Effect of excess pore pressure on the long runout of debris flows over low gradient channels: A case study of the Dongyuege debris flow in Nu River, China

    Effect of excess pore pressure on the long runout of debris flows over low gradient channels: A case study of the Dongyuege debris flow in Nu River, China Zhen-Hua Zhou, Zhe Ren, Kun Wang, Kui Yang, Yong-Jun Tang, Lin Tian, Ze-Min Xu PII: DOI: Reference:

S0169-555X(18)30025-4 doi:10.1016/j.geomorph.2018.01.012 GEOMOR 6294

To appear in:

Geomorphology

Received date: Revised date: Accepted date:

8 June 2017 17 January 2018 17 January 2018

Please cite this article as: Zhou, Zhen-Hua, Ren, Zhe, Wang, Kun, Yang, Kui, Tang, Yong-Jun, Tian, Lin, Xu, Ze-Min, Effect of excess pore pressure on the long runout of debris flows over low gradient channels: A case study of the Dongyuege debris flow in Nu River, China, Geomorphology (2018), doi:10.1016/j.geomorph.2018.01.012

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Effect of excess pore pressure on the long runout of debris flows over low gradient channels: a case study of the Dongyuege debris flow in Nu River, China

Faculty of Civil Engineering and Mechanics, Kunming University of Science and Technology, Kunming 650500, China

SC R

b

of Land Resource Engineering, Kunming University of Science and Technology, Kunming 650000, China

IP

a Faculty

T

Zhen-Hua Zhou a, Zhe Ren b, Kun Wang b, Kui Yang b, Yong-Jun Tang b, Lin Tian b, Ze-Min Xu b,*

*Corresponding author at: Faculty of Civil Engineering and Mechanics, Kunming University of Science and Technology, Kunming 650500, China. Tel.:+86 13629475100. Email address: [email protected] (Ze-Min Xu). Postal address: Room 530 of Civil

NU

Engineering Building in Kunming University of Science and Technology, Kunming, Yunnan, China.

MA

Abstract

Debris flows with long reaches are one of the major natural hazards to human life and property on alluvial fans, as shown by the debris flow that occurred in the Dongyuege (DYG) Gully in August 18, 2010, and caused

D

96 deaths. The travel distance and the runout distance of the DYG large-scale tragic debris flow were 11 km and 9

TE

km, respectively. In particular, the runout distance over the low gradient channel (channel slope < 5°) upstream of the depositional fan apex reached up to 3.3 km. The build-up and maintenance of excess pore pressure in the

debris flow.

CE P

debris-flow mass might have played a crucial role in the persistence and long runout of the bouldery viscous

AC

Experiments to measure pore pressure and pore water escape have been carried out by reconstituting the debris flow bodies with the DYG debris flow deposit. The slurrying of the debris is governed by solid volumetric concentration (SVC), and the difference between the lower SVC limit and the upper SVC limit can be defined as debris flow index (𝐼𝑑 ). Peak value (𝐾𝑝 ) and rate of dissipation (R) of relative excess pore pressure are dependent on SVC. Further, the SVC that gives the lowest rate of dissipation is regarded as the optimum SVC (𝐶𝑣𝑜 ). The dissipation response of excess pore pressure can be characterized by the R value under 𝐶𝑣𝑜 at a given moment (i.e., 0.5 h, 1 h or 2 h later after peak time). The results reveal that a relatively high level of excess pore pressure developed within the DYG debris-flow mass and had a strong persistence capability. Further research shows that the development, peak value and dissipation of excess pore pressure in a mixture of sediment and water are related to the maximum grain size (MGS), gradation and mineralogy of clay-size particles of the sediment. The layer-lattice silicates in clay particles can be the typical clay minerals, including kaolinite, montmorillonite and illite, and also the unrepresentative clay minerals such as muscovite and chlorite.

ACCEPTED MANUSCRIPT Moreover, small woody debris can also contribute to the slurrying of sediments and maintenance of debris flows in well vegetated mountainous areas and the boulders suspended in debris flows can elevate excess pore pressure and extend debris-flow mobility.

T

The parameters, including 𝐼𝑑 , Kp, R and etc, are affected by the intrinsic properties of debris. They, therefore,

IP

can reflect the slurrying susceptibility of sediments, and can also be applied to the research on the occurrence

SC R

mechanisms and risk assessment of other debris flows.

Keywords: Debris flow; Long runout; Low gradient; Excess pore pressure; Clay mineral; Small woody debris 1. Introduction

NU

Debris flows occur when masses of poorly sorted sediments, agitated and saturated with water, surge down slopes, channels and valleys in response to gravitational attraction. They capture social concerns all the time

MA

because they often trigger catastrophic disasters because of the long runout distance and destructive energy (Ng et al., 2015). Although debris flows are common phenomena in mountainous regions (De Hass et al., 2015), the

D

catastrophic mass movements are initially thought to be rare except on very steep slopes, or in active seismic areas

TE

(Takahashi, 2014; Chen et al., 2015; Nocentini et al., 2015; Zhang and Zhang, 2016). It has even been proposed that soil slopes with inclinations ranging from 26°to 45°are generally identified as most prone to the initiation

CE P

of debris flows (Wieczorek, 1987). In terms of transport mechanisms, thus, the occurrence of large-scale catastrophic debris flows on gentle slopes becomes a particularly enigmatic phenomenon. Debris flows have diverse mobility and depositional characteristics (Sohn, 2000; Pellegrino et al., 2015;

AC

Zhou and Ouyang, 2015). Coarse grains in a debris flow mass are supported either by the yield stress of slurries or by dispersive force, which forms two viewpoints to the research into particle support mechanisms (Qian and Wan, 1991). Numerous empirical models and numerical simulations for the transport mechanism of debris flows have been also proposed, mainly relying on input parameters that are often difficult to estimate, such as composition, volume, velocity, and frictional factors (Tang et al., 2012; Sa´nchez et al., 2013; Calvo et al., 2015; Perinotto et al., 2015; Zhang et al., 2015; Gregoretti et al., 2016; Fan et al., 2017). Unlike rock avalanches and sediment-laden water floods, however, solid and fluid forces influence the motion and rheological properties of debris flows simultaneously (Costa, 1988; Iverson, 1997; Klubertanz et al., 2009; Morino et al., 2015). According to various experimental observations and research, pore pressure is considered as the most significant triggering factor for the initiation of debris flows (Pudasaini et al., 2005; Ala and Mathewson, 2015; Guo et al., 2016) and, thus, high and persistent pore pressure in debris flows plays the critical role that leads to the long runout distances of debris flows (Major, 2000).

ACCEPTED MANUSCRIPT In the early time, by using samples of debris flows with different compositions, Pierson (1981) carried out a series of measurements on pore pressure, and the results showed that the average pore pressure could be elevated in excess of the hydrostatic pore pressure for timescales even longer than that of a debris flow. His study has

T

shown that a debris flow possesses the ability to trap pore fluid and, thus, generate excess pore pressure. Similarly,

IP

as concluded by Iverson et al (1997), debris flows can be mobilized by partial or complete liquefaction of a sliding

SC R

mass because of the development of excess pore pressure. Wang and Fei (1999) also concluded that in viscous debris flows, the additional pore pressure can support large fractions of the total solid particles. Lu et al (2010) maintained that the “water film” in soils may serve as a natural sliding surface for post-liquefaction failure.

NU

Thereafter, debris flows may occur on a slope with very gentle slope angle. Kokelaar et al (2014) also proposed that considerable mobility of debris flows can be attributed to pore pressure, because in debris flows, fine-grained

MA

flow-contact zones reduce frictional energy losses and lower flow-substrate permeability so as to enhance pore-fluid pressure retention. In addition, the drum experiments carried out by Kaitna et al (2016) revealed that for

D

debris flows the grain size distribution of the coarse fraction as well as the presence of fines strongly influenced

TE

the presence of sustained excess pore pressure. More recently, considering dilatancy effects (Iverson and George, 2014) in granular materials, François et al (2016) attempted to explain the dynamic response mechanism of fluid

CE P

pressure in debris flows. Based on the concept called “pore pressure feedback” (Iverson, 2005), they pointed out that a change in the fluid pressure may result from a dilation of the granular phase, that induced a sucking of the fluid within the mixture and a diminution of the fluid pressure, thereby increasing the effective friction on the

AC

granular phase. On the contrary, a contraction of the granular phase induced an expulsion of the fluid from the mixture and an increase of the fluid pressure, thereby decreasing the effective friction. The occurrence of dilation or contraction of the granular material depends on whether the solid volumetric concentration (SVC) is respectively higher or lower than a critical value. For the long runout of debris flows over low gradients, as mentioned previously, the generation of excess pore pressure is the prerequisite and its maintenance is the decisive factor. Despite growing attention being paid to the importance of excess pore pressure in debris flows (Iverson, 2003; Okada and Ochiai, 2008; Chen et al., 2010; Li et al., 2015; He et al., 2016) , some scholars use excess pore pressure to explain the transport mechanisms of debris flows over low-gradient channels. These arguments are either speculative or based on the measurement of pore pressure in high-speed debris flows. A lack of research still exists on some specific problems of debris flows in low gradient channels, including the build-up, peak value, maintenance and dissipation of excess pore pressure as well as the influencing factors. The research into these problems can deepen the understanding of high-density

ACCEPTED MANUSCRIPT debris flows that travel long distances over low gradients and other related responses. At 1:30 on August 18, 2010, a large-scale bouldery viscous debris flow, transporting 600,000 m3 of material, erupted from the mouth of DYG Creek on the left bank of Nu River, 584 km northwest of Kunming in Yunnan,

T

China (Fig. 1). The total distance travelled by the DYG debris flow was about 11 km and the total length of the

IP

reach with a gentle gradient between 2° and 5° was as long as 3.3 km. In particular, the moving dam of the snout

SC R

(Iverson et al., 2010) stopped and formed a debris flow dam at about 500 m upstream of the fan apex. It did not cause, however, detrital deposition. Subsequently, the short-lived dam was breached and the remobilized debris flow buried the DYG Village on the former depositional fan, killing 96 people and even damming the Nu River

NU

for a few hours (Fig. 2). This is an indication that a high level of excess pore pressure existed within the debris flow and the debris had a strong water holding capacity because the DYG debris flow could travel more than 3 km

AC

CE P

TE

D

MA

along the low gradient channel and could be remobilized after the stoppage.

Fig. 1. Location of the study site.

In this study, a series of experiments on pore pressure and pore water escape have been conducted on the reconstituted debris flow bodies sampled from the DYG debris flow deposit. Based on the experimental results, the effects of SVC and maximum grain size (MGS) on the generation and maintenance of excess pore pressure and other related problems have been evaluated and discussed. The laboratory findings can explain the slurry forming ability and mobilization mechanism of debris flows that travel long runout distances over low gradient channels, and also can be applied to the risk assessments of debris flows.

SC R

IP

T

ACCEPTED MANUSCRIPT

NU

Fig. 2. The digital elevation model of DYG debris flow gully.

2. Materials

MA

2.1. Sampling

The debris flow detritus tested has been sampled from the unconsolidated deposit in the depositional area

D

adjacent to the mainstream channel of DYG, as shown in Fig. 3. According to our field investigations (Jiang,

TE

2016), the texture of the DYG debris flow deposit was megaphyric, i.e., the grains of the deposit can be divided

CE P

into two main groups: one was the megalock group in which the equivalent grain sizes of boulders are several meters and the largest could be 9 m, and the rest was the matrix group that is made up of the particles smaller than tens of centimeters. Because of the limitation of test capacity, the MGS of the samples collected has been confined

AC

to smaller than 256 mm. Given that the fine particles may be lost because of the leaching action of surface water, only well-preserved debris flow deposit, about 1 m below the deposit surface, was collected as samples. Thereafter, these samples were air-dried and taken back to the laboratory. Through screening, three groups of samples whose MGSs are 2 mm, 20 mm and 40 mm, respectively, have been prepared.

Sampling profile

2015/10/25 Fig. 3. Sampling profile in the depositional area.

ACCEPTED MANUSCRIPT 2.2. Grain size distribution First, to determine the grain size distribution of the DYG debris flow deposit with MGS < 256 mm, 200 kg of debris were collected from the deposition profile (Fig. 3). A system of sieves having seven mesh sizes (i.e., 25 cm,

T

10 cm, 5 cm, 2 cm, 1 cm, 0.5 cm, 0.2 cm) was used in the field sieving test. For the fraction smaller than 0.2 cm,

IP

the laboratory water sieving test of 0.1 cm and 0.075 mm mesh size sieves was used. For the group whose grain

SC R

size was smaller than 0.075 mm, the gradation was measured with a Beckman Coulter LS 13320 (USA) laser diffraction particle size analyzer whose measurement accuracy was 0.04 microns. Using the same sieving method, the grain size distributions of the samples with 2 mm, 20 mm and 40 mm MGSs have been determined. The grain

NU

size distributions and the grading curves of the samples are shown in Fig. 4 and Table 1, respectively. (b)

CE P

TE

D

MA

(a)

Fig. 4. (a) The grain size distribution of the debris flow deposit with MGS < 256 mm. (b) The grain size distributions of the

Table 1

AC

samples with MGS of 2 mm, 20 mm and 40 mm.

Gradation parameters (𝑀𝑧 and 𝐶𝑢 are the median grain size and uniformity coefficient, respectively).

MGS (mm) 2 20 40 256

d10 0.007 0.014 0.014 0.045

Grain size (mm) d16 d30 d50 0.016 0.050 0.155 0.033 0.147 0.571 0.031 0.181 0.790 0.128 0.840 20.000

d60 0.254 1.000 2.300 48.000

d84 0.700 7.200 18.500 154.348

Parameter 𝑀𝑧 0.290 2.601 6.440 58.159

𝐶𝑢 36.286 71.430 164.286 1066.667

2.3. Mineral composition of the fine particles Through repeated suspension, precipitation in water and drying at 65 °C, the soil particles with MGS < 0.005 mm have been extracted, as shown in Fig. 5a. Further, by means of powder XRD (X-Ray Diffraction), the mineral composition of the fine soil particles has been analyzed with a Rigaku®DMAX III A diffractometer equipped with (λ= 1.5406 Å) x-radiation. The scan settings were 5-40° 2θ, 0.02° step size, and 2-s count time per step. The

ACCEPTED MANUSCRIPT result of the analysis shows that the fine particles are mainly composed of quartz (30 wt.%), mica (30 wt.%), plagioclase (15 wt.%) and chlorite (10 wt.%), as shown in Fig. 5b. The typical clay minerals, including illite, montmorillonite and kaolinite (Terzaghi, 1996; Tang et al., 1999; Gao and Yuan, 2001; Xu et al., 2012) are absent

T

in the fine particles of the DYG debris flow deposit. The phyllosilicates contain only mica and a small amount of

IP

chlorite, both of which are collectively called the unrepresentative clay minerals. Clay minerals, however, are

SC R

generally considered as the vital constituents in debris flows because they can influence the yield stress and thereby initiate the mobility of debris flows significantly (Iverson et al., 2000; De Blasio et al., 2011; Yu et al., 2013). Hence, the yield strength theory requires further improvements. (a)

D

MA

NU

(b)

CE P

2.4. Small woody debris

TE

Fig. 5. (a) Fine particles (< 0.005 mm) in the DYG debris flow deposit. (b) Mineral composition of fine particles.

As reported by Luo and Xu (2016), the DYG debris flow occurred in a wildwood area, which was well vegetated with a variety of trees and vegetation. Large amounts of large woody debris that were 1 m or longer in

AC

length and 0.1 m or bigger in diameter (Jackson and Sturm, 2002; Marcus et al., 2002; Faustini and Jones, 2003) were found widely distributed in the debris flow deposit along the upper reaches of the flow path. In addition, small woody debris (< 2 mm in diameter, as shown in Fig. 6) was abundant in the debris flow deposit along the downstream areas, with a total length of up to 3.3 km and a slope between 3-5°. These small woody debris may play an important role in maintaining the long runout of the DYG debris flow along the low gradient channel. (a)

(b)

Fig. 6. (a) Small woody debris in the debris flow deposit of < 2 mm. (b) Picked small woody debris from the debris flow deposit.

ACCEPTED MANUSCRIPT 3. Methods The research into the problems related to excess pore pressure has been carried out using two types of experiments, i.e., measurements of excess pore pressure and pore water escape. The two types of experiments

T

involve three groups of the DYG debris flow deposit with three different MGSs, i.e., 2 mm, 20 mm and 40 mm.

IP

The sediment samples were mixed with water to reconstitute debris flow slurries of different SVCs. The SVC can

Cv 

w

and

s

are the densities of water and solid in the sample respectively, and

the entire sediment-water mixture.

m

is the density of

MA

3.1. Measurement experiments for excess pore pressure

(1)

NU

where

m   w s  w

SC R

be calculated using the following formula:

D

3.1.1. The generation mechanism of excess pore pressure

TE

The variation of the hydrostatic pore pressure, Ps, with the depth, h, in water can be expressed by the

CE P

following relationship:

Ps  wh

(2)

When solid grains are mixed into a fluid, a two-phase system is formed. Thus, part of the load of the solid phase is

AC

temporarily carried by the buoyancy of the fluid phase. As shown by Hampton (1979), pore pressure, P, increases in proportion to the fraction of the load balanced by the interstitial fluid. This increase in pressure is known as excess pore pressure, Pe, which is in addition to the hydrostatic pore pressure, Ps, as follows:

P  Ps  Pe

(3)

The hydrostatic component of pore pressure is defined by Eqn (2). And as Pierson (1981) reported, excess component is defined as follows:

0 Pe  (s  w ) Cvidh h

(4)

where Cvi is the SVC of the grains supported by interstitial fluid. By assuming that the solid particles are dispersed throughoutly into the mixture, Eqn (4) can be simplified to:

ACCEPTED MANUSCRIPT Pe  (s  w )Cvh

(5)

Further, P can be explicitly divided into two parts:

P  wh  (s  w )Cvh

IP

T

(6)

When pore pressure is measured by a piezometer, Eqn (6) can be rewritten as

SC R

P  （  wh  wh ' w h  h '）

(7)

where h is the depth of the piezometer below the debris-flow surface, and h’ is the height of the hydraulic head

NU

above the debris-flow surface. By comparing Eqns (3), (6) and (7), the following formula can be obtained:

MA

Pe  wh '

h ' Cv(s  w )  h w

h '/ h

is known as relative excess pore pressure (Pierson, 1981).

TE

where

(9)

D

and

(8)

CE P

According to Eqns (3) and (9), when the excess pore pressures in a static system completely dissipate, P is close to Ps, which implies that the solid grains supported by the interstitial fluid have settled. To some extent, the rate of particle settling is dependent on the rate of dissipation of excess pore pressure. Hence, the measurements of

AC

pore pressure can yield considerable insights into the maintenance mechanisms of debris flows. 3.1.2. Measurement procedures Through using piezometers, the pore pressures have been measured. The concrete apparatus is shown in Fig. 7. Each piezometer comprises a glass tube whose inner diameter, wall thickness and total length are 5 mm, 1 mm and 500 mm, respectively. Further, scale lines exist at 1 mm intervals on the surface of the glass tube. Three piezometers have been designed to be fixed vertically at 4-cm intervals beginning at 14 cm off the slurry surface along the radial direction. That is, the setting depths of the three piezometers are 14 cm, 18 cm and 22 cm. To prevent fines from entering the piezometers, the inlet ends at the bottom have been covered with a double layer of rough gauze. This allowed only water to enter the piezometers. In order to reduce the effect of the water in the piezometers themselves on the excess pore pressure measurements, the experiments have been conducted in a big, open and nearly vertical-sided cylindrical container whose height is 35 cm, and the top and bottom diameters are

ACCEPTED MANUSCRIPT 33.2 cm and 26.9 cm, respectively. Further, the total accumulated thickness and volume of the slurry for each run

MA

NU

SC R

IP

T

are 29.5 cm and 19,890 cm3, respectively.

TE

D

Fig. 7. The apparatus of measurement experiment for excess pore pressure.

The concrete measurement procedures for each run are as follows:

CE P

(i) Prior to the experiment, the sediments were thoroughly mixed with water by hand for about 30 minutes in a large basin. Then, the density and SVC of the debris flow slurry were measured. After that, the reconstituted slurry was loaded into the apparatus ( Fig. 7).

AC

(ii) Next, the Rhodamine B solution (a red synthetic dye) was injected into the piezometers until the height of the water column equals the corresponding setting depths of the three piezometers. Then the three piezometers were fixed vertically at their specific depths. After that, the Rhodamine B solution was injected into the piezometers again until the height of the water column above the slurry surface was about half of the setting depths of the piezometers, except some runs for the debris flow samples whose MGS were 2 mm. (iii) At that moment, the hydraulic heads in the piezometers would rise quickly. Immediately, the timing began and the first readings of hydraulic heads were taken. In the process of observation, the pore water escaping from the interior to the surface of the slurry was sucked out with a pipette in time. The entire duration of observation for each run lasted for 7.5 hours so that more data can be obtained. From 0 to 60 minutes, 60 to 90 minutes and 90 to 450 minutes, the readings were taken every 1 minute, 2 minutes and 5 minutes, respectively. 3.2. Experiments for pore water escape

ACCEPTED MANUSCRIPT The experiments for pore water escape have been carried out on a simplified tilting rheometer (D'Agostino et al., 2010) which mainly consists of a tilting plane which is 121.5 cm long and 93.5 cm wide. For the experimental conditions to be close to the field conditions in the low gradient DYG channel, using an inclinometer, the inclined

T

plane slope has been fixed to 5° (Fig. 8). To investigate the escape time and the rate of pore water escape, the

IP

sediment-water mixtures with different SVCs and MGSs have been tested in the experiments. The volumes of the

SC R

experimental sediment-water mixtures with the MGS of 2 mm and 20 mm have been set to 200 ml and its starting device has been constructed of a container, whose inner diameter and height are 10 cm and 6.3 cm, respectively. For the samples with MGS of 40 mm, in view of the wider gradation and uniform sampling, its experimental

NU

volume has been set to 2,000 ml and the inner diameter and height of the starting device are 20 cm and 12 cm, respectively. Further, the Rhodamine B was added to the clean water and then mixed with the debris flow deposit.

MA

The Rhodamine B served as a colouring agent so that the experimental phenomenon could be observed clearly

AC

CE P

TE

D

and the escape length of pore water could be measured more accurately.

Fig. 8. The experimental apparatus for pore water escape.

The experimental procedures are the followings: (i) From the start of the experiment, the sediment-water mixtures were mingled by manual stirring for about 5 min to ensure the mixtures were homogeneous. After that, they were loaded into the starting device which was located in the longitudinal axis of the simplified tilting rheometer. (ii) Thereafter, simultaneously the starting device was elevated and the timer was started, the experimental slurry was released, which moved downward in an elongating and thinning manner along the tiling plane without restraint. After a certain period, the front of the slurry stopped at a position on the slope. (iii) After a further period, the red pore water began to escape from the deposit tongue. When the escape lengths of pore water that are defined as the distance between the front of the debris flow slurry and the front of

ACCEPTED MANUSCRIPT the pore water escaping from the slurry reached 5 mm and 10 mm, respectively, the time spent was recorded separately. In addition, using a three-dimensional laser scanner (Scan Station C10) with surface modelling accuracy of 2 mm, produced by Leica Geosystems AG in Switzerland, the morphological parameters of these

T

experimental debris flow deposit were calculated.

SC R

4.1. Results of the measurement experiment for pore pressure

IP

4. Results

The build-up and dissipation processes of excess pore pressure can vary distinctly among the different

NU

slurries. This can be seen in Fig. 9, which shows the variation of the relative excess pore pressure at the depth of 18 cm against time. As a whole, the relative excess pore pressure increases with the increase in the SVC. Further,

MA

for each group of sediment-water mixtures with different MGSs, when the SVC is sufficiently low, the duration curve of the relative excess pore pressure turns sharply concave-downward. This is an indication of the quick

D

escape of pore water. Conversely, when the SVC is sufficiently high, the duration curves are relatively constant,

TE

which indicates that the high excess pore pressure is maintained.

(c)

B

(b)

A

D E

AC

C

CE P

(a)

F

Fig. 9. Variations of the measured relative excess pore pressures in the slurries with time. The measurements have been made at the depth of 18 cm in the slurry columns. (a) Dmax = 2 mm (b) Dmax = 20 mm (c) Dmax = 40 mm. The Letters A, B, C, D, E, F refer to the sharply concave-downward duration curves of the relative excess pore pressure.

ACCEPTED MANUSCRIPT 4.1.1. The peak value of relative excess pore pressure We use “Kp” to denote the peak value of relative excess pore pressure. The experimental value of Kp can be obtained from Fig. 9. If all the solid particles are supported by the fluid in a debris flow, the theoretical value of

h '/ h

in theory), can be calculated by Eqn (9). As shown in Fig. 10, we draw a

T

Kp (equals the maximum value of

IP

comparison between the theoretical Kp and the experimental Kp. For the sediment-water mixtures whose MGS is 2

SC R

mm, the differences between the experimental and theoretical Kp range between 0-17.34%, with an average difference of 9.85% (Fig. 10a). For the sediment-water mixtures whose MGS is 20 mm, the differences between the experimental and theoretical Kp range between 6.81-18.62%, with an average difference of 12.91% (Fig. 10b).

NU

For the sediment-water mixtures whose MGS is 40 mm, the differences between the experimental and theoretical Kp range between 11.33-29.21%, with an average difference of 20.12% (Fig. 10c). The peak value of relative

MA

excess pore pressure generally increases with SVC.

Further, as the MGS increases, the difference between the experimental and theoretical curves gets bigger.

D

The results manifest that not all the weight of the solid particles are borne by the intergranular fluid and the fine

TE

grains show stronger water holding capacity. Hence, the fine grains may play a more critical role than the coarse

(b)

(c)

AC

(a)

CE P

grains in the buildup of excess pore pressure.

Fig. 10. Comparison between the experimental and theoretical curves of peak value of relative excess pore pressure. (a) Dmax = 2 mm. (b) Dmax = 20 mm. (c) Dmax = 40 mm.

4.1.2. Dissipation of relative excess pore pressure By calculating the rate of dissipation of relative excess pore pressure, the effects of SVC and MGS on the maintenance of pore pressure can be further evaluated. The rate of dissipation of relative excess pore pressure can be calculated using the following formula:

R

Kp  Ka Ta  Tp

(10)

ACCEPTED MANUSCRIPT where Tp is the time when Kp occurs, and Ka is the relative excess pore pressure at the observation time,

Ta . Fig. 11a-l show the variations of R with Cv, Ta Tp and MGS. It can be seen that at all the different Ta Tp ,

T

the variations of the rate of dissipation of the relative excess pore pressure with the SVC are generally concave

IP

downward. For instance, for the sediment-water mixture whose MGS is 2 mm, when the SVC increases from 0.41

SC R

to 0.45, the rate of dissipation of the relative excess pore pressure fluctuates quite steadily between 1.85103×10-3 min-1 and 2.72974×10-3 min-1, as shown in Fig. 11j. When Cv reaches 0.46, R decreases sharply and reaches the minimum 0.21743×10-3 min-1 when Cv is 0.52. After that, R begins to increase gradually with Cv. Although the

NU

sediment-water mixture whose MGS is 20 mm does not exhibit obvious regular changes, R reaches the minimum when Cv is 0.68. Subsequently, R increases with increasing Cv (see Fig. 11b, e, h and k). In addition, the results

MA

also show that the rate of dissipation of the relative excess pore pressure depends on the SVC to a large extent, and also varies considerably with the increase in MGS. The larger the MGS, the larger the critical-state SVC at

(b)

(c) Ta – Tp = 60 min

Ta – Tp = 60 min

(d)

AC

CE P

Ta – Tp = 60 min

TE

(a)

D

which the rate of dissipation of the relative excess pore pressure reach the minimum.

(g)

(f)

(e)

Ta – Tp =120 min

(h) Ta – Tp = 180 min

Ta – Tp =120 min

Ta – Tp =120 min

(i) Ta – Tp =180 min

Ta – Tp =180 min

ACCEPTED MANUSCRIPT (k)

(j)

(l)

Ta – Tp =240 min

Ta – Tp =240 min

IP

T

Ta – Tp =240 min

SC R

Fig. 11. Changes with the SVC in the rate of dissipation of relative excess pore water pressure at different times, with respect to the sediment-water mixtures with different MGSs. (a)(d)(g)(j): Dmax = 2 mm. (b)(e)(h)(k): Dmax = 20 mm. (c)(f)(i)(l): Dmax = 40 mm.

NU

4.2. Results of pore water escape experiments

MA

The results of the pore water escape experiments, conducted on the debris flow samples with different MGSs at different SVCs, have been summarized in Table 2. Using the three-dimensional laser scanner, the morphologies of the experimental debris flow deposits have been converted into digital images. Then, using an image analysis

fa ) by the area of the base of the starting device ( A0 ) gives a morphological

TE

the final propagation area (

D

program, the final spread areas of the experimental slurries have been determined, as shown in Fig. 12. Dividing

CE P

parameter, namely the dimensionless spread area ( f ' ), which can be expressed as follows: fa A0

(11)

AC

Table 2

f '

Escape time of pore water and morphological parameters for the experimental slurry deposits with different MGSs. MGS of Material (mm)

Parameter

Cv 2

t1(s) t2(s)

f'

Cv 20

t1(s) t2(s)

f'

Cv

0.43 115 136 12.81 0.48 227 244 7.54

0.46 168 223 7.05 0.50 292 326 4.56

0.48 501 587 5.24

0.51 517 842 4.21

0.53 546 937 3.17

0.57 858 2533 1.05

0.59 1245 / 1.03

0.52 373 433 4.48

0.54 716 986 3.14

0.57 952 1543 2.00

0.59 1163 2810 1.76

0.62 3626 / 1.49

0.59 0.62 0.63 488 492 733 40 520 521 839 6.97 5.84 3.89 f' Notes: t1 and t2 are the times when the escape length of pore water t1(s) t2(s)

0.66 1590 2063 3.02 reached 5

0.60 2074 / 1.03

0.61 / / 1

0.67 0.69 0.70 1983 3973 / 2660 5533 / 2.13 1.23 1.09 mm and 10 mm, respectively, and “/” means after 2 h,

the escape length of pore water escaping from the sediment-water mixtures has not reached 5 mm or 10 mm.

ACCEPTED MANUSCRIPT 150

100

0 (cm)

50

100

150

200

250

300

350

IP

T

50

400

450

500

SC R

Fig. 12. Three-dimensional laser scanned images of the experimental slurry deposits whose MGS is 40 mm. In this figure, the SVC of the debris flow samples increase from left to right, i.e., 0.59, 0.62, 0.63, 0.66, 0.67, 0.69 and 0.70.

NU

4.2.1. The influence of SVC and MGS on the rate of escape of pore water

The experimental data in Table 2 are presented graphically in Fig. 13. Fig. 13a, 13b and 13c show the

(b)

(c)

CE P

TE

D

(a)

MA

variations of t1 and t2 of the tested slurries with MGS of 2mm, 20mm and 40mm, respectively, with SVC.

Fig. 13. Variations of pore water escape time with SVC for (a) Dmax = 2 mm (b) Dmax = 20 mm (c) Dmax = 40 mm.

AC

Further, Fig. 14 shows the relationship between “t2-t1” (the time for the escape length of pore water to increase from 5 mm to 10 mm) and the SVC for the tested slurries with three different MGSs. In this study, the parameter “t2-t1” has been used as a quantitative indication of the rate of escape of pore water. Using the least square method, an exponential function is used to relate the escape time of pore water to the SVC, as follows:

t '  t 2 - t1  AeBCv

(12)

where t’ is the time spent for the escape length of pore water to increase from 5 mm to 10 mm, and A and B are the fitting parameters. The values of A and B are shown in Table 3.

SC R

IP

T

ACCEPTED MANUSCRIPT

NU

Fig. 14. Relationship between the time for the escape length of pore water to increase from 5 mm to 10 mm and the SVC for the three kinds of the slurries with different MGSs. Table 3

Fitting formula

Parameter

t'  AeBCv

A B r

Dmax = 2 mm

Dmax = 20 mm

Dmax = 40 mm

1.67E-5 32.32 0.9926

3.44E-9 45.56 0.9905

2.91E-10 42.49 0.9969

D

Note: r is correlation coefficient.

MA

Fitting formula and corresponding parameters of relationship between the escape time of pore water and the SVC.

TE

As shown in Fig. 13, for the sediment-water mixtures with the same MGS, the amount of time for the escape

CE P

length of pore water to reach 5 mm or 10 mm increases with the SVC. Similarly, as the SVC increases, the time spent for the escaped pore water to move from 5 mm to 10 mm is also longer (Fig. 14). Because all the correlation coefficients are greater than 0.99, as shown in Table 3, this is an indication that in exponential formulas, the

AC

escape of pore water is significantly related to the SVC. For the sediment-water mixtures with the same SVC, the larger the MGS is, the easier it is for pore water to escape. Besides, for each group of sediment-water mixtures with the same MGS, the exponential relationship between the rate of escape of pore water and the SVC shows that a very small increment in SVC may dramatically increase the difficulty of pore water escaping. For example, for the sediment-water mixtures whose MGS is 2 mm, when the SVC varies from 0.43 to 0.57, t’ increases from 21 s to 1675 s, increasing by 2 orders of magnitude. In addition, according to the experimental observations and measurements shown in Table 2, the minimum SVC under which the flow length of the escaped pore water was shorter than 5 mm or 10 mm in the experiment processes increases with the MGS of the samples. All these results show that the fines in the experimental slurries have a greater retarding effect on pore water than the coarse particles. 4.2.2. The influence of SVC and MGS on the morphology of the experimental flow deposits

ACCEPTED MANUSCRIPT For the scan data in Table 2, the dimensionless spread area and the SVC can be related exponentially, as shown in Fig. 15. The values of the fitting parameters for the exponential formulas are displayed in Table 4. From Fig. 15, it can be seen that even with minor changes in the SVC, significant changes occur in the morphological

T

characteristics of the experimental flow deposits. For instance, for the sediment-water mixtures whose MGS is 2

IP

mm, when the SVC increases from 0.43 to 0.61, the dimensionless spread area decreases sharply from 12.81 to

SC R

1.00.

For sufficiently low SVCs, pore water quickly escapes from the tested slurries and some fine grains are transported away by the escaped pore water, leading to an apparent phase separation and eventually long and

NU

narrow spread areas are formed. As the SVC increases, the thickness of the slurry deposit increases accordingly and the dimensionless spread area becomes smaller until it is close to unity. Further, the SVC under which the

MA

dimensionless spread area approaches unity increases with the increase of MGS. According to the experimental results, for the debris flow deposit, its geometry and ability to trap pore fluid

D

have a close relationship with the SVC and MGS. Hence, it can be hypothesized that for the sediment-water

TE

mixtures with a specific MGS, a specific SVC exists at which the solid phase and the fluid phase can mix with each other to an appropriate degree so that the resulting pore water can escape slowly enough to minimize the

AC

flows.

CE P

interfacial contact between the solid grains, thereby enhancing the mobility and propagation ability of debris

Fig. 15. Relationship between the dimensionless spread areas and SVCs of the experimental flow deposits with different MGSs.

ACCEPTED MANUSCRIPT Table 4 Fitting formula and corresponding parameters of relationships between the dimensionless spread areas and SVCs for the experimental flow deposits with different MGSs. Dmax = 2 mm 7363.7 -14.89 0.9715

Dmax = 20mm 1790.5 -11.68 0.9630

Dmax = 40 mm 0.9376 -16.91 0.9376

IP

f '  AeBCv

Parameter A B r

T

Fitting formula

SC R

5. Discussion

As shown in Fig. 9, the excess pore pressure measured in each piezometer, first increases, reaches a maximum, and then decreases. The main differences between the measured pressures at different piezometers are

NU

the maximum pressure and the rate of decrease. The DYG debris flow could move over the low gradient channel, and this suggests that its effective friction angle is only a few degrees, which indicates that a high level of excess

MA

pore pressure existed within the debris flow. The rate of dissipation of excess pore pressure determines the growth rate of effective stress. By this rationale, the development and maintenance of excess pore pressure are crucial in

D

facilitating high mobility and long runout travel of high-density debris flows over low gradients. The main issues

TE

to be discussed in this section include peak values, maintenance and build-up mechanisms of excess pore pressure. 5.1. Peak value of relative excess pore pressure

CE P

As shown in Fig. 9 and Fig. 10, 𝐾𝑝 values are all at a high level, ranging between 0.7679 and 1.0296. Furthermore, significant differences occur in the peak value of relative excess pore pressure among the experimental slurries with different MGSs and that with the same MGS but different SVCs.

AC

Firstly, the larger the MGS of the tested debris flow with the same SVC, the larger the 𝐾𝑝 . The maximum 𝐾𝑝 in the three groups of the reconstituted samples with MGSs of 2 mm, 20 mm and 40 mm, are 0.8706, 1.0172 and 1.0296, respectively. This may be the intrinsic constraint mechanism of the high mobility for the coarse-textured debris flows with continuous gradations (Pierson, 1981; Benda and Cuny, 1990). Secondly, for the tested slurries with the same MGS, the larger the SVC, the larger the 𝐾𝑝 , which is consistent with the change of the solid particle load borne by pore water. It does not mean, however, that the mobility of debris flows is enhanced with the higher SVC (see Fig. 12). This is because, with the higher SVC, the intergranular friction is significantly increased during the movement of debris flows. 5.2. SVC boundary, optimum SVC and debris flow index The experimental results show that for the three groups of the remolded samples with different MGSs, once the SVC is less than a threshold value, the integrated debris-water slurry will not be formed, i.e., upon stop stirring

ACCEPTED MANUSCRIPT the water-debris mixture in the container, the debris immediately starts to settle, forming a double-layer structure comprising an upper layer of turbid liquid and a lower layer of saturated debris. Likewise, during the pore water escape experiments, the low SVCs would cause pore water to escape rapidly and the deposition of debris in situ.

T

Further, in the duration curves of relative excess pore pressure, the quick escape response of pore water is shown

IP

as a sharp peak, after which the relative excess pore pressure decreases rapidly (i.e., line A, B and C in Fig. 9a;

SC R

line D and E in Fig. 9b; line F in Fig. 9c). The threshold value of SVC is called 𝐶𝑣ℎ , which represents the SVC boundary between a hyperconcentrated streamflow (Pierson and Scott, 1985; Lavigne and Suwa, 2004) and a debris flow.

NU

When the SVC > 𝐶𝑣ℎ , no two-layer structure exists in static sediment-water mixtures. As reflected in the duration curves of relative excess pore pressure, the rate of dissipation of pore water is greatly slowed down and

MA

no sharp peak appears. The rate of dissipation of relative excess pore pressure decreases with increasing SVC, leading to a significant enhancement of debris flow suspension competence, which is consistent with the

D

experimental results of Yang et al (2014). On the other hand, when the SVC continues to increases to another

SVC (hereafter known as 𝐶𝑣𝑜 ).

TE

critical value, the rate of dissipation can reach the minimum. In this study, the critical value is called the optimum

CE P

When the SVC > 𝐶𝑣𝑜 , the rate of dissipation begins to be positively correlated with the SVC, as shown in Fig. 11. Under this condition, the solid particles interact with each other and the grain-to-grain contact gradually dominates this static system. This is an indication that for a specific debris with slurry forming ability, an

minimum.

AC

optimum SVC occurs at which the rate of dissipation of relative excess pore pressure can be reduced to the

When the SVC reaches a new critical value after 𝐶𝑣𝑜 , the rate of escape of pore water gets extremely slow but no fluid characteristics display in the sediment-water mixture at the same time, i.e., the mobility of the mixture on the tilting rheometer is lost (Fig. 12). The critical value of SVC is called 𝐶𝑣𝑑 , which denotes the SVC boundary between a debris flow and a granular flow, indicating that pore pressure is no longer elevated in excess of hydrostatic pressure and the weight of debris is nearly all borne by the grain-to-grain contact . Thus, the motion of the water-debris mixture will mainly depend on the momentum transfer in channels with steep gradients. For the three groups of the tested samples with different MGSs (i.e., 2 mm, 20 mm and 40 mm), the differences between 𝐶𝑣𝑑 and 𝐶𝑣ℎ are 0.16, 0.19 and 0.18, respectively. The differences are related to the MGS of the sediments, and also dependent on the gradation and the mineral composition of fine-grained detritus and etc.

ACCEPTED MANUSCRIPT As reported by Johnson and Rodine (1984), for some debris flow deposits, the change of debris from a sticky and immobile substance to a fluid mud with a small increase in the weight of water, is actually a physical manifestation of the small difference between 𝐶𝑣𝑑 and 𝐶𝑣ℎ . Therefore, the difference between the two SVCs

T

can serve as an important indicator on the slurry forming ability of debris. More importantly, this indicator can be

flow. The indicator is named debris flow index, 𝐼𝑑 : 𝐼𝑑 = 𝐶𝑉𝑑 -𝐶𝑉ℎ

(13)

NU

5.3. Rate of dissipation of relative excess pore pressure

SC R

IP

widely applied to old debris flow deposit identification, and also to the risk assessment of a prospective debris

Fig. 14 shows that the larger the MGS of an experimental sample, the higher the rate of escape of pore water,

MA

i.e., the poorer the ability to hold pore water. The increase in MGS causes the increase in average intergranular pore size, thereby expanding the flow paths of pore water. Hence, this is another factor which causes excess pore

D

pressure to dissipate rapidly. Generally, R values of the tested slurries with MGS of 2 mm are much lower than

TE

those with MGS of 20 mm and 40 mm, which indicates that MGS has a great impact on relative excess pore pressure. The coarse particles settle more easily, leading to an increase in the grain-to-grain contact area and it is

CE P

like a domino effect. Thus, the weight of the solid particles previously supported by pore water is transferred to the particle skeleton, resulting in a relatively rapid decrease in the excess pore pressure. For the sediments with the same MGS, R values are affected by the gradation characteristics, especially the

AC

intrinsic factor such as mineral composition of fine-grained debris. The maintenance of excess pore pressure (or called the maintenance of debris flows) plays an important role in the long runout travel of high-density debris flows over low gradient gullies. R values are the key for the research into the mechanism of debris flow travel, and also an important parameter for the risk assessment of debris flows. In practical applications, The Ta Tp in the formula (10) used to calculate R values can be 0.5 h, 1 h or 2 h. To facilitate the comparison between different debris slurries, R values should be measured at the optimum SVC, 𝐶𝑣𝑜 . 5.4. Factors affecting excess pore pressure 5.4.1. Grain size distribution The build-up of the excess pore pressure in the tested materials is mainly related to the gradation. Table 1 shows the gradation parameters of the three materials with different MGSs and it also presents the gradation

ACCEPTED MANUSCRIPT parameters of the debris flow deposit with MSG < 256 mm. It can be seen that they all are extremely poorly sorted (sufficiently well graded in the parlance of engineering). In addition, as stated in section 2.1, the DYG debris flow deposit contains metric-size megablocks that

T

appear to be spatially distributed at random in the finer-grained matrix. These megablocks form the large-scale

IP

phenocrysts, which is a widespread phenomenon in massive debris flows (Rodine and Johnson, 1976; Fannin and

SC R

Wise, 2001; Godt and Coe, 2007). The effect of such megablocks on excess pore pressure of debris flows has been also investigated.

During an experiment to measure pore pressure that has been carried out with the sediment-water mixture

NU

with MGS of 40 mm and SVC of 0.70, two “megablocks” with an equivalent diameter of 70 mm were slowly placed on the slurry surface when the experimental observation time reached 200 minutes and the pore pressure

MA

was decreasing. Immediately, the water tables in the piezometric tubes rose quickly while the “megablocks” remained on the surface. Before loading the “megablocks”, the maximum relative excess pore pressure at the

D

depth of 18 cm was 0.959. After another 29 minutes, the relative excess pore pressure reached a new maximum, 1.011, increasing by 5.42%. Even when the test ended, the relative excess pore pressure remained at a high level

TE

of 0.766 (Fig. 16) and the settlement of the “megablocks” did not occur. The result showed that the floating

CE P

(suspended) boulders in debris flows can elevate excess pore pressure significantly, thereby reducing yield stress

AC

and viscosity coefficient of debris flows.

Fig. 16. The response of relative excess pore pressure to “megablocks”.

5.4.2. Mineral composition of clay particles Many researchers concluded that the fine fractions in debris flow deposits play an important role in the long runout travel of debris flows (Wang and Sassa, 2003; Kokelaar, 2014; Kaitna et al., 2016). The experimental results in this study also support this conclusion. Among the three tested materials whose MGSs are 2 mm, 20 mm

ACCEPTED MANUSCRIPT and 40 mm, respectively, fine fractions < 0.04 mm (D'Agostino et al., 2010) acc….ount for 28.14 wt.%, 17.43 wt.% and 18.34 wt.%, respectively (Fig. 4b). When Ta Tp = 240 min, the minimum rates of dissipation of the relative excess pore pressure are 0.00022 min-1 (Fig. 11j), 0.00115 min-1 (Fig. 11k) and 0.00081 min-1 (Fig. 11l),

T

respectively. This means that the sediment-water mixtures with a larger fraction of fines have a stronger ability to

IP

hold pore water.

SC R

The earlier studies showed that the slurrying of debris was mainly controlled by the typical clay minerals including kaolinite, montmorillonite and illite (Rodine and Johnson, 1976; Jakob, 2005; Kaitna et al., 2007; Remaître et al., 2011). The test materials sampled from the DYG debris flow deposit, however, do not contain

NU

these typical clay minerals and chlorite content is also very low (Fig. 5b). Therefore, presumably, the layer-lattice silicate — mica that is not classified as typical clay mineral (Tang, 2000; Das, 2008) and accounts for 30 wt.% of

MA

clay-size (<0.005 mm) fraction played a key role in the slurrying process of the DYG debris flow. In the alpine environments, the chemical weathering of bedrock is weak and the conditions of forming typical clay minerals

D

could not occur (Sterling and Slaymaker, 2007). Mica is often abundant, however, in the unconsolidated deposit in

TE

alpine mountain catchments and can help these deposits to turn into debris flows. This is of great significance for the assessment of debris flow hazards.

CE P

5.4.3. Small woody debris

As mentioned in section 2.4, small woody debris was abundant in the DYG debris flow deposit along its

AC

downstream path. Using potassium bichromate titrimetric method, we found that the average content of small woody debris within the deposit (MGS = 2 mm) was 1.02 wt.% and the maximum content could reach 1.58 wt.%. Similarly, the average content of small woody debris within the deposit (MGS = 0.075 mm) could even reach up to 1.74%. The presence of small woody debris may have a positive effect on the development and maintenance of excess pore pressure. To further validate the conclusion, the micro-structural characteristics, including the arrangement mode, pore morphology and granular distribution pattern, need to be investigated. Small woody debris specimens were observed under different magnifications with a scanning electron microscopy (SEM). According to the SEM images of the small woody debris, shown in Fig. 17a-h, it can be seen that the surface of the small woody debris is rough and a variety of structures exist including fibrous, honeycomb, mesh and layered structures. Some layered and granular mineral debris was adhered to the surface of the small woody debris and the MGS of a single particle that can be absorbed is up to 50 µm.

ACCEPTED MANUSCRIPT (b)

(c)

(e)

(f)

(g)

(d)

(h)

NU

SC R

IP

T

(a)

Fig. 17. SEM images of the morphology of the small woody debris and the fine particles absorbed on its surface under different

MA

magnifying rates.

The density of fresh woody debris is generally about 0.6 g/cm3. By measurements and calculations, the average density of the small woody debris selected from the DYG debris flow deposit is only 1.085 g/cm3 which

D

is less than the density of the mineral and rock debris, although it adsorbs fine debris that is difficult to remove.

TE

The rough surface (means larger specific surface area) and low density allow the small woody debris to adsorb

CE P

fine particles, especially the granular mineral particles in the debris-water mixture, and suspend them, thereby contributing to the build-up and maintenance of excess pore pressure.

AC

5.5. The high mobility and maintaining mechanisms of the DYG debris flow The DYG debris flow traveled 3.3 km on a low gradient channel. In particular, although the snout moving dam stopped and formed a debris flow dam at about 500 m upstream of the fan apex, this did not cause detrital deposition. Since then, the short-lived dam was breached and the debris flow was remobilized, which buried the DYG Village and once dammed the Nu River. The high mobility, especially strong persistence of the DYG debris flow, was mainly accredited to the generation and maintenance of excess pore pressure in the interior of the debris flow. First and foremost, the strong capability of the DYG debris flow in holding pore water and maintaining excess pore pressure was governed by its poor sorting which made the pore necks tortuous and very narrow. The DYG debris flow deposit does not contain typical clay minerals and the chlorite content is also very low, but it contains mica, especially muscovite which is abundant. In the deposit with the MGS of 2 mm, the volume ratio of muscovite can be identified by naked eyes is about 3 %. Fine muscovite is important for slurry formation. Further,

ACCEPTED MANUSCRIPT the small woody debris with the MGS < 2 mm was found to be widely distributed in the debris flow deposit. Based on the observations, measurements and analyses, small woody debris with the low density, complex structures and large specific surface area, can also contribute to the slurrying and maintenance of the debris flow.

IP

abundant small woody debris, which can extend the mobility of the debris flow.

T

This is a significant finding because the debris flows occurring in well vegetated mountainous areas contain often

SC R

6. Conclusions

Taking the DYG debris flow deposit as the research object, the debris flow bodies have been reconstituted in

NU

the laboratory. To study the build-up, maintenance and dissipation of excess pore pressure, the experiments to measure pore pressure and escape of pore water have been carried out with the sediment-water mixtures with

MA

different maximum grain sizes (MGSs). According to the observations and measurements from these experiments, the following conclusions can be drawn.

(1) The slurrying of the water-debris mixtures is limited by solid volumetric concentration (SVC). The lower

D

SVC limit, 𝐶𝑣ℎ , is the SVC boundary between a hyperconcentrated streamflow and a debris flow. The upper SVC

TE

limit, 𝐶𝑣𝑑 , corresponds to the SVC boundary between a debris flow and a granular flow. The difference between

CE P

𝐶𝑣𝑑 and 𝐶𝑣ℎ (known as debris flow index, 𝐼𝑑 ) can reflect the slurry forming ability of debris. It can be applied to the studies of old debris flows, and also to the risk assessment of prospective debris flows. (2) Peak value (𝐾𝑝 ) and rate of dissipation (R) of relative excess pore pressure for a certain kind of debris are

AC

dependent on SVC. 𝐾𝑝 increases with SVC. With increasing SVC, R initially decreases, and then subsequently increases. Hence, an optimum SVC (𝐶𝑣𝑜 ) exists that allows the rate of dissipation of excess pore pressure to be minimized, thereby facilitating the long run-out travel of debris flows. (3) For different sediments with the same MGS, R value is controlled by the sorting and composition of debris. When applied to the risk assessment of debris flows, R value should be measured at 𝐶𝑣𝑜 , and the interval time of observation after the peak moment can be 0.5 h, 1 h or 2 h. (4) Not all sediments can be reconstituted into an integrated homogeneous slurry. Poor sorting is a prerequisite. The layer-lattice silicates can be the typical clay minerals including kaolinite, montmorillonite and illite, and also the unrepresentative clay minerals such as muscovite and chlorite. Small woody debris can also contribute to the slurrying of sediments and maintenance of debris flows in well vegetated mountainous areas and the boulders entrained in debris flows can elevate excess pore pressure and extend debris-flow mobility.

ACCEPTED MANUSCRIPT (5) The long runout distance of the DYG debris flow over the low gradient channel is related to the internal high excess pore pressure and its strong persistence capability. Acknowledgements

T

This work was supported by the National Science Foundation of China-Yunnan Joint Fund (U1502232,

IP

U1033601) and Research Fund for the Doctoral Program of Higher Education of China (20135314110005). We

SC R

are grateful to the editors and the anonymous reviewers for comments that considerably improved the paper. References

NU

Ala, S., Mathewson, C.C., 2015. Structural control of ground-water induced debris flows. Hydraulics/Hydrology of Arid Lands (H²AL). ASCE, 590-595.

Benda, L.E., Cundy, T.W., 1990. Predicting deposition of debris flows in mountain channels. Can. Geotech. J. 27(4), 409-417.

MA

https://doi.org/10.1139/t90-057

Calvo, L., Haddad, B., Pastor, M., Palacios, D., 2015. Runout and deposit morphology of bingham fluid as a function of initial

D

volume: implication for debris flow modelling. Nat. Hazards 75(1), 489-513. DOI: 10.1007/s11069-014-1334-x Chen, N.S., Zhou, W., Yang, C.L., Hu, G.S., Gao, Y.C., Han, D., 2010. The processes and mechanism of failure and debris flow

TE

initiation for gravel soil with different clay content. Geomorphology, 121(3–4), 222-230. DOI: 10.1016/j.geomorph.2010.04.017 Chen, X., Cui, P., You, Y., Chen, J., Li, D., 2015. Engineering measures for debris flow hazard mitigation in the wenchuan

CE P

earthquake area. Eng. Geol. 194, 73-85. http://dx.doi.org/10.1016/j.enggeo.2014.10.002 Costa, J. E., 1988. Rheologic, geomorphic, and sedimentologic differentiation of water flood, hyperconcentrated flows, and debris flows. In: Baker, V.R., Patton, P.C. (Ed.), Flood Geomorphology. John Wiley, New York, pp. 113–122.

AC

D'Agostino, V., Cesca, M., Marchi, L., 2010. Field and laboratory investigations of runout distances of debris flows in the dolomites (eastern italian alps). Geomorphology. 115(3–4), 294-304. https://doi.org/10.1016/j.geomorph.2009.06.032 Das, B.M., 2008. Advanced Soil Mechanics (3rd edition). New York,Taylor & Francis, pp. 1-6. De Blasio, F.V., Breien, H., Elverhi, A., 2011. Modelling a cohesive-frictional debris flow: an experimental, theoretical, and field-based study. Earth Surf. Process. Landf. 36(6), 753-766. DOI: 10.1002/esp.2101 De Haas, T., Ventra, D., Hauber, E., Conway, S.J., Kleinhans, M.G., 2015. Sedimentological analyses of Martian gullies: the subsurface as the key to the surface. European Planetary Science Congress. 258, 92–108. Fan, L., Lehmann, P., Mcardell, B., Or, D., 2017. Linking rainfall-induced landslides with debris flows runout patterns towards catchment scale hazard assessment. Geomorphology. 280, 1-15. http://dx.doi.org/10.1016/j.geomorph.2016.10.007 Fannin, R. J., Wise, M. P., 2001. An empirical-statistical model for debris flow travel distance. Can. Geotech. J. 38(5), 982-994. DOI: 10.1139/cgj-38-5-982 Faustini, J.M., Jones, J.A., 2003. Influence of large woody debris on channel morphology and dynamics in steep, boulder-r ich mountain streams, western cascades, oregon. Geomorphology 51(1–3), 187-205. https://doi.org/10.1016/S0169-555X(0 2)00336-7

ACCEPTED MANUSCRIPT François, B., Anne, M., Gladys, N. R., 2016. A two-phase two-layer model for fluidized granular flows with dilatancy effects. J. Fluid Mech. 801, 166-221. DOI:10.1017/jfm.2016.417 Gao, D.Z., Yuan, J.Y., 2001. Soil properties and soil mechanics, Beijing, pp. 13-19 (in Chinese). Godt, J.W., Coe, J.A., 2007. Alpine debris flows triggered by a 28 July 1999 thunderstorm in the central Front Range, Colorado.

T

Geomorphology 84, 80-97. https://doi.org/10.1016/j.geomorph.2006.07.009

IP

Gregoretti, C., Degetto, M., Boreggio, M., 2016. Gis-based cell model for simulating debris flow runout on a fan. J. Hydrol. 534,

SC R

326-340. http://dx.doi.org/10.1016/j.jhydrol.2015.12.054

Guo, X.G., Peng, C., Wu, W., Wang, Y.Q., 2016. A hypoplastic constitutive model for debris materials. Acta Geotechnica, Acta. Geotec. 11(6), 1-13. DOI: 10.1007/s11440-016-0494-0

Hampton, M.A., 1979. Buoyancy in debris flows. J. Sediment. Res. 49(3), 753-758. DOI: 10.1306/212F7838-2B24-11D7-86

NU

48000102C1865D

He, S.T., Wang D.J., Chen S., Zhang S.J., Chang S.Q., 2016. Natural consolidation characteristics of viscous debris flow deposition.

MA

J. Mt. Sci. 13(10), 1723-1734. DOI:10.1007/s11629-016-3850-5

Iverson, R.M., 1997. The physics of debris flow. Rev. Geophys. 35(3), 245-296. Iverson, R.M., 2003. The debris-flow rheology myth. In: Rickenmann, D., Chen C.L., (Ed.), Proceedings of the Third International

D

Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction, and Assessment, pp. 303–314.

TE

Iverson, R.M., 2005. Regulation of landslide motion by dilatancy and pore pressure feedback. J. Geophys. Res. Earth Surf. 110 (F2), 273-280. DOI:10.1029/2004JF000268

CE P

Iverson, R.M., Denlinger, R.P., LaHusen, R.G., Logan, M., 2000. Two-phase debris-flow across 3-D terrain: model predictions and experimental tests. In: Wieczorek, G.F., Naeser, N.D. (Ed.), Debris-Flow Hazards Mitigation: Mechanics, Prediction, and Assessment: Proceedings Second International Conference on Debris-flow Hazards Mitigation, Taipei, Taiwan, pp. 521-529.

AC

Iverson, R.M., George, D.L., 2014. A depth-averaged debris-flow model that includes the effects of evolving dilatancy. i. physical basis. Proceedings of the Royal Society A Mathematical Physical & Engineering Sciences, 470(2170), 20130819-20130819. Iverson, R.M., Matthew, L., Lahusen, R.G., Matteo, B., 2010. The perfect debris flow? Aggregated results from 28 large-scale experiments. J. Geophys. Res. Atmos. 115(F3), 438-454. DOI: 10.1029/2009JF001514 Iverson, R.M., Reid M.E., LaHusen, R.G., 1997. Debris-flow mobilization from landslides. Annu. Rev. Earth Planet Sci. 25:85–138 Jackson, C.R., Sturm, C.A., 2002. Woody debris and channel morphology in first- and second-order forested channels in Washington's coast ranges. Water Resour. Res. 38(9), 16-1–16-14. DOI: 10.1029/2001WR001138, 2002 Jakob, M., 2005. A size classification for debris flows. Eng. Geol. 79(3), 151-161. https://doi.org/10.1016/j.enggeo.2005.01.006 Johnson, A.M., Rodine, J.R., 1984. Debris flow. In: Brunsden, D., Prior, D.B., (Ed.), Slope Instability. Wiley, Chichester, pp. 257-361. Kaitna, R., Palucis, M.M., Hill, K.M., Dietrich, W.E., 2016. Effects of coarse grain size distribution and fine particle content on pore fluid pressure and shear behavior in experimental debris flows. J. Geophys. Res. Earth Surf. 121(2), n/a-n/a. DOI:10.10 02/2015JF003725 Kaitna, R., Rickenmann, D., Schatzmann, M., 2007. Experimental study on the rheologic behaviour of debris flow material. Acta.

ACCEPTED MANUSCRIPT Geotec.2, 71-85. DOI: 10.1007/s11440-007-0026-z Klubertanz, G., Laloui, L., Vulliet, L., 2009. Identification of mechanisms for landslide type initiation of debris flows. Engineering Geology, 109(1–2), 114-123.

granular flows. Earth Plant. Sci. Lett. 385(1), 172-180. DOI: 10.1016/j.epsl.2013.10.043

T

Kokelaar, B.P., Graham, R.L., Gray, J.M.N.T., Vallance, J.W., 2014. Fine-grained linings of leveed channels facilitate runout of

IP

Lavigne, F., Suwa, H., 2004. Contrasts between debris flows, hyperconcentrated flows and stream flows at a channel of Mount

SC R

Semeru, East Java, Indonesia. Geomorphology, 61(1–2), 41-58. https://doi.org/10.1016/j.geomorph.2003.11.005 Li, Y., Liu, J.J., Su, F.H., Xie, J., Wang, B.L., 2015. Relationship between grain composition and debris flow characteristics: a case study of the jiangjia gully in china. Landslides. 12(1), 19-28. DOI: 10.1007/s10346-014-0475-z Lundberg, J., Mcfarlane, D.A., 2007. Pleistocene depositional history in a periglacial terrane: a 500 k.y. record from Kents Cavern,

NU

Devon, United kingdom. Geosphere, 3(4), 199. DOI: 10.1130/GES00085.1

Luo, R.Z., Xu, Z.M., 2016. Basin morphology characteristics of the giant debris flow in the Dongyuege gully near the Nujiang River.

MA

Hydrogeology& Engineering Geology. Hydrol Eng. Geol. 43 (6), pp.142-146 (in Chinese). DOI: 10. 16030 / j. cnki. Issn. 1000-3665. 2016. 06. 22

Lu, X.B., Peng, C., Hu, K.H., Zhang, X.H., 2010. Initiation and development of water film by seepage. J. Mt. Sci. 7(4), 361-366.

D

DOI: 10.1007/s11629-010-2052-9

TE

Major, J.J., 2000. Gravity-driven consolidation of granular slurries: implications for debris-flow deposition and deposit characteristics. J. Sediment. Res. 70(1), 64-83. DOI: 10.1306/2DC408FF-0E47-11D7-8643000102C1865D

CE P

Marcus, W.A., Marston R.A., Jr, C.R.C., Gray R.D., 2002. Mapping the spatial and temporal distributions of woody debris in streams of the Greater Yellowstone Ecosystem, USA. Geomorphology. 44(3), 323-335. https://doi.org/10.1016/S016 9-555X(01)00181-7

AC

Morino, C., Conway, S.J., Balme, M.R., Jordan, C., Hillier, J., Sæmundsson, Þ., Argles, T., 2015. The comparison between two airborne LiDAR datasets to analyse debris flow initiation in north-western Iceland. EGU General Assembly Conference (Vol.17). EGU General Assembly Conference Abstracts. pp. 1. Ng, C.W.W., Choi, C.E., Song, D., Kwan, J.H.S., Koo, R.C.H., Shiu, H.Y.K., Ho, K.K., 2015. Physical modeling of baffles influence on landslide debris mobility. Landslides. 12(1), 1-18. DOI: 10.1007/s10346-014-0476-y Nocentini, M., Tofani, V., Gigli, G., Fidolini, F., Casagli, N., 2015. Modeling debris flows in volcanic terrains for hazard mapping: the case study of ischia island (Italy). Landslides. 12(5), 1-16. Okada, Y., Ochiai, H., 2008. Flow characteristics of 2-phase granular mass flows from model flume tests. Eng. Geol. 97(1–2), 1-14. Pellegrino, A.M., Santolo, A.S.D., Schippa, L., 2015. An integrated procedure to evaluate rheological parameters to model debris flows. Eng. Geol. 196, 88-98. DOI: 10.1016/j.enggeo.2015.07.002 Perinotto, H., Schneider, J., Bachèlery, P., Famin, V., Michon, L., 2015. The extreme mobility of debris avalanches: a new model of transport mechanism. J. Geol. Res. Solid. Earth. 120, n/a-n/a. DOI: 10.1002/2015JB011994 Pierson, T.C., 1981. Dominant particle support mechanisms in debris flows at Mt Thomas, New Zealand, and implications for flow mobility. Sedimentology. 28(1), 49-60. DOI: 10.1111/j.1365-3091.1981.tb01662.x

ACCEPTED MANUSCRIPT Pierson, T.C., Scott, K.M., 1985. Downstream dilution of a lahar: transition from debris flow to hyperconcentrated streamflow. Water Res. Res. 21(10),1511-1524. DOI: 10.1029/WR021i010p01511 Pudasaini, S.P., Wang, Y., Hutter, K., 2005. Modelling debris flows down general channels. Nat. Hazards Earth. Syst. Sci. 5(6), 799-819. https://doi.org/10.5194/nhess-5-799-2005

T

Qian, N., Wan Z.H., 1991. Sediment Moving Dynamics. Beijing, Science Press, pp.435-436 (in Chinese).

IP

Remaître, A., Malet, J.P., Maquaire, O., 2011. Geomorphology and kinematics of debris flows with high entrainment rates: a case

SC R

study in the South French Alps. Comptes. Rendus. Geosci. 343(11), 777-794. https://doi.org/10.1016/j.crte.2011.09.007 Rodine, J.D., Johnson, A.M., 1976. The ability of debris, heavily freighted with coarse clastic materials, to flow on gentle slopes. Sedimentology. 23(2), 213-234. https://doi.org/10.1111/j.1365-3091.1976.tb00047.x

Sa´nchez, M.E., Pastor, M., Romana, M.G., 2013. Modelling of short runout propagation landslides and debris flows. In: Georisk

NU

Assessment and Management of Risk for Engineered Systems and Geohazards. Taylorand Francis Group, New York. Sohn, Y.K., 2000. Coarse-grained debris-flow deposits in the miocene fan deltas, se korea: a scaling analysis. Sedimentary Geology.

MA

130(1–2), 45-64. https://doi.org/10.1016/S0037-0738(99)00099-8

Sterling, S., Slaymaker, O., 2007. Lithologic control of debris torrent occurrence. Geomorphology. 86(3), 307-319. https://d oi.org/10.1016/j.geomorph.2006.09.002

D

Takahashi, T., 1980. Debris flow on prismatic open channel. J. Hydraul. Eng. 106(3), 381-396.

TE

Takahashi, T., 2014. Debris flow: mechanics, prediction and countermeasures. CRC Press, Florida. DOI: 10.1201/9780203946282 Tang, B.X., 2000. China's debris flow. Commercial Press, Beijing (in Chinese).

CE P

Tang, C., Zhu, J., Chang, M., Ding, J., Qi, X. 2012. An empirical–statistical model for predicting debris-flow runout zones in the wenchuan earthquake area. Quat. Int. 250(250), 63-73. DOI:10.1016/j.quaint.2010.11.020 Tang, D.X., Liu, Y.R., Zhang, W.S., Wang, Q. 1999. Engineering Geology, Beijing, pp.13-19 (in Chinese).

6-17.

AC

Terzaghi, K., Peck, R.B., Mesri, G., 1996. Soil Mechanics in Engineering Practice (3rd edition). John Wiley & Sons, New Jersey. pp.

Wang, G., Sassa, K., 2003. Pore-pressure generation and movement of rainfall-induced landslides: effects of grain size and fine-particle content. Eng. Geol. 69(1–2), 109-125. https://doi.org/10.1016/S0013-7952(02)00268-5 Wang, Y.Y., Fei, X.J., 1999. Particle support mechanism in viscous debris flows at Jiangjia Ravine, Yunnan, China. Sci. China. Technol. Sci. 29(4), 550-555. https://doi.org/10.1007/BF02917409 Wieczorek, G.F., 1987. Effect of rainfall intensity and duration on debris flows in the central Santa Cruz Mountains, California. In: Costa, J.E, Wieczorek, G.F (Ed.) Debris flows/avalanches: process, recognition, and mitigation. Geological Society of America, Colorado, pp. 93–104. DOI: 10.1130/REG7-p93 Xu, Z.K., Li, T.G., Wan, S.M., Nan, Q.Y., Li, A.C., Chang, F.M., Jiang, F.Q., Tang, Z., 2012. Evolution of east asian monsoon: clay mineral evidence in the western Philippine Sea over the past 700kyr. J. Asian Earth Sci. 60, 188-196. http://dx.doi.org/10.1016/j.jseaes.2012.08.018 Yang, H.J., Wei, F.Q., Hu, K.H., Wang, C.C., 2014. Determination of the suspension competence of debris flows based on particle size analysis. Int. J. Sediment Res. 29(1), 73-81. https://doi.org/10.1016/S1001-6279(14)60023-8

ACCEPTED MANUSCRIPT Yu, B., Ma, Y., Qi, X., 2013. Experimental study on the influence of clay minerals on the yield stress of debris flows. J. Hydraul. Eng. 139(4), 364-373. DOI: 10.1061/(ASCE)HY.1943-7900.0000679. Jiang. Y., 2016. The Basic Characteristics and Occurrence Mechanism of Extra Large Viscous Debris Flow in Donyuege of Nu River. MA Thesis. Kunming University of Science and Technology, Yunnan, China. pp. 23 (in Chinese).

T

Zhang, S., Zhang, L.M., 2016. Impact of the 2008 Wenchuan earthquake in china on subsequent long-term debris flow activities in

IP

the epicentral area. Geomorphology. 276, 86-103. http://dx.doi.org/10.1016/j.geomorph.2016.10.009

SC R

Zhang, W., Chen, J.P., Wang, Q., Yan, J.Y., Guan, S., Chen, X.Y., Wang, F.F., 2015. Velocity and runout determination of a debris flow based on energy conservation: the Dongwopu debris flow in Tianjin, China. Q. J. Eng. Geol. Hydrogeol. 48(1), 5-14. DOI: 10.1144/qjegh2014-036

Zhou, G., Ouyang, C., 2015. Dimensional analysis of natural debris flows. EGU General Assembly Conference (Vol.17). EGU

AC

CE P

TE

D

MA

NU

General Assembly Conference Abstracts.

ACCEPTED MANUSCRIPT Highlights  Donyuege (DYG) debris flow traveled 3.3 km over a gentle-slope channel (3-5°).  Excess pore pressure in DYG debris flow deposits was measured.

T

 Concentration affects dissipation rate (R) of excess pore pressure.

IP

 The optimum concentration allows R to be minimized, facilitating the mobility.

AC

CE P

TE

D

MA

NU

SC R

Debris flow index is brought forward to reflect the slurry forming ability of debris.