What triggers Koyna region earthquakes? Preliminary results from seismic tomography digital array

The cause for prolific seismicity in the Koyna region is a geological enigma. Attempts have been made to link occurrence of these earthquakes with tectonic strain as well as the nearby reservoirs. With a view to providing reliable seismological database for studying the earth structure and the earthquake process in the Koyna region, a state of the art digital seismic network was deployed for twenty months during 1996–97. We present preliminary results from this experiment covering an area of 60 × 80 km² with twenty seismic stations. Hypocentral locations of more than 400 earthquakes confined to 11×25 km² reveal fragmentation in the seismicity pattern — a NE — SW segment has a dip towards NW at approximately 45°, whilst the other two segments show a near vertical trend. These seismic segments have a close linkage with the Western Ghat escarpment and the Warna fault. Ninety per cent of the seismicity is confined within the depth range of 3–10 km. The depth distribution of earthquakes delimits the seismogenic zone with its base at 10 km indicating a transition from an unstable to stable frictional sliding regime. The lack of shallow seismicity between 0 and 3 km indicates a mature fault system with well-developed gouge zones, which inhibit shallow earthquake nucleation. Local earthquake travel time inversion for P- and S-waves show ≈ 2% higher velocity in the seismogenic crust (0–10 km) beneath the epicentral tract relative to a lower velocity (2–3%) in the adjoining region. The high P- and S-wave velocity in the seismogenic crust argues against the presence of high pressure fluid zones and suggests its possible linkage with denser lithology. The zone of high velocity has been traced to deeper depths (≈ 70 km) through teleseismic tomography. The results reveal segmented and matured seismogenic fault systems in the Koyna region where seismicity is possibly controlled by strain build up due to competent lithology in the seismic zone with a deep crustal root.     

Characteristics of ring seismicity in different depth ranges before large and great earthquakes in the Sumatra region

Characteristics of the seismicity in depth ranges 0–33 and 34–70 km before ten large and great (M _w = 7.0−9.0) earthquakes of 2000–2008 in the Sumatra region are studied, as are those in the seismic gap zones where no large earthquakes have occurred since at least 1935. Ring seismicity structures are revealed in both depth ranges. It is shown that the epicenters of the main seismic events lie, as a rule, close to regions of overlap or in close proximity to “shallow” and “deep” rings. Correlation dependences of ring sizes and threshold earthquakes magnitudes on energy of the main seismic event in the ring seismicity regions are obtained. Identification of ring structures in the seismic gap zones (in the regions of Central and South Sumatra) suggests active processes of large earthquake preparation proceed in the region. The probable magnitudes of imminent seismic events are estimated from the data on the seismicity ring sizes.     

Seismic and electrical precursors to the 17-1-1983, M7 kefallinia earthquake,Greece: Signatures of a SOC system

Prior to the 17-1-1983 event, the seismicity of the broader area of the Ionian islands and western Greece exhibited several phenomena interpretable in the context of a self-organised critical system with long range interactions. The regional seismic energy release exhibited power law acceleration towards the time of rupture, the numerical modelling of which yields a time-to-failure of 1983.1 ± 0.2. Time dependent changes were also observed in the b-values, assuming the form of monotonic increase that promptly reversed after the earthquake. This indicates the induction of instability to the region due to the earthquake preparation process, which is consistent with the critical point earthquake model. The critical point model predicts that failure is a co-operative effect occurring at small scale, and cascading from the microscopic to the macroscopic scale. This involves a crack propagation avalanche at the terminal phase of the seismic cycle, the time function of which has been modelled with a limited class of characteristic transient bay-like shapes, featuring a corner frequency and inverse power energy distribution law. Electrification processes due to crack propagation may generate an electrical precursor with similar characteristics. Such a potential precursor has been observed independently on 15-1-1983, approx. 120km from the epicentre. In consequence of our observations, we discuss a model relating seismicity and electrical precursors.     

Mapping of b-value Beneath the Shillong Plateau

The seismic parameter ‘b’ has been computed over rectangular grid of dimension 0.3° ' 0.8° at four depths range: 0-13 km (first layer), 13.1-26 km (second layer), 26.1-39 km (third layer) and 39.1-52 km (fourth layer) beneath the Shillong Plateau area. The four depths were carefully selected based on the crustal structure and distribution of hypocentres. The dimension of each grid was chosen so as to have enough events that can represent the b-value at the respective layer. Finally, two-dimensional mapping was done at these depth-levels considering the respective b-value over each grid. This analysis includes viz., low b-value all through the first layer, and a trend of increasing b-value, which was initially towards north, changes to northwest. Eastern and western parts of the second and third layers document almost moderate b-values, whereas the north-south-oriented central part of layer second is apparently dominated by low b-values, which seems to divide the area broadly into three parallel zones based on b-values. In the deeper part (fourth layer) beneath the Shillong Plateau a moderate b-value that was initially trending towards north becomes high near the northeastern part. This phenomenon may be associated with higher heterogeneity of the medium, and interestingly, this region lies between the lower crust and upper mantle, possibly documents lower degree of seismic coupling, where the Shillong Plateau is being supported by the strong Indian lithosphere at these depths. In addition, minima were noted towards the southern parts of layers first, second and third, which may presumably be related with steeply Bouguer gravity anomaly. It is thus less clear that the occurrence of earthquakes beneath the Shillong Plateau whether is attributed to faults or lineaments at intermediate to deeper level. However, a correlation between high b-values in few parts of each layer and deep-seated minor faults cannot be ruled out.     

Deep roots of seismotectonics of the Far East: The Amur River and Primorye zones

The role of the lateral structure of the lithospheric mantle in the seismotectonics and seismicity of the southern part of the Russian Far East has been investigated. The positions of the epicenters of all the major earthquakes in Sakhalin (M ≥ 6.0), as well as in the Amur region and the Primorye zones (M ≥ 5.0), are defined by the boundaries of the Anyui block of highly ferruginous mantle, which lies at the base of the Sikhote-Alin area. Three cycles of large earthquakes are recognized in the region: the end of the 19th-beginning of the 20th century, the mid-20th century, and end of the 20th-beginning of the 21st century. In the seismic zone of the Amur region (hereafter, the Amur seismic zone), the epicenters of the large earthquakes in each cycle migrate from the SW to NE along the Tan-Lu fault megasystem at a rate of 30–60 km/yr. The specific features of the seismicity of the region are explained by the repeated arrival of strain waves from the west. The waves propagate in the upper part of the mantle and provoke the activation of the deep structure of the region. The detailed analysis of the earthquakes in the Sikhote-Alin area (M ≥ 4.0) in 1973–2009 confirmed the clockwise tectonic rotation of the mantle block. The characteristics of the Primorye zone of deep-focus seismicity at the Russia-China boundary are stated. Since 1973, 13 earthquakes with M ≥ 6.0 have been recorded in the zone at a depth of 300–500 km. This number of earthquakes is at least twice as many as the number of large deep-focus earthquakes elsewhere in the Sea of Japan-Sea of Okhotsk transition zone. The unique genesis of the Primorye seismic zone is related to the additional compression in the seismofocal area due to the creeping of the Anyui mantle block onto the subduction zone during its rotation. The geodynamic implications of the seismotectonic analysis are examined, and the necessity of division of the Amur plate into three geodynamically independent lithospheric blocks is substantiated.     

东南亚地震活动与b 值的时空分布特征

东南亚地区是“21世纪海上丝绸之路”（以下简称“海洋丝路”）的重要组成部分,该区历史上曾发生十余次巨大地震,地震及其次生地质灾害是威胁东南亚地区经济社会发展和国际合作的主要自然灾害。系统梳理该区地震活动的时空分布特征及评估未来灾害风险格局,对于推进“一带一路”倡议实施及区域经济社会可持续发展具有重要意义。文章基于东南亚地区1900年以来M≥5地震的时空分布统计分析和地震b值计算,揭示出该区的地震活动在时间上表现出活跃期与平静期交替变化的特征;空间上表现出明显的聚集效应,成丛性强且主要集中在5个地震统计区内,其中印尼—马来多岛弧盆系地震区和菲律宾群岛地震区的地震活动最为活跃。总体而言,东南亚5个地震区的b值偏低,在0.42~0.91之间。该区内的地震b值也存在时空差异,受大地震事件、俯冲带年龄、活动断裂带和震源深度等众多因素影响,但主控因素在不同区域有所不同。地震b值时空变化特征对区域地震活动预测具有启示作用。上述认识为推进“海洋丝路”工程建设和“一带一路”防灾减灾对策提供了科学支撑。     

Structure and dynamics of the Somma-Vesuvius volcanic complex

Summary We review recently obtained results about the velocity structure of the Somma-Vesuvius (Southern Italy) volcanic complex and present an interpretation of structural features, both at local and regional scale, and of the local seismicity. The local structure of Somma-Vesuvius is reviewed, referring to three depth ranges; i.e. shallow (0–5 km), intermediate (5–15 km) and deep (from 15 km to the upper mantle). The shallow velocity structure is inferred by the joint inversion of shot and local earthquake arrival time data. The main feature pointed out by this inversion is a high-velocity anomaly at the crater axis extending down to a depth of about 5 km. This anomaly can be explained with the presence of residual magma crystallised in the shallow conduits, which accumulated during the last eruptive cycles. The local seismicity is strongly clustered around this anomaly, due to the focusing effect of the rigidity contrast. The space-time seismicity pattern at Somma-Vesuvius is the result of the superposition of background seismicity, mainly due to gravitational instability of the volcanic edifice and to small external stress perturbations, with intense episodic earthquake swarms possibly due to magmatic or hydrothermal activity into the shallow system. The velocity structure in the 10–15 km depth range is characterized by the presence of a low-velocity layer, which has been independently confirmed by multi-channel seismic reflection data and P-Sv conversions from teleseismic waveforms. The study of the deep structure was performed by regional tomography with teleseisms; it confirmed the presence of a low-velocity anomaly underneath the volcano, which appears to have roots at greater depths. The regional structure between the Thyrrenian and the Adriatic sea has been inferred by tomographic inversion of teleseismic arrival times. The main result from this study which is very important for geodynamic interpretations is the first evidence for a continuous subducting slab under the Apennines, in an area where previous models hypothesized a slab window. Received March 3, 2000 revised version accepted July 4, 2001     

Annular seismicity structures and the March 11, 2011, earthquake (Mw = 9.0) in Northeast Japan

The characteristics of seismicity prior to the series of eight very strong earthquakes (M _w = 7.0–9.0) in Northeast Japan are discussed. Ring seismicity structures that appeared prior to all eight events in two depth ranges of 0–33 and 34–70 km are identified. The epicenters of the main shocks were located near areas of crossing or touching of shallow and deep rings. It was shown that the sizes of shallow rings and threshold magnitudes corresponding to seismicity rings grow with the energy of the main shocks. It was noted that the prognosis with respect to the place and magnitude of the catastrophic earthquake on March 11, 2011, had been made before it based on the data obtained prior to July 1, 2009. Use of the new data obtained prior to March 10, 2011, enabled us to specify this prognosis significantly. We obtained correlation dependences of threshold magnitudes on the energy of the main shocks (with a high correlation coefficients). It was shown that the duration of the period for seismicity rings to emerge in the considered region nearly did not depend on magnitude. The nature of annular structures and the possibility of application of their parameters for prognosis of strong earthquakes were discussed.     

Seismic slip rates,potential subsurface rupture areas and seismic potential of the Vienna Basin Transfer Fault

The Vienna Basin Transfer Fault (VBTF) is a slow active fault with moderate seismicity (I _max~8–9, M _max~5.7) passing through the most vulnerable regions of Austria and Slovakia. We use different data to constrain the seismic potential of the VBTF including slip values computed from the seismic energy release during the 20th century, geological data on fault segmentation and a depth-extrapolated 3-D model of a generalized fault surface, which is used to define potential rupture zones. The seismic slip of the VBTF as a whole is in the range of 0.22–0.31 mm/year for a seismogenic fault thickness of 8 km. Seismic slip rates for individual segments vary from 0.00 to 0.77 mm/year. Comparing these data to geologically and GPS-derived slip velocities (>1 mm/year) proofs that the fault yields a significant seismic slip deficit. Segments of the fault with high seismic slip contrast from segments with no slip representing locked segments. Fault surfaces of segments within the seismogenic zone (4–14 km depth) vary from 55 to 400 km². Empirical scaling relations show that these segments are sufficiently large to explain both, earthquakes observed in the last centuries, and the 4th century Carnuntum earthquake, for which archeo-seismological data suggest a magnitude of M ≥ 6. Based on the combination of all data (incomplete earthquake catalog, seismic slip deficits, locked segments, potential rupture areas, indications of strong pre-catalog earthquakes) we argue, that the maximum credible earthquake for the VBTF is in the range M _max = 6.0–6.8, significantly larger than the magnitude of the strongest recorded events (M = 5.7).     

Fine structure of seismotectonics in western Shillong massif,north east India

The epicentral tract of the great Assam earthquake of 1897 of magnitude 8·7 was monitored for about 6 months using an array of portable seismographs. The observed seismicity pattern shows several diversely-oriented linear trends, some of which either encompass or parallel known geological faults. A vast majority of the recorded micro-earthquakes had estimated focal depths between 8–14 km. The maximum estimated depth was 45 km. On the basis of a seismic velocity model for the region reported recently and these depth estimates we suggest that the rupture zone of the great 1897 earthquake had a depth of 11–12 km under the western half of the Shillong massif. Four composite fault plane solutions define the nature of dislocation in three of the seismic zones. Three of them show oblique thrusting while one shows pure dip slip reverse faulting. The fault plane solutions fit into a regional pattern of a belt of earthquakes extending in NW-SE direction across the north eastern corner of the Bengal basin. The maximum principle stress axis is approximately NS for all the solutions in conformity with the inferred direction of the Indian-EuroAsian plate convergence in the eastern Himalaya.     

Pn wave velocity and Moho geometry in north eastern India

Pn velocity has been computed across the NE India and Moho geometry constrained, using regional earthquake travel times recorded by a network of 30 seismological stations operated during February-May 1993. Using an appropriate velocity model and the arrival times at the network stations, preliminary hypocentres of 16 regional earthquakes provided by NEIC were also improved. The average Pn wave velocity in NE India has been found to be 8.5 ±0.2 km/s. It varies from 8.3 to 8.5 km/s beneath the Shillong Plateau, Mikhir hills and Assam valley, which is significantly higher than those in other parts of India. The crustal thickness in NE India is also high, varying from 45–49 km under the Shillong plateau and the adjoining region to 55–65 km in the convergence zone. The presence of a thick crust and high Pn velocity suggests that the lithosphere in NE India is colder, as also indicated by the observed deeper level (45-51 km) seismicity of the region.     

On the Bayesian analysis of the earthquake hazard in the North-East Indian peninsula

The Bayesian extreme-value distribution of earthquake occurrences has been used to estimate the seismic hazard in 12 seismogenic zones of the North-East Indian peninsula. The Bayesian approach has been used very efficiently to combine the prior information on seismicity obtained from geological data with historical observations in many seismogenic zones of the world. The basic parameters to obtain the prior estimate of seismicity are the seismic moment, slip rate, earthquake recurrence rate and magnitude. These estimates are then updated in terms of Bayes’ theorem and historical evaluations of seismicity associated with each zone. From the Bayesian analysis of extreme earthquake occurrences for North-East Indian peninsula, it is found that for T = 5 years, the probability of occurrences of magnitude (M _w = 5.0–5.5) is greater than 0.9 for all zones. For M _w = 6.0, four zones namely Z1 (Central Himalayas), Z5 (Indo-Burma border), Z7 (Burmese arc) and Z8 (Burma region) exhibit high probabilities. Lower probability is shown by some zones namely␣Z4, Z12, and rest of the zones Z2, Z3, Z6, Z9, Z10 and Z11 show moderate probabilities.     

New data on the deep structure of the Northern Vent of the Great Tolbachik Fissure Eruption (1975–1976)

In order to restore the deep structure in the region of the Northern Vent (NV) of the Great Tolbachik Fissure Eruption (GTFE) (1975–1976), low-frequency microseismic sounding was applied. For this purpose accumulation of spectra of the microseismic field was performed in a wide frequency band in 29 points along a linear profile 14 km in length embedded transversely to the fissure eruption. A deep cross section of the Earth’s crust was constructed up to 20 km, reflecting the distribution of relative velocities of transverse seismic waves. The revealed structural heterogeneities were interpreted with consideration of previously known results of complex studies of the eruption. The existence of an abnormal structure at the depths of 2–3 and 7–8 km under the NV GTFE was confirmed, which could be low-depth magma chambers. Deep subvertical low-velocity structures were revealed and spatially registered, which probably feed the conduits of the eruption. It was demonstrated that the ways of possible magma supply to the peripheral chamber at the depth of 2–3 km could be various. For the first time for the zone of areal volcanism, variation of the character of magmatic intrusions was demonstrated at the transition from a crystalline basement to the near-surface depth: subvertical forms are replaced with a system of sills and interesting injections.     

Improved earthquake locations in the Koyna-Warna seismic zone

In estimating the likelihood of an earthquake hazard for a seismically active region, information on the geometry of the potential source is important in quantifying the seismic hazard. The damage from an earthquake varies spatially and is governed by the fault geometry and lithology. As earthquake damage is amplified by guided seismic waves along fault zones, it is important to delineate the disposition of the fault zones by precisely determined hypocentral parameters. We used the double difference (DD) algorithm to relocate earthquakes in the Koyna-Warna seismic zone (KWSZ) region, with the P- and S-wave catalog data from relative arrival time pairs constituting the input. A significant improvement in the hypocentral estimates was achieved, with the epicentral errors <30 m and focal depth errors <75 m i.e. errors have been significantly reduced by an order of magnitude from the parameters determined by HYPO71. The earthquake activity defines three different fault segments. The seismogenic volume is shallower in the south by 3 km, with seismicity in the north extending to a depth of 11 km while in the south the deepest seismicity observed is at a depth of 8 km. By resolving the structure of seismicity in greater detail, we address the salient issues related to the seismotectonics of this region.     

Earthquake forerunner as probable precursor — an example from north burma subduction zone

The Burmese Arc seismic activity is not uniform for its ∼ 1100 km length; only the Northern Burmese Arc (NBA) is intensely active. Six large earthquakes in the magnitude range 6.1–7.4 have originated from the NBA Benioff zone between 1954–2011, within an area of 200 × 300 km² where the Indian plate subducts eastward to depths beyond 200 km below the Burma plate. An analysis on seismogenesis of this interplate region suggests that while the subducting lithosphere is characterized by profuse seismicity, seismicity in the overriding plate is rather few. Large earthquakes occurring in the overriding plate are associated with the backarc Shan-Sagaing Fault (SSF) further east. The forecasting performance of the Benioff zone earthquakes in NBA as forerunner is analysed here by: (i) spatial earthquake clustering, (ii) seismic cycles and their temporal quiescence and (iii) the characteristic temporal b-value changes. Three such clusters (C1–C3) are identified from NBA Benioff Zones I & II that are capable of generating earthquakes in the magnitude ranges of 7.38 to 7.93. Seismic cycles evidenced for the Zone I displayed distinct quiescence (Q₁, Q₂ and Q₃) prior to the 6^th August 1988 (M 6.6) earthquake. Similar cycles were used to forecast an earthquake (Dasgupta et al. 2010) to come from the Zone I (cluster C1); which, actually struck on 4 February 2011 (M 6.3). The preparatory activity for an event has already been set in the Zone II and we speculate its occurrence as a large event (M > 6.0) possibly within the year 2012, somewhere close to cluster C3. Temporal analysis of b-value indicates a rise before an ensuing large earthquake.     

Multi-scale imaging of a slow active fault zone: contribution for improved seismic hazard assessment in the Swiss Alpine foreland

Seismic hazard assessment of slow active fault zones is challenging as usually only a few decades of sparse instrumental seismic monitoring is available to characterize seismic activity. Tectonic features linked to the observed seismicity can be mapped by seismic imaging techniques and/or geomorphological and structural evidences. In this study, we investigate a seismic lineament located in the Swiss Alpine foreland, which was discussed in previous work as being related to crustal structures carrying in size the potential of a magnitude M 6 earthquake. New, low-magnitude (?2.0 ≤ M_L ≤ 2.5) earthquake data are used to image the spatial and temporal distribution of seismogenic features in the target area. Quantitative and qualitative analyses are applied to the waveform dataset to better constrain earthquakes distribution and source processes. Potential tectonic features responsible for the observed seismicity are modelled based on new reinterpretations of oil industry seismic profiles and recent field data in the study area. The earthquake and tectonic datasets are then integrated in a 3D model. Spatially, the seismicity correlates over 10–15 km with a N–S oriented sub-vertical fault zone imaged in seismic profiles in the Mesozoic cover units above a major decollement on top of the mechanically more rigid basement and seen in outcrops of Tertiary series east of the city of Fribourg. Observed earthquakes cluster at shallow depth (<4 km) in the sedimentary cover. Given the spatial extend of the observed seismicity, we infer the potential of a moderate size earthquake to be generated on the lineament. However, since the existence of along strike structures in the basement cannot be excluded, a maximum M 6 earthquake cannot be ruled out. Thus, the Fribourg Lineament constitutes a non-negligible source of seismic hazard in the Swiss Alpine foreland.     

辽宁海城及其邻区地震b值空间分布特征

震级-频度分布（FMD）是地震学研究中最重要的经验公式之一,相关系数b是构造学和地震危险性评估的重要因子,具有表征前震和余震的特性。辽宁省地震多发生在金州断裂附近,自1975年海城7.3级地震发生后,与金州断裂交汇的海城河—大洋河断裂开启活跃模式,其东南端岫岩附近在1999年又发生5.6级地震。近年来盖州附近地震活动也在增强。因此,本文利用b值空间分布特征对海城及其邻区的应力分布特点进行研究。震源定位准确与否直接影响b值计算,双差定位后的数据与常规目录相比具有更高的精度,但是完整性有一定下降。本文收集了中国地震台网1981—2005年的辽宁省地震目录,并进行双差定位,比较分析了常规目录数据和双差数据的b值分布差异,认为在地震密集区,双差定位后的数据可以被用来获得更准确的b值。对主要研究区进行网格划分,使用双差数据,得到b值的水平和垂直分布特征。结果表明：b值为0.6~1.8,随深度增加而降低;岫岩和盖州震区具有较低的b值,意味着具有较高的地震危险性;浑河震区与海城河—大洋河断裂东南方向具有较高的b值,说明该区域未来发生大地震的概率很低;与金州断裂交汇区域的b值在1.0附近,说明该地区应力暂时处于稳定状态,未来具有较低的地震危险性。     

Variations of Seismicity in the Avachinsky Gulf (Kamchatka, Russia)

Variations of seismic mode in the region of the Avachinsky Gulf (Kamchatka, Russia) are considered. Observed anomalies (seismic quiescence, the ring seismicity, reduction of the slope of the earthquake recurrence diagram) provide a basis to consider this region as a place of strong earthquake preparation. The Kamchatka regional catalogues of earthquakes between 1962–1995 were used in the analysis. A reduced seismicity rate is observed during 10 years in an area of 150 km × 60 km in size. During the last five years, in the vicinity of the area considered, earthquakes with M > 5 occurred three times more often than the average over thirty years. It is interpreted as ring seismicity. The block of 220 km × 220~km in size, including the quiescence zone, is characterized by a continuous decrease of the recurrence diagram slope, which has reached a minimum value for the last 33 years in this region.     

Seismic evidence for thick and underplated late Archaean crust of eastern Dharwar craton

The deep crustal structure of eastern Dharwar craton has been investigated through τ-p extremal inversion of P-wave travel times from a network of seismographs recording quarry blasts. Travel times have been observed in the distance range 30–250 km in a laterally homogeneous lithospheric segment Main features of the inferred velocity-depth relationship include: (a) 29 km thick combined upper and middle crust velocity varying from 6 km/s to 7 km/s, with no observable velocity discontinuity in this depth range; (b) a lower crust (∼ 29–41 km) with velocity increasing from 7.0 to 7.3 km/s; (c) an average upper mantle velocity of 8.1 km/s; and (d) presence of a 12 km thick high velocity crustal layer (7.4 – 7.8 km/s) in the depth range 41–53 km, with a distinct velocity gradient marking a velocity increase of 0.4 km/s. The anomalous 53 km thick crust is viewed as a consequence of magmatic underplating at the base of the crust in the process of cratonization of the eastern Dharwar craton during late Archaean. The underplated material reflects here with the velocity of 7–3 to 7–8 km/s below the depth of 40 km. Our proposition of magmatic underplating is also supported by the presence of large scale I-granitoid, a product of partial melting of the upper mantle material.     

Magma plumbing beneath Anak Krakatau volcano, Indonesia: evidence for multiple magma storage regions

Understanding magma plumbing is essential for predicting the behaviour of explosive volcanoes. We investigate magma plumbing at the highly active Anak Krakatau volcano (Indonesia), situated on the rim of the 1883 Krakatau caldera by employing a suite of thermobarometric models. These include clinopyroxene-melt thermobarometry, plagioclase-melt thermobarometry, clinopyroxene composition barometry and olivine-melt thermometry. Petrological studies have previously identified shallow magma storage in the region of 2–8 km beneath Krakatau, while existing seismic evidence points towards mid- to deep-crustal storage zone(s), at 9 and 22 km, respectively. Our results show that clinopyroxene in Anak Krakatau lavas crystallized at a depth of 7–12 km, while plagioclase records both shallow crustal (3–7 km) and sub-Moho (23–28 km) levels of crystallization. These magma storage regions coincide with well-constrained major lithological boundaries in the crust, implying that magma ascent and storage at Anak Krakatau is strongly controlled by crustal properties. A tandem seismic tomography survey independently identified a separate upper crustal (<7 km) and a lower to mid-crustal magma storage region (>7 km). Both petrological and seismic methods are sensitive in detecting magma bodies in the crust, but suffer from various limitations. Combined geophysical and petrological surveys, in turn, offer increased potential for a comprehensive characterization of magma plumbing at active volcanic complexes.