| During Cenozoic time, the West Pacific island-arc and marginal sea region wasvery active in tectonic activity. By the late Tertiary, intensive volcanoes occurredalong the island arc zone, of which the outside exists the deepest trench on the Earth.The westward underthrusting of the West Pacific plate has resulted in the subductionzone beneath the Japanese islands and Japan Sea, named the Japanese subductionzone which is an important part of the subduction zone along the West Pacific margin.Located in the Circum-Pacific seismic and volcanic belt, this region is characterizedby complicated structures and frequent earthquakes, of which the large quakesaccount for20%of the total number of big events in the world. Geographically, theJapanese subduction starts from Kuril Islands in north, extending southward by theJapan Islands and to the Phillipine Sea in south. Along east of the Kuril Islands andHonshu of Japan, the Pacific plate is underthrusting to the Eurasian plate at differentdipping angles. At the Japan trench east of Honshu, distribution of seismic sourcesreveals that the subduction slab is about80-100km thick, dipping29, with maximumplunge depth nearly600km. Due to westward motion and underthrusting of thePacific plate, the northeast Japan island is subject to SEE-NWW directed compression.Focal mechanisms of earthquakes show that the eastern margin of the Japan Sea isalso under this compression, probably associated with the eastward motion of theAmur plate. In such a compressive tectonic setting, most of earthquakes in theJapanese subduction slab are of thrust fault nature. Hypocenter distribution shows thata number of shallow thrust events occur along the Kamchatskiy-Kuril Islands-NorthHonshu, while intermediate and deep seismic sources are found within the subductionzone beneath the island arc. Focal mechanism solutions also suggest the convergencebetween the Pacific and Eurasia plates, implying a compressive stress state of this region. One of the recent examples is the M9.0great earthquake that occurredoffshore east to the Honshu on11march2011, which is the largest event in the historyof Japan. Several tens minutes after this main shock, a huge tsunami with10m-highwaves attacked the eastern cost of Japan. Actually, nine seismic events greater thanM7have taken place along the Japan trench since1973, of which majority took placeat the Japanese subduction zone. Among them, the offshore events of1611,1896and1933caused destructive tsunamis at the Pacific coast of northeastern Japan. So manywell recorded big quakes provide data support to a detailed study of generation andoccurrence of thrusting earthquakes in this subduction zone. Meanwhileunderstanding of thrust-faulting earthquakes in the subduction zone is also helpful forefforts of earthquake prediction research. Previous studies also indicate that theearthquakes on the Japan seduction may have influence on seismicity of Chinamainland. Numerical modeling methods have been developed rapidly and widelyapplied to earthsciences in recent years. In this work, by combination of geologicaland geophysical observations, I use the finite element method to simulate thestick-slip processes of earthquakes in the Japanese subduction slab, and explore themechanisms of thrust faulting earthquakes and deformation features which havetheoretical and applied significance.1Research objectives and methodsThis thesis makes a review on the research status and existing issues ofnumerical modeling of the Japanese subduction zone. Based on various geologicaland geophysical observations, the tectonic setting and deep structure of the study areaare comprehensively described. Using the ANSYS parallel computation system, ahigh-resolution finite element model for the Japanese subduction slab is constructed.With GPS measurements as the boundary constraints, the numerical simulation isperformed to study the generation processes of large earthquakes in the slab andexplore their mechanism.(1) Establish a dynamic model for the Japanese subductaion slab. Based on theANSYS soft ware platform, in combination with geological and geophysical data, atwo-dimensional layered model is constructed, in which each is assigned proper medium attributes. The West Pacific plate underthrusts toward the Eurasia pale at anaverage rate83mm/a with varying dipping angles. The fault behavior of thesubduction zone depends upon dynamic boundaries and rock medium nature as wellas the geometry of the slab. Based on the simulation of the dynamic processes ofstick-slip events on the Japan subduction zone, this work focuses on the effects ofvarying geometry of the slab on the event occurrence. For simulation of stick-slipinstability in the subduction zone, this work constructs four models:1) a slab modelwith bend in the lithosphere;2) a model with bend at the base of lithosphere,3) amodel without bend;4) a model of dipping30without bend. These4models havesame physical parameters. According to deep structures of the subduction zone, theslab models consist of upper and lower layers and shallow and deep portions. Tosimulate the real geology and make simulation more reliable, the lower crust of themodels are assumed to be viscoelastic, assigned by a series of parameters such asYoung’s module, Poisson ratio, viscosity, stratum thickness, and density. In thesimulation calculation, the contact pair concept of the finite element method is used,i.e. contact pairs are fixed at the boundaries between subduction slab and surroundingrock to simulate fault friction. A same gridding manner is adopted for the4models.As this work focuses on the process of stick-slip events on the subduction slab, thearea around the contact boundary is gridded more densely. Such a local intensifiedgridding can enhance the calculation accuracy and reduce computation cost. Sameboundary constraints are imposed on the4models, including displacement andpressure. The applied displacement is determined ased on GPS measurements andtime scales of simulation, while the imposed pressure is the confining pressure atdepth and oceanic water pressure. Gravity is also taken into consideration. First, theinitial stress filed is reconstructed. Then the boundary constraints above mentionedare imposed to calculate the slick-slip motion on the slab.(2) Construct a catalogue of stick-slip events for the slab. Using the numericalmodel mentioned above, considering various factors such as block interaction, platecollision, push of the subduction zone, gravity load, viscous deformation and deepmedia, the tectonic deformation and stick-slip instability of the slab are simulated. Thus a catalogue of large earthquakes for the Japanese subduction zone is synthesized.Similar for each model, such synthesis is based on complex post-processing ofsimulation results. It means to decode the simulation results, extract slip andcoordinates of all nodes on the contact plane. As earthquake mechanism in deepmantle remains unknown, this work only analyzes the events in above300km depth.The simulation time is100,000years, and20000steps are used for calculation.According to the fault slip state, the program identifies whether a sudden slip occursat a step. A maximum slip is chosen to serve the objective event at this step and thecorresponding depth is also determined. In such a way, catalogues of stick-slip eventsas well as their depths can be constructed for each model. Then, slip values andvariations of depths with time are plotted, and their characters are analyzed, resultingin statistics figures for various slab portions and depth ranges. Comparing thesestatistic results of the4models, the effect of slab geometry on simulation is furtheredstudied. Finally, in combination with real records of historic earthquakes, this workanalyzes the common and different points between the simulation results and actualobservations.(3) Study characteristics of tectonic deformation and energy transfer for thrustfaulting earthquakes. Through changing boundary conditions and models, modelingcalculations are repeatedly conducted to invert plate motions and earthquakeoccurrence in the study area. In particular, the effects of modifications in structuraltypes, physical parameters, and boundary conditions on the subduction slab areanalyzed. And by comparison of modeling results and geological, geophysical, GPSand seismic observations, the preferred most reasonable models and their boundaryconditions are determined. Then further calculations are made on these models tosimulate the stick-slip instability process and analyze partitioning and transfer ofstrain and energy between blocks and their influence factors, as wll as thesesimogenesis mechanism in this region.2Primary resultsThe dipping angle of the West Pacific plate may vary when it underthrustsbeneath the Japan Sea. So the four models considering different geometry of the slab are constructed in this work, which consider varied bend features in the slab. Thus theeffects of changing geometry of the slab during the plate subduction can also be takeninto account when stick-slip instability is simulated. Because the earthquakemechanism in deep mantle is more complicated, and it is unclear whether slick slipcan occur there, this work only focuses on the region above depth300km. Fromsimulation and analysis of the models on a time-scale100,000years, this work attainsthe following insights:(1) The analysis of stress-strain curves indicates that the stress in the slabexperiences accumulation, release and re-build up before and after occurrence ofstick-slip events. It implies a process of stick-slip instability characterized by faultlocking, unlocking and re-locking, which represents the process of large earthquakesalong the subduction slab. It also demonstrates that a numerical modeling ofearthquake occurrence processes by fault friction motion is possible.(2) The fault slip from simulated stick slip on the subduction slab is equivalent toseismic dislocation in size. With increasing slip size, the number of events tends todecrease. The events with slip5-10m are the majority, and the number of the eventsexceeding20m decreases greatly, mostly below30m.(3) The simulations of four models indicate that stick-slip events occur mainly inthe slab above100km depth. Most stick-slip events are confined to the depth rangeswhere mechanical nature and fault geometry change considerably, resulting in twoprofound zones which are20-30km deep and around100km depth, respectively. Inthe first zone the events are most and have large magnitudes. Around depth100km,small events are dominated and big quakes are few. This result seems to be consistentwith the real focal distribution in comparison with historic earthquakes documented inthe region.(4) Modifications of dip angle and their depths have effects on the modelingresults. On the model with bend at the lithospheric base, the number of resultedstick-slip events is the largest. On the model without bend and with dip30, thenumber of stick-slip events is the least. It is noted that on the former model, mostrelatively smaller events have slip less than5m, while on the latter model the many events have big slip.(5) Modifications of dip angle and their depths also produce influence on thedepths of events. When the model of dipping30has a bend or no bend, the stick-slipevents occur at a roughly same range of depths, of which most are confined to20-30km depths, in agreement with historical records of earthquakes to some extent.On the model with a bend at the lithospheric base, a lot of small events occur arounddepth200km. On the model without bend and of dip23, the events at depth100kmare more than that around20-30km depths, which are dominantly small ones.(6) The sizes, time intervals and occurrences of slick-slip events from simulationare unified by randomness and sequence. These parameters are sequential(quasi-periodic, characteristic magnitude and dense sesimogenic layer) in macroscope, while random in micro scope (not exactly equal in time, depth and slip size).3. Innovation points and problemsThis thesis has made the following innovational points:(1) It uses the fault contact model and viscoelastic medium to simulate thedynamic processes of stick-slip motion in the subduction slab.(2) It constructs a catalogue of stick-slip events for the Japan subduction zone,and analyzes the spatial and temporary distribution of these events.(3) This work studies the effects of differences in slab geometry on stick-slipfault motion in the subduction slab.As there have been little previous studies on this subject, and limited by softwareand hardware available, this thesis has some problems. For instance, the model issimplified to some extent, few comparisons are made, time step is relatively big, theinitial stress field cannot be evaluated precisely, and the state and activity of deepmantle are unclear. These issues will be addresses in the future work. |
PDF Full Text Request