Surface Microseismic Monitoring Research Articles

Velocity model calibration for surface microseismic monitoring based on a 3D gently inclined layered equivalent model

Abstract Shale gas has become a major source of natural gas production and has received worldwide attention. Hydraulic fracturing is widely performed to stimulate oil and gas wells with considerable success. Given high-precision microseismic (MS) event locations, we can predict the development trend and region of fracturing and evaluate the stimulation effect, thereby providing technical guidance for subsequent exploitation. An accurate velocity model is essential for MS event positioning. However, simple velocity models, such as the uniform or vertical transverse isotropy (VTI) velocity model, are generally applied to calibrate the velocity model. Despite calibration, the VTI model may still face challenges in obtaining accurate MS event locations. Based on the structural characteristics of shale, we propose a novel local velocity model calibration algorithm for surface MS monitoring. To calibrate the velocity model, the actual strata interfaces are replaced with 3D gently inclined planes. We use very fast simulated annealing to concurrently tune the velocity, depth, and angle parameters of the model. Through the assessment of both the stacked amplitude at the position of the perforation shot and the relocation error of the perforation shot, we determine the ideal velocity model. To evaluate the effectiveness of our approach, we conduct experiments on both a synthetic model and a field dataset, and statistically analyze the location error. The results show that the proposed method obviously reduces the perforation shot relocation error and is well-suited for calibrating velocity models that are close to slightly inhomogeneous layered media.

Lighting Up a 1 km Fault near a Hydraulic Fracturing Well Using a Machine Learning-Based Picker

Abstract The development of portable nodal array in the recent years greatly improved the seismic monitoring ability across multiple scales. The dense arrays also directly benefit microseismic monitoring by providing relatively low-cost surface recordings. However, the rapid growth of seismic data is accompanied by the increased demand for efficient seismic phase picking. On the other hand, machine learning-based phase picking techniques achieved high stability and accuracy, showing promising potential to replace human labors and traditional automatic pickers. In this study, we applied a state-of-the-art package on newly collected nodal array data around a hydraulic fracturing well in southwestern China. The array consists of up to 85 nodes with an average station spacing of less than a kilometer. Within the hydraulic fracturing stimulation periods, we detected ∼3000 seismic events with magnitude down to ∼−2. After waveform cross-correlation-based relocation, the 1979 relocated events clearly light up a 1 km long fault structure and several fractures. Furthermore, the frequency–magnitude distribution of the catalog exhibits weak bilinear features with relatively low b-value (0.88) and a moderate coefficient of variation (Cv ∼2). The nature and origin of the observed earthquake cluster are then discussed and defined based on the industrial information, high-resolution earthquake catalog, and basic statistics. Finally, we summarized our experience and provided recommendations for applying similar approaches to other local scale, surface microseismic monitoring scenarios.

A Continuous Data Acquisition System for Three-Component Surface Microseismic Real-Time Monitoring

Surface microseismic monitoring is one of the most critical technologies for monitoring the dynamic subsurface reservoir processes. Due to the weak signals and poor signal-to-noise ratio (SNR), there are two major challenges to monitor the reservoir processes in an accurate and efficient way. One challenge is concerned with weak signals acquisition and how to evaluate the instrument’s abilities to record the microseismic events. The other one is how to realize real-time data monitoring during dynamic processes. Taking these challenges into consideration, a design approach to develop buried sensors consisting of three-component sensors and a low-cost, low-power consumption embedded device for microseismic monitoring data logger with the acquisition, synchronization, record, and transmission functions are presented. The circuits are divided into two printed circuit boards (PCBs). One is for a signal processing circuit, which consists of signal conditioning and analog-to-digital converter (ADC), and the other is for a digital circuit, which consists of the microcontroller and peripherals, including the global navigation satellite system (GNSS), Micro Secure Digital Memory Card (SD card), and Ethernet transceivers. The circuit for signal conditioning and ADC is a combination of a preamplifier, low-pass filter, programmable gain amplifier, and ADC. It has three modes: real-time monitoring, continuous recording, and combination. A software is designed to work in cooperation with the hardware to adjust the working parameters and modes. The self-noise of the data logger is tested in the laboratory, and the monitoring abilities of the data logger are evaluated. Finally, the surface microseismic monitoring system has been tested in the field to evaluate its performance for academic purposes.

First observation of paired microseismic signals during solution salt mining

To monitor the status of cavern roof instability induced by solution salt mining, we deployed a surface microseismic monitoring system for Dongxing salt mine, Dingyuan county, China. The microseismic monitoring system consists of 11 three-component geophones installed in shallow holes of 5 m deep. From 7-month continuous monitoring data, we have detected a novel type of signal that consists of two events. The first event has a long duration (>15 s) and generally has higher frequencies at the beginning. The 2nd event has a shorter duration of ∼3 s with lower frequencies. The two events are separated by ∼30–90 s in time. From May 2017 to November 2017, there are a total of 88 distinct paired signals with two events in sequence. We propose the first event corresponds to a series of fracturing processes on the cavern roof, and microseismic signals associated with different fractures are mixed and overlapped to produce a mixed signal with a longer duration. Due to a series of fracturing processes, some zones of cavern roof become instable and some rock debris may fall from the roof and collide on the cavern floor. This collision can produce a short duration signal, which corresponds to the 2nd event of the paired signal. The further analysis of polarities of the first arrivals for the 2nd event further proves its collision origin. The time interval between two events is related to the time for the falling rock debris through the brine, which is controlled by the cavern height and various physical properties of the rock debris and brine. Through the detailed analysis of paired signals, we can have a better understanding of the cavity development status for solution salt mining.

Technology Focus: Reservoir Surveillance (September 2022)

When it comes to reservoir surveillance, subsurface professionals have a broad array of tools and techniques to understand what is happening in the reservoir. To ensure valuable information is gathered with surveillance, it’s important to first ask “what problem are we trying to solve?” Depending on the problem’s definition and scale, the appropriate surveillance technology will change significantly. Zooming far out, particularly if little data is available for a reservoir, we may be interested in tools and techniques to validate and refine the geological concept for a basin, along with the most likely deposition patterns for sediments. Zooming in to the level of interactions between wells, we need to use a different toolbox that often includes finer-resolution geoscience mapping and engineering methods such as analyzing production, pressure, and fluid-property trends from individual wells. But what’s really exciting is that many new approaches are being developed to understand reservoir behavior and well performance at incredibly high resolution—at the fracture scale or even the pore scale. In this feature, we will examine surveillance technologies that can be applied at the broad scale, midscale, and fine scale. The papers highlighted in this month’s feature apply forward stratigraphic modeling to improve geological models, classical reservoir engineering techniques to understand interwell connectivity in order to optimize waterflooding operations, and exciting new fiber-optic technology to characterize the performance of individual hydraulic fractures in unconventionals. I hope these articles help you to think about “framing” the problem you are trying to solve, like a photographer carefully selecting the right camera settings and lens for a specific shot, in order to select the right tools and methods to add value through surveillance. Recommended additional reading at OnePetro: www.onepetro.org. SPE 202319 - Utilizing Surface Microseismic Monitoring To Improve Understanding of Natural Fractures in Sichuan Shale Gas Play by Cui Jing, Sichuan Changning Natural Gas Development Company, et al. SPE 202837 - Modified Technique To Model Volatile Oil Reservoirs: Implications for Modern Software Programs by Mohamed Ibrahim, Shell Egypt, et al. SPE 201543 - Production Optimization Using a 24/7 Distributed Fiber-Optic DFO Sensing-Based Multiphase Inflow Profiling Capability by Teymur Sadigov, BP, et al.

Deep Learning for Efficient Microseismic Location Using Source Migration‐Based Imaging

AbstractMigration‐based location methods (e.g., time‐reverse imaging based on wave equation, Kirchhoff summation, and diffraction stacking) can effectively locate events of low signal‐to‐noise ratios by stacking waveforms from many receivers. The methods have been widely applied for surface microseismic monitoring. However, these methods may not produce accurate results if there are polarity reversals in the surface records for a double‐couple or even a general moment tensor event. Various imaging conditions have been developed to solve the non‐focus image problem for a non‐explosive source. Here, we propose a deep convolutional neural network to predict a better‐focused image from a regular migration image that contains a quasi‐symmetric pattern in both space and time. To train the network, we first simulate a large number of surface records from sources with various locations and mechanisms. We then compute diffraction stacking images from the records and take the images as the input to the network. We define the corresponding training labels as images (with the same size as the input) with Gaussian distributions centered at the true sources. This network, trained by only synthetic datasets, works well in field data to detect source locations from images for unknown events. Both synthetic tests and field data applications demonstrate that the proposed method can effectively improve diffraction stacking images for efficient microseismic location.

Rapidly Locating a Water–inrush Collapse Column in a Seam Floor: A Case Study

After a mine was flooded by water inrush, a suspected collapse column was speculated based on 3D seismic data, but was not found by drill holes. Seismic source positioning based on different signal arrival times was used to determine the position of the water-inrush point, permitting rapid and accurate grouting and mine restoration. The impact of the water on the roadway roof was used as a continuous and dynamic secondary seismic source, along with surface microseismic monitoring of microseismic events near the inrush collapse column, and two suspected inrush points were rapidly located, which provided the information needed for layout of surface directional detection boreholes. The drilling results indicated that one of the suspected inrush points was a collapse column; its positon had a planar error of 26 m compared to the collapse column controlled by drilling. The inrush point was located 77 m below the seam floor, and the size of the collapse column was 19 m × 7 m. Thus, microseismic monitoring holds promise for rapidly targeting grouting at similar sites.

Detection, location, and source mechanism determination with large noise variations in surface microseismic monitoring

Microseismic monitoring aims at detecting as weak events as possible and providing reliable locations and source mechanisms for these events. Surface monitoring arrays suffer from significant variations of noise levels across receiver lines. When using a large monitoring array, we use a stacking technique to detect microseismic events through maximizing the signal-to-noise ratio (S/N) of the stack. But some receivers with a high noise level do not contribute to improving the S/N of the stack. We have derived a theoretical concept for the proper selection of receivers that best contribute to the stack for a constant strength of a signal across the array. This receiver selection criterion, based on the assumption of constant signal amplitude, provides a robust estimate of the noise threshold level, which could be used to discard or suppress contribution from the receivers that do not improve the S/N of the stack. We found that limiting the number of receivers for stacking improves the location accuracy and reduces the computational cost of data processing. Although the assumption of a constant signal never holds in real-life seismic applications, the noise level varies across the surface receivers in a significantly wider range than the signal amplitude. These noise variations can also increase the uncertainty of the source mechanism inversion and should be accounted for. Synthetic and field data examples show that weighted least-squares inversion with receiver weighting according to the noise level produces more accurate estimates for source mechanisms compared to the inversion that ignores information about noise.

Locating microseismic sources based upon L-shaped single-component geophone array: A synthetic study

We have developed a type of L-shaped single-component geophone array as a single station (L-array station) for surface microseismic monitoring. The L-array station consists of two orthogonal sensor arrays, each being a linear array of single-component sensors. L-array stations can be used to accurately estimate the polarization of first arrivals without amplitude picking. In a synthetic example, we first use segmentally iterative ray tracing (SIRT) method and forward model to calculate the travel time and polarization of first arrivals at a set of L-array stations. Then, for each L-array station, the relative delay times of first arrivals along sensor arrays are used to estimate the polarization vector. The small errors in estimated polarization vectors show the reliability and robustness of polarization estimation based on L-array stations. We then use reverse-time ray-tracing (RTRT) method to locate the source position based on estimated polarizations at a set of L-array stations. Very small errors in inverted source location and origin time indicate the great potential of L-array stations for source localization applications in surface microseismic monitoring.

Research on detector signal receiving network layout optimization model

Fracturing surface micro-seismic monitoring technology is widely used in fracturing surface. Micro- seismic monitoring conducted by means of the internet is composed of a large number of geophones laid on the ground to collect subsurface micro seismic signals and monitor fracturing fractures. With the advantages of a large amount of monitoring data, flexible layout, easy adjustment, and low cost. The technology also has some disadvantages, such as weak signal reception, vulnerable to environmental impact and low signal-to-noise ratio. In order to improve the positioning accuracy and receive as many effective signals as possible, a signal receiving network model of optimal geophone for micro-seismic monitoring on a fractured surface based on improved genetic algorithm is proposed. Through simulation and numerical analysis, the solution model has been optimized to meet the accuracy and quickly solve the optimal geophone array scheme. The results show that the optimal geophone array scheme not only satisfies the positioning accuracy but also achieves better reception of micro-seismic signals as possible with a small number of geophones, which reduces the cost and promotes the industrial application of this technology.

Regularization to Stein unbiased risk estimation denoising and threshold for first-break picking to low signal-to-noise ratio of surface microseismic monitoring — Case of gas shale fracturing

With monitoring of the acoustic emission phenomenon caused by rock deformation and failure, microseismic monitoring has been widely used in the development of unconventional oil and gas fields. Due to the complex environment and diversity types of the noise, the signal energy of surface microseismic monitoring is weak and the signal-to-noise ratio (S/N) of raw data is very low. In the process of data processing, many human resources are needed to discriminate the first-break picking because of the low S/N, and this directly affects the error of microseismic event location. We have adopted the regularization to Stein unbiased risk estimation (R-SURE) algorithm based on the continuous wavelet transform to separate the signal from the noise in different decomposition levels. The regularization factor is the adaptive change of the different geology and fracturing engineering, which is related to shale brittleness, fracturing pressure, and displacement. As a result, the threshold from the R-SURE algorithm is multiresolution in different levels, and the S/N could be improved effectively. In addition, we established the threshold discriminant for picking up the first-break wave of low-S/N data combined with the Akaike information criterion and characteristic function, which compared the maximum absolute value in the time window. The method has good robustness and low computational complexity. The first arrival is automatically and accurately judged, which improves the accuracy of the event location. We successfully applied these methods to the surface microseismic monitoring of shale gas fracturing in several wells in southwest China. The S/N of the raw data has been improved, the effective stimulated reservoir volume and the performance of the gas production are predicted with the results, which provides important technical support for shale gas development in the area.

Comparison of Single-Trace and Multiple-Trace Polarity Determination for Surface Microseismic Data Using Deep Learning

AbstractFor surface microseismic monitoring, determination of the P-wave first-motion polarity is important because (1) it has been widely used to determine focal mechanisms and (2) the location accuracy of the diffraction-stack-based method is improved greatly using polarization correction. The convolutional neural network (CNN) is a form of deep learning algorithm that can be applied to predict the polarity of a seismogram automatically. However, the existing network designed for polarity detection utilizes only individual trace information. In this study, we design a multitrace-based CNN (MT-CNN) architecture using several neighbor traces combined as training samples, which could utilize the polarity information of neighbor sensors in the surface microseismic array. We use 17,227 field seismograms with labeled polarities to train two different neural networks that predict the polarities by a single trace or by multiple traces. The performance of the test set and field example of two CNN architectures shows that the MT-CNN significantly produces fewer polarity prediction errors and leads to more accurate focal mechanism solutions for microseismic events.

Fast-AIC Method for Automatic First Arrivals Picking of Microseismic Event With Multitrace Energy Stacking Envelope Summation

As the signal-to-noise ratio (SNR) of surface microseismic monitoring data is generally low and large, traditional detection and picking algorithms cannot satisfy the real-time and high accuracy to processing. Therefore, a Fast Akaike information criterion (Fast-AIC) algorithm is proposed for microseismic event automatic detection and first arrival time picking. First, an automatic detection method of microseismic events based on multitrace energy stacking is proposed, to avoid the missed detection and false detection in conventional automatic detection methods. Second, the Fast-AIC algorithm is developed by mathematical derivation from the Vector Auto-regressive (VAR-AIC algorithm), to improve the efficiency of first arrival time picking of microseismic signals. Finally, the new method and three other conventional first arrival time picking methods are tested on microseismic monitoring data from a hydraulic fracture site in Shanxi, China. We have found that the new method has the highest picking accuracy and computational efficiency.

An integrated seismic study of a shale reservoir in the Sichuan Basin, onshore China

We describe an integrated seismic evaluation for the Longmaxi shale gas potential in the central part of the Sichuan Basin, onshore China, which benefits from innovations in high-density 3D seismic, borehole-based vertical seismic profiling (VSP), and passive seismic recording for microseismic monitoring. The geology of the basin has a complex history that creates several geophysical challenges, including strong multiple contamination and the requirement for high-resolution data to enable detailed structural and stratigraphic mapping. For the 3D surveys, we adopted a new wide-azimuth and high-density acquisition design. A large part of the processing was devoted to noise removal, optimizing surface-consistent processes including deconvolution, and thorough testing of data-driven internal multiple attenuation. Prestack depth imaging is important for accurate geosteering of long horizontal trajectories within a narrow target window, and our resulting images in depth provide a more planar target structure than with (more routine) time imaging. The more planar true structure was confirmed by drilling. For the borehole seismic calibration, vibroseis was used as the VSP source in preference to the standard use of explosives. This resulted in greater consistency of the waveforms and better amplitude and timing of events with which to constrain the 3D seismic processing and imaging. Surface microseismic monitoring of the stimulation was achieved using a new and bespoke design. The detected microseismic activity helped support and optimize the real-time fracture operations, mitigate casing deformation, obtain in-situ stress fields, and understand the control of preexisting geologic features. The innovations in surface seismic, VSP, and microseismic, along with the integration of other disciplines, contributed to the project's overall success.

Research Note: Frequency domain orthogonal projection filtration of surface microseismic monitoring data

ABSTRACTWe address the problem of increasing the signal‐to‐noise ratio during surface microseismic monitoring data processing. Interference from different seismic waves causes misleading results of microseismic event locations. Ground‐roll suppression is particularly necessary. The standard noise suppression techniques assume regular and dense acquisition geometries. Many pre‐processing noise suppression algorithms are designed for special types of noise or interference. To overcome these problems, we propose a novel general‐purpose filtration method. The goal of this method is to amplify only the seismic waves that are excited in the selected target area and suppress all other signals. We construct a linear projector onto a frequency domain data subspace, which corresponds to the seismic emission of the target area. The novel filtration method can be considered an extension of the standard frequency–wavenumber flat wave filtration method for non‐flat waves and arbitrary irregular receiver‐position geometries. To reduce the effect of the uncertainty of the velocity model, we suggest using additional active shot data (typically the perforation shots), which provide static travel time corrections for the target area. The promising prospects of the proposed method are confirmed by synthetic and semi‐synthetic data processing.

Relative source location using a local equivalent path strategy for surface microseismic monitoring

Relative source location using a local equivalent path strategy for surface microseismic monitoring

Ambient noise tomography of three-dimensional near-surface shear-wave velocity structure around the hydraulic fracturing site using surface microseismic monitoring array

Surface microseismic monitoring array has been widely used to monitor hydraulic fracturing process for developing unconventional oil/gas reservoirs. Here we take advantage of relatively dense surface seismic stations for monitoring the fracturing process of a shale gas reservoir in southwest China, to determine the shallow Vs structure around the fracking site using the ambient noise tomography. This is because having a detailed three-dimensional (3D) Vs model is important for 3D seismic structure inversion. For ambient noise tomography, we use 21 days of continuous ambient noise data to extract Rayleigh wave group and phase dispersion data in the period band of 0.1–0.5 s. The 3D Vs structure is inverted by the direct surface wave tomographic method that directly inverts surface wave dispersion data for 3D Vs model. With the spatial coverage of 4 km by 4 km for this surface seismic array, we can determine the Vs structure down to 400 m. Our inversion results show that shear wave velocity in this region can be up to about 3 km/s above 400 m depth, which is consistent with the lithology of drilling core along the well. The obtained model is further validated by checkerboard test and the close relationship between waveform amplitude and the local velocity structure around stations. It is found that large waveform amplitudes are associated with stations located at shallow low velocity anomalies due to surface reverberations. This application shows that ambient noise tomography using local dense seismic array can be used to reliably determine high-resolution 3D shear velocity structure in the near surface.

Finite-difference modelling of microseismicity associated with a hydraulic-fracturing stimulation in a coalbed methane reservoir

Owing to the large capacity for matrix storage, substantial volumes of methane gas are stored in coal beds in the form of adsorbed gas (Nuccio, 2000). According to Langmuir’s theory that relates the adsorbed gas volume to the reservoir pressure (Langmuir, 1918), a fast dewatering process is required to achieve rapid reduction of the reservoir pressure and thus to release the adsorbed methane gas (Anderson and Simpson, 2003). The effectiveness of this dewatering process depends on the coal permeability, which is mostly related to the aperture and spacing of its natural fractures (or cleats). Coal permeability is highly stress-sensitive. For example, a coal sample with a cleat aperture of 10 μm and a permeability of 1000 millidarcy at surface conditions (no in-situ stress) can be reduced to a cleat aperture of less than 1 μm and a permeability of less than 1 millidarcy at reservoir conditions, assuming 50% lithostatic pressure and a mean pore pressure of 1 MPa (Weniger et al., 2016). This low permeability of coal beds due to fracture closure makes hydraulic fracturing necessary, to increase the fracture frequency and width in order to achieve economic gas production rates; for this reason, hydraulic fracturing has been implemented for stimulation of coalbed methane reservoirs for decades (e.g. Holditch et al., 1988). Microseismic monitoring is an effective method for assessing hydraulic-fracturing treatments and has proven to be a key technology to image the propagation of hydraulic fractures within a reservoir (Maxwell, 2014). Previous microseismic studies in coal-bearing reservoirs, however, have encountered challenges because of the complex velocity model of low-velocity, low-density coal beds interbedded within more rigid clastic units. Consequently, finite-difference (FD) modelling has proven to be a very useful tool to understand complex seismic propagation and interference patterns (Pike, 2014). This case study focuses on the analysis of microseismic data from a CBM reservoir located close to the northern limit of Cesar-Rancheria Basin, Colombia. The microseismic data was acquired at the surface using two perpendicular lines of vertical-component geophones (Figure 1). Clear P-wave arrivals with reverberating waveform character are evident, but no clear S-wave arrivals are discernible in the data. A 2D-FD waveform modelling approach was used to calibrate a velocity model based on data from the reservoir (well logs and seismic-reflection data) and on the recorded microseismic data from the hydraulic-fracturing treatment. A 1D isotropic velocity model was implemented for this case study, as this degree of approximation has been reported to be sufficiently accurate for microseismic event location in similar case studies of surface microseismic monitoring (e.g. Neuhaus et al., 2012). Numerical simulation using a 2D-FD approach is considerably less time-consuming when compared with full 3D-FD modelling. Moreover, the linear alignment of the surface monitoring array, together with the implementation of a 1D isotropic velocity model, makes the implementation of 2D-FD modelling feasible for calibrating the velocity model.

Will your surface microseismic hydraulic fracture monitoring project be successful?

The benefits of microseismic monitoring of hydraulic fracturing are well understood, and continue to grow as acquisition, processing, and interpretation methods improve. Most surveys are conducted using downhole receivers placed close to the reservoir zone. However, for some surveys, a surface monitoring operation can produce good results and provide certain advantages compared to a downhole survey including: • No need for an observation well. • No problems with high reservoir temperatures. • Typically more accurate spatial locations (x,y) of events. • No spatial bias (or distance-to-observation-well bias), which is especially important for large multi-well projects. • Better cost scalability to multi-well projects, and probably lower overall costs for multi-well projects. • More robust source mechanism determination, especially compared to single observation well downhole monitoring. The disadvantages of a surface survey compared to a downhole survey include: • Less precise depth locations of events. • Typically higher costs for a single well project. • Lower sensitivity, and higher risk of failure. The risk of failure, especially in a new area or a new formation, probably deters people from using surface monitoring. Microseismic geophysicists who work for operators describe their discomfort in being unable to answer the question when asked by their management, ‘Will a surface project be successful?’ We present a summary of our experiences with successful and failed projects, we establish a set of metrics for the purposes of comparison, and we provide guidelines for planning successful projects.

Velocity model optimization for surface microseismic monitoring via amplitude stacking

A usable velocity model in microseismic projects plays a crucial role in achieving statistically reliable microseismic event locations. Existing methods for velocity model optimization rely mainly on picking arrival times at individual receivers. However, for microseismic monitoring with surface stations, seismograms of perforation shots have such low signal-to-noise ratios (S/N) that they do not yield sufficiently reliable picks. In this study, we develop a framework for constructing a 1-D flat-layered a priori velocity model using a non-linear optimization technique based on amplitude stacking. The energy focusing of the perforation shot is improved thanks to very fast simulated annealing (VFSA), and the accuracies of shot relocations are used to evaluate whether the resultant velocity model can be used for microseismic event location. Our method also includes a conventional migration-based location technique that utilizes successive grid subdivisions to improve computational efficiency and source location accuracy. Because unreasonable a priori velocity model information and interference due to additive noise are the major contributors to inaccuracies in perforation shot locations, we use velocity model optimization as a compensation scheme. Using synthetic tests, we show that accurate locations of perforation shots can be recovered to within 2m, even with pre-stack S/N ratios as low as 0.1 at individual receivers. By applying the technique to a coal-bed gas reservoir in Western China, we demonstrate that perforation shot location can be recovered to within the tolerance of the well tip location.