Uncertainty In Source Location Research Articles

Identifying the source settings of deep brine leakage from CO2 geological repositories using observations from shallow overlying formations

Shallow groundwater resources overlaying deep saline formations used in carbon storage applications are subjected to a potential contamination threat by CO2/brine leakage via natural or anthropogenically-induced conductive pathways in the confining caprock. Identifying the leakage source location and rate is critical for developing remediation plans and designing corrective actions. Owing to limited information about the flow and transport characteristics of deep regimes and high cost of obtaining data on their response to CO2 injection operation, estimating accurate source settings (i.e., location and rate) can be extremely challenging. Under such conditions, Bayesian inverse frameworks become useful tools to help identify potential leakage patterns. This study tests and validates an ensemble-based data-assimilation approach that reduces the uncertainty in the prior knowledge about source settings through conditioning forward transport models using relatively inexpensive easy-to-acquire shallow zone data. The approach incorporates the newly developed ensemble smoother tool in the inversion code “PEST++” with the transport code “FEFLOW” to perform history matching and uncertainty analysis. A novel parameterization method that allows the disposition of potential source was used to search the leakage location during calibration process. In the absence of field data, the approach was validated using experimental data generated in ∼8 m long soil tank simulating leakage from storage zone migrating to the shallow aquifer. The results show that source location uncertainty can be reasonably reduced using shallow zone data collected from near-surface aquifers. However, more prior information about the system and deeper data are essential to estimate a practical probability range for the leakage rate.

Probabilistic seismic source location and magnitude via inverse analysis of paleoliquefaction evidence

In regions of infrequent moderate-to-large earthquakes, historic earthquake catalogs are often insufficient to provide inputs to seismic-hazard analyses (i.e. fault locations and magnitude–frequency relations) or to inform ground-motion predictions for certain seismic sources. In these regions, analysis of relic coseismic evidence, such as paleoliquefaction, is commonly used to infer information about the seismic hazard. However, while paleoliquefaction studies have been performed widely, all existing analysis techniques require a priori assumptions about the causative earthquake’s location (i.e. rupture magnitude and ground motions can otherwise not be estimated). This may lead to inaccurate assumptions in some settings, and by corollary, erroneous results. Accordingly, this article proposes an inversion framework to probabilistically constrain seismic-source parameters from paleoliquefaction. Analyzing evidence at regional scale leads to (a) a geospatial likelihood surface that constrains the rupture location and (b) a probability distribution of the rupture magnitude, wherein source-location uncertainty is explicitly considered. Simulated paleoliquefaction studies are performed on earthquakes with known parameters. These examples demonstrate the framework’s potential, even in cases of limited field evidence, as well as important caveats and lessons for forward use. The proposed framework has the potential to provide new insights in enigmatic seismic zones worldwide.

Residual physical modeling and semi-supervised source localization

Source localization in reverberant environments remains an open challenge which ML techniques have shown promise in addressing. Real-world acoustic environments contain irregular boundaries, scattering, and diffracting elements (e.g., furniture and uneven surfaces) which are impractical to model using acoustic propagation software. We develop a hybrid approach to acoustic source localization which utilizes both analytical, signal processing-based localization, and machine learning (ML) to provide a data-driven source localization model. In this approach, the output of an analytical source localization model (e.g., SRP-PHAT or MUSIC) is augmented by a neural network (NN) system. This augmented system consists of a variational autoencoder and localization network, which is designed to correct the analytic estimate. The networks are trained via semi-supervised learning, on both labeled and unlabeled inputs. The NNs in this approach parameterize stochastic functions, which facilitate a statistically principled learning approach via variational inference (VI). The stochasticity also helps quantify source location uncertainty under the VI approximate posterior. For simulated and real acoustic data, the hybrid approach generalizes better and is more efficient than ML or analytic approaches alone.

Active Environmental Monitoring and Anomaly Search System for Space Habitat With Markov Decision Process and Active Sensing

For future crewed missions that could last years with limited ground support, the environmental control and life support system (ECLSS) will likely evolve to meet new, more stringent reliability and autonomy requirements. In this work, we focus on improving the performance of the environmental monitoring and anomaly detection systems using Markov decision process and active sensing. We exploit actively moving sensors to develop a novel sensing architecture and supporting analytics, termed Active environmental Monitoring and Anomaly Search System (AMASS). We design a Dynamic Value Iteration policy to solve the path planning problem for the moving sensors in a dynamic environment. To test and validate AMASS, we developed a series of computational experiments for fire search, and we assessed the performance against three metrics: (1) anomaly detection time lag, (2) source location uncertainty, and (3) state estimation error. The results demonstrate that: AMASS provides 10~15 times better performance than the traditional fixed sensor monitoring and detection strategy; ventilation in the monitored environment affects the performance by 6~40 times for any monitoring architecture with fixed or moving sensors; the monitoring performance cannot be fully reflected in a monolithic, single metric, but should include different metrics for the timeliness and spatial resolution of the detection function.

Passive seismic full waveform inversion using reconstructed body-waves for subsurface velocity construction

ABSTRACTFull waveform inversion (FWI) using passive seismic data can use amplitude, phase and travel time information from the data simultaneously. However, at least three challenges are involved in passive seismic full waveform inversion (PSFWI): a low signal-to-noise ratio (SNR), source location uncertainty and an unknown source wavelet. In this study, we propose a method that combines seismic interferometry and a source-independent inversion algorithm to solve these problems. Using seismic interferometry, the original passive seismic data recorded on the surface can be reconstructed into new virtual source records that have a relatively high SNR and certain source location. The source-independent algorithm eliminates the influence of source wavelet error on the final inversion results. Through numerical tests, we discuss the effects of passive source number and recording time on the inversion results and find that increasing the source number or recording time can improve inversion quality. We extract the background velocity model from the results of PSFWI and use it as the initial model of active source FWI. Least square reverse time migration (LSRTM) is then conducted to verify the accuracy of the inverted velocity models. The final results demonstrate that our PSFWI method can construct accurate long-wavelength velocity structures for subsequent active source FWI. The velocity model constructed using our successive inversion strategy can improve the LSRTM results.

Speech Understanding in Noise for Adults With Cochlear Implants: Effects of Hearing Configuration, Source Location Certainty, and Head Movement.

The primary purpose of this study was to assess speech understanding in quiet and in diffuse noise for adult cochlear implant (CI) recipients utilizing bimodal hearing or bilateral CIs. Our primary hypothesis was that bilateral CI recipients would demonstrate less effect of source azimuth in the bilateral CI condition due to symmetric interaural head shadow. Sentence recognition was assessed for adult bilateral (n = 25) CI users and bimodal listeners (n = 12) in three conditions: (1) source location certainty regarding fixed target azimuth, (2) source location uncertainty regarding roving target azimuth, and (3) Condition 2 repeated, allowing listeners to turn their heads, as needed. (a) Bilateral CI users exhibited relatively similar performance regardless of source azimuth in the bilateral CI condition; (b) bimodal listeners exhibited higher performance for speech directed to the better hearing ear even in the bimodal condition; (c) the unilateral, better ear condition yielded higher performance for speech presented to the better ear versus speech to the front or to the poorer ear; (d) source location certainty did not affect speech understanding performance; and (e) head turns did not improve performance. The results confirmed our hypothesis that bilateral CI users exhibited less effect of source azimuth than bimodal listeners. That is, they exhibited similar performance for speech recognition irrespective of source azimuth, whereas bimodal listeners exhibited significantly poorer performance with speech originating from the poorer hearing ear (typically the nonimplanted ear). Bilateral CI users overcame ear and source location effects observed for the bimodal listeners. Bilateral CI users have access to head shadow on both sides, whereas bimodal listeners generally have interaural asymmetry in both speech understanding and audible bandwidth limiting the head shadow benefit obtained from the poorer ear (generally the nonimplanted ear). In summary, we found that, in conditions with source location uncertainty and increased ecological validity, bilateral CI performance was superior to bimodal listening.

Spatial and Temporal Variation of Seismic Attenuation During Hydraulic Fracturing of a Sandstone Block Subjected to Triaxial Stress

AbstractQ factors estimated with spectral ratios describe the attenuation experienced by seismic arrivals in their travel from the source(s) to the receiver(s). When Q is not the same along the travel path of each arrival, the formulation must incorporate two Q factors making the problem indeterminate. Here we show that by assuming that Q is a random variable, the parameters of its probability distribution can be recovered through a mapping workflow using the spectral ratios. An advantage of such a workflow is that uncertainties in source location and velocity model can also be incorporated by representing these elements as random variables. We apply such a methodology to an acoustic emission (AE) data set from a laboratory experiment of hydraulic fracturing in a Colton sandstone block. We assume that the probability distribution of Q is normal. Anisotropic mean (μQ) values correlate with the triaxial stress conditions imposed at the boundaries of the block, with higher stress corresponding with less attenuation. The μQ values estimated within time intervals in the direction perpendicular to the hydraulic fracture plane correlate with stress buildup at the borehole during fracture propagation and borehole pressure relaxation or drop during fracture closure. The standard deviations (σQ) reflect the attenuation heterogeneity within the sampled regions. An increase in heterogeneity is correlated with pressure buildup at the borehole, which we interpret to reflect a dynamic stress gradient formed from the hydraulic fracture toward the surroundings.

Optimal TDOA Sensor-Pair Placement With Uncertainty in Source Location

Source localization based on time-difference-of-arrival (TDOA) measurements from spatially separated sensors is an important problem in sensor networks. In this paper, we study the optimal geometries of a team of networked sensors for the TDOA-based localization problem, taking into account the a priori uncertainty of the source location. Two types of TDOA sensor pairing methods, namely, centralized and decentralized , are considered. Analytic solutions to the optimal sensor geometries in the ideal case are derived in both static and movable source cases. In the movable source case, sensor geometry design is investigated under an extended Kalman filter (EKF) framework. Furthermore, under sensor motion and communication constraints, we extend this paper to an optimal sensor path planning problem. Simulations demonstrate its enhancement to the source localization and tracking performance.

Multipath propagation of low‐frequency radio waves inferred from high‐resolution array analysis

AbstractThe low‐frequency radio sky shows the locations of electromagnetic radio sources with a characteristic dilution of precision. Here we report a thorough high‐resolution analysis of radio waves from low‐frequency (∼20–150 kHz) radio communication transmitters which are recorded with a small aperture array of radio receivers during the day. It is found that the observed dilution of precision results from the array geometry of the radio receivers, a birefringent wave propagation, and the correlated multipath propagation of low‐frequency radio waves. The influence of the array geometry on the dilution of precision is reduced by taking into account the impulse response of the array. This procedure reveals for the very first time the splitting of one single radio source into two distinct source locations separated by ∼0.2°–1.9° which result from a birefringent wave propagation. The two locations are yet more clearly identified by using the polarity of the modulated wave number vectors of the radio waves. This polarity is also used to quantify the dilution of precision arising from correlated multipath propagation which is discriminated against wave number fluctuations arising from the timing accuracy of the radio receivers. It is found that ∼69% of the wave number variability is of natural origin and ∼31% originates from the timing accuracy of the receivers. The wave number variability from correlated multipath propagation results in a standard deviation ∼2–8% relative to the source location. This compact measurement of correlated multipath propagation is used to characterize the uncertainty of source locations in the radio sky. The identification of correlated multipath propagation strongly suggests the existence of very fast processes acting on time scales <1 ms in the D region ionosphere with physically meaningful effects on low‐frequency radio wave propagation. This important result has implications for practical applications in that the observed multipath propagation enables the determination of natural limits for the accuracy of navigation and lightning location methods using low‐frequency radio waves.

Information-theoretic analysis of iterated Bayesian acoustic source localization in a static ocean waveguide.

Fundamental constructs of information theory are applied to quantify the performance of iterated (sequential) Bayesian localization of a time-harmonic source in a range- and time-invariant acoustic waveguide using the segmented Fourier transforms of the received pressure time series. The nonlinear relation, defined by acoustic propagation, between the source location and the received narrowband spectral components is treated as a nonlinear communication channel. The performance analysis includes mismatch between the acoustic channel and the model channel on which the Bayesian inference is based. Source location uncertainty is quantified by the posterior probability density of source location, by the posterior entropy and associated uncertainty area, by the information gain (relative entropy) at each iteration, and by large-ensemble limits of these quantities. A computational example for a vertical receiver array in a shallow-water waveguide is presented with acoustic propagation represented by normal modes and ambient noise represented by a Kuperman-Ingenito model. Performance degradation due to noise-model mismatch is quantified in an example. Potential extensions to uncertain and stochastic environments are discussed.

Locating microseismic events using borehole data

ABSTRACTConstraining microseismic hypocentres in and around hydrocarbon reservoirs and their overburdens is essential for the monitoring of deformation related to hydraulic fracturing, production and injection and the assessment of reservoir security for CO2 and wastewater storage. Microseismic monitoring in hydrocarbon reservoirs can be achieved via a variety of surface and subsurface acquisition geometries. In this study we use data from a single, subsurface, vertical array of sensors. We test an existing technique that uses a 1D velocity model to constrain locations by minimizing differential S‐to‐P arrival times for individual sensors. We show that small errors in either arrival time picks or the velocity model can lead to large errors in depth, especially near velocity model discontinuities where events tend to cluster. To address this issue we develop two methods that use all available arrival times simultaneously in the inversion, thus maximizing the number of potential constraints from to N, where N is the number of phase picks. The first approach minimizes all available arrival time pairs whilst the second approach, the equal distance time (EDT) method defines the hypocentre as the point where the maximum number of arrival time surfaces intersect. We test and compare the new location procedures with locations using differential S‐to‐P times at each individual sensor on a microseismic data set recorded by a vertical array of sensors at the Ekofisk reservoir in the North Sea. Specifically, we test each procedure's sensitivity to perturbations in measured arrival times and the velocity model using Monte Carlo analysis. In general, location uncertainties increase with increasing raypath length. We show that errors in velocity model estimates are the most significant source of uncertainty in source location with these experiments. Our tests show that hypocentres determined by the new procedures are less sensitive to erroneous measurements and velocity model uncertainties thus reducing the potential for misinterpretation of the results.

Decision analysis of emergency ventilation and evacuation strategies against suddenly released contaminant indoors by considering the uncertainty of source locations

In case hazardous contaminants are suddenly released indoors, the prompt and proper emergency responses are critical to protect occupants. This paper aims to provide a framework for determining the optimal combination of ventilation and evacuation strategies by considering the uncertainty of source locations. The certainty of source locations is classified as complete certainty, incomplete certainty, and complete uncertainty to cover all the possible situations. According to this classification, three types of decision analysis models are presented. A new concept, efficiency factor of contaminant source (EFCS), is incorporated in these models to evaluate the payoffs of the ventilation and evacuation strategies. A procedure of decision-making based on these models is proposed and demonstrated by numerical studies of one hundred scenarios with ten ventilation modes, two evacuation modes, and five source locations. The results show that the models can be useful to direct the decision analysis of both the ventilation and evacuation strategies. In addition, the certainty of the source locations has an important effect on the outcomes of the decision-making.

Errors in direction finding of VLF waves caused by a sloping ionosphere

Lightning discharges emit very low frequency (VLF) radio waves commonly called “atmospherics.” In the past, lightning source locations were determined by the triangulation method from recordings at three widely separated stations in Japan. The three direction‐of‐arrival (DOA) lines, however, sometimes formed large triangles resulting in large uncertainties in source location. To approach this problem, the variational treatment is used here in which the sloping ionosphere plays an important role. By obtaining the time required to propagate from the source to the receiver via the Earth‐ionosphere waveguide, it is found that the direction of propagation of a VLF wave can be bent along a parabolic curve. Using this concept of parabolic propagation, the locations of lightning discharges are compared with those determined by the triangular method.

Bayesian statistics applied to the location of the source of explosions at Stromboli Volcano, Italy

Abstract We present a method for determining the location and spatial extent of the source of explosions at Stromboli Volcano, Italy, based on a Bayesian inversion of the slowness vector derived from frequency-slowness analyses of array data. The method searches for source locations that minimize the error between the expected and observed slowness vectors. For a given set of model parameters, the conditional probability density function of slowness vectors is approximated by a Gaussian distribution of expected errors. The method is tested with synthetics using a five-layer velocity model derived for the north flank of Stromboli and a smoothed velocity model derived from a power-law approximation of the layered structure. Application to data from Stromboli allows for a detailed examination of uncertainties in source location due to experimental errors and incomplete knowledge of the Earth model. Although the solutions are not constrained in the radial direction, excellent resolution is achieved in both transverse and depth directions. Under the assumption that the horizontal extent of the source does not exceed the crater dimension, the 90% confidence region in the estimate of the explosive source location corresponds to a small volume extending from a depth of about 100 m to a maximum depth of about 300 m beneath the active vents, with a maximum likelihood source region located in the 120- to 180-m-depth interval.

When enough is enough: The worth of monitoring data in aquifer remediation design

Given the high cost of data collection at groundwater contamination remediation sites, it is becoming increasingly important to make data collection as cost‐effective as possible. A Bayesian data worth framework is developed in an attempt to carry out this task for remediation programs in which a groundwater contaminant plume must be located and then hydraulically contained. The framework is applied to a hypothetical contamination problem where uncertainty in plume location and extent are caused by uncertainty in source location, source loading time, and aquifer heterogeneity. The goal is to find the optimum number and the best locations for a sequence of observation wells that minimize the expected cost of remediation plus sampling. Simplifying assumptions include steady state heads, advective transport, simple retardation, and remediation costs as a linear function of discharge rate. In the case here, an average of six observation wells was needed. Results indicate that this optimum number was particularly sensitive to the mean hydraulic conductivity. The optimum number was also sensitive to the variance of the hydraulic conductivity, annual discount rate, operating cost, and sample unit cost. It was relatively insensitive to the correlation length of hydraulic conductivity. For the case here, points of greatest uncertainty in plume presence were on average poor candidates for sample locations, and randomly located samples were not cost‐effective.

Emission characteristics and source location of the smooth Neptunian kilometric radiation

The observations of the planetary radio astronomy (PRA) experiment aboard Voyager 2 reveal the existence of smooth and bursty radio emission. Both recur in a rather regular pattern with a 16.1‐hour period (the Neptunian spin period). We describe the phenomenology of the smooth component in terms of frequency, polarization, and occurrence in magnetic longitude and latitude. The existence of both right‐handed and left‐handed polarized emissions is consistent with two sources (one in each hemisphere) which radiate independently in the TUX mode. Because Voyager passed Neptune at less than 5000 km from the surface at high northern magnetic latitudes, the radio sources were occulted by the planet near the encounter. We have taken advantage of this occultation to locate the northern hemisphere sources by calculating the radio horizon (based on the offset tilted dipole (OTD2) model) for two spacecraft positions close to the encounter. We find that the northern source is located at high magnetic latitudes δm > 40°. By using a geometrical beaming model which assumes emission in a hollow cone pattern we fit the observed PRA intensity profile. The best fit is obtained for a radio source at L=6, thus confirming δm > 40°. The longitudinal extent of the source is at least 180°, from −90° to +90° magnetic longitude. To locate the southern sources we use the large excursions of Voyager 2 to high southern magnetic latitudes. The source is found to be located also at high magnetic latitudes but is possibly more limited in longitude than the northern source. We estimate the uncertainty in source location due to the very limited knowledge of the magnetic field near the Neptunian surface, and we show that the current OTD2 model is satisfactory for the source locations for the lowest observed frequencies; however, an angular uncertainty of about 20° remains for sources in the northern hemisphere. The observed pattern of the smooth emission is strongly frequency dependent, which is in agreement with the azimuthal asymmetries of the magnetic field as predicted by the OTD2 model.

Optimal Array Processor Performance Trade-offs under Directional Uncertainty

Detection performance in terms of the receiver operating characteristic (ROC) curve is obtained for three specific problems involving either signal or noise source location uncertainty. The array processors are optimum in the sense that they form the likelihood ratio of the outputs of the individual array elements. Performance trade-offs due to signal uncertainty, signal and noise source location uncertainty, and array size are presented.

Prospecting for dead slabs

At subduction zones that have only recently ceased to be active, the lithospheric slab may retain a seismic velocity greater than that of the surrounding mantle even after the slab becomes seismically dead. To seek the subduction zone thought to have been recently active along the western margin of North America, we examined the variation with propagation direction of P-wave travel time residuals from sources at various distances and azimuths to seismograph stations in Washington and California. The uncertainty in source location and origin time was removed by referring the travel-time delay to a nearby station overlying presumably more uniform mantle. An eastward-dipping band of anomalously early arrivals at several stations on the western flank of the Sierra Nevada and California Cascades may imply that a dead slab is present beneath northern California, though a definitive conclusion is premature at present because of a paucity of seismic sources in eastern North America. The position of the dead slab speculatively suggested by the travel-time data is roughly consistent with that predicted by others on the basis of heat flow and geochemistry in the Sierra Nevada, and the southward decrease in the magnitude of the travel-time advance associated with such a slab is in agreement with the history of subduction of the Farallon plate as reconstructed from ocean floor magnetic anomalies and continental tectonic activity.