Search

During the Mw9 Tohoku-Oki earthquake, gradual increases in both ionosphere total electron content (TEC) and geomagnetic declination signals were observed, starting from ~40 minutes before the mainshock, followed by impulsive enhancements ~10 minutes after the mainshock. There have been many studies on pre-seismic TEC enhancements, including their characteristics, debates regarding whether TEC anomalies are real signals or artefacts, and the explainable models, and many studies have reported that the impulsive TEC enhancement was caused by a tsunami-induced neutral atmospheric gravity wave. Since TEC and geomagnetic declination anomalies were synchronized so that their origin should be attributed to the same seismic activities, any models must explain both anomalous phenomena, but not the case considered herein. Compared with the corresponding TEC anomalies, we re-examined the characteristics of geomagnetic variation just before and after the mainshock, focusing on the generation process of the impulsive enhancement immediately after the mainshock. We showed that the observed anomaly could be explained if there are quasi-static electric currents of 20 - 30 kA generated near the epicentre area. The possible mechanism of the current generation is discussed in terms of the ionization process in the atmosphere near the sea surface.

Frequency spectra of seismic waves from a fault rupture reflects the size of the faults, i.e. relatively large amplitudes of long period waves are excited by larger earthquakes. Anomalies in rise times of the fault movements would also influence the spectra. For example, earthquakes characterized by slow faulting, known as tsunami earthquakes, excite large tsunamis for the amplitudes of short-period seismic waves. In this study, we compare amplitudes of long- and short-period atmospheric waves excited by vertical crustal movements associated with earthquake faulting. Such atmospheric waves often reach the ionospheric F region and cause coseismic ionospheric disturbances (CID) observed as oscillations in ionospheric total electron content (TEC), with ground Global Navigation Satellite System (GNSS) receivers. CID often includes long-period internal gravity wave (IGW) components in addition to short period acoustic wave (AW) components. The latter has a period of ~4 minutes and propagate by 0.8-1.0 km/s, while the former has a period of ~12 minutes and propagate as fast as 0.2-0.3 km/s. Here we compare amplitudes of these two different waves for five earthquakes, 2011 Tohoku-oki (Mw9.0), 2010 Maule (Mw8.8), 1994 Hokkaido-Toho-Oki (Mw8.3), 2003 Tokachi-oki (Mw8.0), and the 2010 Mentawai (Mw7.9) earthquakes, using data from regional dense GNSS networks. We found two important features, i.e. (1) larger earthquakes show larger IGW/AW amplitude ratios, and (2) Mentawai earthquake, a typical tsunami earthquake, exhibits abnormally large IGW amplitudes relative to AW amplitudes. These findings demonstrate that earthquakes with longer durations for faulting, or with longer times for vertical crustal movements, excite longer period atmospheric waves such as IGW more efficiently.

Atmospheric waves excited by strong surface explosions, both natural and anthropogenic, often disturb upper atmosphere. In this letter, we report an N-shaped pulse with period ~1.3 minutes propagating southward at ~0.8 km/s, observed as changes in ionospheric total electron content using continuous GNSS stations in Israel and Palestine, ~10 minutes after the August 4, 2020 chemical explosion in Beirut, Lebanon. The peak-to-peak amplitude of the disturbance reached ~2% of the background electrons, comparable to recently recorded volcanic explosions in the Japanese Islands. We also succeeded in reproducing the observed disturbances assuming acoustic waves propagating upward and their interaction with geomagnetic fields.Keywords: Chemical explosion, Beirut, N-shaped pulse, Total electron content

Atmospheric waves excited by strong surface explosions, both natural and anthropogenic, often disturb upper atmosphere. In this letter, we report an N-shaped pulse with period ~ 1.3 min propagating southward at ~ 0.8 km/s, observed as changes in ionospheric total electron content using continuous GNSS stations in Israel and Palestine, ~ 10 min after the August 4, 2020 chemical explosion in Beirut, Lebanon. The peak-to-peak amplitude of the disturbance reached ~ 2% of the background electrons, comparable to recently recorded volcanic explosions in the Japanese Islands. We also succeeded in reproducing the observed disturbances assuming acoustic waves propagating upward and their interaction with geomagnetic fields.

A dense network of ground global navigation satellite system receivers detected ionospheric total electron content (TEC) changes starting similar to 40 min before the 2011 Tohoku-oki (M(w)9.0) earthquake around the ruptured fault, together with the long-lasting postseismic TEC drop. In this paper, we robustly estimate three-dimensional (3-D) distribution of both preseismic and postseismic ionospheric anomalies of the 2011 Tohoku-oki earthquake by tomographic inversions of electron density anomalies. We set up >6,000 blocks, as large as 1.0 degrees (east-west) x 0.9 degrees (north-south) x 60 km (vertical), over the Japanese Islands, the Sea of Japan, and the Korean Peninsula, up to 870 km altitude. By using TEC anomalies of pairs exceeding similar to 1,300 stations and eight satellites obtained using reference curves, we estimated electron density anomalies within individual blocks. The results showed that the preseismic and postseismic anomalies do not overlap in space. The preseismic anomalies are composed of low (similar to 300 km height) positive and high (similar to 600 km height) negative anomalies. They occurred above the land of NE Japan without extending offshore, suggesting its origin related to surface electric charges. On the other hand, the postseismic electron depletion occurred offshore above the region where large coseismic uplift took place. These results demonstrate that the preseismic and postseismic ionospheric anomalies are independent not only temporarily but also spatially and certainly in underlying physical mechanisms. We propose a simple model to explain how surface charges redistribute ionospheric electrons to make the observed preseismic electron density anomalies. Plain Language Summary A dense network of GNSS/GPS receivers found that redistribution of ionospheric electrons started similar to 40 min before the 2011 Tohoku-oki earthquake around the fault. It was also found that a long-lasting electron depletion occurred after the earthquake. We studied the three-dimensional structures of the electron density anomalies immediately before and after the 2011 earthquake. We found that the preseismic anomaly occurred above land, but the postseismic anomaly occurred offshore. The preseismic change is characterized by the simultaneous growth of positive and negative electron density anomalies, while the postseismic change is dominated by an electron decrease. These differences reflect the different physical origins of the preseismic and postseismic anomalies.

Strong typhoons hit the Japanese Islands repeatedly in 2019. Here we study one of these typhoons (2019 #19 Hagibis 915 hPa, 86 casualties) that landed central Japan on Oct.12 (local time) during the Rugby World Cup tournament, using two different space geodetic approaches, i.e. water vapor and crustal deformation. The first approach is the recovery of Precipitable Water Vapor (PWV) using the zenith wet delays (ZWD) estimated by the dense GNSS array in Japan GEONET. Because atmospheric water vapor concentrates in relatively low altitudes, high humidity is often difficult to recognize in ZWDs when the surface altitude is high. To overcome the difficulty, we reconstructed ZWDs, converted to sea-level values, by spatially integrating the tropospheric delay gradient (azimuthal asymmetry of water vapor) vectors. We also calculated convergence of such delay gradients, equivalent to water vapor convergence index (WVCI) proposed by Shoji (2013 Jour. Met. Soc. Japan). We found that very strong rainfall occurs in the region where both reconstructed ZWD and the delay gradient convergence index are high. Next, we studied vertical crustal movements associated with the water load brought by the typhoon, using the two solutions of the GEONET station coordinates, one from the official F3 solution and the other from the UNR data base. We confirmed subsidence down to ~2 cm in multiple regions where severe flood occurred. Such subsidence was observed to recover with a time constant of 1-2 days reflecting rapid drain of rain water to ocean due to large topographic slope and proximity to the sea. We could not identify, however, crustal uplift due to the low atmospheric pressure at the center of the typhoon.

Using the ionospheric total electron content (TEC) data from ground-based Global Navigation Satellite System (GNSS) receivers in Japan, we compared ionospheric responses to five explosive volcanic eruptions 2004–2015 of the Asama, Shin-Moe, Sakurajima, and Kuchinoerabu-jima volcanoes. The TEC records show N-shaped disturbances with a period ~ 80 s propagating outward with the acoustic wave speed in the F region of the ionosphere. The amplitudes of these TEC disturbances are a few percent of the background absolute vertical TEC. We propose to use such relative amplitudes as a new index for the intensity of volcanic explosions. Graphical Abstract

Using ionospheric total electron contents residual data from 146 Global Navigation Satellite System stations in South America, we conducted three-dimensional tomography of ionospheric electron density anomalies immediately before the 2015 Illapel M(w)8.3 earthquake, Central Chile. We used five GPS and five GLONASS satellites and applied continuity constraints to regularize the linear least squares inversion. The reconstructed anomalies are composed of positive and negative regions, at altitudes of similar to 200 km and similar to 400 km, respectively, distributed roughly along the geomagnetic field. This feature suggests that the observed anomalies occurred by the downward E x B drift of electrons due to electric fields within ionosphere, possibly caused by surface positive electric charges. We also discuss the existence of the mirror image anomalies near the geomagnetic conjugate point of the epicenter using stations in Colombia.

We studied ionospheric irregularities caused by midlatitude sporadic-E (Es) in Japan using ionospheric total electron content (TEC) data from a dense GNSS array, GEONET, with a 3D (three-dimensional) tomography technique. Es is a thin layer of unusually high ionization that appears at altitudes of ~ 100 km. Here, we studied five cases of Es irregularities in 2010 and 2012, also reported in previous studies, over the Kanto and Kyushu Districts. We used slant TEC residuals as the input and estimated the number of electron density anomalies of more than 2000 small blocks with dimensions of 20–30 km covering a horizontal region of 300 × 500 km. We applied a continuity constraint to stabilize the solution and performed several different resolution tests with synthetic data to assess the accuracy of the results. The tomography results showed that positive electron density anomalies occurred at the E region height, and the morphology and dynamics were consistent with those reported by earlier studies.

We studied 38 slow slip events (SSEs) in 1997-2016 beneath the Iriomote Island, southwestern Ryukyu Arc, Japan, using continuous Global Navigation Satellite System (GNSS) data. These SSEs occur biannually on the same fault patch at a depth of ~30 km on the subducting Philippine Sea Plate slab with average moment magnitudes (Mw) of ~6.6. Here we show that the slip accumulation rate (cumulative slip/lapse time) of these SSEs fluctuated over a decadal time scale. The rate increased twice around 2002 and 2013 concurrently with earthquake swarms in the Okinawa Trough. This suggests that episodic activations of the back-arc spreading at the Okinawa Trough caused extra southward movement of the block south of the trough and accelerated convergence at the Ryukyu Trench.

Recent observations suggested that ionospheric anomalies appear immediately before large earthquakes with moment magnitudes (Mw) of 8.2 or more. Do similar phenomena precede smaller earthquakes? Here we answer this question by analyzing vertical total electron contents (VTEC) observed near the epicenters before and after 32 earthquakes with Mw7.0–8.0 using data from nearby Global Navigation Satellite Systems stations. To detect anomalies, we defined the reference curves to fit the observed VTEC and considered the departure from the curves as anomalies. In estimating the reference curves, we excluded time windows, prescribed for individual earthquakes considering Mw, possibly affected by earthquakes. We validated the method using synthetic VTEC data assuming both preseismic, coseismic, and postseismic anomalies. Out of the 32 Mw7.0–8.0 earthquakes, eight earthquakes showed possible preseismic anomalies starting 10–20 min before earthquakes. For earthquakes of this Mw range, we can observe preseismic ionospheric changes probably when the background VTEC is large, say 50 TECU (total electron content unit, 1 TECU = 1016 el m−2) or more.

The Gravity Recovery and Interior Laboratory (GRAIL) mission unraveled numbers of features showing strong gravity anomalies without prominent topographic signatures in the lunar Procellarum region. These features, located in different geologic units, are considered to have complex subsurface structures reflecting different evolution processes. By using the GRAIL level-1 data, we estimated the free-air and Bouguer gravity anomalies in several selected regions including such intriguing features. With the three-dimensional inversion technique, we recovered subsurface density structures in these regions.

The Pacific Plate subducts beneath the Philippine Sea Plate along the Izu-Bonin Trench. We investigated crustal movements at the Bonin Islands, using Global Navigation Satellite System and geodetic Very Long Baseline Interferometry data to reveal how the two plates converge in this subduction zone. These islands are located similar to 100 km from the trench, just at the middle between the volcanic arc and the trench, making these islands suitable for detecting signatures of episodic deformation such as slow slip events (SSEs). During 2007-2016, we found five SSEs repeating quasi-periodically with similar displacement patterns. In estimating their fault parameters, we assumed that the fault lies on the prescribed plate boundary, and optimized the size, shape and position of the fault and dislocation vectors. Average fault slip was similar to 5 cm, and the average moment magnitude was similar to 6.9. We also found one SSE occurred in 2008 updip of the repeating SSE in response to an M6 class interplate earthquake. In spite of the frequent occurrence of SSEs, there is no evidence for long-term strain accumulation in the Bonin Islands that may lead to future megathrust earthquakes. Plate convergence in Mariana-type subduction zones may occur, to a large extent, episodically as repeating SSEs.

Growing evidence for ionospheric signatures of impending earthquakes comes from electron content measurements along slanted paths from GPS satellites to multiple ground stations located up to 500 km away from the epicenters. These slant total electron content (STEC) measurements deviate from the classic U-shape pattern, starting about 40 min to over an hour before major earthquakes. Unlike other naturally occurring STEC fluctuations at midlatitudes, we show here that these earthquake-induced deviations are simultaneous over a wide geographical area and do not propagate, thereby indicating a ground-based origin. Prior to the 11 March 2011 Tohoku-Oki earthquake (Mw 9.0), the deviations were as much as 10% of the undisturbed STEC. We argue that such deviations must be due to an electric field-forced rise or fall of the main ionosphere with little change in the vertical electron density profile. Hence, “apparent” is used in the title. We show how stress-related underground electric fields penetrate to 80 km altitude (above which penetration to the main ionosphere easily occurs) with magnitudes high enough to create STEC variations comparable to those observed. Since many thousands of GPS receivers exist worldwide, our theory suggests the possibility of early warning systems that could provide 10 to 20 min notice prior to large earthquakes, after allowing time for signal processing. This theory for prequake-induced STEC fluctuations also explains the ground-based ULF magnetic field data acquired by Fraser-Smith et al. 40 min prior to the Loma Prieta earthquake.

Landward increase of surface velocity has been found for segments adjacent along-strike to megathrust faults after the 2003 Tokachi-oki and the 2011 Tohoku-oki earthquakes, NE Japan. A similar increase of landward velocities was reported for the segments to the north of the rupture of the 2010 Maule earthquake, Chile. We utilize available GNSS data to find such changes for six megathrust earthquakes in four subduction zones, including NE Japan, central and northern Chile, Sumatra, and Mexico to investigate their common features. Our study showed that such increase, ranging from a few mm/yr to ~1 cm/yr, appeared in adjacent segments following the 2014 Iquique (Chile), the 2007 Bengkulu (Sumatra), and the 2012 Oaxaca (Mexico) earthquakes in addition to the three cases. The region of the increased landward movements extends with spatial decay and reach the distance comparable to the along-strike fault length. On the other hand, the temporal decay of the increased velocity is not clear at present. The degree of increase seems to depend on the earthquake magnitude, and possibly scales with the average fault slip in the earthquake. This is consistent with the simple two-dimensional model proposed earlier to attribute the phenomenon to the enhanced coupling caused by accelerated subduction. However, these data are not strong enough to rule out other possibilities.

Using continuous data from ground-based Global Satellite Navigation System (GNSS) receivers in Java and Sumatra, Indonesia, we studied the response of ionospheric total ionospheric electron content (TEC) to the 2010 Nov.5 eruption of the Merapi volcano in central Java. We then compared the results with the case of the 2014 Feb.13 eruption of the Kelud volcano, eastern Java. The TEC showed a quasi-periodic oscillation of a frequency ~4 mHz with average amplitudes of 0.9 and 1.8% relative to background values lasting for ~20 and ~ 120 min for the Merapi and Kelud eruptions, respectively. By comparing the two cases, together with the 2015 April eruption of the Calbuco volcano, Chile, we found the relative TEC oscillation amplitude may scale with the mass eruption rate. This suggests that the product of such TEC oscillation amplitude and the duration provides a new measure for the total volume of the volcanic deposits.

Using regional Global Positioning System (GPS) networks, we studied three-dimensional spatial structure of ionospheric total electron content (TEC) anomalies preceding three recent large earthquakes in Chile, South America, i.e. the 2010 Maule (Mw8.8), the 2014 Iquique (Mw8.2), and the 2015 Illapel (Mw8.3) earthquakes. Both positive and negative TEC anomalies, with areal extent dependent on the earthquake magnitudes, appeared simultaneously 20-40 minutes before the earthquakes. For the two mid-latitude earthquakes (2010 Maule and 2015 Illapel), positive anomalies occurred to the north of the epicenters at altitudes 150-250 km. The negative anomalies occurred further to the north at higher altitudes 200-500 km. This lets the epicenter, the positive and negative anomalies align parallel with the local geomagnetic field, which is a typical structure of ionospheric anomalies occurring in response to positive surface electric charges.

The Izu–Bonin arc lies along the convergent boundary where the Pacific Plate subducts beneath the Philippine Sea Plate. Horizontal velocities of continuous Global Navigation Satellite System stations on the Izu Islands move eastward by up to ~ 1 cm/year relative to the stable part of the Philippine Sea Plate suggesting active back-arc rifting behind the northern part of the arc. Here, we report that such eastward movements transiently accelerated in the middle of 2004 resulting in ~ 3 cm extra movements in 3 years. We compare three different mechanisms possibly responsible for this transient movement, i.e. (1) postseismic movement of the 2004 September earthquake sequence off the Kii Peninsula far to the west, (2) a temporary activation of the back-arc rifting to the west dynamically triggered by seismic waves from a nearby earthquake, and (3) a large slow slip event in the Izu–Bonin Trench to the east. By comparing crustal movements in different regions, the first possibility can be shown unlikely. It is difficult to rule out the second possibility, but current evidence support the third possibility, i.e. a large slow slip event with moment magnitude of ~ 7.5 may have occurred there.

Kilometer-scale fine structures of midlatitude sporadic E (Es) plasma patches have been directly imaged for the first time by an interferogram derived from L band Advanced Land Observation Satellite/Phased Array-type L band Synthetic Aperture Radar data obtained over southwestern Japan. The synthetic aperture radar interferogram captured the eastern part of a large-scale frontal structure of daytime midlatitude Es which spans over 250 km in the east-northeast to west-southwest direction. Fine structures are characterized by frontal and disc-shaped patches which are elongated in the same direction as the large-scale frontal structure. Length and width of the disc-shaped patches are 10–20 km and 5–10 km, respectively, and they are quasi-periodically located with a typical separation of 10–15 km. The Kelvin-Helmholtz instability with the vertical shear of zonal winds is considered to be the most likely candidate for the generation mechanism of the frontal patch and disc-shaped patches aligned in the zonal direction.

International audience

Ionospheric electron enhancement was reported to have occurred similar to 40min before the 2011 Tohoku-oki (M(w)9.0) earthquake, Japan, by observing total electron content (TEC) with Global Navigation Satellite Systems receivers. Their reality has been repeatedly questioned due mainly to the ambiguity in the derivation of the reference TEC curves from which anomalies are defined. Here we propose a numerical approach, based on Akaike's information criterion, to detect positive breaks (sudden increase of TEC rate) in the vertical TEC time series without using reference curves. We demonstrate that such breaks are detected 25-80min before the eight recent large earthquakes with moment magnitudes (M-w) of 8.2-9.2. The amounts of precursory rate changes were found to depend upon background TEC as well as M-w. The precursor times also showed M-w dependence, and the precursors of intraplate earthquakes tend to start earlier than interplate earthquakes. We also performed the same analyses during periods without earthquakes to evaluate the usefulness of TEC observations for short-term earthquake prediction.

We try to detect an unidentified signal from the surface motion at northeastern Japan immediately after the 2011 Tohoku earthquake. A focused frequency range is 10–100 mHz (10–100 s). We find a peaky signal with frequency of about 38 mHz (26 s) based on the horizontal-to-vertical (H/V) spectral ratio using the high-rate GNSS data at 382 GEONET stations. We are not able to identify locality of the signal. The signal appears several minutes after the passing of surface wave fronts. The duration of the signal is about 2 min. Since the origin of the 38 mHz signal is unlikely to be local hydrologic tremors, tectonic tremors, or the tsunami, we speculate that the 38 mHz signal originates from a kind of a characteristic oscillation of Northeastern Japan triggered by the 2011 Tohoku earthquake. A normal-mode simulation implies that high-order radial overtones could create the signal with a spherically-layered velocity structure, however, the detailed mechanism of the signal still remains a mystery.

Coseismic gravity changes stem from (1) vertical deformation of layer boundaries with density contrast (i.e., surface and Moho) and (2) density changes of rocks at depth. They have been observed in earthquakes with Mw exceeding ~8.5 by Gravity Recovery and Climate Experiment (GRACE) satellites, but those of M8 class earthquakes have never been detected clearly. Here we report coseismic gravity change of the 24 May 2013 Okhotsk deep earthquake (Mw8.3), smaller than the detection threshold. In shallow thrust faulting, factor (2) is dominant, while factor (1) remains secondary due to poor spatial resolution of GRACE. In the 2013 Okhotsk earthquake, however, factor (2) is insignificant because they occur at depth exceeding 600 km. On the other hand, factor (1) becomes dominant because the centers of uplift and subsidence are well separated and GRACE can resolve them. This enables GRACE to map vertical ground movements of deep earthquakes over both land and ocean.

Heki (2011) reported that the Japanese Global Positioning System (GPS) dense network detected a precursory positive total electron content anomaly (TEC), with Delta TEC similar ~3 TECU, similar ~40 minutes before the Tohoku-Oki earthquake (Mw 9.0). Similar preseismic TEC anomalies were also observed in the 2010 Chile earthquake (Mw 8.8), 2004 Sumatra-Andaman (Mw 9.2) and the 1994 Hokkaido-Toho-Oki (Mw 8.3). In this paper we apply our improved lithosphere-atmosphere-ionosphere (LAI) coupling model to compute the TEC variations and compare the simulation results with the reported TEC observations. For the Tohoku-Oki earthquake simulations we assumed that the stressed associated current started similar ~40 minutes before the earthquake, linearly increased and reached its maximum magnitude at the time of the earthquake main shock. It is suggested that a dynamo current density of similar ~25 nA m(-2) is required to produce the observed Delta TEC similar ~3 TECU.

We studied ionospheric responses to the 2012 April 11 M-w 8.6 North Sumatra earthquake using total electron content (TEC) measurements with the regional Global Navigation Satellite System network. This earthquake ruptured the oceanic lithosphere off the Indian Ocean coast of North Sumatra, and is known as the largest strike-slip earthquake ever recorded. Coseismic ionospheric disturbances (CIDs) with rapid TEC enhancement of a few TEC units propagated northward with a speed of acoustic waves (similar to 1 kms(-1)). Resonant atmospheric oscillation with a frequency similar to 4 mHz have been found as monochromatic oscillation of TEC lasting for an hour after the main shock and the largest aftershock. We compared CID amplitudes of 21 earthquakes world-wide with moment magnitudes (M-w) 6.6-9.2. They roughly obeyed a law such that CID amplitude increases by two orders of magnitude for the M-w increase of three. The 2012 North Sumatra earthquakes slightly deviated negatively from the trend possibly reflecting their strike-slip mechanisms, that is small vertical crustal movements for their magnitudes.

Using monthly satellite gravimetry data, we studied time-variable gravity field after three M9 class earthquakes, the 2004 Sumatra-Andaman, 2010 Chile (Maule), and 2011 Tohoku-oki earthquakes. The observations showed that the gravity typically (1) decreases coseismically, (2) continues to decrease for a few months, and (3) increases over a longer period. Therefore, postseismic gravity changes have two components with different time constants and polarities. The mechanisms of short- and long-term postseismic gravity changes are not as clear as coseismic changes at the moment, but might be explained to some extent with afterslip and the Maxwell viscoelasticity, respectively. These two components are difficult to discriminate with surface velocity measurements because the forearc area moves trenchward at both stages. They appear in different polarities in gravity, making satellite gravimetry a unique tool to separate them.

[1] We observed two-dimensional structure and time evolution of ionospheric irregularities caused by midlatitude sporadic E (Es) over Japan as positive anomalies of total electron content (TEC) by analyzing the data from the nationwide Global Positioning System (GPS) array. In this paper we report a case study of strong Es observed in the local evening of 21 May 2010, over Tokyo, Japan. In the slant TEC time series, Es showed a characteristic pulse-like enhancement of ~1.5 TEC units lasting for ~10 min. We plotted these positive TEC anomalies on the subionospheric points of station-satellite pairs to study the horizontal structure of the Es irregularity. We confirmed that the irregularity existed at the height of ~106 km by comparing the data of multiple GPS satellites, which is consistent with the local ionosonde observations. The horizontal shapes of the Es irregularity showed frontal structures elongated in E-W, spanning ~150 km in length and ~30 km in width, composed of small patches. The frontal structure appears to consist of at least two parts propagating in different directions: one moved eastward by ~60 m s−1, and the other moved southwestward by ~80 m s−1. Similar TEC signatures of Es were detected by other GPS satellites, except one satellite that had line of sight in the N-S direction which dips by 40–50° toward north, which indicates the direction of plasma transportation responsible for the Es formation. We also present a few additional observation results of strong Es irregularities.

[1] Possible enhancement of ionospheric Total Electron Content (TEC) immediately before the 2011 Tohoku-oki earthquake (Mw9.0) has been reported by Heki (2011). Critical responses to it often come in two stages; they first doubt the enhancement itself and attribute it to an artifact. Second (when they accept the enhancement), they doubt the significance of the enhancement among natural variability of space weather origin. For example, Kamogawa and Kakinami (2013) attributed the enhancement to an artifact falsely detected by the combined effect of the highly variable TEC under active geomagnetic condition and the occurrence of a tsunamigenic ionospheric hole. Here we closely examine the time series of vertical TEC before and after the 2011 Tohoku-oki earthquake. We first demonstrate that the tsunami did not make an ionospheric hole, and next confirm the reality of the enhancement using data of two other sensors, ionosonde and magnetometers. The amplitude of the preseismic TEC enhancement is within the natural variability, and its snapshot resembles to large-scale traveling ionospheric disturbances. However, distinction could be made by examining their propagation properties. Similar TEC anomalies occurred before all the M ≥ 8.5 earthquakes in this century, suggesting their seismic origin.

[1] We detect the acceleration of ice mass depletion in Greenland for the period 1991–2011 from the low-degree gravity field up to degree and order 4 derived from the Satellite Laser Ranging (SLR) data. Between 2003 and 2011, during the era when the GRACE (Gravity Recovery and Climate Experiment) satellite data are available, our SLR results of gravity changes agree well with GRACE showing significant negative patterns in Greenland. Prior to that, the SLR linear trend maps show a near balance in Greenland ice mass during 1991–2002 (after the glacial isostatic adjustment is accounted for using model values). We further confirm the consistency of our SLR results with the vertical crustal uplift of the region observed by the Global Positining System (GPS) which manifests ice mass loading/unloading. Thus, the SLR data series constitute a continuous benchmark for the time history of the Greenland ice mass changes for over two decades.

S U M M A R Y We explore the use of on-land GPS observations to detect deformation due to tsunami propagation near source regions of large interplate earthquakes. Here, we focus on the Mw 9 Tohoku-oki earthquake, which occurred around 14:46 (JST) on 2011 March 11. We consider GPS data in the time span 14:54–15:22 (JST) along the Sanriku coast, where the tsunami had the largest amplitude. The displacement data shows the signatures of large aftershocks as well as post-seismic fault slip (afterslip). These effects are particularly evident in the east component. From the horizontal displacement vectors, we construct a simple fault model for the early phase of the afterslip. Mean slip velocity of the early afterslip reaches 0.1 mm s−1. By compiling the early afterslip velocity of recent interplate earthquakes around that region, we find its increasing trend with the main shock magnitude. This scaling relation may reflect higher stressing rates at edges of larger main shock faults. Separately, we forward calculate land deformation due to tsunami height changes based on a tsunami simulation. Tsunami-induced deformation is only evident in the vertical direction at coastal GPS stations. The predicted subsidence amounts at some coastal stations can account for a large portion of the residuals between the observation and the modelled deformation due to the fault slip.

[1] We studied ionospheric disturbances associated with the two large earthquakes in Sumatra, Indonesia, namely, the 2007 Bengkulu and the 2005 Nias earthquakes, by measuring the total electron contents (TEC) using a regional network of global positioning system (GPS) receivers. We first focus on coseismic ionospheric disturbances (CIDs) of the Bengkulu earthquake (Mw 8.5). They appeared 11–16 min after the earthquake and propagated northward as fast as ~0.7 km/s, consistent with the sound speed at the ionospheric F layer height. Resonant oscillation of TEC with a frequency of ~5 mHz continued for at least 30 min after the earthquake. The largest aftershock (Mw 7.9) also showed clear CIDs similar to the main shock. A CID propagating with the Rayleigh wave velocity was not observed, possibly because the station distribution did not favor the radiation pattern of the surface waves. This earthquake, which occurred during a period of quiet geomagnetic activity, also showed clear preseismic TEC anomalies similar to those before the 2011 Tohoku-Oki earthquake. The positive and negative anomalies started 30–60 min before the earthquake to the north and the south of the fault region, respectively. On the other hand, we did not find any long-term TEC anomalies within 4– 5d ays before the earthquake. Co- and preseismic ionospheric anomalies of the 2005 Nias earthquake (Mw 8.6) were, however, masked by strong plasma bubble signatures, and we could not even discuss the presence or absence of CIDs and preseismic TEC changes for this earthquake.

Interplate thrust earthquakes are usually followed by afterslips, and they let the fore-arc move slowly trenchward. However, we do not know if the subducting oceanic plate is accelerated landward after such earthquakes. The westward velocity of Global Positioning System (GPS) stations in NE Japan show gradient decreasing from east to west reflecting the E-W contractional strain built up by the inter-plate coupling. Here we show that such coupling significantly enhanced (∼1.5 times) after the 2003 Tokachi-Oki earthquake (Mw 8.0), Hokkaido, in the segments adjacent to the ruptured fault. The coupling seems to be further enhanced (∼3 times) after the 2011 Tohoku-Oki earthquake (Mw 9.0). It is unlikely that interplate friction suddenly increased over such a large region, and relatively strong pre-2003 coupling there would not allow such enhancements even if full coupling is attained. Hence they are attributable to the temporary acceleration of the Pacific Plate subduction. We propose a simple 2-dimensional model in which down-dip acceleration of the slab let the force balance rapidly recover promoted by a thin low-viscosity layer on the slab surface. The accelerated subduction would account for temporary activations of regional interplate seismicity after megathrust earthquakes.

Volcanic plumes modify atmospheric temperatures. Wang et al. (2009) reported localized temperature changes associated with the Mt Chaiten eruption in May 2008 using the deviation of GPS radio occultation temperature data from the NCEP global forecast system model. On 14 April 2010, a large scale eruption started under the glacier Eyjafjallajokull in Southern Iceland, and Puyehue-Cordon Caulle (Puyehue), in Chile, started eruption on 4 June, 2011. Here we study instantaneous and localized temperature changes with these two eruptions following the method of Wang et al. (2009). Post-eruption negative temperature anomalies at the 250 hPa plane were clearly observed in the Eyjafjallajokull eruption. In the Puyehue eruption, however, such anomalies were not so clear due possibly to insufficient accuracy of the forecast model. By comparing the temperature profiles in the downwind areas of the two eruptions, we found that significant temperature decreases occurred at ~ 10.5 and ~ 11.5 km altitude in the Eyjafjallajokull and the Puyehue cases, respectively. These results generally agree with Wang et al. (2009), but post-eruption temperature changes seem to be influenced by additional factors, e.g. volcanic explosivity and local climatology.

Acoustic waves from volcanic eruptions are often observed as infrasound in near fields. Part of them propagate upward and disturb the ionosphere, and can be observed with Total Electron Content (TEC) data from Global Navigation Satellite System (GNSS) receivers. Here we report TEC variations after the 13 February 2014 Plinian eruption of the Kelud volcano, East Java, Indonesia, observed with regional GNSS networks. Significant disturbances in TEC were detected with six GNSS satellites, and wavelet analysis showed that harmonic oscillations started at ∼16:25 UT and continued for ∼2.5 h. The amplitude spectrum of the TEC time series showed peaks at 3.7 mHz, 4.8 mHz and 6.8 mHz. Long-wavelength standing waves with a wide range of wavelength trapped in the lower atmosphere are excited by the Plinian eruption. Amplitude spectra of the ground motion recorded by seismometers, however, had frequency components at discrete wave-periods. The condition for the resonant oscillations between the atmosphere and the solid Earth is satisfied only at these discrete wave-period and horizontal wavelength pairs, therefore efficient energy transfer from the atmospheric standing waves to the solid Earth Rayleigh waves occurred at discrete periods and resulted in the harmonic ground motion.

SUMMARY The Arctic Oscillation (AO) controls winter climates in the Northern Hemisphere to a large extent. Positive AO brings lower/higher surface temperature and higher/lower precipitation in high/middle latitude regions, and negative AO reverses the situation. In this study, we investigate signals of anomalous precipitations caused by AO using the data of the Gravity Recovery and Climate Experiment (GRACE) satellites. Wintertime mass deviations inferred from GRACE in the high and middle (boundary ∼55N) latitude regions in Eurasia showed highly positive and negative correlations with the AO indices. This possibly reflects the northward and southward shift of the centre of winter precipitation during the positive and negative phases of AO, respectively. Wintertime mass deviations also showed positive and negative correlations with the AO indices in the northern and southern parts of Greenland, respectively, but the boundary was further to the north, say ∼75N. AO redistributes the water mass as much as ∼1000 Gt between high and middle latitude regions in the Northern Hemisphere. Such mass redistribution causes significant surface deformation by loading large enough to be observed by Global Positioning System. This also causes the shift of the Earth's rotation axis especially towards the Greenwich Meridian large enough to be detected with space geodetic techniques. AO signatures are also derived from the empirical orthogonal function (EOF) analysis, as the first leading mode of GRACE data after excluding seasonal, linear and quadratic components. The EOF analysis also demonstrates that though AO is a main contributor to bring anomalous winter precipitation in the Northern Hemisphere as a whole, in North America, however, influence of the El Nino and Southern Oscillation on the winter precipitation anomaly is larger than AO.

Multipath in global positioning system (GPS) is the interference of the microwave signals directly from satellites and those reflected before reaching the antenna, typically by the ground. Because reflected signals cause positioning errors, GPS antennas are designed to reduce such interference. Recent studies show that multipath could be utilized to infer the properties of the ground around the antenna. Here, we report one such application, i.e. a fixed GPS station used as a snow depth meter. Because the satellite moves in the sky, the excess path length of reflected waves changes at rates dependent on the antenna height. This causes quasi-periodic variations of the amplitude and phase of the received signals. Accumulation of snow reduces effective antenna heights, and we can see it by analyzing multipath signatures. Signal-to-noise ratios (SNR) are often used to analyze multipath, but they are not always available in raw GPS data files. Here, we demonstrate that the geometry-free linear combination (L4), normally used to study the ionosphere, can also be used to analyze multipath signatures. We obtained snow depth time series at a GPS station in Hokkaido, Japan, from January to April in 2009 using L4 and SNR. Then, we compared their precisions. We also discuss mechanisms responsible for the possible underestimation of the snow depth by GPS. Finally, we investigate the possibility of inferring physical conditions of the snow surface using amplitudes of multipath signatures.

International audience; Using data of Japanese dense array of GPS receivers (GEONET) we analyzed variations of total electron content (TEC) over Japan during a period of low solar and geomagnetic activity in April-September 2008. During that time, five large earthquakes with magnitude M > 6.8 occurred around the Japanese Islands. We calculated vertical TEC with time resolution of 1 h and we compared the TEC over Japan with TEC over southwestern USA. In order to detect abnormal signals, we processed daily TEC based on median and quartile of preceding 15 days. Knowing that the decisive role in the ionosphere state is performed by space weather effects, we compared the estimated TEC values with time series of the 8, component of the interplanetary magnetic field (IMF), UV solar radiation, index of geomagnetic activity Dst and planetary index Kp. We have found that though the analyzed time period falls on the minimum of solar activity, the observed positive anomalies in the regional TEC correspond to fluctuations of IMF 13, and to the following increase in Kp and Dst indices. Other case of similar to 25% TEC increase in the duration of four days (DOY: 140-143) is, most likely, caused by increase in the solar UV radiation and coincides with growth of the global TEC. Other positive or negative abnormal TEC variations coincide with fluctuations of the global mean TEC or are of large spatial scale, i.e. much larger than the earthquake preparation zone. Therefore, it is rather difficult to find correlation between the observed anomalies and the five large earthquakes. Our results do not disprove the possibility of precursory phenomena but show the difficulties in identifying earthquake precursors in the ionosphere TEC on the background of TEC changes produced by space weather effects, even during solar minimum

Substantial amount of glacial ice is considered to be melting in the Asian high mountains. Gravimetry by GRACE satellite during 2003–2009 suggests the average ice loss rate in this region of 47 ± 12 Gigaton (Gt) yr − 1 , equivalent to ∼ 0.13 ± 0.04 mm yr − 1 sea level rise. This is twice as fast as the average rate over ∼ 40 years before the studied period, and agrees with the global tendency of accelerating glacial loss. Such ice loss rate varies both in time and space; mass loss in Himalaya is slightly decelerating while those in northwestern glaciers show clear acceleration. Uncertainty still remains in the groundwater decline in northern India, and proportion of almost isostatic (e.g. tectonic uplift) and non-isostatic (e.g. glacial isostatic adjustment) portions in the current uplift rate of the Tibetan Plateau. If gravity increase associated with ongoing glacial isostatic adjustment partially canceled the negative gravity trend, the corrected ice loss rate could reach 61 Gt yr − 1 .