ALBERT

Description: Characterizing the seismicity of Novaya Zemlya and the surrounding Arctic seas requires accurate event-location estimates. Low-magnitude events in this region are currently observed only by a small number of stations in the European Arctic, with a large azimuthal gap, making the accuracy of regional velocity models all the more important. Regional travel-time calibration is difficult given the scarcity of sufficiently well-constrained events. On 11 October 2010, a magnitude 4.5 event occurred close to the northern tip of Novaya Zemlya. This event is significant in that it is the first event in this region to have been recorded both on the relatively recent regional networks and arrays, and also teleseismically with good azimuthal coverage. We examine how well we can constrain the location and origin time using only teleseismic phases. Using only first teleseismic P arrivals, we constrain the epicenter to approximately 76.25° N and 64.75° E but with no depth resolution. Clear depth phases, notably on stations in the southern United States, indicate a depth between 9 and 15 km. This independent hypocenter and origin time estimate allow evaluation of regional phase travel-time prediction using different models. The predicted Sn travel time appears to cause the greatest variability in regional location estimates. The 3D Regional Seismic Travel Times models provide excellent Pn travel-time estimates for Barents Sea paths but may slightly overestimate Sn travel times from this source region. A modified regional 1D velocity model is defined, which best predicts Pn and Sn observations at multiple stations up to 15°. The significance of the regional travel-time models for estimating location is demonstrated for a low-magnitude event on or close to the northern island of Novaya Zemlya in March 2014, recorded with a satisfactory signal-to-noise ratio at only four stations.

Description: Aftershock sequences following very large earthquakes present enormous challenges to near-real-time generation of seismic bulletins. The increase in analyst resources needed to relocate an inflated number of events is compounded by failures of phase-association algorithms and a significant deterioration in the quality of underlying, fully automatic event bulletins. Current processing pipelines were designed a generation ago, and, due to computational limitations of the time, are usually limited to single passes over the raw data. With current processing capability, multiple passes over the data are feasible. Processing the raw data at each station currently generates parametric data streams that are then scanned by a phase-association algorithm to form event hypotheses. We consider the scenario in which a large earthquake has occurred and propose to define a region of likely aftershock activity in which events are detected and accurately located, using a separate specially targeted semiautomatic process. This effort may focus on so-called pattern detectors, but here we demonstrate a more general grid-search algorithm that may cover wider source regions without requiring waveform similarity. Given many well-located aftershocks within our source region, we may remove all associated phases from the original detection lists prior to a new iteration of the phase-association algorithm. We provide a proof-of-concept example for the 2015 Gorkha sequence, Nepal, recorded on seismic arrays of the International Monitoring System. Even with very conservative conditions for defining event hypotheses within the aftershock source region, we can automatically remove about half of the original detections that could have been generated by Nepal earthquakes and reduce the likelihood of false associations and spurious event hypotheses. Further reductions in the number of detections in the parametric data streams are likely, using correlation and subspace detectors and/or empirical matched field processing.

Description: The detectability of low magnitude seismic events in the European Arctic is determined primarily by the small-aperture International Monitoring System arrays ARCES and SPITS. In August 2004, the SPITS array was upgraded to a broadband array with an increase in the sampling rate from 40 to 80 Hz. Most important, however, for the detection and location of small-magnitude seismic events was the deployment of three-component instruments at six of the nine sites. Detection and correct classification of secondary phases are of paramount importance for events observed by only a small number of stations at regional distances; and, in the absence of the strong Lg phases typically observed for continental propagation paths, multiple three-component stations were deemed necessary to exploit the higher S-phase amplitudes anticipated on the horizontal sensors. We demonstrate improved signal-to-noise ratios (SNRs) for S phases on horizontal beams for several events close to Novaya Zemlya. Horizontal component f-k analysis improves direction estimates and phase classification for low-SNR signals. We demonstrate secondary phases that are misidentified by vertical-only f-k analysis but which are correctly classified by three-component array processing. A significant problem with array processing at SPITS is the overlap in slowness space of regional P and S phases. Phase identification is improved greatly by comparing the coherence between vertical traces with the coherence between horizontal traces. Considerations in the routine array processing of SPITS data are reviewed, including the need for elevation corrections in slowness estimation and the need to take into account azimuth-dependent variation of apparent velocity estimates for regional phases.

Description: Microseisms in the period of 2–10 s are generated in deep oceans and near coastal regions. It is common for microseisms from multiple sources to arrive at the same time at a given seismometer. It is therefore desirable to be able to measure multiple slowness vectors accurately. Popular ways to estimate the direction of arrival of ocean induced microseisms are the conventional (fk) or adaptive (Capon) beamformer. These techniques give robust estimates, but are limited in their resolution capabilities and hence do not always detect all arrivals. One of the limiting factors in determining direction of arrival with seismic arrays is the array response, which can strongly influence the estimation of weaker sources. In this work, we aim to improve the resolution for weaker sources and evaluate the performance of two deconvolution algorithms, Richardson–Lucy deconvolution and a new implementation of CLEAN-PSF. The algorithms are tested with three arrays of different aperture (ASAR, WRA and NORSAR) using 1 month of real data each and compared with the conventional approaches. We find an improvement over conventional methods from both algorithms and the best performance with CLEAN-PSF. We then extend the CLEAN-PSF framework to three components (3C) and evaluate 1 yr of data from the Pilbara Seismic Array in northwest Australia. The 3C CLEAN-PSF analysis is capable in resolving a previously undetected Sn phase.

Description: The frequency–wavenumber (fk) and Capon methods are widely used in seismic array studies of background or ambient noise to infer the backazimuth and slowness of microseismic sources. We present an implementation of these techniques for the analysis of microseisms (0.05–2 Hz) which draws on array signal processing literature from a range of disciplines. The presented techniques avoid frequency mixing in the cross-power spectral density and therefore yield an accurate slowness vector estimation of the incoming seismic waves. Using synthetic data, we show explicitly how the frequency averaged broad-band approach can result in a slowness-shifted spectrum. The presented implementation performs the slowness estimations individually for each frequency bin and sums the resulting slowness spectra over a specific frequency range. This may be termed an incoherently averaged signal, or IAS, approach. We further modify the method through diagonal loading to ensure a robust solution. The synthetic data show good agreement between the analytically derived and inferred error in slowness. Results for real (observed) data are compared between the approximate and IAS methods for two different seismic arrays. The IAS method results in the improved resolution of features, particularly for the Capon spectrum, and enables, for instance, Rg and Lg arrivals from similar backazimuths to be separated in the case of real data.

Description: Declared North Korean nuclear tests in 2006, 2009, 2013 and 2016 were observed seismically at regional and teleseismic distances. Waveform similarity allows the events to be located relatively with far greater accuracy than the absolute locations can be determined from seismic data alone. There is now significant redundancy in the data given the large number of regional and teleseismic stations that have recorded multiple events, and relative location estimates can be confirmed independently by performing calculations on many mutually exclusive sets of measurements. Using a 1-D global velocity model, the distances between the events estimated using teleseismic P phases are found to be approximately 25 per cent shorter than the distances between events estimated using regional Pn phases. The 2009, 2013 and 2016 events all take place within 1 km of each other and the discrepancy between the regional and teleseismic relative location estimates is no more than about 150 m. The discrepancy is much more significant when estimating the location of the more distant 2006 event relative to the later explosions with regional and teleseismic estimates varying by many hundreds of metres. The relative location of the 2006 event is challenging given the smaller number of observing stations, the lower signal-to-noise ratio and significant waveform dissimilarity at some regional stations. The 2006 event is however highly significant in constraining the absolute locations in the terrain at the Punggye-ri test-site in relation to observed surface infrastructure. For each seismic arrival used to estimate the relative locations, we define a slowness scaling factor which multiplies the gradient of seismic traveltime versus distance, evaluated at the source, relative to the applied 1-D velocity model. A procedure for estimating correction terms which reduce the double-difference time residual vector norms is presented together with a discussion of the associated uncertainty. The modified velocity gradients reduce the residuals, the relative location uncertainties and the sensitivity to the combination of stations used. The traveltime gradients appear to be overestimated for the regional phases, and teleseismic relative location estimates are likely to be more accurate despite an apparent lower precision. Calibrations for regional phases are essential given that smaller magnitude events are likely not to be recorded teleseismically. We discuss the implications for the absolute event locations. Placing the 2006 event under a local maximum of overburden at 41.293°N, 129.105°E would imply a location of 41.299°N, 129.075°E for the January 2016 event, providing almost optimal overburden for the later four events.