INSTANCE – the Italian seismic dataset for machine learning

The Italian earthquake waveform data are collected here in a dataset suited for machine learning analysis (ML) applications. The dataset consists of nearly 1.2 million three-component (3C) waveform traces from about 50 000 earthquakes and more than 130 000 noise 3C waveform traces, for a total of about 43 000 h of data and an average of 21 3C traces provided per event. The earthquake list is based on the Italian Seismic Bulletin (http://terremoti.ingv.it/bsi, last access: 15 February 2020​​​​​​​) of the Istituto Nazionale di Geofisica e Vulcanologia between January 2005 and January 2020, and it includes events in the magnitude range between 0.0 and 6.5. The waveform data have been recorded primarily by the Italian National Seismic Network (network code IV) and include both weak- (HH, EH channels) and strong-motion (HN channels) recordings. All the waveform traces have a length of 120 s, are sampled at 100 Hz, and are provided both in counts and ground motion physical units after deconvolution of the instrument transfer functions. The waveform dataset is accompanied by metadata consisting of more than 100 parameters providing comprehensive information on the earthquake source, the recording stations, the trace features, and other derived quantities. This rich set of metadata allows the users to target the data selection for their own purposes. Much of these metadata can be used as labels in ML analysis or for other studies. The dataset, assembled in HDF5 format, is available at http://doi.org/10.13127/instance (Michelini et al., 2021).

INSTANCE – the Italian seismic dataset for machine learning

The Italian earthquake waveform data are here collected in a dataset suited for machine learning analysis (ML) applications. The dataset consists of near 1.2 million three-component (3C) waveform traces from about 50,000 earthquakes and more than 130,000 noise 3C waveform traces, for a total of about 43,000 hours of data and an average of 21 3C traces are provided per event. The earthquake list is based on the Italian seismic bulletin (http://terremoti.ingv.it/bsi) of the ``Istituto Nazionale di Geofisica e Vulcanologia'' between January 2005 and January 2020 and it includes events in the magnitude range between 0.0 and 6.5. The waveform data have been recorded primarily by the Italian National Seismic Network (network code IV) and include both weak (HH, EH channels) and strong motion recordings (HN channels). All the waveform traces have a length of 120 s, are sampled at 100 Hz, and are provided both in counts and ground motion units after deconvolution of the instrument transfer functions. The waveform dataset is accompanied by metadata consisting of more than 100 parameters providing comprehensive information on the earthquake source, the recording stations, the trace features, and other derived quantities. This rich set of metadata allows the users to target the data selection for their own purposes. Many of these metadata can be used as labels in ML analysis or for other studies. The dataset, assembled in HDF5 format, is available at http://doi.org/10.13127/instance (Michelini et al., 2021).

Testing hypotheses of stress drop variations with hydraulic fracturing induced seismicity in the Horn River basin

<div> <p>Source parameters can help constrain the causes and mechanics of induced earthquakes. In particular, systematic variations of stress drops of fluid-injection induced seismicity have been interpreted in terms of the role of fluids, differences between tectonic and induced events, and self-similarity. The empirical basis for the variations, however, remains controversial. Here, we test three hypotheses about stress drops with observations of seismicity induced by hydraulic fracturing in the Horn River basin (Canada). First, stress drop is self-similar and independent of magnitude. Second, stress drop increases with distance from the point of fluid injection, which might be expected if in-situ effective stresses increase away from the point of fluid injection. Third, stress drops estimated with empirical Green’s functions (EGFs) are systematically larger than those estimated from direct fits to source models, which is expected if seismic waves attenuate in a frequency-dependent manner or experience site effects.</p> </div><div> <p>We probe the hypotheses with a large microseismic dataset collected during hydraulic fracturing operations in the Horn River shale gas play in British Columbia. 90,000+ seismic events were recorded by three borehole geophone arrays with a moment magnitude range of -3 < M<sub>w </sub>< 0.5. To calculate corner frequencies, we assume small, co-located seismic events can be approximated as EGFs, which effectively remove propagation and site effects from a larger target event. We target 34 M<sub>w</sub> > 0 events and search for EGFs over a 100 m radius for each event, choosing only those EGFs that satisfy multiple quality criteria. This study builds on previous work that estimated stress drops from direct fitting of standard Brune source models and found systematic high frequency resonances recorded by the geophones.</p> </div><div> <p>Of the 34 target events, we retrieve corner frequency and stress drop estimates for 22 events to test the three hypotheses. We observe that stress drop appears relatively constant over M<sub>w </sub>, but the magnitude range (0 < M<sub>w </sub>< 0.5) is currently too limited to draw strong conclusions. Second, stress drop appears to decrease, rather than increase, with distance from the point of injection (with a moderate Pearson’s correlation co-efficient of -0.5 ± 0.2); this could be caused by a direct hydraulic connection causing a reduction of in-situ effective normal stresses distal to the point of injection. Third, we observe no systematic difference between stress drops from direct source fits and EGF-based estimates, although stress drop uncertainties are large compared to standard earthquake source studies because of limited azimuthal coverage and high-frequency instrument resonances. These initial results do not support the systematic variations of stress drop for fluid-injection induced seismicity that have been observed in other datasets.</p> </div>

Predictability of cortico-cortical connections in the mammalian brain

Despite a five-order magnitude range in size, the mammalian brain exhibits many shared anatomical and functional characteristics that should translate into cortical network commonalities. Here we develop a framework employing machine learning to quantify the degree of predictability of the weighted interareal cortical matrix. Data were obtained with retrograde tract-tracing experiments supplemented by projection length measurements. Using this framework with consistent and edge-complete empirical datasets in the macaque and mouse cortex, we show that there is significant amount of predictability embedded in the interareal cortical networks of both species. At the binary level, links are predictable with an Area Under the ROC curve of at least 0.8 for the macaque. At the weighted level, strengths of the medium and strong links are predictable with at least 85-90% accuracy in mouse and 70-80% in macaque, whereas weak links are not predictable in either species. These observations suggest that the formation and evolution of the cortical network at the mesoscale is to a large extent, rule-based, motivating further research on the architectural invariants of the cortical connectome.

The critical role of pore size on depth-dependent microbial cell counts in sediments

Cell counts decrease with sediment depth. Typical explanations consider limiting factors such as water availability and chemistry, carbon source, nutrients, energy and temperature, and overlook the role of pore size. Our analyses consider sediment self-compaction, the evolution of pore size with depth, and the probability of pores larger than the microbial size to compute the volume fraction of life-compatible pores. We evaluate cell counts vs. depth profiles gathered at 116 sites worldwide. Results confirm the critical role of pore size on cell counts in the subsurface and explain much of the data spread (from ~ 9 orders of magnitude range in cell counts to ~ 2 orders). Cells colonize pores often forming dense biofilms, thus, cell counts in pores are orders of magnitude higher than in the water column. Similar arguments apply to rocks.

The Northern Extragalactic WISE × Pan-STARRS (NEWS) catalogue

This study involves two photometric catalogues, AllWISE and Pan-STARRS Data Release 1, which were cross-matched to identify extragalactic objects among the common sources of these catalogues. To separate galaxies and quasars from stars, we created a machine-learning model that is trained on photometric (in fact, colour-based) information from the optical and infrared wavelength ranges. The model is based on three important procedures: the construction of the autoencoder artificial neural network, separation of galaxies and quasars from stars with a support vector machine (SVM) classifier, and cleaning of the AllWISE × PS1 sample to remove sources with abnormal colour indices using a one-class SVM. As a training sample, we employed a set of spectroscopically confirmed sources from the Sloan Digital Sky Survey Data Release 14. Having applied the classification model to the data of crossing the AllWISE and Pan-STARRS DR1 samples, we created the Northern Extragalactic WISE × Pan-STARRS (NEWS) catalogue, containing 40 million extragalactic objects and covering 3/4 of celestial sphere up to g = 23m. Several independent classification quality tests, namely, the astrometric test along with others based on the use of data from spectroscopic surveys show similar results and indicate a high purity (∼98.0%) and completeness (> 98%) for the NEWS catalogue within the magnitude range of 19.0m < g < 22.5m. The classification quality still retains quite acceptable levels of 70% for purity and 97% for completeness for the brightest and faintest objects from this magnitude range. In addition, validation with external data sets has demonstrated the need for using only those sources in the NEWS catalogue that are outside the zone with the enhanced extinction. We show that the number of quasars from the NEWS catalogue identified in Gaia DR2 exceeds the number of quasars previously identified in Gaia DR2 with the use of the AllWISEAGN catalogue. These quasars may be used in future as an additional sample for testing and anchoring the Gaia Celestial Reference Frame.

Catalog of Preinstrumental Earthquakes in Central Mexico: Epicentral and Magnitude Estimations Based on Macroseismic Data

ABSTRACT We present the first parametric catalog of historical earthquakes in Mexico from 1469 to 1912 composed of 323 historical earthquakes. The historical earthquakes were assigned to specific seismotectonic provinces, and attenuation relations of seismic intensity versus distance were calculated using instrumental earthquakes. The intensity data were inverted using a linear regression for the best-fitting magnitude and source location. From the 323 events identified in the historical record, magnitude and source location were determined for 40 earthquakes from 1568 to 1912. The historical subduction earthquakes are distributed uniformly along the coast. There is, however, a conspicuous absence of subduction earthquakes where the great 1985 Michoacán earthquakes took place. The data also show a large number of earthquakes Mw&gt;7 in the presumed Guerrero gap in the past 320 yr. The source parameters of in-slab earthquakes were obtained for 10 earthquakes that took place in the nineteenth and early twentieth centuries. The analysis of completeness of the historical and instrumental International Seismological Centre-Global Earthquake Model catalogs of subduction earthquakes Mw&gt;7.0 show similar values of the slope of the Gutenberg–Richter relation between 1.62 and 1.95. The large b-values appear to reflect the apparently anomalous large number of earthquakes in the magnitude range Mw 7.4–7.7 and an absence of events Mw∼7. This magnitude distribution suggests that the seismicity in the Mexican subduction zone is dominated by characteristic earthquakes in the magnitude range Mw 7.4–7.7, with larger earthquakes Mw&gt;8 showing longer recurrence times. The catalog of historical subduction earthquakes appears to be complete for Mw&gt;7.5. The catalog of crustal earthquakes in the Trans-Mexican volcanic belt is complete since 1568 for events Mw&gt;6.4. Completeness of the catalog of in-slab earthquakes was not estimated due to the short record for this type of event.

Numerical Analysis of Innovative Wind Booster Configurations for Vertical Axis Wind Turbines

Vertical Axis Wind Turbines has been looked upon by researchers as a potential avenue for power generation at domestic level. The Coefficient of Power of Vertical Axis wind Turbine has its limitations mentioned by Betz theory. In this paper three configurations viz one stage, two stage and three stage Wind booster has been designed, modelled and numerically analyzed. A Multistage concept of flaps has been used in wind booster so that wind velocity may increase during each stage. The performance of turbine is assessed by wind booster for Flow velocity, Pressure. For Numerical analysis SST K-epsilon turbulence model has been used. From the analysis it is found that, the two stage wind booster velocity and pressure magnitude range is superior as compared to other two configurations.

A Highly Sensitive and Selective Isobutyraldehyde Sensor Based on Nanosized Sm2O3 Particles

A highly sensitive and selective sensor for isobutyraldehyde (IBD) is demonstrated based on intensive cataluminescence (CTL) emission from the surface of nanosized Sm2O3 particles. The characteristics and optimum conditions for the CTL sensor, including the working temperature, wavelength, and flow rate, were investigated in detail. Under the optimized experimental conditions, the CTL intensity varied linearly with the concentration of IBD, in the two-order-of-magnitude range of 0.015–3.9 μg/mL, with a correlation coefficient (r) of 0.99991 and a limit of detection (LOD), at a signal-to-noise ratio (S/N = 3) of 4.6 ng/mL. The sensor was quite specific: butyraldehyde, methanol, ethanol, acetone, formaldehyde, acetaldehyde, benzene, ethylbenzene, and cumene could not produce significant CTL intensities; specifically, butyraldehyde, ethanol, acetone, and acetaldehyde produced low CTL intensities, with values that were 3.8%, 2.8%, 0.60%, and 0.57% that of IBD. As a test of sensor stability, we found that the relative standard deviation (RSD) of 30 measurements of the CTL at an IBD concentration of 1.6 μg/mL within a period of 72 h was 2.2%, indicating good stability and long service life of the sensor. The sensor was tested against spiked samples containing IBD, and recoveries between 89.7% and 97.4% were obtained with an RSD of 6.1%–8.6%. The performance of the sensor indicated its utility for practical sample analysis.