Urban seismic monitoring in Brasília, Brazil

Urban seismology has gained scientific interest with the development of seismic ambient noise monitoring techniques and also for being a useful tool to connect society with the Earth sciences. The interpretation of the sources of seismic records generated by sporting events, traffic, or huge agglomerations arouses the population’s curiosity and opens up a range of possibilities for new applications of seismology, especially in the area of urban monitoring. In this contribution, we present the analysis of seismic records from a station in the city of Brasilia during unusual episodes of silencing and noisy periods. Usually, cultural noise is observed in high-fequency bands. We showed in our analysis that cultural noise can also be observed in the low-frequency band, when high-frequency signal is attenuated. As examples of noisy periods, we have that of the Soccer World Cup in Brazil in 2014, where changes in noise are related to celebrations of goals and the party held by FIFA in the city, and the political manifestations in the period of the Impeachment trial in 2016, which reached the concentration of about 300,000 protesters. The two most characteristic periods of seismic silence have been the quarantine due to the COVID-19 pandemic in 2020, and the trucker strike that occurred across the country in 2018, both drastically reducing the movement of people in the city.

Assessment of a claimed ultra-low frequency electromagnetic (ULFEM) earthquake precursor

Anomalous ultra-low frequency electromagnetic (ULFEM) pulses occurring before the M5.4 2007 and M4.0 2010 Alum Rock earthquakes have been claimed to increase in number days to weeks prior to each earthquake. We re-examine the previously reported ultra-low frequency (ULF: 0.01-10 Hz) magnetic data recorded at a QuakeFinder site located 9 km from the earthquake hypocenter, as well as data from a nearby Stanford-USGS site located 42 km from the hypocenter, to analyze the characteristics of the pulses and assess their origin. Using pulse definitions and pulse-counting algorithms analogous to those previously reported, we corroborate the increase in pulse counts before the 2007 Alum Rock earthquake at the QuakeFinder station, but we note that the number of pulses depends greatly on chosen temporal and amplitude detection thresholds. These thresholds are necessarily arbitrary because we lack a clear physical model or basis for their selection. We do not see the same increase in pulse counts before the 2010 Alum Rock earthquake at the QuakeFinder or Stanford-USGS station. In addition, when comparing specific pulses in the QuakeFinder data and Stanford-USGS data, we find that the majority of pulses do not match temporally, indicating the pulses are not from solar-driven ionospheric/magnetospheric disturbances or from atmospheric lightning, and lack a common origin. Notably, however, our assessment of the temporal distribution of pulse counts throughout the day shows pulse counts increase during peak human activity hours, strongly suggesting these pulses result from local cultural noise and are not tectonic in origin. The many unknowns about the character and even existence of precursory earthquake pulses means that otherwise standard numerical and statistical test cannot be applied. Yet here we show that exhaustive investigation of many different aspects of ULFEM signals can be used to properly characterize their origin.

Effect of coronavirus lockdowns on the ambient seismic noise levels in Gujarat, northwest India

The Covid-19 pandemic created havoc and forced lockdowns in almost all the countries worldwide, to inhibit social spreading. In India as well, as a precautionary measure, complete and partial lockdowns were announced in phases during March 25 to May 31, 2020. The restricted human activities led to a drastic reduction in seismic background noise in the high frequency range of 1–20 Hz, representative of cultural noise. In this study, we analyse the effect of anthropogenic activity on the Earth vibrations, utilizing ambient noise recorded at twelve broadband seismographs installed in different environmental and geological conditions in Gujarat. We find that the lockdowns caused 1–19 dB decrease in seismic noise levels. The impact of restricted anthropogenic activities is predominantly visible during the daytime in urban areas, in the vicinity of industries and/or highways. A 27–79% reduction in seismic noise ground displacement (drms) is observed in daytime during the lockdown, in populated areas. However, data from station MOR reveals a drastic decrease in drms amplitude both during the day (79%) and night times (87%) since factories in this area operate round the clock. The noise at stations located in remote areas and that due to microseisms, shows negligible variation.

Humming trains as an opportune source for imaging the shallow crust

<div><span>Freight trains are one of the most powerful and persistent seismic sources of cultural noise. They generate tremors equivalent to earthquakes of magnitude 1 that can be detectable up to 100 km distance.  Here, we propose to use the freight train passages as an opportunistic source of noise for passive seismic interferometry (SI). Usually, passive SI relies on blind correlations of long time series of noise for imaging and monitoring purposes. We suggest an alternative method based on noise source characterization, signal and station pairs selection, and specific seismic phase extraction (surface and body waves) for each virtual source to imaging the subsurface. To illustrate our novel method's potential, we show a case study in Canada's mineral exploration context, where we use retrieved body waves to estimate travel time tomography. This noise recovery approach to create valuable sources could be applied for several seismic noise sources and in different contexts improving spatial and temporal resolutions.</span></div>

Method for obtaining response functions from noisy magnetotelluric data using frequency-domain independent component analysis

The magnetotelluric (MT) method has been used for visualizing subsurface resistivity structures and more recently for monitoring resistivity changes. However, electromagnetic data often include cultural noise, which can cause errors in the estimation of MT response functions and subsurface resistivity structure analysis. Frequency-domain independent component analysis (FDICA) offers advantages for MT data processing particularly because this method can extract hidden components in the observed data. These components can be decomposed into natural MT signals and cultural noise so that the noise effect in the recovered MT data is reduced. FDICA is applied to MT data acquired at the Kakioka Magnetic Observatory in Japan. The apparent resistivity and phase curves are obtained with small estimated errors between periods of 7 and 12,000 s, although the length of the time-series data is limited. The curves are smoother than those obtained using a conventional method. Various types of synthetic noise are added to the time series at Kakioka to test the noise-reduction performance of FDICA for MT data with high noise contamination. The results demonstrate that FDICA can be used to estimate MT response functions with high accuracy even under conditions in which more than half of the time-series data are contaminated by noise.

Deep Clustering to Identify Sources of Urban Seismic Noise in Long Beach, California

Ambient seismic noise consists of emergent and impulsive signals generated by natural and anthropogenic sources. Developing techniques to identify specific cultural noise signals will benefit studies performing seismic imaging from continuous records. We examine spectrograms of urban cultural noise from a spatially dense seismic array located in Long Beach, California. The spectral features of the waveforms are used to develop a self-supervised clustering model for differentiating cultural noise into separable types of signals. We use 161 hr of seismic data from 5200 geophones that contain impulsive signals originating from human activity. The model uses convolutional autoencoders, a self-supervised machine-learning technique, to learn latent features from spectrograms produced from the data. The latent features are evaluated using a deep clustering algorithm to separate the noise signals into different classes. We evaluate the separation of data and analyze the classes to identify the likely sources of the signals present in the data. To interpret the model performance, we examine the time–frequency domain features of the signals and the spatiotemporal evolution observed for each class. We demonstrate that clustering using deep autoencoders is a useful approach to characterizing seismic noise and identifying novel signals in the data.

easyQuake: Putting Machine Learning to Work for Your Regional Seismic Network or Local Earthquake Study

We developed a Python package—easyQuake—that consists of a flexible set of tools for detecting and locating earthquakes from International Federation of Digital Seismograph Networks-collected or field-collected seismograms. The package leverages a machine-learning driven phase picker, coupled with an associator, to produce a Quake Markup Language (QuakeML) style catalog complete with magnitudes and P-wave polarity determinations. We describe how nightly computations on day-long seismograms identify lower-magnitude candidate events that were otherwise missed due to cultural noise and how those events are incorporated into the Oklahoma Geological Survey statewide network upon analyst manual review. We discuss applications for the package, including earthquake detection for regional networks and microseismicity studies in arbitrary user-defined regions. Because the fundamentals of the package are scale invariant, it has wide application to seismological earthquake analysis from regional to local arrays and has great potential for identifying early aftershocks that are otherwise missed. The package is fast and reliable; the computations are relatively efficient across a range of hardware, and we have encountered very few (∼1%) false positive event detections for the Oklahoma case study. The utility and novelty of the package is the turnkey earthquake analysis with QuakeML file output, which can be dropped directly into existing real-time earthquake analysis systems. We have designed the functions to be quite modular so that a user could replace the provided picker or associator with one of their choosing. The Python package is open source and development continues.

COVID-19 Societal Response Captured by Seismic Noise in China and Italy

Seismic noise with frequencies above 1 Hz is often called “cultural noise” and is generally correlated quite well with human activities. Recently, cities in mainland China and Italy imposed restrictions on travel and day-to-day activity in response to COVID-19, which gave us an unprecedented opportunity to study the relationship between seismic noise above 1 Hz and human activities. Using seismic records from stations in China and Italy, we show that seismic noise above 1 Hz was primarily generated by the local transportation systems. The lockdown of the cities and the imposition of travel restrictions led to an ∼4–12 dB decrease in seismic noise power in mainland China. Data also show that different Chinese cities experienced distinct periods of diminished cultural noise, related to differences in local response to the epidemic. In contrast, there was only ∼1–6 dB decrease of seismic noise power in Italy, after the country was put under a lockdown. The noise data indicate that traffic flow did not decrease as much in Italy and show how different cities reacted distinctly to the lockdown conditions.