100

kg/m3

Concentration of movable bed

0.65

 

Coefficient of erosion rate

0.0007

 

Coefficient of accumulation rate

0.02

 

Manning’s roughness coefficient

0.03

 

Number of calculation points in channel

87

 

Interval of calculation points in channel

10

m

Results and discussion

River-blocking prediction

Table 3 lists the calculated results of related parameters for the river-blocking prediction. As analyzed above, the height of the protective embankment on the Yuzixi River that passes through the middle of the newly constructed town of Yingxiu is only 4.5 m (Fig. 5). Therefore, if the flood depth in the Yuzixi River rises above 4.5 m, Yingxiu could suffer flood inundation. The calculations show that the peak discharge of a flood should be less than 1496.48 m3/s to maintain safe levels through Yingxiu (Table 3). Based on the downstream evolution equation for a dam-breaking flood (Eqs. 4 and 5) and the calculation of the peak discharge of a dam-breaking flood at the dam site (Eq. 6), the degree of river blockage by a debris flow in the Xiaojia gully can be determined (Fig. 8). The calculations show that the permissible peak discharge of the dam-breaking flood at the dam site should be less than 2018.73 and 2878.31 m3/s under the 1/3 and 1/2 breach modes, respectively (Table 3). Thus, the permissible blocking height of the debris flow barrier dam should be less than 43.09 and 31.64 m under the 1/3 and 1/2 breach modes, respectively (Table 3).
Table 3

The calculating results of related parameters for the river-blocking prediction

Calculation content

Parameters

Unit

Breach mode

1/3 Mode

1/2 Mode

The permissible flood at Yingxiu Town

SP

m2

202.50

202.50

PP

m

54.00

54.00

RP

m

3.75

3.75

vP

m/s

7.39

7.39

QP

m3/s

1496.48

1496.48

The permissible blockage by debris flow

W

104 m3

259.94

140.15

Qbp

m3/s

2018.73

2878.31

HPhP

m

43.09

31.64

The permissible debris flow scale in Xiaojia gully

VP

104 m3

40.03

21.58

Hd

m

16.68

14.95

Ld

m

218.53

188.47

Bd

m

244.31

210.64

Vf

104 m3

30.55

23.62

Vdp

104 m3

70.59

45.21

Qcp

m3/s

784.59

551.77

Further calculations show that the permissible total debris flow volume of a single debris flow event in Xiaojia gully should be smaller 70.59 × 10 and 45.21 × 104 m3 under the 1/3 and 1/2 breach modes, respectively (Table 3). As a result, the peak debris flow discharge should be controlled to be smaller than 784.59 and 551.77 m3/s under the 1/3 and 1/2 breach modes, respectively, to maintain the safety of downstream Yingxiu (Table 3).

As shown in Table 3, the possible peak debris flow discharges in Xiaojia gully were calculated as 790.50, 628.04, and 452.86 m3/s under 1, 2, and 5 % return periods, respectively. By comparing the permissible peak debris flow discharge (Qcp) and the possible peak debris flow discharge (Qc), the necessity of applying mitigation measures can be evaluated (Fig. 8). For the 1/3 break mode, the permissible peak debris flow discharge (784.59 m3/s) is similar to the possible peak debris flow discharge (790.50 m3/s) in the 1 % return period (Table 3). Therefore, for a single debris flow event in Xiaojia gully, it is necessary to take mitigation measures when the debris flow frequency is larger than the 100-year return period (P = 1 %). However, for the 1/3 break mode, the permissible peak debris flow discharge (551.77 m3/s) is between the 20-year (P = 5 %) and 50-year (P = 2 %) return periods for the possible peak debris flow discharge. So, mitigation measures should be applied to decrease the peak discharge when the debris flow frequency is greater than the 50-year return period (P = 2 %). Ultimately, the peak discharge of a single debris flow event with a 100-year return period (P = 1 %) was used for mitigation planning in Xiaojia gully.

Mitigation planning in Xiaojia gully

After the determination of the debris flow peak discharge (Qc = 790.50 m3/s, P = 1 %) that needs to be controlled for a single debris flow in Xiaojia gully, a numerical simulation was carried out as part of the mitigation planning. Usually, the duration of a single debris flow event is short due to the sudden increase–decrease characteristics of the discharge (Kang et al. 2004). In the simulation, the supplied hydrograph of the debris flow in Xiaojia gully was generalized as a pentagonal type (Wu et al. 1993; Liu et al. 2013), that lasted for a period of 1 h based on the 2010 debris flow event. Figure 10a shows the hydrograph used for the simulations.
Fig. 10

The simulation results for a debris flow discharge hydrograph upstream and downstream of the check dams (a); the debris flow peak discharge is compared at various observation points above and below the check dams. The simulation results of debris flow discharge hydrograph on the upstream and downstream of the check dams (a), the debris flow peak discharge comparison at different observation points before and after the check dams

Based on the field investigation combined with the topographic conditions of the channel in the downstream part of the Xiaojia gully, two check dams (closed type, Liu et al. 2013) with effective heights of 18 m (check dam #1, Fig. 6) and 12 m (check dam #2, Fig. 6) were proposed. The effectiveness of controlling the debris flow peak discharge with the two check dams was simulated using the Kanako simulator. The simulation results are shown in Fig. 10. When the debris flow passes through check dam #1 (Figs. 6 and 9), the peak discharge upstream of the dam is 749.4 m3/s (Fig. 10a, obs. 3) and 619.3 m3/s downstream from the dam (Fig. 10a, obs. 4). The debris flow peak discharge reduction rate is 17.36 %. Similarly, for check dam #2, the debris flow peak discharge decreases from 619.3 m3/s upstream of the dam (Fig. 10a, obs. 4) to 504.3 m3/s downstream from the dam (Fig. 10a, obs. 5). The peak debris flow discharge reduction rate of check dam #2 is 18.57 %. Figure 10b shows the peak debris flow discharge comparison before and after the implementation of the check dams. The results indicate a fluctuation in the peak debris flow discharge from upstream to downstream, which decreases from obs. 1 to obs. 5 and increases from obs. 5 to obs. 6 (Fig. 10b). The reason for this is the change in the channel width from upstream to downstream. As shown in Fig. 10b, the debris flow peak discharge has an opposite change tendency with the channel width. In the section of obs. 4 and obs. 5 (Fig. 9), the broad channel, with an average width of 40 m forces the debris flow velocity to drop, and thus, deposition occurs to decrease the peak discharge. For the same reason, the debris flow peak discharge at obs. 6 (channel width of 40.8 m, Fig. 10b) is larger than the peak discharge at obs. 5 (channel width of 53.1 m, Fig. 10b). After comparing simulation results for the gully mouth (Fig. 10b), peak debris flow discharge decreases from 727.3 m3/s before mitigation to 531.2 m3/s after mitigation. The mitigation effect is remarkable, with a reduction rate of 26.96 % as a result of the two check dams in the downstream part of the Xiaojia gully.

After the implementation of the two check dams in the downstream channel of the Xiaojia gully, the peak discharge of a single debris flow event was decreased to 531.2 m3/s, which is smaller than the permissible peak discharge of the 1/3 and 1/2 breach mode (Table 3). Therefore, for the mitigation of a single debris flow with a frequency of P ≤ 1 %, the implementation of two check dams is an effective way to maintain the safety of the downstream town of Yingxiu. However, the debris flow occurrence frequency was very high in Xiaojia gully following the 5.12 Wenchuan earthquake. Debris flows of different scales occurred almost annually after the Wenchuan earthquake. The storage capacity of the two check dams could be quickly used up due to the high frequency of debris flows. Accordingly, the effectiveness of the full check dams is limited for decreasing the debris flow peak discharge to the permissible level for maintaining the safety of the Yingxiu downstream. Thus, additional mitigation works were needed to ensure the safety of human life and property downstream. Based on field investigations of the downstream channel condition in the Xiaojia gully, no appropriate sites could be used to implement an additional check dam. Therefore, a deposition work on the debris fan was recommended to trap debris flow sediments (Fig. 6). According to field topographic conditions and the debris flow peak discharge to be controlled (Table 3), the plain layout of the deposition work was designed (Fig. 6). The retaining wall of the deposition work was designed to be 12-m high, which can provide a 4.6 × 104 m2 depositional area for trapping 56.6 × 104 m3 of debris flow sediments.

Conclusions

Following the Wenchuan earthquake, Xiaojia gully exhibited increases in the frequency and scale of debris flow torrents due to the huge increase of loose solid materials in the catchment. The debris flows in Xiaojia gully have been very active in recent years. If debris flows occur again, the possibility of blocking the main river is very high. The resulting dam-breaking flood could cause damage to the newly constructed town of Yingxiu, which is just 5.4-km downstream of the Xiaojia gully. Therefore, mitigation measures should be applied to decrease the risk level.

A back-calculation method and numerical simulation to assist with mitigation planning in the Xiaojia gully were proposed based on the river-blocking prediction. The peak discharges of permissible floods that maintained the safety of Yingxiu were calculated first. Then, a permissible level of blockage by debris flows was determined based on the back-calculation of flooding resulting from dam breaking. Finally, the permissible scale of debris flows was obtained to facilitate mitigation planning for Xiaojia gully. The peak discharge of the permissible flood (QP) was calculated to be <1496.48 m3/s to ensure the safety of the Yingxiu. Accordingly, under 1/3 and 1/2 breach modes, the permissible peak discharge of a dam-breaking flood at the dam site (Qbp) should be smaller than 2018.73 and 2878.31 m3/s, the permissible blocking height (HP) of the debris flow barrier dam should be less than 43.09 and 31.64 m. Finally, the total volume (Vdp) and the peak discharge (Qcp) of a single debris flow event should be controlled to not exceed 70.59 × 104 m3 and 784.59 m3/s and 45.21 × 104 m3 and 551.77 m3/s, under the 1/3 and 1/2 breach modes, respectively.

After the comparison of the permissible (Qcp) and possible (Qc) peak discharges in Xiaojia gully, the peak discharge of a single debris flow event with a 100-year return period (Qc = 790.50 m3/s, P = 1 %) was used for mitigation planning in the gully. A numerical simulation using the Kanako simulator was carried out to facilitate the mitigation planning. By constructing the two check dams in the downstream channel, the simulation results indicate that the debris flow peak discharge decreases from 727.3 m3/s before mitigation to 531.2 m3/s after mitigation at the gully mouth. For the mitigation of a single debris flow event, the effect is remarkable with a peak discharge reduction rate of 26.96 %. However, the debris flow occurrence frequency was very high in the Xiaojia gully after the 5.12 Wenchuan earthquake. The storage capacity of the two check dams can be quickly used up because of the high frequency of debris flows. Therefore, a deposition work on the debris fan was recommended to trap debris flow sediments. For the mitigation of next possible debris flows, the removal of previous debris flow deposits in the deposition work is very essential for keep the trapping space. Above all, by implementing the two check dams in the downstream channel and the deposition works on the debris fan, debris flow damage can be decreased to an acceptable level for ensuring the safety of Yingxiu Town. Besides the mitigation works, early warning for the debris flow in the Xiaojia gully is also very important. As shown in Fig. 3, the distance between the Yingxiu Town and the gully is about 5.4 km. It takes about 12 min for the out-breaking flood (vP = 7.39, Table 2) reaching to the Yingxiu Town. Therefore, a scientific warning system is very essential for the safety of the Yingxiu Town.

For the above proposed method, a series of empirical equations were used for calculation. Some parameters of the empirical equations such as the roughness coefficient of the Yuzixi River and the Xiaojia gully, the blockage coefficient in the Xiaojia gully were determined empirically. Therefore, the uncertainties could be caused for the calculating results. For example, the average velocity (vP) of the permissible flood passing through the protection embankments at Yingxiu Town can increase about 14.3 % when the roughness coefficient of the Yuzixi River (nP) is determined as 0.035 instead of 0.04 (Eq. 1). Accordingly, the peak discharge of the permissible flood (QP) can increase about 1.14 times, causing the uncertainties for the calculating results of the peak discharge from the dam-breaking flood at the dam site (Qbp) and the permissible peak discharge (Qcp) of debris flow in Xiaojia gully. To make the calculating results more rational and accurate, the determination of some parameters in the used empirical equations needs to be improved in future studies. Due to the uncertainties in the application of the empirical equations, the calculating results need to be verified. However, the available observation data of the Xiaojia gully were not enough to verify the reliability of the calculating results. Subsequent observations and studies will be continued to carry out in the Xiaojia gully for collecting enough data for verifying the calculating results. In addition, the proposed method will also be applied to other real debris flow cases to ensure its accuracy.

Notes

Acknowledgments

This research was financially supported by the Key Projects of the Chinese Academy of Sciences (KZZD-EW-05-01-04), the National Science and Technology Support Program (2012BAC06B02), and the Applicative Fundamental Research Program of Sichuan Province (2012JY0104).

References

  1. Chang DS, Zhang LM, Xu Y, Huang RQ (2011) Field testing of erodibility of two landslide dams triggered by the 12 May Wenchuan earthquake. Landslides 8(3):321–332CrossRefGoogle Scholar
  2. Chen HX, Zhang LM, Chang DS, Zhang S (2012) Mechanisms and runout characteristics of the rainfall triggered debris flow in Xiaojiagou in Sichuan Province, China. Nat Hazards 62(3):1037–1057CrossRefGoogle Scholar
  3. Cui P, Chen XQ, Zhu YY et al (2011) The Wenchuan earthquake (May 12, 2008), Sichuan Province, China, and resulting geohazards. Nat Hazards 56:19–36CrossRefGoogle Scholar
  4. Cui P, Gordon GD, Zhou XH, Zhang JQ (2013a) Scale amplification of natural debris flows caused by cascading landslide dam failures. Geomorphology 182:173–189CrossRefGoogle Scholar
  5. Cui P, Xiang LZ, Zou Q (2013b) Risk assessment of highways affected by debris flows in Wenchuan earthquake area. J Mt Sci 10(2):173–189CrossRefGoogle Scholar
  6. Du C, Yao LK, Shakya S, Li LG, Sun XD (2014) Damming of Large River by debris flow: dynamic process and particle composition. J Mt Sci 11(3):1–10CrossRefGoogle Scholar
  7. Huang RQ, Li WL (2009) Analysis of the geo-hazards triggered by the 12 May 2008 Wenchuan earthquake, China. Bull Eng Geol Environ 68:363–371CrossRefGoogle Scholar
  8. Huang Y, Chen W, Liu JY (2012) Secondary geological hazard analysis in Beichuan after the Wenchuan earthquake and recommendations for reconstruction. Environ Earth Sci 66:1001–1009CrossRefGoogle Scholar
  9. Kang ZC, Li ZF, Luo JT (2004) Research on debris flow in China Beijing. Science Press, Beijing (in Chinese) Google Scholar
  10. Keefer DK (1999) Earthquake-induced landslides and their effects on alluvial fans. J Sediment Res 69(1):84–104CrossRefGoogle Scholar
  11. Liu N, Zhang JX, Lin W et al (2009) Draining Tangjiashan barrier lake after Wenchuan Earthquake and the flood propagation after the dam break. Sci China Ser E Technol Sci 52(4):801–809CrossRefGoogle Scholar
  12. Liu JF, You Y, CHEN XC (2010) Identification of potential sites of debris flows in the upper min river drainage, following environmental changes caused by the Wenchuan earthquake. J Mt Sci 7(3):255–263CrossRefGoogle Scholar
  13. Liu JF, Nakatani K, Mizuyama T (2013) Effect assessment of debris flow mitigation works based on numerical simulation by using Kanako 2D. Landslides 10(2):161–173CrossRefGoogle Scholar
  14. Liu JF, You Y, Chen XQ, Liu JK, Chen XZ (2014) Characteristics and hazard prediction of large-scale debris flow of Xiaojia Gully in Yingxiu Town, Sichuan Province, China. Eng Geol 180:55–67CrossRefGoogle Scholar
  15. Nakatani K, Satofuka Y, Mizuyama T (2007) Development of “Kanako”, a wide use 1-D and 2-D debris flow simulator equipped with GUI. In: DeWrachien D, Brebbia CA, Lenzi MA (eds.). 2nd International Conference on Monitoring, Simulation, Prevention and Remediation of Dense and Debris Flows. Proc. Of 32nd Congress of IAHR, Venice, Italy, 10p,A2.c–182Google Scholar
  16. Nakatani K, Wada T, Satofuka Y, Mizuyama T (2008) Development of “Kanako 2D (Ver.2.00)”, a user-friendly one- and two-dimensional debris flow simulator equipped with a graphical user interface. Int J Erosion Control Eng 1(2):62–72CrossRefGoogle Scholar
  17. Ou GQ (1992) The study on the scale of debris flow. Japan, Kyoto: doctor degree thesis, graduate school of Kyoto University. (in Japanese)Google Scholar
  18. Satofuka Y, Mizuyama T (2005) Numerical simulation on a debris flow in a mountainous river with a sabo dam. J Jpn Soc Erosion Control Eng 58(1):14–19 (in Japanese with English abstract) Google Scholar
  19. Takahashi T, Nakagawa H (1991) Prediction of stony debris flow induced by severe rainfall. J Jpn Soc Erosion Control Eng 44(3):12–19 (in Japanese with English abstract) Google Scholar
  20. Takahashi T, Nakagawa H, Satofuka Y, Kawaike K (2001) Flood and sediment disasters triggered by 1999 rainfall in Venezuela: a river restoration plan for an alluvial fan. J Nat Disaster Sci 23(2):65–82Google Scholar
  21. Tang C, Zhu J, Li WL (2009) Rainfall triggered debris flows after Wenchuan earthquake. Bull Int Assoc Eng Geol 68:187–194CrossRefGoogle Scholar
  22. Tang C, Zhu J, Ding J, Cui FX, Chen L, Zhang JS (2011) Catastrophic debris flows triggered by a 14 August 2010 rainfall at the epicentre of the Wenchuan earthquake. Landslides 8(4):485–497CrossRefGoogle Scholar
  23. Walder JS, O’Connor JE (1997) Methods for predicting peak discharge of floods caused by failure of natural and constructed earthen dams. Water Resour Res 33(10):2337–2348CrossRefGoogle Scholar
  24. Wang F, Cheng Q, Highland L, Miyajima M, Wang H, Yan C (2009) Preliminary investigation of some large landslides triggered by the 2008 Wenchuan earthquake, Sichuan Province, China. Landslides 6(1):47–54CrossRefGoogle Scholar
  25. Wu JS, Tian LQ, Kang ZC (1993) Debris flow and its comprehensive control. Science Press, Beijing (in Chinese) Google Scholar
  26. Xiang LZ, Cui P, Zhong DL, Ge YG, Zhu XH, Yang W (2012) Quantitative hazard assessment of road debris flow in Wenchuan earthquake area: a case study of Xiaojia Ravine in Wenchuan county. J Southwest Jiaotong Univ 47(3):387–393 (in Chinese with English abstract) Google Scholar
  27. Yin Y, Wang F, Sun P (2009) Landslide hazards triggered by the 2008 Wenchuan earthquake, Sichuan, China. Landslides 6(2):139–152CrossRefGoogle Scholar
  28. You Y, Liu JF, Chen XZ (2012a) Debris flow formation conditions and optimal characteristics of drainage canal following Wenchuan earthquake. Environ Earth Sci 65(4):1005–1012CrossRefGoogle Scholar
  29. You Y, Liu JF, Chen XC (2012b) Design of sluiceway channel in a landslide dam triggered by the Wenchuan earthquake. Disaster Adv 5(4):241–249Google Scholar
  30. Zhang WJ (2008) The hydraulics. China Architecture and Building Press, BeijingGoogle Scholar
  31. Zhang HP, Yang N, Zhang YQ, Meng H (2006) Geomorphology of the Minjiang drainage system (Sichuan, China) and its structural implications. Quat Sci 26(1):126–135 (in Chinese with English abstract) Google Scholar
  32. Zhou BF, Li DJ, Luo DF (1991) Guide to prevention of debris flow. Science Press, Beijing (in Chinese) Google Scholar
  33. Zhuang JQ, Cui P, Ge YG (2010) Probability assessment of river blocking by debris flow associated with the Wenchuan Earthquake. Int J Remote Sens 31(13):3465–3478CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2015

Authors and Affiliations

  • Jinfeng Liu
    • 1
  • Yong You
    • 1
  • Xiaoqing Chen
    • 1
  • Xingzhang Chen
    • 2
  1. 1.Key Laboratory of Mountain Hazards and Earth Surface Process/Institute of Mountain Hazards and EnvironmentChinese Academy of SciencesChengduChina
  2. 2.School of Environment and ResourcesSouthwest University of Science and TechnologyMianyangChina

Personalised recommendations