Preparatory Process in Advance of Runaway Fault Rupture through Fluid Injection Observed in Laboratory Experiments Using a Large Specimen of Sub-meter scale

Preparatory Process in Advance of Runaway Fault Rupture through Fluid Injection Observed in Laboratory Experiments Using a Large Specimen of Sub-meter scale | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Preparatory Process in Advance of Runaway Fault Rupture through Fluid Injection Observed in Laboratory Experiments Using a Large Specimen of Sub-meter scale Takatoshi Ito, Koji Aoki, Yusuke Mukuhira, Yasuo Yabe This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4484150/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 22 Nov, 2024 Read the published version in Earth, Planets and Space → Version 1 posted 5 You are reading this latest preprint version Abstract Fault slip is initiated by locally applied fluid pressure, and it can expand unstably over a wide area causing elastic waves having magnitudes that induce felt or destructive earthquakes. Thus, it is important to examine the unstable expansion of initial slips. However, it is hard to reproduce the process by general setup of laboratory experiment such as triaxial loading tests on cylindrical specimens with inclined faults. In this study, we prepared a cubic specimen of sub-meter scale, which was separated into two triangular prisms by a model fault. The specimen was subjected to bi-axial compressions with different magnitudes. A 2D array of strain gauges was embedded beneath the fault plane to measure the changes in shear strain with the fault slip driven by fluid injection. Based on the experimental results, we discussed the features of fault slips that lead to injection-induced earthquake. The strain accumulated around the edge of the fault slipping area. The accumulation increased locally the strain by ~ 10 µε, which was equivalent to ~ 0.1 MPa in shear stress. The fault slipping area expanded gradually first, and it expanded unstably beyond the fluid invasion area ~ 3 s later after the slip initiated. The unstable expansion of initial slips was suppressed due to reducing the initial shear stress on the fault by 0.3 MPa. In this case, the initial shear stress was too small for the additional stress accumulated at the edge of the fault slipping area to overcome the static friction on the fault. Fracturing Injection-induced earthquake Fault slip Laboratory experiment Shear strain Runaway rupture Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Introduction Earthquakes occur when elastic waves pass through subsurface rocks, resulting in shaking of the ground. The elastic waves are produced by unstable deep fault slips, which are induced naturally when shear stress exceeds static friction on faults under tectonic loading. Unstable slips are also induced artificially by injecting fluids into rocks for fracturing (Ellsworth, 2013 ; Morris et al., 2017 ; Kim et al., 2018 ). As earthquakes can threaten socio-economic activities of humans, the mechanisms underlying earthquake occurrence have been actively examined. As in situ observations of unstable slips at deep depths are difficult, a number of attempts have been made to reproduce these slips in the laboratory (Lockner et al., 1982 ; Ohnaka, 1986; Bartlow et al., 2012 ; Guérin-Marthe et al., 2023 ; Selvadurai & Glaser, 2017 ; Yamashita et al., 2021 ; Wang et al., 2020 ; Gori at al., 2021; Cebry & McLaskey, 2021 ; Ji et al., 2022 ; Oye at al., 2022). However, laboratory studies for injection-induced earthquakes are limited compared with those for natural earthquakes, possibly because of the following issues on specimens and observation methods of fault slips in laboratory experiments. In the case of injection-induced earthquakes, fault slip is initiated by fluid pressure applied locally, and it sometimes expands unstably over a wide area such that it results in elastic waves having magnitudes that can induce felt earthquakes. Thus, reproducing the unstable expansion of initial slips in laboratory is important; however, it is hard to reproduce the process by general setup of laboratory experiment such as triaxial loading tests on cylindrical specimens with inclined faults, which have been used for the laboratory studies of natural earthquakes (Bartlow et al., 2012 ; Guérin-Marthe et al., 2023 ). The specimens used for those studies are generally less than 100 mm in diameter, and if such specimens are used to reproduce injection-induced earthquakes, an inlet hole for fluid injection into fault is to be located in the near distance of 50 mm or less from the fault edge. However, fluid pressure propagates quickly over such short distance, making it unfeasible to observe the expansion of initial slips. Another issue related to laboratory experiments is the observation method of fault slips. In laboratory experiments, faults need to be prepared considering the limited allowable size in laboratories. However, slip displacement of faults is proportional to the fault size, and the limited allowable size of fault would make fault slips to be extremely small. Contrastingly, fault slips cause changes in the shear strain on fault. Shear strain is independent of fault size and can be measured easily using strain gauges. Therefore, observing fault slips based on shear strain than displacement is relatively easier in laboratory experiments. This approach was applied by Ohnaka (1986) and allowed to demonstrate the quasistatic and stable nucleation prior to fault instability in natural earthquake studies. The experiment was conducted using a specimen of quadrilateral plate separated into two halves by a sawcut fault. The fault slip was induced by applying compressive stresses to the lateral edges of the specimen. Subsequently, the changes in shear strain associated with fault slip were measured by strain gauges attached on the side of the specimen located close to the fault. This was suitable for an experiment using the fault long in the direction of fault slipping but narrow in its orthogonal direction. However, such fault geometry is not appropriate for conducting experiments of injection-induced earthquakes since fluid pressure propagates quickly over such narrow fault widths as in the case of small cylindrical specimens. The issue of rapid pressure propagation can be avoided by thickening the specimen; however, observing the initiation of fault slips inside the specimen will be difficult by strain gauges attached on the side surface of specimen. Considering these issues, in the present study, we adopted a novel laboratory-based experiment approach to examine fault behavior driven by localized-high fluid pressure. We prepared a cubic specimen separated into two triangular prisms by a model fault with a wide area of about 600 × 700 mm 2 . The specimen was subjected to bi-axial compressions with different magnitudes. A two-dimensional (2D) array of strain gauges was embedded beneath the fault plane to measure the changes in shear strain with the fault slip driven by fluid injection. Based on the experimental results, we discussed the features of fault slip that lead to injection-induced earthquakes. As a result, we found that the observed behavior was described well by the model of injection-induced aseismic slip which transmits an elastic perturbation that triggers seismicity beyond the fluid-pressurized zone. The model has been presented conceptually upon field and laboratory experiments (Elsworth et al., 2016 ; De Barros et al., 2018 ; Bhattchacharya and Viesca, 2019; Cappa et al., 2019 ; Ji et al., 2022 ) and examined numerically in detail through hydromechanical modelling of a slip-weakening and permeable fault response to a fluid injection (Norbeck and Horne, 2017; Wynants-Morel et al., 2020 ; De Barros et al., 2021 ). Materials and Methods Set up of the laboratory experiment Experimental system The experimental system used in this study is shown in Fig. 1 . The system comprised a loading frame to apply compressive loads to a specimen, electric and hand oil pumps to supply oil pressure for the loading, an electric syringe pump to inject fluid into a model fault of specimen, and a data monitoring and acquisition system. The loading frame accommodated specimens with dimensions 600 × 600 × 600 mm 3 . The specimen was subjected independently to biaxial compressive stresses S x and S y in horizontal and vertical directions, respectively, by two pairs of flat jacks. The oil pressure for activating the flat jacks was supplied in principle by an electric pump, and a hand oil pump was used for finely adjusting the oil pressure. The pressure in the pair of flat jacks was equal to the compressive stress subjected to the specimen by the flat jacks. Specimen containing a fault with embedded strain gauges The specimen was a 600 × 600 × 600 mm 3 block of the Mogami andesite with Young’s modulus ( E ) value of 16.3 GPa and a Poisson’s ratio ( n ) of 0.2 (Ito et al., 2022 ). The model fault in the specimen consisted of a throughout-going sawcut oriented at an angle of 45° to the sides of the cubic block. Thus, the sawcut split the block diagonally in two equal halves above and below the fault plane, which are referred to the upper and lower blocks, respectively (Figs. 2 a and b). The normal stress distribution on the fault was checked using a pressure sensitive paper. When the normal stress in a region on the fault was higher than in other regions, the high normal stress region was ground. By repeating these, the fault planes were carefully prepared so that they were in contact as uniformly as possible and the normal stress in the middle region of the fault was rather slightly higher than in its top and bottom edges. We discussed further the normal stress distribution in the discussion section 4.3. The compressive stresses S x and S y applied by the biaxial loading frame caused normal and shear stresses s F and t F on the fault plane given as follows: $${\varvec{\sigma }}_{\varvec{F}}=\frac{{\varvec{S}}_{\varvec{x}}+{\varvec{S}}_{\varvec{y}}}{2}, {\varvec{\tau }}_{\varvec{F}}=\frac{{\varvec{S}}_{\varvec{x}}-{\varvec{S}}_{\varvec{y}}}{2}$$ 1 These relationships show that s F can be increased by maintaining t F , when S x and S y are increased together by the same amount; conversely, t F can be increased by maintaining s F , when S x is increased but S y is decreased by the same amount. An inlet port was set at the center of the fault plane of the upper block for injecting a fluid in the fault (Fig. 2 a). The fluid was supplied by a syringe pump through a tube set in a borehole drilled in the upper block. A groove with width and depth of 5 mm was dug along the fault plane of the upper block in the transverse direction as the groove crossed the inlet port. The groove end was separated by ~ 80 mm away from the open surface of the specimen block, and furthermore, a long rubber seal was installed along the fault edge to prevent fluid outflow from the fault. Thus, the fluid pressure in the groove was not leaked easily to the open surface and maintained while injection. This arrangement ensured that the fluid injected into the fault spread first along the transverse groove and flowed out in parallel with the longitudinal direction of the fault. This flow of fluid will cause a pressure distribution uniform in the transverse direction of the fault. The setup was designed for normal and shear stresses ( s F and t F , respectively) to be also uniform in the transverse direction of the fault. These loading conditions simplified the fault slip to be uniform in the transverse direction of the fault; additionally, the simplification facilitated the observation and understanding of the fault slip. Furthermore, a 2D array of strain gauges was embedded beneath the fault plane to measure the changes in the shear strain with the fault slip driven by fluid injection. To achieve this, three narrow grooves (F, C, and B) were dug along the fault plane according to the arrangement described in Fig. 3 a. These grooves were separated from each other by 140 mm and oriented along the longitudinal direction of the fault. Strain gauges (KFGS-2-120-D31-11, Kyowa Electronic Instruments, 2023 ), as shown in Fig. 3 b, were attached to the side wall of the grooves 5 mm below the fault plane (Fig. 3 c). The strain gauge consisted of a pair of matched gauges with mutually perpendicular orientations to directly monitor the shear strain parallel to the fault plane at each attached location. Notably, the shear strain could be converted to the shear stress by multiplying it and the rigidity of the rock sample, i.e., E /2(1 + n ). The grooves were filled with epoxy resin after attaching the gauge. The resin exhibited a low value of Young’s modulus (~ 3 GPa), which did not affect the deformation of the specimen rock. Lead-wires of the strain gauges were placed along the grooves and were connected to a data logger (UCAM-550A; Kyowa Electronic Instruments, 2023 ). In total, 39 strain gauges were placed, of which 13 were placed in each groove. The data logger synchronously sampled for all channels by a sampling frequency of 10 Hz. The oil pressure in the flat jacks of the biaxial loading frame was acquired by the same logger to eliminate errors in the measurement time of each dataset. The type of strain gauge and the sampling frequency in this study were not designed to detect motions with small amplitudes and high frequencies as of elastic waves. To indicate the location of each strain gauge, we assigned a name in the format of “F/C/B” and “+/-” and “0/1/2 … /6”, where the first letter indicates the respective name of the groove on which the gauge was installed, and the second letter indicates the upper (+) and lower (-) sides, respectively, where the gauge was installed relative to the center line across the fault, and the third letter indicates the distances of 0, 20, 40, 60, 80, 120, and 160 mm from the center line across the fault (see Fig. 3 a), where the gauge was installed. Results Experiment I: Shear stress loading under constant normal stress Figure 4 shows the time variations in S x and S y , and Fig. 5 a shows the resultant stresses s F and t F in Experiment I. The compressive stresses, S x and S y , were increased together first from 0 MPa to 6.5 MPa at the same rate. This manner of loading resulted in s F with the same magnitude as that of S x and S y to maintain t F at zero. After S x and S y increased to 6.5 MPa, S x increased but S y decreased at the same rate. This method of loading increased t F maintaining s F at 6.5 MPa. To achieve this loading, the release valve was opened slightly to gradually decrease the pressure in the pair of the flat jacks for applying S y . Simultaneously, the hand pump was operated manually to increase the pressure in the pair of the flat jacks to apply S x at a rate similar as much as possible to the rate of decreasing S y . The stress variation fluctuated to some extent because the hand pump was operated manually. Figure 6 shows the time variation in the shear strains measured by the 2D array of strain gauges embedded in the fault plane, where the time range was limited to 550–660 s. For the ease of comparison, the plotted values of strain represented the amount of change from the value at a reference time of 550 s. Each graph in Fig. 6 shows the variation measured by one strain gauge and is plotted in accordance with the alignment of corresponding stain gauge in the 2D array. The scale of the vertical axis for strain was the same for all graphs in the figure. The results indicate the dynamic behavior of the fault. A sudden change in the shear strain occurred at 645 s. Subsequently, the shear strains suddenly dropped in all graphs, while the strains increased conversely at the limited locations around B + 1 and B0. This phenomenon can be explained as follows: A dynamic slip of the fault plane occurred entirely at 645 s. Later, the shear stress released by the slip concentrated around B + 1 and B0, since the normal stress on the fault and the resultant friction were larger at these points than at others. Figure 5 b shows the time variations in the normal and shear stresses on the fault plane for the time range of 550–660 s. s F and t F were estimated using Eq. ( 1 ) from S x and S y , which were equal to the oil pressure in the two pairs of flat jacks, respectively. At 645 s, when the measured shear strains changed abruptly, t F also dropped suddenly by a small amount of ~ 0.1 MPa, while s F remained constant around that time. Such changes in s F and t F indicated the occurrence of the dynamic slip of the entire fault for the following reasons: When the entire fault slipped, the distance between the left and right sides of the specimen shortened, and simultaneously the distance between the top and bottom sides of the specimen was extended by the same amount, considering the fault was oriented at 45° to the sides of the specimen. Consequently, the flat jacks for applying S x expanded to decrease the inner pressure and the stress, S x , and those for applying S y compressed to increase the inner pressure and the stress, S y , by the same amount. Such sudden changes in S x and S y abruptly decreased t F , but s F did not change following Eq. ( 1 ). Figure 5 b shows these changes in s F and t F , and this result supported the occurrence of the dynamic slip of the entire fault at 645 s. The shear stress, t F , was 2.4 MPa just before it dropped suddenly. Interesting and locally different features were observed in the strain variation (Fig. 6 ) before the dynamic slip of the entire fault. During the early period, the shear strain increased monotonically at all instances, while the increasing rate changed with locations, with high rates around the middle region but low rates around the top and bottom edges of the 2D array of strain gauges. During the middle period, the time variation in the strains bended, and its slope reduced. This change was associated with the decrease in the increasing rate of t F from ~ 600 s (Fig. 5 b). However, during the later period, although the increasing rate of t F was maintained, the strains changed with time in different manners at different locations. Around the top and bottom edges, the shear strain became constant or decreased gradually. Contrastingly, in the middle region, the shear strain continued to increase rather more steeply. The strain increase in the middle region was considered to have occurred for compensating the decay of the strain accumulation around the top and bottom edges. The phenomena of local and steep increase in strain might have demonstrated the formation of the asperity in seismology. The steep increase in strain followed the slowdown of the increasing rate immediately before the dynamic slip of the entire fault. The slowdown might be associated with the slow slip in the asperity. Experiment II: Injecting fluid under sub-critical stress conditions After disassembling the specimen to check and clean the fault planes, we reinstalled the specimen in the loading frame and conducted Experiment II. Figures 5 c shows the time variation in s F and t F , and injection pressure P . Similar to the procedure of Experiment I, t F started increasing gradually while maintaining s F at 6.5 MPa. The time variation of the measured shear strains is shown in Fig. 7 from 1900 to 2060 s, in the manner same as that described in Fig. 6 . The shear strain increased monotonically in the initial period, and subsequently, the strain started concentrating around the middle region of the 2D array of the strain gauges. This trend was similar to that observed in Experiment I before the dynamic slip occurred in the entire fault. Therefore, the shear stress loading ended at 1980 s and was held by closing the valve on the tube that connected the oil pump and flat jacks. Subsequently, the value of t F was estimated as 2.3 MPa (Fig. 5 d), which was only 0.1 MPa below the critical value (2.4 MPa) required for initiating the dynamic slip of the entire fault in Experiment I. This result indicated good reproductivity of the experiment including the persistent occurrence of the asperity. In the next step, fluid injection into the fault started from 2040 s, and it continued until the pressure reached 5 MPa (Figs. 5 c, d). We used a water-based fluid, with 130 mPa·s viscosity, for injection. The fluid was prepared by adding a small amount of xanthan gum to water (Petri, 2015 ). It was expected that the viscosity of the fluid facilitated localized pressure around the injection port because of increasing flow resistance through the fault. At ~ 2045 s, that is 5 s after starting fluid injection, the shear strain changed suddenly, as shown in Fig. 7 . Figure 5 d shows the time variation in s F and t F from 1900 to 2060 s. As shown in the figure, the shear stress suddenly dropped at 2045 s and was consistent with the time when the shear strain changed suddenly. The sudden change in t F and the measured strain indicated that the dynamic slip of the entire fault occurred at 2045 s in this experiment. The strain change and fault slip before the dynamic slip of the entire fault have been discussed comprehensively later. Experiment III: Injecting fluid under sub-critical stress conditions Upon checking and cleaning the fault planes, we conducted Experiment III. s F was set as 6.5 MPa as in Experiment II, but t F was set as 2.0 MPa, which was 13% smaller than that in reported in Experiment II (2.3 MPa). Figure 5 e shows the time variation in s F and t F , and P . The fluid injection into the fault started from 2710 s in the same manner as in Experiment II by maintaining s F and t F values. In ~ 8 s, the injection pressure reached 5 MPa, which was maintained thereafter. The time variation from 2600 to 2740 s in the measured shear strains is shown in Fig. 8 . The shear strains were stable until fluid injection. Some strain variations occurred later; for example, the strain decreased at C-1 and C-2 but increased at C0, C + 1, and C + 2. However, the variation was limited in the middle region of the 2D array of strain gauges, and sudden changes in the strain were not reported at all instances as observed previously in Experiments I and II. Similar was the case for the time variations in s F and t F , as shown in Fig. 5 f, from 2600 to 2740 s. t F was stable from 2710 s for more than 10 s after starting fluid injection. Based on these observations, we concluded that the dynamic slip of the entire fault did not occur in this experiment. Discussion Fault slipping estimated from observed distribution of shear strain Fault behavior associated with fluid injection has been previously examined in numerical studies of Norbeck & Horne ( 2018 ) and Wynants-Morei et al. (2020). These studies identified a general feature related to the fault slip, shear stress, and pressure changes along the fault. A typical example is shown in Figs. 9 a - c, which was obtained by our simulation on a 2D hydromechanical modelling of a slip-weakening and permeable fault response to fluid injection. The theories for the modelling are basically the same as those of Norbeck & Horne ( 2018 ) and Wynants-Morei et al. (2020). It was assumed for the simulation that the fault was subjected initially to shear stress, normal stress and fluid pressure of 2.5, 17.5 and 10 MPa, respectively, a local fluid injection into the fault increased the pressure linearly to 15 MPa, and the friction coefficient decreased from 0.6 down to 0.55 with fault slip velocity. The results of Figs. 9 a - c demonstrated the role of fluid and friction on the injection-induced seismicity described conceptually in other previous studies (Elsworth et al., 2016 ; De Barros at al., 2018; Bhattchacharya and Viesca, 2019; Cappa et al., 2019 ; Ji et al., 2022 ). Injection pressure propagated from the injection point with time as shown in Fig. 9 a. The increased pressure reduced friction between fault planes and induced fault slipping. The fault slip released shear stress locally, and the released shear stress led to their accumulation around the edge of the slipping fault area for overall stress balance (Fig. 9 b). Conversely, the time variation in shear stress is shown in Fig. 9 c for four locations at different distances from the injection point. The results showed two types of variations (A and B) as illustrated in Fig. 9 d, depending on the location. Type A variation appeared around locations where the fault slip initiated, and Type B variation appeared at locations where they departed from the location of slip initiation, and the fault slip occurred later. In Type A variation, the shear stress decreased, approaching a lower limit associated with the dynamic friction of the fault. Type B variation showed a peak value during the middle period, and during the peak period, the edge of the fault slipping area passed the location. The peak value increases gradually as its location departs away from the location of slip initiation. Note that shear strain changed in the same manner with shear stress, since they were directly proportional to each other. This implied that the time variation in shear strain could be characterized in the same manner as described in Fig. 9 d. We compared the expected stress/strain variation of Fig. 9 d with the observations of Experiment II. Figure 10 shows the details of strain change during the fluid injection for Column F of the 2D array of strain gauges, where the scale of the vertical axes for strain was the same for all graphs in the figure but different from that of Fig. 7 . The time variations in t F and P have been shown together for comparison. Figure 10 shows that the shear stress suddenly dropped at 2045.5 s accordingly to the dynamic slip of the entire fault. Note that in the figure, t F does not seem to decrease instantaneously; that is, it decreased along a slope with not an infinite but a finite gradient, since the data were sampled and plotted discretely at the sampling frequency of 10 Hz. The shear strain at F0 and F-1 decreased from ~ 2042.7 s similar to that for Type A variation; therefore, the fault slip was considered to initiate from a region around F0 and F-1 at that time. The shear strain at the other locations around F0 and F-1 varied in the manner of Type B variation. Each variation showed a peak value, which appeared later, as the location was away from F0 and F-1. This indicated that the fault slipping area expanded gradually with time far from F0 and F-1 to F + 4 and F-4 following the peak as indicated by arrows in Fig. 10 . Subsequently, the dynamic slip of the entire fault occurred at 2045.5 s, which was ~ 3 s later after the slip initiated. The propagation speed of the Type B variation was ~ 20–30 mm/s on average, which is much slower than a typical speed of the dynamic rupture propagation (~ 1.3 km/s; 80% of S-wave speed). These suggested the presence of a quasi-static or aseismic preparatory process before the dynamic or seismic fault slipping induced by fluid injection that took few seconds. Figure 11 shows a photograph of the fault plane captured immediately after the experiment. The strain gauges were installed along three parallel grooves filled with epoxy resin, which corresponded to the three columns, F, C, and B, of the 2D array of strain gauges. The gauge locations are indicated by black dots. The area where the injected fluid invaded along the fault can be seen clearly from the color differences. The invasion area almost reached the F + 4 and F-4 locations along Column F. This result was consistent with the extension of fault slipping area estimated as above from the variation in the measured strain. Moreover, the fluid invasion area was ~ 20% of the area of the entire fault. Thus, in the dynamic slip of the entire fault, the fault slip area expanded instantaneously from 20 to 100% beyond the fluid invasion area. Further, since the dynamic slip of the entire fault occurred ~ 5 s after the fluid injection started at 2040 s, a migration speed of the injected-fluid front was estimated to be ~ 16 mm/s on average. This speed was smaller than the propagation speed of ~ 20–30 mm/s of Type B variation initiated at ~ 2042.7 s. Thus, the aseismic slip initiated ~ 3 s after the fluid injection, the slip area expanded faster than the injected-fluid area, and the seismic slip occurred at last when the slip front got across the injected-fluid front. Such a fault behavior follows well the model proposed recently on field experiments and numerical simulations (Elsworth, 2016; De Barros at al., 2018; Bhattchacharya and Viesca, 2019; Cappa et al., 2019 ; Ji et al., 2022 ). The model shows that the injection-induced aseismic slip transmits an elastic perturbation that triggers seismicity beyond the fluid-pressurized zone. Conditions for runaway rupture Fluid injection was conducted in both Experiments II and III. However, the dynamic slip of the entire fault occurred in the former experiment but not in the latter. It is reasonable to consider that the dynamic slip was associated with a difference in the experimental condition between them, which was the shear stress t F set before fluid injection. In Experiment II, the t F value was set at 2.3 MPa, indicating that the fault was ready to slip entirely. As described about the results in Fig. 10 in the previous section, the shear strain around the edge of the fault slipping area locally increased compensating the reduction in shear strain supported on the fault plane along the slip area. The amount of shear strain that increased locally tended larger as the slip area became larger. Such tendency was consistent with the numerical simulations in Fig. 9 c. When the fault slip area started to expand instantaneously beyond the fluid invasion area, the amount of shear strain that increased locally was ~ 10 µe in shear strain, which was equivalent to ~ 0.1 MPa in shear stress. This amount was small but sufficient for the total sum of shear stress to exceed the static friction on the fault. Consequently, if the slip area expanded further, the shear stress would increase around the edge of the expanded slip area. These coupled phenomena will occur spontaneously without additional fluid injection, in contrast to the preparatory process that was driven by the fluid pressure, as illustrated in Figs. 12 a–d, where µ s and µ d (assuming µ s > µ d ) are the static and dynamic friction coefficients of the fault, respectively, and s f and s f 0 are effective normal stress and its initial value, respectively. Subsequently, the slip area will expand entirely over the fault assuming that the shear stress is nearly balanced with the static friction on the fault. This type of fault slipping has been referred as the runaway rupture (Norbeck & Horne, 2018 ; Ji et al., 2022 ). The mechanism shown in Figs. 12 a–d suggests that the runaway rupture tends to occur more easily, as the shear stress on fault approaches the static friction and the dynamic friction departs to be smaller from the static one. The numerical simulation of Norbeck & Horne ( 2018 ) accordingly showed that the critical state of shear stress on fault is not sufficient but the friction condition of µ s > µ d is also necessary for the runaway rupture. The dynamic slip of the entire fault in Experiment II possibly occurred following the sequence shown in Figs. 12 a–d. However, in Experiment III, the shear stress was set 2 MPa before fluid injection with a difference of 0.3 MPa from that of Experiment II. The difference appears to be too large to be overcome by the accumulation of shear stress around the edge of slip area. Therefore, the dynamic slip of the entire fault did not occur possibly in Experiment III. Wynants-Morel et al. ( 2020 ) discussed in detail the effect of initial shear stress t F 0 on the injection-induced slip of fault. To do this, they carried out numerical simulations changing the initial Shear Capacity Utilization, SCU, from 53–71%. The SCU is defined as the value of t F 0 divided by the static friction f 0 = µ s ( s F 0 – P 0 ), and s F 0 and P 0 are the initial normal stress on fault and the initial fluid pressure in fault, respectively. In their all cases of simulation, the pressure front and the Type B variation of shear stress moved away with time from the injection point similarly to Fig. 9 a-c. However, two different behaviors were observed depending on SCU. For the cases with SCU less than 64%, the Type B variation remained behind the pressure front. On the other hand, for the cases with SCU larger than 68%, the Type B variation was first behind the pressure front, then the Type B variation accelerated and outpaced the pressure front. The reason was concluded that for the cases with SCU larger than 68%, the t F 0 was so large to exceed the dynamic friction f d 0 = µ d ( s F 0 – P 0 ) and then the stress drop ( t F 0 – f d 0 ) provided additional potential energy to drive the slip front growth. These results of Wynants-Morel et al. ( 2020 ) and the result of Norbeck & Horne ( 2018 ) described above fit well with the results of our laboratory experiments assuming that the f 0 and f d 0 took a value of ~ 2.4 MPa and a smaller value between 2 and 2.3 MPa, respectively. Thus, the runaway rupture occurred in the case with t F 0 > f d 0 as Experiment II, but the rupture was limited around the fluid injection place in the case with t F 0 < f d 0 as Experiment III. Shear strain/stress redistribution with loading Here, we discuss the variations in shear strain observed while setting the shear stress. In Experiment II, t F increased to 2.3 MPa, maintaining s F at 6.5 MPa. s F and t F estimated by Eq. ( 1 ) from the flat jack pressures, S x and S y , represent the stresses averaged over the fault plane. Normal and shear stresses varied locally more or less than the estimated values of s F depending on the local conditions of contact between the fault planes. Thus, the normal and shear stresses acting locally on the fault were referred and denoted as the local normal stress s and the local shear stress t , respectively. For the normal stress s , its distribution was actually checked by using a pressure sensitive paper for a range of 2.5–10 MPa (Prescale, LW type, Fujifilm Co., 2024). Then, the paper was inserted between the fault planes and compressed by applying S x and S y with the same value of 6 MPa to the specimen. Figure 13 shows a photograph of the paper after the measurement, where the strain gauge locations are indicated by black dots. The color appears red where normal stress is applied, and the color density varies according to the amount of normal stress. Note that no color indicates not no stress but a stress less than the lower limit of measurement range, i.e. 2.5 MPa in this case. It can be seen from Fig. 13 that s was distributed uniformly in the middle region between + 4 and − 4 locations of the strain gauges overlapping with the region of injected-fluid invasion in Experiment II (see Fig. 11 ), while it became smaller around + 6 and − 6 locations of the strain gauges. In the summarized process for Experiment II shown in Fig. 7 , the shear strain increased monotonically everywhere in the initial period, while the increasing rate changed with locations as it was high around the middle region between + 4 and − 4 locations of the strain gauges but low around + 6 and − 6 locations of them. Subsequently, the strain began to concentrate around the middle region. The difference in the strain variation may be caused by the difference in local conditions of contact between the fault planes as appeared in Fig. 13 . In the regions where the fault planes were more firmly in contact, the local normal stress and static friction were larger, and the fault became harder to slip and accept more shear strain in the surrounding rock body. This phenomenon can be understood more clearly by replacing shear strain with shear stress in the directly proportional relationship. Figure 12 e illustrates the paths of local stress changing at three locations A, B, and C on the fault. Stage 1 shows the stress state at the end of s F loading. t was zero everywhere, and s was assumed to differ at each location. With t F loading, t also increased, and at Stage 2, t at the location A with the smallest s was the first to reach the static friction on the fault shown by the diagonal line with a slope of the static friction coefficient µ s . Subsequently, the fault slipped locally at A, and later, t at A could not increase further; however, t at locations B and C continued to increase rather at a higher rate since they transferred the shear stress, which should have been stored at A. The same phenomenon possibly occurred in the loading process of Experiments I and II. Thus, the strain varied in different manners in different locations. With further t F loading, the stress state could reach Stage 3 wherein t was locally balanced with the static friction, while the value of the static friction somewhat differed at each location. In Experiment II, the stress state before fluid injection was considered at Stage 3, and therefore, the runaway rupture occurred even while the local normal stress was not uniform. This may also be observed for Experiment I. The strain varied until the occurrence of the dynamic slip of the entire fault similar to that in Experiment II. The stress state at Stage 3 was established finally and resulted in the runway rupture. Thus, Experiment I may demonstrate a type of preparedness process for natural earthquakes. In contrast to Experiments I and II, the stress state in Experiment III should be in between Stages 2 and 3. The stress state should be on the line of the static friction in top and bottom edges of the fault where the normal stress was low. On the other hand, although the shear stress in the middle of the fault might be high, it should be much lower than the static friction level, because the normal stress was high in this portion. Therefore, even when the normal stress was reduced by the fluid injection or the shear stress was increased by the accumulation in the middle of the fault, the preparatory process could not be sustained or grow to the runaway rupture. Therefore, Experiment III argues that we could mitigate risk of destructive earthquakes induced by injection, if we would measure the fault strength as well as the stress state along a fault in advance and would adjust the fluid pressure carefully. We argue here the same physics from opposite perspective argued in Wynants-Morel et al. ( 2020 ) and De Barros et al. ( 2021 ) that for a high-criticality stressed fault, the aseismic slip propagation is self-sustained and outpaces the pressurized zone, on the contrary, for a low-criticality stressed fault, the aseismic slip stays driven by the effective stress decrease and remains within the pressurized zone. Conclusions In this study, we successfully conducted experiments to detect a preparatory process in advance of a dynamic slip of the entire fault induced by fluid injection. We found that fault slipping area expanded gradually away from the location of fluid injection. Dynamic slip occurred few seconds after the initiation of fault slipping, and subsequently, the fault slip area expanded instantaneously from 20 to 100% of the model fault. The shear stress should have been nearly balanced with the static friction everywhere on the fault just before the dynamic slip. This critical condition should have been developed along with local slipping of the fault during shear stress loading on the fault. Under critical conditions, the fault slipping area expanded to increase the shear stress around its edge, and the increase drove its further expansion. These coupled phenomena occurred spontaneously without additional fluid injection. Further, the dynamic slip of entire fault possibly occurred after this sequence of events. In contrast, when the fault was loaded under subcritical conditions, the preparatory process induced by the fluid injection was quickly arrested, probably due to an insufficient shear stress accumulation in a high normal stress (high static friction) region on the fault. This suggests that the measurement of frictional properties of the fault in addition to the stress state along the fault should provide key information to reduce risks caused by earthquakes induced by the fluid injection. Numerical simulations allow us to figure out fault slip behavior in detail. However, there is unavoidable ambiguity resulting from simplifications in its modelling and difficulties to give all simulation parameters correctly. In the present study, we succeeded to observe directly the behavior that the injection-induced aseismic slip transmitted along the fault an elastic perturbation to lead the seismic slip. This is the first attempt in the study of the injection-induced earthquake, to the author’s knowledge. The results and the proposed method of laboratory experiment will contribute for understanding the injection-induced earthquake. Declarations Ethics approval and consent to participate Not applicable. Availability of data and materials The dataset of laboratory experiments is available at Ito et al. (2023). Competing interests The authors declare that they have no competing interest. Funding This work was supported by Japan Society for the Promotion of Science (JSPS KAKENHI) Grant Number 16H04612 and JP21H05201. Authors' contributions TI designed this study. KA and YM contributed to experiments and data analysis. YY joined the interpretation. All authors contributed to and approved the manuscript. Acknowledgments The authors wish to thank Prof. D. Swenson, Mr. K. Kaneta, Mr. S. Inoue and Mr. K. Yokoyama for their assistance in the numerical simulations and the experiments. This work was supported by JSPS KAKENHI Grant Number 16H04612 and JP21H05201. We also wish to thank Editage (www.editage.com) for English language editing. Authors' information 1 Institute of Fluid Science, Tohoku University, Sendai, Japan. 2 Research Center for Prediction of Earthquakes and Volcanic Eruptions, Graduate School of Science, Tohoku University, Sendai, Japan References Bartlow NM, Lockner DA, Beeler NM (2012) Laboratory triggering of stick-slip events by oscillatory loading in the presence of pore fluid with implications for physics of tectonic tremor. J Geophys Research: Solid Earth 117(B11411). 10.1029/2012JB009452 Bhattacharya P, Viesca RC (2019) Fluid-induced aseismic fault slip outpaces pore‐fluid migration. Science 364(6439). 10.1126/science.aaw7354 Cappa F, Scuderi MM, Collettini C, Guglielmi Y, Avouac J-P (2019) Stabilization of fault slip by fluid injection in the laboratory and in situ. Sci Adv 5(3). 10.1126/sciadv.aau4065 Cebry SBL, McLaskey GC (2021) Seismic swarms produced by rapid fluid injection into a low permeability laboratory fault. Earth Planet Sci Lett 557(116726). 10.1016/j.epsl.2020.116726 De Barros L, Wynants-Morel N, Cappa F, Philippe D (2021) Migration of fluid-induced seismicity reveals the seismogenic state of faults. J Geophys Research: Solid Earth 126(11). 10.1029/2021JB022767 De Barros L, Guglielmi Y, Rivet D, Cappa F, Duboeuf L (2018) Seismicity and fault aseismic deformation caused by fluid injection in decametric in-situ experiments. CR Geosci 350(8). 10.1016/j.crte.2018.08.002 Ellsworth WL (2013) Injection-induced earthquakes. Science 341(6142). 10.1126/science.1225942 Elsworth D, Spiers CJ, Niemeijer AR (2016) Understanding induced seismicity. Science 354(6318). 10.1126/science.aal2584 Fujifilm Co (2024) https://www.fujifilm.com/us/en/business/industrial-materials/measurement-film/prescale , April 16, 2024 Gori M, Rubino V, Rosakis AJ, Lapusta N (2021) Dynamic rupture initiation and propagation in a fluid-injection laboratory setup with diagnostics across multiple temporal scales. Proc Natl Acad Sci USA 118(51). 10.1073/PNAS.2023433118 Guérin-Marthe S, Kwiatek G, Wang L, Bonnelye A, Martínez-Garzón P, Dresen G (2023) Preparatory slip in laboratory faults: Effects of roughness and load point velocity. J Geophys Research: Solid Earth 128(4). 10.1029/2022JB025511 Ito T, Kumazawa A, Tezuka K, Ogawa K, Yokoyama T, Funato A (2022) Experimental and numerical study on a two-stage coring method for stress measurement: Application to deep and high-temperature geothermal wells. Geothermics 100. 10.1016/j.geothermics.2021.102333 Ito T, Aoki K, Mukuhira Y, Yabe Y (2023) Preparatory process in advance of runaway fault rupture through fluid injection observed in laboratory experiments using a. 10.5281/zenodo.8348064 . large specimen of sub-meter scale [Data set]Zenodo Ji Y, Hofmann H, Duan K, Zang A (2022) Laboratory experiments on fault behavior towards better understanding of injection-induced seismicity in geoenergy systems. Earth Sci Rev 226. 10.1016/j.earscirev.2021.103916 Kim K-H, Ree J-H, Kim Y, Kim S, Kang SY, Seo W (2018) Assessing whether the 2017 Mw 5.4 Pohang earthquake in South Korea was an induced event. Science 360. 10.1126/science.aat6081 Kyowa Electronic Instruments (2023) https://www.kyowa-ei.com/jpn/product/index.html , Sept. 2, 2023 Lockner DA, Okubo PG, Dieterich JH (1982) Containment of stick-slip failures on a simulated fault by pore fluid injection. Geophys Res Lett 9(8). 10.1029/GL009i008p00801 Morris AP, Ferrill DA, Walter GR, Price AM, Smart KJ, Skoumal RJ, Brudzinski MR, Currie BS (2017) Lessons learned from the Youngstown, Ohio induced earthquake sequence from January 2011 to January 2012. J Rock Mech Geotech Eng 9(5). 10.1016/j.jrmge.2017.03.016 Norbeck JH, Horne RN (2018) Maximum magnitude of injection-induced earthquakes: A criterion to assess the influence of pressure migration along faults. Tectonophysics 733(9). 10.1016/j.tecto.2018.01.028 Ohnaka M, Kuwahara Y, Yamamoto K, Hirasawa T (1986) Dynamic breakdown processes and the generating mechanism for high-frequency elastic radiation during stick-slip instabilities. In: S. Das, J. Boatwright, C.H. Scholz (Eds.), Earthquake Source Mechanics , American Geophysical Union Geophysical Monograph Series 37. 10.1029/GM037p0013 Ohnaka M, Kuwahara Y (1990) Characteristics features of local breakdown near a crack-tip in the transition zone from nucleation to unstable rupture during stick-slip shear failure. Tectonophysics 175. 10.1016/0040-1951(90)90138-X Oye V, Stanchits S, Babarinde O, Bauer R, Dichiarante AM, Langet N, Goets-Allmann B, Frailey S (2022) Cubic–meter scale laboratory fault re–activation experiments to improve the understanding of induced seismicity risks. Sci Rep 12. 10.1038/s41598-022-11715-6 Petri DFS (2015) Xanthan gum: A versatile biopolymer for biomedical and technological applications. J Appl Polym Sci 132. 10.1002/APP.42035 Selvadurai PA, Glaser SD (2017) Asperity generation and its relationship to seismicity on a planar fault: a laboratory simulation. Geophys J Int 208(2). 10.1093/gji/ggw439 Wang L, Kwiatek G, Rybacki E, Bonnelye A, Bohnhoff M, Dresen G (2020) Laboratory study on fluid-induced fault slip behavior: The role of fluid pressurization rate. Geophys Res Lett 47(6). 10.1029/2019GL086627 Wynants-Morel N, Cappa F, De Barros L, Ampuero JP (2020) Stress perturbation from aseismic slip drives the seismic front during fluid injection in a permeable fault. Journal of Geophysical Research: Solid Earth 125. doi.org/10.1029/2019JB019179 Yamashita F, Fukuyama E, Xu S, Kawakata H, Mizoguchi K, Takizawa S (2021) Two end-member earthquake preparations illuminated by foreshock activity on a meter-scale laboratory fault. Nat Commun 12. 10.1038/s41467-021-24625-4 Supplementary Files EPSItoGraphicalAbst.jpg Cite Share Download PDF Status: Published Journal Publication published 22 Nov, 2024 Read the published version in Earth, Planets and Space → Version 1 posted Editorial decision: Major Revision 24 Jun, 2024 Reviewers agreed at journal 05 Jun, 2024 Reviewers invited by journal 04 Jun, 2024 Editor assigned by journal 01 Jun, 2024 First submitted to journal 27 May, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4484150","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":310420116,"identity":"4c0dc85c-9e02-4317-9273-8648204ebe71","order_by":0,"name":"Takatoshi Ito","email":"data:image/png;base64,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","orcid":"https://orcid.org/0000-0003-1992-1224","institution":"Tohoku University: Tohoku Daigaku","correspondingAuthor":true,"prefix":"","firstName":"Takatoshi","middleName":"","lastName":"Ito","suffix":""},{"id":310420117,"identity":"5c930750-223f-4bde-9841-ea2c10ae85be","order_by":1,"name":"Koji Aoki","email":"","orcid":"","institution":"Tohoku University: Tohoku Daigaku","correspondingAuthor":false,"prefix":"","firstName":"Koji","middleName":"","lastName":"Aoki","suffix":""},{"id":310420118,"identity":"a9faf77a-eb5f-42ea-93ff-302862f7f2e0","order_by":2,"name":"Yusuke Mukuhira","email":"","orcid":"","institution":"Tohoku University: Tohoku Daigaku","correspondingAuthor":false,"prefix":"","firstName":"Yusuke","middleName":"","lastName":"Mukuhira","suffix":""},{"id":310420119,"identity":"feef67a1-8b80-415e-980d-e61b71f70a26","order_by":3,"name":"Yasuo Yabe","email":"","orcid":"","institution":"Tohoku University: Tohoku Daigaku","correspondingAuthor":false,"prefix":"","firstName":"Yasuo","middleName":"","lastName":"Yabe","suffix":""}],"badges":[],"createdAt":"2024-05-27 10:13:42","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4484150/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4484150/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s40623-024-02092-7","type":"published","date":"2024-11-22T15:57:50+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":58753196,"identity":"fb0543a8-2421-4af5-92ec-361b5cb6d06c","added_by":"auto","created_at":"2024-06-20 16:17:12","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":481066,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"EPSItoFigure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4484150/v1/e1e23fc6ebc2ed8c3c9c1920.jpg"},{"id":58753191,"identity":"e9935718-c035-4b21-b3fc-796b6bea421b","added_by":"auto","created_at":"2024-06-20 16:17:12","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":303650,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"EPSItoFigure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4484150/v1/15c7d08d706494931f77804b.jpg"},{"id":58753192,"identity":"dfd65203-b449-4b17-b040-1ddf910bb849","added_by":"auto","created_at":"2024-06-20 16:17:12","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":426481,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"EPSItoFigure3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4484150/v1/eb5c68c45baca8e83ef80bc4.jpg"},{"id":58754249,"identity":"b8138b8f-f6dc-4064-8cab-6c7c62ef29f9","added_by":"auto","created_at":"2024-06-20 16:25:12","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":220483,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"EPSItoFigure4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4484150/v1/6d08039a8a5d3a587c246d3e.jpg"},{"id":58753202,"identity":"79c6ede6-6a64-47ae-b4a4-88227975a17c","added_by":"auto","created_at":"2024-06-20 16:17:12","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":453531,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"EPSItoFigure5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4484150/v1/4f3682dfd5507747194e07fd.jpg"},{"id":58753199,"identity":"54eab7fd-710c-4183-a632-ce21eacfd9d0","added_by":"auto","created_at":"2024-06-20 16:17:12","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":395625,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"EPSItoFigure6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4484150/v1/9853f605b1ffae07bd3e9c1b.jpg"},{"id":58753205,"identity":"f695701c-b1fe-49fc-a4d7-044690e5be1f","added_by":"auto","created_at":"2024-06-20 16:17:12","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":418945,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"EPSItoFigure7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4484150/v1/b486750b9c876b5a641134e9.jpg"},{"id":58754248,"identity":"5f5f0b45-d288-4116-82cc-83e017262abb","added_by":"auto","created_at":"2024-06-20 16:25:12","extension":"jpg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":374722,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"EPSItoFigure8.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4484150/v1/400f4812dad6f32dabae6b36.jpg"},{"id":58753195,"identity":"39cada6d-7004-4698-a5b2-9a732fcad9db","added_by":"auto","created_at":"2024-06-20 16:17:12","extension":"jpg","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":453096,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"EPSItoFigure9.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4484150/v1/765a78f46b23d392b840f654.jpg"},{"id":58753206,"identity":"6a1eeaa0-1070-4bbe-807a-fd75db684978","added_by":"auto","created_at":"2024-06-20 16:17:13","extension":"jpg","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":318104,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"EPSItoFigure10.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4484150/v1/d7ef682373038fdef3672ce5.jpg"},{"id":58753198,"identity":"c3fa1f95-f1a3-4514-ada4-ff42880cc470","added_by":"auto","created_at":"2024-06-20 16:17:12","extension":"jpg","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":355923,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"EPSItoFigure11.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4484150/v1/676883561df465c540cbb03a.jpg"},{"id":58754251,"identity":"e43c40ff-fb00-47df-a21a-34bb76164c28","added_by":"auto","created_at":"2024-06-20 16:25:12","extension":"jpg","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":579210,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"EPSItoFigure12.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4484150/v1/4b735fb1c78f98ba2d472fda.jpg"},{"id":58753200,"identity":"2eb67b17-e1d7-4054-b9b4-6436d8f065ec","added_by":"auto","created_at":"2024-06-20 16:17:12","extension":"jpg","order_by":13,"title":"Figure 13","display":"","copyAsset":false,"role":"figure","size":547751,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"EPSItoFigure13.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4484150/v1/408ecb051e94ec9d8bf8f555.jpg"},{"id":69834943,"identity":"cd5f57ea-959e-41f7-9853-ed9ce45a3baa","added_by":"auto","created_at":"2024-11-25 16:10:38","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":5984666,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4484150/v1/6a36785a-d101-4492-82e1-d7de8f0a96cb.pdf"},{"id":58754250,"identity":"ccdbbfd2-700f-47dd-b293-6380b1aa683d","added_by":"auto","created_at":"2024-06-20 16:25:12","extension":"jpg","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":117802,"visible":true,"origin":"","legend":"","description":"","filename":"EPSItoGraphicalAbst.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4484150/v1/3af6dfdbc0bcbb1fc8caa765.jpg"}],"financialInterests":"","formattedTitle":"Preparatory Process in Advance of Runaway Fault Rupture through Fluid Injection Observed in Laboratory Experiments Using a Large Specimen of Sub-meter scale","fulltext":[{"header":"Introduction","content":"\u003cp\u003eEarthquakes occur when elastic waves pass through subsurface rocks, resulting in shaking of the ground. The elastic waves are produced by unstable deep fault slips, which are induced naturally when shear stress exceeds static friction on faults under tectonic loading. Unstable slips are also induced artificially by injecting fluids into rocks for fracturing (Ellsworth, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Morris et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Kim et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). As earthquakes can threaten socio-economic activities of humans, the mechanisms underlying earthquake occurrence have been actively examined. As in situ observations of unstable slips at deep depths are difficult, a number of attempts have been made to reproduce these slips in the laboratory (Lockner et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e1982\u003c/span\u003e; Ohnaka, 1986; Bartlow et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Gu\u0026eacute;rin-Marthe et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Selvadurai \u0026amp; Glaser, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Yamashita et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Wang et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Gori at al., 2021; Cebry \u0026amp; McLaskey, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Ji et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Oye at al., 2022). However, laboratory studies for injection-induced earthquakes are limited compared with those for natural earthquakes, possibly because of the following issues on specimens and observation methods of fault slips in laboratory experiments.\u003c/p\u003e \u003cp\u003eIn the case of injection-induced earthquakes, fault slip is initiated by fluid pressure applied locally, and it sometimes expands unstably over a wide area such that it results in elastic waves having magnitudes that can induce felt earthquakes. Thus, reproducing the unstable expansion of initial slips in laboratory is important; however, it is hard to reproduce the process by general setup of laboratory experiment such as triaxial loading tests on cylindrical specimens with inclined faults, which have been used for the laboratory studies of natural earthquakes (Bartlow et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Gu\u0026eacute;rin-Marthe et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The specimens used for those studies are generally less than 100 mm in diameter, and if such specimens are used to reproduce injection-induced earthquakes, an inlet hole for fluid injection into fault is to be located in the near distance of 50 mm or less from the fault edge. However, fluid pressure propagates quickly over such short distance, making it unfeasible to observe the expansion of initial slips.\u003c/p\u003e \u003cp\u003eAnother issue related to laboratory experiments is the observation method of fault slips. In laboratory experiments, faults need to be prepared considering the limited allowable size in laboratories. However, slip displacement of faults is proportional to the fault size, and the limited allowable size of fault would make fault slips to be extremely small. Contrastingly, fault slips cause changes in the shear strain on fault. Shear strain is independent of fault size and can be measured easily using strain gauges. Therefore, observing fault slips based on shear strain than displacement is relatively easier in laboratory experiments. This approach was applied by Ohnaka (1986) and allowed to demonstrate the quasistatic and stable nucleation prior to fault instability in natural earthquake studies. The experiment was conducted using a specimen of quadrilateral plate separated into two halves by a sawcut fault. The fault slip was induced by applying compressive stresses to the lateral edges of the specimen. Subsequently, the changes in shear strain associated with fault slip were measured by strain gauges attached on the side of the specimen located close to the fault. This was suitable for an experiment using the fault long in the direction of fault slipping but narrow in its orthogonal direction. However, such fault geometry is not appropriate for conducting experiments of injection-induced earthquakes since fluid pressure propagates quickly over such narrow fault widths as in the case of small cylindrical specimens. The issue of rapid pressure propagation can be avoided by thickening the specimen; however, observing the initiation of fault slips inside the specimen will be difficult by strain gauges attached on the side surface of specimen.\u003c/p\u003e \u003cp\u003eConsidering these issues, in the present study, we adopted a novel laboratory-based experiment approach to examine fault behavior driven by localized-high fluid pressure. We prepared a cubic specimen separated into two triangular prisms by a model fault with a wide area of about 600 \u0026times; 700 mm\u003csup\u003e2\u003c/sup\u003e. The specimen was subjected to bi-axial compressions with different magnitudes. A two-dimensional (2D) array of strain gauges was embedded beneath the fault plane to measure the changes in shear strain with the fault slip driven by fluid injection. Based on the experimental results, we discussed the features of fault slip that lead to injection-induced earthquakes. As a result, we found that the observed behavior was described well by the model of injection-induced aseismic slip which transmits an elastic perturbation that triggers seismicity beyond the fluid-pressurized zone. The model has been presented conceptually upon field and laboratory experiments (Elsworth et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; De Barros et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Bhattchacharya and Viesca, 2019; Cappa et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Ji et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) and examined numerically in detail through hydromechanical modelling of a slip-weakening and permeable fault response to a fluid injection (Norbeck and Horne, 2017; Wynants-Morel et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; De Barros et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eSet up of the laboratory experiment\u003c/h2\u003e \u003cdiv id=\"Sec4\" class=\"Section3\"\u003e \u003ch2\u003eExperimental system\u003c/h2\u003e \u003cp\u003eThe experimental system used in this study is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e1\u003c/span\u003e. The system comprised a loading frame to apply compressive loads to a specimen, electric and hand oil pumps to supply oil pressure for the loading, an electric syringe pump to inject fluid into a model fault of specimen, and a data monitoring and acquisition system. The loading frame accommodated specimens with dimensions 600 \u0026times; 600 \u0026times; 600 mm\u003csup\u003e3\u003c/sup\u003e. The specimen was subjected independently to biaxial compressive stresses \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e in horizontal and vertical directions, respectively, by two pairs of flat jacks. The oil pressure for activating the flat jacks was supplied in principle by an electric pump, and a hand oil pump was used for finely adjusting the oil pressure. The pressure in the pair of flat jacks was equal to the compressive stress subjected to the specimen by the flat jacks.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eSpecimen containing a fault with embedded strain gauges\u003c/h2\u003e \u003cp\u003eThe specimen was a 600 \u0026times; 600 \u0026times; 600 mm\u003csup\u003e3\u003c/sup\u003e block of the Mogami andesite with Young\u0026rsquo;s modulus (\u003cem\u003eE\u003c/em\u003e) value of 16.3 GPa and a Poisson\u0026rsquo;s ratio (\u003cem\u003en\u003c/em\u003e) of 0.2 (Ito et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The model fault in the specimen consisted of a throughout-going sawcut oriented at an angle of 45\u0026deg; to the sides of the cubic block. Thus, the sawcut split the block diagonally in two equal halves above and below the fault plane, which are referred to the upper and lower blocks, respectively (Figs.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003ea and b). The normal stress distribution on the fault was checked using a pressure sensitive paper. When the normal stress in a region on the fault was higher than in other regions, the high normal stress region was ground. By repeating these, the fault planes were carefully prepared so that they were in contact as uniformly as possible and the normal stress in the middle region of the fault was rather slightly higher than in its top and bottom edges. We discussed further the normal stress distribution in the \u003cspan refid=\"Sec10\" class=\"InternalRef\"\u003ediscussion\u003c/span\u003e section 4.3.\u003c/p\u003e \u003cp\u003eThe compressive stresses \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e applied by the biaxial loading frame caused normal and shear stresses \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e on the fault plane given as follows:\u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e\n$${\\varvec{\\sigma }}_{\\varvec{F}}=\\frac{{\\varvec{S}}_{\\varvec{x}}+{\\varvec{S}}_{\\varvec{y}}}{2}, {\\varvec{\\tau }}_{\\varvec{F}}=\\frac{{\\varvec{S}}_{\\varvec{x}}-{\\varvec{S}}_{\\varvec{y}}}{2}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eThese relationships show that \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e can be increased by maintaining \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e, when \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e are increased together by the same amount; conversely, \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e can be increased by maintaining \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e, when \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e is increased but \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e is decreased by the same amount.\u003c/p\u003e \u003cp\u003eAn inlet port was set at the center of the fault plane of the upper block for injecting a fluid in the fault (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003ea). The fluid was supplied by a syringe pump through a tube set in a borehole drilled in the upper block. A groove with width and depth of 5 mm was dug along the fault plane of the upper block in the transverse direction as the groove crossed the inlet port. The groove end was separated by ~\u0026thinsp;80 mm away from the open surface of the specimen block, and furthermore, a long rubber seal was installed along the fault edge to prevent fluid outflow from the fault. Thus, the fluid pressure in the groove was not leaked easily to the open surface and maintained while injection. This arrangement ensured that the fluid injected into the fault spread first along the transverse groove and flowed out in parallel with the longitudinal direction of the fault. This flow of fluid will cause a pressure distribution uniform in the transverse direction of the fault. The setup was designed for normal and shear stresses (\u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e, respectively) to be also uniform in the transverse direction of the fault. These loading conditions simplified the fault slip to be uniform in the transverse direction of the fault; additionally, the simplification facilitated the observation and understanding of the fault slip.\u003c/p\u003e \u003cp\u003eFurthermore, a 2D array of strain gauges was embedded beneath the fault plane to measure the changes in the shear strain with the fault slip driven by fluid injection. To achieve this, three narrow grooves (F, C, and B) were dug along the fault plane according to the arrangement described in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e3\u003c/span\u003ea. These grooves were separated from each other by 140 mm and oriented along the longitudinal direction of the fault. Strain gauges (KFGS-2-120-D31-11, Kyowa Electronic Instruments, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e3\u003c/span\u003eb, were attached to the side wall of the grooves 5 mm below the fault plane (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e3\u003c/span\u003ec). The strain gauge consisted of a pair of matched gauges with mutually perpendicular orientations to directly monitor the shear strain parallel to the fault plane at each attached location. Notably, the shear strain could be converted to the shear stress by multiplying it and the rigidity of the rock sample, i.e., \u003cem\u003eE\u003c/em\u003e/2(1\u0026thinsp;+\u0026thinsp;\u003cem\u003en\u003c/em\u003e). The grooves were filled with epoxy resin after attaching the gauge. The resin exhibited a low value of Young\u0026rsquo;s modulus (~\u0026thinsp;3 GPa), which did not affect the deformation of the specimen rock. Lead-wires of the strain gauges were placed along the grooves and were connected to a data logger (UCAM-550A; Kyowa Electronic Instruments, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). In total, 39 strain gauges were placed, of which 13 were placed in each groove. The data logger synchronously sampled for all channels by a sampling frequency of 10 Hz. The oil pressure in the flat jacks of the biaxial loading frame was acquired by the same logger to eliminate errors in the measurement time of each dataset. The type of strain gauge and the sampling frequency in this study were not designed to detect motions with small amplitudes and high frequencies as of elastic waves. To indicate the location of each strain gauge, we assigned a name in the format of \u0026ldquo;F/C/B\u0026rdquo; and \u0026ldquo;+/-\u0026rdquo; and \u0026ldquo;0/1/2 \u0026hellip; /6\u0026rdquo;, where the first letter indicates the respective name of the groove on which the gauge was installed, and the second letter indicates the upper (+) and lower (-) sides, respectively, where the gauge was installed relative to the center line across the fault, and the third letter indicates the distances of 0, 20, 40, 60, 80, 120, and 160 mm from the center line across the fault (see Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e3\u003c/span\u003ea), where the gauge was installed.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eExperiment I: Shear stress loading under constant normal stress\u003c/h2\u003e \u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e4\u003c/span\u003e shows the time variations in \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e, and Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e5\u003c/span\u003ea shows the resultant stresses \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e in Experiment I. The compressive stresses, \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e, were increased together first from 0 MPa to 6.5 MPa at the same rate. This manner of loading resulted in \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e with the same magnitude as that of \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e to maintain \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e at zero. After \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e increased to 6.5 MPa, \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e increased but \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e decreased at the same rate. This method of loading increased \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e maintaining \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e at 6.5 MPa. To achieve this loading, the release valve was opened slightly to gradually decrease the pressure in the pair of the flat jacks for applying \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e. Simultaneously, the hand pump was operated manually to increase the pressure in the pair of the flat jacks to apply \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e at a rate similar as much as possible to the rate of decreasing \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e. The stress variation fluctuated to some extent because the hand pump was operated manually.\u003c/p\u003e \u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e6\u003c/span\u003e shows the time variation in the shear strains measured by the 2D array of strain gauges embedded in the fault plane, where the time range was limited to 550\u0026ndash;660 s. For the ease of comparison, the plotted values of strain represented the amount of change from the value at a reference time of 550 s. Each graph in Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e6\u003c/span\u003e shows the variation measured by one strain gauge and is plotted in accordance with the alignment of corresponding stain gauge in the 2D array. The scale of the vertical axis for strain was the same for all graphs in the figure. The results indicate the dynamic behavior of the fault. A sudden change in the shear strain occurred at 645 s. Subsequently, the shear strains suddenly dropped in all graphs, while the strains increased conversely at the limited locations around B\u0026thinsp;+\u0026thinsp;1 and B0. This phenomenon can be explained as follows: A dynamic slip of the fault plane occurred entirely at 645 s. Later, the shear stress released by the slip concentrated around B\u0026thinsp;+\u0026thinsp;1 and B0, since the normal stress on the fault and the resultant friction were larger at these points than at others.\u003c/p\u003e \u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e5\u003c/span\u003eb shows the time variations in the normal and shear stresses on the fault plane for the time range of 550\u0026ndash;660 s. \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e were estimated using Eq.\u0026nbsp;(\u003cspan refid=\"Equ1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) from \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e, which were equal to the oil pressure in the two pairs of flat jacks, respectively. At 645 s, when the measured shear strains changed abruptly, \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e also dropped suddenly by a small amount of ~\u0026thinsp;0.1 MPa, while \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e remained constant around that time. Such changes in \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e indicated the occurrence of the dynamic slip of the entire fault for the following reasons: When the entire fault slipped, the distance between the left and right sides of the specimen shortened, and simultaneously the distance between the top and bottom sides of the specimen was extended by the same amount, considering the fault was oriented at 45\u0026deg; to the sides of the specimen. Consequently, the flat jacks for applying \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e expanded to decrease the inner pressure and the stress, \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e, and those for applying \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e compressed to increase the inner pressure and the stress, \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e, by the same amount. Such sudden changes in \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e abruptly decreased \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e, but \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e did not change following Eq.\u0026nbsp;(\u003cspan refid=\"Equ1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Figure\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e5\u003c/span\u003eb shows these changes in \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e, and this result supported the occurrence of the dynamic slip of the entire fault at 645 s. The shear stress, \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e, was 2.4 MPa just before it dropped suddenly.\u003c/p\u003e \u003cp\u003eInteresting and locally different features were observed in the strain variation (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e6\u003c/span\u003e) before the dynamic slip of the entire fault. During the early period, the shear strain increased monotonically at all instances, while the increasing rate changed with locations, with high rates around the middle region but low rates around the top and bottom edges of the 2D array of strain gauges. During the middle period, the time variation in the strains bended, and its slope reduced. This change was associated with the decrease in the increasing rate of \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e from ~\u0026thinsp;600 s (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e5\u003c/span\u003eb). However, during the later period, although the increasing rate of \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e was maintained, the strains changed with time in different manners at different locations. Around the top and bottom edges, the shear strain became constant or decreased gradually. Contrastingly, in the middle region, the shear strain continued to increase rather more steeply. The strain increase in the middle region was considered to have occurred for compensating the decay of the strain accumulation around the top and bottom edges. The phenomena of local and steep increase in strain might have demonstrated the formation of the asperity in seismology. The steep increase in strain followed the slowdown of the increasing rate immediately before the dynamic slip of the entire fault. The slowdown might be associated with the slow slip in the asperity.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eExperiment II: Injecting fluid under sub-critical stress conditions\u003c/h2\u003e \u003cp\u003eAfter disassembling the specimen to check and clean the fault planes, we reinstalled the specimen in the loading frame and conducted Experiment II. Figures\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e5\u003c/span\u003ec shows the time variation in \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e, and injection pressure \u003cem\u003eP\u003c/em\u003e. Similar to the procedure of Experiment I, \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e started increasing gradually while maintaining \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e at 6.5 MPa. The time variation of the measured shear strains is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e7\u003c/span\u003e from 1900 to 2060 s, in the manner same as that described in Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e6\u003c/span\u003e. The shear strain increased monotonically in the initial period, and subsequently, the strain started concentrating around the middle region of the 2D array of the strain gauges. This trend was similar to that observed in Experiment I before the dynamic slip occurred in the entire fault. Therefore, the shear stress loading ended at 1980 s and was held by closing the valve on the tube that connected the oil pump and flat jacks. Subsequently, the value of \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e was estimated as 2.3 MPa (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e5\u003c/span\u003ed), which was only 0.1 MPa below the critical value (2.4 MPa) required for initiating the dynamic slip of the entire fault in Experiment I. This result indicated good reproductivity of the experiment including the persistent occurrence of the asperity.\u003c/p\u003e \u003cp\u003eIn the next step, fluid injection into the fault started from 2040 s, and it continued until the pressure reached 5 MPa (Figs.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e5\u003c/span\u003ec, d). We used a water-based fluid, with 130 mPa\u0026middot;s viscosity, for injection. The fluid was prepared by adding a small amount of xanthan gum to water (Petri, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). It was expected that the viscosity of the fluid facilitated localized pressure around the injection port because of increasing flow resistance through the fault. At ~\u0026thinsp;2045 s, that is 5 s after starting fluid injection, the shear strain changed suddenly, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e7\u003c/span\u003e. Figure\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e5\u003c/span\u003ed shows the time variation in \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e from 1900 to 2060 s. As shown in the figure, the shear stress suddenly dropped at 2045 s and was consistent with the time when the shear strain changed suddenly. The sudden change in \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e and the measured strain indicated that the dynamic slip of the entire fault occurred at 2045 s in this experiment. The strain change and fault slip before the dynamic slip of the entire fault have been discussed comprehensively later.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eExperiment III: Injecting fluid under sub-critical stress conditions\u003c/h2\u003e \u003cp\u003eUpon checking and cleaning the fault planes, we conducted Experiment III. \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e was set as 6.5 MPa as in Experiment II, but \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e was set as 2.0 MPa, which was 13% smaller than that in reported in Experiment II (2.3 MPa). Figure\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e5\u003c/span\u003ee shows the time variation in \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e, and \u003cem\u003eP\u003c/em\u003e. The fluid injection into the fault started from 2710 s in the same manner as in Experiment II by maintaining \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e values. In ~\u0026thinsp;8 s, the injection pressure reached 5 MPa, which was maintained thereafter. The time variation from 2600 to 2740 s in the measured shear strains is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e8\u003c/span\u003e. The shear strains were stable until fluid injection. Some strain variations occurred later; for example, the strain decreased at C-1 and C-2 but increased at C0, C\u0026thinsp;+\u0026thinsp;1, and C\u0026thinsp;+\u0026thinsp;2. However, the variation was limited in the middle region of the 2D array of strain gauges, and sudden changes in the strain were not reported at all instances as observed previously in Experiments I and II. Similar was the case for the time variations in \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e5\u003c/span\u003ef, from 2600 to 2740 s. \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e was stable from 2710 s for more than 10 s after starting fluid injection. Based on these observations, we concluded that the dynamic slip of the entire fault did not occur in this experiment.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eFault slipping estimated from observed distribution of shear strain\u003c/h2\u003e \u003cp\u003eFault behavior associated with fluid injection has been previously examined in numerical studies of Norbeck \u0026amp; Horne (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) and Wynants-Morei et al. (2020). These studies identified a general feature related to the fault slip, shear stress, and pressure changes along the fault. A typical example is shown in Figs.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e9\u003c/span\u003ea - c, which was obtained by our simulation on a 2D hydromechanical modelling of a slip-weakening and permeable fault response to fluid injection. The theories for the modelling are basically the same as those of Norbeck \u0026amp; Horne (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) and Wynants-Morei et al. (2020). It was assumed for the simulation that the fault was subjected initially to shear stress, normal stress and fluid pressure of 2.5, 17.5 and 10 MPa, respectively, a local fluid injection into the fault increased the pressure linearly to 15 MPa, and the friction coefficient decreased from 0.6 down to 0.55 with fault slip velocity.\u003c/p\u003e \u003cp\u003eThe results of Figs.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e9\u003c/span\u003ea - c demonstrated the role of fluid and friction on the injection-induced seismicity described conceptually in other previous studies (Elsworth et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; De Barros at al., 2018; Bhattchacharya and Viesca, 2019; Cappa et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Ji et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Injection pressure propagated from the injection point with time as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e9\u003c/span\u003ea. The increased pressure reduced friction between fault planes and induced fault slipping. The fault slip released shear stress locally, and the released shear stress led to their accumulation around the edge of the slipping fault area for overall stress balance (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e9\u003c/span\u003eb). Conversely, the time variation in shear stress is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e9\u003c/span\u003ec for four locations at different distances from the injection point. The results showed two types of variations (A and B) as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e9\u003c/span\u003ed, depending on the location. Type A variation appeared around locations where the fault slip initiated, and Type B variation appeared at locations where they departed from the location of slip initiation, and the fault slip occurred later. In Type A variation, the shear stress decreased, approaching a lower limit associated with the dynamic friction of the fault. Type B variation showed a peak value during the middle period, and during the peak period, the edge of the fault slipping area passed the location. The peak value increases gradually as its location departs away from the location of slip initiation. Note that shear strain changed in the same manner with shear stress, since they were directly proportional to each other. This implied that the time variation in shear strain could be characterized in the same manner as described in Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e9\u003c/span\u003ed.\u003c/p\u003e \u003cp\u003eWe compared the expected stress/strain variation of Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e9\u003c/span\u003ed with the observations of Experiment II. Figure\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e10\u003c/span\u003e shows the details of strain change during the fluid injection for Column F of the 2D array of strain gauges, where the scale of the vertical axes for strain was the same for all graphs in the figure but different from that of Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e7\u003c/span\u003e. The time variations in \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003eP\u003c/em\u003e have been shown together for comparison. Figure\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e10\u003c/span\u003e shows that the shear stress suddenly dropped at 2045.5 s accordingly to the dynamic slip of the entire fault. Note that in the figure, \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e does not seem to decrease instantaneously; that is, it decreased along a slope with not an infinite but a finite gradient, since the data were sampled and plotted discretely at the sampling frequency of 10 Hz. The shear strain at F0 and F-1 decreased from ~\u0026thinsp;2042.7 s similar to that for Type A variation; therefore, the fault slip was considered to initiate from a region around F0 and F-1 at that time. The shear strain at the other locations around F0 and F-1 varied in the manner of Type B variation. Each variation showed a peak value, which appeared later, as the location was away from F0 and F-1. This indicated that the fault slipping area expanded gradually with time far from F0 and F-1 to F\u0026thinsp;+\u0026thinsp;4 and F-4 following the peak as indicated by arrows in Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e10\u003c/span\u003e. Subsequently, the dynamic slip of the entire fault occurred at 2045.5 s, which was ~\u0026thinsp;3 s later after the slip initiated. The propagation speed of the Type B variation was ~\u0026thinsp;20\u0026ndash;30 mm/s on average, which is much slower than a typical speed of the dynamic rupture propagation (~\u0026thinsp;1.3 km/s; 80% of S-wave speed). These suggested the presence of a quasi-static or aseismic preparatory process before the dynamic or seismic fault slipping induced by fluid injection that took few seconds.\u003c/p\u003e \u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e11\u003c/span\u003e shows a photograph of the fault plane captured immediately after the experiment. The strain gauges were installed along three parallel grooves filled with epoxy resin, which corresponded to the three columns, F, C, and B, of the 2D array of strain gauges. The gauge locations are indicated by black dots. The area where the injected fluid invaded along the fault can be seen clearly from the color differences. The invasion area almost reached the F\u0026thinsp;+\u0026thinsp;4 and F-4 locations along Column F. This result was consistent with the extension of fault slipping area estimated as above from the variation in the measured strain. Moreover, the fluid invasion area was ~\u0026thinsp;20% of the area of the entire fault. Thus, in the dynamic slip of the entire fault, the fault slip area expanded instantaneously from 20 to 100% beyond the fluid invasion area. Further, since the dynamic slip of the entire fault occurred\u0026thinsp;~\u0026thinsp;5 s after the fluid injection started at 2040 s, a migration speed of the injected-fluid front was estimated to be ~\u0026thinsp;16 mm/s on average. This speed was smaller than the propagation speed of ~\u0026thinsp;20\u0026ndash;30 mm/s of Type B variation initiated at ~\u0026thinsp;2042.7 s. Thus, the aseismic slip initiated\u0026thinsp;~\u0026thinsp;3 s after the fluid injection, the slip area expanded faster than the injected-fluid area, and the seismic slip occurred at last when the slip front got across the injected-fluid front. Such a fault behavior follows well the model proposed recently on field experiments and numerical simulations (Elsworth, 2016; De Barros at al., 2018; Bhattchacharya and Viesca, 2019; Cappa et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Ji et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The model shows that the injection-induced aseismic slip transmits an elastic perturbation that triggers seismicity beyond the fluid-pressurized zone.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eConditions for runaway rupture\u003c/h2\u003e \u003cp\u003eFluid injection was conducted in both Experiments II and III. However, the dynamic slip of the entire fault occurred in the former experiment but not in the latter. It is reasonable to consider that the dynamic slip was associated with a difference in the experimental condition between them, which was the shear stress \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e set before fluid injection. In Experiment II, the \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e value was set at 2.3 MPa, indicating that the fault was ready to slip entirely. As described about the results in Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e10\u003c/span\u003e in the previous section, the shear strain around the edge of the fault slipping area locally increased compensating the reduction in shear strain supported on the fault plane along the slip area. The amount of shear strain that increased locally tended larger as the slip area became larger. Such tendency was consistent with the numerical simulations in Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e9\u003c/span\u003ec. When the fault slip area started to expand instantaneously beyond the fluid invasion area, the amount of shear strain that increased locally was ~\u0026thinsp;10 \u0026micro;e in shear strain, which was equivalent to ~\u0026thinsp;0.1 MPa in shear stress. This amount was small but sufficient for the total sum of shear stress to exceed the static friction on the fault. Consequently, if the slip area expanded further, the shear stress would increase around the edge of the expanded slip area. These coupled phenomena will occur spontaneously without additional fluid injection, in contrast to the preparatory process that was driven by the fluid pressure, as illustrated in Figs.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e12\u003c/span\u003ea\u0026ndash;d, where \u003cem\u003e\u0026micro;\u003c/em\u003e\u003csub\u003e\u003cem\u003es\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003e\u0026micro;\u003c/em\u003e\u003csub\u003e\u003cem\u003ed\u003c/em\u003e\u003c/sub\u003e (assuming \u003cem\u003e\u0026micro;\u003c/em\u003e\u003csub\u003e\u003cem\u003es\u003c/em\u003e\u003c/sub\u003e \u0026gt; \u003cem\u003e\u0026micro;\u003c/em\u003e\u003csub\u003e\u003cem\u003ed\u003c/em\u003e\u003c/sub\u003e) are the static and dynamic friction coefficients of the fault, respectively, and \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003ef\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003ef\u003c/em\u003e0\u003c/sub\u003e are effective normal stress and its initial value, respectively. Subsequently, the slip area will expand entirely over the fault assuming that the shear stress is nearly balanced with the static friction on the fault. This type of fault slipping has been referred as the runaway rupture (Norbeck \u0026amp; Horne, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Ji et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The mechanism shown in Figs.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e12\u003c/span\u003ea\u0026ndash;d suggests that the runaway rupture tends to occur more easily, as the shear stress on fault approaches the static friction and the dynamic friction departs to be smaller from the static one. The numerical simulation of Norbeck \u0026amp; Horne (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) accordingly showed that the critical state of shear stress on fault is not sufficient but the friction condition of \u003cem\u003e\u0026micro;\u003c/em\u003e\u003csub\u003e\u003cem\u003es\u003c/em\u003e\u003c/sub\u003e \u0026gt; \u003cem\u003e\u0026micro;\u003c/em\u003e\u003csub\u003e\u003cem\u003ed\u003c/em\u003e\u003c/sub\u003e is also necessary for the runaway rupture. The dynamic slip of the entire fault in Experiment II possibly occurred following the sequence shown in Figs.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e12\u003c/span\u003ea\u0026ndash;d. However, in Experiment III, the shear stress was set 2 MPa before fluid injection with a difference of 0.3 MPa from that of Experiment II. The difference appears to be too large to be overcome by the accumulation of shear stress around the edge of slip area. Therefore, the dynamic slip of the entire fault did not occur possibly in Experiment III.\u003c/p\u003e \u003cp\u003eWynants-Morel et al. (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) discussed in detail the effect of initial shear stress \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e0\u003c/sub\u003e on the injection-induced slip of fault. To do this, they carried out numerical simulations changing the initial Shear Capacity Utilization, SCU, from 53\u0026ndash;71%. The SCU is defined as the value of \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e0\u003c/sub\u003e divided by the static friction \u003cem\u003ef\u003c/em\u003e\u003csub\u003e0\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;\u003cem\u003e\u0026micro;\u003c/em\u003e\u003csub\u003e\u003cem\u003es\u003c/em\u003e\u003c/sub\u003e (\u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e0\u003c/sub\u003e \u0026ndash; \u003cem\u003eP\u003c/em\u003e\u003csub\u003e0\u003c/sub\u003e), and \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e0\u003c/sub\u003e and \u003cem\u003eP\u003c/em\u003e\u003csub\u003e0\u003c/sub\u003e are the initial normal stress on fault and the initial fluid pressure in fault, respectively. In their all cases of simulation, the pressure front and the Type B variation of shear stress moved away with time from the injection point similarly to Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e9\u003c/span\u003ea-c. However, two different behaviors were observed depending on SCU. For the cases with SCU less than 64%, the Type B variation remained behind the pressure front. On the other hand, for the cases with SCU larger than 68%, the Type B variation was first behind the pressure front, then the Type B variation accelerated and outpaced the pressure front. The reason was concluded that for the cases with SCU larger than 68%, the \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e0\u003c/sub\u003e was so large to exceed the dynamic friction \u003cem\u003ef\u003c/em\u003e\u003csub\u003e\u003cem\u003ed\u003c/em\u003e0\u003c/sub\u003e = \u003cem\u003e\u0026micro;\u003c/em\u003e\u003csub\u003e\u003cem\u003ed\u003c/em\u003e\u003c/sub\u003e (\u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e0\u003c/sub\u003e \u0026ndash; \u003cem\u003eP\u003c/em\u003e\u003csub\u003e0\u003c/sub\u003e) and then the stress drop (\u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e0\u003c/sub\u003e \u0026ndash; \u003cem\u003ef\u003c/em\u003e\u003csub\u003e\u003cem\u003ed\u003c/em\u003e0\u003c/sub\u003e) provided additional potential energy to drive the slip front growth. These results of Wynants-Morel et al. (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) and the result of Norbeck \u0026amp; Horne (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) described above fit well with the results of our laboratory experiments assuming that the \u003cem\u003ef\u003c/em\u003e\u003csub\u003e0\u003c/sub\u003e and \u003cem\u003ef\u003c/em\u003e\u003csub\u003e\u003cem\u003ed\u003c/em\u003e0\u003c/sub\u003e took a value of ~\u0026thinsp;2.4 MPa and a smaller value between 2 and 2.3 MPa, respectively. Thus, the runaway rupture occurred in the case with \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e0\u003c/sub\u003e \u0026gt; \u003cem\u003ef\u003c/em\u003e\u003csub\u003e\u003cem\u003ed\u003c/em\u003e0\u003c/sub\u003e as Experiment II, but the rupture was limited around the fluid injection place in the case with \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e0\u003c/sub\u003e \u0026lt; \u003cem\u003ef\u003c/em\u003e\u003csub\u003e\u003cem\u003ed\u003c/em\u003e0\u003c/sub\u003e as Experiment III.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eShear strain/stress redistribution with loading\u003c/h2\u003e \u003cp\u003eHere, we discuss the variations in shear strain observed while setting the shear stress. In Experiment II, \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e increased to 2.3 MPa, maintaining \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e at 6.5 MPa. \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e estimated by Eq.\u0026nbsp;(\u003cspan refid=\"Equ1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) from the flat jack pressures, \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e, represent the stresses averaged over the fault plane. Normal and shear stresses varied locally more or less than the estimated values of \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e depending on the local conditions of contact between the fault planes. Thus, the normal and shear stresses acting locally on the fault were referred and denoted as the local normal stress \u003cem\u003es\u003c/em\u003e and the local shear stress \u003cem\u003et\u003c/em\u003e, respectively. For the normal stress \u003cem\u003es\u003c/em\u003e, its distribution was actually checked by using a pressure sensitive paper for a range of 2.5\u0026ndash;10 MPa (Prescale, LW type, Fujifilm Co., 2024). Then, the paper was inserted between the fault planes and compressed by applying \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003eS\u003c/em\u003e\u003csub\u003e\u003cem\u003ey\u003c/em\u003e\u003c/sub\u003e with the same value of 6 MPa to the specimen. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e13\u003c/span\u003e shows a photograph of the paper after the measurement, where the strain gauge locations are indicated by black dots. The color appears red where normal stress is applied, and the color density varies according to the amount of normal stress. Note that no color indicates not no stress but a stress less than the lower limit of measurement range, i.e. 2.5 MPa in this case. It can be seen from Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e13\u003c/span\u003e that \u003cem\u003es\u003c/em\u003e was distributed uniformly in the middle region between +\u0026thinsp;4 and \u0026minus;\u0026thinsp;4 locations of the strain gauges overlapping with the region of injected-fluid invasion in Experiment II (see Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e11\u003c/span\u003e), while it became smaller around +\u0026thinsp;6 and \u0026minus;\u0026thinsp;6 locations of the strain gauges.\u003c/p\u003e \u003cp\u003eIn the summarized process for Experiment II shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e7\u003c/span\u003e, the shear strain increased monotonically everywhere in the initial period, while the increasing rate changed with locations as it was high around the middle region between +\u0026thinsp;4 and \u0026minus;\u0026thinsp;4 locations of the strain gauges but low around +\u0026thinsp;6 and \u0026minus;\u0026thinsp;6 locations of them. Subsequently, the strain began to concentrate around the middle region. The difference in the strain variation may be caused by the difference in local conditions of contact between the fault planes as appeared in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e13\u003c/span\u003e. In the regions where the fault planes were more firmly in contact, the local normal stress and static friction were larger, and the fault became harder to slip and accept more shear strain in the surrounding rock body. This phenomenon can be understood more clearly by replacing shear strain with shear stress in the directly proportional relationship. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e12\u003c/span\u003ee illustrates the paths of local stress changing at three locations A, B, and C on the fault. Stage 1 shows the stress state at the end of \u003cem\u003es\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e loading. \u003cem\u003et\u003c/em\u003e was zero everywhere, and \u003cem\u003es\u003c/em\u003e was assumed to differ at each location. With \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e loading, \u003cem\u003et\u003c/em\u003e also increased, and at Stage 2, \u003cem\u003et\u003c/em\u003e at the location A with the smallest \u003cem\u003es\u003c/em\u003e was the first to reach the static friction on the fault shown by the diagonal line with a slope of the static friction coefficient \u003cem\u003e\u0026micro;\u003c/em\u003e\u003csub\u003e\u003cem\u003es\u003c/em\u003e\u003c/sub\u003e. Subsequently, the fault slipped locally at A, and later, \u003cem\u003et\u003c/em\u003e at A could not increase further; however, \u003cem\u003et\u003c/em\u003e at locations B and C continued to increase rather at a higher rate since they transferred the shear stress, which should have been stored at A. The same phenomenon possibly occurred in the loading process of Experiments I and II. Thus, the strain varied in different manners in different locations. With further \u003cem\u003et\u003c/em\u003e\u003csub\u003e\u003cem\u003eF\u003c/em\u003e\u003c/sub\u003e loading, the stress state could reach Stage 3 wherein \u003cem\u003et\u003c/em\u003e was locally balanced with the static friction, while the value of the static friction somewhat differed at each location. In Experiment II, the stress state before fluid injection was considered at Stage 3, and therefore, the runaway rupture occurred even while the local normal stress was not uniform. This may also be observed for Experiment I. The strain varied until the occurrence of the dynamic slip of the entire fault similar to that in Experiment II. The stress state at Stage 3 was established finally and resulted in the runway rupture. Thus, Experiment I may demonstrate a type of preparedness process for natural earthquakes.\u003c/p\u003e \u003cp\u003eIn contrast to Experiments I and II, the stress state in Experiment III should be in between Stages 2 and 3. The stress state should be on the line of the static friction in top and bottom edges of the fault where the normal stress was low. On the other hand, although the shear stress in the middle of the fault might be high, it should be much lower than the static friction level, because the normal stress was high in this portion. Therefore, even when the normal stress was reduced by the fluid injection or the shear stress was increased by the accumulation in the middle of the fault, the preparatory process could not be sustained or grow to the runaway rupture. Therefore, Experiment III argues that we could mitigate risk of destructive earthquakes induced by injection, if we would measure the fault strength as well as the stress state along a fault in advance and would adjust the fluid pressure carefully. We argue here the same physics from opposite perspective argued in Wynants-Morel et al. (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) and De Barros et al. (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) that for a high-criticality stressed fault, the aseismic slip propagation is self-sustained and outpaces the pressurized zone, on the contrary, for a low-criticality stressed fault, the aseismic slip stays driven by the effective stress decrease and remains within the pressurized zone.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusions","content":"\u003cp\u003eIn this study, we successfully conducted experiments to detect a preparatory process in advance of a dynamic slip of the entire fault induced by fluid injection. We found that fault slipping area expanded gradually away from the location of fluid injection. Dynamic slip occurred few seconds after the initiation of fault slipping, and subsequently, the fault slip area expanded instantaneously from 20 to 100% of the model fault. The shear stress should have been nearly balanced with the static friction everywhere on the fault just before the dynamic slip. This critical condition should have been developed along with local slipping of the fault during shear stress loading on the fault. Under critical conditions, the fault slipping area expanded to increase the shear stress around its edge, and the increase drove its further expansion. These coupled phenomena occurred spontaneously without additional fluid injection. Further, the dynamic slip of entire fault possibly occurred after this sequence of events.\u003c/p\u003e \u003cp\u003eIn contrast, when the fault was loaded under subcritical conditions, the preparatory process induced by the fluid injection was quickly arrested, probably due to an insufficient shear stress accumulation in a high normal stress (high static friction) region on the fault. This suggests that the measurement of frictional properties of the fault in addition to the stress state along the fault should provide key information to reduce risks caused by earthquakes induced by the fluid injection.\u003c/p\u003e \u003cp\u003eNumerical simulations allow us to figure out fault slip behavior in detail. However, there is unavoidable ambiguity resulting from simplifications in its modelling and difficulties to give all simulation parameters correctly. In the present study, we succeeded to observe directly the behavior that the injection-induced aseismic slip transmitted along the fault an elastic perturbation to lead the seismic slip. This is the first attempt in the study of the injection-induced earthquake, to the author\u0026rsquo;s knowledge. The results and the proposed method of laboratory experiment will contribute for understanding the injection-induced earthquake.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe dataset of laboratory experiments is available at Ito et al. (2023).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interest.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by Japan Society for the Promotion of Science (JSPS KAKENHI) Grant Number 16H04612 and JP21H05201.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTI designed this study. KA and YM contributed to experiments and data analysis. YY joined the interpretation. All authors contributed to and approved the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors wish to thank Prof. D. Swenson, Mr. K. Kaneta, Mr. S. Inoue and Mr. K. Yokoyama for their assistance in the numerical simulations and the experiments. This work was supported by JSPS KAKENHI Grant Number 16H04612 and JP21H05201. We also wish to thank Editage (www.editage.com) for English language editing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; information\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003csup\u003e1\u003c/sup\u003eInstitute of Fluid Science, Tohoku University, Sendai, Japan. \u003csup\u003e2\u003c/sup\u003eResearch Center for Prediction of Earthquakes and Volcanic Eruptions, Graduate School of Science, Tohoku University, Sendai, Japan\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eBartlow NM, Lockner DA, Beeler NM (2012) Laboratory triggering of stick-slip events by oscillatory loading in the presence of pore fluid with implications for physics of tectonic tremor. J Geophys Research: Solid Earth 117(B11411). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1029/2012JB009452\u003c/span\u003e\u003cspan address=\"10.1029/2012JB009452\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBhattacharya P, Viesca RC (2019) Fluid-induced aseismic fault slip outpaces pore‐fluid migration. Science 364(6439). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1126/science.aaw7354\u003c/span\u003e\u003cspan address=\"10.1126/science.aaw7354\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCappa F, Scuderi MM, Collettini C, Guglielmi Y, Avouac J-P (2019) Stabilization of fault slip by fluid injection in the laboratory and in situ. Sci Adv 5(3). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1126/sciadv.aau4065\u003c/span\u003e\u003cspan address=\"10.1126/sciadv.aau4065\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCebry SBL, McLaskey GC (2021) Seismic swarms produced by rapid fluid injection into a low permeability laboratory fault. Earth Planet Sci Lett 557(116726). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.epsl.2020.116726\u003c/span\u003e\u003cspan address=\"10.1016/j.epsl.2020.116726\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDe Barros L, Wynants-Morel N, Cappa F, Philippe D (2021) Migration of fluid-induced seismicity reveals the seismogenic state of faults. J Geophys Research: Solid Earth 126(11). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1029/2021JB022767\u003c/span\u003e\u003cspan address=\"10.1029/2021JB022767\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDe Barros L, Guglielmi Y, Rivet D, Cappa F, Duboeuf L (2018) Seismicity and fault aseismic deformation caused by fluid injection in decametric in-situ experiments. CR Geosci 350(8). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.crte.2018.08.002\u003c/span\u003e\u003cspan address=\"10.1016/j.crte.2018.08.002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEllsworth WL (2013) Injection-induced earthquakes. Science 341(6142). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1126/science.1225942\u003c/span\u003e\u003cspan address=\"10.1126/science.1225942\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eElsworth D, Spiers CJ, Niemeijer AR (2016) Understanding induced seismicity. Science 354(6318). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1126/science.aal2584\u003c/span\u003e\u003cspan address=\"10.1126/science.aal2584\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFujifilm Co (2024) \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.fujifilm.com/us/en/business/industrial-materials/measurement-film/prescale\u003c/span\u003e\u003cspan address=\"https://www.fujifilm.com/us/en/business/industrial-materials/measurement-film/prescale\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e, April 16, 2024\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGori M, Rubino V, Rosakis AJ, Lapusta N (2021) Dynamic rupture initiation and propagation in a fluid-injection laboratory setup with diagnostics across multiple temporal scales. Proc Natl Acad Sci USA 118(51). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1073/PNAS.2023433118\u003c/span\u003e\u003cspan address=\"10.1073/PNAS.2023433118\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGu\u0026eacute;rin-Marthe S, Kwiatek G, Wang L, Bonnelye A, Mart\u0026iacute;nez-Garz\u0026oacute;n P, Dresen G (2023) Preparatory slip in laboratory faults: Effects of roughness and load point velocity. J Geophys Research: Solid Earth 128(4). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1029/2022JB025511\u003c/span\u003e\u003cspan address=\"10.1029/2022JB025511\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIto T, Kumazawa A, Tezuka K, Ogawa K, Yokoyama T, Funato A (2022) Experimental and numerical study on a two-stage coring method for stress measurement: Application to deep and high-temperature geothermal wells. Geothermics 100. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.geothermics.2021.102333\u003c/span\u003e\u003cspan address=\"10.1016/j.geothermics.2021.102333\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIto T, Aoki K, Mukuhira Y, Yabe Y (2023) Preparatory process in advance of runaway fault rupture through fluid injection observed in laboratory experiments using a. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.5281/zenodo.8348064\u003c/span\u003e\u003cspan address=\"10.5281/zenodo.8348064\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. large specimen of sub-meter scale [Data set]Zenodo\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJi Y, Hofmann H, Duan K, Zang A (2022) Laboratory experiments on fault behavior towards better understanding of injection-induced seismicity in geoenergy systems. Earth Sci Rev 226. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.earscirev.2021.103916\u003c/span\u003e\u003cspan address=\"10.1016/j.earscirev.2021.103916\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim K-H, Ree J-H, Kim Y, Kim S, Kang SY, Seo W (2018) Assessing whether the 2017 Mw 5.4 Pohang earthquake in South Korea was an induced event. Science 360. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1126/science.aat6081\u003c/span\u003e\u003cspan address=\"10.1126/science.aat6081\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKyowa Electronic Instruments (2023) \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.kyowa-ei.com/jpn/product/index.html\u003c/span\u003e\u003cspan address=\"https://www.kyowa-ei.com/jpn/product/index.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e, Sept. 2, 2023\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLockner DA, Okubo PG, Dieterich JH (1982) Containment of stick-slip failures on a simulated fault by pore fluid injection. Geophys Res Lett 9(8). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1029/GL009i008p00801\u003c/span\u003e\u003cspan address=\"10.1029/GL009i008p00801\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorris AP, Ferrill DA, Walter GR, Price AM, Smart KJ, Skoumal RJ, Brudzinski MR, Currie BS (2017) Lessons learned from the Youngstown, Ohio induced earthquake sequence from January 2011 to January 2012. J Rock Mech Geotech Eng 9(5). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jrmge.2017.03.016\u003c/span\u003e\u003cspan address=\"10.1016/j.jrmge.2017.03.016\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNorbeck JH, Horne RN (2018) Maximum magnitude of injection-induced earthquakes: A criterion to assess the influence of pressure migration along faults. Tectonophysics 733(9). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.tecto.2018.01.028\u003c/span\u003e\u003cspan address=\"10.1016/j.tecto.2018.01.028\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOhnaka M, Kuwahara Y, Yamamoto K, Hirasawa T (1986) Dynamic breakdown processes and the generating mechanism for high-frequency elastic radiation during stick-slip instabilities. In: S. Das, J. Boatwright, C.H. Scholz (Eds.), \u003cem\u003eEarthquake Source Mechanics\u003c/em\u003e, \u003cem\u003eAmerican Geophysical Union Geophysical Monograph Series\u003c/em\u003e 37. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1029/GM037p0013\u003c/span\u003e\u003cspan address=\"10.1029/GM037p0013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOhnaka M, Kuwahara Y (1990) Characteristics features of local breakdown near a crack-tip in the transition zone from nucleation to unstable rupture during stick-slip shear failure. Tectonophysics 175. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/0040-1951(90)90138-X\u003c/span\u003e\u003cspan address=\"10.1016/0040-1951(90)90138-X\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOye V, Stanchits S, Babarinde O, Bauer R, Dichiarante AM, Langet N, Goets-Allmann B, Frailey S (2022) Cubic\u0026ndash;meter scale laboratory fault re\u0026ndash;activation experiments to improve the understanding of induced seismicity risks. Sci Rep 12. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41598-022-11715-6\u003c/span\u003e\u003cspan address=\"10.1038/s41598-022-11715-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePetri DFS (2015) Xanthan gum: A versatile biopolymer for biomedical and technological applications. J Appl Polym Sci 132. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/APP.42035\u003c/span\u003e\u003cspan address=\"10.1002/APP.42035\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSelvadurai PA, Glaser SD (2017) Asperity generation and its relationship to seismicity on a planar fault: a laboratory simulation. Geophys J Int 208(2). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/gji/ggw439\u003c/span\u003e\u003cspan address=\"10.1093/gji/ggw439\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang L, Kwiatek G, Rybacki E, Bonnelye A, Bohnhoff M, Dresen G (2020) Laboratory study on fluid-induced fault slip behavior: The role of fluid pressurization rate. Geophys Res Lett 47(6). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1029/2019GL086627\u003c/span\u003e\u003cspan address=\"10.1029/2019GL086627\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWynants-Morel N, Cappa F, De Barros L, Ampuero JP (2020) Stress perturbation from aseismic slip drives the seismic front during fluid injection in a permeable fault. \u003cem\u003eJournal of Geophysical Research: Solid Earth\u003c/em\u003e 125. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003edoi.org/10.1029/2019JB019179\u003c/span\u003e\u003cspan address=\"10.1029/2019JB019179\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYamashita F, Fukuyama E, Xu S, Kawakata H, Mizoguchi K, Takizawa S (2021) Two end-member earthquake preparations illuminated by foreshock activity on a meter-scale laboratory fault. Nat Commun 12. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41467-021-24625-4\u003c/span\u003e\u003cspan address=\"10.1038/s41467-021-24625-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"earth-planets-and-space","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"epsp","sideBox":"Learn more about [Earth, Planets and Space](http://earth-planets-space.springeropen.com)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/epsp/default.aspx","title":"Earth, Planets and Space","twitterHandle":"@SpringerOpen","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Fracturing, Injection-induced earthquake, Fault slip, Laboratory experiment, Shear strain, Runaway rupture","lastPublishedDoi":"10.21203/rs.3.rs-4484150/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4484150/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eFault slip is initiated by locally applied fluid pressure, and it can expand unstably over a wide area causing elastic waves having magnitudes that induce felt or destructive earthquakes. Thus, it is important to examine the unstable expansion of initial slips. However, it is hard to reproduce the process by general setup of laboratory experiment such as triaxial loading tests on cylindrical specimens with inclined faults. In this study, we prepared a cubic specimen of sub-meter scale, which was separated into two triangular prisms by a model fault. The specimen was subjected to bi-axial compressions with different magnitudes. A 2D array of strain gauges was embedded beneath the fault plane to measure the changes in shear strain with the fault slip driven by fluid injection. Based on the experimental results, we discussed the features of fault slips that lead to injection-induced earthquake. The strain accumulated around the edge of the fault slipping area. The accumulation increased locally the strain by ~\u0026thinsp;10 \u0026micro;ε, which was equivalent to ~\u0026thinsp;0.1 MPa in shear stress. The fault slipping area expanded gradually first, and it expanded unstably beyond the fluid invasion area\u0026thinsp;~\u0026thinsp;3 s later after the slip initiated. The unstable expansion of initial slips was suppressed due to reducing the initial shear stress on the fault by 0.3 MPa. In this case, the initial shear stress was too small for the additional stress accumulated at the edge of the fault slipping area to overcome the static friction on the fault.\u003c/p\u003e","manuscriptTitle":"Preparatory Process in Advance of Runaway Fault Rupture through Fluid Injection Observed in Laboratory Experiments Using a Large Specimen of Sub-meter scale","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-06-20 16:17:07","doi":"10.21203/rs.3.rs-4484150/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Major Revision","date":"2024-06-25T00:41:26+00:00","index":"","fulltext":""},{"type":"reviewerAgreed","content":"","date":"2024-06-05T09:02:02+00:00","index":0,"fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-06-04T10:49:04+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-06-01T04:23:24+00:00","index":"","fulltext":""},{"type":"submitted","content":"Earth, Planets and Space","date":"2024-05-27T06:13:17+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"earth-planets-and-space","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"epsp","sideBox":"Learn more about [Earth, Planets and Space](http://earth-planets-space.springeropen.com)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/epsp/default.aspx","title":"Earth, Planets and Space","twitterHandle":"@SpringerOpen","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"9a49ae4a-edce-4d1d-b6f3-3469ff62a51f","owner":[],"postedDate":"June 20th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2024-11-25T16:03:08+00:00","versionOfRecord":{"articleIdentity":"rs-4484150","link":"https://doi.org/10.1186/s40623-024-02092-7","journal":{"identity":"earth-planets-and-space","isVorOnly":false,"title":"Earth, Planets and Space"},"publishedOn":"2024-11-22 15:57:50","publishedOnDateReadable":"November 22nd, 2024"},"versionCreatedAt":"2024-06-20 16:17:07","video":"","vorDoi":"10.1186/s40623-024-02092-7","vorDoiUrl":"https://doi.org/10.1186/s40623-024-02092-7","workflowStages":[]},"version":"v1","identity":"rs-4484150","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4484150","identity":"rs-4484150","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}