SUMMARY We use data from recently installed broad‐band seismographs on the islands of Crete, Gavdos, Santorini, Naxos and Samos in the Hellenic subduction zone to construct receiver function images of the crust and upper mantle from south of Crete into the Aegean Sea. The stations are equipped with STS‐2 seismometers and they are operated by GFZ Potsdam, University of Chania and ETH Zürich. Teleseismic earthquakes recorded by these stations at epicentral distances between 35° and 95° have been used to calculate receiver functions. The receiver function method is a routinely used tool to detect crustal and upper‐mantle discontinuities beneath a seismic station by isolating the P–S converted waves from the coda of the P wave. Converted P–S energy from the oceanic Moho of the subducted African Plate is clearly observed beneath Gavdos and Crete at a depth ranging from 44 to 69 km. This boundary continues to the north to nearly 100 km depth beneath Santorini island. Because of a lack of data the correlation of this phase is uncertain north of Santorini beneath the Aegean Sea. Moho depths were calculated from primary converted waves and multiply reflected waves between the Moho and the Earth's surface. Beneath southern and eastern Crete the Moho lies between 31 and 34 km depth. Beneath western and northern Crete the Moho is located at 32 and 39 km depth, respectively, and behaves as a reversed crust–mantle velocity contrast, possibly caused by hydration and serpentinization of the forearc mantle peridotite. The Moho beneath Gavdos island located south of Crete in the Libyan Sea is at 26 km depth, indicating that the crust south of the Crete microcontinent is also thinning towards the Mediterranean ridge. This makes it unlikely that part of the crust in Crete consists of accreted sediments transported there during the present‐day subduction process which began approximately 15 Ma because the backstop, i.e. the boundary between the current accretionary wedge of the Mediterranean ridge and the Crete microcontinent, is located approximately 100 km south of Gavdos. A seismic boundary at 32 km depth beneath Santorini island probably marks the crustal base of the Crete microcontinent. A shallower seismic interface beneath Santorini at 20–25 km depth may mark the depth of the detachment between the Crete microcontinent and the overlying Aegean subplate. The Moho in the central and northern Aegean, at Naxos and Samos, is observed at 25 and 28 km depth, respectively. Assuming a stretching factor of 1.2–1.3, crustal thickness in the Aegean was 30–35 km at the inception of the extensional regime in the Middle Miocene.

[1] We have found seismic evidence for the existence of narrow (about 200 km diameter) and hot plumes in the upper mantle above the 410 km discontinuity. They are located mainly under ocean islands, not however, under the central parts of major continents. These results are obtained from comparing globally the observed travel time differences between Ps conversions and SS precursors from the discontinuity at 410 km depth, using published data and new observations from ocean island stations. Comparison of SS precursors and Ps conversions, however, leads also to indications of significantly fewer narrow plumes in the upper mantle transition zone between 410 and 660 km depth, than in the upper mantle above 410 km depth.

No abstract

The high-rate kinematic Precise Point Positioning (PPP) of the Global Navigation Satellite System has become an effective method for monitoring crustal deformation caused by earthquakes. In this contribution, the method of GPS/GLONASS PPP with the receiver clock modeling is applied in active seismic deformation monitoring for the first time. With the modeling method, the short-term vertical positioning accuracy of 2-4 mm that usually cannot be obtained by standard PPP is achieved. Our PPP results confirm that the positioning accuracy is improved due to the increase of GLONASS observations compared to the GPS-only solution. Based on the external seismic data and the high-rate GPS/GLONASS data for the 2011 Japan earthquake and 2010 and 2015 Chile earthquakes, comparative analyses concerning receiver clock modeling are carried out. The results show that a high degree of decorrelation between the height position estimates and receiver clock offsets can be obtained by using the receiver clock modeling. The short-term accuracy of the GPS-based vertical displacements is improved to the level of about 4.4 mm, and the short-term accuracy of better than 4 mm for the GPS/GLONASS-combined vertical displacements is achievable. Furthermore, the weak vertical signals that are not detected by standard PPP can be captured with the modeling of highly stable receiver clock. Hirahara et al.

S U M M A R YThe debate concerning thermal plumes in the Earth's mantle, their geophysical detection and depth characterization remains contentious. Available geophysical, petrological and geochemical evidence is at variance regarding the very existence of mantle plumes. Utilizing P-to-S converted seismic waves (P receiver functions) from the 410 and 660 km discontinuities, we investigate disposition of these boundaries beneath a number of prominent hotspot regions. The thickness of the mantle transition zone (MTZ), measured as P660s−P410s differential times (t MTZ ), is determined. Our analyses suggest that the MTZ thickness beneath some hotspots correlates with the plume strength. The relationship between t MTZ , in response to the thermal perturbation, and the strength of plumes, as buoyancy flux B, follows a power law. This B-t MTZ behavior provides unprecedented insights into the relation of buoyancy flux and excess temperature at 410-660 km depth below hotspots. We find that the strongest hotspots, which are located in the Pacific, are indeed plumes originating at the MTZ or deeper. According to the detected power law, even the strongest plumes may not shrink the transition zone by significantly more than ∼40 km (corresponding to a maximum of 300-400 • excess temperature).