金曜日セミナー 2013 年度の予定／記録

Earthquake damage depends in part on the velocity of the rupture front. In 1973, Burridge demonstrated theoretically that in-plane shear ruptures (mode-II) could propagate at velocities higher than the shear wave velocity (Cs) and up to the compressional wave velocity (Cp). Since then, so-called supershear ruptures (Vr >Cs) have been observed during large strike-slip earthquakes. The stress and geometric conditions leading to the transition between sub-Rayleigh and supershear ruptures have been investigated both theoretically and experimentally on brittle polymers using photoelasticity. Although these experiments successfully illustrate supershear ruptures, the lack of experiments on rock samples limits the ability to understand these rare events observed in nature. Here we report results from stick-slip experiments conducted on saw cut Westerly granite samples, which serve as proxies for crustal rocks, during triaxial loading (where the principal stresses, σ1 > σ2 = σ3). Stick-slip experiments and earthquake mechanisms are analogous in nature because they both result from rapid frictional sliding along pre-existing faults resulting in partial or total stress drop. In our experiments, the stress conditions were typical of the upper crust, ranging from 10 to 150 MPa in normal stress acting on the pre-existing fault. In total, more than 250 stick slip events were recorded. For each event, we inverted the rupture velocity directly using high frequency acoustics as a tracking tool. Supershear ruptures systematically occur when the normal stress exceeds 45 MPa with resulting stress drops of the order of 3-25 MPa, comparable to the stress drops inferred by seismology for crustal earthquakes. In our experiments, the sub-Rayleigh to supershear transition length is at most a few centimeters, suggesting that the rupture of asperities along a fault may propagate locally in supershear. In turn, these sudden accelerations/decelerations could play an important role in the generation of high frequency radiation and the overall rupture energy budget.

The thinness of fault slipping zones and the paucity of observed melts implies very low dynamic friction compared to the overburden pressure (less than 0.05 for a meter of slip at 10 km). However, if static friction was comparably low, then the crust could not support observed topographic relief. Strong-rate-weakening friction seems to be a plausible explanation for these seemingly conflicting observations. Strong-rate-weakening friction leads to slip-pulses with extremely complex failure dynamics; strong positive feedback between the slip and the friction produces multi-scale chaos. Unfortunately, 3-d continuum problems with strong-rate-weakening friction are numerically intractable. Therefore we (Ahmed Elbanna and I) investigated the much simpler problem of 1-d spring block sliders with strong-rate-weakening-friction. We show that the system produces power-law complexity. That is, the pre-stress evolves into a state that is heterogeneous at all scales. Since the pre-stress and the events are spatially heterogeneous, we must generalize our definition of "strength." We define "stress-based strength" to be the spatial average of the pre-stress in a failure region, and we define "work-based strength" to be the average work per unit of deformation. We show that these strengths are not the same. Furthermore, we show that the larger the event (or system), the smaller the strength. We show that the strength decreases as a power with the size; the exponent of this relation is related to the dynamic heterogeneity of the system. Since the model is homogeneous, all complexity is dynamic. Earthquakes are so gentle because the Earth is so big. Finally we show a surprising new energy transport equation that reproduces the chaotic behavior of the full numerical simulation. The equation is multi-scale and many orders of magnitude faster than the full numerical system.

The East African Rift (EAR) provides a natural laboratory to study continental rifting and breakup processes as they occur. The EAR encompasses the youngest continental flood basalt province that is dynamically linked to the largest deep mantle anomaly, the African superplume. The large number of ongoing projects demonstrates the growing interest of the international Earth Sciences community to identify the processes responsible of the lithospheric weakening that leads to strain localisation and the continental break-up. The recent EAGLE and AFAR projects (NERC) greatly helped to characterize the late stage of magmatic rifting processes. These projects focus on systems already extremely influenced by significant lithospheric thinning and resultant magmatism. More recently French (ColiBrea) and US (Crafti) projects were funded to study the first 5 My of a magmatic rift initiating in thick (>150 km) continental lithosphere, where one can directly image and detect fault and magma interactions, as well as the role of inherited and rheological heterogeneities of the lithosphere on rift localisation. Finally the US Geoprism initiative is planning a program to study the EAR. In all these experiments, magnetotellurics surveys were carried out to complement our structural understanding of the crust and mantle in this region. Our group is involved in all these experiments and projects and participated to additional studies thus providing a substantial sampling of the Electric EAR from the Afar region in Ethiopia close to oceanisation, to Tanzania at early stage break up.

Earthquakes strongly cluster in space and time, driven both by earthquake-to-earthquake triggering and underlying physical processes, such as tectonic stress loading, increased pore pressure, etc. To understand the general characteristics of earthquake clustering from a large dataset of earthquakes, I analyze seismicity in southern California. I use a high-resolution earthquake catalog based on waveform cross-correlation to study the spatial-temporal distribution of earthquakes. Parameters based on event location, magnitude and occurrence time are computed for isolated seismicity clusters. Spatial migration behavior is modeled using a weighted-L1-norm method. Aftershock-like event clusters do not exhibit significant spatial migration compared with earthquake swarms. Two triggering processes are considered for swarms: slow slip and fluid diffusion, which are distinguished based on a statistical analysis of event migration. The results suggest fluid-induced seismicity is found across southern California, particularly within geothermal areas. In the Salton Sea geothermal field (SSGF), a correlation between seismicity and fluid injection activities is seen. Spatial-temporal variations of earthquake stress drops are investigated in different regions, and a distance-dependence of stress drop from the injection source is found in the SSGF, suggesting the influence of increased pore pressure. Temporal variation of stress drops within mainshock source regions shows that foreshocks and earthquake swarms have lower stress drops than background seismicity and aftershocks. These results, combined with the spatial migration observed for some large foreshock sequences, suggests an aseismic transient process is likely involved in foreshock triggering.

九州地方は、フィリピン海(PHS)スラブが沈み込んでおり火山活動の活発な地域である。九州地方の火山活動の特徴として、火山フロント上に阿蘇から霧島まで110kmにわたって火山のない非火山地域が存在することや、多様な化学組成の火山岩が分布することが挙げられる。このような火山の分布や火山岩の化学組成の成因を説明しマグマの生成過程を理解するためには、PHSスラブによってどのように流体が輸送されているのかを明らかにすることが重要である。 　私たちは、九州地方に防災科学技術研究所が設置したHi-netの観測点と、気象庁、九州大学、鹿児島大学、京都大学が設置したJ-arrayの観測点の波形データをもとにレシーバ関数(RF)を作成した。そして、不連続面の傾斜に応じた地震波の屈折を考慮してRFをスタック(Abe et al., 2011, GJI)することで、大陸モホ面、海洋モホ面およびPHSスラブ上面に相当する速度不連続面を検出し、その形状を推定した。 　九州地方の中部(32-33°N)・南部(31-32°N)では、海洋モホ面に相当する不連続面が70-90kmの深さまで検出され、その深さまで海洋地殻がスラブマントルに対して低速度であることがわかった。北部(33-34°N)では、その不連続面が60kmより深部では検出されなかった。このことから、PHSスラブの海洋地殻は、九州地方の中部・南部では70-90kmの深さまで含水鉱物を含んでおり、北部では60kmよりも浅部で脱水していると解釈できる。 　大陸モホ面における速度コントラストから、直下のマントルの速度を推定したところ、九州地方中部では前弧側のマントルウェッジに非常に低速度(Vs<3.2km/s)の領域が存在することがわかった。S波速度の値から、この低速度領域には蛇紋岩および間隙水が含まれていると考えられる。九州地方の南部と北部では、このような低速度領域は検出されなかった。 　このように推定された流体分布をもとに、九州地方の脱水流体の移動および火山活動について以下のように解釈することができる。九州地方南部ではマントルウェッジの前弧側に流体が分布せず、中部ではマントルウェッジの前弧側に含水領域が存在することから、海洋地殻から脱水した流体は南部では背弧側へ、中部では前弧側へと移動する傾向があると考えられる。九州地方北部では、海洋地殻が浅部で脱水し多量の流体がマントルに持ち込まれない。九州地方南部では、マントルウェッジを背弧側へと移動した脱水流体によって火山活動が引き起こされ、中部・北部では、背弧側に十分な流体が供給されないことにより、非火山地域が形成されるか、あるいはスラブ脱水流体に汚染されていない火山岩が分布する。

Ground deformation measurements are a promising tool to forecast severe volcanic unrest and eruptions. Prior to an eruption, magma ascents towards the Earth's surface. When new magma arrives in a magma storage zone (magma chamber), this leads to the pressurization of the reservoir which in most cases is associated with the inflation of the volcanic edifice. I will present observations of volcano inflation and other types of deformation using satellite-based Interferometric Synthetic Aperture Radar (InSAR), largely supported by data of the Japanese ALOS-1 satellite. I will discuss arc volcanoes in Indonesia and Mexico and hot-spot volcanoes in the Galapagos islands and in Hawaii. For the Indonesian volcanoes I report on several cases of precursory inflation, in which InSAR, if conducted in real time, could have provided eruption warnings. The basaltic Galapagos volcanoes behave differently. They are very dynamic with multiple volcanoes inflating and deflating simultaneously, out of phase, before, or after eruptions. The observations provide new insides into the volcanic systems, such as information on the mode of subsurface magma migration, and on the connectivity of the volcanic plumbing systems at depth. References: Chaussard, E. and F. Amelung, Precursory inflation of shallow magma reservoirs at west Sunda volcanoes detected by InSAR, Geophys. Res. Lett., doi:10.1029/2012GL053817, 2012 Bagnardi, M. and F. Amelung, Space-geodetic evidence for multiple magma reservoirs and subvolcanic lateral intrusions at Fernandina Volcano, Galapagos Islands J. Geophys. Res., 117, B10406, doi:10.1029/2012JB009465, 2012 Baker, S. and F. Amelung, Top-down inflation and deflation at the summit of Kīlauea Volcano, Hawaii observed with InSAR, J. Geophys. Res., doi:10.1029/2011JB009123, 2012 http://www.geodesy.miami.edu/publications.html

Institut de Physique du Globe de Paris, Paris, France. Actually invited to Earthquake Research Institute, Tokyo, Japan. Abstract: The recent tsunamigenic Tohoku earthquake (2011) strongly affirms, again, after the 26 December 2004, the necessity to open new paradigms in oceanic monitoring. Detection of ionospheric anomalies following the Sumatra earthquake tsunami (e.g., Occhipinti et al. 2006) demonstrated that ionosphere is sensitive to earthquake and tsunami propagation: ground and oceanic vertical displacement induces acoustic-gravity waves propagating within the neutral atmosphere and detectable in the ionosphere. Observations supported by modelling proved that tsunamigenic ionospheric anomalies are deterministic and reproducible by numerical modeling via the ocean/neutral-atmosphere/ionosphere coupling mechanism (Occhipinti et al., 2008). To prove that the tsunami signature in the ionosphere is routinely detected we show here perturbations of total electron content (TEC) measured by GPS and following tsunamigenic eartquakes from 2004 to 2010 (Rolland et al. 2010, Occhipinti et al., 2013), nominally, Sumatra (26 December, 2004 and 12 September, 2007), Chile (14 November, 2007), Samoa (29 September, 2009) and the recent Tohoku-Oki (11 Mars, 2011). In addition to GPS/TEC observations close to the epicenter and measured by GEONET network, new exciting measurements in the far-field were performed by Airglow measurement in Hawaii: those measurements show the propagation of the IGWs induced by the Tohoku tsunami in the Pacific Ocean (Occhipinti et al., 2011). Based on the observations close to the epicenter, mainly performed by GPS networks located in Sumatra, Chile and Japan, we highlight the TEC perturbation observed within the first hour after the seismic rupture. This perturbation contains informations about the ground displacement, as well as the consequent sea surface displacement resulting in the tsunami. In this talk we present all this new tsunami observations in the ionosphere and we discuss, under the light of modelling, the potential role of ionospheric sounding in the oceanic monitoring and future tsunami warning system by GPS, Airglow and OTH radar (Coisson et al., 2011). All ref. @ www.ipgp.fr/~ninto

An improved understanding of the connection between seismic behavior and fault-zone rheology at depth is an essential step toward understanding the underlying mechanics of the faulting process. We investigate the seismicity along the northernmost creeping section of the San Andreas fault near San Juan Bautista (SJB), California, by systematically examining spatiotemporal behaviors of characteristically repeating earthquakes (CRE). We analyzed waveforms from over 12,000 local SJB earthquakes (1984-2012) that occurred within a 15-km range from the epicenter of the 12 August 1998 Mw 5.1 SJB earthquake. This 1998 SJB earthquake was the largest historic earthquake in the SJB area and was associated with a large slow slip event. Our analysis identified about 600 CRE sequences (0.1 < M < 3.5) distributed in the seismogenic crust extending over a depth range of 10 km. We used the empirical method of Nadeau and McEvilly (1998, BSSA) to estimate the amount of fault slip around CRE sequences. Our main findings are as follows. 1) A spatial variation of the fault slip rate in 1984-1998 inferred from the CREs identifies a low-creep (0.1-0.2 cm/year) or locked asperity near the 1998 Mw 5.1 earthquake. 2) In the postseismic period following the 1998 Mw 5.1 mainshock, the CRE activity was significantly increased in a zone at a depth of 5-10 km about 2-7 km northwest of the 1998 SJB mainshock, which indicates a triggering of substantial aseismic slip induced by the 1998 SJB mainshock. 3) Assuming a rate-strengthening friction law, the fault frictional parameter (a-b)sigma (sigma is the effective normal stress) was estimated, by inverting the 10-year of the postseismic slip time series inferred from CREs. The resulting (a-b)sigma ranges from 0.01 MP to 0.05 MPa with a 0.1 MPa of the mean Coulomb stress change determined from our coseismic slip model for the 1998 SJB mainshock.

Episodic aseismic slip events observed on major tectonic fault systems constitute a new mode of deformation in earthquake cycles along active plate boundaries. In subduction zones, slow slip events, sometimes accompanied by low-frequency seismic tremors, recur several months to years with total slips of order mm to cm downdip of the megathrust seismogenic zones. In the uppermost 3–5 km of continental strike-slip faults, a significant fraction of the total slip is typically accommodated by stable sliding, or fault aseismic creep. Creep can be continuous or episodic, lasting only a few hours, and varies throughout an earthquake cycle and from one fault to another. This talk will discuss numerical models based on a rate- and state-dependent friction law developed to investigate the physical mechanism of aseismic slip events in two different tectonic environments. I will show that in subduction zones episodic SSEs arise spontaneously around the friction stability transition at the downdip end of the seismogenic zone, with recurrence intervals similar to observations under near-lithostatic pore pressure. Fault gouge dilatancy, an effective strengthening mechanism under high pore pressure, affects the occurrence of SSEs and, more importantly, stabilizes downdip rupture propagation of megathrust earthquakes. On the contrary, the wide variability in observed shallow creep characteristics on strike-slip faults in California cannot be readily explained by a classic mechanical model which attributes creep events to the transition from unconsolidated sediments to lithified rocks at depth. Instead, an additional unstable layer embedded within the shallow, stable zone, is necessary in order to explain the geodetic observations on a number of significant fault segments in California. Such an unstable layer may result from shallow fine-scale lithological heterogeneities on the continental strike-slip faults.

A glacier emits seismic waves during a variety of processes like when it undergoes basal stick-slip sliding, calves icebergs, or drains a supraglacial lake. One particularly prominent emitter has been the Whillans Ice Plain (WIP), which may be the most seismically active glacier in the world, in that its bi-daily stick-slip events generate seismic waves observable up to 1000 km away. I will present new field observations from the WIP detailing the rupture process and suggest that the rupture speed may provide not only insight into glacier basal conditions, but basic rupture physics of tectonic earthquakes. Next, I will show that the dynamics at Store Glacier, an outlet glacier in West Greenland, are influenced by ocean tides and the seasonal removal of compacted icebergs and sea ice during the spring. Finally, I highlight the connection between the solid earth and cryosphere by showing that surface waves from distant, large earthquakes, trigger icequakes across Antarctica. The work yields information regarding the sensitivity of ice and tectonic processes occurring on various spatial and temporal timescales. Detailed study of many of these processes is still needed before ice sheet models can realistically forecast ice sheet response to climate forcing.

雪崩は、斜面上に積もった雪が重力の作用により肉眼で識別し 得る速さで流れ下る現象と定義される。春先に多く発生する全層雪 崩は、一般 に低速で「流れ型」である場合が多いが、冬季の表層雪崩は「煙型」へと発達 し、時速300km以上に達することもある。雪崩発生危 険度の予測としては、気象 庁が発表する「雪崩注意報」があるが、気温と積雪深（または降雪深）による経 験則から導かれていることに加え、対象 とする領域が比較的広いため、地方自 治体、道路、スキー場管理者からは、よりきめの細かい情報の提供を望む声が大 きい。本セミナーでは、ま ず、上記の背景のもと、気象情報から積雪構造の変 質と物性を算出し、その結果に基づいて雪崩発生危険度を予測する試みを紹介す る。続いて、動 き出した雪崩の内部構造の解明を目的とした、人工爆破による 雪崩発生等による野外観測、フルード数を考慮した小スケールの実験、さらには モデ ルを用いた雪崩の流下経路と速度、到達点を予測する試み等に関して、最 近の動向も含めてお話しする予定である。

The Nankai subduction zone has repeatedly generated great earthquakes in excess of Mw 8. Because great earthquakes at convergent plate margins are believed to occur both at the plate interface and on mega-splay faults, seismic reflection data were acquired to reveal the structure of these seismogenic faults. However, we cannot clearly interpret the evolution of such a mega-splay fault system from the reflectivity images alone because of the low signal-to-noise ratio at the landward side of the active mega-splay fault, which may be due to rock consolidation as well as multiple reflections. Furthermore, the transition zone between the landward mega-splay fault (inner wedge) and the seaward décollement (outer wedge) is unclearly imaged because of the topographic relief at the outer ridge. The structures and pressure conditions of the transition zone are critical to evaluate coseismic rupture propagation close to the trough axis that may lead to tsunami generation. The 1605 Keicho earthquake (Mw7.9) is well known as a tsunami earthquake in the Nankai Trough, and was characterized by coseismic rupture close to the trough axis. The tsunami of the 2011 Tohoku-oki earthquake was also generated because of rupture propagation close to the trench. Kamei et al. [2012 EPSL, 2013 GJI] has applied frequency-domain Waveform Tomography (WT) to controlled source Ocean Bottom Seismometer (OBS) data, and retrieved high-resolution P-wave velocity images of the mega-splay fault system. By exploiting recorded seismic waveforms beyond their first arrivals, the WT method achieves a much higher resolution than that of conventional traveltime tomography methods, and resolves the transition zone between mega-splay fault and seaward décollement. Here, we explore the evolution of the mega-splay fault and its relationship to the décollement based on pore pressure distribution. We applied the methodology in Tsuji et al. [2008, JGR] in order to estimate the pore pressure around the Nankai mega-splay fault from the WT-derived seismic velocity model by integrating logging data and laboratory-derived data. Our results suggest that a high pore pressure zone at the footwall side of the deep mega-splay fault continues to the seaward region close to the trough axis. The over pressure zone of the footwall of the basal mega-splay fault (between the mega-splay fault and crustal surface) has almost constant values. This high pore pressure distribution indicates the possibility of coseismic rupture propagation from the deep mega-splay fault to the seaward trough region. In this presentation, we further show crustal stress variation derived from seismic anisotropy analysis for the same OBS data as well as walkaround VSP data [Tsuji et al., 2011 Geophysics, 2011 G-cubed]. In these studies, we estimated principal stress orientation and its magnitude within the Nankai accretionary prism. Finally we show the studies of monitoring seismic velocity variation using seismic interferometry [e.g., Minato et al., 2012 GRL] to introduce our recent research objective (i.e., monitoring stress and pore pressure). References Kamei, R., G. Pratt, and T. Tsuji, Waveform Tomography Imaging of a Megasplay Fault System in the Seismogenic Nankai Subduction Zone, Earth and Planetary Science Letters, 317-318, 343-353, 2012. Kamei, R., G. Pratt, and T. Tsuji, On acoustic waveform tomography of wide-angle OBS data - Strategies for preconditioning and inversion, Geophysical Journal International, 194, 1250-1280, doi:10.1093/gji/ggt, 2013. Minato, S., T. Tsuji, S. Ohmi, and T. Matsuoka, Monitoring seismic velocity change caused by the 2011 Tohoku-oki earthquake using ambient noise records, Geophysical Research Letters, 39, L09309, 2012. Tsuji, T., R. Hino, Y. Sanada, K. Yamamoto, J-O. Park, T. No, E. Araki, N. Bangs, R. von Huene, G. Moore, and M. Kinoshita, In situ stress state from walkaround VSP anisotropy in the Kumano basin southeast of the Kii Peninsula, Japan, Geochem. Geophys. Geosyst., 12, Q0AD19, doi:10.1029/2011GC003583, 2011. Tsuji, T., J. Dvorkin, G. Mavko, N. Nakata, T. Matsuoka, A. Nakanishi, S. Kodaira, and O. Nishizawa, Vp/Vs ratio and shear-wave splitting in the Nankai Trough seismogenic zone: Insights into effective stress, pore pressure and sediment consolidation, GEOPHYSICS, 76, No.3, WA71-WA82, 2011. Tsuji, T., H. Tokuyama, P. Costa Pisani, and G. Moore, Effective stress and pore pressure in the Nankai accretionary prism off the Muroto Peninsula, southwestern Japan, J. Geophys. Res., 113, B11401, doi:10.1029/2007JB005002, 2008. Forthcoming seminars:

In discrete element method (DEM) simulations, the tangential spring force has the limit according to the Coulomb friction law. And generally, the coefficient of static friction is used as the frictional parameter. However, when the normal force is impulsive, such as the case of collision, it is doubtful to use the parameter based on static friction. Especially, for polygonal DEM, friction due to the rotation which is induced by the inertia coupling, such as the case of eccentric collision, is very complex but not studied well. Then, we performed two dimensional eccentric collision experiment using polygon plates on a frictionless air table. The dynamics of the polygon was analyzed from the kinematic data measured by using high-speed video camera.

Prolonged extension of the continents results in deep elongated rift valleys and in rift-centered magma ponding zones at the crust-mantle boundary. Some rifts present very scarce surface volcanism, others show abundant volcanism within the rift, others show, counterintuitively, off-rift volcanic fields offset by tens of kilometers from the source of magma at depth. The reasons behind the surface distribution of magmatism in rifts remain enigmatic. Here we demonstrate, with numerical models of dyke propagation, that the gravitational unloading induced by crustal thinning and rift border faulting can steer ascending dykes from the deep magma ponding zone towards the rift sides, causing eruptions at tens of kilometers outside the rift border. Our models predict that the presence and location of surface volcanism is linked to the geometry of the mass deficit in the rift graben. Furthermore, they predict the formation of stacked magmatic sills in the lower crust above the magma ponding zone, and the along-rift propagation of shallow dykes in rifting events. Gravitational unloading, which facilitates decompression melting of the asthenosphere, is fundamental in controlling the transfer of magma from depth to the surface. Our predictions are consistent with the distribution of volcanism from rift zones around the Earth.

原子核乾板とは、荷電粒子の飛跡が通過する際に発する微弱な光を記録する一種の写真フィルムである。荷電粒子が臭化銀結晶に当たると、結晶中に潜像核という非常に小さな銀の塊が発生する。現像処理すると、潜像核が成長し大きな塊(１ミクロン以下)になる。光学顕微鏡を通してこれを見ると、荷電粒子が通った跡が点々とした線として残っている。大気中で生成されたミューオンも、直線的な飛跡として原子核乾板に記録される。 　飛跡を記録するために電源が不要である一方で、時間分解能を持たないという一面も持つ。他のミューオン検出器には見られない、非常にユニークな特徴を持つ検出器である。この特徴を用いて、浅間山火口浅部や昭和新山溶岩ドームの密度構造イメージングがなされた(Tanakaet al., 2007)。 　現在、原子核乾板に簡単なからくりを付与することで、時間情報付きの観測を可能する検出器を開発中である。その他、一時間弱で写真フィルムを用いた火山透視について理解できるネタを色々提供する。

活動中の火山に対し，活動現場の付近で観測することは，後の噴火予測や住 民の避難計画，砂防計画の策定を行う上で，非常に有用である．しかしなが ら，一般に，活動中の火山には，人が立ち入ることができず，また，事前に 設置したセンサが火山活動中に動作するとは限らない．そこで，東北大学で は，芝浦工業大学，工学院大学，土木研究所，国土交通省らと共に，遠隔操 作にて火山の不整地を走行し，火山噴火による緊急的な土砂災害対策に必要 となる降灰等の調査を近距離から実施することが可能な，移動探査ロボット の研究開発を，2009年より進めてきた．これまでの研究開発ならびにフィー ルド試験により，電動マルチコプタをベースとした飛行ロボットにより，小 型軽量の地表移動ロボットを運搬し，地表移動ロボットの遠隔操作によって 探査を行う，飛行ロボットと地表移動ロボットの複合システムが有効である ことが分かってきた．そこで，本発表では，現在想定している無人探査シナ リオと探査アプローチについて紹介し，2013年3月，9月に行った，この複合 システムによる遠隔探査に関するフィールド試験について報告する．

気象学や海洋学を中心に発展してきたデータ同化は、シミュレーションモデルと 観測データをベイズ統計学の枠組みで融合するための計算基盤技術である。 シミュレーションを実施しながら、時々刻々と入力される観測データを 挿入することにより、モデルパラメータおよび状態変数の逐次推定を行う。 データ同化の最も身近な例としては日々の天気予報が挙げられ、例えば台風の 進路予測において、確率密度関数の平均である台風の中心位置の軌跡だけでなく、 確率密度関数の標準偏差を中心位置が存在する可能性が高い予報円として 表現することは、逐次ベイズフィルタによる推定プロセスが端的に表れたものと言える。 現在のデータ同化の枠組みでは、起こり得る現象をある程度想定した上で、シミュ レーションモデルを与える必要がある。逆に言えば、所与のモデルで想定されていな い現象については、いくらデータ同化を実施しても再現することはできないため、 現在のデータ同化はシミュレーション駆動の「演繹型モデリング」手法であると言える。 台風の例で言えば、発生後の規模の推移と進路についてはモデルによって予測可能で あるが、台風の発生そのものについてはモデルに内包されていないために予測不可能であり、 もっぱら観測データに頼らざるを得ない。もっとも、台風の発生域は日本のはるか南 方の太平洋上にほぼ限られることもあって、我が国のほとんどの住民にとっての興味の対 象は台風の規模と進路の予測であり、発生そのものについてはなかなか興味が注がれにくい。 しかしながら、固体地球科学においては地震や火山噴火といった非線形現象が 興味の対象となりやすく、当然ながらデータ同化によってこれらのイベント発生その ものを予測することは極めて困難である。それでも、海洋プレート境界域における大地震発 生サイクルを再現する試み等、固体地球科学へもデータ同化が徐々に浸透し始めている。 一方、地球物理観測網の急速な発展や、携帯電話を始めとする各種センサーデータの活用は、 固体地球科学分野においても「ビッグデータ時代」が到来したことを感じさせており、 爆発的に増加した観測データをいかに調理するかが問われている。このような超高次元の時 空間スケールを持つ膨大な観測データともなれば、人間の頭脳によって物理モデルを 構築することは、もはや限界を超えた作業であると言って良い。 このように、シミュレーションモデルによって表現しきれない現象を追い求め、 かつビッグデータを取り扱うためには、従来のシミュレーション駆動型の演繹的 モデリングによるアプローチはもちろん必要であるものの、それと同時に、観測データから モデルを構築するデータ駆動型の「帰納的モデリング」も併せて必要となってくる。 発表においては、データ同化の基礎と応用について述べ、現在発表者が目論んでいる モデル不完全性を示すシステムノイズをスパースモデリングによって分析する シミュレーション／データ両駆動型データ同化の創出について述べる。

沈み込み帯で生起するジオダイナミクスを明らかにすることは、地震発生の長期予測の基礎として重要である。2011年の東北日本太平洋沖地震は、稠密な地球物理観測網の中で発生した超巨大地震であり、巨大な地震時変位に伴う地殻応答の研究は、沈み込み帯のダイナミクスの理解にとって極めて重要な情報をもたらすことになる。この地震によって明確に示されたことは、プレート内地震(内陸地震)とプレート間地震の強い相関関係である。内陸地震の長期予測は、独立した断層系の挙動と見るのではなく、プレート境界を含むシステムとしての取り扱いが必要である。こうした問題を解明するために、我々は地殻・マントル構造のモデル化と、三次元数値実験を通じて、東北太平洋沖地震後の地殻活動と、長期間地殻変動を理解するための研究を進めている。 　日本列島の構造モデルについては、自然地震・制御震源などのデータを用いて、多数の研究者と共同でコミュニティモデルの構築を進めている。巨大地震後の粘性緩和を含む応力変化に伴う、内陸の震源断層やプレート境界の応力変化を精度よく定量化するためには、特に震源断層のモデル化が急務である。例えば、日本海沿岸に発達する大規模な活断層では、日本海形成時の中絶リフトによって決定された活断層-震源断層の形状や、地殻構成岩石の性質が、地震活動に大きな影響を与えていることが近年の調査で明瞭になってきた。現在、日本海の津波波源断層の形状把握のための調査を実施しており、数年以内に日本列島周辺地域の断層形状モデルの作成を予定している。 　島弧スケールの長期間の地殻変動については、変動地形学・第四紀地質学に加えて近年の地殻構造探査により、新たな活構造の抽出やひずみ速度の推定が可能になってきた。東北日本南部では、背弧側の活断層で消費されるひずみ速度が前弧側に比べて明確に大きく、また背弧域に顕著な沈降域がある、といった特徴が認められる。沈降帯の地下では薄化した地殻と下部地殻〜上部マントルにP波速度の高速度異常が分布しており、長期間地殻変動とリソスフェアの構造の関係を示唆している。 　東北日本太平洋沖地震による地殻変動については、三次元有限要素法により検討を加えている。太平洋プレートとフィリピン海プレートの形状を取り入れて、超巨大地震以降の粘性緩和による日本列島域の大局的な地殻変動変化を計算した。地震時の地殻変動に関しては、弾性構造による違いが見られるが、せいぜい1割程度である。一方、地震後では、粘性構造によって粘性緩和による応力伝播のパターンに明らかな違いが見られ、特に低温・高粘性のスラブの存在が地殻内の応力伝播に与える影響の大きさを示している。地震後の地殻変動データとモデル計算の比較により、地下の粘性構造を拘束していくことを当面の目標としているが、将来的には断層形状を取り込んだクーロン応力の変化による地震発生の長期予測につなげて行きたい。 　なお、本講演では、断層形状を中心とした構造モデルについて佐藤が、長期間地殻変動については石山が、数値実験については橋間が紹介する。

We summarize new measurements of seismic anisotropy and its relation to stress and fault zone structure in the crust and mantle of New Zealand, which sits on the NE/SW striking transpressional boundary region between the Pacific and Australian plates. We particularly focus on new results from an OBS deployment around the South Island of New Zealand, from an aftershock deployment recording the 2010-2011 Canterbury earthquake sequence, and from a combined controlled source/passive source study of the Wellington/Wairarapa region.

New understanding about heat transfer deep inside the Earth is presenting new challenges to understanding the state of the early Earth, and how it evolved to the present structure. High core-mantle heat flow, more extensive CMB reactions, the absence of a rock record during the Hadean, and the geochemical evidence for extensive mixing as well as ancient reservoirs must all be reconciled with a new generation of geodynamical models. These models must be framed in a way that allows us to better understand the coupling between Earth's various physical systems, and based on this understanding of the Earth's interior as an open system that exchanges matter and energy with the surface environment we can begin to approach big scientific questions regarding interactions with other open systems, such as life on the Earth's surface. In such a new framework of open systems, it will be even more difficult to draw ad hoc boundaries around components and separate them into living and non-living components. But through such paradigm shifts we may begin to find new ways of looking at the entire Earth system which should benefit all the disciplinary sciences concerned with various aspects of this system.

Recent broadband seismograph deployments are greatly increasing the seismic coverage of Antarctica and leading to the development of new higher resolution seismic models of the continent, which is poorly understood due to limited geological exposure. The new results reveal that East Antarctica represents an ancient continent, with a geological age of mid-Proterozoic or earlier and with highlands supported by thick continental crust. In contrast, West Antarctica shows low upper mantle seismic velocities consistent with Cenozoic extension and possible mantle plumes. We use these models to constrain parameters important for ice sheet models such as heat flow and mantle viscosity. The inferred mantle viscosity varies by several orders of magnitude, and should have a first order effect on glacial isostatic adjustment (GIA). Inferred mantle viscosity in West Antarctica is much lower than used in recent GIA models, and limits the GIA response to ice sheet mass changes to the last several hundred years. In addition, the recent discovery of an active subglacial volcano suggests volcanism and associated high geothermal heat flow may have an important effect on the ice sheet. Seismic data also provides important constraints on the movement and dynamics of ice sheets. A wide variety of new, relatively uncharacterized seismic sources related to ice movement have been recorded. We use co-located seismographs and GPS receivers deployed on the Whillans Ice Stream to obtain very broadband records of ice velocity during the twice-daily stick-slip events, which involve a region larger than 100 km and occur over a time period of about 20 minutes. These records, combined with observations of far-field Rayleigh waves, reveal that the stick-slip motion is controlled by high friction regions similar to earthquake fault asperities located along the grounding line. Other seismic sources in the upper 50 m of the East Antarctic ice sheet result from fracturing and cracking in the firn layer, facilitating air circulation through the upper part of the ice sheet.

この6年間地震研究所に籍を置かせてもらい、日本の地震予知研究の歴史を調査研究してきた。日本の地震予知研究の歴史 は、1880年の日本地震学会の創設と同時に始まる。1965年の第一次地震予知計画の開始、ないしは62年の“ブループリント”作成から始まったわけでは ない。この130年間、地震発生のメカニズムの理解に関しては大き な進展が見られた。しかし、地震の発生予測に関してはそれほど大きな進展はなかった。いろいろな意味で同じことの繰り返しの歴史で あったようにも見える。予知研究の歴史を簡単に紹介するとともに、地震の発生予測に大きな進展が見られないのは何故なのか、私見を述 べてみたい。