Geology topics

Robert E. Anthony

Publications and source records attributed to Robert E. Anthony.

At least 37 records · Page 2Linked to original sources

USGS2022

Classifying Worldwide Standardized Seismograph Network records using a simple convolution neural network

The U.S. Geological Survey (USGS) maintains an archive of 189,180 digitized scans of analog seismic records from the World‐Wide Standardized Seismograph Network (WWSSN). Although these scans have been made public, the archive is too large to manually review, and few researchers have utilized large numbers of these records. To facilitate further research using this historical dataset, we develop a simple convolutional neural network (CNN) that rapidly (∼4.75 s/film chip) classifies scanned film chip images (called “chips,” because they are individually cut segments of 70 mm film) into four categories of “interestingness” to earthquake seismologists based on the presence of earthquakes and other seismic signals in the record: “no interest,” “little interest,” “interest,” and “high interest.” The CNN, dubbed “Seismic Analog Record Network” (SARNet), can identify four types of seismic traces (“no events,” “minor events,” “major events,” and “errors”) in 200 × 200 pixel subcrops with an accuracy of 92% using a confidence threshold of 85%. SARNet then converts 100 random subcrops from each film chip into the overall classification of interestingness. In this task, SARNet performed as well as expert human classifiers in determining the film chip’s overall interest grade. Applying SARNet to 34,000 film chips in the WWSSN archive found that 21% of the images were of “high interest” and had an “indeterminate” rate of only 4%. Thus, the need for the manual review of images was reduced by 79%. Sorting of film chips derived from SARNet will expedite further exploration of the archive of digitized analog seismic records stored at the USGS.

Seismological Research Letters↗

USGS2022

Characteristics, relationships and precision of direct acoustic-to-seismic coupling measurements from local explosions

Acoustic energy originating from explosions, sonic booms, bolides and thunderclaps have been recorded on seismometers since the 1950s. Direct pressure loading from the passing acoustic wave has been modelled and consistently observed to produce ground deformations of the near surface that have retrograde elliptical particle motions. In the past decade, increased deployments of colocated seismometers and infrasound sensors have driven efforts to use the transfer function between direct acoustic-to-seismic coupling to infer near-surface material properties including seismic velocity structure and elastic moduli. In this study, we use a small aperture (≈600 m) array of broadband seismometers installed in different manners and depths in both granite and sedimentary overburden to understand the fundamental nature and repeatability of seismic excitation from 1 to 15 Hz using horizontally propagating acoustic waves generated by 97 local (2–10 km) explosions. In agreement with modelling, we find that the ground motions induced by acoustic-to-seismic coupling attenuate rapidly with depth. We confirm the modelled relation between acoustic and ground motion amplitudes, but show that within one acoustic wavelength, the uncertainty in the transfer coefficient between seismic and acoustic energy at a given seismic station increases linearly with separation distance between the seismic and acoustic sensor. We attribute this observation to the rapid decorrelation of the infrasonic wavefield across small spatial scales and recommend colocating seismic and infrasound sensors for use in studies seeking to invert for near-surface material properties. Additionally, contrary to acoustic-to-seismic coupling theory and prior observations, we find that seismometers emplaced in granite do not record retrograde elliptical particle motions in response to direct pressure loading. We rule out seismometer tilt effects as a likely source of this observations and suggest that existing models of acoustic-to-seismic excitation may be too simplistic for seismometers placed in high rigidity materials.

Geophysical Journal International↗

USGS2022

Improved resolution across the Global Seismographic Network: A new era in low-frequency seismology

The Global Seismographic Network (GSN)—a global network of ≈150 very broadband stations—is used by researchers to study the free oscillations of the Earth (≈0.3–10 mHz) following large earthquakes. Normal‐mode observations can provide information about the radial density and anisotropic velocity structure of the Earth (including near the core–mantle boundary), but only when signal‐to‐noise ratios at very low frequencies are sufficiently high. Most normal‐mode observations in the past three decades have been made using Streckeisen STS‐1 vault seismometers. However, these sensors are no longer being manufactured or serviced. Candidate replacement sensors, the Streckeisen STS‐6 and the Nanometrics T‐360GSN, have been recently installed in boreholes, postholes, and vaults at several GSN stations and GSN testbeds. In this study, we examine normal‐mode spectra following three M w "> M w 8 earthquakes in 2021 and from one M w "> M w 8.2 earthquake in 2014 to evaluate the change in GSN low‐frequency performance on the vertical component. From this analysis, we conclude that the number of GSN stations capable of resolving normal modes following M w "> M w 8 earthquakes has nearly doubled since 2014. The improved observational capabilities will help better understand the radial velocity and density estimates of the Earth.

The Seismic Record↗

USGS2022

Seismic background noise levels across the continental United States from USArray Transportable Array: The influence of geology and geography

Since 2004, the most complete estimate of background noise levels across the continental U.S. was attained using 61 broadband seismic stations to calculate power spectral density (PSD) probability density functions. To improve seismic noise estimates across the U.S., we examine vertical component seismic data from the EarthScope USArray Transportable Array seismic network that rolled across the U.S. and southeastern Canada between 2004 and 2015 and form a large (10 TB) PSD database from 1679 stations that contains no smoothing or binning of the spectral estimates. Including station outages, our database has a mean of 98.9% data completeness, and we present maps showing the spatial and temporal variability of seismic noise in six bands of interest between 0.2- and 75-s period. At 0.2 s period, seismic noise across the eastern U.S. is predominantly anthropogenically generated and may be subsequently amplified more than 20 decibels in the sandy and water-saturated sediments of the southeastern U.S. Coastal Plain and Mississippi Embayment. In these sediments, 1 s noise shows similar amplification and is generated through a variety of mechanisms including cultural activity throughout Kentucky and the southeastern Appalachian Mountains, lake waves around the Great Lakes, and ocean waves throughout New England, the Pacific Northwest, and Florida. Both 0.2 and 1 s noise levels are the lowest in the Intermountain West portion of the U.S. We attribute this to a combination of installations on crystalline rocks and reduced population density. Finally, we find that sensors emplaced in sandy, water-saturated sediments observe median, diurnal variations in vertical component power at 18 to 75 s period, which we infer arise through local deformation driven by pressure variations. Ultimately, our results underscore that for shallow (<5 m depth) sensor installation, bedrock provides superior broadband noise performance compared to unconsolidated sediments.

Bulletin of the Seismological Society of America↗

USGS2022

Atmospheric waves and global seismoacoustic observations of the January 2022 Hunga eruption, Tonga

The 15 January 2022 climactic eruption of Hunga volcano, Tonga, produced an explosion in the atmosphere of a size that has not been documented in the modern geophysical record. The event generated a broad range of atmospheric waves observed globally by various ground-based and spaceborne instrumentation networks. Most prominent was the surface-guided Lamb wave (≲0.01 hertz), which we observed propagating for four (plus three antipodal) passages around Earth over 6 days. As measured by the Lamb wave amplitudes, the climactic Hunga explosion was comparable in size to that of the 1883 Krakatau eruption. The Hunga eruption produced remarkable globally detected infrasound (0.01 to 20 hertz), long-range (~10,000 kilometers) audible sound, and ionospheric perturbations. Seismometers worldwide recorded pure seismic and air-to-ground coupled waves. Air-to-sea coupling likely contributed to fast-arriving tsunamis. Here, we highlight exceptional observations of the atmospheric waves.

Hunga volcano↗

USGS2021

Local variations in broadband sensor installations: Orientations, sensitivities, and noise levels

As seismologists continue to place more stringent demands on data quality, accurately described metadata are becoming increasingly important. In order to better constrain the orientation and sensitivities of seismometers deployed in U.S. Geological Survey networks, the Albuquerque Seismological Laboratory (ASL) has recently begun identifying true north with a fiber optic gyroscope (FOG) and has developed methodologies to constrain mid-band, vertical component sensitivity levels to less than 1% in a controlled environment. However, questions remain regarding the accuracy of this new alignment technique as well as if instrument sensitivities and background noise levels are stable when the seismometers are installed in different environmental settings. In this study, we examine the stability and repeatability of these parameters by reinstalling two high-quality broadband seismometers (Streckeisen STS-2.5 and Nanometrics T-360 Global Seismographic Network (GSN) version) at different locations around the ASL and comparing them to each other and a reference STS-6 seismometer that stayed stationary for the duration of the experiment. We find that even in different environmental conditions, the sensitivities of the two broadband seismometers stayed stable to within 0.1% and that orientations attained using the FOG are generally accurate to within a degree. However, one install was off by 5° due to a mistake made by the installation team. These results indicate that while technology and methodologies are now in place to calibrate and orient a seismometer to within 1°, human error both during the installation and while producing the metadata is often a limiting factor. Finally, we find that background noise levels at short periods (0.1–1 s) become noisier when the sensors are emplaced in unconsolidated materials, whereas the noise levels at long periods (30–100 s) are not sensitive to local geological structure on the vertical components.

Pure and Applied Geophysics↗

USGS2021

Modeling seismic network detection thresholds using production picking algorithms

Estimating the detection threshold of a seismic network (the minimum magnitude earthquake that can be reliably located) is a critical part of network design and can drive network maintenance efforts. The ability of a station to detect an earthquake is often estimated by assuming the spectral amplitude for an earthquake of a given size, assuming an attenuation relationship, and comparing the predicted amplitude with the average station background noise level. This approach has significant uncertainty because of unknown regional attenuation and complications in computing small event power spectra, and it fails to account for the specific capabilities of the automatic seismic phase picker used in monitoring. We develop a data‐driven approach to determine network detection thresholds using a multiband phase picking algorithm that is currently in use at the U.S. Geological Survey National Earthquake Information Center. We apply this picking algorithm to cataloged earthquakes to determine an empirical relationship of the observability of earthquakes as a function of magnitude and distance. Using this relationship, we produce maps of detection threshold using station spatial configuration and station noise levels. We show that quiet, well‐sited stations significantly increase the detection capabilities of a network compared with a network composed of many noisy stations. Because our method is data driven, it has two distinct advantages: (1) it is less dependent on theoretical assumptions of source spectra and models of regional attenuation, and (2) it can easily be applied to any seismic network. This tool allows for an objective approach to the management of stations in regional seismic networks.

Seismological Research Letters↗

USGS2021

A review of timing accuracy across the Global Seismographic Network

The accuracy of timing across a seismic network is important for locating earthquakes as well as studies that use phase‐arrival information (e.g., tomography). The Global Seismographic Network (GSN) was designed with the goal of having reported timing be better than 10 ms. In this work, we provide a brief overview of how timing is kept across the GSN and discuss how clock‐quality metrics are embedded in Standard for Exchange of Earthquake Data records. Specifically, blockette 1001 contains the timing‐quality field, which can be used to identify time periods when poor clock quality could compromise timing accuracy. To verify the timing across the GSN, we compare cross‐correlation lags between collocated sensors from 1 January 2000 to 1 January 2020. We find that the mean error is less than 10 ms, with much of the difference likely coming from the method or uncertainty in the phase response of the instruments. This indicates that timing across the GSN is potentially better than 10 ms. We conclude that unless clock quality is compromised (as indicated in blockette 1001), GSN data’s timing accuracy should be suitable for most current seismological applications that require 10 ms accuracy. To assist users, the GSN network operators have implemented a “gsn_timing” metric available via the Incorporated Research Institutions for Seismology Data Management Center that helps users identify data with substandard timing accuracy (the 10 ms design goal of the GSN).

Seismological Research Letters↗

USGS2021

Six decades of seismology at South Pole, Antarctica: Current limitations and future opportunities to facilitate new geophysical observations

Seismograms from the South Pole have been important for seismological observations for over six decades by providing (until 2007) the only continuous seismic records from the interior of the Antarctic continent. The South Pole, Antarctica station has undergone many updates over the years, including conversion to a digital recording station as part of the Global Seismographic Network (GSN) in 1991 and being relocated to multiple deep ( ⁠ > 250    m "> > 250 m >250 m ⁠ ) boreholes 8 km away from the station in 2003 (and renamed to Quiet South Pole, Antarctica [QSPA]). Notably, QSPA is the second most used GSN station by the National Earthquake Information Center to pick phases used to rapidly detect and locate earthquakes globally, and has been used for a variety of glaciological and oceanography studies. In addition, it is the only seismic station on the Earth where low‐frequency ( ⁠ < 5    mHz "> < 5 mHz <5 mHz ⁠ ), normal‐mode oscillations of the planet excited by large earthquakes can be recorded without influence from Earth’s rotation, and most of the direct effects of the solid Earth tide vanish. However, the current sensors are largely 1980s vintage, and, while able to make some lower‐frequency observations from earthquakes, the borehole sensors appear unable to resolve ambient ground motions at frequencies lower than 25 mHz due to instrument noise and contamination from magnetic field variations. Recently developed borehole sensors offer the potential to extend background noise observations to below 3 mHz, which would substantially improve the fidelity and scientific value of seismic observations at South Pole. Through collaboration with the IceCube Neutrino Observatory, the opportunity exists to emplace a modern very broadband seismometer near the base ( ⁠ > 2    km "> > 2 km >2 km depth) of the Antarctic ice cap, which could lead to unprecedented seismic observations at long periods and facilitate a broad spectrum of Earth science studies.

Seismological Research Letters↗

USGS2021

Rayleigh wave amplitude uncertainty across the Global Seismographic Network and potential implications for global tomography

The Global Seismographic Network (GSN) is a multiuse, globally distributed seismic network used by seismologists, to both characterize earthquakes and study the Earth’s interior. Most stations in the network have two collocated broadband seismometers, which enable network operators to identify potential metadata and sensor issues. In this study, we investigate the accuracy with which surface waves can be measured across the GSN, by comparing waveforms of vertical‐component Rayleigh waves from M w "> M w Mw 6 and larger events between collocated sensor pairs. We calculate both the amplitude deviation and correlation coefficient between waveforms at sensor pairs. In total, we make measurements on over 670,000 event–station pairs from events that occurred from 1 January 2010 to 1 January 2020. We find that the average sensor‐pair amplitude deviation, and, therefore, GSN calibration level, is, approximately, 4% in the 25–250 s period band. Although, we find little difference in sensor‐pair amplitude deviations as a function of period across the entire network, the amount of useable data decreases rapidly as a function of increasing period. For instance, we determined that just over 12% of records at 250 s period provided useable recordings (e.g., sensor‐pair amplitude deviations of less than 20% and sensor‐pair correlation greater than 0.95). We then use these amplitude‐estimate deviations to identify how data coverage and quality could be limiting our ability to invert for whole Earth 3D attenuation models. We find an increase in the variance of our attenuation models with increasing period. For example, our degree 12 attenuation inversion at 250 s period shows 32% more variance than our degree 12 attenuation model at 25 s. This indicates that discrepancies of deep‐mantle tomography between studies could be the result of these large uncertainties. Because these high uncertainties arise from limited, high‐quality observations of long‐period ( ⁠ > 100    s ">

Bulletin of the Seismological Society of America↗

USGS2020

Towards understanding relationships between atmospheric pressure variations and long-period horizontal seismic data: A case study

Variations in atmospheric pressure have long been known to introduce noise in long-period (>10 s) seismic records. This noise can overwhelm signals of interest such as normal modes and surface waves. Generally, this noise is most pronounced on the horizontal components where it arises due to tilting of the seismometer in response to changes in atmospheric pressure. Several studies have suggested methodologies for correcting unwanted pressure-induced noise using collocated microbarograph records. However, how applicable these corrections are to varying geologic settings and installation types (e.g. vault versus post-hole) is unclear. Using coefficients obtained by solving for the residuals of these corrections, we can empirically determine the sensitivity of instruments in a specific location to the influences of pressure. To better understand how long-period, pressure-induced noise changes with time and emplacement, we examine horizontal seismic records along with barometric pressure at five different Global Seismographic Network stations, all with multiple broadband seismometers. We also analyse three Streckeisen STS-2 broadband seismometers, which are collocated with a microbarograph, at the Albuquerque Seismological Laboratory. We observe periods of high magnitude-squared-coherence (γ 2 -coherence; γ 2 > 0.8) between the seismic and pressure signals which fluctuate through time, frequency, and even between seismic instruments in the same vault. These observations suggest that these tilt-generated signals are highly sensitive to very local (<10 m) site effects. However, we find that in cases where instruments are not located at a large depth (<100 m), the pressure-induced noise is polarized in a nearly constant direction that is consistent with local topographic features or the geometry of the vault. We also find that borehole instruments at a large depth (>100 m) appear to be unaffected by pressure-loading mechanisms outlined by Sorrells (1971) but possibly by Newtonian attraction. Correlating the induced-noise polarization direction with times of high coherence, we work to identify sensors that are ultimately limited by pressure-induced horizontal noise as well as period bands that can benefit from pressure corrections. We find that while the situation is complex, each sensor appears to have its own unique response to pressure. Our findings suggest that we can determine empirical relationships between pressure and induced tilt on a case by case basis.

Geophysical Journal International↗

USGS2020

Magnetic field variations in Alaska: Recording space weather events on seismic stations in Alaska

Seismometers are highly sensitive instruments to not only ground motion but also many other nonseismic noise sources (e.g., temperature, pressure, and magnetic field variations). We show that the Alaska component of the Transportable Array is particularly susceptible to recording magnetic storms and other space weather events because the sensors used in this network are unshielded and magnetic flux variations are stronger at higher latitudes. We also show that vertical‐component seismic records across Alaska are directly recording magnetic field variations between 40 and 800 s period as opposed to actual ground motion during geomagnetic events with sensitivities ranging from 0.004 to 0.48    ( m / s 2 ) / T "> 0.48 ( m / s 2 ) / T 0.48 (m/s2)/T ⁠ . These sensitivities were found on a day where the root mean square variation in the magnetic field was 225 nT. Using a method developed by Forbriger (2007 , his section 3.1), we show that improving vertical seismic resolution of an unshielded sensor by as much as 10 dB in the 100–400 s period band using magnetic data from a collocated three‐component magnetometer is possible. However, due to large spatial variations in Earth’s magnetic field, this methodology becomes increasingly ineffective as the distance between the seismometer and magnetometer increases (no more than 200 km separation). A potential solution to this issue may be to incorporate relatively low‐cost magnetometers as an additional environmental data stream at high‐latitude seismic stations. We demonstrate that the Bartington Mag‐690 sensors currently deployed at Global Seismographic Network sites are not only acceptable for performing corrections to seismic data, but are also capable of recording many magnetic field signals with similar signal‐to‐noise ratios, in the 20–1000 s period band, as the observatory grade magnetometers operated by the U.S. Geological Survey Geomagnetism Program. This approach would densify magnetic field observations and could also contribute to space weather monitoring by supplementing highly calibrated magnetometers with additional sensors.

Alaska↗

USGS2020

Installation and performance of the Albuquerque Seismological Laboratory small-aperture posthole array

The Global Seismographic Network (GSN) has been used extensively by seismologists to characterize large earthquakes and image deep earth structure. Although the network’s original design goals have been met, the seismological community has suggested that the incorporation of small-aperture seismic arrays at select sites may improve performance of the network and enable new observations. As a pilot study for this concept, we have created a 500 m aperture, nine-element broadband seismic array around the GSN station ANMO (Albuquerque, New Mexico) at the U.S. Geological Survey Albuquerque Seismological Laboratory (ASL). The array was formed by supplementing the secondary borehole seismometer (90 m depth) at ANMO with eight additional 2.6 m posthole sites. Each station’s seismometer was oriented using a fiber optic gyroscope to within 2.0° of north. Data quality, particularly on the vertical components, is excellent with median power levels closely tracking the secondary sensor at ANMO at frequencies lower than 1 Hz. Horizontal component data are more variable at low frequencies (< 0:02 Hz), with the type of installation and local geography appearing to strongly influence the amount of tilt-induced noise. Throughout the article, we pose several fundamental questions related to the variability and precision of seismic wavefield measurements that we seek to address with data from this array. In addition, we calculate the array response and show a few examples of using the array to obtain back azimuths of a local event and a continuous narrowband noise source. The apparent velocity of the event across the array is then used to infer the local P-wave velocity at the ASL. Near real-time data collected from the array along with collocated meteorological, magnetic, and infrasound data are freely available in near-real time from the Incorporated Research Institutions for Seismology Data Management Center.

New Mexico↗

USGS2020

How processing methodologies can distort and bias power spectral density estimates of seismic background noise

Power spectral density (PSD) estimates are widely used in seismological studies to characterize background noise conditions, assess instrument performance, and study quasi‐stationary signals that are difficult to observe in the time domain. However, these studies often utilize different processing techniques, each of which can inherently bias the resulting PSD estimates. The level of smoothing, the size of the data window, and the method used for actually estimating the spectral content can all have strong influences on PSD estimates and background noise statistics. We show that although smoothing reduces the variance of the PSD estimate, the corresponding decrease in frequency resolution can eliminate or distort features of interest. For instance, popular software packages such as Incorporated Research Institutions for Seismology Modular Utility for STAatistical kNowledge Gathering (MUSTANG) and earlier versions of Portable Array Seismic Studies of the Continental Lithosphere Quick Look eXtended (PQLX), which were designed for data quality control and are effective in that regard, are less suitable for scientific studies that require accurate resolution of spectral peaks, even for peaks as broad as the primary ( ⁠ ∼ 14    s "> ∼ 14 s period) and secondary ( ⁠ ∼ 7    s "> ∼ 7 s period) microseisms. We also demonstrate how the 1 and 3 hr data windows used in MUSTANG and PQLX can be strongly influenced by energy generated from moderate‐size ( ⁠ M > ∼ 4.8 "> M > ∼ 4.8 ⁠ ) teleseismic earthquakes. The ubiquity of these events is likely skewing median ambient‐noise estimates by as much as 5 dB upward, for periods of 10–50 s at high‐quality broadband stations. Finally, we illustrate that many of the discrepancies between global low‐noise models are attributable to processing methodologies rather than fundamental differences in the underlying seismic data.

Seismological Research Letters↗

USGS2020

Global quieting of high-frequency seismic noise due to COVID-19 pandemic lockdown measures

Human activity causes vibrations that propagate into the ground as high-frequency seismic waves. Measures to mitigate the COVID-19 pandemic caused widespread changes in human activity, leading to a months-long reduction in seismic noise of up to 50%. The 2020 seismic noise quiet period is the longest and most prominent global anthropogenic seismic noise reduction on record. While the reduction is strongest at surface seismometers in populated areas, this seismic quiescence extends for many kilometers radially and hundreds of meters in depth. This provides an opportunity to detect subtle signals from subsurface seismic sources that would have been concealed in noisier times and to benchmark sources of anthropogenic noise. A strong correlation between seismic noise and independent measurements of human mobility suggests that seismology provides an absolute, real-time estimate of population dynamics.

Science↗

USGS2019

Improvements in seismic resolution and current limitations in the Global Seismographic Network

Station noise levels play a fundamental limitation in our ability to detect seismic signals. These noise levels are frequency-dependent and arise from a number of physically different drivers. At periods greater than 100 s, station noise levels are often limited by the self-noise of the instrument as well as the sensitivity of the instrument to non-seismic noise sources. Recently, station operators in the Global Seismographic Network (GSN) have deployed several Streckeisen STS-6A very broadband borehole seismometers. These sensors provide a potential replacement for the no-longer-produced Streckeisen STS-1 seismometer and the GeoTech KS-54000 borehole seismometer. Along with showing some of the initial observational improvements from installing modern very broadband seismometers at depth, we look at current limitations in the seismic resolution from Earth tide periods 100,000 s (0.01 mHz) to Nyquist at most GSN sites (0.02 s or 50 Hz). Finally, we show the potential for improved observations of continuously excited horizontal Earth hum as well as the splitting of very long-period torsional modes as a result of installing instruments at depth. Both of these observations make use of the low horizontal noise levels which are obtained by installing very broadband borehole seismometers at depth with noise levels similar to the Streckeisen STS-1.

Geophysical Journal International↗

USGS2019

Rapid station and network quality analysis for temporary deployments

Seismic station data quality is commonly defined by metrics such as data completeness or background seismic noise levels in specific frequency bands. However, for temporary networks such as aftershock deployments or induced seismicity monitoring, the most critical metric is often how well the station performs when recording events of interest. A timely measure of station performance can be used for real-time network maintenance and to help make decisions about which stations may need to be moved or are redundant. We develop new event-based methods to assess station and network performance, including estimating network magnitude of completeness, determining station signal-to-noise ratios as a function of earthquake magnitude, and computing relative station amplitudes. At times, a complete catalog of local seismic events may not exist, such as in an aftershock deployment where hundreds to thousands of small earthquakes may be happening and catalog generation efforts cannot keep up. To overcome this, we use an envelope of the average energy recorded by the network to identify events of interest. We find that the log amplitude of events identified using this technique scales linearly with local earthquake magnitudes. This suggests that this approach can be used to determine seismicity rates and detection thresholds.

Seismological Research Letters↗

USGS2019

Characteristics and spatial variability of wind noise on near-surface broadband seismometers

By coupling with the ground, wind causes ground motion that appears on seismic records as noise across a wide bandwidth. This wind-generated noise can drown out important features such as small earthquakes and prevent observation of normal modes from large earthquakes. Because the wind field is heterogeneous at local scales due to structures, diurnal heating, and topography, wind-induced seismic noise may be different on seismometers installed just meters apart. We have investigated the spatial variability of wind-induced noise using two weather sensors separated by approximately ~100 m and co-located with one deep borehole and four near-surface broadband seismometers. We found that at longer periods (>5 s), increasing wind speed causes increases in noise on the horizontal components of seismometers. While this has been previously observed, we also measured a γ2-coherence of less than 0.2 between the wind speed, wind direction, and the pressure recorded by our weather stations. We further observed a loss of coherence between the vertical components of our seismometers from 8 s to 20 s period. The amplitude of the drop-in coherence appears to depend on the substrate surrounding the seismometer. Based on two previously-developed theoretical models, we found that the local material surrounding the sensor could be amplifying the wind-generated noise. We also investigated the frequency dependence of wind-induced noise and found that the dominant source of high-frequency seismic noise at some sites could be anthropogenic rather than induced by wind. Additionally, we estimated the linear relationship between the root mean squares (RMS) of wind speed and RMS seismic velocity for all sensors, finding substantial variability between different installments. A more detailed understanding of the complex processes by which wind-induced noise is generated can inform the installation of sensors and the development of methods for mitigation of these effects, thus improving the overall quality of seismic data.

New Mexico↗