The Evolution Characteristics of Different Deformation Modes of Shear Slip Surface Based on Acoustic Emission Measurements

: To investigate the acoustic emission (AE) precursor detection of landslide failures, a model test aiming 14 at reproducing the typical shear surface deformation of different landslide modes was designed. The evolution 15 characteristics of the AE signals were analyzed in terms of AE count, cumulative AE count, AE correlation 16 diagrams, and corresponding time-frequency properties. The test results show that for the progressive deformation 17 mode, the AE count experiences a low-level period, an active period and a rapid increase period, and the 18 distribution of correlation diagram hits concentrates in a relatively small scale and then gradually scatters. There is 19 low frequency signals firstly and then high frequency signals, and the energy proportion of the high-frequency 20 signals shows an increasing tendency. For the sudden deformation mode, the magnitude of AE count increases 21 sharply, leading to the cumulative AE count curve rises steeply, and correlation diagram hits distribution turns into 22 relatively scattering rapidly. Furthermore, the high frequency signals and high energy proportion appear much 23 earlier than that of the progressive deformation mode. For steady deformation mode, however, the acoustic 24 emission activity is quite active in the initial stage, the cumulative AE count curve rises sharply and then 25 maintains relatively flat trend, and correlation diagram hits distribution scatters firstly, then the signal hits 26 distribution begins to concentrate in a relatively small scale. There are intensive high-frequency hits and high 27 energy proportion earlier, and later they tend to decay in response to smaller magnitudes of movement. 28 Comprehensive use of multiple features can help identify landslide deformation patterns more accurately under 29 complex natural conditions, which may provide a promising reference for the field warning monitoring of the 30 diverse landslide failures. 31

conditions. Therefore, monitoring the physico-mechanical parameters and their related characteristics during the 58 slope deformation process is very important to judge the damage degree and the disaster mechanism of the 59 landslide. Acoustic emission (AE) has recently been recognized as the feasible technique to monitor landslides 60 movement, which could contain the rich information of the entire landslide evolution process. For decades, 61 researchers have used the measurement and quantification of acoustic emission generated by the sliding 62 deformation of soil slopes to continuously develop monitoring technology strategies (Koerner et al. 1981; 63 Chichibu et al. 1989;Dixon et al. 2003; Shiotani 2006, Smith et al. 2014). These methods usually use waveguides 64 to provide a low attenuation propagation path of acoustic emission signals. Later, researchers developed active 65 waveguide devices (Dixon et al. 2007; Smith et al. 2015), and abundant acoustic emission signals are generated by 66 the backfill material. When the slope is sliding, the soil acts on the backfill material, causing particle-particle 67 according to the creep test results of rock and soil, that is, the possible landslide types and behavior characteristics under different stress conditions. Considering the long-term strength and creep characteristics of rheological 112 mechanics, the possible mechanical state of the landslide during different evolution stages is analyzed. 113 During the landslide deformation process, the unbalanced sliding force is the essential cause of the landslide 114 evolution. The unbalanced sliding force refers to the difference value between the sliding force p F and the 115 resistance against sliding R F , which is recorded as Mainly, the evolution trend of geotechnical materials creep deformation is determined by the relationship 126 between the shear stress  and long-term strength  of the sliding band soil. 127 When the shear stress is slightly greater than long-term strength    , the slope will undergo continuous 128 gentle deformation and may undergo a deformation stage with low creep rate. As the deformation continues to 129 evolve, coupled with the influence of external factors such as rainfall, the sliding resistance force provided by the 130 sliding zone continues to decrease. Gradually, the sliding force is far greater than the anti-sliding force provided 131 by the sliding zone soil. Later, as the deformation process continuously evolves, it will enter the accelerating 132 deformation stage eventually, causing the slope body instability and damage, which shows the progressive 133 landslide mode throughout all the process. 134 If the slope is subjected to strong external factors abruptly, such as earthquake or heavy rainfall, the sliding 135 force may suddenly increase, also the anti-sliding force provided by the sliding band soil loses rapidly, probably 136 the sliding force is much greater than the anti-sliding force in a short time, i.e.     , thus the landslide 137 deformation rapidly increases and eventually evolves into an accelerating deformation stage. Landslide occurs 138 short-term movement with very high deformation rate, causing the slope body instability and damage in a quite 139 short time. The time from the deformation occurrence to the final slope instability is relatively short, and the 140 rapidly accelerating deformation stage appears without experiencing longer process of low deformation rate, and 141 sudden landslides case is prone to occur. larger deformation probably occurs early. However, with the gradual deformation evolution and the external effects attenuation or disappearance, the sliding force will decrease, when the sliding stress does not exceed the 145 long-term strength of the slide zone soil     , the landslide deformation mainly appears stable creep under this 146 stress condition, which will undergo long term tiny creep movement, it might maintain a stable state at last, so the 147 steady landslide case may appear. 148 3 Experimental system and test procedure 149

Experimental apparatus 150
The shear apparatus was designed to simulate the shear surface deformation of the soil slope sliding, which 151 comprised two steel boxes, each box with the dimensions 0.2×0.2×0.3 m. And the bottom box was fixed on the 152 heavy steel pedestal to prevent movement. For each test, the lubricant was applied on the two boxes interface to 153 reduce the friction force. The apparatus was filled with soil to represent the element of the soil slope. Besides, the 154 steel waveguide could be installed through the soil column with the granule backfills. 155 In the test, the servo loading system is CMT-5205, the electro-hydraulic servo test machine, equipped with 156 precision oil pump and electro-hydraulic servo valve to perform automatic control, allowing the more 157 experimental displacement variables could be controlled. The maximum test load is 300 kN, and the displacement 158 rate is adjustable within the range of 0.001～500 mm/min, which fully meets the experimental requirements. The 159 set program is used to maintain a specified displacement rate, then the upward loading device pulls the wire rope, 160 and the wire rope is horizontally connected to the top box to induce the shearing formation in the soil column, so 161 as to control the displacement behaviors of the shear sliding surface, as shown in Fig. 3. 162 163 Fig. 3 Experimental apparatus 164

Soil and backfills material 165
The 80 mm diameters hole was reserved by the PVC tube matched the size. An active waveguide, a 30 mm 166 dia., 1.0 m long steel tube with 5 mm wall thickness was installed in the hole with the columnar backfills. Then 167 the soil element model was prepared using the layered filling method. The filling height of each layer was 15 cm, 168 and the soil was uniformly compacted with a constant external force. and the specific particle parameters are 169 showed in the Table 1. 170  , among the series tests on the waveguide with different backfills (glass sand, marble gravel, river  173 sand), the glass sand backfills case shows remarkable variation characteristics for the cumulative AE count, which 174 is closely related to the sliding deformation process. Besides, there is prior AE detection sensitively when using 175 the glass sand backfills compared to the other cases during the initial stage with extremely slow magnitude 176 deformation rate. Moreover, the most evident quantitative correlational relationship between the cumulative AE 177 count and the shear deformation, and the priority AE detection, it is the potential application for detecting 178 pre-failure processes on soil slope, and providing timely information on the movement status of the various soil 179 slope. Subsequently, a series of displacement-controlled shear tests were conducted on active waveguide models 180 with glass sand granular backfills, to study the AE detection characteristics under different shear deformation 181 movements, and the specific particle parameters are showed in the Table 2. 182

AE measurement system 185
The DS2AE equipment was adopted, which can collect and display AE signal waveforms and parameters in 186 real time. The acoustic emission sensor is a Nano-30 resonant high-sensitivity sensor with a frequency range of 187 125~750 kHz. The acoustic emission sampling threshold value was 50 dB to effectively reduce the noise impact. 188 The waveform sampling rate was 1MSPS. The sensor was fixed by white tape, and the coupling agent was applied 189 between the sensor and the waveguide to reduce the signal attenuation. During the loading process, the 190 displacement-time information and acoustic emission waveforms were collected synchronously. 191

Test procedure 192
The displacement-time functions (Table 3~5) were designed to reproduce the evolution process of different 193 landslide modes, all tests were controlled by displacement rate. Through the control system of test machine, input 194 the displacement rate parameters at different moments, and the wire rope linked to the shear box will respond in 195 time, which can accurately control the deformation rate of the shear surface according to the given parameters 196 formulas. The displacement-time curve of the shear sliding surface was automatically controlled, as shown in the 197 Fig. 4. Thus, the testing process was stable and a smooth displacement-time curve could be obtained. After the 198 tertiary stage of operation, the testing program automatically terminated, and the shear deformation process was 199 completed. 200    an active period and a rapid increase period. In the initial stage, there is small magnitude of AE count relatively, 226 then the AE count is getting more active notably. When the shear displacement enters the later accelerating stage, 227 the acoustic emission bursts, thus the AE count increases sharply. For the sudden deformation mode, the evolution 228 process reaching to larger deformation is much shorter than that of progressive deformation mode. In the initial 229 deformation stage, the acoustic emission activities are getting more active quickly. In the rapid deformation stage 230 subsequently, the AE count increases sharply in a short time due to larger deformation rate, and the evolving 231 curve of cumulative AE count rises much steeply. For the steady deformation mode case, in the initial stage with 232 rapid deformation rate, significant deformation suddenly appears in this stage, the acoustic emission activity is 233 quite active, and the curve of cumulative AE counts rises sharply, which is distinctive to the previous two cases. 234 Subsequently, the deformation rate begins to decrease gradually, the number of the acoustic emission signals 235 decreases, and the curve of cumulative AE counts slows down gradually. In the last stage with a quite tiny 236 deformation rate, the acoustic emission activity maintains a slight low-level, and the curve of cumulative AE 237 counts is relatively flat trend, exhibiting a "quiet period". 238 Besides, the stage Ⅱ, 300-450s; (c) stage Ⅲ , 450-700s 294 Fig. 9 presents the correlation diagram of AE amplitude-energy in the progressive deformation mode. In the 295 initial stage, the acoustic emission activity maintains at a low level, the signal amplitude and energy are both at a 296 relatively low-level, and the signal hits distribution concentrates in a relatively small scale. As the deformation 297 rate increase, the acoustic emission signal has a larger magnitude of AE amplitude and energy. Relatively, the 298 amplitude-energy distribution range begins to gradually expand widely, with a large number of high-amplitude 299 and low-energy, low-amplitude and high-energy acoustic emission signals, and there is a noteworthy phenomenon 300 that the signal hits distribution has relatively scattered during final stage. stage Ⅱ 200-500s; (c) stage Ⅲ , 500-800s 317 Fig. 11 shows the correlation diagram of AE amplitude-energy in the steady deformation mode. In the early 318 stage with rapid deformation rate, and there is significant deformation instantly. In this stage, the acoustic 319 emission activity is quite active, and the acoustic emission signals have greater magnitude of energy and 320 amplitude. Relatively, the amplitude-energy has a large distribution range, whereas the signal hits distribution is 321 relatively scattered. Then the deformation rate gradually decreases, thus the acoustic emission signal amplitude 322 and energy value both gradually decrease, and the signal distribution scale begins to decreases gradually. In the 323 later stage with tiny deformation rate, the signal points distribution concentrates in a small scale. band has a small distribution range at the initial stage, and its frequency range is mainly between 25~150 kHz. 342 Gradually the number of the acoustic signals in the dominant frequency band increases. After entering the later 343 accelerating stage with rapid deformation rate, the signal intensity increases significantly, note that it clearly 344 exhibits the strong tendency of serial high-frequency band distribution appearing in 300~350 kHz. 345 For the sudden deformation mode, in the early deformation stage, the acoustic emission signal is dominated 346 by relatively low frequencies, with a narrow frequency domain. And its dominant frequency domain is also below 347 150 kHz, a small number of signal hits appear in the low dominant frequency bands. In a short time, as the 348 deformation rate increases rapidly, the acoustic emission events begin to occur frequency contents over a wide 349 range, and high-frequency signals appear significantly ranging from 300~350 kHz. 350 For the steady deformation case, the deformation rate is relatively large in the early stage, the deformation 351 value increases sharply, and the acoustic emission signal is particularly active with high signal intensity. At this 352 stage, some high-frequency hits with a dominant frequency into 300~350 kHz arise earlier. As the deformation 353 rate gradually decreases, also the number of high-frequency signal gradually decreases. In the later stage with a 354 tiny deformation rate, there is a noteworthy phenomenon only low-frequency signals exciting, and a relatively 355 small number of signal hits appear in this stage. During this tiny deformation stage, the frequency domain is From the serial tests, it's found that the dominant frequency presents evolution characteristics during 358 different shear deformation stages. When the deformation rate is slow, there are only relatively low frequency 359 signals under 150 kHz, but no high frequency signals. These acoustic emission signals emitted are probably 360 derived from the mutual compression of the backfill particles in this stage. As the deformation increase gradually, 361 the column of granular backfills also deforms more, and this deformation behavior induces relatively high 362 frequency levels of AE signal propagating along the waveguide. There is the mutual compression between the 363 particles, also with some frictional sliding of grains, which leads to an increase in the number of acoustic emission 364 events. When the deformation rate increases rapidly, namely getting into the accelerating deformation stage, not 365 only the low frequency and intermediate frequency signals appear more drastically, also coupling with the 366 continual appearance of high-frequency signals between the 300~350 kHz, which is quite distinct from that in low 367 deformation rate stages. Probably because the host soil deformation rate accelerates, the friction force between the 368 internal particles increases due to the high-confining stress, and thus generating intense interparticle friction. Even 369 the high frequency signals are partially derived from the amounts of backfill particle fracture by the greater 370 compact stress. As the consequence, these activities release the acoustic signal in the high dominant frequency. 371

Energy ratio analysis 372
From the above analysis, the frequency domain shows different characteristic with the shear deformation, 373 particularly the evolutionary trend of high-frequency signals in response to the deformation behavious. By 374 analyzing the trend of the energy proportion of high-frequency signals in different deformation stages, we further 375 discuss correlation between the deformation modes and the energy ratio evolution of acoustic emission signal. 376 Table 6 shows the reconstructed frequency bands signal decomposed up to the third layer after the wavelet packet 377 decomposition, totaling 8 frequency bands. 378  proportion significantly increases with the shear sliding deformation. In the initial period, the energy proportion 400 exhibits the relatively minimum value, not exceeding 5%; as the deformation rate increases, the high-frequency 401 energy ratios tend to increase, and the proportion value has a certain rise greater than 5%; During deformation 402 accelerating stage, the energy proportion of high-frequency signals increases significantly, and the energy 403 proportion is around 20%, or even higher. For the sudden deformation mode, the high-frequency signal energy 404 proportion increases significantly with the rapid increase of the deformation rate, and the magnitude of energy 405 proportion is around 20% close to the later accelerating stage. For the steady deformation mode, the deformation 406 rate is large in the rapid deformation stage early, and the energy proportion of high-frequency signal increases rapidly. However, as the deformation rate gradually decreases later, the energy proportion of high-frequency begins to decrease. At last the tiny deformation stage, the proportion of high-frequency energy is at a relatively 409 low-level. Noteworthily, the proportion of high-frequency energy has experienced a process of the rapid increase 410 firstly and then the gradual decrease in response to the deformation behavious. 411 It follows that the physical interaction variation between the backfills particles appears during the different 412 deformation stages, which will lead to the energy proportion evolution of the acoustic emission signal emitted. To 413 some extent, this statistical energy proportion can be used as identification and monitoring methods, to alert the 414 user that the soil slope has deformed in which condition, from a low displacement rate to a high magnitude, or 415 from a high rate then to a low rate, further to be used in landslide modes recognition and risk management, it is 416 critical to enhance predictability and early warning precision. The test results show that evolution characteristics of acoustic emission are difference in response to the 433 applied shear deformation modes. For the progressive mode, the low deformation rate maintains for greater time 434 magnitudes in early stages, thus the smaller AE count continues a long time until the accelerating deformation 435 stage, the number of AE count starts to increase rapidly. For the sudden mode, the AE count presents a rapid 436 growth trend in a short time, and the curve of cumulative AE count rises rapidly over time. For the steady mode, 437 the AE count increases sharply in initial stage, as the deformation rate gradually decreases, the AE count 438 gradually decreases, showing a distinction trend of rapid increase firstly and then gradual decrease. 439 In addition, for the signal hits distribution characteristic of AE duration-count and amplitude-energy 440 correlation, for the both of progressive mode and sudden mode, the AE hits distribution presents the evolution 441 process form the concentrated hits distribution in the small scale early to the scattered hits distribution in large 442 scale relatively in the later accelerating deformation stage, distinctively the evolution process experiences a quite 443 short time for the sudden mode. For the steady mode, however, the signal scattered distribution gradually evolves 444 into a concentrated small-scale distribution in response to applied deformation behavior over time. The 445 evolutionary transformation tendency from the small-range to large-range, or from large-range to small-range, 446 indicates that landslide movement is undergoing different deformation states. These results demonstrate that deformation movement states, this information can provide identification information and early detection of 449 deformation evolution behavious that soil slopes experience during landslide movement process. 450 Besides, Fig. 14~15 focus on the evolution process of the AE detection amplitude and duration parameters. 451 The AE amplitude and AE duration are integrated as an identification indicator to understand the overall 452 deformation behavior of slope movements, which reveals relevant evolution characteristics to different 453 deformation modes. It is clear that both the AE detection amplitude and duration evolutions experience a low-level period, an 461 active period and a rapid increase period in progressive deformation mode. For the sudden deformation mode, the 462 evolution time from a low magnitude to a high magnitude is often much shorter than that of a progressive 463 deformation mode. For the steady deformation mode, in the initial rapid deformation stage, the magnitude of both 464 AE amplitude and duration parameters is larger, as the deformation rate gradually decreases, the parameters 465 magnitude shows a decay trend, smaller than that in the previous stage. These AE detection parameters increase to 466 the relatively larger magnitude during the monitoring process, revealing that the landslide deformation rate is 467 getting larger, probably leading to the soil slope entering a dangerous period. 468 Overall, few studies have focused on the frequency domain characteristic relevant to deformation evolution 469 stages of landslide failures directly. By analyzing the frequency domain evolution characteristics of the AE signals 470 during the different deformation stages, the results demonstrate that the frequency domain characteristic is a good 471 discriminant indicator identifying the deformation stage, which has the distinct perception characteristic. 472 According to the frequency domain indicator of the AE detection signals, which is closely related to the shear 473 deformation modes, especially in the accelerating deformation stage, it may directly determine whether the soil 474 slope is during the rapid deformation stage. Frequency domain has the promising potential to become an efficient 475 early warning indicator for such failures. For instance, during the shear deformation stage with rapid deformation accelerating stage is quite distinct from the low deformation rate stage. When the shear deformation rate appears 479 larger magnitude, note that there is wide frequency band with high-frequency signal hits. Once the deformation 480 rate gradually decreases, the number of high-frequency signals gradually decrease, or even high-frequency signals 481 disappear correspondingly. 482 As the sliding soil mass begins to move, glass granular shearing takes place within the active waveguide at 483 the shear surface. Under the rapid deformation rate, as the active waveguide resists shear and bending, the backfill 484 deforms around the waveguide. The reactions from the host soil cause the pressures along the active waveguide to 485 increase. Reactions from both the host soil and the waveguide cause the confining pressures in the backfill to 486 increase, this behaviour causes the intensively interparticle friction, even the particles fracture, emitting the high 487 frequency AE signals. This finding can contribute to identify the rapid landslide deformation and further improve 488 slope earning management. Furthermore, it is noteworthy that the energy proportion of the high-frequency band in 489 the range of 312.5~500 kHz significant increases with the rapid shear deformation. For both progressive 490 deformation mode and sudden mode, the energy proportion of high-frequency signal shows an increasing trend, 491 and the energy proportion occupies to approximately 20% close to the late stage with rapid deformation rate. Just 492 the evolution process of sudden mode up to the high energy proportion is much faster than the progressive mode. 493 For the steady mode, however, the proportion of high-frequency energy has experienced the evolution trend with 494 rapid increase firstly and then gradual decrease. 495 In summary, different AE detection parameter indexes can be utilized for the landslides movement 496 identification and monitoring, in this study, the AE count, cumulative AE count, duration-count scattergram, 497 amplitude-energy scattergram, dominant frequency characteristic, energy proportion of high-frequency signals 498 overall process are explored to identify the evolution process of the different deformation modes. Specifically, we 499 summarize the evolution characteristics of AE detection parameters during different deformation modes, as shown 500 in Table 7. 501  For the general steep slope with loose soil or gravel soil subject to continuous long-term rainfall, if the 507 typical AE parameters, such as the count rate, amplitude, duration, present the trend from low level to the high 508 level gradually in a long time, and the correlation duration-count and amplitude-energy scattergrams both extend 509 to large scale gradually, notably when the relatively high frequency signals appear constantly, the progressive 510 landslide instability will possibly occur. 511 For high-steep slope with siltized intercalations underneath the sliding mass, impacted by the triggering 512 conditions of slope toe excavation, heavy rainstorm or earthquake, once the magnitude of AE parameters all 513 increase from the low level to the high level rapidly in a short time during the process of monitoring, and the 514 correlation scattergrams show the tendency of scattered expansion rapidly, also the relatively high frequency 515 signals appear intensively with high energy proportion, then the slope failure is about to occur, presenting the 516 sudden landslide mode. 517 Conversely, for gentle terrain underneath the soft layer potentially affecting by fluctuation of reservoir water 518 level, when the AE count, amplitude, duration present the trend of a rapid increase firstly and then a gradual 519 decrease with the time, and these correlation diagram hits distribution scatters firstly and then the signal hits 520 distribution begins to concentrate, also the number of high-frequency signal gradually decreases or disappear, the 521 cumulative AE counts curve begins to maintain relatively flat trend, thus the possibility of the occurrence of 522 steady landslide mode will be high, the slope probably does not occur instable failure. 523 The instability process of soil slope is complicated and can be influenced by many factors, such as geological 524 conditions, hydrological conditions, geotechnical properties. In the natural field situation, just based on few AE 525 detection parameter characteristics, it is indeed difficult and unrealistic to judge the deformation stage of different 526 landslide movements, to accurately identify ultimate sliding deformation mode. Through the analysis processing 527 of continuous AE monitoring data, we can obtain important identifications reference by synthetic application of 528 these AE detection evolution characteristics, comprehensive use of multiple features can help to improve the 529 accuracy of landslide deformation stage, thus we can identify landslide deformation patterns more accurately, 530 which is significance to the early landslide warning. An integrated discriminant criterion based on multiple AE 531 features may be a promising method of guiding slope evolution modes monitoring. 532 Generally, the slope instability and collapse does not occur instantly, but through accumulation of unstable 533 influencing factors in a certain period, slope sliding movement experiences an evolutionary process from deformation evolution trend of the soil slope sliding in different evolution period is essential to identify landslide 536 modes and predict soil slope failure early. Considering the natural slopes with more complex geological 537 conditions, these AE detection evolution characteristics can be used as a guide for on-site monitoring methods. 538 Further research is required to apply these methods to field monitoring strategy, sending the monitoring 539 information to the relevant geological management department so that geologists can take relevant measures, thus 540 a sensitive, continuous, remote and real-time monitoring system pattern based on multiple AE features may be 541 developed in the future for the identification reference and early warning of different landslide modes. 542 Furthermore, combining with other monitoring parameters, such as displacement, rainfall, pore water pressure, 543 cracking and sliding velocity, a monitoring and early warning platform for landslide disasters can be established 544 based on multi-source heterogeneous monitoring data, to effectively improve the identification and early warning 545 accuracy for landslide geological disaster. 546

Conclusions 547
In this paper, a model test for reproducing the typical shear surface deformation was designed, the displacement, 548 AE data, AE correlation diagram, and the corresponding frequency domain characteristics were obtained through 549 experiments with different deformation modes. The evolutionary characteristics of AE detection and possible 550 precursors were analyzed, the primary conclusions are drawn: 551 (1) In response to applied deformation behaviours, the relationship between the shear deformation and cumulative 552 AE count demonstrates strong consistency. Once the deformation increases sharply in the accelerating stage, the 553 cumulative AE count curve shows a steeply upward trend. For the progressive deformation mode, the AE count 554 experiences a low-level period, an active period and a rapid increase period gradually. For the sudden deformation 555 mode, the AE count increases sharply, and the cumulative AE count curve rises much steeply. For the steady 556 deformation mode case, the acoustic emission activity is quite active in the early stage, and the cumulative AE 557 count curve rises sharply in the initial stage. Subsequently, the acoustic emission activity maintains a relatively 558 low level, the cumulative AE count curve shows a relatively flat trend. 559 (2) For the progressive deformation and sudden deformation mode, correlation hits distribution of the AE 560 duration-count and amplitude-energy concentrated in a relatively small scale in the initial stage with low 561 deformation rate. The signal hits distribution becomes relatively scattered and expands into a large scale as the 562 deformation rate increases. For the sudden deformation mode, the evolution period towards scattered signal hits 563 distribution is much shorter than that of progressive deformation mode. Distinctive to the previous two cases, the 564 correlation hits distributions are relatively scattered in a large range firstly, then the signal hits distribution begins 565 to narrow gradually, at last, the signal points distribution is relatively concentrated in a very small scale. 566 (3) The AE signal detection during the applied movement process of different deformation modes exhibits 567 evolution characteristics as for the dominant frequency domain. Under the rapid deformation rate, not only the 568 number of the low frequency and intermediate frequency signals increases drastically, while the continuous high 569 frequency signals also significantly increase. For the progressive and the sudden deformation modes, the 570 frequency domain presents low frequency signals firstly then other additional high frequency. In contrast, for the 571 steady case, series high-frequency signal hits occur in the early stage, while there are no high-frequency signals 572 appear anymore over time. The evolution pattern of high frequency domain may be effective indexes for different 573 landslides, indicating that the landslide may be about to enter a rapid deformation stage. 574 (4) From the statistical trend of the energy percentage of the high frequency band between the 312.5~500 kHz energy proportion of the high-frequency signals increases significantly during the rapid deformation stage, and the 578 energy proportion occupies up to approximately 20% close to the later stage. For the steady deformation mode, 579 the energy proportion of high-frequency signal increases to a larger value rapidly in the initial stage. The energy 580 proportion of high-frequency begins to decrease, experiencing a process of a rapid increase firstly and then a 581 gradual decrease over time as the deformation rate gradually decreases. Note that the energy proportion parameter 582 of the high frequency band has the potential to be used as the identifying reference to the risk management of the 583 landslide early monitoring. 584 585 between fracture geometrical morphology and acoustic emission power spectrum characteristics. Bulletin of Figure 1 Part of regional landslide hazards in the world Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.  The AE count, cumulative AE count and shear surface displacement for east test (a) progressive deformation mode;(b) sudden deformation mode;(c) steady deformation mode The AE dominant frequency evolution and shear displacement in the entire test process(a) progressive deformation mode;(b) sudden deformation mode;(c) steady deformation mode