Deformation Forecasting of Huangtupo Riverside Landslide in the Case of Frequent Microseisms

Journal of Earth Science, Vol. 27, No. 1, p. 160 166, February 2016 ISSN 1674-487X Printed in China DOI: 10.1007/s12583-016-0617-4 Deformation Forecasting of Huangtupo Riverside Landslide in the Case of Frequent Microseisms Jiwei Jiang* 1, Wei Xiang 2, Wei Zhang 1, Jiajun Pan 1 1. Key Laboratory of Geotechnical Mechanics and Engineering of Ministry of Water Resources, Yangtze River Scientific Research Institute, Wuhan 430010, China 2. Faculty of Engineering, China University of Geosciences, Wuhan 430074, China ABSTRACT: Ever since the impoundment of Three Gorges Reservoir (TGR), the seismicity in head region of TGR has increased significantly. Coupled with wide fluctuation of water level each year, it becomes more important to study the deformation forecasting of landslides beside TGR. As a famous active landslide beside TGR, Huangtupo riverside landslide is selected for a case study. Based on long term water level fluctuation and seismic monitoring, three typical adverse conditions are determined. With the established 3D numerical landslide model, seepage-dynamic coupling calculation is conducted under the seismic intensity of V degree. Results are as follows: 1. the dynamic water pressure formed by water level fluctuation will intensify the deformation of landslide; 2. under seismic load, the dynamic hysteresis is significant in defective geological bodies, such as weak layer and slip zone soil, because of much higher damping ratios, the seismic accelerate would be amplified in these elements; 3. microseisms are not intense enough to cause the landslide instability suddenly, but long term deformation accumulation effect of landslide should be paid more attention; 4. in numerical simulation, the factors of unbalance force and excess pore pressure also can be used in forecasting deformation tendency of landslide. KEY WORDS: landslide, frequent microseisms, deformation forecasting, multi-field coupling. 0 INTRODUCTION Since the impoundment of Three Gorges Reservoir (TGR) began from 2003, the seismicity has increased more than 10 times, especially in head region of the reservoir (from Three Gorges Dam site to Badong County) (Han and Rao, 2004). The TGR area is a geological disasters-prone area in China. Whether frequent microseisms threaten the stability of landslide is worthy of in-depth research. The reservoir-induced seismicity (RIS) has been acknowledged to be related with the fluctuation of water level, hydrogeological conditions, as well as the scale of reservoir (Baisch et al., 2010; Durá-Gómez and Talwani, 2010; Wu, 1981; Simpson, 1976). In China, for the problems of RIS in TGR, Wang et al. (2009) got the conclusion that the possible highest seismic magnitude can reach to 5.0 (M5.0); based on engineering geological analogy, the seismic intensity of RIS in TGR can reach to VI degree (the China Seismic Intensity Scale is classified into 12 degrees, from I for insensible to XII for landscape reshaping, refer to GB/T17742-2008 (China National Standardization Management Committee, 2009)), especially in Badong County (Li et al., 2005). Additional, Numerical simulation is verified effectively in studying the dynamic response of slope, spatially in *Corresponding author: jiangjw1023@163.com China University of Geosciences and Springer-Verlag Berlin Heidelberg 2016 Manuscript received April 9, 2014. Manuscript accepted September 5, 2014. forecasting the stability and deformation tender (Ni et al., 2014). In this paper, Huangtupo Landslide, which is located in Badong County beside TGR is selected as the research object, the RIS and water level fluctuation data in TGR are collected and analyzed firstly, and then the 3D multi-field coupling numerical simulation is carried out to research the deformation mechanism and dynamic response under the conditions of frequent microseisms. 1 THE CHARACTERISTICS OF RIS IN TGR The incomplete statistics shows at least 71 times of RIS, of which the seismic magnitude are greater than M2.0 occurred in head region of TGR in the last 5 years, and the characteristics can be obtained as follows. 1. The earthquake epicenter is concentrated in some areas, such as Badong County, where Huangtupo Landslide is located (31 2 17 N, 110 22 59 E). 2. The epicenter is shallow, basically within 10 km, even less than 5 km. Although the magnitude of some RIS is only M2.0 to M3.0, the maximum seismic intensity can reach to V degree. 3. From the monitored water level fluctuation data and seismic occurrence time (Fig. 1), we can conclude: (1) when the water level is high, the water load increases rapidly and the frequency of RIS becomes high; (2) because of the hysteresis of unloading effect, the frequency of RIS is high during rapid water level decrease and keeps stable at lower water level; (3) during the flood season of Yangtze River, the dispatching of Jiang, J. W., Xiang, W., Zhang, W., et al., 2016. Deformation Forecasting of Huangtupo Riverside Landslide in the Case of Frequent Microseisms. Journal of Earth Science, 27(1): 160 166. doi:10.1007/s12583-016-0617-4. http://en.earth-science.net

Deformation Forecasting of Huangtupo Riverside Landslide in the Case of Frequent Microseisms 161 TGR causes the water level fluctuate violently and the frequency of RIS high. These three typical working conditions will be taken into account in the following multi-field coupling numerical simulation. 2 ESTABLISHING 3D NUMERICAL MODEL OF HUANGTUPO LANDSLIDE The authors measured 4 engineering geological profiles for Huangtupo Landslide (Jiang, 2012). One of them is shown in Fig. 2. Based on these geological profiles, the 3D numerical model is established by FLAC 3D (developed by ITASCA Company). As Fig. 3 shows, the numerical geological model consists of 6 816 grid points and 30 934 elements, and it is separated by 8 lithological classes. In multi-field coupling numerical calculation, the mechanics calculation uses Mohr-Coulomb model, seepage calculation uses porous media model, and the dynamic load is applied at the bottom of the model with the acceleration versus time data. The vertical deformation is fixed at the bottom of the model. Moreover, dynamic free-field boundary is applied around the model, which is a specialized dynamic boundary developed in FLAC 3D. The physico-mechanical parameters are obtained by indoor tests. Such as the shear strength for soil samples is achieved by consolidated-undrained triaxial test, and with the advanced dynamic triaxial apparatus by GDS Company, the dynamic parameters are obtained. All necessary calculation parameters are listed in Table 1. 3 CALCULATION SPECIFICATIONS So far, the highest scale of seismic intensity is probably between IV to V degrees, and the corresponding peak acceleration is 0.22 0.44 m/s 2. So, in our research, the seismic load refers to the seismic intensity of V degree, which is a little higher than any RIS occurred in TGR. Meanwhile, according to the monitored data, each RIS generally lasts no longer than 5 s, but in a short period (such as in 2 d), the RIS happens 2 to 3 times continuously. So, in order to reflect the characteristics of frequency, a complete sequence of seismic wave composed of 3 separate microseisms (5 s for each) is determined. Generally, the period of seismic wave is between 0.2 and 1.0 s. So, the seismic wave with the frequency greater than 5 Hz should be filtered out; and after the necessary correction, the seismic wave curves of a complete sequence are obtained (Fig. 4). As mentioned above, the regulars of RIS have been 180 175 170 Water level fluctuation curve Seismogenic time points Water level (m) 165 160 155 150 145 140 2008-08-01 2009-02-17 2008-09-05 2010-03-24 2010-10-10 2011-04-28 Date (yyyy-mm-dd) Figure 1. The monitored water level fluctuation data and seismic occurrence time (from Aug. 1, 2008 to Jun. 7, 2011). Table 1 Physico-mechanical parameters of rock and soil in Huangtupo Landslide No. Density Poisson Cohesion (kpa) φ ( ) E d (MPa) λ K (m/d) 1 2 100 0.32 80 22 200 0.12 15 2 1 750 0.36 35 17.2 46 0.21 2 3 1 950 0.36 24 16.2 57 0.28 2 4 2 050 0.33 45 18.5 82 0.23 2 5 2 100 0.28 60 28.0 340 0.15 10 6 2 550 0.25 550 32.0 1 600 0.12 12 7 2 500 0.30 400 28.0 1 300 0.12 15 8 2 650 0.20 1 200 40.0 3 200 0.05 --- φ. Friction angle; E d. dynamic elastic modulus; λ. damping ratio; K. permeability coefficient.

162 Jiwei Jiang, Wei Xiang, Wei Zhang and Jiajun Pan 350 15º Altitude (m a.s.l.) 300 250 200 150 100 50 Bedrock Soft-rock Crackrock Cataclastic Gravel Slip zone and weak (argillization) layer Designed water level 175 m 0 100 200 300 400 500 600 700 800 900 1 000 Length (m) Figure 2. Typical geological profile of Huangtupo landslide. Figure 3. 3D numerical model of Huangtupo Landslide. Accelerate (m s ) -2 05. 03. 00. - 03. (a) Accelerate (m s ) -2 015. 010. 005. 0.00 005. - 010. (b) - 05. 000. 500. 10. 00 15. 00 Lasting time (s) - 015. 0 5 10 15 Lasting time (s) Figure 4. The corrected seismic wave curves of a complete sequence. (a) Corrected vertical acceleration versus time data; (b) corrected horizontal acceleration versus time data. summarized and can be simplified as the following three conditions: water level rises from 145 to 175 m in 30 days (145 175 m in 30 d for short) for Condition 1, 175 145 m in 60 d for Condition 2, and 145 165 145 m in 30 d for Condition 3. Based on initial steady seepage field, the transient seepage field and seepage displacement caused by water fluctuation can be achieved (Fig. 5). For Condition 1, when the water level rises rapidly, the additional pore pressure has not transmitted to deep part of landslide completely. The violent fluctuating of water level induces the horizontal seepage vector in upper part of landslide increases rapidly and the maximum cumulative displacement is about 4.33 cm (Fig. 5). For Condition 2, the violent drawdown makes the seepage

Deformation Forecasting of Huangtupo Riverside Landslide in the Case of Frequent Microseisms 163 Figure 5. The contour of pore pressure and landslide displacement under three seepage conditions. force much bigger than that in Condition 1. Refer to Fig. 5, the deformation extends to slip zone in deeper part, and the maximum cumulative displacement during this process is about 5.97 cm. For Condition 3, due to the hysteresis of seepage activity, the cushioning effect of seepage inhibits the increase of displacement during the process of violent fluctuation of water level, and the maximum cumulative displacement is only 4.81 cm, close to that in Condition 1. After seepage calculation, the seismic load showing in Fig. 4 can be applied. Figure 6 shows the displacement contour and deformation mechanism at the end of 15 s. The relative larger displacement occurs on the surface gravel. Refer to Figs. 2 and 3, the defective geological bodies (argillization interlayer and weak layer) play the decisive role in deformation of landslide. In the case of Condition 2, the displacement above argillization interlayer is 18 cm at least, whereas only 9 cm at most under it. On the right part of Fig. 6, the deformation mechanism of defective geological bodies is listed specially. It s clear that in the front part of argillization interlayer, many elements are undergoing the shear plastic deformation process (shear-n in Fig. 6); and at the back edge, the tension deformation mechanism (tension-n in Fig. 6) is obvious. It means the deformation of landslide is controlled by defective geological bodies, especially the argillization interlayer. Taking Condition 2 for example, 4 typical nodes located in argillization interlayer (Point 1), weak layer (Point 2), slip zone (Point 3) and on the surface (Point 4) separately (refer to Fig. 7) are monitored during the whole 15 s. Compared with original seismic wave, the hysteresis is obvious. For the nodes in argillization interlayer (Fig. 8a), weak layer (Fig. 8b) and slip zone (Fig. 8c), the average amplification factors of acceleration response are about 3, 2 and 2 times, respectively. The damping ratios for these defective geological bodies are much higher, they should adapt to the applied dynamic load, and more kinetic energy is absorbed. But for the node on surface and belongs to the group of cataclastic rock (Fig. 8d), shape of response curve can nearly meet the original seismic wave, and the amplifying effect is not so obvious. Still taking Condition 2 for example to illustrate the characteristic of displacement in defective geological bodies (Fig. 9), and distributions of typical nodes refer to Fig. 7, conclusion can be as follows.

Jiwei Jiang, Wei Xiang, Wei Zhang and Jiajun Pan 164 Figure 6. The contour of displacement and deformation mechanism at the end 15 seconds. shear-n means shear failure; tension-n means tension failure. Figure 7. The location sketch of four monitoring nodes. 1. If the seismic intensity reaches to V degree, according to the monitoring nodes, the overall deformation of landslide is significant, not only in shallow part, but also the slip zone in deep part. 2. In the initial stage, the increment of displacement is much bigger. The displacement keeps increasing later, but the increment becomes slighter. For the typical nodes in argillization interlayer and weak layer, the maximum cumulative displacement reaches 21.5 and 19.7 cm separately. This displacement should be a threat to the stability of gravel above argillization interlayer and weak layer. 3. Unbalance force is closely related with displacement. As Fig. 10 shows, the maximum unbalance force fluctuates violently from the 7th second, which means the frequent microseism is not equal to several separate seismic activities, and unbalance force dominates the tendency of displacement. But in the following time (after the 7th second), the fluctuation range of unbalance force keeps stable, which means that when the seismic intensity reaches to V degree, the landslide will not be in failure immediately. 4. The deformation of landslide is also related to excess pore pressure. The pore pressure of the above typical nodes are also monitored (Fig. 11). In the initial 5 s, the pore pressure increases slightly, but then, for the node in argillization interlayer (Point 1), the pore pressure increases rapidly, and finally increases by 110%; meanwhile, for the node in weak layer (Point 2), the pore pressure only increases by about 25%. Refer to Fig. 7, Point 1 locates in the deep part and far from exposure face, the dissipation of pore pressure is not easy; but due to Point 2 close to exposure face, the pore pressure can be dissipated, so it does not increase rapidly.

Deformation Forecasting of Huangtupo Riverside Landslide in the Case of Frequent Microseisms Figure 8. The comparison between applied seismic travel-time curve and the dynamic response curve (Condition 2). Figure 9. Displacement response curves of typical nodes under seismic action (Condition 2). Figure 10. Unbalance force response curve of Point 1 under seismic action (Condition 2). 165

166 Jiwei Jiang, Wei Xiang, Wei Zhang and Jiajun Pan Pore pressure/meter water column 50 40 30 20 10 0 0 Point 1 Point 2 5 10 15 Lasting time (s) Figure 11. Pore pressure response curves of typical nodes under seismic action (Condition 2). 4 CONCLUSIONS In this paper, Huangtupo Landslide is selected as research object to study the deformation mechanism and dynamic response in the case of frequent microseisms, coming to the following conclusions. 1. Water level fluctuation exerts adverse impacts on the deformation tendency of landslide, and the authors also have obtained the similar conclusions with another landslide beside TGR in former (Jiang et al., 2011), so the impact of seepage conditions can t be ignored. 2. The defective geological bodies have much higher damping ratios, seismic wave can be amplified and dynamic hysteresis is significant in these elements, so the defective geological bodies play the decisive role in the deformation of landslide. 3. Having undergone a complete sequence of RIS (15 s), the maximum accumulation displacement for the landslide is 21.7 cm. Although the frequent microseisms will not cause the landslide failure suddenly, more attention should be paid to the long term deformation accumulation of landslide. 4. Violent fluctuation of unbalance force and increasing in excess pore pressure also can be used in indicating the deformation tendency of the landslide. Finally, the accumulative effect should be a key problem that threatens the stability of landslide, so the deformation monitoring and risk prediction for typical landslides beside TGR are necessary. ACKNOWLEDGMENTS This study was financially supported by the National Natural Science Foundation of China (Nos. 51409011 and 51309029), and the Basic Scientific Research Operating Expenses of Central-Level Public Academies and Institutes (Nos. CKSF2014057/YT and CKSF2015051/YT). The final publication is available at Springer via http://dx.doi.org/10.1007/s12583-016-0617-4. REFERENCES CITED Baisch, S., Vörös, R., Rothert, E., et al., 2010. A Numerical Model for Fluid Injection Induced Seismicity at Soultz-Sous-Forêts. International Journal of Rock Mechanics and Mining Sciences, 47(3): 405 413. doi:10.1016/j.ijrmms.2009.10.001 China National Standardization Management Committee, 2009. General Administration of Quality Supervision, Inspection and Quarantine of the People s Republic of China: GB/T 17742-2008. Standards Press of China, Beijing. 2 Durá-Gómez, I., Talwani, P., 2010. Reservoir-Induced Seismicity Associated with the Itoiz Reservoir, Spain: A Case Study. Geophysical Journal International, 181(1): 343 356. doi:10.1111/j.1365 246x.2009.04462.x Han, X. G., Rao, Y. Y., 2004. Analysis of Genesis of Reservoir- Induced Microquakes in Badong Reach of Three Gorges. Crustal Deformation and Earthquake, 24(2): 74 77 Jiang, J. W., 2012. Research on the Deformation Mechanism and Dynamic Response of Typical Landslides in Three Gorges Reservoir in Case of Frequent Microseisms: [Dissertation]. China University of Geosciences, Wuhan (in Chinese with English Abstract) Jiang, J. W., Xiang, W., Zhang, X. Y., 2011. Research on Mechanical Parameters of Intact Sliding Zone Soils of Huangtupo Landslide Based on CT Scanning and Simulation Tests. Chinese Journal of Rock Mechanics and Engineering, 30(5): 1025 1033 Li, P., Li, Y. J., Wang, M. E., 2005. A Study of Reservoir Induced Earthquake in the Three Gorges Area. Engineering Science, 7(6): 14 20 Ni, W. D., Tang, H. M., Liu, X., et al., 2014. Dynamic Stability Analysis of Wedge in Rock Slope Based on Kinetic Vector Method. Journal of Earth Science, 25(4): 749 756 Simpson, D. W., 1976. Seismicity Changes Associated with Reservoir Loading. Engineering Geology, 10(2 4): 123 150. doi:10.1016/0013-7952(76)90016-8 Wang, Q. L., Yao, Y. S., Xia, J. W., et al., 2009. Application of Statistical Forecasting Model to the Prediction of the Three Gorges Reservoir Induced Seismicity. Seismology and Geology, 31(2): 287 294 Wu, J. N., 1981. Rock Softening and Reservoir-Induced Seismicity. South China Journal of Seismology, 1(1): 84 95 (in Chinese with English Abstract)